UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Holocene sedimentary history of Chilliwack Valley, Northern Cascade Mountains Tunnicliffe, Jon Francis 2008

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2008_spring_tunnicliffe_jon.pdf [ 18.35MB ]
Metadata
JSON: 24-1.0066305.json
JSON-LD: 24-1.0066305-ld.json
RDF/XML (Pretty): 24-1.0066305-rdf.xml
RDF/JSON: 24-1.0066305-rdf.json
Turtle: 24-1.0066305-turtle.txt
N-Triples: 24-1.0066305-rdf-ntriples.txt
Original Record: 24-1.0066305-source.json
Full Text
24-1.0066305-fulltext.txt
Citation
24-1.0066305.ris

Full Text

Holocene Sedimentary History of Chilliwack Valley, Northern Cascade Mountains by Jon Francis Tunnicliffe B.A. (Hons), University of Western Ontario, 1995 M.Sc, University of Northern British Columbia, 2000 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF Doctor of Philosophy in The Faculty of Graduate Studies (Geography)  The University Of British Columbia (Vancouver, Canada) January, 2008 c Jon Francis Tunnicliffe 2008  Abstract  I seek to reconstruct the balance between sediment storage and yield across multiple drainage basin scales in a large (1 230 km2 ) watershed in the Northern Cascade range, British Columbia and Washington. Chilliwack Valley and surrounding area has been the site of numerous studies that have detailed much of its Quaternary sedimentary history. In the present study this information is supplemented by reconstruction of the morphodynamic trajectory of the river valley though the Holocene Epoch, and development of a sediment transfer model that describes the relaxation from the Fraser glaciation. The total Holocene sediment yield is estimated from basins across several scales using field and remotely sensed evidence to constrain the historical mass balance of delivery to higher order tributary basins. Rates of hillslope erosion are estimated using a diffusion-based relation for open slopes and delimitating the volume evacuated from major gully sources. Digital terrain models of paleo-surfaces are constructed to calculate total sediment erosion and deposition from tributary valleys and the mainstem. Chilliwack Lake has effectively trapped the entire post-glacial sediment load from the upper catchment (area = 334 km2 ), allowing to compare this ‘nested’ system with the larger catchment. Rates of lake sediment accumulation are estimated using sediment cores and paleomagnetism. These are compared with accumulation rates in the terminal fan inferred from radiocarbon dating of fossil material, obtained by sonic drilling in the apex gravels. A sediment budget framework is then used to summarize the net transfer of weathered material and glacial sediments from the hillslope scale to the mainstem. The long-term average sediment yield from the upper basin is 62 ± 9 t/km2 /yr; contemporary yield is approximately 30 t/km2 /yr. It is found that only 10-15% of the material eroded from the hillslopes is delivered to mouths of the major tributaries; the remaining material is stored at the base of footslopes and within the fluvial sedimentary system. Since the retreat of Fraser Ice from the mouth of the valley, Chilliwack River delivered over 1.8 ± 0.21 km3 of gravel  ii  Abstract and sand to Vedder Fan in the Fraser Valley. In the sediment budget developed here, roughly 85% of that material is attributed to glacial sources, notably the Ryder Uplands and glacial valley fills deposited along the mainstem, upstream of Tamihi Creek. In tributary valleys, local base-level has fallen, leading to the evacuation of deep glacial sedimentary fills. Many of the lower reaches of major tributaries in upper Chilliwack Valley (e.g. Centre and Nesakwatch Creeks) remain primarily sediment sinks for slope-derived inputs, since base-level fall has not been initiated. In distal tributaries (Liumchen, Tamihi and Slesse creeks), paraglacial fans have been incised or completely eroded, entrained by laterally active channels. A transition from transport-limited to supply-limited conditions has been effected in many of these reaches. Slesse Creek has struck an intermediate balance, as it continues to remobilize its considerable sediment stores. It functions today as the sedimentary headwaters of Chilliwack Valley. Using grain size data and fine-sediment geochemical data gathered from Chilliwack River over the course of several field seasons, a simple finite-difference, surface-based sediment transport model is proposed. The aim of the model is to integrate the sediment-balance information, as inferred from estimates of hillslope erosion and valley storage, and physical principles of sediment transport dynamics to reproduce the key characteristics of a system undergoing base-level fall and reworking its considerable valley fill during degradation. Such characteristics include the river long profile, the river grain-size fining gradient, the percentage of substrate sand, and the diminution of headwater granite lithology in the active load. The model is able to reproduce many of the characteristics, but is not able to satisfy all criteria simultaneously. There is inevitably some ambiguity as to the set of parameters that produce the “right” result, however the model provides good insight into long-term interactions among parameters such as dominant discharge, grain size specifications, abrasion rates, initial topography, hiding functions, and hydraulic parameters.  iii  Table of Contents  Abstract  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ii  Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  iv  List of Tables  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  viii  List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  x  Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxvi 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  1  1.1  Problem Statement  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  1  1.2  The Study Basin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  5  1.2.1  Physiography  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  8  1.2.2  Regional Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  9  Thesis Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  11  2 Hillslope and Tributary Sediment Stores . . . . . . . . . . . . . . . . . . . .  13  1.3  2.1  Introduction  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  13  2.2  Data Sources and Associated Errors . . . . . . . . . . . . . . . . . . . . . . .  15  2.2.1  15  The Magnitude of Error  . . . . . . . . . . . . . . . . . . . . . . . . .  2.3  Network Structure and Process Domains  . . . . . . . . . . . . . . . . . . . .  2.4  Sediment Deposition in Lower-Order Catchments  2.5  Sediment Source Areas  17  . . . . . . . . . . . . . . .  21  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  26  2.5.1  Large Bedrock Failures  . . . . . . . . . . . . . . . . . . . . . . . . . .  2.5.2  Gullies and Diffusive Slope Processes  2.5.3  Sediment Source Areas: Surficial Materials and Gullied Terrain  . . . . . . . . . . . . . . . . . . . . .  26 27 28 iv  Table of Contents 2.5.4  . . . . . . . . . . . . . . . . . .  33  The Fluvial Domain: The Lower Tributary Valleys . . . . . . . . . . . . . . .  38  2.6.1  . . . . . . . . . . . . . . . . . . . . . . .  41  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  43  3 Chilliwack Lake . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  46  2.6 2.7  Sediment Source Areas: Open Slopes Lower Tributary Valley Fills  Discussion  3.1  Study Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  47  3.2  Seismic Methodology  49  3.3  Interpretation of the seismic record  3.4  Fan Deltas  3.5  Ground Penetrating Radar Surveys  . . . . . . . . . . . . . . . . . . . . . . .  56  3.6  Lake Cores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  61  3.6.1  Lake Core Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . .  62  3.6.2  Tephra and Other Disturbance Layers . . . . . . . . . . . . . . . . . .  63  3.6.3  Magnetic Parameters  . . . . . . . . . . . . . . . . . . . . . . . . . . .  64  3.6.4  Palaeomagnetism  . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  69  Rates of Sediment Accumulation in the Holocene Epoch . . . . . . . . . . . .  70  4 Evolution of Chilliwack Valley Mainstem . . . . . . . . . . . . . . . . . . . .  77  3.7  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  50  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  54  4.1  Initial Conditions  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  78  4.2  Mid Valley Fill . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  81  4.3  Glacio-Lacustrine Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . .  82  4.4  Lower Valley Fill . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  85  4.4.1  Ryder Lake Upland Moraine Complex . . . . . . . . . . . . . . . . . .  86  4.5  Vedder Fan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  90  4.6  Architecture of the Vedder Fan . . . . . . . . . . . . . . . . . . . . . . . . . .  92  4.7  Well-log database  95  4.8  Apex Gravels - Core Descriptions  4.9  Chronology and Volumetric Estimation 4.9.1  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Isopach Diagrams  4.10 Discussion  . . . . . . . . . . . . . . . . . . . . . . . .  Introduction  99  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105  5 Characterization of Valley Sediments 5.1  . . . . . . . . . . . . . . . . . . . . .  95  . . . . . . . . . . . . . . . . . . . . . . 108  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108  v  Table of Contents 5.2  Sampling of Tributary and Mainstem Gravels . . . . . . . . . . . . . . . . . . 109 5.2.1  5.3  5.5  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114  Lithology and Geochemistry 5.3.1  5.4  Fining Patterns  . . . . . . . . . . . . . . . . . . . . . . . . . . . 119  Coarse Clast Lithology  Silt Geochemistry  . . . . . . . . . . . . . . . . . . . . . . . . . . 119  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123  5.4.1  Methods  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123  5.4.2  Factor Analysis  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126  Summary and Conclusions  . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131  6 A Morphodynamic Model of Postglacial River Evolution 6.1  Introduction  6.2  One-dimensional representation  6.3  6.4  6.5  6.6  . . . . . . . . . 133  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 . . . . . . . . . . . . . . . . . . . . . . . . . 136  6.2.1  The Degrading River Valley System . . . . . . . . . . . . . . . . . . . 137  6.2.2  Bed Shear Stress Distribution  6.2.3  The Active Layer  . . . . . . . . . . . . . . . . . . . . . . 139  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140  Model Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 6.3.1  Grid Resolution and Hydraulics  . . . . . . . . . . . . . . . . . . . . . 141  6.3.2  Abrasion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144  6.3.3  Mass Balance  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145  Boundary Conditions  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146  6.4.1  Basin Hydrology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147  6.4.2  Stratigraphy and Bedrock  6.4.3  Upstream Feed and Tributary Inputs  6.4.4  Time and Intermittency . . . . . . . . . . . . . . . . . . . . . . . . . . 156  Model Performance  . . . . . . . . . . . . . . . . . . . . . . . . 150 . . . . . . . . . . . . . . . . . . 152  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158  6.5.1  Long-Profile Adjustments . . . . . . . . . . . . . . . . . . . . . . . . . 160  6.5.2  Textural Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164  6.5.3  Clast Fining and Abrasion  6.5.4  Subsurface Sand Content . . . . . . . . . . . . . . . . . . . . . . . . . 167  6.5.5  Growth of Vedder Fan  6.5.6  Suspended Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172  . . . . . . . . . . . . . . . . . . . . . . . . 165  . . . . . . . . . . . . . . . . . . . . . . . . . . 169  Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173  vi  Table of Contents 7 Conclusions 7.1  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178  Process Domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 7.1.1  Terminal Deposits of Chilliwack Valley  7.1.2  Mainstem Deposits  . . . . . . . . . . . . . . . . . 181  . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182  7.2  Textural Evolution of the Mainstem . . . . . . . . . . . . . . . . . . . . . . . 183  7.3  Models of the Holocene Fluvial System  . . . . . . . . . . . . . . . . . . . . . 184  Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186  vii  List of Tables  2.1  Estimated volumes eroded from gullied morainal and colluvial cover . . . . .  33  2.2  Estimated volumes eroded from planar to convex slopes . . . . . . . . . . . .  37  2.3  Volume of glacigenic fill evacuated from major tributaries. . . . . . . . . . . .  43  3.1  Volumetric estimates for catchment erosion and fan deltas bedload delivery since deglaciation, not including outwash stores. . . . . . . . . . . . . . . . . .  58  3.2  Radiocarbon ages from Chilliwack Lake. . . . . . . . . . . . . . . . . . . . . .  62  3.3  Sediment delivery to Chilliwack Lake: Volumetric estimates . . . . . . . . . .  74  3.4  Mineral sediment delivery to Chilliwack Lake: Late Holocene (<2 000 years BP) estimates based on core chronology. . . . . . . . . . . . . . . . . . . . . .  4.1  Net bulk volume eroded from the mainstem between Chilliwack Lake and Borden Creek . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  4.2  74  81  Estimated bulk erosion volumes for three different assumed topographic configurations in the Lower Chilliwack Valley in post-glacial time. The true value is assumed to be intermediate between I & II. The bounds of minimum net erosion are shown in Figure 4-8 . . . . . . . . . . . . . . . . . . . . . . . . . .  4.3  88  Radiocarbon ages from drilling at Vedder Fan. All samples were dated by conventional radiometric technique. Intercept ages and age range in calendar years before AD 1950. The age ranges in parentheses represent 1σ error limits. The ages were determined using the INTCAL98 database. . . . . . . . . . . .  4.4  Volumetric estimates of Vedder Fan composition  (m3  × 106 )  99  based on well log  records. Values are cumulative over the time intervals indicated. The overall expected accuracy of the estimates is ±16%, based on the range of possible spatial extents and depths of the fan. Values in brackets indicate additional volumes attributable to the Sumas Lake Basin. . . . . . . . . . . . . . . . . . 105  viii  List of Tables 4.5  Bed material yield below Vedder Crossing - lowest, mean and highest annual averages over the course of study periods for three investigations using different methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106  5.1  Correlation scores among 19 elements in fine grained (< 63µ) sediments (both active channel and hillslope sources) Chilliwack Valley. Values greater than 0.65 or less than -0.65 are highlighted. . . . . . . . . . . . . . . . . . . . . . . 125  5.2  Factor Scores from elements in the factor analysis using Varimax rotation. Principal loadings for each element are highlighted. . . . . . . . . . . . . . . . 129  5.3  Mixing estimates based on Fe-Factor and associated raw element variables. Values indicate the approximate proportional addition to the mainstem sediments at each tributary junction. +∞ scores indicate scores greater than one; −∞, less than zero. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130  6.1  Representative volumes of the eroded stratigraphic units in the lower mainstem. Volume is tabulated within 360 m cells, each having a specified width and vertical walls, and thus some spatial (volumetric) resolution is lost. Volumes for an alternate model configuration with larger accumulations within the lower valley are indicated in parentheses. . . . . . . . . . . . . . . . . . . 152  6.2  Summary of static and varied model parameters . . . . . . . . . . . . . . . . 159  ix  List of Figures  1-1 A spatio-temporal view of sediment transfer across the landscape, adapted from Church [1996]. The box indicates the range of scales that are addressed in this thesis - river distances of up to 100 km, and timescales ranging from 10 years of river measurements out to 14 ka cal. B.P.). The annotations in the diagram indicate appropriate modes of explanation across various spatiotemporal scales. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  3  1-2 Conceptual picture of the transit of a large-scale sedimentary disturbance through space and time (after Church [2002]). The disturbance decays quickly away from proximal watersheds. Dispersion is due to deposition of material at tributary junctions and in sedimentary reservoirs along the system, vegetation of formerly active sources, and differential mobility of grains in the mixture. The figure illustrates a “primary” disturbance of post-glacial time, upon which smaller-scale perturbations are super-imposed. . . . . . . . . . . . . . . . . .  5  1-3 Chilliwack Valley and surrounding area. . . . . . . . . . . . . . . . . . . . . .  7  1-4 Lithologic assemblages of the study area. . . . . . . . . . . . . . . . . . . . . .  10  2-1 Definition of the reference post-glacial surface. Material deposited on top of the post-glacial surface constitutes an input to the post-glacial budget, material eroded from the valley wall and glacial fill are considered to be outputs. The reference postglacial topography is V0 . . . . . . . . . . . . . . . . . . . . . . .  14  2-2 Chilliwack River Valley. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  18  2-3 Three different basin scales. The left side of (a) shows linear, first- and secondorder debris flow channels, next to a 3rd order alpine colluvial basin, in the headwaters of the Slesse basin. Grid spacing is 500 m. (b) 5th-order Bear Creek, with a developed fluvial network (grid spacing is 1 km). Note the relatively small outlet fan in the latter. . . . . . . . . . . . . . . . . . . . . . .  19 x  List of Figures 2-4 Histograms showing the distribution of upstream area within valley network links, from headmost channels to the Chilliwack River at Vedder Crossing. The mean area for each order class is shown with a cross (+). The process domains identified by Brardinoni and Hassan [2006] are overlaid in grey. The total number of links for each order is indicated. . . . . . . . . . . . . . . . .  20  2-5 Diagrams illustrating estimation of fan deposits using the technique of Campbell and Church [2003] and using CAD shapes. . . . . . . . . . . . . . . . . .  23  2-6 Estimated Holocene bulk sediment deposition (sand and coarser) plotted against contributing catchment area.  Catchments are coloured according to their  Strahler order. Two reference lines are shown, indicating that the envelope of maximum deposition scales approximately to the power of 1.5. A similar exponent is used for scaling the size of fans and deltas from larger fluvial basins. 24 2-7 Pierce Creek, July 2003. Source area of the 1996 event is shown with an arrow 25 2-8 Data from Kirchner et al. [2001] (black diamonds) with data from the Chilliwack Valley overlain (open circles). Sediment accumulation estimates from Chilliwack Valley display more scatter, but they are largely consistent with the long-term average rates of specific sediment yield calculated by Kirchner et al. [2001]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  26  2-9 Histogram of gradient distributions in gullied topography and on planar to convex hillslopes (vegetated and bedrock) in the first to fifth order catchments of Chilliwack Valley. A limiting gradient for the vegetated open slopes is estimated to be 1.4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  28  2-10 Photos of (a) Holocene channel incision and (b) exposed sidewall deposits of till and glacio-lacustrine material. (c) is a stratigraphic section from a left-bank tributary to Lower Slesse Creek. Total incision is over 25 m, but the exposed 10 m section gives some idea of the complex nature of the initial (Pleistocene) hillside stratigraphy. Some similar sections were found in nearby tributaries, but exposure of such a stratigraphic record is uncommon. . . . . . . . . . . .  30  2-11 Photo and figure of a landslide headscarp at Foley Creek. This is the initiation point of a debris flow gully, incised into a sandy, compacted till deposit. . . .  31  xi  List of Figures 2-12 The distribution of hillslope gradients within the gullied topography of Chilliwack Valley. Contours indicate the cumulative density of points, in increments of 0.2. The slopes at the initiation points for 492 landslides in the EBA database are overlaid. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  32  2-13 Hillslope relaxation response. Histograms indicate the distribution of DEM slopes that are found within the gullied zones. The modern (TRIM) DEM is shown in darker gray; the inferred post-glacial distribution is outlined in black. The shift in the curve describing the distribution of slopes indicates that the modern (TRIM) surface is left steeper after the morainal and colluvial material has been evacuated, due to establishment of steep gully sidewalls.  . . . . . .  34  2-14 Results from the subtraction of DEM surfaces in gullied zones, over the entire Chilliwack Valley. Graph shows the distribution of maximum depths of eroded volume. Linear scale for bar chart is shown on the left, log scale for the curve is on the right. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  34  2-15 Cumulative erosion from gullied hillslope sources in Chilliwack Valley by slope class. Maximum rates of erosion occur in the range of 35-40◦ . Mean depth of vertical erosion increases with slope. Depths of erosion are indicated with bounds of one standard deviation. . . . . . . . . . . . . . . . . . . . . . . . .  35  2-16 Volumetric estimates of volumetric erosion vs. fan volume (bedload) are compared in a number of basins within the Chilliwack Valley, to assess the agreement between estimates. Symbols are the same as in previous figure. . . . . .  37  2-17 (a) Combined specific erosion potential for bedrock, gullied and forested slopes, calculated for a proportion of all links in Chilliwack Valley (K1soil = 0.006, Sc = 1.4 for forested terrain, K1rock = 0.003, Sc = 2 for bedrock, assumed deposit specific density of 1.6 kg/m3 ). A maximum rate of erosion is attained in some catchments smaller than 1 km2 . (b) Inferred rates of tributary coarse sediment yield are plotted with the envelope of hillslope sediment mobilization shown in (a), above (grey region). Dashed line indicates the upper bound of the region with a doubling of the diffusion coefficients. . . . . . . . . . . . . . . . . . . .  39  xii  List of Figures 2-18 Longitudinal profiles of 6 major tributaries in the Chilliwack Valley. Vertical exaggeration is 20x. Black triangles indicate the transition from the colluvial process domain to the fluvial domain. White triangles indicate secondary knickpoints, conditioned either by glacial erosion in the master valley (Liumchen and Tamihi), or, in the case of Foley Creek, by a large landslide. . . . .  40  2-19 Longitudinal section of evacuated glaciofluvial valley fill, superimposed on the modern river profile in Liumchen, Tamihi and Slesse Creeks. Bedrock features, such as hanging glacial sills (described in the text) are shown in light gray. The Slesse Creek fill overlies a lacustrine layer near the junction with the mainstem. Vertical grid spacing = 100 m . . . . . . . . . . . . . . . . . . . . . . . . . . .  42  2-20 (a) Upper strata of the valley fill in Foley Creek ( 350 m upstream of Foley mouth). Exposed is a mix of glaciofluvial and debris flow deposits. (b) Upper strata of the glaciofluvial valley fill in Foley Creek (∼1.2 km upstream of Foley mouth). The coarse fluvial beds are approximately 45 m above the modern channel. The majority of the boulders are most likely Mt. Barr granite, indicating the headwater provenance of the bulk of the sediments (c) Truncated remains of a fan deposit that once interfingered with glaciofluvial fill in lower Tamihi Creek. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  44  3-1 Chilliwack Lake, with TRIM digital terrain model. Lines A, B, C and D indicate seismic sections discussed below. . . . . . . . . . . . . . . . . . . . .  48  3-2 Figure showing mid-lake CHIRP data on the North end of ‘C’ transect, Figure 3-1, and the continuous drape of lacustrine beds leading up onto the edge of the Paleface fan delta. See site (b) in the next figure for the larger setting within the lake sediments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  49  3-3 Composite seismic image of Chilliwack Lake (centreline transect, 9 km) showing major depositional units along the axial section. Five labelled features are discussed in the text: (a) continuous, laminated Holocene lacustrine sediments, (b) sandy outwash material (c) distorted reflectors within Pleistocene strata, (d) Paleface Creek delta, and (e) Depot Creek delta, showing portions of the pre-Fraser glaciation topography (see lateral cross-section, Figure 3-4). Vertical exaggeration is roughly 45x, though seismic velocity through the lower strata is likely somewhat higher than in the lacustrine zone. . . . . . . . . . .  52  xiii  List of Figures 3-4 Cross-section B (see Figure 3-1), the closest transect to the down-valley extent of Chilliwack Lake, and the deepest section (120 m water depth). Units a, b and c correspond to units labeled in the previous figure. Top figure shows the un-migrated data, with the characteristic ‘bow-tie’ structures. The migrated section, below, shows the synform structure at depth, though distortion is introduced to other parts of the trace. The valley walls are planar; curvature at the upper edges of the trace is due to turning of the boat and the array. .  53  3-5 Location of subsurface surveys at Depot and Paleface Creek fan deltas. Underwater topography was surveyed using sonar on a 100 m grid. Lake core locations and seismic trace paths are indicated. Section ’D’ refers to the seismic trace in Figure 3-6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  55  3-6 Section D (see Figures 3-1 and 3-5). Two seismic techniques used on the same profile at the toe of Depot Creek Fan: CHIRP (above) and acoustic air gun (below). The CHIRP record shows a detailed picture of the uppermost unit visible in the air gun profile. The airgun record shows the deeper strata . . .  56  3-7 Profile C (see Figure 3-1) with superimposed seismic record. Perspective view, facing south east, shows the relationship among the upper beds of lacustrine material, outwash beds, and several antiform reflectors at depth that likely indicate the structure of buried Pleistocence fans emerging from Paleface and Depot Creeks. The sand limit on the fans is inferred from the CHIRP record. Transverse seismic transects are indicated with black lines. . . . . . . . . . . .  57  3-8 GPR transects at Paleface fan delta. (a)Location map for the GPR transects. (b) Down-dip section. 0 m indicates the lake shore. (c) Strike section, moving NNE to SSW. A radiowave velocity of 0.07 m ns− 1 is assumed. This makes the limit of penetration 35 m (500 ns). The lower bounds of the sediment package were not detected. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  59  3-9 Isopach diagrams, (a) Paleface and (b) Depot Creeks. A hypothetical postglacial surface (top of outwash unit ‘c’, Figure 3-3), built with a 4th order polynomial grid, is subtracted from the modern topography of the lake fan deltas to yield an estimate of total coarse (sand and gravel) sediment accumulation from these catchments over the Holocene Epoch. The fines component of the fan deltas extends much further out into the lake and are estimated in a separate analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  60  xiv  List of Figures 3-10 Site map: Chilliwack Lake with the location and description of seven cores recovered from the lake. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  61  3-11 CHWK-06 - Vibracore from the deepest section of Chilliwack Lake. Dark bands in the column at left indicate discrete beds with concentrations of coarser clastic material and/or fine organics. The upper portion of the core may have undergone some compaction during the coring operation, interrupting the otherwise coherent gradient of increasing density with depth. . . . . . . .  64  3-12 Cross-core comparisons of two parameters, water (%) and loss on ignition (LOI). 65 3-13 Ternary diagram showing compositional fields for a number of Holocene volcanic sources within the Cascades and Coast Mountains. Points (Xs) show microprobe readings taken on shards from the Chilliwack Lake tephra layer. .  65  3-14 Three cores from Chilliwack Lake show a number of fire episodes over time. From left to right: CHWK-06 (Distal 2), CHWK-05 (Distal 1), and CHWK-04 (Paleface). Positive X-Rays (4 mAs at 75 kV) reveal very fine low-density ash layers. Based on paleomagnetic dating and stratigraphic sequence, two event beds are correlated across two distal cores and a third from as far away as Palefacefan delta. The events span an approximate range of 1 350 to 1 750 cal. years B.P. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  66  3-15 Percentage of coarse (>16 mm) clastic lithologies found in active bars, Upper Chilliwack River (open squares) and major tributary sources (Bear and Indian Creeks, black diamonds). There is a disproportionate representation of volcanic lithologies in the channel, despite the mostly granitic source material provided by the major lower tributaries, which represent 41% of the upper catchment drainage area.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  67  3-16 Graph showing the relative increase in iron and elements associated with heavy minerals and magnetic oxides within the silt fraction of channel sediments (Upper Chilliwack River) and lake sediments. Fe values range from 3 to 5%. Here they have been log-ratio transformed, centered on zero. Individual elements Cr, Mg, Ti and V have been similarly transformed, then summed, to create a composite index. Individually, V has the strongest down-valley gradient, and Ti the weakest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  68  xv  List of Figures 3-17 Susceptibility, measured at 5 cm intervals along the length of each core. There is an evident longitudinal gradient, with magnetic mineral concentration increasing down-valley.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  68  3-18 Sedigraph grain size curves from lake core sediment samples, Chilliwack Lake  69  3-19 Traces of magnetic inclination and declination from four cores are shown in the first panel, as well as their average. These are the raw (unsmoothed) data, interpolated onto a standardized 200-element array to facilitate crosscore statistics. Standard deviation among the magnetic readings are shown in the middle panel. The shorter core extends only to 2 500 B.P., and thus sample size diminishes prior to this date. The final panel shows Fish Lake and Mara Lake datasets compared with the Chilliwack series.  . . . . . . . . . . .  71  3-20 Paleomagnetic chronology (calendar years) mapped to sediment core depth (cm). Red dots indicate fire disturbance events recorded in the deepest core. Elevated rates of sedimentation are inferred at the Upper Delta (CHWK-03) based on radiocarbon dating of an organic layer at 85 cm (180 ±40 years BP - roughly A.D. 1670-1800; radiocarbon calibration for this era is only loosely constrained). Before approximately 3 200 BP there appears to be a slightly elevated rate of accumulation. Reduced major axis regression within the distal cores indicates a significant change in slope (α=0.01), though there are fewer core samples deeper than 2 m to confirm this shift in trend. . . . . . . . . . .  72  3-21 Isopach map showing the inferred depth of lacustrine sediment since the end of outwash deposition and the onset of lacustrine conditions at Chilliwack Lake, based on seismic and sonar surveys. The seismic transects are overlaid in white. 73 3-22 (a) Volume of lacustrine sediment with depth. Uncertainty bounds are indicated by multiple profile lines. Model is based on seismic imaging of the sediments and follows assumptions outlined in the text. (b) Variation of specific dry weight density of sediments with depth in each of the lake cores. The logarithmic trend lines indicate the expected pattern of density with depth. .  73  3-23 Fine sediment yield for Chilliwack Lake shown in comparison to other lakes in British Columbia. Dots size indicates the proportion of basin glacial cover. Data collated from Gilbert et al. [1997]; Desloges and Gilbert [1998]; Schieffer et al. [1999]; Hodder et al. [2006]. . . . . . . . . . . . . . . . . . . . . . . . . .  75  xvi  List of Figures 4-1 Six-part deglacial history of Chilliwack Valley: (a)-(c) depicts the retreat phase of the final stage of the Sumas ice lobe, active floodplain outlined in green, 11 200 to 10 500  14 C  years B.P. (d) A remnant lobe of ice in Sumas Valley,  following ice retreat, 10 500 - 10 000  14 C  years B.P. (e) Sumas Lake fills the  depression left by the ice, basin fills throughout Holocene (f) The modern landscape: City of Chilliwack and Sumas Valley. Chilliwack River has been channelized into the Vedder Canal. Adapted from Saunders et al. [1987], Cameron [1989] and Clague et al. [1997]. . . . . . . . . . . . . . . . . . . . . .  80  4-2 Isopachs of the estimated net erosion from the mid-valley, at Slesse confluence. The maximum erosion depth is 106 m. Shallow erosion is shown across the intact remnants of Larson’s Bench, after a pre-erosional surface was fit to it. .  82  4-3 Longitudinal survey of terraces between Foley and Tamihi Creeks. Degradation appears to be hinged on a knick point close to the outlet of Foley Creek. Terrace flights have been grouped as high (red), middle (light blue) and low (green). Points surveyed on the terraces are coloured accordingly; points on the modern river are coloured dark blue. River topography is taken from the BC TRIM DEM breaklines . . . . . . . . . . . . . . . . . . . . . . . . . . . .  83  4-4 Fine sediment source material is delivered episodically from deep glacio-lacustrine deposits in the mid-valley reaches. . . . . . . . . . . . . . . . . . . . . . . . .  84  4-5 Larson’s Bench is incised by a late Holocene channel. The formerly active layer of the incising channel overlies sandy deltaic fill on the left; lacustrine clay on the right. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  85  4-6 Cross section showing the available evidence (exposures and well logs) to describe the Pleistocene and post-glacial valley fill. The modern channel is demarcated with a dashed line. Upper planform map shows detailed terrain mapping [Armstrong, 1980; Ryder, J.M. and Associates, 1995] that outlines major lacustrine and morainal strata. Ryder Lake Upland is depicted with a veneer of aeolian material and thick till accumulations beneath. Points on the plan map indicate exposures and well logs that are graphed in the section detail. Vertical exaggeration is 20x. There is assumed to be considerably more lateral complexity within the valley’s deposits (and eroded volume) than can be reconstructed with this evidence. . . . . . . . . . . . . . . . . . . . . . . .  86  xvii  List of Figures 4-7 Reconstructed geometry of the major valley landforms in Lower Chilliwack Valley, downstream of Tamihi Creek. (a) shows a composite set of cross sections of the valley near Ryder Creek, 3 km downstream from the crest of Tamihi Moraine. The shaded area represents the zone of minimum net erosion of glacial material (scenario III, see Figure 4-8 and text). (b) is a longitudinal valley cross-section looking North, encompassing the former locality of Tamihi moraine. A mid-valley lake elevation is shown, based on the assumed elevation of the delta front at the distal end of the Larson’s Bench sandur. Cross-section lines from Ryder Upland showing the maximum height of the remnant moraine are in the background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  89  4-8 Scenario III, with minimum bulk volume erosion from valley sidewall sediment sources, including the Ryder Lake Upland. Contours of erosion and deposition are in increments of 15 m. Total erosion is 344 ±45 × 106 m3 . Deposition evident at the base of Ryder Upland is approximately 35 × 106 m3 . . . . . . .  90  4-9 Figure showing a band of elevations between 1 and 35 m a.s.l. in Sumas Valley, Vedder Fan and the Fraser River. The bounds of Figure 4-14, showing City of Chilliwack Drilling work (2006), are highlighted. Note the semi-circular geometry of the Vedder Fan, and the relatively low-lying surrounding topography. TRIM BC digital elevation data. . . . . . . . . . . . . . . . . . . . . . . . . .  91  4-10 Historical planform of Chilliwack River north of Vedder Crossing, ca. 1891 (top) and 1991 (bottom). An interlinked network of channels (four of the larger threads are labelled) alternately occupied and abandonned various section of the fan as it evolved. The historical figure was derived from an ordnance survey of the Chilliwack area. The modern map was generated from BC TRIM mapping. 93 4-11 Model of alluvial fan growth, after Blair and McPherson [1994]. The lower bounding surface slopes slightly upward, as deposition keeps pace with a slowly rising base level. In the case of Vedder Fan this corresponds to active deposition on the Fraser Valley floor over the course of the Holocene. . . . . . . . . . . .  94  4-12 Gravel quarry near Vedder Road at Watson. Looking westward: flow was from left to right. Photo credit Vic Galay, Northwest Hydraulic Consultants. . . .  96  4-13 Gravel quarry near Vedder Road at Watson. Photo is looking southward, and the direction of flow was out of the page. Photo credit Vic Galay, Northwest Hydraulic Consultants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  96  xviii  List of Figures 4-14 Locations of City of Chilliwack drilling operations, 2003-2006 (1995 aerial photography). See Figure 4-9 for map location with respect to the larger fan area.  97  4-15 Grain size distribution from a sampling of units recovered from sonic drilling. Representative fractions > 32 mm could not be effectively recovered from the core samples, but the fractions that were examined effectively show the bimodal nature of the fan deposits. . . . . . . . . . . . . . . . . . . . . . . . . .  98  4-16 Cross-section of Vedder Fan, based on examination of sonic core cuttings and assembled well logs. Letters refer to locations of photos in Figure 4-17. Narrower logs with solid colours are database wells with minimal detail and have not been examined. Dashed lines indicate interpolated surfaces with an associated date based on radiocarbon samples. Most elevations are approximate, unless significant figures are indicated. . . . . . . . . . . . . . . . . . . . . . . 100 4-17 Photos of sonic core material, illustrating facies assemblages that are indicated in Figure 4-16. Wood dated at 10 125 ±245 cal. years B.P. was recovered from the silt unit shown in (i) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 4-18 Regression curve through multiple radial sections of Vedder Fan . . . . . . . . 103 4-19 Isopach diagrams of gravel, sand and lacustrine material within the stratigraphic bounds of Vedder Fan and surrounding area. Bounds of former Sumas Lake are shown in black. See text for description of numbered localities in 4-19c.104 4-20 Cumulative volumetric growth of Vedder Fan from 11 035 BP to present, based on the assumed stratigraphic relationships indicated in Figure 4-16. Error bounds indicate 16% error assumed in the calculations. . . . . . . . . . . . . . 106 5-1 Sampling bar material in the lower Chilliwack Valley (site # 111-04) . . . . . 111 5-2 Sampling sites within Chilliwack Valley. (a) Lower Valley and (b) Upstream of the Slesse Creek confluence. Open circles indicate surface sampling only, and filled circles indicated both surface and bulk sampling. . . . . . . . . . . 112 5-3 Duplicate samples taken at (a) just upstream of the Tamihi Bridge (113-04), and (b) a large bar complex upstream of Borden Creek (135-04). Lines indicate the relative difference among samples for each size fraction, relative to the first of the two duplicates (‘0’ datum). Samples are not significantly different within each sedimentary link, but show a closer affinity than between links. . . . . . 113 5-4 Grain size distributions for bulk and Wolman samples within the Chilliwack mainstem (a,b), Tributaries (c,d) and Glacial Tills (e). . . . . . . . . . . . . . 115 xix  List of Figures 5-5 Mean, standard deviation, skewness and kurtosis of gravels (Ψ scale) sampled in Chilliwack Valley. (a) shows the statistical moments of the full distribution, including fines, (b) shows the same statistics with the distribution truncated at 2 mm (no fines). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 5-6 Downstream fining along Chilliwack Valley Mainstem. (a) shows the cumulative fractions of the subsurface sediments from Chilliwack Lake to the end of the gravels in Vedder Canal. Samples from Vedder Crossing to the canal were taken by Y. Martin and BC Environment (1989-1991). (b) shows the fining pattern for surface samples. No samples were taken beyond Vedder Crossing. (c) is similar to (b), except the fining gradient is shown from the mid-reaches of Slesse Creek (grey shading) to Vedder Crossing. The comparison shows clearly the geographic source terrain of the coarse material in the mid-reaches of Chilliwack River. Dashed lines indicate regression fits to the data downstream of Slesse Creek. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 5-7 Percentage sand (<2 mm) in surface and subsurface deposits, plotted along the length of Chilliwack River. . . . . . . . . . . . . . . . . . . . . . . . . . . 120 5-8 (a) Downstream variation in percentage granite composition of streambed sediments. The size composition of the total granite percentage is broken into 3 classes: 8-22 mm, 32-90 mm and 128 mm and larger. Relatively few granite pebbles were recorded, while a large number of boulders are evident. The dark line indicates the relative percentage of granitic terrain upstream (not including the drainage above Chilliwack Lake). (b) Shows the percentage of granitic clasts in the 32-90 mm category relative to both upstream granitic terrain (black) and percentage of granitic boulders (light gray). This is an index of how much mobile granitic material we might expect to see in the channel versus the observed. See text for discussion. The Sternberg relation for diminution of grain size from 90 mm to 32 mm is shown (dotted line, axis to the right). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121  xx  List of Figures 5-9 Bivariate plot showing the major geochemical domains among channel sediments and glacial deposits in the Chilliwack Valley. Among the raw element data, K and Fe provide the best discriminating potential. Fine channel sediments from batholith sources (upper left) and Chilliwack Group or Cultus Group sources (lower right) are reasonably distinct. Mainstem alluvium plots as an intermediate field, and glacial deposits plot to the lower left. A few seemingly anomalous points are evident: (1) a sample taken from glaciolacustrine bluffs at the Slesse/Chilliwack confluence. There is a strong affinity here for headwater lithologies (2) two samples taken from the Slesse Park landslide, which show a concentrated (proximal) mix of the two source terranes. (3) is a distinct sandy bed from Slesse Park landslide that clearly has a strong indication of Slesse source material. There is also a ‘mainstem alluvium’ point here that was sampled downstream of the Nesakwatch Creek confluence, illustrating its strong influence on mainstem sediment composition. (4) is from Borden Creek, whose source terrain appears relatively high in silica and low in indicator elements.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127  5-10 (a) shows the fine sediment samples and geochemical data from source rocks cast in Fe-K factor space. There is clearly an influence from both sources on the elemental composition of the sediments. Source rock information comes from compilations in Richards [1971], Sevigny and Brown [1989] and Tepper and Kuehner [2004]. (b) is a vector diagram illustrating the magnitude and direction of each element in Fe-K factor space. (c) shows the relation between mainstem samples (open circles below Slesse Creek, closed circles joined by lines above, distance to Vedder Crossing indicated) and tributary sediments (crosses represent the mean of all samples within each tributary). There is a clear influence of granitic lithologies on the upper mainstem sediments that diminishes for samples lower in the valley. See text for further discussion. . . 128 5-11 Longitudinal pattern of ‘K’-Factor variation, in comparison to the relative proportion of granitic source terrain upstream. Peaks in the gradient are evident at Slesse and Tamihi Creeks, and one near 30 km, a short distance above Alison Pool. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 6-1 Figure illustrating 1-D representation of channel floodplain . . . . . . . . . . 139  xxi  List of Figures 6-2 Flow chart for the ACRONYM-based Chilliwack sediment routing model. Initial and boundary conditions are fed by the user (upper left), then hydraulics and sediment transport are calculated, generating a new bed configuration. The model is updated, and continues in an iterative process. Some of the key variables and adjustable parameters are indicated. GSD refers to grain size distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 6-3 Mean of average daily flows, 1977-2005, at Chilliwack Lake, Chilliwack Canyon, Slesse Creek and Vedder Crossing. . . . . . . . . . . . . . . . . . . . . . . . . 148 6-4 Partial series for average daily flows at Vedder Crossing, based on the 19522005 gauging period. There are two distinct flood regimes for spring/summer and fall/winter. The mean annual flood for the fall/winter is estimated to be a daily average of 320 m3 /s . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 6-5 Figure illustrating the relative contribution of various daily mean flows to total suspended sediment transport. . . . . . . . . . . . . . . . . . . . . . . . . . . 150 6-6 (a)Long profile with stratigraphy. The posited initial longitudinal profile is at the top of the stratigraphic reconstruction. Tamihi moraine has either filled the lower valley (‘High Moraine Scenario’), or only partially filled it. The river flows across the remains of the drained lake and upper delta. While the boundaries are more gradational in reality, they are represented as discrete sections in the model. The modern profile is shown below in gray. Bedrock and bouldery knickpoints are indicated. (b) Graph of the volume of glacial material that was stored above the modern profile within each stratigraphic grid cell, taking account of the valley floodplain width. . . . . . . . . . . . . . 151 6-7 Isolation of sand, gravel and boulder modes within the grain-size distribution, using curve-fitting techniques. The boulder lag mode potentially masks the shape of the upper limit on the gravel mode. . . . . . . . . . . . . . . . . . . 153 6-8 a) the average among all fluvial sand modes isolated from surface gravel samples taken along the length of Chilliwack River. b) illustrates how varying the sand mode for a given sample alters the shape of the grain size distribution. The mixing proportion ranges from -16 to +16% of the initial distribution. . 153  xxii  List of Figures 6-9 Schematic of the model grid in the lower valley. Each computational cell represents 360 m of river length, and the total active width of the floodplain. Sediment source points are indicated with arrows. The large valley wall slump upstream of the Tamihi moraine (mid-lower right) is modelled as an example of a large discrete input to the channel. . . . . . . . . . . . . . . . . . . . . . 155 6-10 Gravel and sand sediment budget for the Chilliwack model domain below Nesakwatch Creek confluence. The net bulk volume of outwash material eroded from the major valleys has been derived in previous chapters. Three alternative scenarios are presented for the total volume of material eroded from the lower valley at Tamihi Moraine. Bedload quantities are based on the relations developed between basin size and volumetric bedload yield developed in Chapters 2 & 3, partitioning the total yield to discount tributary valley fills and account for only gravel and sand from hillslope sources. . . . . . . . . . . . . 157 6-11 Polynomial curves describing the sediment volume / outlet elevation relationship for Slesse and Tamihi Creeks. The outlet elevation is shown as a relative scale. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 6-12 Space-time diagrams illustrating long-term transport trends in Chilliwack Valley. (a) shows the balance of degradation and aggradation across space and time. Rates of change are in units of vertical metres per Model Year, with contours showing 1 m increments, up to a maximum range of 6 m. The elevation at the downstream boundary is the Fraser River floodplain, rising at a rate of 40 cm per model-year. (b) shows the corresponding unit bedload flux rates in the model. On the left axis, time in Model Years, Flood Years and an hypothetical scale for calendar years is shown.  . . . . . . . . . . . . . . . . . 161  6-13 Longitudinal profile, showing successive bed elevations over the course of a model run. Bed elevations at node 44 (downstream of Borden Creek) are shown in geometric series through time (inset). . . . . . . . . . . . . . . . . . 162  xxiii  List of Figures 6-14 (a) Illustration of the range of concavity that may be achieved during degradation. The worst-case (RMS = 4) is a planar profile between the knickpoint and the lower valley. (b) The RMS difference between river profile and model results, for a range of specified discharges. Parameters were freely varied within the ranges specified in Table 6.2 to generate a suite of model runs. Bubble size indicates the RMS difference for the longitudinal subsurface sand profile. Results generally indicate that the profile fit improves with lower discharge, and often improves at the expense of the sand profile RMS. Circle indicates reference runs, which had a relatively coarse composition and a low abrasion value. All of the lowest points have high abrasion rates. . . . . . . . . . . . . 163 6-15 Tributaries that are likely to disrupt the downstream fining pattern will plot above Rice’s discriminant function Rice [1998] . . . . . . . . . . . . . . . . . . 165 6-16 Modelled fractional reduction in grain size for the final equilibrium channel, compared with field results. The subsurface samples from the field are compared with the modelled bedload for a range of abrasion values. . . . . . . . . 166 6-17 The effect of abrasion on the long-profile fit: with a higher abrasion parameter specified, the long-profile fit improves accordingly. The RMS difference between modelled and field measurements of the subsurface sand content (bubble size) profile worsens, as a surplus of abraded material is added to the mix. Reference runs are circled; a coarser gravel mix achieves a better fit to the long profile, without having to specify a higher abrasion coefficient.  . . . . . . . . 168  6-18 Subsurface sand percentage (<2 mm) from field samples compared with the modelled bedload. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 6-19 (a) Cumulative deposition on Vedder Fan over the course of model runs with varying parameter values. (b) The effect of variations in abrasion (0 to 0.040), sand/boulder content (±10% and 5% variation, respectively), D50 (range through 1 phi mode), profile form (3 scenarios) and tributary flux (1 to 3x) are shown relative to the reference run (‘100%’) to illustrate the range of variability in the time taken to deliver 1.5 × 109 m3 , based on the parameter specification.  170  xxiv  List of Figures 6-20 Cumulative deposition on an idealized Vedder Fan over the course of the reference model run. Model stratigraphy of Vedder Fan, showing D50 at left and the proportion of the total sediment yield contributed by mainstem gravels represented in the fan at right. Blue contours indicate fan topography at even 2.5 ka (approx.) intervals. The modern topography is overlaid in red. . . . . . 171 6-21 Bedload transport gradient in space and time, illustrating changing specific transport rate along the length of the channel during the process of downcutting through the valley sediment stores and re-equilibration of the channel. . 174 6-22 Bedload transport rates at Vedder Crossing throughout a reference model run. The lower three lines indicate flux rates for material sourced from upstream of the model grid, from tributaries (dashed), and from the mainstem channel. The sum of these fluxes is represented by the upper black line. . . . . . . . . 175 7-1 Contemporary fluvial suspended sediment yield as a function of drainage area [Church and Slaymaker , 1989], overlaid with long-term estimates of coarse sediment delivery in Chilliwack Valley. Black points and open circles are from Church and Slaymaker [1989]; see their text for description of the different sediment yield regimes. Coloured points are from the present study (see Chap. 2). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181  xxv  Acknowledgements  I would like to thank professor Michael Church for his insight, advice, and his perseverance in attempts at uncovering the articulate and meticulous scientist that is buried, somewhere, deep within me. The pages ahead are much clearer for his efforts, though I take full responsibility for any lingering lapses in accuracy and logic. I also would thank my committee members Olav Slaymaker and John Clague for their enthusiasm and support in this endeavour. Brett Eaton, Marwan Hassan, Andre Zimmerman, Dave Campbell, Josh Caulkins and Dave Lutzi have been the best of supportive colleagues during my time at UBC, and I owe them a debt of gratitude for their help on many fronts, both in the field and in our many discussions and cogitations. During my time at NIWA, Murray Hicks and Jeremy Walsh provided much added insight into sediment budgets and hydraulic models, and I thank them both for sharing their experience and ideas. My deepest acknowledgements go to my wife Kristiann and son Max, who kept everything else on course while Dad was distracted with writing and research ventures. It has been a great adventure, and continues to be so, thanks to Kristiann. The are many people to thank for their assistance along the way. The chapters that follow could not have come to fruition without their help. Bruce Thomson provided much of the background: the maps, the data, and in particular, his unparalleled knowledge of the area was a very important asset in the development of this work. Rob Huggins (Geometrics, Sunnyvale, CA) provided much help and expertise in conducting the seismic transects on Chilliwack Lake and seismic refraction surveys on Larson’s bench. Geometrics generously provided a GEODE data acquisition system, a geophone streamer ’eel’ and other essential equipment. Bob Gilbert (Queens) kindly lent his CHIRP system for subsurface profiling of the lake. Brian Menounos and Melanie Grubb (UNBC) provided much assistance and helpful advice. Brian spent a great deal of his time on the geophysical work  xxvi  Acknowledgements and supplied the vibracoring equipment for the cores recovered from Chilliwack Lake. I would also like to thank staff at BC Parks, for facilitating several survey campaigns on Chilliwack Lake. Randy Enkin and Judith Baker (GSC Pacific) introduced me to the world of paleomagnetism, and carried out all of the susceptibility measurements using their OLGA susceptibility meter. The results helped significantly with respect to establishing the chronology of sediment deposition in Chilliwack Lake. All of the seismic data processing was done using the computers and Globe Claritas software in Ron Clowes’ seismology research group (EOS). I would like to thank Phil Hammer and Jounada Oueity for their help in sorting out the many intricacies of data processing. Ted Hickin and Michael Roberts (SFU) generously lent their equipment and software for the GPR studies of Paleface delta. The field work was carried out with Natalie Helmstetter and BJ Kelly, and I thank them both for their time in the field and discussions of GPR technique. Thanks to Vic Levson at the British Columbia Geological Survey for making available his exceptional database of geophysical logs and water wells in the Chilliwack area. Janet DeMarcke (City of Chilliwack) and Dan Emerson (Emerson Consulting) kindly let me take part in their research on the groundwater conditions and stratigraphy of Vedder Fan. They both supplied much data and provided access to vibracore drill cores during the City’s drilling campaign. I would also like to thank Kathryn Black (UBC) who carried out much of the analytical work on the drill cores, and made significant contributions to the development of the subsurface model. The geochemical tracing component in the thesis work was inspired through discussions with W.K. Fletcher. Thanks go to Richard Friedman and Jim Mortensen for discussions and making their lab space and resources available. Field work was a large undertaking that couldn’t have been done without much assistance. In a addition to the colleagues mentioned above who were recruited at regular intervals to heft buckets of gravel around the bars of Chilliwack River, I would like to thank Dason Commodore, Sydney Kjellander, Jason Rempel, and Dev Khurana. Staff at Area Services Unit, Chilliwack CFB were particularly helpful in providing access to their properties in Chilliwack Valley. A Natural Sciences and Engineering Research Council of Canada grant to M. Church supported this research.  xxvii  Chapter 1 Introduction  1.1  Problem Statement  The last glacial cycle has left a distinct imprint of erosion and deposition in the Canadian Cordillera. In British Columbia and other formerly glaciated regions, the phenomenon of paraglacial sedimentation [Church and Ryder , 1972; Ballantyne, 2002b] has been well documented. In the large mountain valleys, sediment that was originally eroded by the advancing ice continues to work its way through the sedimentary network, many thousands of years after deglaciation [Church and Slaymaker , 1989]. A large volume of sediment derived from relict glacial landforms has a continuing influence on rates of sedimentation in large valleys. Most dating evidence that has been recovered from paraglacial landforms such as outwash terraces, fans and lake deltas [Ryder , 1971; Jackson et al., 1982; Clague, 1986; Brooks, 1994; Ballantyne and Benn, 1994; Ballantyne, 2002a] indicates that the sedimentary relaxation following deglaciation follows an approximately exponential pattern. Initially high rates of sediment transfer from proglacial zones gradually taper off as glaciers retreat and meltwater discharge and sediment supply diminish. The sediment mass is then gradually reworked within the drainage network. As upstream sediment supply is gradually exhausted and meltwater discharge decreases, aggraded landforms such as glacial outwash and alluvial fans become incised and remobilized material is carried downstream. Their initially unvegetated surface renders them particularly vulnerable to erosion. The effects of base-level fall generated by retreating ice gradually propagate upstream. Incision is eventually checked by vegetation and stream bed armouring in the postglacial period. Episodes of renewed downcutting occur intermittently, inducing smaller-scale cycles of aggradation and degradation [Schumm, 1973; Church, 2002; Dadson and Church, 2005]. 1  Chapter 1. Introduction The effects of glacial disturbance on a large valley are complex and time transgressive - it requires several thousand years for the postglacial sedimentary regime to become fully established in large (>1 000 km2 ) valleys. The fluvial system experiences a downward shift in transporting capacity as sediment transport processes move toward a meta-equilibrium with postglacial hydroclimatic conditions [Baker , 1983; Knox , 1983; Chatters and Hoover , 1992; Kesel et al., 1992]. As the river responds to the changes imposed by falling base-level, it may do so by adjusting one or many channel characteristics [Schumm, 1973; Schumm and Rea, 1995]. Equilibration of the fluvial system is governed by sediment supply. A fundamental control over river regime is the rate and calibre of material supplied to the mainstem from its wider catchment. Many studies of long-term river metamorphosis have focused on a landscape element at a particular scale (e.g. valley floodplain), without due consideration of the sedimentary conditions in the larger fluvial network. Measurements of flux, or reconstruction by means of proxies, from a single point or reach in a watershed do not provide the necessary quantitative insights into linkages among storage elements that govern fluvial response on bars, floodplains, hillslope and fan deposits. At the larger scale, the fluvial system seems to be subject to complex (linked, causal) changes in configuration over long periods that are difficult to deal with using equilibrium transport models. Figure 1-1 shows a plot of time-space relations and some of the different virtual velocities at which water and sediment traverse the landscape. The virtual velocity of active channel sediments is of order 100 m/yr [Church, 1996]. However, at greater spatial and temporal scales, there is a tendency for this velocity to diminish considerably due to increasing opportunities for deposition at junctions and breaks in slope, resulting in extended periods spent in storage. Thus, one may speculate that the long term rates of transit for river gravels are closer to 1 to 10 m/yr. As the spatiotemporal domain under consideration extends to larger and larger scales, contingencies arise whereby established equilibrium river regimes may be modified. Complications include climate fluctuations, sediment storage, and any adjustment or accident that results in a change of the geometric configuration of the channel or drainage network, limiting or enhancing the rate of sediment flux. The problem of contingency emphasizes the need for renewed investment in historical approaches to landscape studies. Sugden et al. [1997] and Summerfield [2005] point out that the lack of studies in this vein has hindered progress in developing our understanding of long-term landscape evolution. This is presently an impor-  2  Chapter 1. Introduction global  10 10  4  3  chaotic  regional 10 10  no detection  2  al  -1  10  Distance (km)  10 10 10 10 10 10  e tiv ac  1  0  10  -1  0  stochastic -2  10  -3  1  m  yr  i uv  se  m di  en  t)  fl  (  -1  m  yr  -1  m  yr contingent  deterministic  -4  -5  -6  10  -8  second  10  -6  minute  10  -4  hour  10  -2  day  10  0  year  10  2  10  4  10  6  10  8  Time (years)  Figure 1-1: A spatio-temporal view of sediment transfer across the landscape, adapted from Church [1996]. The box indicates the range of scales that are addressed in this thesis - river distances of up to 100 km, and timescales ranging from 10 years of river measurements out to 14 ka cal. B.P.). The annotations in the diagram indicate appropriate modes of explanation across various spatio-temporal scales.  tant frontier in geomorphological research, as it challenges the discipline to model and predict change. Sediment budgets have commonly been used to evaluate the relative magnitude of sedimentary contributions from different sources within a watershed in order to better understand and manage transfer of material through the drainage network [Dietrich and Dunne, 1978; Reid and Dunne, 1991; Benda and Dunne, 1997; Malmon et al., 2003]. In a long-term approach to modelling an evolving landscape, it is possible to apply a similar mass-balance framework to constrain the evolution of storage elements and the availability of sediment to the fluvial system over the course of time. In some mountainous regions, tectonic uplift is considered an important factor in the Holocene time frame, but in the Chilliwack Valley this effect is considered to be relatively marginal. By reconstructing the history of a glaciated valley and investigating the geometry, sedimentology and chronology of landforms, information on the long-term rates of transfer may be 3  Chapter 1. Introduction inferred. Furthermore, we may understand better the mechanisms of the relaxation response as the mass of coarse materials is reworked and transported onwards and out of the basin. This will generate a dataset that may be used to test hypotheses of landscape evolution in formerly glaciated terrain. The analogy of a cascade of reservoirs has been used in a number of studies [Church and Slaymaker , 1989; Warburton, 1993; Ballantyne, 2002a; Church, 2002]. Conceptually, disturbance translates from upper catchments to lower trunk valleys, very much as a flood hydrograph, though the time scale is considerably longer and transit is complicated by a number of factors. Smaller disturbances distributed through Holocene time will be superimposed on the primary paraglacial signal. Basins at different spatial scales may respond to disturbance on different temporal scales (Figure 1-2). A perturbation visited upon a single low-order catchment will not have a large impact downstream. However, if disturbance from several catchments simultaneously impacts the sediment balance of a higher-order system, there may be a detectable surge in the trunk channel. The hypothesis pursued in this thesis is that the pattern of postglacial sediment yield is strongly conditioned by (1) the volume and geometry of aggraded glacial stratigraphy in the valley, (2) the rate and timing of sediment delivered from tributaries, (3) the rate of base-level fall in the mainstem, and (4) the size grading of material delivered to the mainstem channel. While the first three items have often been dealt with in other studies, this work seeks to analyze all four boundary conditions as an integrated whole, in light of recent advancements that have been made in modelling the transport of gravel mixtures. The following chapters develop the boundary conditions as fully as possible in order to test the sensitivity of these controls on the Holocene river morphodynamics in Chilliwack Valley. I propose that the sedimentology of the glacial valley fill in Chilliwack Valley has had a central and persistent influence on the development of the post-glacial river system, most notably on the form of the final long profile, the asymptotic transport equilibrium that is achieved, and its characteristic response to disturbance. If properly parameterized, a physically-based numerical 1-dimensional sediment transport model with appropriate conservation of mass and treatment of multiple grain size fractions provides a good laboratory for testing some of the theorized notions of paraglacial system response. The transit of the reworked glacial material through the watershed is moderated by deposition at tributary junctions and in sedimentary reservoirs along the system, the vegetation of formerly active sources, and the differential mobility of grains in the mixture. The textural  4  Chapter 1. Introduction  Rate of Sediment Flux in Relation to the Subaerial Norm  3  2 1 km2  10 km2  1 12 ka 100 km2 8 ka  Time (post-glacial)  4 ka  Catchment Area  1000 km2 Present  Figure 1-2: Conceptual picture of the transit of a large-scale sedimentary disturbance through space and time (after Church [2002]). The disturbance decays quickly away from proximal watersheds. Dispersion is due to deposition of material at tributary junctions and in sedimentary reservoirs along the system, vegetation of formerly active sources, and differential mobility of grains in the mixture. The figure illustrates a “primary” disturbance of post-glacial time, upon which smaller-scale perturbations are super-imposed.  character of the network is an important moderating factor. Residual lag (boulder) modes condition the river bed in a number of important ways, most notably in the stability of the bed structure, hydraulic roughness, and in the evolution of the long profile. The relative quantities of sand in the system will also condition transport and dictate rates of response to disturbance. The modern river is the unique product of the source material yielded by glaciation.  1.2  The Study Basin  Chilliwack River has a drainage area of 1230 km2 (not including the Cultus Lake / Columbia Valley drainage), with most of the major headwaters situated in the North Cascades National Park in Washington State (Figure 1-3). It is the only major left-bank tributary of the Fraser River in the Lower Mainland, descending from the North Cascade Mountains through a broad 5  Chapter 1. Introduction glacial valley, emerging to form an expansive alluvial fan on the floor of Fraser Valley. The river is still actively eroding many remnant glacial and paraglacial deposits. Coarser deposits of reworked glacial material remain in place throughout the drainage network. A great deal of Holocene alluvium remains in storage in the lower reaches [McLean, 1980], and the river adopts a wandering habit through these sedimentation zones. The river has experienced a primarily degradational trajectory throughout the course of the Holocene Epoch. The length of Chilliwack River is roughly 80 km, with a lake in the upper valley truncating the lower mainstem length to 50 km. The pattern of sediment yield in British Columbia established by Church and Slaymaker [1989] indicates that all basins larger than 1 km2 are still to some degree under the influence of British Columbia’s glacial legacy. Based on the observed behaviour of several large-scale sedimentary disturbances, Church [2002] estimates that the rate of propagation (in years) of a large disturbance traveling as a dispersive wave scales approximately with area (in km2 , t ∝ Ad ), much more slowly than the virtual velocity of active channel sediments. The scaling relation for distance along the channel, L(in km), is approximately L ∝ t1/2 . Considering the timescale of interest is 13 000 years, catchments that are larger than 10 000 km2 or have a mainstem length greater than 100 km should still be experiencing transit of the end-glacial disturbance. Catchments that are somewhat smaller are likely to be still experiencing the asymptotic tail of the late-glacial disturbance. In his study of the smaller Squamish Valley (3 850 km2 ), Brooks [1994] concludes that the Squamish Valley has eroded the majority of its reworked glacial sediment. Evidence of the evolution of the drainage network following Fraser Glaciation is remarkably well preserved. Traces left from glaciers moving through the valleys indicate interlinkages in time and space among tributaries, the trunk valley, and neighbouring drainages. Many relict features such as sandar, kame terraces, glaciolacustrine deposits and lateral moraines remain clearly discernible and have been documented extensively [Chubb, 1966; Easterbrook , 1971; Munshaw , 1978; Clague and Luternauer , 1982; Saunders, 1985; Saunders et al., 1987; Easterbrook , 1992; Watson, 1999]. An advantage of a system with a primarily degradational history is that the additional complexity introduced by episodes of significant aggradation does not hinder the historical interpretation of landforms. Dating has been successfully carried out on a number of the major glacial landforms. Year zero for the sediment budget is approximately 13 300 calendar years ago (∼11 400  14 C  years  B.P.), when the final layers of deltaic sands were deposited on the shores of an unnamed mid valley lake, and the glacier at the head of the valley retreated from its terminal moraine and  6  Chapter 1. Introduction  Figure 1-3: Chilliwack Valley and surrounding area.  became uncoupled from the outwash sandur [Clague and Luternauer , 1982; Saunders, 1985; Saunders et al., 1987]. In the lower valley, wood from an ice-contact face at Cultus Lake has been dated at 11 300 ± 100  14 C  years B.P. [Lowdon and Blake, 1978].  The Chilliwack Valley has been subject to several episodes of glaciolacustrine sedimentation, when blockages at the mouth of the valley caused the formation of a lake [Clague and Luternauer , 1982; Saunders, 1985; Saunders et al., 1987]. These deposits are relatively common in the watersheds that drain into the Fraser Lowland. Ward and Thomson [2004] point out that, in valleys such as the Chehalis and the Chilliwack, the ice accumulation area is quite small in comparison to that of the Fraser Lobe and thus the valley glaciers responded differently to climatic changes. During advance, the valleys lagged behind the Fraser lobe whereas, during retreat, the valleys responded more quickly to climatic amelioration. This resulted in glacial lakes forming during both advance and retreat phases [see also Johnsen and Brennand , 2004].  7  Chapter 1. Introduction  1.2.1  Physiography  Climatic conditions in the valley span a gradient from a warm maritime climate in the west to a colder continental climate in the east. Climate also varies significantly by elevation. The lower elevation valley bottom areas in the east are characterized by warm, dry summers and moist, cool winters with moderate snowfall; valley bottoms in the west are similar in summer but have moist, mild winters with little snowfall. At higher elevations, summers are short, cool and moist with long, moist and cold winters [George et al., 2005]. Mean annual precipitation in the period 1961-2005 is 137 cm at the fish hatchery near Slesse Creek, with roughly 50% of the precipitation falling in four months from October to January. Up to 10% of the total annual precipitation may fall as snow in the valley bottoms. At higher elevations, there may be deep snow cover for up to 9 months. The valley floor and lower tributaries are generally designated as Coast Western Hemlock (CWH) in the BC Biogeoclimatic Scheme [Meidinger and Pojar , 1991]. This zone primarily contains coniferous forests, or ‘temperate rainforests’ composed of western hemlock (Tsuga heteropylla) and western red cedar (Thuja plicata). Other species include Pacific silver-fir (Abies amabilis) and Alaska yellow cedar (Chamaecyparis nootkatensis) in wetter areas, and Douglas-fir (Pseudotsuga menziesii), western white pine (Pinus monticola) and bigleaf maple (Acer macrophyllum) in drier areas. Between approximately 900 and 1800 m above sea level, the Mountain Hemlock (MH) ecological zone has mountain hemlock (Tsuga mertensiana) and Pacific silver-fir, with some yellow cedar and sub-alpine fir (Abies lasiocarpa). Active forestry has been ongoing since the 1910s, and by the 1930s, much of the easily accessible flat had been harvested, and operations have extended onto progressively steeper terrain since then [Hay & Co. Consultants, 1992]. Roughly 15% of the forested land base on the Canadian side had been logged by the late 1980s [Jordan, 1990], and George et al. [2005] estimate that 36% of forested land in the valley presently has growth that is younger than 60 years. The headwaters of Chilliwack Valley, including Chilliwack Lake and its tributaries, are within a granitic batholith (Figure 1-4). Radium and Centre Creeks and the headwaters of Nesakwatch, Slesse and Foley Creeks are similarly granitic. Chilliwack Batholith was emplaced between terranes of metamorphosed phyllites and amphibolites in Eocene to Pliocene time. The batholith is the largest pluton in the Cascade arc (960 km2 ), and consists of at least 40 plutons of granite, granodiorite, tonalite and quartz diorite composition (57% to 78% SiO2 , Tepper et al. [1993]). Chilliwack Batholith also contains stocks of gabbro and diorite, as well as rhyolite and dacite of the Hannegan Formation, located in the headwaters of the 8  Chapter 1. Introduction Chilliwack Valley. These are likely Pliocene in age [Richards, 1971; Tabor et al., 2003; Tepper and Kuehner , 2004]. The mid section of the valley consists primarily of Chilliwack Group lithologies [Cairnes, 1944; Monger , 1966; Brown, 1987]. This unit is an intra-oceanic island arc assemblage characterized by an upper volcanic sequence of pyroclastic rocks and lesser basalt, basaltic andesite, and dacite flows built on a clastic package of volcanic litharenite, slate, limestone and chert [Sevigny and Brown, 1989; Tabor et al., 2003]. Rocks have experienced sub-greenschist to greenschist facies metamorphism. Limestone units within the Chilliwack Group are evident in the mid- to lower reaches of the valley, with some notable outcrops across from the outlet of Chipmunk Creek, and near the mouth of Slesse Creek. The Cultus Group underlies much of the lower portion of the valley, as well as portions of Chipmunk and Foley creeks. Cultus rocks consist of pelites, siltstones, fine-grained sandstones and thin limestone beds. A number of other bedrock units crop out in the study area. These are summarized in Figure 1-4.  1.2.2  Regional Studies  The present study follows other Cordilleran studies of large-scale sediment budgets that are inherently represented in valley landforms. Some important recent contributions are those of Brooks [1994] in the Squamish Valley (3850 km2 ), and Jordan and Slaymaker [1991] and Friele et al. [2005] in the Lillooet Valley (3150 km2 ) in the Coast Mountains. Work by Kovanen and Easterbrook [2001] in the Nooksack Valley and Collins et al. [2003] in the Nooksack, Skagit, and nine other major rivers draining the western slopes of the Washington Cascades have also shed light on the large-scale reworking of the landscape following deglaciation. In his study, Brooks concludes that reworked glacial sediments are a considerable component of the historical sediment budget, but not a dominant component in the late Holocene river load of Squamish River. The contrast among these valleys is instructive. Both basins in the Coast Mountains and numerous major valleys in the Cascades are affected by volcanic activity, and thus have experienced major disturbance and sediment inputs from both active and reworked sources [Friele and Clague, 2002]. In particular, studies in Lillooet Valley suggest that repeated volcanic episodes have come to dominate long-term sedimentary processes in the valley. Four landslides in excess of 1 x 106 m3 have occurred in the last century alone [Friele et al., 2005]. 9  0  Figure 1-4: Lithologic assemblages of the study area.  2.5  5  0  10 15 Kilometers  20  (Tomyhoi)  Borden  Indian Creek  Bear Creek  Depot Creek  49°00’ N  2  Brush Creek  Paleface Creek  Skagit Gneiss Complex Cretaceous - Tertiary Banded gneiss, some mafic orthogneiss,mafic migmatite, ultramafic rock, tonalite  Chilliwack Batholith Oligocene - Pliocene Intermediate granodiorite, tonalite to quartz diorite, some gabbro.  Eocene Conglomerate  Easy Creek  Little Chilliwack  Chilliwack Batholith  Hannegan Formation  (Silessia)  Post Creek  Radium  Volcanic Rocks Miocene-Pliocene Andesitic to dacitic breccia, tuff  Yellow Aster Gneiss  Foley Creek  Airplane Creek  Slesse Creek  Chipmunk Creek  Slolicum Schist  Chilliwack Group  Yellow Aster Gneiss pre-Devonian (?) Well-layered pyroxene gneiss, calc-silicate gneiss and associated marble and meta-igneous rocks. Some quartz-rich tonalites.  Liumchen Creek  Tamihi Creek  Cultus Formation  Slollicum Schist Cretaceous - Tertiary Greenschist-grade mafic to intermediate volcanics, phyllite, minor conglomerate.  Shuksan Amphibolites Cretaceous - Tertiary Mid-amphibolite to andalusitegrade metamorphic rocks  Shuksan Amphibolites  e ntr Ce ek h Cre atc kw a s k Ne ree C  Nooksack Formation Jurassic-Cretaceous Argillite, sandstone, conglomerate, dacitic volcanic rocks  Cultus Formation Late Triassic to Late Jurassic Fine-grained sandstone Tuffaceous siltstone Limestone beds and lenses Dacitic tuff and flows  Calcareous Rocks  Clastics: Phyllite, sandstone, semischist, greenstone or greenschist, minor chert, conglomerate  Volcanics: Calc-alkalic island arc, predominantly mafic volcanic flows and breccias.  Chilliwack Group Silurian - Permian  Chapter 1. Introduction  10  Chapter 1. Introduction The collapse of volcano flanks in the valley at 8700 and 4400 years BP is estimated to have delivered 600 x 106 m3 of material to the valley. In the Squamish Valley, a debris avalanche from Mount Cayley (4.8 ka BP) instantaneously delivered 200 x 106 m3 of material. Brooks points out that this is more material than all of the reworked glacial sediment from any given Squamish River tributary. Collins et al. [2003] state: Voluminous lahars from eruptions of Glacier Peak volcano inundated the Cascade drainages of the Sauk, Skagit, and Stillaguamish Rivers (Beget, 1982); remnants of lahar deposits since incised by fluvial erosion can be found in each of these three valleys. At least 60 Holocene lahars moved down valleys heading on Mt. Rainier (Hoblitt et al. 1988). Mt. Rainier’s National Lahar travelled from the Nisqually River to Puget Sound less than 2200 years BP. In comparison, the Chilliwack drainage appears to have been relatively quiescent throughout the Holocene. Although very large bedrock failures have been documented (some as large as 1.8 x 106 m3 ; Thomson [1999]), these rare deliveries are not as mobile within the drainage network as the lahars and landslides described above. Given, then, the relatively subdued pace of geomorphic processes throughout the Holocene, the valley provides an important opportunity to observe the course of relaxation from glacial disturbance with minimal largescale overprinting by other perturbations. Accordingly, this thesis will explore some of the likely scenarios for the subsequent evacuation of glacial sediments, given the evidence that remains.  1.3  Thesis Structure  In order to address the question of sediment transfer on the Holocene time scale, it is essential to develop a quantitative framework that bounds the system through time and space. This is achieved by analyzing the geometry, mass, provenance and chronology of the deposits along the length of the system. Chronology is established using radiocarbon dating, providing an important time-frame for the deposition of valley landforms. While dating evidence is relatively sparse, further insights have been gained by applying numerical models to evaluate potential rates of sediment mobilization throughout the Holocene Epoch using data from the various study components. These different avenues of investigation can provide important checks on estimates of sediment yield and evacuation of stored glacial material within the system. 11  Chapter 1. Introduction Chapter Two examines the network structure of the valley, and estimates volumetric erosion and deposition of glacial and hillslope material within the postglacial time frame. Approximate volumes of erosion and deposition are calculated from air photos, digital elevation models and land surveys. Chapter Three provides a detailed sediment budget at Chilliwack Lake for the postglacial period. The bounding volume of the lake bed layers and deltas provide an integrated picture of sediment yield at the scale of 5th to 7th order catchments within the valley network. The history of deposition in the lake is constrained using evidence from paleomagnetism and radiocarbon dating. Chapter Four presents architectural and volumetric analysis of major landforms in the lower valley. Isopachs of erosion are generated for the major landforms, allowing estimation of the long term yield to the lower valley and Vedder Fan. Large glacial deposits such as Ryder Lake Upland and Larson’s Bench are the dominant components of the postglacial sediment budget. The geometry of Vedder Fan provides some constraints on summary yield from the catchment. Chapter Five looks at the geochemical, lithological and sedimentological character of sediments in source areas and along the length of the mainstem. The grain size distributions of major landforms and active channel sediments along the drainage network are characterized to further establish the pathways of sediment transport. Important landform elements have been sampled in order to determine the relative sequestration or transfer of various grain-size fractions. Source ascription and rates of mixing for fine sediments are presented, for both fine and coarse sediments. Evidence from geochemical analyses and patterns of grain lithology along the mainstem are used to supplement the mass balance calculations. Chapter Six integrates the above information in order to develop a temporal and spatial picture of sediment transport within the valley mainstem. Using conventional hydraulics and sediment transport equations, it is possible to model the passage of the paraglacial sediment wave using a simple 1D finite-difference framework. Certain boundary conditions are assumed, and the sensitivity of those assumptions and other parameters is examined.  12  Chapter 2 Hillslope and Tributary Sediment Stores  2.1  Introduction  This chapter examines the volumetric balance of erosion and deposition of material from Holocene weathering processes, as well as from primary (outwash and morainal material) and secondary (colluvium, paraglacial fans and valley fill) remobilization of glacial sediment stores. The thesis seeks to develop the post-glacial sediment budget for Chilliwack Valley moving from the hillslope scale to the broad river reaches of the lower valley. The aim of this chapter is to estimate the long term yield from the major tributaries (5th and 6th order) to the mainstem. Following a summary of methods and potential errors in the analysis, the chapter is broken into three major topics: (1) the structure of the valley network, (2) sediment mobilization and deposition at the hillslope scale, out to the scale of the major tributaries, and (3) estimation of the volume of glacial fills in the lower reaches of the major tributaries. Many of the landforms within the upper drainage network of Chilliwack Valley retain aspects of their early Holocene, post-glacial configuration allowing some assessment of the total volume of sediment evacuated over time. Steep drift-mantled slopes have delivered abundant material to the lower slopes and valleys immediately following deglaciation. Till has been incised and mobilized as debris flows and debris slides over the postglacial period. Various authors have proposed that valley-side till may have been eroded, redeposited and stabilized within a timescale of decades or centuries rather than millennia (Eyles et al. [1988]; Evans and Clague [1988]; Jackson et al. [1989]; Ballantyne and Benn [1996]; Harrison [1996]; Curry [2000]). Diffusive erosional processes have subsequently smoothed topographic breaks and lowered landform surfaces [Putkonen and O’Neal, 2005]. Much of the mass eroded from hillslopes and glacial landforms, particularly the coarser 13  Chapter 2. Hillslope and Tributary Sediment Stores  Erosion (Output) Deposition (Input) Post-Glacial Surface - V0  Valley Fill  Bedrock  Figure 2-1: Definition of the reference post-glacial surface. Material deposited on top of the postglacial surface constitutes an input to the post-glacial budget, material eroded from the valley wall and glacial fill are considered to be outputs. The reference postglacial topography is V0 .  sediment fractions, remains within the valley network, modified and remobilized occasionally by mass wasting and by the fluvial system. Rates of sediment sorting and transfer are dictated by the relative position of the deposit within the network hierarchy: erosive capacity is controlled by the total upstream area and local slope, which may be tied to the history of base-level fall in the downstream catchment. Headward migration of degradational trends appears to be a strong controlling factor in the erosion of glacial and paraglacial landforms in many Chilliwack tributaries. The proposed methodology for analyzing the valley sediment budget is to construct a postulated pre-erosional (i.e. end-Pleistocene) surface and to calculate erosion from, and deposition upon, such a surface (Figure 2-1). A similar approach has been used in other studies in which a simple and well-dated initial surface is available (notably Ollier and Brown [1971]; Gorler and Giese [1978]; Ibbeken and Schleyer [1991]). Given the reasonably good understanding of the end-glacial topography in Chilliwack Valley, this technique holds some promise for unravelling the magnitude of erosional and depositional volumes within the Holocene Epoch. Where possible, maximum and minimum estimates are generated in order to assess the magnitude of potential errors.  14  Chapter 2. Hillslope and Tributary Sediment Stores  2.2  Data Sources and Associated Errors  Most of the analysis of the sediment drainage network and landforms was done using the TRIM (British Columbia Terrain Resource Information Management [1996] digital basemap) provincial dataset (sheets 092H001-004 and 092H011-013) and United States Geological Survey 10 m DEMs. Calculations of erosion and deposition of landforms was mainly carried out using the TRIM and USGS DEMs of the valley topography, supplemented with some topographic field surveys and air photo measurements. The drainage network was extracted from the DEM dataset using ESRI’s ArcGIS hydrological toolbox. Calculation of terrain curvature was carried out using the Landserf package developed by Wood [2004]. The TRIM data are intended for work at 1:20 000 scale. Elevation points are 50-75 m apart, though 3D breakline vectors are more detailed in areas of high relief and complex topography. The suggested grid resolution of a digital elevation model from the point and breakline data is 25 m. Terrain mapping by J.M. Ryder and Associates [1995] was used to assist in the delineation of glacial deposits on the Canadian side of the border. This dataset is also intended for use at the 1:20 000 scale, and was not intended for high-resolution interpretation of landforms. Higher resolution (ca. 1:10 000) mapping by Rode [unpublished] and Thomson [unpublished] was consulted for the Ryder Lake Upland and Foley Creek, respectively. An inventory of over 1 100 landslides and other features such as bedrock failures was compiled using a single-year (1996) set of air photos [EBA Engineering 2002]. This provided a GIS-ready dataset that furnished information on the types and spatial distribution of mass failures. Measurement errors are a critical part of the analysis, yet the true nature of the variance in measurements is difficult to ascertain. Quantities measured are areas and volumes, as well as slopes of hillsides and fan deposits. Depths observed in the field at exposures can be easily measured to within ± 2-3%, as can surface area, subject to correct interpretation of the deposit stratigraphy. However, in the absence of natural exposures or geophysical soundings, estimates of deposit depths often require judgement and guesswork.  2.2.1  The Magnitude of Error  Errors involved in DEM analysis can be divided into at least three major categories: (1) DEM spot elevation and breakline errors, (2) surface interpolation and processing errors and (3) errors of geomorphic interpretation.  15  Chapter 2. Hillslope and Tributary Sediment Stores Spot elevations collected by semi-automated photogrammetric methods are subject to a number of errors, which may be exacerbated by factors such as micro-relief, dense vegetation, shading or sloping ground. The accuracy specification of the TRIM product for spot elevation points is a Linear Mapping Accuracy Standard (LMAS) for the elevation of less than 5 metres, corresponding to a maximum RMSE of 3 metres [GDBC , 1992]. This likely varies over the landscape, depending on the overall roughness of the terrain. Given the high spatial autocorrelation of points in elevation datasets [Liu and Jezek , 1999] the error term may be somewhat less than 3 m. Some systematic error is evident in the study area DEM, most notably ‘stripes’ associated with photogrammetric lines [Albani and Klinkenberg, 2003] that impart a characteristic herringbone pattern to the topography. Spot elevations (mass points) and breaklines from the TRIM dataset were used to construct digital elevation models. A great variety of methods is available for surface interpolation, and there is no clear consensus on the best methods for mountainous terrain [Fisher and Tate, 2006]. Estimates of valley landform volumes consist of subtracting a DEM of the modern landscape from an approximation of the post-glacial landscape. Some error resides within the modern DEM, however, in most cases the larger error - which is, moreover, not exactly specifiable - will be in the coordinates of the posited post-glacial surface. Taylor [1997] shows the propagation of error in the general case of a parameter q derived from variables x, ..., z measured with uncertainties ∂x, ..., ∂z. Provided the uncertainties in x, ..., z are random and independent, the uncertainty in q may be calculated as:  dq =  ∂q dx ∂x  2  + ... +  ∂q dz ∂z  2  (2.1)  Many of the errors are random, and are likely compensating. The composite uncertainty terms for deposit area and depth are calculated for each landform and propagated in quadrature along the drainage network. The total error associated with volumetric estimates of each landform is determined by: Et =  Ez 2 + Eplan 2 + Eintrp 2  (2.2)  where Ez 2 is the error incurred in deriving the vertical bounds of the deposit, Eplan 2 is the error related to delimiting the planform boundaries, and Ez 2 is a composite error term related to errors of geomorphic interpretation. Volumetric estimates of landforms are most sensitive to the estimation of depth z of 16  Chapter 2. Hillslope and Tributary Sediment Stores aggraded deposits such as alluvial fans and valley fills. The error term is treated as random, however there does tend to be some systematic bias toward large estimates of deposit depth and thus higher volumes. The bedrock boundaries are assumed to follow the curvature of the hillslope above. In the case of fans and debris cones, the base of the deposit at the distal toe is interpreted to be the lowest point in the adjacent channel. Eroded valley fills have been reconstructed as longitudinally curved, laterally planar surfaces. There is relatively less error in this case, since the reconstructed surface represents the average elevation of the deposit. In some cases, particularly where valley walls show a record of mass flow deposits and the elevation of the former alluvial surface is uncertain, a larger uncertainty term is assigned. Generally, given clear evidence of the former surface, the volumetric error for valley fills is approximately ±5%. Errors of interpretation are consequences of the many assumptions regarding the postglacial landscape that must be made. These may include assessments of the geometry of the initial deposit, interpretation of a stratigraphic sequence, and/or inferences of postdepositional reworking. A further problem is the attribution of older tills and drift sequences in the valley to a younger event. Obviously many of these errors are difficult to assign outright, and must be applied with judgement. Planimetric area (x, y) of landforms can be established with an estimated precision of ±2%. The depth of the deposit may be determined with a precision of roughly ±10%. Smaller cones or fans are delineated by as few as 12-15 topographic points. The relative volumetric error in this case is potentially greater than for larger features. Given additional potential errors in defining the lower bounding geometry and interpretation of the extents, estimation of fan volumes typically carry an error of about 20%.  2.3  Network Structure and Process Domains  The spectrum of channel types within the tributary valleys ranges from steep first-order gullies with no channel bifurcation to larger fluvial basins with an evolved, branched network (Figure 2-2). Figure 2-3a shows examples of steep, linear hillslope gullies that deliver material directly to the alluvial floor of a 6th order channel. Sediment is stored at breaks in slope at tributary junctions, leading to higher apparent rates of yield from the lower-order catchments, whereas more storage occurs within larger basins having a developed fluvial system. Many of the larger, 6th and 7th order fluvial basins have an elongate, trellis form (Centre,  17  5  10  15  (Tomyhoi) (Silessia)  Depot Creek  49°00’ N  Indian Creek  Bear Creek  Paleface Creek  Brush Creek  Chilliwack Lake  Easy Creek  Little Chilliwack  Radium  Post Creek  Figure 2-2: Chilliwack River Valley.  20 km  Slesse Creek  Pierce Creek  e n tr Ce ek h Cre a tc w k sa k Ne ree C  0  (Damfino)  Tamihi Creek  121°50’ W  Borden  Chipmunk Creek  Foley Creek  121°30’ W  Liumchen Creek  Vedder Crossing  Airplane Creek  Chapter 2. Hillslope and Tributary Sediment Stores  18  Chapter 2. Hillslope and Tributary Sediment Stores  3rd - Order ‘Convergent’ Basin  1st - Order Hillslope Gullies  5th - Order Fluvial Basin  Figure 2-3: Three different basin scales. The left side of (a) shows linear, first- and second-order debris flow channels, next to a 3rd order alpine colluvial basin, in the headwaters of the Slesse basin. Grid spacing is 500 m. (b) 5th-order Bear Creek, with a developed fluvial network (grid spacing is 1 km). Note the relatively small outlet fan in the latter.  Nesakwatch, Slesse, Depot), while others exhibit a more dendritic form (Paleface, Liumchen, and Tamihi). The trellis form lends itself to a particular regime of sediment transport. Stream power grows incrementally along the tributary length, as opposed to a more punctuated regime in a dendritic configuration (cf. Strahler [1964]; Gregory and Walling [1976]; Benda et al. [2004]). Material is discharged periodically from weathered bedrock exposures, gullies and zero- to second-order basins along the master channel. With delivery from the lowerorder systems to a channel several order higher, there is an abrupt change in channel slope which leads to deposition on the footslopes. There may also be a greater degree of de-coupling between hillslopes and the main channel where a floodplain has developed in the higher order channel. Network topology for Chilliwack Valley was extracted from the BC TRIM DEM using Tarboton’s [1991] suite of ArcGIS tools, TauDEM. Links are segments of the channel network between junctions. Link position is specified by Horton’s ordering system, as modified by Strahler [Strahler , 1952, 1957]. Optimum resolution of the valley network was achieved with  19  Chapter 2. Hillslope and Tributary Sediment Stores 10  10  Chilliwack Valley, n=404 1 026  9  10  1.5274  y = 5436 e 2 R = 0.9993  8  10  Upstream Area (m2)  Distal fluvial (Glacial trough)  1 311 2 343  3 169 5 641  7  10  Colluvial (Valley step)  11 327 6  10  Hanging Fluvial  n = 25 222  5  10  Colluvial (Cirque walls)  All Chilliwack Network Links Mean Link Area by Order  4  10  Hillslope 3  10  0  1  2  3  4  5  6  7  8  9  10  Link Strahler Order  Figure 2-4: Histograms showing the distribution of upstream area within valley network links, from headmost channels to the Chilliwack River at Vedder Crossing. The mean area for each order class is shown with a cross (+). The process domains identified by Brardinoni and Hassan [2006] are overlaid in grey. The total number of links for each order is indicated.  Chilliwack River at Vedder Crossing as an 8th-order channel. On 1:50 000 scale NTS maps, Chilliwack River is a 5th order channel, so the network is specified in greater detail. Horton [1945] identified a number of important scaling relationships among quantities such as the number of streams, length of streams, and slope of streams within a drainage network. Schumm [1956] refined a similar relationship for catchment area. He noted that the scaling among link areas for successive orders is approximately constant. This is shown for Chilliwack Valley in Figure 2-4. There is a systematic change in upstream catchment area relative to the position of each network link within the drainage network. Within the lower order links, catchment size is lognormally distributed across roughly two orders of area magnitude, reducing to one order of magnitude at 7th order. The log-normal distribution appears to break down in higher orders, presumably due to the greatly decreasing sample size. A regression through the mean basin area points indicates that scaling is quite consistent across five orders of area magnitude. Given the systematic change in channel slope and upstream contributing area (thus chan20  Chapter 2. Hillslope and Tributary Sediment Stores nel flow) with scale, there are distinctive process domains within each hierarchic level of the network. Such process domains were initially proposed by Montgomery and FoufoulaGeorgiou [1993], and have subsequently been amended to included formerly glaciated landscapes by Brardinoni and Hassan [2006]. The characteristic area scales for each of the process domains have been superimposed on the diagram (Figure 2-4). It is difficult to assign definitive process specifications or provide precise bounding spatial scales across such a diverse landscape; however, Brardinoni’s generalization of how sedimentary processes are conditioned by the large-scale imprint of glaciation on the landscape does hold up well in Chilliwack Valley.  2.4  Sediment Deposition in Lower-Order Catchments  Colluvial and alluvial deposits have been identified throughout Chilliwack Valley, using air photos, terrain maps and TRIM data. In the following analysis, the depositional volume in fans that are judged to be transport-limited and/or decoupled from their master channel is taken as an estimate of coarse sediment yield over postglacial time. These apparent yields are compared among catchment order and area, in order to estimate a bounding envelope to represent the magnitude of the total bedload/mass wasting yield from upland basins. Many deposits have a multi-phase history. Many fans in Chilliwack Valley are polygenetic, built by a variety of active processes, ranging through debris flows, snow avalanches, and fluvial transport. In some catchments the sediment and flow regimes have changed dramatically, leaving a smaller, low-gradient modern fan inset into remnants of a high-gradient paraglacial fan. Ryder [1971] notes that the rate at which fan dissection proceeds is likely to depend upon the erosive capacity of the tributary stream. Fan volumes and geometry are thus indicators of the colluvial/alluvial reservoir state and history, essentially, the equilibrium achieved between hillslope sediment yield and transport efficiency within the fluvial network. Alluvial reaches that are degrading will carry away their sediment stores, leaving smaller, incised fans. In Chilliwack Valley there are several states found: • Transport-limited basins with coarse-grained alpine fans: exposed bedrock faces in the upper basins have enhanced rates of weathering and sediment delivery. These fans are mostly decoupled from the transport system, though some cutting of the fan toe may occur close to the higher-order receiving channel. Accumulation rates were probably quite high in the late-glacial period, as glacial ablation debuttressed many steep slopes. 21  Chapter 2. Hillslope and Tributary Sediment Stores Rates have probably been significantly lower over the course of the Holocene Epoch. Fans tend to have a steep, straight to concave slope profile. A number of these features grade into debris or alluvial cones. • Aggraded paraglacial mixed-transport fans: high-gradient fans built by both fluvial and mass wasting processes that have aggraded throughout the Holocene Epoch. The system is generally transport-limited, emptying into a trunk channel that is still choked with glacigenic material. Nesakwatch and Centre Creeks and some headwater basins have a number of examples. There are also many examples of such fans emptying into lower-gradient hanging valleys - the ‘sink-colluvial’ domain identified by Brardinoni and Hassan [2006]. • Degraded paraglacial mixed-transport fans: Since lowering of base level in tributaries, much of the glacial valley fill in the mid to lower tributaries has been evacuated. Fluvial channels have trenched the fan-head and carved away at the base of the fan, sometimes leaving only a few terraced remnants. A modern fan usually develops in the incised remains. Slesse and Tamihi have some examples of relict fans that have been cut away by the mainstem. Some valleys have had renewed episodes of fan aggradation after recent disturbances, most notably logging [Millar , 2000]. • Fluvial fans: outlets of higher-order basins have well-established fluvial fans, which are likely influenced by debris-flow events, but these are not a major construction agent. In some cases there may be relatively little inheritance from early Holocene paraglacial processes. Fan profiles are typically flat to convex. There are numerous intermediate states; the interplay between basin size, order, slope and geometry, and characteristic grain size determine the balance achieved. 135 relatively intact fans have been identified throughout the valley, which incorporate both remobilized glacial material and Holocene weathering products, and thus potentially indicate relatively high rates of deposition over time compared to non-glaciated regions. Rates of yield have diminished over post-glacial time, and patterns of trees and shrub vegetation on fans or lichens on talus slopes highlight the large inactive zones on many of the deposits. Finer sediments have moved onward as washload and suspended load. The volume of each fan has been approximated using CAD geometric shapes (Figure 2-5). The methodology is based on the technique used by Campbell and Church [2003] in 22  Chapter 2. Hillslope and Tributary Sediment Stores  Figure 2-5: Diagrams illustrating estimation of fan deposits using the technique of Campbell and Church [2003] and using CAD shapes.  Lynn Valley, Southern Coast Mountains. An improvement afforded by CAD shapes is the ability to refine curvature to match the fan’s frontal slope, to adjust to compound slopes on the bounding valley wall, and to construct complex, coalescing fans. Accumulations in talus slopes were generally assumed to have a triangular prismatic shape. The cross sectional geometry was interpolated over a linear or curved path along the lower slope. The largest source of error remains the lower bounding surface, which is usually approximated from the valley wall geometry and the height of the valley floor. It is possible that more volume is obscured below as the valley walls may not be planar, and the deposit may interfinger with floodplain and valley fill. Figure 2-6 displays the relation between fan volume and upstream catchment surface area. Points are distinguished by basin order. Each landform ostensibly represents 13 000 years of coarse material accumulation on the lower slopes of valley walls. It is assumed that finer material moves onwards, and there is variable preservation of fine material in the fans. The relation exhibits considerable scatter which can be attributed to measurement error, as well as a complex combination of geologic and morphometric variables. Grouping the basins by geology or by slope class explains only a relatively small proportion of the fan-size variability. Variables related to hypsometry and vegetative cover presumably must also play a role, due to the accelerated weathering in exposed environments at higher elevations. There is a scale-dependent transition from catchments with steep, linear morphology to drainage basins with greater bifurcation and thus sediment storage; this results in two distinct patterns in the dataset (Figure 2-6). The smaller, steeper low-order basins are able to funnel 23  Chapter 2. Hillslope and Tributary Sediment Stores 1E+9 y = 0.015x  Estimated Deposit Volume (m 3)  2nd Order  1.5  y = 5E-05x  1.5  3rd Order  1E+8  5  4th Order 3  5th Order  4  6th Order  1E+7 7th Order 1 2  1E+6  1: Pierce Creek 2: Brush Creek 3: Paleface Creek 4: Depot Creek 5: Upper Chilliwack River  1E+5  1E+4 1E+4  1E+5  1E+6  1E+7  1E+8  1E+9  Upstream Catchment Area (m 2)  Figure 2-6: Estimated Holocene bulk sediment deposition (sand and coarser) plotted against contributing catchment area. Catchments are coloured according to their Strahler order. Two reference lines are shown, indicating that the envelope of maximum deposition scales approximately to the power of 1.5. A similar exponent is used for scaling the size of fans and deltas from larger fluvial basins.  and accelerate the transit of water and debris, whereas slope breaks within higher order catchments presumably tend to promote shorter step lengths relative to the drainage basin size, thus more sorting and storage occurs. Points 3,4 and 5 in Figure 2-6 indicate end-point sedimentation in Chilliwack Lake. Pierce Creek is a steep 5th order basin (6.8 km2 in area (Figure 2-2) near the Slesse Creek confluence, that has deposited most of its post-glacial load in a conical fan on a terraced glaciofluvial plain 50 m above the mainstem. The total fan volume is approximately 2.3 x106 m3 . In a single debris flow event, the creek delivered roughly 63 000 m3 of material, with a recurrence interval estimated to be in the range of 50-100 years Thomson [1999]. The source area of this debris flow was in the steep lower portion of the basin; the rest of the basin is otherwise relatively stable and there is considerable storage of material. Examining Figure 2-6 (‘1’), Pierce Creek sits closer to the fluvial basin trend, with more subdued rates of sediment delivery over the long term. Information on the total Holocene yield from upland catchments in Cordilleran British Columbia is relatively scarce. There are, however, a number of data sets that emerge from 24  Chapter 2. Hillslope and Tributary Sediment Stores  Figure 2-7: Pierce Creek, July 2003. Source area of the 1996 event is shown with an arrow  modern, decadal-scale studies with which to make comparisons. Many of these studies have focussed on the relatively softer, volcanic terrain, which may have yielded rates an order of magnitude greater than the metamorphic and granitic terrain considered here. In the Lillooet Valley, roughly 150 km to the North in the Southern Coast Mountains, Jordan and Slaymaker [1991] estimated an average magnitude and frequency of debris flows from basins as follows. For “small” debris flows, an average yield of 20 000 m3 every 10 years is assumed, 25% of which is delivered onward to the channel. For “large” debris flows, an average yield of 50 000 m3 , at the same frequency was used. At several sites in the southwestern British Columbia, including Lillooet, Squamish, and Lower Fraser valleys, Jakob and Bovis [1996] estimated volumetric output from 34 notable debris-flow basins, ranging in size up to 15 km2 . The magnitude of sediment delivery was correlated with a number of morphometric characteristics. The authors noted that most debris flows occur with magnitudes in the range of 25 000 to 30 000 m3 for a single event - with large flows ranging up to 200 000 m3 . Smaller events may not be detected in the dendrochronological and depositional record. Recurrence intervals were commonly less than 10 years. Another data set with which we can compare is that of Kirchner et al. [2001], Figure 2-8. Their database provides estimates of fan accumulation over the course of up to 30 ka in the unglaciated Idaho Batholith. The long-term estimates of averaged annual accumulation 25  Chapter 2. Hillslope and Tributary Sediment Stores  Specific Yield (T/km2/a-1)  104  103  Vedder Fan  102  101  100 104  105  106  107  108  109  1010  Upstream Area (m 2)  Figure 2-8: Data from Kirchner et al. [2001] (black diamonds) with data from the Chilliwack Valley overlain (open circles). Sediment accumulation estimates from Chilliwack Valley display more scatter, but they are largely consistent with the long-term average rates of specific sediment yield calculated by Kirchner et al. [2001].  in Chilliwack Valley seem to be in line with the pattern in Kirchner’s dataset, although the effects of glaciation in the Chilliwack dataset probably result in higher apparent rates.  2.5 2.5.1  Sediment Source Areas Large Bedrock Failures  Rocks of the Chilliwack Group and Cultus Formation are noted as being somewhat more susceptible to large-scale mass failure than are crystalline lithologies, but this is not a highly significant trend [Thomson, 1999]. The Chilliwack Valley does not have large outcrops of tuffs, breccias and other volcanic lithologies that have often been implicated as indicators of enhanced debris flow activity [Rood , 1984; Jordan, 1994; Jakob and Bovis, 1996; Jakob, 2005]. Some of the larger failures in the basin, such as the Pierce Creek landslide, are related to seams of limestone running through the Chilliwack Group. Others are related to planes of weakness found within phyllitic rocks. 26  Chapter 2. Hillslope and Tributary Sediment Stores The nature of tectonic building by thrust faulting in the region has led to some degree of asymmetry in valley sectional profiles [Groulx , 1993]. The southwest-facing sides of the Airplane, Lower Chipmunk, Nesakwatch and Centre Creek valleys are part of a thick, tabular body of amphibolite rock that conformably overlies sedimentary strata [Monger , 1966]. The hanging wall of this body is more susceptible to erosion, which appears to have led to oversteepening over the course of glaciation and enhanced gully dissection over time. This may account for higher total rates of erosion in Airplane, Foley, Nesakwatch and Centre Creeks. Twenty two historic bedrock failure zones are described by Thomson [1999] and EBA [2002] in the Canadian portion of the study area, totalling 4.5 km2 , for an overall density of roughly 0.6 percent. Rates of delivery from deep-seated bedrock sliding in Chilliwack Valley are lower than in the neighbouring Nooksack drainage, where Easterbrook et al. [2007] documented at least 13 large events that have occurred in postglacial time. Several of the landslides cover 10 km2 , and the largest (the Church Mountain Slide) delivered an estimated 2.83 × 108 m3 at approximately 2 500 BP [Carpenter , 1993]. In Chilliwack Valley, some of the larger source areas are roughly 0.5 km2 , with historical deliveries on the order of several tens of millions of cubic metres. Onward mobilization within the fluvial network in the post-glacial era is generally quite limited, however, there is often a profound impact on patterns of sediment transport and channel configuration. Foley Creek is a notable example, with a mid-valley lake (0.1 km2 in area) created behind the debris of a large bedrock slide. A lake delta and a “forced” alluvial channel [Montgomery and Buffington, 1997] have been created upstream of the lake.  2.5.2  Gullies and Diffusive Slope Processes  In order to match the observed patterns of detrital deposition in the upland basins with likely source areas, gross rates of mobilization were estimated throughout the catchment. The hillslope areas were divided into three major type areas (gullies, soil-mantled, and bedrock terrain) where the active geomorphic processes and regime of sediment delivery are likely to be different. Bedrock areas were distinguished from vegetated zones using Landsat spectral imagery; gullied zones were identified using measures of planform concavity extracted from the TRIM DEM. Gullied terrain is found primarily within the soil mantled hillslopes of the valley, but also includes portions of the bedrock zone. It is assumed that a majority of the volume evacuated from lower post-glacial hillslopes is represented by these planform concavities, often carved 27  Chapter 2. Hillslope and Tributary Sediment Stores  Vegetated Open Slopes Bedrock Open Slopes Convergent Gullies  6%  Sc = Critical Slope  Percent Plan Area  8%  4%  2%  0 0  0.5  1  1.5  2  Gradient  Figure 2-9: Histogram of gradient distributions in gullied topography and on planar to convex hillslopes (vegetated and bedrock) in the first to fifth order catchments of Chilliwack Valley. A limiting gradient for the vegetated open slopes is estimated to be 1.4  into till blankets, colluvium on steep slopes and into jointed, weathered bedrock. The second two zones are planar to divergent hillslopes, which are less active than gullied areas, but occupy a considerably greater proportion of the landscape. Gullied terrain accounts for roughly 14% of the valley hillslope area, the bedrock zone accounts for 16%, and vegetated slopes 70%. On soil-mantled hillslopes, both slow, quasi-continuous diffusive processes and more episodic shallow failures deliver sediment to the drainage network. The slow diffusive processes include creep, slope wash, tree-throw and bioturbation. Figure 2-9 shows the distribution of slopes within the three zones: convergent gullies, and convex-to-planar vegetated and bedrock terrain. The vegetated zone has only a very small fraction of its area on slopes greater than 1.4 (54.5◦ ), while gullied topography and open-slope bedrock zones both extend into a steeper range. This would suggest that there is a threshold slope within the soil-mantled zone beyond which much greater rates of failure tend to occur. This is explored further below; the behaviour is consistent with a number of non-linear models of hillslope evolution [Howard , 1994; Roering et al., 1999; Martin, 2000; Montgomery and Brandon, 2002; Dietrich et al., 2003], and this type of model is used here to approximate rates of flux from the planar slopes within the valley.  2.5.3  Sediment Source Areas: Surficial Materials and Gullied Terrain  Blankets of morainal material and colluvium cover most of the mid- to lower slopes in the major Chilliwack Valley tributaries. During alpine glaciation, ice lobes descended from cirque 28  Chapter 2. Hillslope and Tributary Sediment Stores basins and moved out to fill the 5th to 8th order valleys, building moraines at their lateral extents. Lateral gullies and debris flow basins delivered sediment to the sides of the glacier. Many of the deposits on valley walls have been gullied and incised, remobilizing relict glacial material to the lower footslopes and to the fluvial system. Gullied sections of these deposits reveal quite variable depths; road cuts show depths of up to several tens of metres. Gullies on steep valley walls and interfluves tend to have simple, linear geometry with limited branching in the upper reaches. Hillslope gully cross-sectional geometry is typically a ‘V’ shape with vertical depths typically on the order of 5 to 30 m and widths of roughly 10 to 60 m (Figure 2-10a). These planform concavities are regions of groundwater and overland flow convergence that promote the most effective sediment mobilisation. Exposures of incised materials in gullies reveal more than uniform morainal drift. An example from a second-order tributary is shown in Figure 2-10(b) and (c). Remnants of similar stratigraphy at about the same elevation (lower to mid-slope) are found throughout neighbouring tributaries. Significant deposition and ponding against ice occurred over the course of glaciation, leaving thick deposits of glaciolacustrine material, as well as a chaotic mix of colluvium, slope wash and fluvial deposits. As the ice pulled away, rapid incision likely occurred in the steeper reaches of each gully. Stratigraphy is also visible in the active headscarps of failures throughout the watershed. The headwall sections of some headscarps show a compacted diamicton, with subrounded to angular, glacially abraded material. A number of headscarps examined show an upper, dry weathered horizon and a wet lower section (Figure 2-11). There may be preferential failure of finer grained materials that retain moisture better than coarser grained deposits which are better drained (B. Thomson, pers. comm.). There is significant weathering of these materials as well, gradually reducing stability over time. Materials fail by headward retreat, with episodic transport of the sediment mass over time. Material from slope wash and creep may fill declivities, gradually building up over time. This material may be released as a debris flow following the occurrence of major rainstorm event, once a critical limit of accumulation has been reached. EBA [2002] conducted a study of mass wasting in the Chilliwack Valley, identifying over 1 100 modern and historical slides. This inventory provides a window of landslide activity that is visible on the landscape in 1996 aerial photography. The age of the slides has been roughly estimated, and the inventory includes historically reactivated features such as large, chronic bedrock failure zones.  29  Chapter 2. Hillslope and Tributary Sediment Stores  a)  c) C  S  S  G  448m  Top of sequence, weakly bedded Clasts up to 30mm. Sandy matrix with marginal soil development. Clay layer with drop stones 25mm+ Colluvial diamict. Poorly sorted with clasts 256mm+. Mostly subrounded/subangular meta-seds. Some fluvial bedding toward lower contact  447  446 Chaotic, unsorted sequence of highly weathered material overlain unconformably by clay layer. Abundant evidence of Fe leaching through profile from moisture in upper layer. Bedding dipping gently upwards, down-valley  445  Sand and pebble fluvial units very well sorted. Intermitted band of clay along the section.  444  Bedded fluvial gravels and silt. Rounded clasts 50mm+  b) 443  Silt lense (5-10cm) dipping steeply downvalley.  442  Assorted fluvial bedding particles up to 150mm. Bed of silt with wavy and broken bands of interlensing sand. Stones up to 15mm  441  440 ?  ? ? ?  ?  ?  Coarse diamict with boulders up to 300mm+. No apparent bedding, same pattern of Fe leaching as above - moisture emerging from overlying silt layer. Highly weathered.  Figure 2-10: Photos of (a) Holocene channel incision and (b) exposed sidewall deposits of till and glacio-lacustrine material. (c) is a stratigraphic section from a left-bank tributary to Lower Slesse Creek. Total incision is over 25 m, but the exposed 10 m section gives some idea of the complex nature of the initial (Pleistocene) hillside stratigraphy. Some similar sections were found in nearby tributaries, but exposure of such a stratigraphic record is uncommon.  30  Chapter 2. Hillslope and Tributary Sediment Stores  Upper, Oxidized Layer Saturated Zone  Figure 2-11: Photo and figure of a landslide headscarp at Foley Creek. This is the initiation point of a debris flow gully, incised into a sandy, compacted till deposit.  Most shallow landslides occur within the convergent, gullied topography described above. The bounds of such areas were delineated using indices of planform convergence such as may be generated using tools such as LandSerf [Wood , 2004] and ArcGIS. In Landserf, a 60 x 60 m window was used to identify zones with negative plan-form curvature (i.e. concavity). A threshold value was then found (-0.25) to extract the gully boundaries - spurious depressions were also picked up, but these could be reduced using a low-pass or majority filter. 90% of the debris-flow tracks identified in the EBA database are found within these zones. The channel within these gullied zones is steep, and the gullies extend only for short distances, generally less than a kilometre. Figure 2-12 is a plot of upstream contributing area versus slope for all of the gullied (convergent) topography within Chilliwack Valley. Most of the points cluster near a gradient of 0.7 (35◦ ), with lower slopes for larger gully basin areas. Most of the material on slopes steeper than 1.2 to 1.5 (50◦ - 56◦ ) has already failed. Headscarp points of historic debris-flow gullies identified in the EBA database are overlain, and would appear to occupy roughly similar space in the landscape. These slopes were interpolated from a 25 m DEM and are thus approximate. Gullies and incised terrain in the original DEM were ‘filled’ by interpolating across the boundaries of the zones of planform convergence. The resultant smoothed topography consti31  Chapter 2. Hillslope and Tributary Sediment Stores  0  Gradient (m/m)  10  -1  10  10  2  10  3  10  4  10  5  10  6  2  Upstream Catchment Area (m )  Figure 2-12: The distribution of hillslope gradients within the gullied topography of Chilliwack Valley. Contours indicate the cumulative density of points, in increments of 0.2. The slopes at the initiation points for 492 landslides in the EBA database are overlaid.  tutes a putative ‘pre-erosional’ surface. An index of gully erosion volumes was then generated by subtracting the original DEM (i.e. modern) surface from the pre-erosional (i.e. constructed post-glacial) surface. The procedure involves some important assumptions about where erosion has occurred in Holocene time, and is subject to errors inherent in working with a 25 m2 grid representation of the landscape, particularly at hillslope scale. The method nevertheless provides a spatially distributed estimate of the total mass eroded from gullies (Table 2.1). Some tributaries, notably Nesakwatch Creek, have higher rates of specific erosion that may be attributed to a denser network of gullies within the drainage. The volume of gully erosion is consistent with (equal to or greater than) the volume of coarse material estimated to have been deposited at the base of the slope for 66% of the depositional sites discussed in Section 2.4 (summarized in Figure 2-6). Outlet deposition is of course an imperfect measure of the total sediment output, and does not account for finer sediment deliveries that have moved onward. There are additional contributions from planar 32  Chapter 2. Hillslope and Tributary Sediment Stores to divergent slopes are important to the sediment balance as well, and these are calculated below. Table 2.1: Estimated volumes eroded from gullied morainal and colluvial cover  Tributary Liumchen Tamihi Slesse Chipmunk Foley Nesakwatch Centre Paleface Depot Upper Chilliwack  Eroded Bulk Volume (m3 × 106 ) 34 ±4.7 97 ±13 152 ±20 33 ±4.5 71 ±9.7 96 ±13 57 ±7.8 38 ±5.2 66 ±9.0 177 ±23  Specific Bulk Volume (m3 /km2 ) 48 ±6.6 57 ±7.7 72 ±9.8 57 ±7.7 70 ±9.5 130 ±17 110 ±15 79 ±10 88 ±12 72 ±9.8  Slope classes represented within the gullied zones are shown in Figure 2-13. The overall distribution in the modern DEM is steeper than the “post-glacial” surface because material has been eroded, leaving steeper channels and side-walls in the gullies. The EBA database is overlain (lighter gray) to show where active erosion is occurring in the modern landscape. The total planimetric area occupied by each erosion depth class within the gullied zones is shown in Figure 2-14. The inferred depths of erosion over the Holocene are typically in the range of 1-8 vertical metres (mean = 6.3, median = 4.4), with some depths ranging to over 30 m. Cumulative volumetric erosion among 17 slope classes is shown in Figure 2-15.  2.5.4  Sediment Source Areas: Open Slopes  Surficial failures on open to convex slopes do not leave the same prominent signature on the landscape as gullies, but are nevertheless an important mechanism for sediment delivery over the Holocene time scale. Shallow landsliding occurs along the interfluves of some of the major valleys and on many of the steep valley slopes. The gullied zones described above are also pathways for material delivered from these upslope areas. In the alpine areas of the catchment, conditions are generally weathering-limited and slopes may attain relatively steep gradients. Broad swaths of exfoliating bedrock surfaces are evident in the alpine zone, and large accumulations of sediment are found at the base of many bedrock slopes. Products of continuous processes such as ravel and spalling, and of 33  Chapter 2. Hillslope and Tributary Sediment Stores  120 Reconstructed (Postglacial) DEM Slopes within Hollows and Gullies  100000  100 Modern (TRIM) DEM Slopes within Hollows and Gullies  80000  80  60000  60  40000  40  20000  20 EBA Database EBA Database  0 0  10  20  30  40  Number of Active Headscarps (EBA)  Topographic Points within Gullies  120000  0 50  60  70  80  90  Slope Angle (q)  Figure 2-13: Hillslope relaxation response. Histograms indicate the distribution of DEM slopes that are found within the gullied zones. The modern (TRIM) DEM is shown in darker gray; the inferred post-glacial distribution is outlined in black. The shift in the curve describing the distribution of slopes indicates that the modern (TRIM) surface is left steeper after the morainal and colluvial material has been evacuated, due to establishment of steep gully sidewalls.  1E+6  1.5E+5  2  1E+4  Number of 10m pixles (Log)  1E+5  2  Number of 10m pixles (Linear)  2E+5  1E+5  1E+3 5E+4  0E+0  1E+2 0  5  10  15  20  25  30  35  Eroded Volume (Vertical Meters)  Figure 2-14: Results from the subtraction of DEM surfaces in gullied zones, over the entire Chilliwack Valley. Graph shows the distribution of maximum depths of eroded volume. Linear scale for bar chart is shown on the left, log scale for the curve is on the right.  34  8E+8  80  6E+8  60  4E+8  40  2E+8  20  0E+0  0  MeanDepth of Vertical Gully Incision (meters, w/Std Dev)  100  05  1E+9  510 10 -1 5 15 -2 0 20 -2 5 25 -3 0 30 -3 5 35 -4 0 40 -4 5 45 -5 0 50 -5 5 55 -6 0 60 -6 5 65 -7 0 70 -7 5 75 -8 0 80 -8 5  Cumulative Erosion (m3)  Chapter 2. Hillslope and Tributary Sediment Stores  Slope Category  Figure 2-15: Cumulative erosion from gullied hillslope sources in Chilliwack Valley by slope class. Maximum rates of erosion occur in the range of 35-40◦ . Mean depth of vertical erosion increases with slope. Depths of erosion are indicated with bounds of one standard deviation.  freeze-thaw contribute to the sediment cascade on lower, soil-mantled slopes, or may fall into long-term storage. Overall, alpine zones do not provide as much material to the drainage network as processes operating downslope, in gullies and on soil-mantled surfaces. Forested slopes do not attain the same steep hillslope gradients as alpine zones, presumably due to the lower cohesive strength of the surficial material. The histogram of hillslope gradients (Figure 2-9) shows some indications of a stability threshold, suggesting that a nonlinear model may be most appropriate for representation of slow, continuous processes such as creep, slope wash, tree-throw and bioturbation and more episodic, shallow failures. In this type of model, erosion rates increase in a nearly linear manner for a shallow range of slopes, then rapidly increase for steeper sections, up to a critical gradient. Beyond this threshold, rates are limited by supply. The equation proposed by Roering et al. [1999] is qs =  K1 ∇z 1 − (|∇z|/Sc )2  (2.3)  in which qs is the volumetric sediment transport rate, ∇z is the local slope, Sc is the critical slope for failure, and K1 is the diffusion coefficient, in units of m2 /yr. This equation was initially intended for use on soil-mantled terrain, calibrated using a relatively finely spaced (4 m) DEM [Roering et al., 1999]. Roering et al. found that the 35  Chapter 2. Hillslope and Tributary Sediment Stores local erosion rate was dependent on the local curvature of the landscape, and this curvature may not be accurately resolved on coarser DEMs, such as employed here. It has been found, however, that the equation has wider applicability to bedrock terrain at a regional scale, over millennial timescales [Montgomery and Brandon, 2002]. Its use here is intended as an approximate index of erosional potential, with the magnitude of the diffusion term constrained by the estimated volumes of depositional forms along the length of the drainage network. The equation is applied to the TRIM and USGS DEMs of the study area, calculating the potential erosion for the local slope at each grid cell within planar bedrock and vegetated zones, integrated over 13 ka. This summation of flux rates provides an index of the total volumetric sediment load mobilized within each major catchment, neglecting for a moment the effects of intermediate storage (Table 2.2). The fit between this index value and the volumes estimated to have been deposited at each cone/fan deposit (Figure 2-6) is much better than any other single index, particularly for lower-order basins. In their work in the Oregon Coast Range, Roering et al. [1999] found the optimum fit to their hillslope flux data was a diffusivity term (K1 ) range of 0.0031 to 0.0045 m2 /yr. The critical slope value, (Sc ), was found to be have an optimum range of 1.2 to 1.35. In their model fit, the critical slope value was found to be less sensitive than the diffusivity term. In order to balance hillslope erosion with the depositional volumes found on hillslopes throughout Chilliwack Valley, a diffusivity range of 0.006 was selected for the soil mantled areas, and 0.003 for the bedrock zones. Long-term flux rates for the North Cascades landscape are expected to be higher due to the tills that blanket many of the hillslopes. The critical slope chosen for the soil mantled zone was 1.4. A limiting gradient of 2 was chosen for the bedrock zone, however, since the slopes do extend to a much steeper range, a constant rate was assumed beyond that limit. Figure 2-16 shows the volumetric estimates of catchment channel erosion versus the fan volume for a number of basins. If points plot above the 1:1 line in the figure (deposition > erosion), this would indicate that either erosion is under-estimated or deposition is overestimated. As the basin scale increases, storage becomes a more prominent effect, and the quality of the match between erosion and yield is reduced. The ratio between erosion and storage is highly variable, with most 2nd order basin fans holding 10% to 100% of the estimated erosion, and higher order catchments (5-6th order) generally capturing less than 10%. Uncertainty for measurements of all the fan features is conservatively estimated (i.e. worst case) as ± 30%.  36  Chapter 2. Hillslope and Tributary Sediment Stores Table 2.2: Estimated volumes eroded from planar to convex slopes  Tributary Liumchen Tamihi Slesse Chipmunk Foley Nesakwatch Centre Paleface Depot Upper Chilliwack  Eroded Bulk Volume (m3 × 106 ) 197 ±32 391 ±57 513 ±73 125 ±19 253 ±35 182 ±23 125 ±16 110 ±14 154 ±19 528 ±71  Specific Bulk Volume (m3 /km2 ) 279 ±45 228 ±33 244 ±34 214 ±33 248 ±35 247 ±31 240 ±30 225 ±29 204 ±26 216 ±29  108 2nd Order  1:1  3rd Order  Volume Deposited (m 3)  4th Order  0  1 :1  5-6th Order  107  1 :2  106  105  104 105  106  107  108  109  Volume Eroded (m 3)  Figure 2-16: Volumetric estimates of volumetric erosion vs. fan volume (bedload) are compared in a number of basins within the Chilliwack Valley, to assess the agreement between estimates. Symbols are the same as in previous figure.  Figure 2-17a presents the summation of the volume of material eroded from each grid cell in each basin, according to Equation 2.3, with no intervening storage, routed from 3rd to 8th order links. The process domains across basin scales exert a distinctive influence on the  37  Chapter 2. Hillslope and Tributary Sediment Stores pattern. Catchments with areas of up to 5-6 km2 (3rd - 4th order) show higher variance in yield, but also achieve higher delivery rates due to steeper slopes. The scatter in the pattern of specific erosion decreases at larger basin scales due to the integration of yields across a much larger contributing area. Floodplains begin to emerge in smaller tributaries such as Centre Creek and the Upper Chilliwack River as catchment plan area grows to roughly 35 km2 . Figure 2-17b shows how storage varies across the basin scales, shaping the overall pattern of specific yield in the valley. Deposition volumes (Figure 2-6) are plotted under the envelope of erosion potential. Rates of mobilization in the headwaters are relatively high, due to higher slopes, accelerated weathering, and the large load of glacial material that has been evacuated from basins. Moving from 3rd up to 6th order catchments, the specific yield of sediment is at least one order of magnitude lower, highlighting the deposition of material in fans and colluvial or alluvial reservoirs along the drainage network. The long-term yield from 5th, 6th and 7th order basins, such as Paleface, Depot and Upper Chilliwack Creek are plotted on the diagram. These are terminal deposits, emptying into Chilliwack Lake, and are estimated to represent essentially the entire yield from the Holocene Epoch (Chapter 3). The pattern of yield downstream in the 6th to 8th order catchments is one of gradually increasing specific yield over the long term. This is inferred to be the result of past remobilization of large glacial deposits that filled the lower valley. The extents of these deposits are explored in the following section.  2.6  The Fluvial Domain: The Lower Tributary Valleys  Examination of the physiography of the 6th and 7th order Chilliwack tributary valleys (see Figure 2-2) reveals that intensive erosion by valley glaciers in the lower-order basins has provided an abundant sediment supply that filled the lower portion of the larger tributary valleys and the mainstem beyond. The slope of each channel reduces gradually as it emerges from the rocky alpine headwaters. Figure 2-18 shows a comparison among 6th and 7th order tributary channel long-profiles, illustrating the results of interplay between bedrock structure, glacial and fluvial erosion. There is an evident knick point in most channels, marking a transition from an undulating or roughly convex-up profile, to the typically smooth concaveup curve of a fluvial channel (Figure 2-18). This also indicates the transition from the dominant influence of episodic transport processes in the headwaters to fluvial transport in the mid to lower valley. The transition occurs at a point in the basins with upstream areas  38  Chapter 2. Hillslope and Tributary Sediment Stores  Specific Erosion (T/km2/yr)  3x103  103  Floodplains Develop  102 104  105  106  107  108  2  109  Upstream Catchment Area (m )  104  Hanging Fluvial  Colluvial (Cirque Walls)  Fluvial Remobilization  Valley Step  Specific Yield (T/km2/a-1)  K1soil = 0.006, K1rock = 0.003 103  Chilliwack Lake 2  10  2nd Order 3rd Order 4th Order  1  10  5th Order 6th Order 7th Order 8th Order  100 104  105  106  107  Upstream Catchment Area (m 2)  108  109  Figure 2-17: (a) Combined specific erosion potential for bedrock, gullied and forested slopes, calculated for a proportion of all links in Chilliwack Valley (K1soil = 0.006, Sc = 1.4 for forested terrain, K1rock = 0.003, Sc = 2 for bedrock, assumed deposit specific density of 1.6 kg/m3 ). A maximum rate of erosion is attained in some catchments smaller than 1 km2 . (b) Inferred rates of tributary coarse sediment yield are plotted with the envelope of hillslope sediment mobilization shown in (a), above (grey region). Dashed line indicates the upper bound of the region with a doubling of the diffusion coefficients. 39  Chapter 2. Hillslope and Tributary Sediment Stores  100 m  4 km  h  un k  c at  i ih  ip m  w ak  m Ta  Ch  s Ne  sse Sle  re  y  03  le  Fo  0.  Li um  ch  en  nt  Ce  02  0.  1 0.0 0.05 0.001  Figure 2-18: Longitudinal profiles of 6 major tributaries in the Chilliwack Valley. Vertical exaggeration is 20x. Black triangles indicate the transition from the colluvial process domain to the fluvial domain. White triangles indicate secondary knickpoints, conditioned either by glacial erosion in the master valley (Liumchen and Tamihi), or, in the case of Foley Creek, by a large landslide.  of 30-40 km2 in the 6th order channels and 110 km2 in the 7th order ones. Bars and floodplains begin to develop in the distal reaches of 6th and 7th order valleys. There is usually a steep mid-section with a coarse bed and very little channel storage. The larger grain size fractions of poorly sorted debris introduced to the upland reaches are typically sequestered in long-term storage while finer material moves downstream. These reaches are supply-limited, with structured boulder beds and logjams along much of their length (“threshold” channels, [Church, 2006]). The substrate is a coarse lag derived from glacial sources and mass wasting. Much of the coarse detritus delivered from upland basins is delayed from onward transfer, resting at breaks in slope on lower valley walls or at tributary junctions. The transition point from bedrock or boulder substrate to labile alluvium has likely migrated downstream over the course of the Holocene, following the evacuation of glaciofluvial sediment stores. Valleys above the degradational limit on the mainstem such as Nesakwatch and Centre Creeks have not undergone a fall in base-level, with the attendant incision and evacuation of glacial sediments. Though there has been considerable evacuation of hillslope sediment stores 40  Chapter 2. Hillslope and Tributary Sediment Stores through the action of debris-flows, the configuration of these valleys remains a useful analog for reconstructing the ‘initial conditions’ of tributary valleys further downstream. Major valleys such as Liumchen, Tamihi, Slesse and Foley must have remained in a comparable aggraded state for a short time following deglaciation.  2.6.1  Lower Tributary Valley Fills  During deglaciation greatly elevated sediment production from valley glaciers filled the lower reaches of the valley tributaries with sandy, heterogeneous fluvial fill. Considerable evidence remains of the former surface in most of the valleys. In the case of Liumchen and Tamihi Creeks, the tributary channels were graded to the Pleistocene (late-glacial) mainstem channel, which ran alongside the ice lobe that filled Chilliwack Valley, some 150 m above the modern channel. It is not clear how long this configuration lasted, but there are remnant deposits of considerable size left in place from this paleo-Chilliwack River between Tamihi Creek and Cultus Lake [see Saunders, 1985; Saunders et al., 1987]. Figure 2-19 shows the longitudinal section geometry of the fills; the elevation indicated at the valley outlet are presumed to have been continuous with mainstem deposits. Since the disappearance of the glaciers and a fall in base-level, the five large catchments downstream of Nesakwatch Creek have evacuated significant quantities of alluvial fill from their lower reaches. Liumchen and Tamihi evacuated their valley fill relatively soon after the retreat of the Fraser Lobe from its terminus at Tamihi Moraine. Slesse, Chipmunk, Foley and the mid-valley section (present-day Larson’s Bench) must have followed a short time later, though the chronology has not been established. Liumchen presently has relatively little storage in the lower valley: material was likely evacuated quickly and streambed armouring set in soon afterwards. A bedrock sill near the outlet of Tamihi Valley is a remnant of the lip of this once-hanging valley (Figure 2-19). The river has regraded to the mainstem by incising through this feature, though clearly the bedrock incision was not wholly accomplished in the Holocene. The notch has likely developed over the course of several glaciations. The Tamihi Creek Forest Service Road travels along a high bench adjacent to the channel, and numerous truncated fans from lateral basins are evident along the length of the valley. There are substantial remnants of glaciofluvial terraces in the headwaters that have been gradually excavated. The total depth of the glacial fill in the canyon of lower Tamihi Creek appears to have been as great as 150 m, which leads to very large volumetric estimates (Figure 2-20c, Table 2.3). 41  Chapter 2. Hillslope and Tributary Sediment Stores  Elevation (8x exaggeration)  Liumchen Creek  elev. 210m  Tamihi Creek elev. 250m  Slesse Creek elev. 310m  2000  4000 6000 8000 10000 Distance from Chilliwack Mainstem (m)  12000  Figure 2-19: Longitudinal section of evacuated glaciofluvial valley fill, superimposed on the modern river profile in Liumchen, Tamihi and Slesse Creeks. Bedrock features, such as hanging glacial sills (described in the text) are shown in light gray. The Slesse Creek fill overlies a lacustrine layer near the junction with the mainstem. Vertical grid spacing = 100 m  In Slesse Creek, the former elevation of the glacial fill is evident in a number of exposures in the lower portion of that valley [Saunders et al., 1987]. Deltaic deposits stand 70 m above the modern channel. Bedded glaciolacustrine sediments lie beneath the deltaic sands at an elevation of 255-270 m. Glacial Slesse Creek evidently prograded into a mid-valley lake [Saunders et al., 1987]. Evidence from Foley Creek is particularly clear (Figure 2-20a). The valley fill is exposed in several sections, and shows a sandy matrix with a preponderance of granitic boulders that emanated from the headwaters. The fluvial facies indicate a high-energy aggrading channel (Figure 2-20b). Given the elevation and gradient of the exposed beds, baselevel for the tributary would be consistent with the elevation on the top of Larson’s Bench. Ryder Creek is not a tributary valley fill, but it represents a significant amount of sediment 42  Chapter 2. Hillslope and Tributary Sediment Stores Tributary Liumchen Tamihi Slesse Chipmunk Foley - Lower Foley - Upper Ryder Creek  Evacuated Fill (m3 x106 ) 16.0 ±2.4 128.8 ±22 78.0 ±14 6.0 ±0.8 7.1 ±0.5 8.2 ±0.5 79.0 ±6.5  Table 2.3: Volume of glacigenic fill evacuated from major tributaries.  eroded from the Ryder Lake Upland (see Chapter 4). It is included in this section as a significant, discrete tributary input that is directly coupled to the mainstem. The total volume eroded from each valley fill is presented in Table 2.3. The error term incorporates uncertainty due to alternate possible configurations of the fill. Maximum values assume continuity of the surface across the channel; minimum values assume the fills sloped toward the modern valley floor. For most valleys, the true number is probably closer to the maximum value. In some tributaries, the volume of valley fill is comparable to the total hillslope erosion estimated for the catchment over the postglacial period. This evacuated fill is a major term in the tributary sediment budget, and has presumably had a major influence on the sediment dynamics in the lower tributary valleys, and points downstream, over time.  2.7  Discussion  This chapter has examined the relative balance of storage and erosion throughout the drainage network of Chilliwack Valley, in an attempt to identify the important linkages and process zones in the Holocene sediment cascade. The pattern that emerges from this analysis is a landscape that has deposited large quantities of material at the outlet of 1st to 3rd-order catchments with basin areas generally less than one square kilometre. This is consistent with observations of Brardinoni and Hassan [2006], who showed that the hillslope colluvial domain was a distinct area of deposition within formerly glaciated catchments. Material tends to accumulate along the base of steep valley slopes, with varying proportions of onward transfer, depending on the coupling with the fluvial system. In the soil-mantled hillslopes of the Oregon Coast Range, Reneau and Dietrich [1991]  43  Chapter 2. Hillslope and Tributary Sediment Stores  a)  b)  c)  Figure 2-20: (a) Upper strata of the valley fill in Foley Creek ( 350 m upstream of Foley mouth). Exposed is a mix of glaciofluvial and debris flow deposits. (b) Upper strata of the glaciofluvial valley fill in Foley Creek (∼1.2 km upstream of Foley mouth). The coarse fluvial beds are approximately 45 m above the modern channel. The majority of the boulders are most likely Mt. Barr granite, indicating the headwater provenance of the bulk of the sediments (c) Truncated remains of a fan deposit that once interfingered with glaciofluvial fill in lower Tamihi Creek.  44  Chapter 2. Hillslope and Tributary Sediment Stores described a landscape that appears to be in approximate equilibrium, showing rates of specific yield between 50 and 200 t/km2 /yr across scales of less than 1 km2 to over 1 000 km2 . They suggested that transitory departures from this pattern might be possible due to, for instance, climate cycles that induce an augmentation in rates of landsliding. Roering et al. [1999] indicate that the Oregon Coast Range may be in an equilibrium at the millennial scale, since rates of uplift are approximately balanced by the rates of inferred hillslope erosion. The rates of bedrock lowering established in the Oregon studies ranges from 0.05 to 0.15 mm/yr. The overall rates of sediment erosion established for Chilliwack Valley, using a model similar to Roering’s, are on the upper end of the scale, ranging from 0.10 to 0.16 mm/yr averaged over the whole basin. The higher rates are due to the large quantities of glacial tills on hillslopes that have been recruited to the network over the postglacial period. Much of the glacial contribution is derived from gully incision on till blankets or morainal deposits along valley walls. Rates of basin yield are lower than the rates of mobilization because of storage along the length of the system. Over the Holocene Epoch material has accumulated along the footslopes of the steeper 1st to 3rd order drainage segments. There is a decline in the net storage in 4th order catchments, as material is mobilized onward. At the scale of 5th order catchments and beyond, there is an increase in specific rates of sediment delivery due to remobilization of stores of glacigenic material that collected in the lower reaches of each catchment. These deposits continue to strongly influence transport conditions in the distal reaches of major valleys, roughly 20 to 200 km2 in size. In the next chapter, the estimates of post-glacial yield at the scale of 5th to 7th order catchments is explored in further detail at Chilliwack Lake.  45  Chapter 3 Chilliwack Lake  Having established approximate rates of sediment mobilization in the upper tributaries of the Chilliwack Valley, this chapter examines quantities of sediment yield at the next downstream step, that of the major 6th and 7th order valleys. Chilliwack Lake provides an opportunity to estimate summary rates of yield, since it has trapped nearly all of the post-glacial load arriving from Upper Chilliwack basin (187.5 km2 ), Depot Creek (57.7 km2 ) and Paleface Creek (37.9 km2 ). Establishing the rates of yield for fine and coarse sediment fractions at this scale provides some additional insights into the storage term in the sediment cascade between the hillslope scale and that of the tributary fluvial system. A significant term in Chilliwack Valley’s post-glacial sediment budget is the stores of outwash that accumulated in the lower tributary valleys. These valleys reached a maximum state of glaciofluvial aggradation prior to the retreat, at the valley outlet, of Fraser Ice (see Chapter 4). Following deglaciation and a fall in the valley mainstem base-level, these fills would have been rapidly incised, leaving only a few terraced remnants along the valley walls. In Chilliwack Lake, subaqueously-deposited glacial outwash sediments are evident in the seismic stratigraphy beneath the Holocene lacustrine sediments. The boundary between these two units marks the transition between delivery of outwash material during deglaciation and the development of graded, post-glacial fluvial systems in tributary watersheds. The thickness of the lacustrine deposit has been estimated using sonar and seismic profiling techniques; rates of deposition are verified using lake cores from the upper 3 m of lake sediments. Radiocarbon and paleo-magnetic dating of the lake cores reveals the chronology of deposition in the last 5 000 years. The lake core record highlights the pervasive influence of debris flow activity and large inputs of organic debris. Long-term rates of bedload accumulation are less certain, but may be estimated based on the geometry of three major deltas building out into the lake. 46  Chapter 3. Chilliwack Lake  3.1  Study Area  Chilliwack Lake is approximately 9.2 km in length, impounded behind a large moraine. The moraine is a vestige from the last glacial cycle and represents a terminal position of the Chilliwack Valley glacier [Clague and Luternauer , 1982]. Since its inception at the end of the last glaciation, the lake has been an effective sink for nearly all sediment delivered from the upper valley, and thus it provides an important record for the catchment sediment budget. According to the reservoir trap efficiency rating curves of Brune [1953], based on the ratio between reservoir capacity (860 ×106 m3 ) and mean annual water inflow (630 ×106 m3 ), very close to 100% of the total sediment delivery to the lake is trapped. Only a small fraction of fine sediment reaches the lake outlet. The sedimentary strata within the lake bed record the pace of fine sediment deposition since the retreat of the Chilliwack Valley glacier at 13.3 ka cal. B.P. The fan-deltas that have formed at the outlet of the upper Chilliwack basin and two major tributaries hold the total Holocene record of coarse sediment yield. The lake is surrounded by steep rugged valley walls and half a dozen small basins that deliver occasional debris flows directly to the lake. Outflow from the lake during nival floods averages 45 m3 /s, though daily averages have been as high as 100 m3 /s (HYDAT, 2007). Annual precipitation is on the order of 3 000 mm. Winters often have significant snowfall, but the lake surface has very seldom frozen over in the past century. The lake elevation is stable at 621 m a.s.l. At 11.6 km2 in surface area, Chilliwack Lake (Figure 3-1) is small in comparison with many other ‘large’ Cordilleran lakes (>10 km) that have been the subject of study [Gilbert, 1975; Eyles et al., 1990; Desloges and Gilbert, 1991; Eyles et al., 1991; Desloges and Gilbert, 1994; Eyles and Mullins, 1997; Gilbert et al., 1997; Desloges and Gilbert, 1998]. Its upstream contributing area is also proportionately much smaller than other large lakes, and upstream glacial sedimentary stores were not as extensive. Rates of sediment yield in British Columbia mountain lakes scale to some extent with the total glacierized area in the basin [Desloges and Gilbert, 1998]. Annual rates of deposition typically range from a few millimetres to about 2 cm sediment/year. Specific yields vary significantly among the sampled population, from 30-500 t/km2 /year. Only a small percentage of Chilliwack Lake’s contributing area (2.7%) has permanent icefields or small glaciers, thus the rates of suspended sediment delivery are relatively low in comparison with other large lakes that have extensive glacial cover. The geometry of Chilliwack Lake is simple compared to larger lakes that have multiple basins, sills and bedrock-controlled topography. Gilbert et al. [2006] have divided large Cordilleran lakes into three distinct groups: (1) 47  Chapter 3. Chilliwack Lake  A B Radium Creek  Silverhope Creek Paleface Creek  C D Centre Creek  US Border Depot Creek  Little Chilliwack  Creek  Bear Creek  Figure 3-1: Chilliwack Lake, with TRIM digital terrain model. Lines A, B, C and D indicate seismic sections discussed below.  those which contain only small amounts of sediment distributed as discontinuous deposits in the deepest parts of the lake, (2) those which received abundant sediment during late Pleistocene deglaciation but very much less during the Holocene Epoch and (3) those which contain thick deposits of sediment begun during deglaciation and continued through the Holocene. Chilliwack Lake evidently belongs in the second group, having a large, distinct package of outwash material that most likely accumulated in the immediate postglacial period, and relatively modest contributions thereafter. The pattern of sedimentation appears to be uniform and the history of the lake is fairly well understood. It thus presents a good case for developing a long-term lacustrine sedimentation model.  48  Chapter 3. Chilliwack Lake Water Depth 80 m 100 m  250 m  90 m  100 m  Paleface Fan  110 m  120 m  Figure 3-2: Figure showing mid-lake CHIRP data on the North end of ‘C’ transect, Figure 3-1, and the continuous drape of lacustrine beds leading up onto the edge of the Paleface fan delta. See site (b) in the next figure for the larger setting within the lake sediments.  3.2  Seismic Methodology  Profiling of the lake was undertaken using three different seismic methods in order to assess the geometry of lake deposits since the close of the Fraser glaciation. Sounding of the lake had previously been done by R. Gilbert and M. Church (pers. comm.) in 1989, using a Datasonics Bubble Pulse system. They were able to resolve a number of features that hinted at the sedimentary history of the lake, and they speculated about the possible origins of multiple recorded strata. A first survey was undertaken in September 2003 using a CHIRP II system. The CHIRP II has a single MASSA TR-75 transducer with a 4 kHz source and a CAP-6600 receiver. This provides a relatively low frequency signal and a high-resolution image of the upper sedimentary strata. Penetration was typically on the order of 15-20 m. The imagery highlights the lacustrine bedding in the lake, which appears to be distinct and continuous. Figure 3-2 shows an example of how depositional layers drape evenly over the lake bed and over distal deltaic sediments from Paleface Creek, at the right. The sediments are interstratified with the alluvial fans and coarser rock fall deposits that project into the lake. Debris flow basins have built coarse-grained cones out into the lake. A second survey was undertaken during the same period using a Bubble Pulse SBP 510 system (source frequency = 400 Hz), the unit that was used in 1989. This time, however, a 24 channel hydrophone streamer was used, rather than the conventional Datasonics Eel.  49  Chapter 3. Chilliwack Lake Hydrophone sensors in the streamer were located at 2 m intervals, allowing for a much broader multi-channel array than the Eel. Analog signals from the streamer were recorded with the Geometrics GEODE. Recording the shots with the GEODE allows the opportunity for multichannel post-processing algorithms, stacking and migration. Finally, to sound out the geometry of the lower strata, a BOLT acoustic airgun was used with a streamer similar to the second survey. This was done in June 2004. Resolution from the airgun record provides only a rudimentary picture of finer structures (see a direct comparison of the two systems below in Figure 3-6), however it offers an excellent record of major boundaries within the lake sediments as well as the bedrock interface.  3.3  Interpretation of the seismic record  The limit of penetration of the CHIRP system shows the likely extent of finer-grained lacustrine materials. The higher energy sources highlight the boundaries between lake sediments and underlying outwash material, as well as bedrock boundaries. Multiple reflections obscure the record in several places, however the presence of deeper structures in the lowest strata is confirmed by multiple surveys. Penetration was likely hindered in many areas by the presence of gas and organic layers that damped the energy of the seismic signal. The presence of these layers was confirmed in the lake cores, two of which contained thick (>10 cm) organic beds. Holocene lacustrine sediments attain an average thickness of approximately 9.5 m, more in locations near the lake deltas (up to 15 m), and in the most distal basin. This estimate is subject to errors inherent in modelling changes in seismic velocity with depth. Relatively clear waters are assumed. Velocity modelling from the CHIRP system was automated by the software and so it is difficult to ascertain the parameter sensitivity. Model parameters for velocity changes with depth from the airgun data are more explicit; however the vertical resolution of the data was too coarse to enable much refinement of depth estimates for the upper layer of lacustrine material. It is assumed that the relative precision of the CHIRP II record is such that the potential errors in estimates of sediment thickness are roughly ±0.5 m. Three major facies can be distinguished from the seismic records (Figure 3-3). It is speculated that the lowest consists of advance outwash material, possibly with valley sediments that pre-date the advance of the Chilliwack alpine glacier. A middle facies would then correspond to sandy, subaqueously deposited late glacial outwash material sedimented during ice retreat, overlain by the third facies consisting of fine-grained, Holocene lacustrine sediments.  50  Chapter 3. Chilliwack Lake The boundaries that separate the three groups are distinct, though there are numerous minor reflectors discernible within each unit. The zone beneath the Holocene lacustrine unit (as resolved by the airgun survey) is stratified but the facies appear less ordered. The dispersion of seismic energy within this unit indicates coarser material than lake sediments, consistent with sandier material. Figure 3-3 shows the air gun record from a longitudinal profile taken along the 9 km centreline of the lake, from the impounding moraine to the Upper Chilliwack River Delta. Unit ‘a’ represents the Holocene lacustrine record. In the upper section of the lake, distal portions of Paleface and Depot Creek deltas are revealed (‘d’ and ‘e’). The reflectors from unit ‘c’ disappear below the upward curve of the moraine at the distal end of the lake. The terminus of the Upper Chilliwack glacier likely remained in place against the moraine for an extended period, shedding outwash into the lower valley over the course of the final phase of Fraser glaciation. The lower boundary of unit ‘b’ appears to be conformable with unit ‘c’, suggesting that it was deposited as a sandy, subaqueous outwash following the retreat and downwasting of the Chilliwack ice lobe. Long, inclined reflectors originating from Paleface Fan are discernible, indicating interstratification of fan sediments with the outwash. There are at least two sub-units within ‘b’ up-valley of Paleface Creek, indicating more complex interactions among Upper Chilliwack, Depot and Paleface Creeks. The down-lake, distal fringe of the Depot Creek fan has a down-valley distal ridge morphology roughly 7 m high that appears to be built upon a structure deeper within the outwash package. The lowest zone (‘c’) appears to have large-scale disruption structures within it. However a reflector in the most distal cross section (Figure 3-4, see Figure 3-1, line ‘b’) shows a diagnostic “bow-tie” reflector that indicates a steep synform structure, consistent with a buried channel. It is interesting to note in Figure 3-4 that the disrupted structure is mostly confined to the center of the valley, and is surrounded by sediments that have undisturbed layering. The structure is picked up intermittently in the long profile section, often giving the impression of distorted structures in the lower unit. I assume this to be the Pleistocene valley floodplain that was buried by the advancing alpine glacier. The outwash that overlies the Pleistocene valley floodplain (Figure 3-4, layer ‘b’) constitutes the initial topography for the post-glacial period, and represents an important datum for the volumetric analysis and sediment budget that follows. By 13 300 ± 180 cal. B.P. the final strata of outwash were deposited on the sandur that extends from Chilliwack Lake Moraine to downstream of Slesse Creek [Saunders, 1985].  51  Two-Way Travel Time (s)  c  a  Distal  b  4  d  6  Distance Along Lake Centreline (km)  Proximal  e  8  Figure 3-3: Composite seismic image of Chilliwack Lake (centreline transect, 9 km) showing major depositional units along the axial section. Five labelled features are discussed in the text: (a) continuous, laminated Holocene lacustrine sediments, (b) sandy outwash material (c) distorted reflectors within Pleistocene strata, (d) Paleface Creek delta, and (e) Depot Creek delta, showing portions of the pre-Fraser glaciation topography (see lateral cross-section, Figure 3-4). Vertical exaggeration is roughly 45x, though seismic velocity through the lower strata is likely somewhat higher than in the lacustrine zone.  0.3  0.2  0.1  2  Chapter 3. Chilliwack Lake  52  Chapter 3. Chilliwack Lake  Figure 3-4: Cross-section B (see Figure 3-1), the closest transect to the down-valley extent of Chilliwack Lake, and the deepest section (120 m water depth). Units a, b and c correspond to units labeled in the previous figure. Top figure shows the un-migrated data, with the characteristic ‘bow-tie’ structures. The migrated section, below, shows the synform structure at depth, though distortion is introduced to other parts of the trace. The valley walls are planar; curvature at the upper edges of the trace is due to turning of the boat and the array.  53  Chapter 3. Chilliwack Lake After this phase, the glacier was de-coupled from the moraine, and began to retreat and/or break up. A lake would have quickly filled the resultant depression behind the moraine, and a regime of lacustrine sedimentation would have been established. Bedlevel at Paleface and Depot Creek would have fallen as the ice departed, and deltas would have begun building into the lake. Not long after, a delta would have begun prograding into the lake from the Upper Chilliwack River. Given the bounding geometry of the present lakehead delta floodplain, the upstream end of an early Holocene Chilliwack Lake may have been up to 4 km south of its present shoreline.  3.4  Fan Deltas  Upper Chilliwack River and the two other major tributaries to Chilliwack Lake have large fan deltas that have prograded into the lake. At the climax of the Fraser Glaciation, Depot and Paleface Creek catchments would have had glaciers that merged with the principal lobe of the Chilliwack glacier. At the close of the glacial period, an outwash fill would have built up in each basin, and has since been eroded and become incorporated into the depositional package at the mouth of each of the valleys. The fan deltas have continued building and prograding into Chilliwack Lake throughout the Holocene Epoch. These depositional packages, then, represent approximately 13.3 ka of continuous accumulation of sand and gravel. The boundary between glacially derived bedload and Holocene delivery cannot be definitively resolved given the resolution of the seismic data. Figure 3-5 shows sonar bathymetry of the fan deltas and the location of the seismic transects. Water-borne seismic and land-based ground penetrating radar (GPR) transects reveal something of the architecture of the deltas, particularly where they grade from gravels to sands to lacustrine silts and clays. At the interface of sands and silts, upper sheets of sand ‘shade out’ the finer-grained strata below, obscuring the nature of the contact. Figure 3-6 shows two data sets from different profiling techniques carried out along the same transect path. Figure 3-6 (top) shows the image from the CHIRP survey, which highlights the numerous fine-scale reflectors within the lacustrine beds. Note the dispersion with depth of the CHIRP signal, most likely attributable to the coarser outwash material. Figure 3-6 (lower) shows the boundaries between the lacustrine material, outwash, the lower valley fill, and a lower reflector that is assumed to be bedrock. In the more proximal section of the profile there is a package between Holocene delta sands and the bedrock reflector that is  54  c Se 0  ‘D ’ tio n  5 km  CHWK-02  Paleface Creek  CHWK-04  Figure 3-5: Location of subsurface surveys at Depot and Paleface Creek fan deltas. Underwater topography was surveyed using sonar on a 100 m grid. Lake core locations and seismic trace paths are indicated. Section ’D’ refers to the seismic trace in Figure 3-6.  Percussion Core  Seismic Line  Depot Creek  CHWK-03  Bathymetric Contour (5m)  Sonar Line  Upper Chilliwack Delta  Chapter 3. Chilliwack Lake  55  Chapter 3. Chilliwack Lake Distance from Shore (m)  800  600  Sand-Silt Transition  CHIRP - Depth (m)  40m  Air Gun - Two-Way Travel Time (s)  Delta Slope  400  Lacustrine Deposits  60m  Upper Outwash Boundary  80m 0.10  Upper Outwash Boundary  Outwash Fan Bedrock(?)  0.20  Lower Boundary (?)  Figure 3-6: Section D (see Figures 3-1 and 3-5). Two seismic techniques used on the same profile at the toe of Depot Creek Fan: CHIRP (above) and acoustic air gun (below). The CHIRP record shows a detailed picture of the uppermost unit visible in the air gun profile. The airgun record shows the deeper strata  probably alluvium from the older Pleistocene fan that emerged from Depot Creek. Figure 3-7 shows a 3-dimensional perspective view of the long-profile seismic survey. The record shows that the uniform upper lacustrine layer overlies complex, interpenetrating layers of outwash sediments. The record at Paleface Creek shows curved reflectors that are likely outwash fans.  3.5  Ground Penetrating Radar Surveys  GPR surveys were conducted on the subaerial surface of the Paleface Creek fan delta in an attempt to discern the subsurface structure of the landform (Figure 3-8). All GPR data were collected with Simon Fraser University’s 400 V Sensors and Software pulseEKKO IV radar system. 50-MHz antennae were used with 2 m spacing and moved broadside perpendicular in 0.5 m intervals. 56  Chapter 3. Chilliwack Lake  Depot Creek Paleface Creek  n Sa Moder  it nd Lim  3km 2km 1km  0km Figure 3-7: Profile C (see Figure 3-1) with superimposed seismic record. Perspective view, facing south east, shows the relationship among the upper beds of lacustrine material, outwash beds, and several antiform reflectors at depth that likely indicate the structure of buried Pleistocence fans emerging from Paleface and Depot Creeks. The sand limit on the fans is inferred from the CHIRP record. Transverse seismic transects are indicated with black lines.  57  Chapter 3. Chilliwack Lake Much of the subaerial portion of the fan is heavily forested, making GPR deployment difficult. Surveys were carried out on the sandy 100 m margin at the lake shore. The GPR transects indicate thick sandy foreset sequences. Not surprisingly, a lower bounding horizon to the outer deltaic strata could not be detected. The fan is built out over steep underlying topography, so the lower surface is beyond the ∼35 m limit of detection with GPR, likely on the order of 80-100 m depth (see Figure 3-9). The geometric model of the fan package remains an estimate based primarily on the water-borne seismic surveys. Estimates of the total coarse bedload sediment delivery in the Holocene Epoch are shown in Table 3.1. Volume estimates were established by subtracting ‘immediate post-glacial’ DEM surface from the contemporary one. The post-glacial surface was constructed using the seismic profiles, and incorporates errors due to seismic velocity calculations, and any possible misinterpretations of the seismic stratigraphy. This information then provides a basis for establishing the volumetric bounds of the total detrital load that has been delivered to the lake system since deglaciation. Rates of potential volumetric hillslope mobilization (developed in the previous Chapter) would appear to indicate that about 10% or less of the mobilized material reaches the terminal deposit. This is not surprising, given the storage within the hanging valleys and footslope accumulations along the length of the tributary catchments. It is difficult to provide an accurate assessment of deposition from the smaller lateral basins; it may be larger than indicated. Table 3.1: Volumetric estimates for catchment erosion and fan deltas bedload delivery since deglaciation, not including outwash stores.  Upper Chilliwack Delta Depot Delta Paleface Delta Smaller Catchments  Contributing Area (km2 ) 187.5 57.7 37.9 39.9  Est. Post Glacial Erosion (Ch. 2; m3 × 106 ) 516-772 167-241 102-155 127-195  Bedload Volume (m3 × 106 ) 70-110 24-28 14-18 ∼8  The next section examines results from lake cores, allowing better estimation of grain size distribution, specific density, and rates of accumulation. Thus a time-integrated, mass-based estimate of yield can be established.  58  c)  100m  616500 5433500  PF102  Gravel  PF100 Transect  Sand  Paleface Fan  PF101  Approximate limit of gravels  50m  616900 5433600  616900 5434000  70m  500  400  300  200  100  0m  b)  Abandonned Channel Gravel  Sandy Foresets  Sand  PF100  120m Sand  PF103  50m  100m  170m  Gravel Limit  Gravels, cobbles  Figure 3-8: GPR transects at Paleface fan delta. (a)Location map for the GPR transects. (b) Down-dip section. 0 m indicates the lake shore. (c) Strike section, moving NNE to SSW. A radiowave velocity of 0.07 m ns− 1 is assumed. This makes the limit of penetration 35 m (500 ns). The lower bounds of the sediment package were not detected.  500  400  300  200  100  0m  rve  ismic Su  Lake Se  PF 100  y Line C5  Chilliwack Lake  Two-Way Travel Time (ms)  PF1 03  01 PF1  616500 5434000  Two-Way Travel Time (ms) Active Channel  a)  Chapter 3. Chilliwack Lake  59  Chapter 3. Chilliwack Lake  Figure 3-9: Isopach diagrams, (a) Paleface and (b) Depot Creeks. A hypothetical post-glacial surface (top of outwash unit ‘c’, Figure 3-3), built with a 4th order polynomial grid, is subtracted from the modern topography of the lake fan deltas to yield an estimate of total coarse (sand and gravel) sediment accumulation from these catchments over the Holocene Epoch. The fines component of the fan deltas extends much further out into the lake and are estimated in a separate analysis.  60  Chapter 3. Chilliwack Lake  Chilliwack Lake Core Locations  (05) CHWK-06 (Distal 2) Vibracore, ~3.5 m Deepest portion of lake 124m depth  Bathymetry = 20m contours  (05) CHWK-04 (Paleface) Percussion core, ~2.4 m ~1km from mouth of Paleface Creek, 101 m depth  (05) CHWK-02 (Depot) Percussion core, ~1.8 m ~1km from mouth of Depot Creek, 66 m depth  80m  120m  (05) CHWK-07 (Distal 3) Vibracore, ~2 m 117 m depth  100m  60m  (05) CHWK-01 (Moraine) Percussion core, ~1.3 m Most distal site, 116 m depth  (05) CHWK-05 (Distal 1) Percussion core, ~1.9 m 117 m depth (05) CHWK-03 (Delta) Percussion Core, ~1.5 m ~1km from main delta, 64 m depth  0  2 km  Figure 3-10: Site map: Chilliwack Lake with the location and description of seven cores recovered from the lake.  3.6  Lake Cores  To supplement the reconstruction of the Holocene history of Chilliwack Lake, seven lake sediment cores - five percussion cores (CHWK-01 to 05) and two vibracores (CHWK-06 and 07) - were retrieved from Chilliwack Lake (Figure 3-10). CHWK-07 penetrated to a depth of 2 m, however the core barrel became twisted during penetration of the lake bottom, and it is thought that the recovered material may not be a truly vertical section; only limited analyses were done on this core. The cores were split lengthwise and photographed. As the lake cores dried, more photographs were taken (after 2 days and 10 days) to capture any further contrast and details that might emerge from the lake stratigraphy. The cores were subjected to standard analyses to determine water content, sediment dry weight and specific density, organic content (loss on ignition), and mass magnetic susceptibility: 1.5 cc samples were taken at 5 cm intervals to measure water and organic content. Low-field susceptibility measurements were taken every 2 cm on each of the cores using a GF Instruments SM20 hand-help susceptibility meter with a 5 cm coil and a sensitivity of 10−6 SI. Oriented samples taken from the lake cores in small, 2 cc plastic cylinders at a 5 cm 61  Chapter 3. Chilliwack Lake sampling interval were submitted for analysis in the OLGA J-meter coercivity spectrometer at the Pacific Geoscience Centre, at Sidney, B.C.  3.6.1  Lake Core Descriptions  CHWK-01 was recovered on the lake-floor slope that rises on the moraine at the downvalley end of the lake. This core was relatively short (1.3 m) but effectively demonstrates the influence of debris flow activity in the lake over time. Seven thick (up to 3 cm) beds of relatively coarse sands (granules up to 2-3 mm) were evident, indicating the contributions from the terminal moraine and high-energy events originating from hillslopes adjacent to the lake. The longest cores came from the deepest sections of the lake. Two vibracores and one percussion core were pulled up near 120 m water depth. A small (20 g) wood chip recovered from the very bottom of the deepest core (3.46 m, CHWK-06) yielded an AMS radiocarbon date of 4 400 ±50 years B.P. (4 870 to 5 040 cal. B.P.). All of the cores have a rich record of forest fire events, indicating at least twenty episodes over the course of 5 000 years. Two percussion cores were taken on distal sections of the Paleface and Depot Creek fandeltas (see Figure 3-5). One percussion core was extracted from the delta at the head of the lake. The core near Paleface Creek (CHWK-04) was the second-longest core recovered, with just over 2.4 m of silty material. The core taken nearest the head of the lake (CHWK-03) intercepted a thick organic bed, making it impossible to recover material beyond 1.5 metre depth. An organic bed at 84 cm depth yielded a conventional radiocarbon date of 180 ±40 years BP (roughly A.D. 1660-1800; radiocarbon calibration for this era is only loosely constrained) which indicates very rapid rates of sedimentation. Table 3.2: Radiocarbon ages from Chilliwack Lake.  Lab Code Beta- 213922 Beta- 213923  Core CHWK-03 CHWK-06  Depth (cm) 84 346  Measured Age (14 C yr BP) 180 ± 40 4 400 ± 50  Calibrated Age2 (cal yr BP) 280; 180; 150; 10; 0; (0-290) 4970; (5040-4870)  2 — Intercept ages and age range in calendar years before AD 1950. The age ranges (in parentheses) represent 1σ error limits. Beta- 213922 was dated by conventional radiometric technique, Beta- 213923 was dated by AMS. The ages were determined using the INTCAL98 database.  Figure 3-11 and Figure 3-12 show the measured physical properties for CHWK-06 (the 62  Chapter 3. Chilliwack Lake distal lake basin) and among all cores. There does not appear to be any major temporal trend in the loss on ignition data, however there is a somewhat greater concentration of organic material in the more distal cores. Water content decreases with depth, likely due to compaction and dewatering of sediments. Assuming the observed trend continues to greater depths, this may provide some explanation for the loss of seismic energy with depth in Chilliwack Lake. Sediments were generally massive, showing no indication of seasonal laminae or varves and very little visible variation with depth. Even with drying, the cores yielded little vertical variation other than beds with fine and/or coarse organics that were related to discrete sedimentation events.  3.6.2  Tephra and Other Disturbance Layers  A volcanic ash layer was found in two of the cores. Cores CHWK-04 and CHWK-06 yielded a single thick (3-5 mm) layer of ash at nearly identical depths (1.95 and 1.97 m). Fine volcanic glass from the ash was concentrated by density separation (methyl iodide, S.G. 2.2). This material was analyzed using a Cameca SX-50 scanning electron microprobe (M. Raudsepp, UBC EOS Electron Microbeam / X-Ray Diffraction Facility) with readings from individual shards indicating a likely provenance from Mt. St. Helens. 16 of 21 shards yielded readings that plot decisively within the Mt. St. Helens ‘Y’ Fe-K-Ca compositional field (Figure 313). The remaining grains are widely scattered due to selection of impure glass shards or misidentification of grains. It is expected that this tephra must have been from the midHolocene Yn eruptive period. Published dates vary, but indicate a date of roughly 3 700 cal. yrs B.P. [Westgate, 1977; Luckman et al., 1986; Vogel et al., 1990; Mullineaux , 1996]. Slabs of sediment taken from the lake cores were submitted for X-ray analysis at UBC Hospital. Positive X-rays (4 mAs at 75 kV) reveal very fine low-density ash layers. The layers are often found above or interbedded with charcoal-laden layers (Figure 3-14). Presumably, burnt organic material continued to accumulate in the lake sediments for some time (decades) after the actual fire event [Hallett et al., 2003]. These event beds provide additional checks on the radiocarbon and paleomagnetic chronology (see below). Figure 3-14 shows positive X-ray and colour photographs of CHWK-04, -05 and -06. Three distinct events within CHWK-05 and -06 appear to be correlative, based on relative position and dating evidence. CHWK05 and -06 are separated by less than a kilometre. CHWK-04 is 2.15 km up-valley from CHWK-05 and registers at least one event that appears to be correlated with the other two. 63  Chapter 3. Chilliwack Lake 3  Dry Wt. Density (g/cm ) Water Content (%) 0  1.2  1.4  0.45 0.6 0.75  Loss on Ignition (%) 0.05  0.1  0.15  Susceptibility 0.5  1  1.5  0  50  50  100  100  150  150  200  200  250  250  300  300  350  350  Figure 3-11: CHWK-06 - Vibracore from the deepest section of Chilliwack Lake. Dark bands in the column at left indicate discrete beds with concentrations of coarser clastic material and/or fine organics. The upper portion of the core may have undergone some compaction during the coring operation, interrupting the otherwise coherent gradient of increasing density with depth.  3.6.3  Magnetic Parameters  Magnetic measurements were taken on the lake cores in order to identify any spatial or temporal trends in lithology and grain size that might signal changes in sedimentation regime. Measurements were also used to reveal the record of paleosecular variation in the earth’s magnetic field over time for use in paleomagnetic dating. Iron oxides from plutonic and volcanic sediment source areas are an important component in the detrital composition of Chilliwack Lake sediments. Bedload material from Upper Chilliwack Creek has a significant component of dacite rock from dikes and shallow pods that have cut across the Hannegan Volcanic Complex (D. Tucker, pers. comm.). Despite the 64  Chapter 3. Chilliwack Lake  Water Content (%) 45%  55%  Loss on Ignition  65%  75%  0  5%  10%  15%  20%  0  50  50  100  100  150  150  Depth (cm)  Depth (cm)  35% 0  200  CHWK01 - Moraine Wall CHWK02 - Depot Creek CHWK03 - Delta CHWK04 - Paleface CHWK05 - Mid Lake CHWK06 - Deep Lake CHWK07 - Mid Lake (2)  200  250  250  300  300  350  350  Figure 3-12: Cross-core comparisons of two parameters, water (%) and loss on ignition (LOI). K2O  60  Glacier 30 Pk.  Bridge River Mazama 40  FeO  50  Mt. St. Helens  30  50  Ca0  Figure 3-13: Ternary diagram showing compositional fields for a number of Holocene volcanic sources within the Cascades and Coast Mountains. Points (Xs) show microprobe readings taken on shards from the Chilliwack Lake tephra layer.  65  Chapter 3. Chilliwack Lake  Figure 3-14: Three cores from Chilliwack Lake show a number of fire episodes over time. From left to right: CHWK-06 (Distal 2), CHWK-05 (Distal 1), and CHWK-04 (Paleface). Positive XRays (4 mAs at 75 kV) reveal very fine low-density ash layers. Based on paleomagnetic dating and stratigraphic sequence, two event beds are correlated across two distal cores and a third from as far away as Palefacefan delta. The events span an approximate range of 1 350 to 1 750 cal. years B.P.  relatively small map-area of this sediment source within the granitic Chilliwack Batholith, it has evidently provided abundant material to the upper catchment over time. A ternary diagram showing the lithological composition of coarse (> 16 mm) clastic channel material in the Upper Chilliwack mainstem is shown in Figure 3-15. Depot and Paleface Creeks both drain granitic terrain and evidently provide further enrichment of iron within the lacustrine sediments. The silt fraction from channel sediments in Upper Chilliwack River and cores in Chilliwack Lake were submitted for geochemical (ICPMS) analysis (see Chapter 5). Results show that typical iron content within the silt fraction of the lake sediments is on the order of 3-5% (Figure 3-16). There is a general pattern of increasing iron concentration from the source river silts to the distal lake material. Figure 66  Chapter 3. Chilliwack Lake 100  Tributary Basins 20  40  (%  60  )  Me  itic  tam  orp  60  an Gr  hic  (%  )  80  40  80  20  Mainstem Sediments 100 0  0 20  40  60  80  100  Volcanic (%)  Figure 3-15: Percentage of coarse (>16 mm) clastic lithologies found in active bars, Upper Chilliwack River (open squares) and major tributary sources (Bear and Indian Creeks, black diamonds). There is a disproportionate representation of volcanic lithologies in the channel, despite the mostly granitic source material provided by the major lower tributaries, which represent 41% of the upper catchment drainage area.  3-16 shows that there is also a higher concentration of elements associated with magnetite, chromite, and other heavy minerals (Cr, Mg, Ti and V). A composite index of these elements, all of which have quite similar behaviour in the system, was generated by summing centered, log-ratio transformed values for individual elements (see Chapter 5). This pattern is mirrored to some extent in readings of magnetic susceptibility from each of the lake cores (Figure 3-17). There is a systematic increase in susceptibility from the delta to distal reaches. Bulk susceptibility is a proxy for both the relative concentration of magnetic minerals and the effective grain size. It would appear that grain-size variation exerts the strongest effect on the magnetic behaviour of the sediment. It is not expected that the slightly increasing concentration in iron exerts a primary effect on the magnetic readings, but it probably does enhance some of the down-lake contrast observed. Grain size information from the lake cores obtained from sedigraph analysis is shown in Figure 3-18. Fining and sorting processes are very effective beyond the upper lake delta. Some samples from CHWK-03 (taken roughly 670 m from the lake head) show up to 25% fine sand composition and a full range of silt sizes, while samples from Paleface (CHWK-04, 67  Chapter 3. Chilliwack Lake  1  Cr + Mg + Ti + V  0.6  0.2  1.4 km 4.7 km  -0.2  -0.6 3.2 km  12.3 km  Upper Chilliwack River CHWK03 - Delta CHWK 02 - Depot Creek CHWK - 04 - Paleface Creek Distal Cores  -1 8.9 km  -1.4 -0.3  -0.2  -0.1  0  0.1  0.2  0.3  Fe  Figure 3-16: Graph showing the relative increase in iron and elements associated with heavy minerals and magnetic oxides within the silt fraction of channel sediments (Upper Chilliwack River) and lake sediments. Fe values range from 3 to 5%. Here they have been log-ratio transformed, centered on zero. Individual elements Cr, Mg, Ti and V have been similarly transformed, then summed, to create a composite index. Individually, V has the strongest down-valley gradient, and Ti the weakest.  0  50  Core Depth (cm)  100  150  200 CHWK-03 - Delta  250  CHWK-02 - Depot CHWK-04 - Paleface CHWK-05 - Distal 1  300  CHWK-06 - Distal 2 CHWK-01 - End Moraine  350 9E-8  1E-7  1.1E-7  1.2E-7  1.3E-7  1.4E-7  1.5E-7  1.6E-7  cp- Bulk Magnetic Susceptibility  Figure 3-17: Susceptibility, measured at 5 cm intervals along the length of each core. There is an evident longitudinal gradient, with magnetic mineral concentration increasing down-valley.  68  Chapter 3. Chilliwack Lake Grain Size (mm) 0.0039  0.01563  0.0625  0.25 0  Percent Finer by Weight  75  Coring Sites CHWK-02 - Depot CHWK-03 - Delta CHWK-04 - Paleface CHWK-06 - Distal 2  25  Clay 50  50  Silt  Sand  25  0 -12  Percent Coarser by Weight  0.000977 100  75  -10  -8  -6  -4  100 -2  Grain Size (y)  Figure 3-18: Sedigraph grain size curves from lake core sediment samples, Chilliwack Lake  roughly 4 km from lake head) and beyond show material graded primarily in the 2-11 micron fraction (clay and very fine silts). Sediment at CHWK-06, 6.7 km from the upper lake delta, shows only a minor degree of further sorting.  3.6.4  Palaeomagnetism  A chronology was established for the deepest cores (CHWK-04, -05 and -06) using the variation in magnetic inclination and declination. Curves established from the Chilliwack cores were compared with the closest available calibrated paleomagnetic curves, namely Fish Lake, CA [Verosub et al., 1986] and Mara Lake, BC [Turner , 1987]. The magnetic inclination and declination trends extracted from the Chilliwack cores were fit to the established data sets using a ‘least-error’ fitting technique. A master chronology was built from the deepest vibracore, CHWK-06 (3.5 m). Although the core wrapped itself in the tube no less than three complete revolutions, the torsion of the core sample appears to have been uniform, and declination data were recovered using a simple de-trending algorithm. All 69  Chapter 3. Chilliwack Lake inclination data were centered on zero. A program was coded for LabView, whereby the raw inclination and declination data were interpolated to a 200-element array. The user can then manually stretch the linked inclination and declination series into place by eye to match the Fish Lake or Mara Lake series, guided by a sum of square errors term. The fit that yielded the least error, and matched the tephra and  14 C  dates, was a starting point for further refinement. The minimum-error fits for all  cores are superimposed in Figure 3-19. The average of these curves is the best estimate of the local paleomagnetic curve. Agreement in some periods is better than others, as shown by a standard deviation that was calculated for the inclination and declination estimates at each point in the time series. The final match with the Mara Lake and Fish Lake records is shown in Figure 3-19 at right. The Fish Lake series shows a much better fit, since that curve had the best initial fit during the curve-fitting process. Thus the final, averaged form of the Chilliwack curves is highly influenced by the Fish Lake data.  3.7  Rates of Sediment Accumulation in the Holocene Epoch  Based on the chronology developed from tephra and the paleomagnetic and radiocarbon data, it is possible to reconstruct the rates of sediment accumulation over time. Figure 3-20 shows the depth of sediment accumulated versus calendar year. The Paleface core (CHWK-04) record extends to approximately 3 600 cal. B.P. (2.4 m), and the core from the deepest portion of the lake (CHWK-06) extends to 5 200 cal. B.P. (3.5 m). Overall, there are relatively slow rates of accumulation in distal portions of the lake (average of 0.7 mm/year) and more rapid rates near the delta (5 mm/year). The overall interpretation is that rates of accumulation on the lake floor are relatively stable over the decadal to century scale. This is despite the evident variability of sediment delivery shown in numerous events beds, and among the varied geomorphic lake floor settings of the cores. The long-term gross rates of lacustrine sedimentation inferred from the lake cores are in good agreement with the interpretation of the CHIRP seismic record. The minimum expected thickness of the distal lacustrine package (above the outwash), based on accumulation rates of 0.7 mm/year, would be 9.45 m. The lake floor thickness varies somewhat, with the maximum distal thickness ranging up to 10 m, and accumulations near the fans reaching 15 m. The mean estimated thickness along the central transect of the lake is 9.8, with one standard deviation of ±1.6 m.  70  Chapter 3. Chilliwack Lake Averaged Chilliwack Core  ‘Agreement’ among cores  Results shown with Fish and Mara Lake 0  0  Cal. Years B.P.  -1000  -1000  CHWK-02 CHWK-04 CHWK-05 CHWK-06 Average  Std. Deviation Inclination Declination  -2000  -2000  Series Match Mara Lake Fish Lake Chilliwack  -3000  -3000  MSH’Y’ ~ 3600 Cal. B.P.  -4000 15 30 45 60 75 90  Inclination (F)  -4000  -90  -30  30  Declination (q)  90  0  50  Std. Deviation  15 30 45 60 75 90 -90  Inclination (F)  -30  30  90  Declination (q)  Figure 3-19: Traces of magnetic inclination and declination from four cores are shown in the first panel, as well as their average. These are the raw (unsmoothed) data, interpolated onto a standardized 200-element array to facilitate cross-core statistics. Standard deviation among the magnetic readings are shown in the middle panel. The shorter core extends only to 2 500 B.P., and thus sample size diminishes prior to this date. The final panel shows Fish Lake and Mara Lake datasets compared with the Chilliwack series.  Following the assumption that the outwash surface is represented by the seismic reflector described in the previous section, an isopach map is shown in Figure 3-21 showing the thickness of the overlying Holocene lacustrine accumulation. The horizontal boundaries of the lacustrine sediment package can be accurately delineated at each of the transects, however it has to be estimated for intermediate areas. Assuming that the boundaries are delineated with an accuracy that is within ± 50 m, the surface area of the lacustrine deposit is 10.8 ± 0.45 km2 . This is then combined with the expected vertical accuracy to produce the total volumetric error. A model of the expected vertical error term is shown in Figure 3-22a. The total bulk volume of Holocene fine sediment accumulation is 106 ± 13 ×106 m3 . A breakdown is shown in Table 3.3. The dry-weight density for most lakefloor sediments in the proximal zone ranges from 1.28 to 1.55 t/m3 . In distal zones the density ranges from 1.18 to 1.44 t/m3 (Figure 371  Chapter 3. Chilliwack Lake 0 CHWK02-Depot Creek CHWK03-Lake Delta CHWK04-Paleface Creek CHWK05-Distal 1 CHWK06-Distal 2 Fire Disturbance Radiocarbon Date (cal yr. BP, 1ó)  -50  0.56 mm/yr.  Core Depth (cm)  -100  -150  -200  0.79 mm/yr. -250  -300  -350  -400 0  -1000  -2000  -3000  -4000  -5000  -6000  Calendar Date (B.P.)  Figure 3-20: Paleomagnetic chronology (calendar years) mapped to sediment core depth (cm). Red dots indicate fire disturbance events recorded in the deepest core. Elevated rates of sedimentation are inferred at the Upper Delta (CHWK-03) based on radiocarbon dating of an organic layer at 85 cm (180 ±40 years BP - roughly A.D. 1670-1800; radiocarbon calibration for this era is only loosely constrained). Before approximately 3 200 BP there appears to be a slightly elevated rate of accumulation. Reduced major axis regression within the distal cores indicates a significant change in slope (α=0.01), though there are fewer core samples deeper than 2 m to confirm this shift in trend.  22b). Organic material is estimated to constitute roughly 10% of the volume. Based on the relationship amongst density, depth and longitudinal position established from the lake cores, a simple three-dimensional model was developed to estimate the total mass of fine sediment in the lake. For the deeper portion of the deposit, it is expected that the average bulk density of the lake sediment has increased with time due to compaction and consolidation. It is assumed that the density profile roughly follows a logarithmic curve [Lane and Koelzer , 1943; Gill , 1988] and that the maximum density to be expected is near 1.65 t/m3 (further assuming constant grain size composition with depth). The total mass of fine lacustrine material is then calculated to be approximately 144 ± 18.6 ×106 Mg, indicating an average deposit density of 1.36 t/m3 . Based on the above calculations, the long term (13 300 year) specific annual yield of fine lacustrine sediment for the whole contributing basin is 32 ± 4 t/km2 /yr. Including the  72  Chapter 3. Chilliwack Lake  0.0m  3  6  9  12  15  Lacustrine Sediment Thickness (m)  Figure 3-21: Isopach map showing the inferred depth of lacustrine sediment since the end of outwash deposition and the onset of lacustrine conditions at Chilliwack Lake, based on seismic and sonar surveys. The seismic transects are overlaid in white. (a)  (b)  0  0  -2  100  200  -6  300  Depth (cm)  Depth of 1/2 m strata  -4  -8 -10  400  CHWK-02 - Depot CHWK-03 - Delta CHWK-04 - Paleface CHWK-05 - Distal 1 CHWK-06 - Distal 2  500 -12  600  -14  ?  700  -16  800  -18 0  1E+6  2E+6  3E+6  4E+6  Volume (m 3)  5E+6  6E+6  1  1.1  1.2  1.3  1.4  1.5  1.6  1.7  1.8  Density (g/cm3)  Figure 3-22: (a) Volume of lacustrine sediment with depth. Uncertainty bounds are indicated by multiple profile lines. Model is based on seismic imaging of the sediments and follows assumptions outlined in the text. (b) Variation of specific dry weight density of sediments with depth in each of the lake cores. The logarithmic trend lines indicate the expected pattern of density with depth.  73  Chapter 3. Chilliwack Lake Table 3.3: Sediment delivery to Chilliwack Lake: Volumetric estimates  Medium Term (4 000 ka) Yield Deltaic Sediments Lacustrine Sediment Long Term (13 000 ka) Yield Deltaic Sediments Lacustrine Sediments Outwash Deposition Total Long Term  Estimated Volume (×106 m3 )  Mean Density (Mg/m3 )  Mass (×106 Mg)  ? 34.5± 2.5  1.48  51  1.8 1.48 ?  190 146  106 ± 13 98.5 ± 6.5 140 ± 20 344.5 ± 24  bedload material from three fan deltas adds roughly another 30 t/km2 /yr. In the last 2 000 years, the fine sediment yield has been about 21 t/km2 /yr (Table 3.4), plus an unknown amount of bed material yield. This places it decidedly at the low end of the spectrum with respect to other large (>10 km2 ) Cordilleran lakes, though this is to be expected considering the small glacier coverage in the basin and relatively small upstream catchment area. Figure 3-23 shows Chilliwack Lake plotted within the growing dataset of lake information collected by researchers in the past few years. Table 3.4: Mineral sediment delivery to Chilliwack Lake: Late Holocene (<2 000 years BP) estimates based on core chronology.  Chilliwack Lake  Average Deposition Rate (mm/yr) 0.48 ± 0.04  Inferred Specific Yield (t/km2 /yr) 20.9 ± 2.0  There are no historical measurements of suspended sediment yield from the Upper Chilliwack River. However, the medium term yield as inferred from the lake cores appears to be the same order of magnitude as the modern Silverhope Creek, a neighbouring catchment to the Chilliwack, with similar catchment area (350 km2 ) and physiography. Monitoring of suspended sediment there suggests a contemporary average specific fine sediment yield of 11.7 t/km2 /yr, consistent with patterns of sediment yield in the Cordillera in post-Neoglacial time [Schieffer et al., 1999]. The resolution of the paleomagnetic curve averages 60 years between sampling points (5 cm). This temporal scale likely incorporates a significant range of variability, from multiple geomorphically effective floods to large scale changes in global circulation patterns (PDO, 74  Chapter 3. Chilliwack Lake 10000 Stave Meziadin 1000  Ape  Berg  Bowser  2  Specific Sediment Yield (T/km /yr.)  Nostetuko Hector  100  Green  Lillooet  Mud Lake  Chilko  Chephren Maggie  Quiniscoe 10  Harrison  Chilliwack  Glacier  Moose  Woods Clayoquot  Pyramid Kite  1  Klept Ash 0.1  Glacier Cover (%) 85%  Middle  50%  0.01  10%  Gallie  0%  0.001 0.01  0.1  1  10  100  1000  10000  100000  2  Contributing Catchment Area (km )  Figure 3-23: Fine sediment yield for Chilliwack Lake shown in comparison to other lakes in British Columbia. Dots size indicates the proportion of basin glacial cover. Data collated from Gilbert et al. [1997]; Desloges and Gilbert [1998]; Schieffer et al. [1999]; Hodder et al. [2006].  ENSO). The presence of large, distinct beds that are presumably from discrete events emphasizes the high variability of sediment delivery on the annual-to-decadal scale. At the century scale, however, much of the noise appears to be damped. Sedimentation rates appear stable, and do not show a prominent signal from Neoglacial advance, though a slight increase in sedimentation rates is hinted at prior to 3 200 BP. Further work is needed to assess this, but it appears that Neoglacial effects did not overwhelm the system and/or were buffered by storage effects upstream of the lake. Chilliwack Lake captures essentially the entire Holocene depositional record from 3 major basins. In the Strahler ordering scheme established in Chapter 2, Paleface, Depot and Upper Chilliwack Creeks are 5th, 6th and 7th order catchments, respectively. This is a spatial scale within the drainage where storage of material becomes relatively important. The total deposition in the lake, including the lower outwash layer, accounts for roughly 10-15% of the sediment estimated to have been mobilized within all contributing catchments.  75  Chapter 3. Chilliwack Lake A majority of material eroded from hillslopes is left in storage between hillslope scales and the major tributary, or distal, fluvial domain.  76  Chapter 4 Evolution of Chilliwack Valley Mainstem  Most of the summary post-glacial sediment yield from Chilliwack Valley is contained in a large alluvial fan that extends in a semi-circular arc from Vedder Crossing, at the piedmont of the Cascade Range, to the Fraser River. Much of the sediment mass in Vedder Fan is composed of glacial material eroded from the glacially aggraded mainstem and the tributary valleys. Previous chapters have examined long-term sediment fluxes at the hillslope and tributary scales; here we look at the mass balance of post-glacial erosion and sedimentation at the scale of the mainstem valley. These are 7th and 8th order links according to the Strahler ordering system established in Chapter 2. In the first section of this chapter, the volume of the glacial valley fill is estimated by fitting a spline mesh across the remnant surfaces and calculating the bounding volume. The depth of valley fill that has been evacuated is over 80 m in places and the breadth is over 1 km. In the postglacial evolution of the valley, then, the lower tributary fills and mainstem deposits are a dominant component of the sediment budget. Given the uncertainty in the topography of the lower valley immediately following final retreat of the Fraser ice, maximum and minimum estimates are given. Many of the major valley depositional units were laid down in complex association with older deposits, and thus definitive landform interpretation and ascription of precise bounding volumes remains difficult. Nevertheless, an approximate estimate of the total volume evacuated from the lower valley can be made. The second part of this chapter examines the historical development of Vedder Fan at the outlet of the valley, using well logs and radiocarbon dates to infer the internal architecture of the fan and probable rates of accumulation. Sonic drilling work was undertaken by the City of Chilliwack in 2003-2006, and a rich database of historical well logs was assembled. The information has assisted in the development of a 3-dimensional chrono-stratigraphic model 77  Chapter 4. Evolution of Chilliwack Valley Mainstem of the fan. An idealized post-glacial Fraser Valley floor was reconstructed, and the volume of sediment deposited thereupon was estimated. A final section compares the eroded volumes from the major valley sediment sources with the total calculated fan volume.  4.1  Initial Conditions  Before introducing the post-glacial sediment budget of the lower valley, a brief overview of the regional glacial history is presented. Radiocarbon dates in the text that follows are in 14 C  years, in keeping with normal practice in Quaternary studies, whereas later discussions  use calibrated ages in order to properly estimate rates of sediment flux. The Puget Lobe of the Cordilleran Ice Sheet reached its maximum extent 14 500-15 000 14 C  years ago [Easterbrook , 1992; Porter and Swanson, 1998; Booth et al., 2003]. By 13 000  14 C  BP, a large calving embayment began to develop in the Strait of Georgia and as ice  retreated, glaciomarine drift and other marine deposits accumulated over a landscape that had not yet rebounded from isostatic depression [Clague et al., 1997]. Armstrong et al. [1965] designated this interval as the Everson Interstade, ending at about 12 000  14 C  years BP.  Relative sea levels lowered substantially - up to 200 m in some parts of the Fraser Lowland - as a result of isostatic rebound of the deglaciating land surface [Hutchinson et al., 2004]. Relative sea level fell from 150 m elevation to about −15 m between 12 000 9 900  14 C  14 C  yr BP to  yr BP. There was then a slow rise to near the modern level some 3 000 years later.  There were at least two late readvances of the ice sheet in the Lower Fraser Valley (Figure 4-1), marking the Sumas Stade. The number of the advances and their timing remain unclear [Clague et al., 1997; Kovanen and Easterbrook , 2002]. With the advance of the Sumas ice, glacial sub-lobes extended into the lower Coquihalla, Hatzic and Chilliwack Valleys. During this period a large train of outwash extended from Chilliwack Lake, past Slesse Creek to a short distance downstream of Borden Creek (Figure 4-1a). The sandur appears to have prograded downvalley into a large and long-lived lake that was ponded by Fraser Ice. The upper valley glacier had retreated east (upstream) of Slesse Creek by 11 900  14 C  years  BP. Fraser ice that entered the lower valley advanced up-valley toward Tamihi Creek and downstream toward the Columbia Valley. Evidence suggests that erosion of lower Chilliwack Valley mainstem sediments, and the building of Vedder Fan, began immediately following the recession of the Fraser ice lobe in the Lower Mainland [see Easterbrook , 1971; Saunders, 1985, Figure 4-1c] . Ice-free conditions  78  Chapter 4. Evolution of Chilliwack Valley Mainstem in the lower valley commenced after approximately 11 000  14 C  years B.P., based on the  radiocarbon chronology at Tamihi moraine developed by Saunders et al. [1987] and Clague et al. [1997]. After this date, Tamihi moraine was abandoned and ice gradually disappeared above Vedder Crossing. The date of the final withdrawal of the ice sheet from the Fraser Valley is not well constrained; Kovanen and Easterbrook have proposed two readvances of the Fraser ice between 11 400 and 10 250  14 C  years (their Phases SIII and SIV), introducing the  possibility that the ice sheet remained in the Fraser Valley for some time after the deposition of material at Tamihi Moraine. As the ice retreated from the low divide of metasedimentary rock between the Chilliwack River and the Fraser Valley, the Chilliwack River would have begun to flow northward to the Fraser River. With the retreat of the ice, the effects of the baselevel fall in Chilliwack Valley would have propagated rapidly upstream. Much of the initial downcutting in Chilliwack tributaries was accomplished in the early part of the Holocene Epoch. High sediment loads would have resulted from mass movement on valley slopes laden with glacial debris. Material was initially eroded from the valley floor of Liumchen Creek, later followed by Tamihi Creek, Slesse and Foley Creeks. There is a sequence of terraces between Foley Creek and Slesse Creek that record the stages of incision in the channel. Evidence at Post Creek and upper Foley Creek emphasizes that channels were sometimes shaped by catastrophic flooding - j¨okulhlaups - as ice dams released large volumes of meltwater that had collected in proglacial lakes and valleys [Goff and Hicock , 1995]. In the Fraser Lowland, Cameron [1989] discussed the possibility of a wasting ice lobe sitting in the Sumas Valley in the final phase of deglaciation (Figure 4-1d), leaving a large slackwater basin that would eventually become Sumas Lake. To the southwest, the Nooksack River began building an alluvial fan toward the Sumas Valley (Figure 4-1e). The Chilliwack fan built outward, prograding into Sumas Lake, and also possibly contributing to the northward deflection of the Fraser River [Cameron, 1989]. The distal edges of the fan interfinger with floodplain sediments of the Fraser River. The initial topographic state for the post-glacial sediment budget investigation, and the routing model that follows in Chapter 6, is considered to be the landscape as it was immediately prior to the onset of baselevel fall (Figure 4-1d). At that point the Fraser lobe no longer obstructed the valley at Vedder Crossing, and the alpine Chilliwack glacier had receded to the headwaters of the valley. Constraints on the exact date are poor, but for the purposes of estimating rates of sediment evacuation in the sediment budget, a date of 11 100  14 C  years,  79  Tamihi Moraine  Chilliwack Valley Sumas Fan  Sumas Valley  Sumas Lake  (f)  City of Chilliwack  (e)  (d)  Vedder Fan  0  15 km  Figure 4-1: Six-part deglacial history of Chilliwack Valley: (a)-(c) depicts the retreat phase of the final stage of the Sumas ice lobe, active floodplain outlined in green, 11 200 to 10 500 14 C years B.P. (d) A remnant lobe of ice in Sumas Valley, following ice retreat, 10 500 - 10 000 14 C years B.P. (e) Sumas Lake fills the depression left by the ice, basin fills throughout Holocene (f) The modern landscape: City of Chilliwack and Sumas Valley. Chilliwack River has been channelized into the Vedder Canal. Adapted from Saunders et al. [1987], Cameron [1989] and Clague et al. [1997].  (c)  (b)  Cultus Lake  (a)  Columbia Valley  Sumas Ice  Chapter 4. Evolution of Chilliwack Valley Mainstem  80  Chapter 4. Evolution of Chilliwack Valley Mainstem or 13 000 calendar years, is used. A 500 year error in this date would introduce a 4% error in the whole Holocene sediment flux data.  4.2  Mid Valley Fill  Because many remnants of the mid-valley glacial topography remain intact, it is possible to estimate the total volume of sediment evacuated from Larson Sandur, which extended from present day Chilliwack Lake to a short distance downstream from Borden Creek. Several terrace surfaces representing different historical stages of valley evolution were rendered in ArcGIS using data from field reconnaissance and published data, in order to recreate the endPleistocene topography. Subtracting the DEM grid representing modern topography from this yields approximate volumes of evacuated material. Modern fan deposition and other accumulations were accounted for in the resulting volume, to establish the net erosion. Figure 4-2 shows an isopach diagram of the total volume estimate to have been evacuated from Larson Sandur and the valley fills in lower Slesse, Foley and Chipmunk Creeks. Some erosion from the surface intact remnant topography is assumed. The total volume eroded here (Table 4.1 does not include the glacio-lacustrine unit that underlies the sandur, as this is tabulated in the following section. Table 4.1: Net bulk volume eroded from the mainstem between Chilliwack Lake and Borden Creek  Reach Chilliwack Lake to Slesse Lower Slesse Valley Slesse to Tamihi Sum  Eroded Sandur Vol.(m3 x106 ) 15.3 ±2.1 5.0 ±0.7 25.3 ±5.2 45.6 ±5.7  The reach extending from Slesse Creek to Ranger Run (located 100 m west of the edge of Figure 4-2) consists of floodplain that has been constructed from the reworked mass of sediment originating from Foley Creek, Larson Sandur and Slesse Creek. There were likely several generations of paraglacial floodplain that have progressively coarsened with succeeding transport episodes. There is now a broad (up to 1 km wide) floodplain from which material is recruited by a laterally active channel. Figure 4-3 shows the surveyed extents of terraces between Foley and Tamihi Creeks. Terrace surfaces on the North and South sides of the valley were surveyed using Paulin altimeters. The surveys revealed only sparse evidence of high terraces within the sandur 81  Chapter 4. Evolution of Chilliwack Valley Mainstem  593000 .000000  596000 .000000  599000 .000000  602000 .000000  Foley Creek  ’s on s r La  h nc e B  5438000.000000  5438000.000000  5441000.000000  Chipmunk Creek  5441000.000000  590000 .000000  Net Erosion Depth (m)  20.1 - 30 30.1 - 40 40.1 - 50  Slesse Creek 1  2  50.1 - 60 60.1 - 70 70.1 - 80 80.1 - 90  4 Kilometers  5432000.000000  0  90.1 - 106  590000 .000000  593000 .000000  5435000.000000  10.1 - 20  596000 .000000  599000 .000000  5432000.000000  5435000.000000  0 - 10  602000 .000000  Figure 4-2: Isopachs of the estimated net erosion from the mid-valley, at Slesse confluence. The maximum erosion depth is 106 m. Shallow erosion is shown across the intact remnants of Larson’s Bench, after a pre-erosional surface was fit to it.  remnants, but there are numerous terrace flights at lower elevation. It is expected that initial downcutting occurred quite quickly, and once the channel became sufficiently armoured, it took longer to incise and rework the sediment supplied from within the deposit and from upstream (see Chapter 6).  4.3  Glacio-Lacustrine Deposition  Exposures of laminated clay- and silt-dominated sequences are found throughout the mainstem drainage (Figure 4-4), recording the existence of at least two, most likely several, lakes over the course of the late Pleistocene Epoch. Glaciolacustrine facies are quite variable, mak82  Chapter 4. Evolution of Chilliwack Valley Mainstem Foley Chipmunk Creek Creek  400m  Borden Creek  Slesse Creek n’s Larso  300m  h Benc  Sandur/ Delta  Outwash Surface Upper Terraces 200m  Mid Terraces Lower Terraces Water Level  Lacustrine Deposits  30  25  20  15  10  Distance Upstream from Vedder Crossing (km)  Figure 4-3: Longitudinal survey of terraces between Foley and Tamihi Creeks. Degradation appears to be hinged on a knick point close to the outlet of Foley Creek. Terrace flights have been grouped as high (red), middle (light blue) and low (green). Points surveyed on the terraces are coloured accordingly; points on the modern river are coloured dark blue. River topography is taken from the BC TRIM DEM breaklines  ing definitive interpretation of lacustrine sequences quite difficult. Numerous interpretations have been made regarding the extent and interconnection of lakes [Chubb, 1966; Munshaw , 1978; Clague and Luternauer , 1982; Hicock et al., 1982; Saunders et al., 1987]. A glaciolacustrine sequence is found in the lower 10 km of the valley (see Figure 4-6, Hicock et al. [1982]), up to 16 m thick, and overlying outwash gravels from the advance phase of the Fraser glaciation (20 to 21 ka  14 C  B.P.). Exposures of a second major sequence,  representing several episodes of glaciolacustrine sedimentation, extend from a short distance (approx. 1 km) upstream of the Slesse Creek confluence to near the Tamihi Creek confluence [Clague and Luternauer , 1982; Saunders, 1985; Saunders et al., 1987]. Present exposures in lower Slesse Creek show the lake extended some distance up Slesse Valley and have been dated to 11 900 ka  14 C  B.P. [Saunders et al., 1987]. Well logs from the Slesse Hatchery  (Provincial Well Log Database: 23325 & 23326) indicate that the laminated silts extend to 83  Chapter 4. Evolution of Chilliwack Valley Mainstem  Figure 4-4: Fine sediment source material is delivered episodically from deep glacio-lacustrine deposits in the mid-valley reaches.  over 70 m depth . Dates recovered by [Saunders et al., 1987] further downstream along the mainstem near the Tamihi confluence indicate a similar age: 11 600 ka  14 C  B.P.  Assuming valley-wide extents for the above sequences, the total volume of eroded glaciolacustrine clays and silts in the mainstem and Slesse drainages is at least 122 ±26 × 106 m3 , a significant portion of the lower valley budget. In recent times, individual episodes of erosion in the glaciolacustrine material are estimated to deliver over 104 m3 to the channel [Thomson, 1999]. One source area in particular was estimated to have delivered as much as 1.8 x 106 m3 of material [Thurber Engineering, 1997]. These are some of the most prominent slope movements along the valley mainstem.  84  Chapter 4. Evolution of Chilliwack Valley Mainstem  Figure 4-5: Larson’s Bench is incised by a late Holocene channel. The formerly active layer of the incising channel overlies sandy deltaic fill on the left; lacustrine clay on the right.  4.4  Lower Valley Fill  The Lower Chilliwack Valley has a complex history of glacio-fluvial, glacio-lacustrine and glacial till deposition. The most prominent deposits in the lower valley are the Ryder Lake Upland and the Tolmie Upland (Figure 4-6). In post-glacial time, these thick moraine deposits have been subject both to river incision and to diffusive slope processes that have carved a network of large gullies into their flanks. The sediment mass has been largely stabilized under the influence of vegetation, however there are still shallow failures that continue to shed material to the floodplain below. In the early postglacial evolution of the valley, the upland moraine complexes have been a very important sediment source to the mainstem due to the direct mode of delivery to the channel (no intervening storage) and the large volumes of material delivered. The Ryder Lake Upland comprises a thick sequence of early Fraser Glaciation outwash sediments overlain by silts and a compact massive diamicton, most likely subglacial till. The upper sequence of sediments is fine-textured and contain an assortment of sub-angular to sub-rounded, often striated clasts. Quartzite lithologies found in the Ryder Lake Upland suggests a Fraser Valley provenance [Saunders et al., 1987], although the contribution of Chilliwack source material is not entirely clear (See Chapter 5). The Tolmie Upland is part of the ‘Upper Terrace’ sequence identified by Saunders [1985]  85  Chapter 4. Evolution of Chilliwack Valley Mainstem (‘L2b’ surface). The genesis of the Tolmie Upland is uncertain, however, assuming that the gullies within the deposit have been eroded in post-glacial time, the minimum volume eroded from the source area has been estimated to be 80 ±13 × 106 m3 . Saunders also identified a large failure in the same terrace sequence, which deposited a large quantity of material in the river valley. The net erosion there is estimated to be 29 ±4.4 × 106 m3 .  4.4.1  Ryder Lake Upland Moraine Complex  According to Saunders [1985], the Ryder Lake Upland was a principal deposition site for the Fraser Valley ice lobe, and it is likely the largest single source area in the postglacial sediment budget. The depth of the deposit is over 100 m in places, based on water well records and exposures (Figure 4-6). The compaction of the material indicates little or no secondary movement occurred after till formation. Lodgement and meltout processes were likely largely responsible for their deposition. Near the up-valley extent of the Ryder Lake Upland, there is the residual crest of a large moraine [Saunders et al., 1987]. Two important problems in determining the exact magnitude of this contribution are establishing the sequence of events that unfolded during glaciation, and the true historical bounding geometry of the deposit. The first point has been addressed by Saunders et al. [1987], who have described the terrace geomorphology of the lower valley. In their reconstruction of late-glacial valley conditions, the Fraser Lowland ice mass advanced into the valley and deposited large quantities of till. A subsequent readvance of the ice overrode and compacted the till, and left the end moraine near Tamihi Creek. The mass of subglacial till extended from Tamihi Moraine in the east to a point roughly 7 km downstream, where a delta was built out toward a large lake in the modern Cultus Lake basin. A river floodplain 250-300 m wide was built along the southern margin of the glacier. Based on this reconstruction of the Ryder Lake Upland till, the second problem (bounding geometry) is dealt with here as three possible scenarios: (i) a relatively deep deposit, making Figure 4-6 (facing page): Cross section showing the available evidence (exposures and well logs) to describe the Pleistocene and post-glacial valley fill. The modern channel is demarcated with a dashed line. Upper planform map shows detailed terrain mapping [Armstrong, 1980; Ryder, J.M. and Associates, 1995] that outlines major lacustrine and morainal strata. Ryder Lake Upland is depicted with a veneer of aeolian material and thick till accumulations beneath. Points on the plan map indicate exposures and well logs that are graphed in the section detail. Vertical exaggeration is 20x. There is assumed to be considerably more lateral complexity within the valley’s deposits (and eroded volume) than can be reconstructed with this evidence.  86  Elevation a.s.l. (m)  0  0  200  400  21 400 21 600  51506  CH-1  34588  37909  CH-1A (RC)  32336  32308  57836  ?  Ryder Upland Till Deposit  59530  28078  utwash  Older Silt Sequence  34830  O Advance  56031  Ryder Creek  Thin Aeolian Veneer  CH-2 (LS)  WATSON-9  CH-1B (SS)  35877  Youth Cen.  87  Sand  Gravel  Distance Along Transect (m)  Borden  Chubb, 1966  Diamicton  Clay  Silt  20000  11 900  11 700  ?  Campground Bluffs  Upper Valley Delta/Sandur  Gray Bluffs  Clague et al. 1997 11 400  Well Logs - Legend  Glacio-Lacustrine (Multiple Episodes)  Bedrock  Profile  Valley Wall  er Modern Riv  20 190  CH-3  Tolmie Bluffs  WATSON-10  10000  35900  Clague et al. 1997  11 600  11 800  11 200 11 300  11 200  11 500  Crest of Tamihi Moraine  Tolmie Upland  ansect  Lower Valley Tr  Ryder Upland  Outwash Exposure at Pierce Creek  Chapter 4. Evolution of Chilliwack Valley Mainstem  Chapter 4. Evolution of Chilliwack Valley Mainstem Table 4.2: Estimated bulk erosion volumes for three different assumed topographic configurations in the Lower Chilliwack Valley in post-glacial time. The true value is assumed to be intermediate between I & II. The bounds of minimum net erosion are shown in Figure 4-8  Scenario I II III  Description Valley-wide fill Intermediate fill Minimum net erosion  Estimated Net Erosion (m3 x106 ) 1 350 855 309  a valley-wide fill at the elevation of the ‘middle terrace’ of Saunders et al. [1987], (ii) a deposit of intermediate depth, closer to the mid-elevations of the lower valley, and (iii) a minimum estimate of Lower Valley erosion, closer to the valley floor. The associated volumes for each of these configurations is shown in Table 4.2. Cross-sections showing the extent of these three scenarios are shown in Figure 4-7. The third scenario appears less likely, based on the geometry of the subglacial surface on Ryder Lake Upland and Tamihi Moraine remnants, but it represents the minimum estimate of the net erosion from lateral valley deposits. Figure 4-8 shows the eroded volume from the valley sidewalls, based on a reconstruction of the deposit with a spline mesh surface (grey zone in Figure 4-7a). The wall angle of least-eroded portions of both the Ryder Lake Upland and the southern valley wall show a similar angle of repose. Fluvial reworking has likely undercut the slopes repeatedly, as the river cut down through the deposit and sediment mass was reworked by Chilliwack River. It is assumed that the lower terrace surface that dams Cultus Lake, sitting approximately 20 m above the modern Chilliwack River floodplain, was once a continuous surface across the broad lower reach of the river above Vedder Crossing (Saunders et al. [1987], see Figure 4-1). This represents between 62 ± 8 m3 x106 of gravelly material. Additional quantities of sediment delivered over the course of the Holocene from the Columbia Valley have mostly been intercepted by Cultus Lake. Compared to some of the fluxes detailed in the previous chapter, a dominant proportion of the Holocene sediment budget is derived from the former glacial valley fill. At least 1 km3 of material, possibly as much as 1.6 km3 , has eroded from the valley mainstem, representing roughly 60-75% of the sediment budget. From the perspective of the modern sediment transport rates, the fluxes required to evacuate that volume would have been much higher than modern rates. The volume of outwash material estimated to have been evacuated from lower Slesse Creek alone, for instance, represents over 1 000 years of gravel bedload transport from 88  Chapter 4. Evolution of Chilliwack Valley Mainstem  Elevation (m a.s.l.)  400  Extents of the late glacial ice mass, Chilliwack Valley Glaciofluvial surface  Ryder Lake Upland I  200  L3 ‘Middle’ Terrace  Ice  II L4 ‘Lower’ Terrace III Modern Floor  North  South  ?  0  1  2  3  4  (a) Tamihi Moraine 400  Ryder Upland  L3 Surface  Lake Elevation  I  200  II  Ice Moraine  III  Longitudinal Valley Section, Looking North 0  1 2 Cross Section Distance (km)  3  (b) Figure 4-7: Reconstructed geometry of the major valley landforms in Lower Chilliwack Valley, downstream of Tamihi Creek. (a) shows a composite set of cross sections of the valley near Ryder Creek, 3 km downstream from the crest of Tamihi Moraine. The shaded area represents the zone of minimum net erosion of glacial material (scenario III, see Figure 4-8 and text). (b) is a longitudinal valley cross-section looking North, encompassing the former locality of Tamihi moraine. A mid-valley lake elevation is shown, based on the assumed elevation of the delta front at the distal end of the Larson’s Bench sandur. Cross-section lines from Ryder Upland showing the maximum height of the remnant moraine are in the background.  89  Deposition  Chapter 4. Evolution of Chilliwack Valley Mainstem  120 m 90 60 30  Erosion  0 -30 -60 -90 -120 m  Figure 4-8: Scenario III, with minimum bulk volume erosion from valley sidewall sediment sources, including the Ryder Lake Upland. Contours of erosion and deposition are in increments of 15 m. Total erosion is 344 ±45 × 106 m3 . Deposition evident at the base of Ryder Upland is approximately 35 × 106 m3 .  the entire Chilliwack Basin (roughly ∼50 000 m3 yr−1 ).  4.5  Vedder Fan  The alluvial fan at the outlet of the Chilliwack Valley, known as Vedder Fan, is a Holocene landform that has built out over the outwash and floodplain surface of the Fraser Valley (Figure 4-9). In post-glacial time, distributary rivers have delivered a mixture of coarse sediment from the Chilliwack Valley. The switching of the channel back and forth through the action of avulsions and reactivation of secondary channels has gradually built up a semicircular cone of sand and gravels that is more than 35 m thick near the apex. The Chilliwack River was in the process of reworking and incising through accumulated glaciofluvial fill in the lower valley when the ice retreated and construction of the fan was  90  Chapter 4. Evolution of Chilliwack Valley Mainstem  Chilliwack Mountain  Elevation Valley Floor 1-4 4-6 6-8 8 - 10 10 - 12 12 - 13 13 - 15 15 - 17 17 - 19 19 - 21 21 - 23 23 - 25 25 - 28 28 - 30 30 - 32 32 - 35  -  ain unt  o er M  d Ved 0  0.5  1  2 Kilometers  Figure 4-9: Figure showing a band of elevations between 1 and 35 m a.s.l. in Sumas Valley, Vedder Fan and the Fraser River. The bounds of Figure 4-14, showing City of Chilliwack Drilling work (2006), are highlighted. Note the semi-circular geometry of the Vedder Fan, and the relatively lowlying surrounding topography. TRIM BC digital elevation data.  initiated. The lowest possible elevation of the fan apex is on the bedrock sill of a ‘V-notch’ in the range front, only a few metres below the present channel - roughly 30 m above sea level. If the glaciofluvial fill above Vedder Crossing was intact at the time that fan-building began, it is possible that the apex could have been up to 10 m higher, as discussed by McLean [1980]. Based on interpretation of the modern radial profile of the fan, it is quite likely that there has been some complex reworking of the upper region. Material from the apex has been incised and redistributed out along the fan, much in the manner described by Schumm [1977]. Judging from historical maps of the fan area (Figure 4-10, top), the Thuwlman / Lhqueleq / Qwelkwaltem channel complex (Chilliwack / Luckakuck / Atchelitz, see Bowman [1992]; Schaepe [1999]) was a multi-threaded wandering gravel bed river. There was a distributary  91  Chapter 4. Evolution of Chilliwack Valley Mainstem network of slough channels that would have been reactivated periodically as the river avulsed from one side of the fan to the other in the course of aggradation. The Sto:lo Indian word for the river at this time was Tswelmuh (Th’ewlmel, Schaepe [1999]), the term Tsel meaning to “go away”. This is an appropriate name for a river that changes its course, as the Chilliwack system often does. The historical Chilliwack River developed a network of small channels that drained westward to Sumas Lake [Orchard , 1983]. Sumas Lake was a large body of water that changed its extent seasonally, depending principally on the flows in the Fraser River. At low water, the lake was 3 m deep, with a length of 10 km and width of 6 km. During spring freshets and winter rains the lake swelled to a length of over 25 km and a depth of 10 m [Thom and Cameron, 1996]. The last of the lake waters were drained in 1924 to make way for agricultural lands. Cameron [1989] notes that silt and sand facies recovered from drill cores within the Sumas valley are quite variable, indicating a high energy environment within the lake basin at times. The lake was bounded by Nooksack/Sumas River fan deposits to the southwest and the Vedder Fan to the northeast (Figure 4-1e).  4.6  Architecture of the Vedder Fan  The basement of most of the Chilliwack/Sardis area and the Sumas Valley is a thick blue clay unit [Halstead , 1986; Cameron, 1989]. None of the drilling at the fan apex attained depths that approached this horizon. Logs from a petroleum test hole place the lower limit of gravels in the area at 24 m below grade (−12 m a.s.l.) near mid-fan, roughly 4.3 km out from Vedder Crossing. Gravels are underlain by sand to 43 m below sea level, and then by clay to 400 m b.s.l. A deep borehole further east (11.6 km from Vedder Crossing, Bos Trout Farm) found a similar stratigraphy: sand and gravel to 14 m b.s.l., sand to 52 m b.s.l., and then layers of clay, silts and fine sands to bedrock at 489 m b.s.l. Presumably the sands are glacial outwash and deltaic material, overlying the deep glacio-marine or marine fill that underlies much of the Fraser Valley [see Armstrong, 1984]. The bounds of the Vedder Fan have been delimited by Armstrong [1980], based on surface expression of sediments and an evaluation of drilling logs. This work has been supplemented by Levson et al. [1996] and Monahan and Levson [2003], who have provided greater detail in the course of reviewing available well records and geotechnical drilling reports to assess the earthquake hazard for the Chilliwack area. Maps and sections from these reports have  92  Chapter 4. Evolution of Chilliwack Valley Mainstem  Thuwélman (Chilliwack)  Qwelkwaltem (Atchelitz) Lhéqueleq (Luckakuck)  Th’éwálmel (Vedder)  Sumas Lake  Chilliwack Mountain  Vedder Canal  Figure 4-10: Historical planform of Chilliwack River north of Vedder Crossing, ca. 1891 (top) and 1991 (bottom). An interlinked network of channels (four of the larger threads are labelled) alternately occupied and abandonned various section of the fan as it evolved. The historical figure was derived from an ordnance survey of the Chilliwack area. The modern map was generated from BC TRIM mapping.  93  Chapter 4. Evolution of Chilliwack Valley Mainstem provided the foundation for the work that follows. The pattern of fan growth is probably similar to the model proposed by Blair and MacPherson (1994, Stage 1, Figure 4-11). These authors have proposed a sequence of fan development that begins with a system dominated by mass-wasting processes, gradually evolving into a low-gradient fluvial system. Stage 1 in the case of the Chilliwack would have been heavily influenced by the very large pulse of glacial material that was delivered from the catchment in the early stages of development.  Figure 4-11: Model of alluvial fan growth, after Blair and McPherson [1994]. The lower bounding surface slopes slightly upward, as deposition keeps pace with a slowly rising base level. In the case of Vedder Fan this corresponds to active deposition on the Fraser Valley floor over the course of the Holocene.  It is assumed that the construction of Vedder Fan proceeded in a regime of minimal tectonic influence, and that subsidence only occurs by compaction under the weight of the sediment delivered. The shape of the nascent fan may have been influenced by stagnant ice in the Fraser Valley; it is possible that a significant proportion of the fan material is stored in the remnant basin that became Sumas Lake. Near-surface (<5 m) apex facies were exposed in section at a number of construction sites around the town of Sardis. An example is shown in Figure 4-12, where the upper 5 m (approx.) of a section displays a coarse base of fluvial gravels with sandy fill above. The gravel packages are most commonly massive, with abundant sands in matrix-supported units. Drill holes and exposures close to the apex and mid-fan show that deep fills of bedded uniform sands are not restricted to more distal sections but can be found throughout the fan head as well (Figure 4-12 and 4-13). Thick sand layers are typically the product of overbank sedimentation and infilling of sloughs. Minor silt layers are common throughout. The outcrops shown in the figure are similar to exposed banks in the lower portion of the Chilliwack Valley, above Vedder Crossing, where the river has maintained a wandering to 94  Chapter 4. Evolution of Chilliwack Valley Mainstem braided habit. The Vedder Fan has a relatively low gradient by comparison with other larger paraglacial fans in BC. The modern proximal fan slope is 0.022; channel slopes in the Vedder River range from 0.0046 in the first 3 km below Vedder Crossing to 0.00035 in Vedder Canal.  4.7  Well-log database  A database of well logs was developed using records from the British Columbia Ministry of Water, Land and Air Protection (WLAP), and the extensive collection of geotechnical logs compiled by Levson et al. [1996]. Further additions were obtained from Emerson Groundwater and the City of Chilliwack [Emerson, 2003]; the latter maintains a network of monitoring wells and has carried out a number of drilling projects in the last ten years. Automated text recognition and lithology coding of the logs was accomplished using the LDBuilder software package that was developed by a research group led by N. Schuurman and D. Allen, Simon Fraser University. The lithology units were reviewed to ensure consistency. More than two thirds of the database was eventually discarded, leaving 255 holes that met criteria of depth (at least 6 m deep) and detail needed to develop a representative 3-dimensional model. Information was imported into RockWorks to enable 3-dimensional interpolation of the lithology units. Rockworks employs an inverse-distance weighting algorithm with a horizontal bias to interpolate among the lithologic units in the bore holes. It further uses a randomized blending technique to simulate the likely horizontal transitions between boreholes. 3-D pixels (known as voxels) used in the model measured 150 x 150 x 0.5 m. Relatively few deep holes have been drilled within the proximal, gravelly areas of Vedder Fan due to difficult drilling conditions. The subsurface information database was enriched considerably by sonic drilling carried out in 2003-2006 in order to improve monitoring of the City’s drinking water supply. Drilling efforts were focused near the fan apex; Figure 4-14 shows the locations of the drill holes. All of the cores from the drilling campaign were made available for inspection and sampling. The cores were generally deep, achieving depths of 20 m on average. The deepest cores extended to over 50 m depth.  4.8  Apex Gravels - Core Descriptions  Material from the drill cores was classified according to sediment calibre and whether the unit was matrix- or clast- supported. A small number of framework gravel deposits appeared 95  Chapter 4. Evolution of Chilliwack Valley Mainstem  Figure 4-12: Gravel quarry near Vedder Road at Watson. Looking westward: flow was from left to right. Photo credit Vic Galay, Northwest Hydraulic Consultants.  Figure 4-13: Gravel quarry near Vedder Road at Watson. Photo is looking southward, and the direction of flow was out of the page. Photo credit Vic Galay, Northwest Hydraulic Consultants. 96  572,500  573,000  0  0.25  573,500  0.5  574,000  1  574,500  575,000  575,500  576,000  1.5  5,441,000  Chapter 4. Evolution of Chilliwack Valley Mainstem  Kilometers  5,440,500  UNSW-06 MW-1999-7 MW-2005-07  BC-11424  MW-2003-3  MW-2003-2 PW-1964-3  MW-2003-4 BC-75340  MW-1995-4 5,439,500  MW-2006-2 MW-2006-1  5,440,000  MW-2003-1  MW-2006-4  ²  5,438,500  Vedder Crossing  5,439,000  MW-2005-2  Figure 4-14: Locations of City of Chilliwack drilling operations, 2003-2006 (1995 aerial photography). See Figure 4-9 for map location with respect to the larger fan area.  in the drill cores, though it was sometimes uncertain whether some of the matrix had been partly washed out in the course of removing the sediment from the core barrel. The sonic coring tends to disrupt sedimentary structure and clast orientation, and so classification remains somewhat tentative. Nevertheless, the recovered samples generally showed good representation of the sedimentary layers, and yielded valuable information such as patterns of bedding and vertical fining, lithology, and the nature of active geomorphic processes on the fan over time. This 1-dimensional view does not provide a diagnostic reconstruction of the fan history, but it does provide a few hints as to the evolution of the fan over time. Cobbles were recovered from a few units, but clasts coarser than 256 mm were generally quite rare. The barrel of the sonic core was 6 inches (152.4 mm) in diameter; perhaps two or three stones were recovered that exceed that diameter (the drill effectively bored through larger stones rather than displacing them, so that their presence, if not their full dimensions, 97  Chapter 4. Evolution of Chilliwack Valley Mainstem 0.063 mm 75  0.250 mm  1 mm  4 mm  Sand  16 mm  64 mm  4  6  Gravel  Percent Finer  50  25  0 -4  -2  0  2  Grain Size (Y)  Figure 4-15: Grain size distribution from a sampling of units recovered from sonic drilling. Representative fractions > 32 mm could not be effectively recovered from the core samples, but the fractions that were examined effectively show the bimodal nature of the fan deposits.  was reasonably well represented). Figure 4-15 shows a plot of the grain size distribution from a number of drill cores. Sands consistently had a modal size close to 0.5 mm. Much like the modern channel gravels on the fan (see Chapter 6), there is a bimodal character to most of the gravels with a deficit of material in the range of 1-8 mm. Figure 4-16 shows a cross-section of Vedder Fan based on the cores from City of Chilliwack drilling (Figure 4-14), as well as a number of other well logs (which were not viewed or sampled) within three kilometres of the fan apex, arranged as a cross section. Higher energy river environments (coarse, well-sorted matrix-depleted gravels) account for roughly 8% of the total examined core length. Approximately 35% of the record is dominated by lowenergy environments that exhibit graded sands and silts with occasional, unsorted gravel. The remainder consists of intermediate states with generally abundant matrix and varying degrees of sorting that represents a mix of bed material load and washload. Sands with occasional gravel in the lowest unit encountered (MW03-1, MW06-1 and 4) have a slightly darker hue and a relatively greater proportion of quartzite clasts than other units encountered. Pebbles found in the sediment are not otherwise much different from Chilliwack lithological makeup (similar granitic, volcanic and metasedimentary types). This unit is interpreted to be chiefly composed of reworked tills. Based on the dating evidence recovered from the cores, this is probably part of the initial pulse of till evacuated from the 98  Chapter 4. Evolution of Chilliwack Valley Mainstem Table 4.3: Radiocarbon ages from drilling at Vedder Fan. All samples were dated by conventional radiometric technique. Intercept ages and age range in calendar years before AD 1950. The age ranges in parentheses represent 1σ error limits. The ages were determined using the INTCAL98 database.  Lab Code Beta-213924 Beta-215942 Beta-215943 Beta-215944  Core Twin Rinks (S.End) Unsworth Road Watson & Tyson, MW05 Garrison Crossing, MW06  Depth (m) 16.5 11.0 27.4 36.6  Measured Age (14 C yr BP) 8 800 ± 60 5 420 ± 40 9 020 ± 100 9 700 ± 60  Calibrated Age (cal yr BP) 9 880 (9 710-9 920) 6 250; 6 200 (6 190-6 280) 10 190 (9960-10230) 11 160 (11 120-11 190)  valley. Material at depth (below −10 m a.s.l.) that is closer than 1 km to the apex is presumed to be part of the till package as well. However, there are some uniform framework gravel facies in MW05-07 (see Figure 4-17a) with well-rounded clasts that would suggest active sorting processes, and thus perhaps a lower-gradient fluvial system. The lithologies encountered here are typical of the modern upper Chilliwack River, rather than Fraser provenance.  4.9  Chronology and Volumetric Estimation  Wood was found in four different cores at the base of sand or silty units that were evidently low energy environments. Enough wood was recovered to enable conventional radiocarbon analysis. The wood at Twin Rinks (MW03-4) was recovered at a depth of 18 m below grade, however the drill hole collar was located on the floor of an old gravel quarry, adding approximately 7 m to the inferred stratigraphic horizon (see Figure 4-16). The depth and date of wood from this unit is consistent with woody material recovered from MW05-7 sited 725 m away (Figure 4-14). Dates from wood at these sites were (2σ) 9 865 ±285 and 10 125 ±245 cal. B.P. (Table 4.3). A deep hole at Chilliwack Area Services Unit (across the street, 285 m from MW05-7) yielded wood at 8 m below sea level that was dated at 11 035 ±175 cal. B.P. Isochron curves representing temporally bounded surfaces in the deposit are shown on the cross-section diagram (Figure 4-16). Their elevation is assumed to be constant within the radial swath of the fan, though this is clearly a simplification. Sequential paleo-surfaces of Chilliwack Fan were generated as a series of axially symmetric conical grids that are based on the isochrons presented in Figure 4-16. The distal slope of each surface is smoothly graded to intercept inferred elevations of the Sumas Valley and 99  -40  -30  -20  -10  0  10  20  30  a  )  m  g  1  Darker, Sandy Diamict  f  c  +/-  /-  2  14-C Date  2 mm  Sand  Estimated Modal Grain Size  8 mm Silt/Clay  Sand Clay  16 mm Gravel w/ sand matrix  Gravel  Water Wells +64mm  10 000 BP +  )  hy (TRIM  Topogra p +/-  6230 BP  Framework Gravel  11035 BP  i  e  Modern  Distance from Vedder Crossing (km)  d  b  ) ) ) ) ) ) ) ) on m m m m . m m m m ) Tys 8 it 4 g 1 . 6 63 d 4 3 28 ntre v. 2 Un .4 Rd. v. 3 res . 3 sin 9. n R v. 2 7m & .2 . 0 v v 2 2 v o 3 r le s le C le le Ce (ele ices le v. ls v. on (e Cro lev. dde (e ura (e le i (e le ts (e (e W 2 s -4 s 1 am 6-1 erv (e a -4 n 24 a (e e 3 ith 3- k 3 nk 3- e 7 W S 06 iso -4 9 V 14 3 L 4- Ke W0 Rin W0 Ri W0 Ch 200 ea -0 r of 1 63 5 1 20 arr 6 5 9 6 5 M in M in M ar 0 e 5 Ar G e W 5 PW 521 20 orn Tw N Tw M . of 4 C E  3  21 v. le (e oad 6 -0 h R t SW or N U nsw U  )  Gravel Limit  m  Figure 4-16: Cross-section of Vedder Fan, based on examination of sonic core cuttings and assembled well logs. Letters refer to locations of photos in Figure 4-17. Narrower logs with solid colours are database wells with minimal detail and have not been examined. Dashed lines indicate interpolated surfaces with an associated date based on radiocarbon samples. Most elevations are approximate, unless significant figures are indicated.  Elevation (m a.s.l.)  40  35 v. le ty (e per 2 - ro 05 P 20 FB C  Chapter 4. Evolution of Chilliwack Valley Mainstem  100  20 cm  Framework Beds  b  c  More Matrix  d  e f  Lower Strata  g h  Silt Beds  i  Figure 4-17: Photos of sonic core material, illustrating facies assemblages that are indicated in Figure 4-16. Wood dated at 10 125 ±245 cal. years B.P. was recovered from the silt unit shown in (i)  a  Chapter 4. Evolution of Chilliwack Valley Mainstem  101  Chapter 4. Evolution of Chilliwack Valley Mainstem the Fraser River floodplain. There is some evidence that supports placing an upper distal bounding surface of the fan at 6 800 cal. B.P. near 5 m a.s.l. Cameron (1989) found Mazama ash (approx. 7 680 cal. B.P.) at 0 and at −4 m a.s.l. in the Sumas Valley. A drill log on the northern distal edge of Vedder Fan records the presence of volcanic ash at 6 m a.s.l., though it is not explicitly identified as Mazama. Wood at 12 m a.s.l. was recovered on Unsworth Road, approximately 2.95 km northwest of Vedder Crossing, that was dated at 6 230 ±60 years cal. B.P. The lowest bounding surface is an approximation of Fraser Valley topography circa 11 000 B.P. The potential error inherent in defining the vertical extents of this boundary is much larger than in the horizontal (i.e. the horizontal extents are quite large relative to the vertical). If, for example, the elevation of the bounding surface was consistently overestimated by 2 m, the resultant volumetric error would be roughly 10% of the total volume. The horizontal boundaries present some difficulties as well, in that the lateral extent of the Fraser floodplain is not well constrained as the Vedder Fan builds. An accurate estimate of the lower bounds of the fan is required, and the model presented here is based on the best available evidence. The base of the unit generally conforms to boundaries that have been proposed or inferred by Cameron (1989, ‘Unit 4’), Levson et al. [1996] and Emerson [2003] and does appear to be in good agreement with most of the well logs in the database. The radial geometry of the fan may be approximated as a level, symmetric radial cone, rotated through 180 degrees. An averaged profile of the fan was generated by taking 30 radial profiles from apex to distal edge across the TRIM data in even 5.3 degree increments, extending 6 km down slope. An exponential curve was fit to the multiple topographic measurements taken at successive radii from Vedder Crossing (Figure 4-18). The conical regression surface was subtracted from the modern topography. The east side showed slightly more positive residuals (i.e. topography is generally lower) and the west side showed more negative. The gross discrepancy is a matter of perhaps 16 ×106 m3 on either side - not insubstantial, but small compared to the total volume. On the whole, errors on the regression cone cancel out to a residual error of 5 million m3 or 0.22% of the total volume.  4.9.1  Isopach Diagrams  Using the established geometry of the fan, with lower bounds near −12 m a.s.l., a 3D model was constructed using the well database described above. The possibility of a basin extending 102  Chapter 4. Evolution of Chilliwack Valley Mainstem 45  -0.0001989x  40  y = 36.21e  Elevation a.s.l. (m)  35 30 25 20 15 10 5 0  500  1000  1500  2000  2500  3000  3500  4000  4500  5000  5500  6000  Distance from Vedder Crossing (m)  Figure 4-18: Regression curve through multiple radial sections of Vedder Fan  to −30 m depth, within the bounds of the former Sumas Lake, is considered as well. Three major lithologic categories (gravel, sand, silt/clay) were distilled from many log entries, to develop a general picture of the fan architecture and the relative thickness of each major category (Figure 4-19). Gravel units include some proportion of sand, while sand units represent relatively uniform sand. The spatial density and depth of subsurface data for the study area are not optimal for this form of analysis, and a number of interpolation effects are evident. The results shown here should be considered highly generalized. A series of diagrams show the approximate thickness of lacustrine, sand and gravel units between the posited lower bounding surface of the fan and the modern topography. This volume is 40 m thick near Vedder Crossing, and tapers to less than 20 m at distal regions. The boundary of historical Sumas Lake is shown in the diagrams. Here sand and silt deposits are depicted as having a thickness of up to 40 m; based on drilling work and radiocarbon analysis by Cameron [1989], accumulations of Chilliwack-derived material may be greater. Gravel (Figure 4-19b) is predictably quite thick at the apex of the fan, and there appears to be a lobe of coarse material extending to the northeast. The bounds of this lobe are tentative, since there are relatively few deep holes to the north and northwest to constrain the true gravel thickness there. Silt and clay units (Figure 4-19c) are concentrated in four different zones, labeled on the diagram: (1) Sumas Lake deposits (2) Fraser silts and clays (3) a small but deep deposit on the southwest distal edge of Vedder Fan, and (4) a lacustrine zone in eastern Chilliwack.  103  Chapter 4. Evolution of Chilliwack Valley Mainstem  a)  Surficial Geology - after Levson et al., 1995 Proximal Fan  Distal Fan  MidFan  Lacustrine Sediments  Isopach Diagrams of Vedder Fan from Borehole Logs Depth of Fill 1m  10m  20m  30m  40m  b)  c) 4 2  1  Gravel d)  3  Silt/Clay  e) 0.00030  0.00015  0.00 Well Depth (m)  Sand  Confidence  Figure 4-19: Isopach diagrams of gravel, sand and lacustrine material within the stratigraphic bounds of Vedder Fan and surrounding area. Bounds of former Sumas Lake are shown in black. See text for description of numbered localities in 4-19c.  104  Chapter 4. Evolution of Chilliwack Valley Mainstem There is apparently a thick clay lens (6-8 m thick) at the center middle section of Vedder Fan, but this is based on a two relatively isolated core logs; there is an insufficient density of holes nearby to adequately assess the extent of this deposit. The pattern of sandy units (Figure 4-19d) shows a prominent swath of thick deposits across most of the mid- to distal fan. Shallower sand deposits are also evident at what must have been the delta of Sumas Lake. Sumas Lake itself has quite thick accumulations of sand. Spatial density of the holes varies from roughly 20 /km2 to 0.2 /km2 . A map showing the point density in the model at different points across the fan is shown in Figure 4-19e. The depth of the well log is indicated by point size. The number of well points that fall within a 1.5 km radius are weighted by their depth, then totaled and divided by the area of the circle, giving a spatially distributed index of well log coverage, and thus confidence in the subsurface model. Based on the assumed extent of the Chilliwack Fan, volumes were calculated for gravel, sand and silt/clay fractions. Typical bulk densities for these materials are assumed and density differentials with depth are taken into account. Volume has been stratified into different historical periods detailed in Figure 4-20, and resulting estimates are shown in Table 4.4. Table 4.4: Volumetric estimates of Vedder Fan composition (m3 × 106 ) based on well log records. Values are cumulative over the time intervals indicated. The overall expected accuracy of the estimates is ±16%, based on the range of possible spatial extents and depths of the fan. Values in brackets indicate additional volumes attributable to the Sumas Lake Basin.  Period to 11 035 BP to 10 000 BP to 6 230 BP to Present  4.10  Gravel 14.5 90.1 696 1 020  Sand 9.4 36.7 383 555 (765)  Clay/Silt 8.7 21.5 436 768 (872)  Discussion  The summary volume of Vedder Fan has been examined in order to better constrain the output term for the larger sediment budget model. The bulk volumetric output of gravel, sand and silt/clay fractions appears to balance most of the net eroded bulk volume that has been established in the previous chapters, to a first approximation. The chronostratigraphic model shows an initially heightened phase of growth, followed by reduced aggradation on the 105  Chapter 4. Evolution of Chilliwack Valley Mainstem 2.8E+9  Sumas Lake Deposition (?)  Cumulative Volume (m 3)  2.4E+9 2E+9 1.6E+9  Silt / Clay  1.2E+9  Sand 8E+8 4E+8 0 -12000  Gravel  -10000  -8000  -6000  -4000  -2000  0  Calendar Years B.P.  Figure 4-20: Cumulative volumetric growth of Vedder Fan from 11 035 BP to present, based on the assumed stratigraphic relationships indicated in Figure 4-16. Error bounds indicate 16% error assumed in the calculations.  fan. A large proportion of the volumetric output from the valley consists of sandy material derived from the deep glacial deposits in the lower and middle reaches of Chilliwack Valley. This would appear to constitute much of the lower fan stores. The sand and gravel volume in Vedder Fan is estimated to be 1.57±0.25×109 m3 or roughly 3.0 ×109 t, assuming some compaction and a bulk density of 1.9. Since deglaciation, this averages out to a bed material yield of 121±19 ×103 m3 /yr. The long term average since 6 200 cal. B.P. appears to be 78.8±9 ×103 m3 /yr, a figure that is higher than the mean, but within the range of modern estimates made by McLean [1980]; Martin [1992] and Ham [1996] (Table 4.5). Table 4.5: Bed material yield below Vedder Crossing - lowest, mean and highest annual averages over the course of study periods for three investigations using different methodologies  Bed Material Load Estimate McLean (1980), 1940-1976 Martin (1992) 1981-1990 Ham (1996) 1952-1991  Low (m3 /yr) 19 000 ±8 670 4 900 ±3 900  Mean (m3 /year) 58 000 ± 36 600 ±5 600 54 600 ±10 000  High (m3 /yr) 76 000 ± 156 800 ±16 700 178 000 ±10 000  The net volume of glacial material (outwash and till) that has been eroded from the valley mainstem and lower tributaries was estimated to be 1.4 ×109 m3 (middle range estimate, see above). This represents roughly 90% of the fan volume, suggesting that either the estimated 106  Chapter 4. Evolution of Chilliwack Valley Mainstem volume eroded from the lower valley is high or that transfer of material to the Fraser is higher than anticipated. I expect the answer is a combination of the two; there is large uncertainty in the bounds of post-glacial topography in the lower valley. The contribution of glacial material is nevertheless the most significant portion of the early fan’s development, the river load has been gradually supplemented with more modern weathering products over post-glacial time. For a comparison among scales, the total volumetric post-glacial yield to Chilliwack Lake, including outwash deposition, was 344.5 ± 24 ×106 m3 . The specific bulk rate of yield is thus 79 ± 9 m3 /km2 /yr over 13 000 years. The total bulk volume estimated for Vedder Fan indicates a post-glacial specific rate of yield almost twice that of Chilliwack Lake: 146 ± 23 m3 /km2 /yr for the entire basin, due mostly to erosion of very large glacial stores in the lower valley. Examination of the long-term fine sediment yield component from both Chilliwack Lake and the lower valley show a similar disparity. The amount of fine sediment identified in Vedder Fan would suggest approximate bulk specific yields of 48 m3 /km2 /yr (67 t/km2 /yr) over the long term. The Holocene lacustrine specific yield at Chilliwack Lake was 21.8 m3 /km2 /yr (32 t/km2 /yr). The deep glacio-lacustrine deposits in the mid-valley represent several centuries, quite likely millennia, of fine sediment accumulation that continues to be eroded from the valley. The Upper Chilliwack basin would have had considerable glacigenic material, but not nearly as much as the lower valley. Monitoring of the fine sediment load at Vedder Crossing shows the contemporary yield is 106 ±60 t/km2 /yr [Church et al., 1989], somewhat higher than the inferred long-term average. This is possibly due to the effects of land use within the valley. The averaging effect over the latter 6 200 years of the Holocene leaves no capacity to resolve the contributions from distinct episodes of accelerated sedimentation such as the events of the Neoglacial. Nevertheless, the transport estimates presented here assist in verifying that modern rates of sediment yield are not significantly different from those of the past, and that the aggradational process on the fan may be effectively modelled based on the bulk transport rating relations that have been established in the historic hydroclimatic regime.  107  Chapter 5 Characterization of Valley Sediments  5.1  Introduction  Previous chapters explored the total volumetric exchange that has occurred in the Chilliwack drainage network since the close of Fraser Glaciation. This chapter examines the character of the valley mainstem, in particular, the sedimentology and geochemistry of channel sediments. These investigations illustrate the longitudinal gradients of sediment calibre and fine sediment chemistry, which provide some important constraints on rates of mixing among the many sediment sources in the valley. Sediment characteristics such as grain sorting, clast lithology, and major and trace element chemistry reveal the present meta-equilibrium state of the system, which continues to evolve on a scale of centuries to millennia. It is beyond the scope of this thesis to characterize the full range of depositional environments in the valley, but some important patterns emerge from representative sites. Tributaries deposit sediment into a relatively confined valley, and have built extensive fan deposits that are inter-stratified with mainstem alluvium. Sediment is also recruited to the mainstem from large deposits of floodplain and glacigenic material along the length of the channel. The calibre and lithology of materials supplied to the mainstem vary according to the source, and there tends to be a dispersal trail of both coarse clastic material and fine sediments downstream of each major source or tributary junction. This chapter endeavours to examine the relative influence of the major tributaries upon the river below Chilliwack Lake, as observed in the patterns and gradients of grain size, lithology and geochemistry. As outlined in previous sections, the middle mainstem section and lower tributaries of Chilliwack Valley have been in a prolonged state of degradation; incision of the post-glacial deposits has proceeded to a depth of over 80 m in the mainstem, and tens of metres in the 108  Chapter 5. Characterization of Valley Sediments lower tributaries. In the middle mainstem the river has abandoned its broad glacial outwash floodplain, carving itself a narrow canyon within Pleistocene deltaic and lacustrine beds. A fining trend is evident in the valley [McLean, 1980] though the pattern is subject to frequent interruptions by lateral inputs from individual tributaries. With the extensive reworking of the alluvial deposits and the incorporation of glacial cobbles and boulders, a relatively coarse lag has developed on the bed of Chilliwack River. The lag deposit has an important influence on the evolution of the channel long-profile in that it increases the hydraulic roughness of the channel, and may also contribute to increased concavity of the long profile (see Chapter 6). Steep, bouldery riffles can be found on the downstream edges of some tributary fans; coarse accumulations of boulders tend to develop up to 1 km downstream of the major tributary fans. Much of the material coarser than the 128-256 mm (i.e. boulder) fraction appears to have been effectively trapped within the valley (escaping eventually after weathering or abrasion), while the finer grades are carried beyond the valley to the piedmont fan. Thus there is a marked partitioning of the sediment budget with respect to differential sediment mobility acting over long periods.  5.2  Sampling of Tributary and Mainstem Gravels  Surface (Wolman) and subsurface (bulk volume) sampling was carried out in Chilliwack Valley in July and August, 2004. Bars were typically sampled at low flows when the maximum surface area was accessible for sampling. Representative surface and subsurface sampling of gravel bars of the mainstem and lower tributaries is wrought with a number of difficulties. Bars represent a palimpsest of different depositional environments [Kellerhals and Church, 1978; Bluck , 1982; Shaw and Kellerhals, 1982; Rice, 1996], from sheets of gravel to blankets of fines laid down in waning flows to broad fields of static armour deposited by much larger, historical flows. The variability across the width of the broad, wandering channels of Chilliwack Valley is often considerable. Ideally, surface sampling sites are large enough to sample 400 stones in a 12 m2 (0.5 m grid) area, yet remain within a homogenous facies unit. Following recommendations of Shaw and Kellerhals [1982], Rice [1996], Bunte and Abt [2001] and others, sites were selected that represented the coarsest active environment that could be isolated, irrespective of location on the bar surface, i.e. not necessarily at bar heads. The grain size distribution is invariably  109  Chapter 5. Characterization of Valley Sediments influenced by relict materials. Clearly armoured areas were avoided, though lag boulders and cobbles were frequently present, potentially unduly skewing the estimate of the surface composition. Some bars have material that was finer than the ambient river substrate, due to the deposition of transient bedload gravel sheets. Some sites were clearly influenced by engineering works that have either disrupted the substrate or have changed local flow conditions. Despite reasonable measures taken to target active channel gravels, the grain size distribution measured is not always exclusively the product of modern channel processes. Grain size samples were taken from 18 sites in two major tributaries (Foley and Slesse), 12 sites in the Upper Chilliwack River and 32 sites along the maintem below Chilliwack Lake. Photographs were taken of all bar facies, using a 1 m2 photo template for scale. Grid-by-number (Wolman) samples were made using a 30 m tape measure. Stones that fell under each metre increment (greater spacings were used in the presence of larger caliber material) were identified by lithology and sized using a square-hole gravel sizing template with 1/2Ψ intervals. The finest clast size category was 8 mm, and patches of finer material were noted as ’fines’. For optimal characterization of the overall grain size distribution, up to 400 observations were recorded where sufficient bar area was exposed. Sampling grids were not always square, but an attempt was made to follow a grid pattern that matched the boundaries of the targeted depositional unit with minimized longitudinal or transverse bias. A representative sample of 3-4 kg of surface material, including fines and grains up to 22 mm was collected by hand. This was returned to the lab for sieving and determination of the ’fines’ content of the surface fraction and refinement of the overall distribution. A hybrid grid-based surface grain size distribution was generated following the procedures developed by Fripp and Diplas [1993] and outlined in Bunte and Abt [2001]. Axial measurements of boulders larger than 256 mm were made in order to assess the distribution of the coarse population of clasts within the sampling unit. This information was used to improve the representation of coarse material in the upper tail of the size-distribution curve. At most sites, subsurface bulk (volume-by-weight) samples were collected subsequent to Wolman sampling. The surface layer was removed to a depth of the largest stone in the active surface layer of the sampled unit. Material from the subsurface was then excavated until 750 to 1200 kg were collected (the largest sample was 1800 kg). The largest stones on the bar averaged about 40 kg or roughly 4% of the total sample weight. All except for three exceptionally coarse sites met the 5% precision criterion (for gravels with clasts greater than 128 mm) specified by Church et al. [1987].  110  Chapter 5. Characterization of Valley Sediments  Figure 5-1: Sampling bar material in the lower Chilliwack Valley (site # 111-04)  Subsurface samples were sieved in 1/2 Ψ intervals in the field to 16 mm (Figure 5-1). The remaining finer material was split and a small sample (3-4 kg) was returned to the laboratory for further analysis. Weight corrections were applied to the bulk sample measurements based on the water content of the subsamples. Corrections were typically on the order of 4-6%. Sampling frequency of bar material was better than 2-3 sites per kilometre in lower valley reaches, and as seldom as one site in a 4 km stretch through the canyon where access was difficult and exposures of bar material were infrequent. Figure 5-2 shows the locations of gravel sampling sites in the valley. Two sites were chosen to assess how representative a subsurface grain size grading at a site may be, relative to upstream and downstream links. The sites chosen were (1) just upstream of the Tamihi Bridge (113-04), and (2) a large bar complex upstream of Borden Creek (135-04). Replicate samples were taken from within the same facies unit, although the duplicate site was invariably somewhat finer. The pit was dug to the usual depth, and results showed similar curves. Figure 5-3 shows the differences among each grain size fraction for 4 111  180-04  190-04  87-04 87-04  156-04  88-04 88-04  > ! > > ! ! ! > > ! ! > 112-04 112-04 E E  625 1,250  2,500  3,750  > ! 162-04  E ( E > !  153-04  5,000 Meters  176-04  414-04  413-04  FOL9  FOL8  ( (  AIR3  408-04  E E E E  409-04  410-04  > !  406  405  > !  166-04  ((  FOL6  FOL10  ETRC#1 ETRC#1  FOL5  172-04  AIR1  418 418  407-04  138-04  > !  !( >  FOL12  EEEE  ETRC#2 ETRC#2  > ! > !  122-04 122-04  E E ETRC#3 E ETRC#3 E E E  > ! ! > ! > ! >  422-04 422-04 93-04  E EE EE > ! E  114-04 114-04 113-04 113-04  E EE E 420 420 419 419  421-04 421-04  FOL14 FOL13  106-04 106-04  > ! > > ! ! ! > > ! ! > >! ! > ! ! > >  E  163-04 424-04 163-04 423-04 424-04 082-04 423-04 082-04 118-04 118-04  150-04 150-04 E 95-04 148-04 95-04 148-04 131-04  > ! >  142-04 142-04  > !  169-04  Bulk - Wolman - Geochem  Wolman - Geochem  Bulk - Wolman  Wolman  Geochem  Sampling Points  E 404403 E E E  FOL12  E (  134-04 134-04  > ! >! ! > ! > > ! > ! > ! > ! > 135-04 > 128-04 135-04 128-04  125-04 125-04  Figure 5-2: Sampling sites within Chilliwack Valley. (a) Lower Valley and (b) Upstream of the Slesse Creek confluence. Open circles indicate surface sampling only, and filled circles indicated both surface and bulk sampling.  0  >  > > ! ! 144-04 > > ! ! > > 135-04 ! ! 175-04 > 142-04 >  134-04  92-04 92-04  98-04 98-04  85-04 ! 85-04 >  96-04 107-04 107-04 99-04 99-04 96-04  91-04 91-04 97-04 97-04  > !> > ( ( > > 90-04 90-04  > ! > !  83-04 83-04  Chapter 5. Characterization of Valley Sediments  112  Chapter 5. Characterization of Valley Sediments 16  32  64  128  256  6  7  8  64  128  256  6  7  8  10% Duplicate U/S - 093-04 U/S - 163-04 U/S - 123-04 D/S - 095-04 D/S - 094-04 D/S - 107-04  Percent Difference  6%  2%  -2%  -6%  Upstream of Tamihi Bridge -10% 4  5  Grain Size (Y) 16  32  10% Duplicate U/S - 153-04 U/S - 156-04 U/S - 180-04 D/S - 129-04 D/S - 126-04 D/S - 123-04  Percent Difference  6%  2%  -2%  -6%  Upstream of Borden Creek -10% 4  5  Grain Size (Y)  Figure 5-3: Duplicate samples taken at (a) just upstream of the Tamihi Bridge (113-04), and (b) a large bar complex upstream of Borden Creek (135-04). Lines indicate the relative difference among samples for each size fraction, relative to the first of the two duplicates (‘0’ datum). Samples are not significantly different within each sedimentary link, but show a closer affinity than between links.  to 7.5 Ψ, relative to the first sample of the two duplicates. The two sets of duplicate samples were taken at the upstream and downstream boundaries of the sedimentary link that spans the length of river between Slesse Creek and Tamihi Creek on the mainstem. There were 6 bulk samples taken along intervening bars, at an average spacing of 1.5 km. Figure 5-3(a) shows the duplicate sample taken upstream of Tamihi bridge. All fractions for the duplicate sample tend to plot within the range of variance for samples taken upstream (i.e. within the sedimentary link). The samples taken upstream of Borden Creek show more of a mixture, however, here also, there appears to be more of an affinity for the within-link mixtures, particularly in the finer range. Thus, while bulk samples generally show a more stable down-valley fining pattern than surface samples, it must be kept in mind that the at-a-site variability may be at least as great 113  Chapter 5. Characterization of Valley Sediments as the within-link variability, particularly for the coarser (>64 mm) range of grain sizes. The cumulative grain size distributions for tributary and mainstem gravels are shown in Figure 5-4 (a-e). This includes samples from Slesse Creek, Foley Creek and Upper Chilliwack River. Within the mainstem samples, it can be seen clearly that the surface samples overall show much greater variability than the subsurface samples, particularly in the gravel range. The range of standard deviation (σ) values for the individual surface samples shows a wider range than the subsurface samples (Figure 5-5). Many surface samples are better sorted than the subsurface and show much greater skewness and kurtosis. This effect is reduced when the fine fraction (<2 mm) is trimmed from the distribution, and the fractions re-normalized (Figure 5-5(b)). The bed grain size distributions show relatively poor sorting (σ = 0.65 to 2.2), which is well within the range of sorting coefficients reported from other gravel bed river studies [see Ferguson and Paola, 1997]. Bulk samples of glacial deposits were taken at several sites in the valley for an approximate characterization in the sediment routing model (Chapter 6) and for contrast with the channel sediments. Not surprisingly, the texture is highly variable and therefore difficult to characterize at the basin scale. There tends to be a richer sand component, with a mode in the 1/2 to 2 mm in range, often with a secondary mode in the coarser fractions. Examples of several deposits are shown in Figure 5-4(e) for comparison with channel sediments. As fluvial networks develop over glacial parent material, selective transport carries away sands and silts leaving a concentration of coarser clasts, illustrating their relative ubiquity within the deposits. Their true percentage is nevertheless quite difficult to accurately characterize from these relatively small bulk samples. When compared to channel sediments, the till deposits sampled have similar variance, but lower skewness and kurtosis, indicating a lack of fluvial sorting (Figure 5-5). Positive skewness indicates a deficit of finer material (<1 mm). Only a few of the tills show any sign of a distribution skewed toward the finer material; all of the fluvial material in tributaries and the mainstem are skewed to the coarser end. The highest skewness and kurtosis values are attributable to samples taken higher in the catchments where a lag of bouldery material emphasizes the coarse mode or even introduces a pronounced third mode to the distribution.  5.2.1  Fining Patterns  The geometric mean grain size of surface (composite Wolman) samples from alluvial reaches along the length of the mainstem ranges from 28 to 200 mm. The spatial scale of sampling is 114  Chapter 5. Characterization of Valley Sediments  b) 0%  90%  10%  80%  20%  70%  30% 40%  60%  Silt  50%  Sand  Boulders  Gravel  50%  40%  60%  30%  70%  20%  80%  Cobbles  10%  Mainstem - Bulk Sample  100%  0%  80%  20%  60%  40%  40%  60%  20%  80%  Percent Coarser  100%  Percent Finer  Mainstem - Wolman Sample  Percent Coarser  Percent Finer  a)  90%  0  0  100% -6  -4  -2  0  2  4  6  8  10  12  100% -6  -4  -2  0  Grain Size (Y)  c)  2  4  6  8  10  12  Grain Size (Y)  d)  Tributaries - Wolman Sample  100%  0%  100%  80%  20%  80%  Tributaries - Bulk Sample 0%  20%  60%  20%  80%  0 -4  -2  0  2  4  6  8  10  40%  60%  20%  100% -6  40%  80%  Slesse Ck.  0  12  Percent Coarser  40%  60%  100% -6  -4  -2  0  Grain Size (Y)  2  4  6  8  10  12  Grain Size (Y)  Percent Finer  e)  Till Sources - Bulk Sample  100%  0%  80%  20%  60%  40%  40%  60%  20%  80%  0  Percent Coarser  40%  Percent Finer  60%  Percent Coarser  Percent Finer  Foley Ck.  100% -6  -4  -2  0  2  4  6  8  10  12  Grain Size (Y)  Figure 5-4: Grain size distributions for bulk and Wolman samples within the Chilliwack mainstem (a,b), Tributaries (c,d) and Glacial Tills (e).  115  Chapter 5. Characterization of Valley Sediments  Mean - Standard Dev.(σ)  a)  Skewness - Kurtosis  4  12  Channel Subsurface Channel Surface Till / Glacial Sources 9  Kurtosis  Standard Deviation  Full Distribution  3  2  1  6  3  Channel Subsurface Channel Surface Till / Glacial Sources 0  0 -4  0  4  8  -1  0  Mean Grain Size (Y) 4  2  3  2  3  12  Channel Surface Channel Subsurface Till / Glacial Sources  Channel Surface Channel Subsurface Till / Glacial Sources  9  Kurtosis  3  Standard Deviation  Truncated Distribution  b)  1  Skewness  2  1  6  3  0  0 -4  0  4  Mean Grain Size (Y)  8  -1  0  1  Skewness  Figure 5-5: Mean, standard deviation, skewness and kurtosis of gravels (Ψ scale) sampled in Chilliwack Valley. (a) shows the statistical moments of the full distribution, including fines, (b) shows the same statistics with the distribution truncated at 2 mm (no fines).  116  Chapter 5. Characterization of Valley Sediments sufficient to provide only an approximate picture of the fining pattern between the tributary sources (Figure 5-6). Despite scatter, there is a coherent pattern of downstream fining, particularly within major sedimentary links. The grain size distributions in the middle and upper portions of the mainstem below Chilliwack Lake are largely controlled by lateral inputs from tributaries and Pleistocene deposits, however once the channel passes the coarse-grained tributary junction with Slesse Creek, a more persistent fining pattern develops. The fining pattern among various grain sizes can be best appreciated when viewed as individual cumulative fractions, plotted along the length of the valley (see Figure 5-6a). There is minimal influence exerted on the subsurface fining pattern by coarse tributary inputs and other sediment sources when compared to the gradient for the surface grain size distribution. The surface pattern also shows a somewhat steeper rate of fining. There is a step-change in the surface pattern at Slesse Creek, where the material becomes considerably coarser. The fining coefficient (α in the equation D/Do =exp(-αx), accounting here for both sorting and abrasion) for surface material between Chilliwack Lake and Slesse Creek is 0.020. The fining rate steepens between Slesse Creek and Vedder Canal, and is estimated to be 0.081 for the surface D50 and be 0.033 for the subsurface D50 . The discrepancy is much less for the D84 : 0.036 for surface and 0.034 for subsurface. These fining rates are all significantly less for the degrading mainstem channel than for the aggrading Vedder Fan, where rates are on the order of 0.2 [Ferguson et al., 2001]. Figure 5-6(c) shows the same section of the lower valley, this time with the mid-reaches of Slesse Creek (14-25 km) shown. Here it can be seen that there is coarse material in the lower reach of Slesse Creek that feeds into the mainstem, explaining the step-change in the previous graph. Material upstream of the bridge at Lower Slesse Creek is dominated by boulders, with patchy collections of gravel; thus a representative grain-size distribution is difficult to obtain upstream of 16 km. The subsurface grain size trend shows a relatively stable fining pattern between Foley Figure 5-6 (facing page): Downstream fining along Chilliwack Valley Mainstem. (a) shows the cumulative fractions of the subsurface sediments from Chilliwack Lake to the end of the gravels in Vedder Canal. Samples from Vedder Crossing to the canal were taken by Y. Martin and BC Environment (1989-1991). (b) shows the fining pattern for surface samples. No samples were taken beyond Vedder Crossing. (c) is similar to (b), except the fining gradient is shown from the mid-reaches of Slesse Creek (grey shading) to Vedder Crossing. The comparison shows clearly the geographic source terrain of the coarse material in the mid-reaches of Chilliwack River. Dashed lines indicate regression fits to the data downstream of Slesse Creek.  117  Vedder  Liumchen  Tamihi  Borden  Slesse  Foley Chipmunk  Centre  100%  Percent Coarser  Nesakwatch  Chapter 5. Characterization of Valley Sediments  2 mm 4 mm 8 mm 16 mm 32 mm  75%  64 mm  50%  25% 128 mm 256 mm  0 0  8  16  24  32  40  48  56  64  Distance Downstream from Chilliwack Lake (km)  Percent Coarser  100%  2 mm 4 mm 8 mm 16 mm 32 mm  75%  64 mm  No Data Collected  50%  25% 128 mm 256 mm  0 0  8  16  24  32  40  48  56  64  Distance Downstream from Chilliwack Lake (km)  Percent Coarser  100%  2 mm 4 mm 8 mm 16 mm  75%  32 mm  50%  No Data Collected  64 mm  25%  128 mm 256 mm  0 8  16  24  32  40  48  56  64  Distance Downstream from Slesse Headwaters (km)  118  Chapter 5. Characterization of Valley Sediments Creek and roughly 4 km below Vedder Crossing, with an evident spike just below Slesse Creek, where a bouldery section interrupts the pattern. There is another spike at 40.8 km, which may possibly be attributable to engineering works on one of the bars that was sampled. The coarse distribution at 45.3 km is due to a large lag deposit below Liumchen Creek. Material below Vedder Crossing was collected by Martin [1992], who sampled 19 bars along the length of the Vedder River and canal. The data were supplemented in Ferguson et al. [2001] with samples from the Vedder Canal in 1998 [Bloomer , 2000] and seven subsurface samples taken by BC Ministry of Environment personnel after a major flood in 1989. This distribution has potentially changed slightly in the intervening twelve years, as the gravel from the Chilliwack River progrades into the canal. The overall pattern shows the rapid diminution of grain sizes at the distal edge of Vedder Fan. Figure 5-7 shows the percentage of sand content (material < 2 mm) in surface and subsurface gravel samples. There is a gradual increase in sand content moving downstream, likely related to recruitment of sandy bank material and Pleistocene deposits, as well as clast abrasion. There is more scatter in the relation for surface samples, which is not surprising given the considerable variability of surface facies. The higher surface sand content tends to be found in the mid- to lower reaches.  5.3  Lithology and Geochemistry  As part of the sediment budget investigations, clast lithology was recorded in the course of collecting surface Wolman samples. There is a systematic decrease in the granitic (granodiorite) content of surface sediments from headwaters to valley mouth. The geochemistry of fine, interstitial sediments from the subsurface was also analysed in order to establish if some partitioning of the budget could be established by developing a mixing model based on major and minor element “fingerprints”. Given that the geological contrast between headwater areas and the lower valley is relatively high, it was anticipated that some measure of the rates of downstream sediment mixing could be obtained.  5.3.1  Coarse Clast Lithology  Granitic clasts provide a good marker lithology along the length of the valley below Chilliwack Lake (see Physiography section, Chapter 1). There is an exponential decline in the concentration of coarse granitic clasts within the channel from headwaters to piedmont. 119  Chapter 5. Characterization of Valley Sediments  Vedder Crossing  40%  Subsurface Sand Surface Sand 30%  20%  10%  0 0  10  20  30  40  50  60  Distance Downstream from Chilliwack Lake (km)  Figure 5-7: Percentage sand (<2 mm) in surface and subsurface deposits, plotted along the length of Chilliwack River.  Metasedimentary lithologies such as phyllites, slates and cherts are the most voluminous additional lithologies downstream, followed by various volcanics, metamorphic rocks, including an assortment of greenstone, gneiss, conglomerate. Figure 5-8a shows the pattern of dilution of coarse clastic granitic material along the mainstem. The dispersal pattern of granite is primarily due to the renewal of the sediment load by clasts of Chilliwack and Cultus Group lithologies and the exchange of granitic material within the storage reservoirs along the length of the drainage over time. The percentage of upstream granitic terrain, versus metamorphic terrain, was calculated at each step along the length of the mainstem, using ArcView’s Flow Accumulation algorithm. Area accumulation begins at Chilliwack Lake and in the headwaters of the lower valley tributaries. At each tributary junction, there may be a positive or negative step in relative percentage of granitic terrain upstream, and downstream from Foley Creek the balance is less than 50%. The longitudinal pattern of granite percentage in the channel shows a consistent renewal of granitic clast composition near each tributary junction, followed by downstream attenuation. Despite a decrease in the balance of upstream granitic terrain at tributaries such as Foley 120  Chapter 5. Characterization of Valley Sediments  32-90+ mm  Non-Granite  Liumchen  Tamihi  Borden  Upstream Fractional Area  50%  Centre  75%  Slesse  Percent Granite Composition  128+ mm  Foley  Nesakwatch  100%  Slesse Creek 25%  8-32+ mm  Granite  a)  Foley Creek  0 0  8  16  24  32  40  48  56  Distance Downstream from Chilliwack Lake (km)  60  25%  Grain Size (mm)  Liumchen  Tamihi  Borden  80  Nesakwatch  50%  0 0  Slesse  Grain size diminution á = 0.04  100  Foley  75%  Centre  100%  Mobile Fraction Lag Fraction  b)  40  20 8  16  24  32  40  48  56  Distance Downstream from Chilliwack Lake (km)  Figure 5-8: (a) Downstream variation in percentage granite composition of streambed sediments. The size composition of the total granite percentage is broken into 3 classes: 8-22 mm, 32-90 mm and 128 mm and larger. Relatively few granite pebbles were recorded, while a large number of boulders are evident. The dark line indicates the relative percentage of granitic terrain upstream (not including the drainage above Chilliwack Lake). (b) Shows the percentage of granitic clasts in the 32-90 mm category relative to both upstream granitic terrain (black) and percentage of granitic boulders (light gray). This is an index of how much mobile granitic material we might expect to see in the channel versus the observed. See text for discussion. The Sternberg relation for diminution of grain size from 90 mm to 32 mm is shown (dotted line, axis to the right).  121  Chapter 5. Characterization of Valley Sediments and Tamihi, there is still a recharge of granitic material. Most of the granitic clasts counted belong to a lag population (128 mm and greater), while the size fraction that is associated with coarse active bedload (32-90 + mm) is relatively depleted in granites. The finer fractions (8-32 + mm) form a smaller percentage yet. Smaller clasts are more mobile, and thus move through the system at a greater pace. The cumulative percentages from the three size classes are plotted in Figure 5-8a. The factors that influence the basin-wide lithological patterns are numerous and complex. In glacial terrain, the reworking of mainstem material such as glacial outwash, floodplains, fans and tills complicates the picture further. There is a large accumulation of granitic boulders at the outlet of Foley Creek, for instance, that is derived primarily from an exposure of relict sandur that is closely coupled with the mainstem. The volume of remnant glacial deposits and floodplain storage is considerable in relation to volumes of sediment transfer and therefore the dispersion pattern is persistent and not subject to transitory perturbations. The balance of granitic terrain upstream appears to explain most of the lithologic composition in the coarser fractions. That is, granitic boulders appear on the bed roughly in proportion to the relative area of granitic terrain upstream. The behaviour of the 32-90 + mm fraction is more complex. Figure 5-8b shows the percentage of granite in this size range, normalized by the boulder range (gray line) and by the percentage of granitic terrain upstream (black), which are roughly similar. This provides an approximate index of how much mobile granitic material we might expect to see in the channel versus the observed. The graph reveals a trend of relative depletion above Slesse Creek, and then of enrichment downstream to Vedder Crossing. Fining processes may be responsible for the relative diminution of headwater granites in the upper portion of the mainstem between Chilliwack Lake and Slesse Creek. Sampling of surface gravels suggests a Sternberg coefficient of 0.02 for surface D50 within that reach, and 0.081 downstream of Slesse Creek (see above). If a 90 mm clast were worn down to less than 32 mm in 25 km, the Sternberg coefficient would be just over 0.04 which is high for the river above Slesse Creek, but within range of observed rates in the system. Downstream of Slesse Creek, the relative supply of granite material is replenished, presumably due in part to reactivation of glacial material from the numerous sources along the length of the mainstem.  122  Chapter 5. Characterization of Valley Sediments  5.4  Silt Geochemistry  The geochemistry of modern channel subsurface interstitial silts shows patterns that are similar to those found in the above analysis, particularly with respect to the interaction among sediments derived from headwater areas and those from tributary sources. The influence of tributaries sources on the mainstem geochemistry is relatively strong, particularly upstream of the Slesse Creek confluence. Below Slesse Creek, the geochemical patterns become more stable, again emphasizing the importance of the reworking of stored material along the valley mainstem. The silt fraction (< 63µ) was chosen for analysis since it travels in suspension and is generally well mixed within downstream sedimentary reservoirs. Many studies have shown that the geochemical composition of silts offers a reasonably integrated picture of the fine sediment sources upstream (Cullers et al. [1987]; Shilts [1995]; Collins et al. [1997]; Dirszowsky [2004]). The capacity for differentiating upstream contributions of course depends strongly on the contrast in bedrock and parent materials found within each catchment. Some sources may tend to be over-represented, while others with weaker contrast may not be detected.  5.4.1  Methods  In the course of sampling subsurface material in bars, sandy material was collected by hand from just below the active layer and put aside for geochemical analysis. Additional samples were collected at intervals along the mainstem, in tributaries, and in a number of prominent sediment source areas such as the glaciolacustrine exposures of the lower valley. Samples collected from sediment source areas were taken from undisturbed, unweathered C-horizon (parent material) deposits. At several sites, 4-5 samples were collected within a span of approximately three to four channel-widths in order to examine the natural variability of the geochemical signature in different fluvial environments. A total of 20 of 89 samples were collected as field duplicates. Samples were oven dried at 100◦ C for 8 hours, then disaggregated (where necessary) and sieved to 63µ. Eleven blind duplicates were prepared for submission to different assay laboratories in order to assess possible error inherent in laboratory work. Samples were submitted to a suite of geochemical analyses using inductively coupled plasma mass spectrometry (ICP-MS) at both ACME labs and ALS Chemex (Vancouver). ALS Chemex offered a 27 element package (ME-ICP61), and ACME provided 37 elements in  123  Chapter 5. Characterization of Valley Sediments their Group 1EX package. Both laboratories performed a total digestion of the samples by 4-acid (HNO3 -HClO4 -HF-HCl) techniques. Element determination was done using ICP-MS. Control reference standards and analytical duplicates were routinely inserted into sample suites to monitor and assess accuracy and precision of laboratory analytical results. Upon examination of the results, Ag, Be, Bi, Cd, Mo, Sb and W were discarded outright, as they were close to the detection limit and did not offer good discriminating potential. From an initial dataset of 27 element variables, 20 are then left for analysis. The performance of these variables was analysed for both analytical and sampling errors. Using the method of Thompson and Howarth [1978], the analytical error was estimated from 11 duplicate sediment samples that were submitted to different laboratories. Precision was better than ±10% at the 95% confidence level for most elements. Cr, Mn, Ti and V were close to that precision threshold, and As, Co, Mo, and S were definitely over that limit. Field duplicates from three different sampling sites (Center Creek, Tamihi Creek and the mainstem below Liumchen Creek) were examined in order to establish the range of within-site variability. One-way analysis of variance (ANOVA) was used to verify whether the variability of the element concentrations within sampling sites was greater than between the sites. The null hypothesis (no difference between sites) was rejected for all elements except Pb (α=0.05). As, Cu, and P generally had low F scores. Post-hoc multiple comparisons using Scheff´e’s test indicate that Mg, Mn, Ni, and Sr remained the best discriminating elements for these sites, having significantly different means at the α=0.01 level. Ca, Co, and Na all failed to reject the null hypothesis of differing means at α=0.10. Element concentration data were first log-transformed and centered to a zero mean to correct for the characteristic skew of geochemical data. Correlation among the elements was then examined (see Table 5.1). There are strong (>0.7) positive associations amongst metal elements Fe, Co, Cr, Mg, Ni, V and Zn. Metals and iron oxide minerals are introduced in distinctive proportions from the various source terranes in the Chilliwack Valley, particularly the mid to lower valley tributaries such as Foley, Slesse, Tamihi and Liumchen. Potassium is an effective proxy for plutonic felsic materials in the valley, particularly orthoclase feldspar, a very common mineral in granitic terrain. Basins within the Chilliwack Batholith such as Centre Creek and the Upper Chilliwack Valley show greater concentrations of K, Al, Na and P. Positive, though weaker, associations can be found among these elements.  124  Chapter 5. Characterization of Valley Sediments Table 5.1: Correlation scores among 19 elements in fine grained (< 63µ) sediments (both active channel and hillslope sources) Chilliwack Valley. Values greater than 0.65 or less than -0.65 are highlighted.  Al As Ba Ca Co Cr Cu Fe K Mg Mn Na Ni P Pb Sr Ti V Zn  Al As Ba Ca Co Cr Cu Fe K Mg Mn Na Ni P Pb Sr Ti V Zn  Al 1.000 0.182 0.337 -0.114 -0.008 -0.322 0.152 0.227 0.680 0.122 0.441 0.366 -0.257 0.467 0.384 0.079 0.064 0.032 0.363  As  Ba  Ca  Co  Cr  Cu  Fe  K  1.000 0.222 -0.179 0.493 0.248 0.339 0.559 -0.010 0.476 0.489 -0.371 0.311 0.248 0.253 -0.327 0.351 0.389 0.505  1.000 -0.470 -0.011 -0.049 0.279 0.359 0.341 0.018 0.365 -0.226 0.011 0.050 0.277 -0.367 0.382 0.247 0.653  1.000 0.042 0.250 -0.062 0.083 -0.181 0.341 0.038 0.403 0.180 0.341 -0.344 0.456 0.058 0.141 -0.420  1.000 0.665 -0.031 0.518 -0.432 0.772 0.333 -0.613 0.772 0.031 -0.157 -0.387 0.324 0.524 0.139  1.000 0.004 0.568 -0.701 0.773 0.140 -0.487 0.945 -0.054 -0.492 -0.343 0.554 0.666 -0.014  1.000 0.370 0.116 0.130 0.447 0.123 -0.037 0.324 0.384 -0.237 0.222 0.276 0.760  1.000 -0.202 0.703 0.665 -0.316 0.512 0.379 -0.149 -0.525 0.787 0.910 0.485  1.000 -0.336 0.228 0.533 -0.588 0.408 0.652 0.394 -0.318 -0.454 0.240  Mg  Mn  Na  Ni  P  Pb  Sr  T  V  Zn  1.000 0.476 -0.289 0.817 0.292 -0.221 -0.297 0.522 0.674 0.171  1.000 0.000 0.209 0.524 0.179 -0.243 0.523 0.456 0.602  1.000 -0.538 0.439 0.183 0.596 -0.236 -0.382 -0.174  1.000 -0.057 -0.380 -0.291 0.444 0.551 0.018  1.000 0.211 0.136 0.245 0.249 0.236  1.000 0.213 -0.367 -0.352 0.503  1.000 -0.573 -0.602 -0.466  1.000 0.785 0.388  1.000 0.337  1.000  125  Chapter 5. Characterization of Valley Sediments  5.4.2  Factor Analysis  Two trends appear to be at work in the geochemical data: a ‘metals’ component and an ‘aluminosilicate’ component. A bivariate plot showing Fe vs. K (Figure 5-9) gives the best initial separation of the various deposits in the Chilliwack Valley. Borden Creek lithologies do not appear to have significant intrusive igneous minerals or iron oxide concentrations. A variety of multivariate techniques were explored to uncover the underlying structure of the element variables in a framework that was consistent with the lithological environment of Chilliwack Valley. R-mode factor analysis using principal components was used because results could be interpreted in a clear geological sense. Given the poor discriminating ability of many trace elements (described above), an optimum set of 11 elements was used for the factor analysis: Al, Ba, Ca, Fe, K, Mn, Na, P, Ti, V, Zn. It was found that three factors account for the majority (∼80%) of common variance in the data (Table 5.2). The first factor (“Fe-factor”) strongly weights the metal-associated elements: Fe, Mn, Ti and V. The second factor (“K-factor”) applies a large, negative weight to the aluminosilicate component: Al, K, Na, and P. The third factor places largest weights on Ba, Ca and Zn, though the geological interpretation of this factor is not as straightforward. A Varimax rotation of the scores emphasizes the elements further. The first two factors correspond reasonably well to the two major lithologic groups in the valley: metasedimentary rocks of the Chilliwack and Cultus Groups versus granitic rocks from the Chilliwack Batholith. Minerals containing elements of the metals group (Fe, Mn, Ti and V) can be found in both lithologic groups, though concentrations appear to be systematically higher in the metasedimentary terrain, particularly Slesse Creek. Figure 5-10(a) shows the excellent separation of the bedrock sources based on the sediment factor scores. 5-10(b) shows a directional plot of the factor scores. Figure 5-10(c) shows the geochemical data plotted along the two factor axes. Samples from Chilliwack Valley mainstem channel plot along the diagonal axis, showing the balanced influence of the two sources on the composition of the fine sediment. The tributary influence in the upper mainstem is stronger than it is further downstream, below Slesse Creek. Samples taken from above Slesse Creek (tied by a line and shown with their chainage in kilometres above Vedder Crossing, Slesse = 24.1 km) tend to show points that sway toward the compositional field of their upstream source basin. In the lower valley, an equilibrium develops along the unit ratio line, with points moving from proximal to distal. Samples with higher element concentrations plot closer to the ‘proximal’ end of the spectrum. 126  Chapter 5. Characterization of Valley Sediments  2 Batholith Tributaries Chwk/Cultus Group Tributaries Mainstem Alluvium Pleistocene Deposits  2  1  K (%)  1.5  3  1 4  0.5 2  4  6  8  Fe (%) Figure 5-9: Bivariate plot showing the major geochemical domains among channel sediments and glacial deposits in the Chilliwack Valley. Among the raw element data, K and Fe provide the best discriminating potential. Fine channel sediments from batholith sources (upper left) and Chilliwack Group or Cultus Group sources (lower right) are reasonably distinct. Mainstem alluvium plots as an intermediate field, and glacial deposits plot to the lower left. A few seemingly anomalous points are evident: (1) a sample taken from glaciolacustrine bluffs at the Slesse/Chilliwack confluence. There is a strong affinity here for headwater lithologies (2) two samples taken from the Slesse Park landslide, which show a concentrated (proximal) mix of the two source terranes. (3) is a distinct sandy bed from Slesse Park landslide that clearly has a strong indication of Slesse source material. There is also a ‘mainstem alluvium’ point here that was sampled downstream of the Nesakwatch Creek confluence, illustrating its strong influence on mainstem sediment composition. (4) is from Borden Creek, whose source terrain appears relatively high in silica and low in indicator elements.  127  Chapter 5. Characterization of Valley Sediments  a)  b)  3  K-Factor  K-Factor  Alluvium Batholith Enclave Chwk Group Elbow YAC  1  1  V 0 Ti  Fe  -1  Zn  Ca Ba  Mn Na  P Al -3  6  4  2  0  -2  -4  -1 1  -6  K 0  Fe-Factor  -1  Fe - Factor  1.5  Distal  Chilliwack/Cultus Group (Metamorphic) Source  c)  Pleistocene Deposits  -0.5  Lower Slesse  Tamihi  0.5  Upper Slesse  31.8 34.8  37.7  Nesakwatch  Upper Chilliwack River  Centre  40.7  Foley  Chilliwack Lake  Below Chilliwack Lake Below Slesse Creek Group Centroid  43.4  Batholith (Crystalline) Source  Fe, Ti, V, Mn  Proximal -1.5 2.5  Vedder Fan  29.5  Al, K, Na, P  K-Factor (Aluminosilicates)  Liumchen & Chipmunk  1.5  0.5  -0.5  -1.5  -2.5  Fe - Factor (Metals)  Figure 5-10: (a) shows the fine sediment samples and geochemical data from source rocks cast in Fe-K factor space. There is clearly an influence from both sources on the elemental composition of the sediments. Source rock information comes from compilations in Richards [1971], Sevigny and Brown [1989] and Tepper and Kuehner [2004]. (b) is a vector diagram illustrating the magnitude and direction of each element in Fe-K factor space. (c) shows the relation between mainstem samples (open circles below Slesse Creek, closed circles joined by lines above, distance to Vedder Crossing indicated) and tributary sediments (crosses represent the mean of all samples within each tributary). There is a clear influence of granitic lithologies on the upper mainstem sediments that diminishes for samples lower in the valley. See text for further discussion.  128  Chapter 5. Characterization of Valley Sediments Table 5.2: Factor Scores from elements in the factor analysis using Varimax rotation. Principal loadings for each element are highlighted.  Element Al Ba Ca Fe K Mn Na P Ti V Zn  Fe-Factor 0.09790 0.26910 0.20380 0.94400 -0.35790 0.65780 -0.29480 0.40310 0.87710 0.93930 0.43320  K-Factor -0.79420 -0.24070 -0.15020 -0.08430 -0.84720 -0.50110 -0.69050 -0.72970 0.04710 0.15710 -0.31050  Third Factor 0.26460 0.76230 -0.88750 0.12340 0.27730 0.19510 -0.43930 -0.25420 0.11040 0.01660 0.68240  The actual balance, based on the raw element data, is approximately 1:2.2, with the K-factor wielding the dominant influence in the mainstem fine sediment chemistry. The influence of the tributaries injecting fresh K- and Fe-rich material into the channel diminishes downstream, presumably buffered by the sediment supplied from within the mainstem. The sample from the canyon near Larson’s Bench (29.5 km) plots close to the lower Slesse Creek field possibly because they are both reworking sediment derived from the same outwash deposit. Samples taken in higher reaches of Slesse Creek show much more affinity for the Chilliwack Group lithology. Liumchen and Chipmunk Creek plot within the same field since the Cultus Formation crops out in large portions of each catchment. The longitudinal picture (Figure 5-11) shows the K-factor tracking the upstream granitic terrain area trend quite closely. Downstream from Slesse Creek, the element concentration of the fine sediment appears to be more readily influenced near the junctions with the larger tributaries than the coarse sediment (Figure 5-8a). The influx of fine tributary silts results in a high spike at each confluence, which diffuses rapidly downstream. The tributaries clearly exert some influence on the K-factor along the length of the mainstem. There is insufficient contrast, however, to derive good estimates of mixing. The Fe-factor and associated elements, on the other hand, show good contrast and thus are a better choice for estimating the proportion of material contributed from each catchment. Predictably, Slesse and Tamihi emerge as very strong contributors to the mainstem fine sediment budget.  129  Chapter 5. Characterization of Valley Sediments  -0.5  0.5  Foley  25%  0 Nesakwatch  50%  0  Aluminosilicate (’K’-) Factor  Liumchen  Tamihi  Slesse  75%  Borden  -1  Centre  Percent Granite Terrain Upstream  100%  1  0  8  16  24  32  40  48  56  Distance Downstream from Chilliwack Lake (km)  Figure 5-11: Longitudinal pattern of ‘K’-Factor variation, in comparison to the relative proportion of granitic source terrain upstream. Peaks in the gradient are evident at Slesse and Tamihi Creeks, and one near 30 km, a short distance above Alison Pool.  To estimate the relative contributions of silt from each tributary to the mainstem, a simple mixing model [Peart and Walling, 1986; James, 1991] was used: Ptrib = 100 ×  Cds − Ctrib Cus − Ctrib  (5.1)  where Ptrib is the proportion of material contributed by the tributary, Cus , Cds , and Ctrib are the element concentrations in the upstream mainstem, downstream mainstem, and tributary reaches, respectively. Table 5.3 shows the estimates of mixing at four major tributary junctions. The estimates at each junction vary; the Fe-Factor likely provides the most balanced estimate, as it is based on 11 elements. Table 5.3: Mixing estimates based on Fe-Factor and associated raw element variables. Values indicate the approximate proportional addition to the mainstem sediments at each tributary junction. +∞ scores indicate scores greater than one; −∞, less than zero.  Tributary Center Nesakwatch Slesse Tamihi  Fe 0.55 0.23 0.93 0.76  Mn 0.34 0.18 +∞ 0.52  Ti 0.25 0.31 0.43 0.59  V 0.56 −∞ 0.80 0.80  Zn 0.45 0.38 0.79 0.54  Fe-Factor 0.39 0.18 0.82 0.69  Distal sediments show increasing silica content, and decreasing concentrations of other major elements and minor elements. Geochemical samples taken from Pleistocene glaciolacustrine exposures show a clearly dominant batholith influence above Slesse Creek, with 130  Chapter 5. Characterization of Valley Sediments increasing influence of Chilliwack and Cultus Group source terrain downstream. The provenance of the upland glacial tills remains less certain. It is interesting to note that the few (n=7) samples taken from the Ryder and Tolmie Uplands are scattered but geochemically indistinguishable. It was expected that sediments from the upper sections of Tamihi Slide (Ryder Lake Upland) would show more of an exotic fingerprint, since it was likely built of sediment from Fraser Valley sources. Quartzite clasts of Fraser provenance are found in the Ryder Lake Upland [see Saunders et al., 1987], though they were not noted in the Tolmie Uplands. Quartzite is mostly silica and essentially geochemically inert with respect to the elements under consideration, and therefore perhaps the foreign lithics do not have a detectable signal. As a whole the ‘Glacial Tills’ group (Figure 5-10c) shows the characteristic mainstem factor ratio, and the element concentrations at some sites are not significantly different from the lower valley channel sediments, suggesting a strong Chilliwack influence. A very few (n=3: 7.3, 31.7, 80.5 m depth) samples taken from the Vedder Fan cores suggest a fairly stable fan source composition over time.  5.5  Summary and Conclusions  Previous chapters have examined the total volume of material eroded throughout Chilliwack Valley, and this chapter has sought to refine the analysis by considering the partitioning of the sediment budget among sedimentological and geochemical lines. This provides a number of important quantities for the sediment routing models in the following chapter. The fining pattern between Chilliwack Lake and Slesse Creek is markedly different from the one between Slesse Creek and Vedder Crossing. The upper portion of the valley has finer material, and a more subdued fining gradient. As noted previously by Ham and Church [2000], Slesse Creek is the sedimentary headwaters of Chilliwack Valley, providing the majority of coarse clastic material and setting the fining gradient for the lower valley. With the emplacement of the Chilliwack Lake moraine and the evacuation of glacial sedimentary stores from Larson’s Bench and lower Foley Creek, the middle mainstem has been adjusted to behave essentially as a tributary. The catchment area ratio at the confluence of Slesse Creek and the Chilliwack mainstem is 1:4. Average flows gauged at the mainstem canyon (HYDAT, 2005; 08MH103, 29 km above Vedder Crossing) are 3.4 times the flows from lower Slesse (HYDAT, 2005; 08MH056,  131  Chapter 5. Characterization of Valley Sediments 25 km above Vedder Crossing). However, with the truncation of the sedimentary network at Chilliwack Lake, the ratio of sediment recruitment areas is 1:1 (Slesse = 47.25%). There has presumably been an ongoing adjustment in the balance of sediment delivery from the two basins throughout the Holocene Epoch. Downstream of Slesse Creek, the channel surface facies become quite coarse and show much variability. The coarser fractions within the active channel have clearly been winnowed from the glacial outwash deposits, and have been deposited near tributary junctions in broad bouldery bars. The implications of this process are explored in the following chapter. The bouldery lag appears to exert some influence on the development of the concavity of the river long profile. The granitic portion of the boulders is distributed throughout the drainage roughly in proportion to the granite source area upstream. Upstream of Slesse Creek, the more mobile population of granite materials appears to decline, relative to the source areas upstream. This is interpreted to arise as a result of selective transport and abrasion of material. Downstream of Slesse Creek, however, there appears to be a resurgence of granitic clasts in the 32 to 90 mm range. It is speculated that this may be remobilized glacial material that is recruited from lateral deposits along the length of the mainstem. The sand percentage in the channel subsurface increases from roughly 10% to 20%, indicating a gradual enrichment of the sand fraction over 50 km through the recruitment of weathered materials and glacigenic sediments, and through abrasion of gravels and coarser clasts. Sediment geochemistry shows that the major suspended sediment sources to the mainstem are Slesse and Tamihi Creeks. The supply of freshly eroded minerals from the tributaries below Chilliwack Lake is gradually diluted downstream, to be replenished at Slesse and Tamihi confluences. While estimates of mixing at the confluences suggest high local rates of contribution (70 to 80%), the signal decays quickly downstream due to mixing with the larger pool of mainstem sediments. These observations help to elucidate some of the complex linkages in the sedimentary cascade that has evolved during the process of river degradation in Chilliwack Valley. The long-term sorting patterns strongly reflect the glacial legacy of the catchment. Stores of glacial material are continually recruited by the fluvial network and routed through the valley, lengthening and complicating the post-glacial relaxation response.  132  Chapter 6 A Morphodynamic Model of Postglacial River Evolution  6.1  Introduction  In the previous chapters, I have estimated the volumes of material evacuated from Pleistocene and Holocene sediment stores and the approximate rates of bedload sediment flux from the tributary catchments in Chilliwack Valley. The present chapter uses these estimates to parameterize a numerical model that provides an exploratory framework for analyzing the paraglacial response and the history of subsequent Holocene yield. The model is used to estimate the fractional rates and patterns of bedload (gravel and sand) transport and profile evolution in the lower alluvial reaches of Chilliwack River. The aim of the modelling exercise is to track the evolution of the post-glacial river over time. By simulating the sediment cascade numerically, it becomes possible to explore hypotheses regarding the development of the fining pattern in river gravels, and the growth rate of valley landforms In the conceptual model of paraglacial sedimentation [Church and Ryder , 1972], maximum sediment transfer occurs soon after deglaciation as rivers begin to erode aggraded glacial landforms. Glacial valley fills and fan deposits are incised, and falling base-level in tributaries releases stores of sediment that have collected over the glacial period. Material that is initially entrained in tributaries is redeposited in the valley of higher order streams, then later remobilized [Church and Slaymaker , 1989]. In a large dendritic network, the pattern of response at the valley outlet represents the cumulative response of several tributaries and a large number of smaller contributing sediment sources. Super-imposed on the paraglacial response are disturbances that operate over smaller spatial and temporal domains [Church, 133  Chapter 6. A Morphodynamic Model of Postglacial... 2002]. There have been readvances of the glaciers over the Holocene Epoch, for instance, the imprint of which is evident in many valleys of the Coast Mountains. Some very large discrete inputs to the valley, such as failures from a valley wall [Dadson and Church, 2005], are known to have occurred in Chilliwack Valley though the timing and magnitude of these inputs remain uncertain. In the century to millennial scale, the river system operates in a somewhat buffered state - forcing agents such as large floods and debris flows will have a distinct effect, however the system is effectively insulated from rapid overall change as large depositional zones may act to damp large disturbances. A river such as the Chilliwack may rework over half of its floodplain width on a timescale of roughly 150 years [Martin and Church, 2004]. The rate of translation of a major perturbation is modulated by storage and size segregation of materials through differential transport, which give rise to significant dispersion over time and downstream distance. Routing a large-scale sediment wave at the millennial scale requires appropriate parameterization and calibration of the many transfer and storage zones along the fluvial network that affect the transit time. Storage reservoirs such as broad floodplain reaches, as well as material stored behind debris dams, at breaks in slope, or at tributary junctions all act as ‘rate-limiting’ steps in the transit of the disturbance. Vegetation plays a role in moderating the amount of material available for recruitment, and constrains the rate of channel migration. It is these filters and rest stations in the landscape that give rise to the dispersion of the sedimentary disturbance downstream. The hypothesis presented earlier in the thesis maintains that the post-glacial fluvial system reaches a distinctive, asymptotic equilibrium that bears the imprint of glacial processes. A simple (idealized) numerical model is presented in the following sections using the database of Holocene erosion volumes from Chilliwack Valley. The goal of the modelling exercise is to reproduce several aspects of the paraglacial sediment wave and the longitudinal development of the lower valley mainstem using the evidence available to constrain the flux of material and the evolution of the fining gradient. The model can be effectively used to synthesize many of the quantities estimated and topographic relationships inferred earlier. The model does not seek to reconstruct the historical trajectory of the system, but rather to examine the role of several key parameters and driving forces. The alluvial valley floodplain, bounded by steep valley walls, is a ubiquitous feature of the Canadian Cordillera. In terms of model construction, it is well suited to 1-dimensional representation, particularly if the total sediment delivery is reasonably well constrained.  134  Chapter 6. A Morphodynamic Model of Postglacial... Many of the essential parameters remain simplified estimations. Arrival at the modern basin conditions is by no means an indication that the numerous mutually interacting processes have been captured effectively in the model. The model represents a highly averaged view of Holocene sedimentation, and there is undoubtedly a problem of “equifinality” [Schumm, 1991] in that the landscape generated through the model may achieve a likeness of the modern landscape through a different combination of processes than actually occurred. The model can be used, however, to emphasize some of the major controls on the sedimentology of the channel and the development of the longitudinal profile. The model is used to evaluate semi-quantitatively the following: 1. The relative response timescales of textural (bed substrate) and topographic (longitudinal profile) adjustments, and how they vary with the tributary/mainstem ratios of discharge, sediment flux and sediment calibre. 2. The feedback between the calibre of sediment fed to the system and (i) the resultant rates of down-cutting through the aggraded glacial stratigraphy and (ii) development of longitudinal concavity in the river bed profile. 3. The role of lateral floodplain storage zones in moderating the rate of passage of the paraglacial wave. 4. The relative rates of gravel sorting and abrasion required to replicate the fining patterns and subsurface sand content observed today in the mainstem and on the fan. 5. The relative influence of sediment delivery from tributaries and the mainstem on the aggraded stratigraphy of Vedder Fan, the terminal sediment delivery point. Very little is known about the boundary and/or forcing conditions over time, and yet explicit boundary conditions must be specified. Only the beginning and end states of the system are suitably characterized. However, given our understanding of the general hydrology and sedimentology of the system, it is possible to gain some understanding of the evolution of the sedimentary system based on conventional time-scale hydraulics and sediment transport equations. A sediment routing model has been developed to achieve a synthesis and integration of the estimated storage mass and bedload fluxes reported in this thesis. To assess the agreement between field measurements and numerical output, there must be a suitable metric. Much 135  Chapter 6. A Morphodynamic Model of Postglacial... of the field data collected and inferences developed throughout this thesis are used to define the initial state of the system. In reality, the number of measurements is limited, and the variability of the alluvial system over time is great. There is a further challenge of interpreting the model output and comparing this to the modern river. Observation and measurement of both independent and dependent quantities are laden with inferences and assumptions. The degree to which the assumptions hold in complex environments cannot be established a priori. Information such as reach-averaged grain size or lithological variation is subject to some modeller interpretation. Other measurements, such as the elevation of points along the long profile, are easier to assess. After a description of the model components, a framework for quantitatively assessing agreement between the model output and field measurements is described below.  6.2  One-dimensional representation  1D computations of river bed evolution at the decadal to millennial scale have typically been carried out assuming uniform sediment, often using the diffusion equation or other slopedependent approximations of gravel transport. River models have been used in a number of different ways to simulate the long-term response of channels to different forcing agents such as tectonics [Paola et al., 1992; Rivenaes, 1997], base-level fall [Begin et al., 1981; Begin, 1988; Doyle and Harbor , 2003], and climate change [Tucker and Slingerland , 1997]. Recent work has been done by Braun et al. [1999] on modelling the glacial-fluvial transition. More sophisticated sediment transport models have arisen with better understanding of the differential mobility of sediments [Parker , 1990; Wilcock and Crowe, 2003], though routing models that use these multiple fraction transport equations are not routinely applied to largescale studies. One exception is the study by Gasparini et al. [2004], who examined the role of sediment composition in shaping the longitudinal profile of steepland rivers. At the shorter, event-scale, numerous models that track mass conservation using a transfer function among individual grain size fractions and bed elevation have been very successful in reproducing both textural changes and observed changes in bed topography [Ribberink , 1987; Rahuel et al., 1989; van Niekerk et al., 1992; Hoey and Ferguson, 1994; Cui et al., 1996]. Recent progress has been made in representing steady-state inputs of different magnitude along the length of channel [Ferguson et al., 2006]. Many of the models referred to above were designed to represent either aggrading or  136  Chapter 6. A Morphodynamic Model of Postglacial... equilibrium systems. Two important demonstrations of the degrading response are those of Talbot and Lapointe [2002] and Ferguson et al. [2006], who show the importance of armouring in a degrading system. The response is generally rapid and, among other effects, it acts to diminish transport rates as the finer material is evacuated.  6.2.1  The Degrading River Valley System  In a 1D framework, downstream divergence of sediment transfer, and thus evolution of the degrading bed profile, is governed by erosion of the bed substrate. As the bed is eroded and the mainstem lowers, new supplies of sediment are added to the active load. By expanding the formulation to include the active floodplain width, a 1D model may simulate a number of features inherent in a larger valley system over extended timescales (centuries to millennia). This alluvial system generally behaves as a series of interconnected reservoirs whose active lateral floodplain extent, and thus exchange volume, is bounded by valley walls of bedrock and glacial deposits. For an equilibrium profile, it is assumed that floodplain erosion within each reach is approximately balanced by lateral accretion and overbank deposition. A degrading system tends to remain confined by the high terraces that it develops. As the system evolves, much of the active floodplain width is reworked within a few model timesteps. Channel hydraulics are based on the reach-averaged slope, and sediments from lateral sources are mixed with the active bed in proportion to their rate of delivery, i.e. mixing is assumed to be complete and instantaneous. Throughout the process of relaxation from the glacial disturbance and accompanying degradation, rivers exhibit a process of integral readjustment among channel morphology, topography (river profile) and texture of the river bed. The river planform may change from a braided configuration to a more confined, meandering one, changing the hydraulic geometry and sediment transporting characteristics. The upper channel reaches steepen and the concavity of the river profile increases. Finer material is winnowed away, leaving a progressively coarser active layer in the main channel. In the model, changes in bed elevation are tied to a specified ‘active width’ of the floodplain. This is probably a reasonable representation of many reaches in Chilliwack Valley. There are a few terraced remnants left behind from early stages of degradation, but the river reworks the majority of the material within the mainstem as it incises its valley deposits through the Holocene Epoch. In rivers with broad floodplains that are actively incising, down-cutting would be expected to occur relatively quickly in a restricted portion of the 137  Chapter 6. A Morphodynamic Model of Postglacial... valley cross-section. This would be followed by bank failure and lateral reworking of the substrate. Recent research into the behaviour of the large alluvial sediment mass released during a dam removal has refined understanding of this process [Cantelli et al., 2004; Cui et al., 2006a]. After reworking the material, the channel widens again toward a new equilibrium state with a lower streamwise slope. The specified floodplain width varies along the valley, but is constant through time; an improvement to the model would be to incorporate a changing floodplain width as incision progresses. The sediment continuity framework for sand materials and coarser on the bed is based on a modified version of the Exner equation as follows: ∂η wc 1 ∂qs =− Ω ∂t wv 1 − λ ∂x  (6.1)  where η is the channel bed elevation, λ is the deposit porosity, Ω is the reach sinuosity, qs is unit bedload discharge, x is the distance downvalley, and  wc wv  is the ratio of channel width  to valley width (Figure 6-1). Transport is assumed to be uniform within the channel width. The equation is schematized in an explicit first-order finite-difference form with a backwarddifference scheme for the temporal elevation derivative and a forward difference scheme for the spatial load derivative [Parker , 2006]. There is the option of specifying a time-series for changing sinuosity over time, as the river becomes increasingly constrained, however this has not been pursued here. Pre-erosional stratigraphy is programmed into the model so that as the channel erodes, the stratigraphy is entrained and becomes part of the active channel sediments. In the model, erosion occurs only within the vertical extents of the active floodplain width. Many of the deposits have more expansive lateral extents that later fail and deliver material to the river. This process is not adequately represented in the model formulation, and thus additional sediment inputs are added at each node for an extended period after incision of deposits such as Tamihi Moraine and Larson’s Bench. With respect to this formulation it is important to emphasize that, over millennia, the final sedimentology of the river system is highly influenced by the programmed stratigraphy. If the substrate grain characteristics are considered to be relatively uniform, then this will influence the composition of the active load as degradation occurs, and will possibly inhibit the development of downstream sorting. Whatever the specification of the subsurface grain size distribution, its imprint is reinforced along the length of the mainstem.  138  Chapter 6. A Morphodynamic Model of Postglacial... ValleyAxis, Model Grid  Sinuosity  Eroded Stratigraphy  Δx  Active Floodplain Channel  Storage Layers  Active Floodplain Width  Wv  Wc  Channel Width  Figure 6-1: Figure illustrating 1-D representation of channel floodplain  6.2.2  Bed Shear Stress Distribution  Sediment routing models may effectively capture the co-evolution of river profile and bed texture; however, over long time periods there are changes to the channel cross section and floodplain surface elevation that introduce significant variability to transport relations [Carson and Griffiths, 1989; Wilcock and McArdell , 1993; Paola and Seal , 1995; Haschenburger and Wilcock , 2003]. Parameterizing the variability of depth, velocity and shear stress across the channel and floodplain are common problems that researchers have identified as research priorities in order build better 1-D models. There tends to be substantial lateral variation in shear stress, τ , across the width of the channel, and simple averaging tends to underestimate total transport rates. This creates a problem for both specification of “reach-averaged” substrate conditions, and interpretation of model output. Paola [1996]; Paola et al. [1999]; Nicholas [2000] and Ferguson [2003] have attempted to refine representation of shear stress in a single dimension using parameterized distributions to represent the variability of shear stress across the channel. Each grid node of the model may then be sub-divided into high- to low-energy environments, with appropriate evolution of sedimentary facies. In the present work, the effect of locally higher fluxes at lower water 139  Chapter 6. A Morphodynamic Model of Postglacial... discharge is probably important in the initial phases of bed degradation in the postglacial Chilliwack River. A further complication is of course the many contingencies encountered in a river in flood. It is difficult to account for the effects of log jams or diversion of flow by landslide deliveries. Scouring of channel banks is often associated with debris inputs that act to divert flow against the bank [Cooper , 1977]. At the millennial scale, it is assumed to be the average flood magnitudes and not the extremes that generate the dominant geomorphic signal, and that these are adequately captured in the model.  6.2.3  The Active Layer  The surface layer is distinguished from the substrate below it in order to allow for armouring and differential grain sorting [Hirano, 1971]. The active layer thickness is commonly associated with the D90 of the bed material (i.e. 2.5 · D90 ), and this convention was followed in the model that follows. This is probably true for the case of straight-walled flumes, however in natural rivers it is expected that there is considerable variation in active depth of erosion in the stream bed. It has also been specified as a function of bedform height [Deigaard , 1980; Armanini and Di Silvio, 1988] and as a function of the bed shear stress [e.g. van Niekerk et al., 1992], and a probabilistic approach has been adopted in characterizing its depth [DeVries, 2002; Wong et al., 2007]. The depth of the active layer is not a sensitive parameter in terms of transport rate or the topographic evolution of the system, however, it is relatively important in terms of the rate of textural change and the persistence and/or rate of propagation of a disturbance along the system. With respect to long-term bed evolution models, this is important because it dictates the residence time of material in the system. As will be discussed below, the rate of transit of tracer lithologies in the model may be directly related to the specified depth of the active layer.  6.3  Model Development  The model used to simulate the evolution of Chilliwack River is a variant of the ACRONYM models, developed by Gary Parker, Yantao Cui, and colleagues. The hydraulics, stratigraphy and tributary feed components were developed by Yantao Cui. Jeremy Walsh (NIWA) translated the code to Borland Delphi, making improvements to the interface and to the code 140  Chapter 6. A Morphodynamic Model of Postglacial... structure. I have modified the program code to include components to account for floodplain width, long-profile knickpoints, clast lithology, terminal fan stratigraphy, and routines for clast abrasion. I then modified the model to set up the general boundary conditions for Chilliwack River. A further alteration has been implementation of the Wilcock and Crowe [2003] equation, which takes account of the fact that the percentage of sand in the bed material plays an important role in modulating the rates of bedload transport. Thus, an enriched sand fraction tends to render gravel more mobile; this is an important consideration in the paraglacial environment where the sand fraction constitutes a significant proportion of many sediment sources.  6.3.1  Grid Resolution and Hydraulics  The optimal cross-section resolution of longer time-scale 1D river models tends to be on the order of several channel widths, incorporating the length of an alternate bar or pool-riffle sequence. Each cell thus estimates the average behaviour of the reach and thus resolution is lost of many finer scale processes. Quasi-steady flows are assumed, ignoring variations in flow depth and velocity with time ( ∂h ∂t  and  ∂u ∂t ).  Since the model is run in steady-flow conditions there are few discontinuities in  the hydraulic energy gradient; super-critical flows are occasionally seen in the initial stages of down-cutting through the glacial stratigraphy when slopes may become quite steep. The equilibrium channel has bed slopes ranging up to 0.025. Channel width is set using the standard hydraulic relationship w = aQb with the b exponent set to 0.50 [Ham, 1996], increasing downstream with step changes at tributaries. The timestep of the model was varied over a range of values from 10 seconds to 60 minutes; longer timesteps resulted in numerical instability during the early phase of the model. A 60 minute timestep exceeds the Courant-Friedrich-Levy (CFL) stability criterion [Cunge et al., 1980] by a factor of 10; however, since steady flows were used, this was considered to be an acceptable approximation, and timesteps finer than 60 s did not show significantly different results. The model is decoupled (“asynchronous”), first calculating hydraulics based on the bed configuration, then estimating sediment transport and updating the bed profile in an iterative process. This is an acceptable approximation since the celerity of any deformation of the bed and the longitudinal sorting of bed sediments is orders of magnitude slower than water flow. Transient sediment disturbances such as mass failures are ignored in this initial treatment of 141  Chapter 6. A Morphodynamic Model of Postglacial...  Channel configuration (long profile, width, slope, bed & surface GSD, stratigraphic GSD)  Update  Upstream feed: Q, feed GSD Tributaries & sed. sources Q, GSD  qw,Dx  h,ô  Hydraulic submodel step backwater, steady flow Roughness height, roughness constant  Active layer depth, critical reference shear, hiding coeff. Sediment erosion: Wilcock & Crowe (2003)  Calculate exchange fractions  Exner exchange routine for each size fraction & lithology.  Net deposition Net erosion  Figure 6-2: Flow chart for the ACRONYM-based Chilliwack sediment routing model. Initial and boundary conditions are fed by the user (upper left), then hydraulics and sediment transport are calculated, generating a new bed configuration. The model is updated, and continues in an iterative process. Some of the key variables and adjustable parameters are indicated. GSD refers to grain size distribution  the postglacial system. First-order numerical methods are generally used in this situation, and are all that can reasonably be applied given the sparse data on the system over time. The length-scale of any backwater effects is quite small compared to the grid spacing. Backwater zones do occur at tributary junctions and in side channels, however, given the relatively steep channel slopes in Chilliwack River they are never greater than one grid cell length. A flow diagram for the model is shown in Figure 6-2. The initial channel bed profile, width, slope, grain size distributions (GSD) for stratigraphy, tributaries and other lateral sediment sources are specified by the user. The hydraulic subroutine computes flow based on the discharge and channel geometry specified. The friction slope, Sf , is calculated using a Keulegan-type formulation (Cui et al. [1996]): qw h = 2.5 ln 11 ks h ghSf  (6.2)  142  Chapter 6. A Morphodynamic Model of Postglacial... where h is flow depth, qw is unit discharge, and ks is the characteristic roughness height, equivalent to: ks = 2 · Dsg 21.28σ  (6.3)  Dsg denotes the geometric mean grain size of surface gravel and σ is the arithmetic standard deviation of the grain size of the surface material on the ψ scale. Dsg 21.28σ is an approximation of grains size of the 90-th percentile (D90 ). The model uses the standard backwater formulation when the Froude number (F = √ qw / gh) at a given cell passes a threshold level, Frn = 0.75. Once this threshold has been exceeded, the quasi-normal flow assumption is applied (Eqn. 6.4). This assumption has been used in many numerical models for the transport of heterogeneous sediment, e.g., Deigaard [1980]; Parker [1991]; Cui and Parker [2005], and studies have demonstrated that the quasinormal assumption adequately represents the fully decoupled equations in the case of flows with high Froude number [Cui et al., 1997; Ferguson, 2003]. Under the quasi-normal assumption, the St. Venant shallow water equations can be simplified as the following backwater equation. S0 −Sf , 1−Fr2  dh dx  =  Sf  = S0 ,  F ≤ Frn F > Frn  (6.4)  where S0 denotes the slope of the the bed along the channel. The Wilcock and Crowe [2003] transport equation was selected for use in the model because of its treatment of multiple grain size fractions and its recognition of the role of sand content in the surface layer in controlling entrainment. The formulation also employs a hiding function that accounts for the greater relative mobility of coarser fractions due to their protrusion into the flow. The bed grain size distribution (GSD) in the model is resolved into 12 size fractions, in ψ intervals from 0.25 mm to 512 mm. 0.25 mm is chosen as the threshold of suspension, however in the flood flows simulated here, it is expected that even larger grains (up to 8 mm often travel in long saltating trajectories. In the Wilcock-Crowe equation, the more important parameter is the overall sand content, since the proportion of sand in the bed surface conditions the dimensionless reference shear stress for entrainment [see Wilcock and Crowe, 2003]. The reference shear stress is defined as the shear stress necessary to initiate a small transport rate over the bed; critical shear stress is assumed to be a constant fraction of reference shear stress. On a sand-poor bed (¡10 per cent sand), the interlocked gravel 143  Chapter 6. A Morphodynamic Model of Postglacial... framework dominates and inhibits entrainment of both sand and gravel. Knickpoints in the model were set to a fixed elevation. Aggradation could occur at these nodes, and the bed was allowed to subsequently degrade back to the specified elevation, but not below it. When the knickpoint node was at its minimum, material from upstream was bypassed to the next node, unless there was a positive divergence in the bedload flux gradient, and material deposited. A default resistance coefficient was retained at each knickpoint for the purposes of hydraulic calculations after the surface layer had been flushed out.  6.3.2  Abrasion  Observed patterns of downstream fining are most commonly regarded as a combination of sorting and abrasion, depending on the length of river under consideration, lateral sediment sources, resistance of clasts to abrasion, changes in bed slope, and the present state of active aggradation or degradation. Higher proportional rates of sorting are commonly associated with aggrading systems, such as alluvial fans and braided rivers [Krumbein, 1942]. Alluvial fans are commonly noted as having an exponential rate of decrease in the size of grains, up to several orders of magnitude higher than rivers, due mainly to differential transport. Degrading systems may suppress the development of downstream sorting to some extent since the material in active transport is mixed with material freshly recruited from the substrate. Tributaries notably complicate simple downstream fining concepts. In the numerical model, the mechanism of abrasion is assumed to be the binary collision of particles in transport. Volume is lost from the bedload grains as well as from grains composing the surface layer as they are struck by over-passing grains. Abrasion from particle collisions typically generates sand and silt. Following Parker [1991], mass is conserved as the grains are successively reduced in size and yield silt material in the process; the silt is considered to be lost to suspension. Abrasion rates for material that is freshly recruited to the channel are often higher than for worked fluvial gravels [Adams, 1979]. Much of the breakdown process may actually occur within relict deposits by chemical weathering; freshly eroded sediment that has been sitting in a deposit for thousands of years is then more susceptible to breakage [Jones and Humphrey, 1997; Lewin and Brewer , 2002]. Based on data compilations in Shaw and Kellerhals [1982]; Rice [1996] and Lewin and Brewer [2002], the observed coefficient of abrasion and diminution in rivers (the effects are combined) range from 0.001 km−1 for more resistant lithologies, such as chert, to 0.1 km−1 for weathered and/or less competent lithologies. 144  Chapter 6. A Morphodynamic Model of Postglacial... As Parker states in his presentation of the model, the physics of the process are not fully captured. It is certain that abrasion is more complex than simply a linear relation between volumetric loss of material and travel distance; however, until the process is better characterized this formulation provides the most pragmatic model term to account for the net effect of clast abrasion and diminution. The rate of particle volume change for a given ψ class is dV = −β · vb · V dt  (6.5)  where β is the abrasion coefficient and vb is the transport velocity of the sediment grains. Particle volume as expressed in ψ units is described by the following: 1 V = π23ψ 6  (6.6)  Again following Parker [1991] and using the chain rule to equate the two, 1 dψ = β(ψ)vb dt 3 · ln(2)  (6.7)  Fj is an estimate of the proportional area of the j-th size-fraction exposed to grain collisions (a function of grain diameter, Dj , Equation 6.8). Fj =  Fj / Dj (Fj / Dj )  (6.8)  The final form of the abrasion term is then combined with the equation for mass continuity (Eqn. 6.1) to compute the transfer of mass for each grain size (j) and lithology (k) with each time step.  6.3.3  Mass Balance  The equation for sediment continuity among grain size fractions is taken from Cui and Wilcox [2005] and Stillwater Sciences [2002].  145  Chapter 6. A Morphodynamic Model of Postglacial...    (A)    (C)    ∂(La Fj ) ∂(η − La )   + ∂(qG pj ) + (1 − λ)fG Bw  + f Ij  ∂t  ∂t ∂x   (B)    (F)  (G)    (D) (E)   pj + Fj pj+1 + Fj+1  βqG    = ql βqG ( pj + Fj ) + −  3 ln(2)   ψj+1 − ψj ψj+2 − ψj+1   (6.9)  where λ is the porosity of the channel bed deposit, fG is the volumetric fraction of gravel in the bed, Bw is the sediment-transporting channel width, η denotes the elevation of the bed, qG is the unit transport rate of gravel, β is the fluvial abrasion coefficient, pj is the volumetric fraction of the j-th size fraction in the bedload, Fj is the volumetric fraction of the j-th size fraction in the surface layer, Fj is the particle number fraction of the j-th size range in the surface layer, fIj is the volumetric fraction of the j-th size range in the interface between bedload and the channel deposit, and La is the surface layer (or active layer) thickness. ql is the lateral input rates (e.g. from tributaries and sediment sources). Some of the terms in the equation are (A), the gravel mass change of the j-th range through the change in grain size distribution within the active layer; (B), the gravel mass flux of the j-th range through the change in thickness of the deposit; (C), the gravel mass flux of the j-th range through transport; (D), loss of bedload of the j-th range to silt by abrasion; (E), the loss of surface gravel of the j-th size range to silt by abrasion; (F), the transfer of gravel in the j-th range to the next finer range by abrasion; (G), the transfer of gravel to the j-th range from the next coarser range by abrasion.  6.4  Boundary Conditions  The model grid begins a short distance downstream of the Nesakwatch confluence with Chilliwack River and extends some 47.2 kilometers to the Fraser River. This domain encompasses the region of major degradation along Chilliwack River, the outlets of 6 major tributaries, and includes the aggrading Vedder Fan. A uniform grid spacing of 360 m was adopted, cap146  Chapter 6. A Morphodynamic Model of Postglacial... turing four to six channel widths in each computational cell. Sinuosity (Ω) averages 1.16 along the length of the grid. The initial conditions for the model are set to a time when Fraser ice has retreated from Vedder Crossing, base-level falls, and the river begins to re-sculpt its longitudinal profile. There is some uncertainty as to the bounding topography of the lower valley deposits at this time; several different scenarios were run to assess the magnitude of this uncertainty in terms of the flux rates, response time and sedimentological conditions. The model takes proportionally longer to evacuate a larger sediment package (see below), but there are no discernible effects otherwise; the dominant effect is the release of tonnes of sandy outwash as the river regrades its profile. After the erosion of the major outwash deposits from the tributaries and mainstem, it requires 200-300 years for the river to attain an equilibrium configuration. The upstream boundary is set at a fixed elevation (410 m); there is even today relatively minor incision of the post-glacial surface upstream of Nesakwatch Creek. The upstream node delivers water and sediment starting at relatively high rates, tapering off over the course of the first 20 model-years. The downstream boundary is one of zero bedload transport, as all coarse material is assumed to be deposited at Vedder Fan. The elevation at the downstream boundary is the Fraser River floodplain, rising at a rate of 40 cm per model-year (corresponding to roughly 1.4 mm per calendar year). The model has numerous sediment inputs along the system, which are difficult to calibrate independently, but are based on the established sediment budget framework. Rates of delivery from glacial deposits and smaller tributary valleys follow a simple exponential decay function that approximates the temporal pattern of yield after base-level fall (see discussion of baselevel fall below). It is assumed that rates of sediment delivery from hillslopes are heightened for some time after the base-level falls in each tributary, then continue at a steady rate throughout the Holocene.  6.4.1  Basin Hydrology  The hydrology of the modern Chilliwack River is well characterized, and details can be found in summaries by McLean [1980]; Jordan [1990]; Hay & Co. Consultants [1992]; Rood [1995] and Ham [1996]. Spring and Fall/Winter have distinct flooding patterns, with the winter floods showing a pattern of higher, flashier peaks. The spring nival melt tends to be of a longer duration, with rain-on-snow events generating high flows. 147  Chapter 6. A Morphodynamic Model of Postglacial...  Average Daily Flows, 1977-2005  Discharge (m3/s)  1-April  125  Vedder Crossing  1-October  150  100  Canyon 75  50  Slesse Creek  25  Chilliwack Lake 0 0  34  68  102  136  170  204  238  272  306  340  374  Julian Day  Figure 6-3: Mean of average daily flows, 1977-2005, at Chilliwack Lake, Chilliwack Canyon, Slesse Creek and Vedder Crossing.  Figure 6-3 shows the long-term mean of average daily flows at four gauging stations in the basin: Chilliwack Lake, Chilliwack River Canyon, Slesse Creek and Vedder Crossing. The gauge at Vedder Crossing has the longest record, going back to 1911. The records shown are from 1977 to 2006, and represent a common gauging period for all stations, and also a period of relatively stable climatic conditions in the Pacific Northwest. Channel geometry for the modern channel is known at four stations within the basin: the outlet of Chilliwack Lake, a river canyon above Slesse Creek, Slesse Creek, and at Vedder Crossing. Flow additions to the mainstem from each tributary are measured as incremental additions to the mainstem specific discharge, scaled by catchment area. Mean water-surface width in the lower valley above Vedder Crossing at high flows (>250 m3 · s−1 ) is approximately 100 m (Ham and Church [2000]). Average daily flows at Chilliwack Lake can be used to interpolate the flows at points between the lake and Vedder Crossing, within a standard error of 5%, using the ratio of their upstream contributing areas: QV ed = QCL  AV ed ACL  0.955  (6.10)  where QV ed and QCL are the discharge at Vedder Crossing and Chilliwack Lake, respec148  Chapter 6. A Morphodynamic Model of Postglacial... 700  2.33 Year MAF (Daily Mean, Fall/Winter) 3 = 320 m /s  650 600  Discharge (m3/s)  550 500 450 400 350 300  Fall/Winter Series  250  Spring/Summer Series  200 1  2  3  4  5 6 7 8 10  20  30  40 50 6070  100  Recurrence Interval (Years)  Figure 6-4: Partial series for average daily flows at Vedder Crossing, based on the 1952-2005 gauging period. There are two distinct flood regimes for spring/summer and fall/winter. The mean annual flood for the fall/winter is estimated to be a daily average of 320 m3 /s  tively, and AV ed and ACL are the drainage basin areas upstream of each gauging site. In the lower valley, the threshold for significant sand transport appears to be roughly 100 m3 /s, with gravel transport conditions generally setting in when flows exceed 250 m3 /s (see McLean [1980] and Ham [1996]). Bankfull flows with a return interval of 2-3 years are about 350 m3 /s (Figure 6-4). Significant morphologic change often occurs in events when maximum instantaneous flows reach over 500 m3 /s [Ham, 1996]. Based on ten years of suspended sediment gauging at Vedder Crossing, 50% of the suspended sediment load is carried by daily average flows less than 220 m3 /s and over 75% is carried by daily average flows less than 350 m3 /s (Figure 6-5). A single flood, in December of 1975, delivered more sediment in two days (approx. 183 254 Mg · d−1 ) than previous yearly totals . Peak sediment concentrations were over 1000 g · L−1 , at least an order of magnitude higher than previously sampled high flows. Although this is a very brief “snapshot” of transport work over the millennial scale, it provides a starting point for determining the relative effectiveness of sediment transporting flows. It is expected that bedload transport follows a roughly similar pattern. Steady discharge in model runs varies from set flows of 180 m3 /s to 500 m3 /s. Although the regime of flood frequency has changed over the course of the Holocene, it is assumed that 149  Chapter 6. A Morphodynamic Model of Postglacial... 1E+5  100%  7.5E+4  75%  1.37E+5 1.07E+5 5E+4  50%  2.5E+4  25%  Cumulative Percent  Sediment Load (Tonnes)  1.83E+5  60 0  56 0  52 0  48 0  44 0  40 0  36 0  32 0  28 0  24 0  20 0  16 0  12 0  80  40  0  0  Daily Average Flow Category (m3/s)  Figure 6-5: Figure illustrating the relative contribution of various daily mean flows to total suspended sediment transport.  peak flows and the overall threshold for transport for bedload transport have not changed dramatically. It is further assumed that discharge from tributaries remains in concert with the mainstem, though this is an obvious simplification.  6.4.2  Stratigraphy and Bedrock  Figure 6-6a shows the assumed stratigraphy of the mainstem, based on valley architectural elements explored in Chapter 4 (Figure 4-6). These units are set as the initial bed stratigraphy with individual layers 5 m in thickness. As the top active layer is eroded by a depth dz from above, material is progressively mined from the stratigraphy, adding to the active layer to maintain a constant thickness. The stratigraphy is assumed to be uniform across the width of the valley. The volume of material in each stratigraphic class is summarized in Table 6.1 and in Figure 6-6b. For simplicity, a uniform bulk density of 1800 kg/m3 is assumed for all units. Bedrock knickpoints impose a characteristic shape to the river long-profile, and their implementation in the model is a key boundary condition. Two major bedrock knickpoints (Chipmunk confluence and Vedder Crossing) are shown in the profile (Figure 6-6a). Boulder accumulations may act as softer knickpoints that slow erosion of the river at certain points. These occur below tributary outlets, and also near the base of remnant Tamihi moraine. In the numerical model, the bed may aggrade at the bedrock points, but cannot erode below them. Once the river degrades to bedrock, sediment is routed directly through the reach with no further bed lowering. 150  0E+0  10 E+6  20 E+6  30 E+6  b)  a)  0  5,000  10,000  ‘High Moraine Scenario’  Vedder Crossing  Outwash Gravels  ‘High Moraine Scenario’  15,000  20,000  25,000  Distance Upstream from Vedder Crossing (m)  Tamihi Moraine  Glacio-lacustrine Zone  Sandur/Deltaic Sediments Overlying Glacio-lacustrine  30,000  35,000  Chipmunk Confluence  40,000  Figure 6-6: (a)Long profile with stratigraphy. The posited initial longitudinal profile is at the top of the stratigraphic reconstruction. Tamihi moraine has either filled the lower valley (‘High Moraine Scenario’), or only partially filled it. The river flows across the remains of the drained lake and upper delta. While the boundaries are more gradational in reality, they are represented as discrete sections in the model. The modern profile is shown below in gray. Bedrock and bouldery knickpoints are indicated. (b) Graph of the volume of glacial material that was stored above the modern profile within each stratigraphic grid cell, taking account of the valley floodplain width.  Volume (m3) per grid cell  Sandur Outwash with Cobble Units  Chapter 6. A Morphodynamic Model of Postglacial...  151  Chapter 6. A Morphodynamic Model of Postglacial... Table 6.1: Representative volumes of the eroded stratigraphic units in the lower mainstem. Volume is tabulated within 360 m cells, each having a specified width and vertical walls, and thus some spatial (volumetric) resolution is lost. Volumes for an alternate model configuration with larger accumulations within the lower valley are indicated in parentheses.  Stratigraphic Unit Distal Gravel Outwash Tamihi Moraine and Till Glaciolacustrine Silts Sandy Deltaic Sediments Mid-Valley Gravel Outwash Total  6.4.3  Sediment Volume (m3 × 106 ) 84.5 135.3 (323.8) 295.7 156.0 60.8 732.3 (920.8)  Upstream Feed and Tributary Inputs  Grain Size Inputs An important issue for a Holocene sediment routing model is how to specify grain size inputs for fluvial sources in the model, when most of the information collected comes from modern deposits, or relict deposits whose geomorphological context is not entirely clear, given only partial exposure. The fluvial inputs should represent equilibrium transport conditions, ideally, free from inherited modes that may be over-represented, such as a lag mode introduced from relict glacial deposits. Boulders tend to accumulate on the bed, in obvious disproportion to their presence in the active bedload. Feeding the modern load to the model typically results in an over-armoured surface. Figure 6-7 shows surface sample 121-04 from Allison Pool, 17.0 km above Vedder Crossing, representative of much of the mid-valley river bed. There are three prominent modes, which were isolated using curve-fitting algorithms (fityk, Wojdyr [2007]) that employ a number of optimization functions to yield the best possible model fit. The sand and boulder modes are approximated by the lognormal distribution; the gravel mode is a split-lognormal distribution, with a trailing fine tail and a steeper coarse tail. Isolating these modes provides a means to generate representative synthetic grain size distributions by addition or subtraction of the appropriate mode (Figure 6-8). The modes may be programmatically varied in order to explore the sensitivity of model runs to increased sand or boulder content.  152  Chapter 6. A Morphodynamic Model of Postglacial...  20%  Field Data  % finer  Boulder ‘Lag’ Mode  Sum of Modes  15%  Limit of Entrainment 10%  Gravel Mode  5%  Sand Mode  0 -6  -4  -2  0  2  4  6  8  10  Grain Size (y)  Figure 6-7: Isolation of sand, gravel and boulder modes within the grain-size distribution, using curve-fitting techniques. The boulder lag mode potentially masks the shape of the upper limit on the gravel mode.  16%  16%  12%  % finer  % finer  12%  8%  8%  4%  4%  0 -6  +0.16 +0.12 +0.08 +0.04 0 -0.04 -0.08 -0.12 -0.16  -4  -2  0  2  Grain Size (y)  4  6  8  0 -6  -4  -2  0  2  4  6  8  10  Grain Size (y)  Figure 6-8: a) the average among all fluvial sand modes isolated from surface gravel samples taken along the length of Chilliwack River. b) illustrates how varying the sand mode for a given sample alters the shape of the grain size distribution. The mixing proportion ranges from -16 to +16% of the initial distribution.  153  Chapter 6. A Morphodynamic Model of Postglacial... Tributary Mixing and Lowering of Base-Level In Chilliwack Valley, as the glacial valley fill was transferred out of the lower reaches, it is assumed that there was a wave of degradation that proceeded upstream from Liumchen Creek to Foley Creek. As the mainstem degraded, tributary base-levels fell and released large quantities of fill from the lower reaches of each tributary (Chapter 4). Figure 6-9 shows the model grid in the lower valley, each cell representing 360 m of river length and the full extent of the Holocene floodplain. Sediment input points are highlighted with arrows. Most of the major tributaries in the lower valley clearly underwent a fall in base-level once the ice retreated and the mainstem valley incised. Output from tributaries following base-level fall is assumed to be initially quite high, tapering gradually over the course of the Holocene as supply diminished. Perfect mixing is assumed at tributary junctions, which is a simplification of real river junctions, where backwater effects, complex storage patterns and fan construction complicate the patterns of transfer. The lower tributaries potentially incised through large stores of outwash in a few calendar decades or less, following the rate of downcutting in the mainstem very closely. Meigs et al. [2006] report rates of base-level fall in tributary catchments of up to 30 m/year in the vicinity of Tyndall Glacier, Alaska, followed soon after by evacuation of the valley fill. It is usually assumed that the effects of base-level fall are primarily manifest as changes in slope. There are accompanying changes in channel width and bed texture that will extend or accelerate the response time. Doyle and Harbor [2003] have identified two important factors that dictate the rate of response: the characteristic grain size and downstream aggradation. With respect to the latter point, it is assumed that there is relatively little downstream aggradation for the tributaries, as the rate of downcutting in the mainstem must have been relatively rapid. With an increasing cover of vegetation on valley landforms and stabilization of steep slopes on glacial deposits, the rates of sediment recruitment to the mainstem may have diminished significantly. The estimated volume evacuated from each sediment source and tributary catchment has been described in previous chapters, and is summarized here (Figure 6-10). Sediment from the tributaries is fed to the model at a baseline rate that is based on the long-term estimates which incorporate mass wasting and weathering processes. As may be observed in the figure, glacial fill along the mainstem is the most significant sediment source in the model. In the long term, the tributary valleys only contribute about 20-25% of the total volume to the 154  Chapter 6. A Morphodynamic Model of Postglacial...  Foley Creek  Chipmunk Creek  Larson’s Bench Pierce Creek Nesakwatch Creek 0  625 1,250  2,500  3,750  5,000  Slesse Creek  Hillslope Source Fluvial Source Vedder Fan  Moraine Complex  Sweltzer Creek  Liumchen Creek  Tolmie Uplands Valley Wall Failure  ²  Figure 6-9: Schematic of the model grid in the lower valley. Each computational cell represents 360 m of river length, and the total active width of the floodplain. Sediment source points are indicated with arrows. The large valley wall slump upstream of the Tamihi moraine (mid-lower right) is modelled as an example of a large discrete input to the channel.  155  Chapter 6. A Morphodynamic Model of Postglacial... post-glacial sediment cascade. In the initial stages of the model, the baseline rate of tributary sediment delivery is augmented with the large quantities of glacial valley fill that had previously aggraded the lower portions of those valleys. Erosion of the fill is based on an outlet elevation / basin storage function (Figure 6-11) that was generated by estimating the volume of the outwash fill under successive long profiles during downcutting. Figure 6-11 shows two example curves from Slesse and Tamihi Creeks. As the mainstem degrades, baseline rates of delivery from the tributary are supplemented with the outwash fill material until the entire volume is evacuated. Tying the rate of tributary outwash feed to the confluence bed elevation allows a feedback to arise between the mainstem and tributaries. If the incision rate in the mainstem is proceeding quickly, there is a process of mutual adjustment, as the outwash contribution from the tributary valley increases and slows the lowering of the mainstem profile. Alternately, if transport capacity in the mainstem is satisfied, no incision of outwash occurs in the tributary and no additional inputs are added to the mainstem.  6.4.4  Time and Intermittency  One of the intended goals of modelling is to examine the timescale of transition from postglacial to modern fluvial transport regimes. Time becomes a somewhat relative quantity in the model, since all computations are width-averaged, and the model is set to run in steady flood conditions. As such, it is difficult to translate from ‘model-time’ to conventional calendar years. Matching model output to the evolution of the valley mainstem depends on the intermittency and duration of geomorphically effective floods. For instance, if such floods occur over 24 hours in one year, on average, then one model-year would represent 365 calendar years. Over the course of Holocene time, the real-time value of a model-year varies with long term climate cycles. It is impossible to resolve the effect of individual floods on the the evolution of the mainstem, however it is possible to simulate decade-scale surges in flood discharge or sedimentation rates. Palaeoenvironmental reconstructions of the immediate post-glacial climate of the Lower Mainland suggest that it was relatively dry and tundra-like [Mathewes and Heusser , 1981]. Moister conditions followed soon after, and coniferous forests were established [Walker and Pellatt, 2003]. The period from 10 000 to 6 700  14 C  yr BP is referred to as the xerothermic  interval [Mathewes, 1985] in BC. Maximum summer warmth occurred at ca. 8 000  14 C  yr  BP. After 6 700 there was a gradual shift to cooler, more moist conditions in the mid to late 156  4.3  5.0  Vedder Fan  36 km  32.0  24.5  11.9  Slesse  Liumchen  Little Tamihi  Tamihi  Borden  Sweltzer  Pierce  0 km  Feed  8  1,938 - max  8 9  3  6  (309-1350)  Lower Basin (63)  Liumchen (16)  Tamihi (128)  Tolmie Uplands (109)  Borden (7)  3  6  Sweltzer (2.4)  Liumchen (10)  Little Tamihi (0.64)  208 +/- 34 x10 m x10  Ryder (6.0)  Tamihi (38)  Borden (3.8)  Slesse (54)  Pierce Ck (0.8)  (43)  Bedload Yield from Weathering/ Hillslope Sources Foley (35) Chipmunk (15)  Larson’s Bench (15) Slesse (78)  897 x10 m x10  Ranger Run (60)  1 x 10  Bulk Volume 3 6 (m x10 ) 5 x 10  Ryder (79)  1 x 10  5 x 10  7  Foley (15) Chipmunk (6)  (12)  Sand and Gravel Glacial Valley Fill  36 km  30 km  24 km  18 km  12 km  6 km  0  Figure 6-10: Gravel and sand sediment budget for the Chilliwack model domain below Nesakwatch Creek confluence. The net bulk volume of outwash material eroded from the major valleys has been derived in previous chapters. Three alternative scenarios are presented for the total volume of material eroded from the lower valley at Tamihi Moraine. Bedload quantities are based on the relations developed between basin size and volumetric bedload yield developed in Chapters 2 & 3, partitioning the total yield to discount tributary valley fills and account for only gravel and sand from hillslope sources.  Ryder  Eroded Valley Fill  Chipmunk  Foley  Chapter 6. A Morphodynamic Model of Postglacial...  157  Chapter 6. A Morphodynamic Model of Postglacial... 1E+8  8E+7  Storage Volume (m 3)  Slesse Creek 6E+7  Tamihi Creek 4E+7  2E+7  y=-1.646´105-7.392´104x+1.203´104x2 y=-1.513´106+1.89´105x+5528x2  0  0  20  40  60  80  100  120  Elevation Above Modern Outlet (m)  Figure 6-11: Polynomial curves describing the sediment volume / outlet elevation relationship for Slesse and Tamihi Creeks. The outlet elevation is shown as a relative scale.  Holocene [Mathewes and Heusser , 1981]. In the modern climate, Chilliwack River exceeds an average daily discharge of 250 m3 /s discharge (average daily flow) for 3.5 days of the year (84 hours). Assuming for a moment that this mean holds over the longer time frame, 13 000 years may be approximated as 124.5 years of model flood time. While this assumption is problematic for a number of reasons, primarily owing to the non-linear relation between sediment transport and channel flow, it is used here as an approximate frame of reference for interpreting model results.  6.5  Model Performance  The sensitivity of the model to input parameters is explored below. There are three observable quantities for evaluating the model performance, namely, the equilibrium river bed long profile, the fining gradient of individual size fractions of the river gravels, and the subsurface sand content. These three river features are of course highly interrelated; a criterion for the best parameter performance is a minimum RMS difference between the longitudinal gradient measured in the field and results from the representative model grid. Results from model runs are compared; improvement in one category often results in worsening in another. The best model runs achieve a balance among the three criteria while using physically reasonable 158  Chapter 6. A Morphodynamic Model of Postglacial... parameters. Table 6.2 lists some of the major parameters in the model. Water discharge and sediment feed are primary controls on the evolution of the channel, and a range of plausible values has been examined. Addition or removal of sand and coarse-grained modes is carried out as a means of assessing how important the sedimentology of the feed and stratigraphic source material may be. Finally, abrasion rates have been varied through a range of possible values. For all runs the variables were randomly varied through their expected range, MonteCarlo style, in order to generate a collection of transport estimates. These results were then analysed to infer the relative influence of each of the parameters, their optimal range, and their overall role in the development of the river profile. Table 6.2 shows the range of values used; those identified as percentages reflect variation of parameters from the best available estimates of the boundary conditions. Table 6.2: Summary of static and varied model parameters  Fixed Parameters  Minor Variations Profile Variation Fully Varied Parameters  Model Parameter Transport Equation Resistance Formulation Grain Size Fractions Roughness Height Interfacial Exchange (χ) Storage Layer Deposit Porosity (Λ) Channel Width Reach Length (dx) Time Step (dt) Lower Valley Deposit Geometry Sediment Feed Water Discharge Coarse-Grained Mode Sand Mode Abrasion  Value Wilcock and Crowe [2003] Keulegan-type 12; 0.25 mm to 512 mm 2 · Dg90 0.7 for aggradation 5 m Fixed height 0.4 w = aQb , b = 0.50 360, 500 m × Ω 60, 3600 s High Moraine or Intermediate to Low 0.5x to 3x Budgeted Quantities 200 m3 /s to 500 m3 /s -5% to +5% -16% to +16% 0 to 0.06 km−1  Reference Run  360 m 3600 s Low 1x 225 m3 /s +3% -5% 0.02 km−1  An optimal set of parameters was found such that the RMS error terms were balanced, and a the best overall fit to topographic and sedimentary gradients was achieved. The parameters are outlined in the ’Reference Run’ section of Table 6.2. Output from several model runs using these parameters was chosen to demonstrate the general behaviour of the model. Figure 6-12a 159  Chapter 6. A Morphodynamic Model of Postglacial... shows a space-time diagram of aggradation and degradation in the model over a period of 140 model-years. Degradation sets in immediately upon the fall in base-level, and carries on for roughly 60 model-years (roughly 6 000 calendar years), at which time an equilibrium profile develops. Further adjustments of the long profile arise from the complex interplay between floodplain width, bedrock control points and recruitment of sediment to the valley. Minor degradation is evident downstream of tributary junctions, and some aggradation upstream. There is a mix of lowering and raising of the bed downstream of the knickpoint at 4.5 km. Corresponding unit bedload flux rates are shown in Figure 6-12b. Initially, rates of transport increase downstream, where material is recruited to the channel in increasing quantities. Rates are lower in the upper valley, until incision of the delta upstream of Slesse Creek takes place and rates gradually increase. Base-level fall at the tributary outlets triggers additional inputs from tributaries to the system, gradually subsiding within the first ten model-years. A patch of aggradation downstream of the Liumchen confluence is visible at the outset of the record. In the model, Vedder Fan responds quickly to the initial pulse of material from upstream, with higher rates continuing after transport rates in the lower mainstem have fallen somewhat. This is presumably due to continued supply from the broad lower reach, and evolution of the initially steep bed on the apex of the fan. The elevated transport rates extend 5-6 km downstream from the apex of the fan.  6.5.1  Long-Profile Adjustments  An equilibrium profile begins to emerge as the model bed approaches the elevation bounds of the modern river topography (Figure 6-13). The runs that produce the closest match with the profile had one or more of the following characteristics: (1) lower unit discharge, and thus more selective transport, (2) a higher abrasion coefficient, and/or (3) grain size distributions specified with a slightly more bouldery composition. The latter condition is not a strong influence unless the modelled flow depth is relatively low, and then it becomes quite important. The three conditions result in a greater profile concavity and deeper erosion of the stratigraphy. If boulders are over-represented in the model, then the lag becomes too great, and very coarse accumulations dominate the distribution over time. There is a clear threshold where the bed essentially ‘locks up’ and ceases to evolve its profile or bed surface substrate. A positive feedback develops between longitudinal fining on the bed surface and greater 160  Time - ‘3.5 Flood-Day’ Years  Present  1ka  3ka  5ka  7ka  9ka  11ka  140  120  100  80  60  40  20  0  0  5  Slesse Creek 10  15  -0.05  0  Tamihi Creek 20  25  Time - Model Years  40  Vedder Crossing 35  45  140  120  100  80  60  40  20  0  b)  0  5  4  10  Slesse Creek  5  15  20  6  25  30  7  35  8  40  Unit Bedload Transport (10-3, m 3/m/s)  Distance (km) Along Model Grid (Vedder Crossing = 36 km)  30  Tributaries & Discrete Sources  0.15  Vedder Fan  0.10  Tamihi Creek  Aggrading (m/yr) 0.05  9  45  Vedder Fan  Bedrock Knickpoint  Figure 6-12: Space-time diagrams illustrating long-term transport trends in Chilliwack Valley. (a) shows the balance of degradation and aggradation across space and time. Rates of change are in units of vertical metres per Model Year, with contours showing 1 m increments, up to a maximum range of 6 m. The elevation at the downstream boundary is the Fraser River floodplain, rising at a rate of 40 cm per model-year. (b) shows the corresponding unit bedload flux rates in the model. On the left axis, time in Model Years, Flood Years and an hypothetical scale for calendar years is shown.  Present  1ka  3ka  5ka  7ka  9ka  11ka  13ka  -0.10  Vedder Crossing  Degrading (m/yr) -0.15  Growth of Fan Front Aggrading Downstream Boundary  13ka  Time - Calendar Years  a)  Chapter 6. A Morphodynamic Model of Postglacial...  161  Chapter 6. A Morphodynamic Model of Postglacial... 450 400  Sandur Delta Front  350  Elevation (m)  300  Lacustrine Bed  300 280  250  Residual Moraine  260  200 240  150  Node 44  220  100 200  50 180  0  0  0  2 ka  10  4 ka  6 ka  20  8 ka  30  10 ka  12 ka  40  50  60  70  80  90  100  Model Grid Cells (360m)  Figure 6-13: Longitudinal profile, showing successive bed elevations over the course of a model run. Bed elevations at node 44 (downstream of Borden Creek) are shown in geometric series through time (inset).  erosion of the mid-valley deposits, and the dominant controls that emerge from model runs are abrasion rate (β) and discharge. Beyond governing the overall rate of the bed evolution, the specified water discharge affects concavity by controlling selective entrainment of available bed material. At higher values of bed shear stress the bedload size distribution approaches that of the surface layer, so that both the vertical and streamwise sorting are suppressed. At lower flows the opposite effect arises. In an aggrading system, this process is reinforced along the channel via deposition of the bedload. In the degrading case, however, a greater balance of material is eroded from the channel subsurface and lateral sources, strongly imparting the character of the channel boundaries upon the active surface composition. One way to achieve effective sorting in a degrading channel is by running flows that are below the threshold of mobilization for the full bed distribution. Another possibility, not pursued in this work, is that smaller transitory pulses of aggradation superimposed on the overall degradation trend may give rise to patterns of longitudinal fining. Figure 6-14a illustrates the different levels of concavity achieved in the various model runs. Agreement with the modern profile is measured as the RMS difference along the model grid. The bed is fixed at the upstream knickpoint near Chipmunk Creek and at the bedrock sill at Vedder Crossing. A planar profile with no concavity results in an RMS value of 4. Figure 6-14b shows a pattern of improving fit with decreasing discharge. The best matches (lowest 162  Chapter 6. A Morphodynamic Model of Postglacial...  a)  450  Knickpoint  400  RMS = 4 RMS = 1 RMS = 0  Elevation (m)  350 300 250  Initial Profile  200 150 100 50 0  0  10  20  30  40  50  60  70  80  90  100  Model Grid Cells (360m) 10  b)  ‘Planar Profile’  1  0.5  Better Fit  Maximum Profile Concavity 0.1 200  250  300  Mean Annual Flood  Long Profile RMS  Worse 5 Fit  0.140 0.005  350  400  Sand Profile RMS  450  500  Model Discharge at Vedder Crossing (m3/s)  Figure 6-14: (a) Illustration of the range of concavity that may be achieved during degradation. The worst-case (RMS = 4) is a planar profile between the knickpoint and the lower valley. (b) The RMS difference between river profile and model results, for a range of specified discharges. Parameters were freely varied within the ranges specified in Table 6.2 to generate a suite of model runs. Bubble size indicates the RMS difference for the longitudinal subsurface sand profile. Results generally indicate that the profile fit improves with lower discharge, and often improves at the expense of the sand profile RMS. Circle indicates reference runs, which had a relatively coarse composition and a low abrasion value. All of the lowest points have high abrasion rates.  RMS) in each flow category have the highest rates of abrasion. The scatter in the graph is due to the simultaneous variation of the other variables to generate a suite of model runs for comparison. Water and sediment inputs from lateral sources act to promote concavity as well, though these are secondary controls in this model formulation. The long profile, as derived from BC TRIM vector points, is a highly generalized repre163  Chapter 6. A Morphodynamic Model of Postglacial... sentation of the bed and may contain some artifacts that do not reflect true bed elevations. There are a few irregularities that are possibly related to ‘soft’ knickpoints, consisting of boulder accumulations such as found at Tamihi moraine or Alison Pool. The profile fit could potentially be improved by better grain-size specifications at these points in the model.  6.5.2  Textural Response  As the river cuts down through valley fill stratigraphy, the reworking of this material constitutes a greater sediment source than inputs from tributaries. In the case of Chilliwack River, sandy outwash, lacustrine silts and deltaic fill constitute a considerable proportion of the initial degradation of the mainstem. Material is quickly winnowed away by high postglacial flows, progressively sorting the alluvium and developing a coarsened active layer in the mainstem. A unique aspect of deglaciated terrain is the presence of large boulders in the drainage network, that become important structural elements in the active bed [Fahnestock and Haushild, 1962]. As sediment supply diminishes, the bed becomes armoured with a lag from the evacuated sediment stores. Profile adjustments reach an asymptotic rate, and the fining pattern is set by the higher-calibre deliveries from lateral sources. Based on gravel sampling work, there is some evidence that the mainstem surface fining pattern is interrupted by Slesse, Tamihi, Foley, Liumchen and Chipmunk Creeks. Ferguson et al. [2006] note there are three parameters that control the interruptions of the fining gradient along the mainstem. These are the ratio of tributary and mainstem discharges (QR =  Qtrib Qmain ),  the bedload flux ratio (FR) and the grain size ratio (DR). In the model, QR is  assumed to be proportional to the area ratio, and is a fixed quantity for all runs. The tributaries that are likely to disrupt the fining gradient may be distinguished using the discriminant function developed by Rice [1998] (Figure 6-15). Rice’s figure highlights the tributaries that have a high QR and are thus predisposed to having a strong sedimentological influence. FR and DR are varied in the experimental runs, and these determine the final balance. The ratio of bedload flux from tributaries to that of the mainstem varies systematically over the course of the Holocene Epoch, and thus there tends to be greater dominance of a particular tributary source than others in some periods of the rivers’ evolution. Given the ranges of FR and DR specified in the Chilliwack model, the flux ratio appears to exert the strongest control; aggrading and/or coarsening trends downstream of tributary confluences 164  Chapter 6. A Morphodynamic Model of Postglacial... 10  Slesse  5  (Distal Slope x Area)  Tamihi Chipmunk Foley  Liumchen 1  Tributaries that Disrupt Fining  Borden 0.5  Dis  Pierce Ryder  0.1 0.01  0.05  crim  inan  t Fu  0.1  ncti  on  0.5  1  Relative Basin Area (A T/AM)  Figure 6-15: Tributaries that are likely to disrupt the downstream fining pattern will plot above Rice’s discriminant function Rice [1998]  occur early in the post-glacial period when the flux ratios reach values above 0.5. The grain-size ratio also varies throughout the course of the Holocene as conditions move from finer bedload inputs to a more armoured state. The relative magnitude of the shift is not as great as the flux ratio, and thus the effect tends to be of second order. Slesse has the highest value for all three ratios and invariably interrupts the mainstem fining pattern in all runs.  6.5.3  Clast Fining and Abrasion  Results from the equilibrium end-point of the reference run with four different rates of abrasion are shown in Figure 6-16. Fining rates are shown for each whole ψ interval. The fining rate shows a good overall match among the fractions, however there are some important discrepancies. As stated earlier, the model provides a broadly averaged picture, and does not account for lateral variability in shear stress. It is this variability and bed patchiness that presumably account for much of the ‘noise’ in the field data. Furthermore, the model does not appear to account very well for the dispersal of residual boulder material left stranded along the mainstem. 165  64 mm  32 mm  256 mm  8 mm  28  12  20  β = 0.030  28  36  36  44  44  52  52  256 mm  50%  75%  100%  0 -12  8 mm  64 mm  32 mm  16 mm  4  4  12  20  28  β = 0.015  20  28  Model Grid (km, Nesakwatch to Vedder Fan)  12  β = 0.050  Model Grid (km, Nesakwatch to Vedder Fan)  Field Subsurface Modelled Bedload  -4  -4  Nesakwatch Nesakwatch  36  36  44  44  52  52  Figure 6-16: Modelled fractional reduction in grain size for the final equilibrium channel, compared with field results. The subsurface samples from the field are compared with the modelled bedload for a range of abrasion values.  Model Grid (km, Nesakwatch to Vedder Fan)  0 -12  256 mm  4  20  Model Grid (km, Nesakwatch to Vedder Fan)  12  256 mm  -4  4  64 mm  32 mm  25% 128 mm  64 mm  32 mm  16 mm  -4  8 mm 16 mm  25% 128 mm  50%  75%  100% Foley Foley  β = 0.00  128 mm  0 -12  25%  50%  75%  100%  0 -12  Nesakwatch Nesakwatch  8 mm  16 mm  25% 128 mm  50%  75%  Chipmunk Chipmunk  Percent Coarser  Percent Coarser  Percent Coarser Percent Coarser  Centre  Centre  Slesse Slesse  Foley Foley  Borden Borden  Chipmunk Chipmunk  Centre Centre  Slesse Slesse  Tamihi Tamihi  Borden Borden  Liumchen Liumchen  Tamihi Tamihi  Vedder Vedder  Liumchen Liumchen  Vedder Vedder  100%  Chapter 6. A Morphodynamic Model of Postglacial...  166  Chapter 6. A Morphodynamic Model of Postglacial... The abrasion parameter (β) that appears to best reproduce the longitudinal fining in the system is approximately 0.030 km−1 , a value that is in the higher ranges for river gravels, and roughly an order of magnitude higher than granite abrasion rates in laboratory mills [see Shaw and Kellerhals, 1982; Lewin and Brewer , 2002]. Abrasion rates in the river are likely to be relatively high; phyllite becomes a dominant component of the river load, and is less competent than granite. Many of the clasts have spent extended periods in storage, and are thus more susceptible to breakage. Another possibility is that this result is related to the specification of sorting in the model. According to the formulation, sorting will occur by selective deposition in an aggrading environment. In the present model configuration, the river is degrading and does not achieve the necessary conditions for sorting to develop. In the model, the abrasion term accounts for almost all of the grain size diminution upstream of Vedder Fan. When β is set to zero (Figure 6-16), there is very little diminution of grain size along the channel until the fan, which is aggrading throughout the model run. Further indication of the importance of the abrasion term in the model is illustrated in Figure 6-17, where the minimum RMS fit of the long profile is shown to be highly responsive to the abrasion coefficient.  6.5.4  Subsurface Sand Content  The proportion of sand in the bed substrate changes downstream (though there is significant scatter) growing from roughly 10% downstream of Chilliwack Lake to near 20% at Vedder Crossing. On Vedder Fan, sand content reaches as high as 30%. Fitting the longitudinal sand profile to the model data is difficult due to the scatter, and thus detection of a superior model fit is less robust than for the long-profile case. Figure 6-18 shows the subsurface sand content for four runs with different abrasion constants. The β variable effectively governs the rate at which clasts are diminishing in volume and generating fines as a function of downstream distance. The optimum match is closest to β = 0.030 (RMS = 0.0073). Given the scatter among the field data points, it is difficult to discern the best match among results using the various β coefficients. The overall gradient shown in the β = 0.030 run, above Vedder Crossing, seems to have the best representation.  167  Chapter 6. A Morphodynamic Model of Postglacial...  4  Long Profile RMS  3  2  No Abrasion  1  0.140 Sand Profile RMS 0.005 0 -0.02  Optimum Fit of Gravel Fining Pattern  0.00  0.02  0.04  0.06  0.08  Abrasion Coefficient (km-1)  Figure 6-17: The effect of abrasion on the long-profile fit: with a higher abrasion parameter specified, the long-profile fit improves accordingly. The RMS difference between modelled and field measurements of the subsurface sand content (bubble size) profile worsens, as a surplus of abraded material is added to the mix. Reference runs are circled; a coarser gravel mix achieves a better fit to the long profile, without having to specify a higher abrasion coefficient.  b=0.140, RMS = 0.0216  40%  b=0.070 RMS = 0.0094 b=0.030 RMS = 0.0073  20%  b=0.0 RMS = 0.0091  Vedder Crossing  Sand % by Weight  30%  10%  0 0  10  20  30  40  50  60  Distance Downstream from Chilliwack Lake (km)  Figure 6-18: Subsurface sand percentage (<2 mm) from field samples compared with the modelled bedload.  168  Chapter 6. A Morphodynamic Model of Postglacial...  6.5.5  Growth of Vedder Fan  Vedder Fan is approximated as an initially steep bed that gradually regrades itself as sediment arrives from upstream. The lateral bounds of the floodplain model grow with distance from Vedder Crossing, approximating the radial geometry of the fan. Given the symmetrical geometry of the fan, the mean bed elevation for a given radius can be specified for the breadth of the fan [Parker et al., 1998]. The model grid follows the line of direct descent on the fan. The channel is represented as a single-thread river throughout the simulation, working its way back and forth across the fan surface. This is a simplification of the real case, which would have been a braided system for much of the model time. The effects of multiple channels, varying morphology, changing shear stress distributions, and avulsion have been given consideration in other fan models [Paola et al., 1999], but are not pursued here. Another process that has evidently played a role in shaping the fan has been trenching of the fan head followed by lateral erosion and reworking [Schumm et al., 1987] which, again, is not simulated in the present model. Subsidence over the course of the Holocene is assumed to be negligible, though there is likely some consolidation of material over time. Valley yield is tabulated from the valley outlet at Vedder Crossing. The total coarse (bedload) sediment yield calculated in Chapter 4 was estimated to be roughly 1.6 km3 of sediment. With the inclusion of fine sediments, the total volume is 2.4 km3 . This provides some constraints on the model; as the cumulative output of sand and gravel approaches 1.6 km3 , the mainstem has degraded to its present configuration in most runs. The cumulative curve showing the model delivery of sediment is shown in Figure 6-19a. The pattern of growth is most strongly affected by model discharge and the specification of parameters such as abrasion rate. The effect of varying grain size, abrasion rates and tributary flux are indicated by brackets that span the potential model time taken to deliver 1.6 km3 . If discharge is constrained to 250 m3 /s, and abrasion kept to a plausible value of 0.03, the remaining sedimentological variables (sand and coarse-grained modes) introduce variability of roughly 10% to the growth rate over time (Figure 6-19b). As described above, all coarse material from the valley is deposited at Vedder Fan, and the downstream boundary (Fraser River floodplain) rises at a rate corresponding to 1.7 mm per year [Wooldridge, 1999]. A uniform rate of rise is used in the present formulation, although it is expected that early postglacial rates of Fraser River aggradation would have been higher, and tapered off over time. The shape of the fan stratigraphy is naturally strongly influenced 169  Chapter 6. A Morphodynamic Model of Postglacial...  2E+9  Cumulative Volume Deposited (m 3)  a)  Tributary Flux  Reference Run  1.5E+9  1E+9  Grain Size 5E+8  Abrasion  0 0  20  40  60  80  120  140  Model Time Step 120  b) Relative Range (%)  Abrasion 110  Sand/ Boulder Mixture  D50  Profile Form  Tributary Flux  100  90  80  Figure 6-19: (a) Cumulative deposition on Vedder Fan over the course of model runs with varying parameter values. (b) The effect of variations in abrasion (0 to 0.040), sand/boulder content (±10% and 5% variation, respectively), D50 (range through 1 phi mode), profile form (3 scenarios) and tributary flux (1 to 3x) are shown relative to the reference run (‘100%’) to illustrate the range of variability in the time taken to deliver 1.5 × 109 m3 , based on the parameter specification.  170  Chapter 6. A Morphodynamic Model of Postglacial... D50, Ø 2  3  Proportion of Mainstem Lithology  4  5  6  7  0.1  0.3  0.5  0.7  0.9  40 40  D50, Surface Layer Facies on Vedder Fan  30  30 20  20 10 10  ~70% Mainstem Source  0 0 -10 -10 -20 -20 0  1  2  3  4  5  6  7  8  9  Tributary Material  Initial Outwash 0  1  2  3  4  5  6  7  8  9  Distance Downstream from Vedder Crossing (km)  Figure 6-20: Cumulative deposition on an idealized Vedder Fan over the course of the reference model run. Model stratigraphy of Vedder Fan, showing D50 at left and the proportion of the total sediment yield contributed by mainstem gravels represented in the fan at right. Blue contours indicate fan topography at even 2.5 ka (approx.) intervals. The modern topography is overlaid in red.  by the rate of base-level rise. The depositional stratigraphy of the fan is assumed to be a combination of active layer and subsurface layer facies deposited on the fan. These may be reconstructed from the model run data. Two plots of the model fan cross-sectional architecture are shown in Figure 6-20. The surface facies are shown, illustrating the changing fining pattern on the modelled fan over time. The actual depositional record has a finer character overall, with an additional component of overbank and slackwater deposits, which are not dealt with in the model. The pattern of yield in the model over time shows a large spike of sandy material immediately following the release of sediment stored in the mainstem, followed later by a secondary pulse of somewhat coarser material from tributary stores (Figure 6-20, at right). The timing and size of the second peak is dictated to a great extent by the specified output magnitude and coarseness of the tributary material. Gravelly material takes longer to traverse the system; the amplitude of the disturbance wave is diminished and the wavelength is much longer. In the case of finer tributary source runs, the virtual velocity of all sediments is similar, and thus the disturbance tends to move as a single wave with greater amplitude. At the end of the run shown in the figure, transport rates have diminished, and roughly 70% of the material being delivered to the fan is derived from the mainstem stratigraphy, the balance being 171  Chapter 6. A Morphodynamic Model of Postglacial... delivered from hillslope, tributary, and glacial sources. The modern fan (after channelization) has D50 values that range from 35 mm at Vedder Crossing to 10 mm at the entrance to the Vedder Canal, roughly 8 km downstream. The model surface layer has a somewhat finer mixture, but fines at a similar gradient. The fan stratigraphy is not intended to be a detailed representation of the fan architecture but rather is compiled to provide an heuristic, integrated picture of the signature output from the paraglacial period. Fine sediments carried with bedload are deposited at the distal end of the model domain; no attempt has been made to represent the deposition of suspended sediments. The summary record emphasizes the major events in the Holocene Epoch, but has insufficient resolution to highlight the smaller events that occurred. The overall pattern is consistent with details reconstructed from the drilling work (Chapter 4), such as the location of outwash sands and coarse units near the apex. There is little evidence to confirm the strong pulse of tributary material shown in the early phase of post-glacial fan development. This is quite possibly due to the algorithm used to feed tributary sediments to the mainstem; transport capacity is satisfied by tributary sources before erosion of the mainstem deposits. Without more detailed provenance information, it is difficult to provide quantitative assessment of the agreement between the model and fan stratigraphy, or to suggest constraints that could be used to improve to the model.  6.5.6  Suspended Load  The total volume of suspended sediment stored in Vedder Fan (Chapter 4) was estimated at approximately 770 ×106 m3 (±22%). The most important single identifiable source for this material is Pleistocene glacio-lacustrine deposits in the mainstem, estimated to be roughly 122 ×106 m3 (Table 6.1). Sediment geochemistry indicates that the headwater contribution to the lower valley fine sediment composition is probably less than 25% in the modern load. It is difficult to ascertain the overall influence of material from the morainal gully complexes due to their indistinct geochemical character. The fine material generated from abrasion in the lower valley model was estimated to be 270 ×106 m3 , given an abrasion coefficient of 0.03. If we assume that 30% of the eroded mass from other deposits in the valley (776 ×106 m3 ) was silt or finer, this increases the total mass to 1.17 ×109 ±0.21. Losses to the Fraser are expected, and this initial analysis would suggest roughly 0.4 ×106 m3 of material, or 34% of fine material mobilized from the valley have passed beyond the fan over time. 172  Chapter 6. A Morphodynamic Model of Postglacial... Church et al. [1989] established an average yield of 0.29 Mg km−2 · d−1 , or roughly 76 000 m3 · a−1 bulk volume assuming that sediment density is 1.7 Mg m−3 . Integrating the shortterm mean over 13.3 ka provides a mass (1.03 ×109 m3 ) that is within the margin of error of the above estimate for the total evacuated fine sediment from the valley. This suggests either an underestimation of fine sediment yield from the catchment, or relatively stable rates over time. While the former should not be ruled out, the evidence from Chilliwack Lake indicates fairly stable patterns of yield over the course of the late Holocene. Presently, rates of fine sediment yield are elevated due to land use effects and, on the longer time scale, it may be a period of elevated channel erosion into the glacio-lacustrine stratigraphic basement of the valley.  6.6  Discussion and Conclusions  A simple sediment transport model has been developed to illustrate some general aspects of the relaxation from large scale disturbance on the distal fluvial segment of Chilliwack River. In the model, Chilliwack River is coupled with its valley deposit, entraining material from underlying stratigraphy as it incises and lowers. Transport and sorting of the deposit gravels, and gradually waning sediment supply from upstream, act to coarsen the bed and lower rates of transport. Despite the many simplifications inherent in this 1D finite-difference framework, it offers a number of important insights into the mass balance and bed evolution of the postglacial system over time. The observed character of the modern river is the product of 13 000 years of abrasion, sorting and reworking of glacial deposits. The most evident traces of this process are the systematic change in grain size, clast lithology and sand content along the length of the river. The cause-and-effect that give rise to such features may be readily observed in a small stream or laboratory flume, however the controlling conditions are far more complex in a large system with transient deliveries of sediment from a network of tributaries. The numerical model provides results that come to approximate the patterns we see in nature, but they are conditioned strongly on our imperfect estimates of the input parameters and boundary conditions [Oreskes et al., 1994]. Evacuation of the upper sandur sediment stores and tributary valley fills is sufficiently delayed that there is only slight aggradation in the broad reach downstream. Storage volume along each step of the model constrains the rate of change. With a series of broad reservoirs  173  Chapter 6. A Morphodynamic Model of Postglacial...  Transport Rate 3 (m /m/s) 0.007 Vedder Crossing 0.005  0.003 U/S of Foley Creek  12 Present  40  8 30 Time (ka)  20  4 10 0 Deglaciation  Distance Downstream (km)  Figure 6-21: Bedload transport gradient in space and time, illustrating changing specific transport rate along the length of the channel during the process of downcutting through the valley sediment stores and re-equilibration of the channel.  along the system, the translation of a disturbance through the system is potentially slowed and diffused. Figure 6-21 shows the changing bedload transport gradient along the length of the valley: rates initially increase downstream, then settle into an equilibrium. Tracer lithologies within the model allow some accounting of the changing sediment source over time. Figure 6-22 shows the relative contribution of material from the lower tributaries, the upper catchment (Chilliwack Lake to Nesakwatch Creek) and mainstem sources (bed and stratigraphy) over time, accounted at Vedder Crossing. In the initial stages of the model, material supplied from tributaries provides a great majority of the bedload, tapering quickly over time. As the sediment available from the tributaries and upstream decreases, there is a  174  Chapter 6. A Morphodynamic Model of Postglacial...  100000  80000  Total Discharge  3  Transport Rate (m , bulk annual yield)  120000  60000  Mainstem Lithology  40000  Tributary Lithology 20000  Upstream Lithology 0 -14000  -12000  -10000  -8000  -6000  -4000  -2000  0  Years Before Present  Figure 6-22: Bedload transport rates at Vedder Crossing throughout a reference model run. The lower three lines indicate flux rates for material sourced from upstream of the model grid, from tributaries (dashed), and from the mainstem channel. The sum of these fluxes is represented by the upper black line.  gradually increasing proportion of the bedload flux that is derived from mainstem sediment stores. Since storage within tributary fans is not explicitly considered, there is likely more complexity in the interactions between tributary and the mainstem than what is reproduced here. Sediment flux is cast in terms of ‘bulk annual yield’, which, in this run, settles out to roughly 65 000 m3 . The error propagated through the model places an error of roughly 15% on the final estimate, however the shape of the transport curve over time among runs is relatively consistent; a sustained pulse of high sediment transport during post-baselevel adjustment, followed by a tapering limb that falls by 25-30%. The long-term average inferred from the volume of coarse material stored in Vedder Fan (Chapter 4) was 121±19 ×103 m3 /yr, dropping to 78.8±9 ×103 m3 /yr after 6 200 cal. BP. Estimates of modern average rates are 36 600 to 58 000 ×103 m3 /yr. After the significant postglacial transfers of sediment following baselevel fall in the model, the long profile attains a new equilibrium which, in absence of further perturbations of the system, remains relatively static. In reality the profile is of course in a ‘meta-equilibrium’, in the sense that there are many transitory disturbances super-imposed on this balance, which are not represented in the model. Rates of transport drop significantly, following the initial 175  Chapter 6. A Morphodynamic Model of Postglacial... evacuation of sediment from the lower reaches. Since the stabilization of the long profile, most of the morphologic work in the river has been the shifting of material within the alluvial deposits on the valley floor, repeatedly reworking the channel sediments. Observations of the river on the modern time scale [McLean, 1980; Ham, 1996] have emphasized that the dominant signal in the lower reaches is transient cycles of aggradation and degradation as floodplain material is built up, eroded and sent out to the terminal fan. Due to the constraint of steady discharge in the model, this phenomenon has not been resolved. The quantity and calibre of material fed to the model governs the rate of post-glacial relaxation, and the equilibrium form of the channel profile. Finer, sandier material fed to the mainstem hastens the pace of equilibration, however the final form of the profile does not achieve an appropriate degree of concavity. What becomes clear from the model runs shown above is that there is an important feedback between profile concavity and the observed fining pattern [Sinha and Parker , 1996]; if some concavity is not achieved, then fining tends to be suppressed. Conversely, anything that enhances the selective transport of material achieves greater concavity. The best replication of the modern profile used either unrealistically high rates of abrasion, a low ‘threshold’ discharge, a coarse, heterogeneous mixture, or some combination of these three. The lower discharge is not far above the threshold for the initiation of gravel transport, leading to enhanced sorting of sediments along the channel. No single optimum set of model parameters emerges from the model runs. Exploring the effects of parameters such as discharge, grain size, sand content and abrasion is difficult because of their complex interactions. One parameter may only be sensitive when one or several others are close to a particular threshold. Another parameter that could be altered, but was not investigated in this study, is the hiding coefficient in the equation of Wilcock and Crowe [2003]. It is quite likely that the physics governing the transport and deposition of the coarser fractions is incompletely specified. The interaction of the generalized parameters in the model might lead us to assume cause-and-effect relationships that do not play out in reality. Nevertheless, the model is good enough to extract some important lessons regarding the post-glacial relaxation of a bouldery system such as Chilliwack River. The budget of eroded material in the model is dominated by the large fills along the mainstem. Output from tributary valleys (including glacial valley fills) represents only 25% of the global post-glacial sediment budget (Figure 6-10). It is also noteworthy that the error  176  Chapter 6. A Morphodynamic Model of Postglacial... term for the lower valley fill is larger than all other glacial material in the valley combined. In light of such uncertainties, results from the model must be interpreted accordingly. With better understanding of sediment recruitment along the system, and constraints of erosion rates by new dating methods, it may be possible to simulate shorter time-scale perturbations and fluctuations in transport regime. This will be a necessary step to bridging among scales (see discussion in Chapter 1). With better representation of reach-averaged channel morphology, the timescales and spatial patterns of response to disturbance can be assessed. Some other points for refinement are representation of river knickpoints, base-level fall in tributaries, and storage and mixing at tributary junctions. The transit time of material is delayed to some extent with use of a suitable cell width specification, however, there should be a better means of representing the sequestration of material into long-term storage. Finally, the question of abrasion vs. sorting has not been fully resolved here, and requires further theoretical work to better characterize the case of channels that are degrading over the long term.  177  Chapter 7 Conclusions  The postglacial landscape of the Canadian Cordillera is different in a number of important respects from other temperate landscapes that were not covered by Pleistocene ice sheets. After each glacial phase, the mountain landscape has emerged, sculpted and steepened, with footslopes, valleys and lowlands draped in deep morainal cover. Much of the drainage network form is inherited from the scouring action of glaciers, leaving ‘under-fit’ post-glacial streams in broad, U-shaped valleys. The regime of mass wasting activity provides an episodic supply of remobilized glacial material and weathered detritus to the fluvial system. Large stores of sediment remain along the valley footslopes, collected from glacial deposits on upper slopes and from rapidly wearing bedrock faces via snow avalanches and debris flow activity. The preceding chapters have sought to use a mass-balance framework in order to analyze the relations between sedimentary processes occurring at different temporal and spatial scales across the formerly glaciated Chilliwack watershed. The landscape functions as a series of process domains with disjointed linkages in both time and space. Routing sediment from its source, and estimating delivery volumes along the river network emphasizes that the behaviour of the system cannot be understood on a time scale of decades or centuries, but rather must be considered in terms of millennia at least, given the transit rates of material through the landscape [Schumm, 1977; Church, 2002]. The essential problem with such a large system is that, while its behaviour may be explained in terms of immanent laws of physics, the trajectory of the system through time consists of a sequence of configurational changes that cannot be reconstructed solely by our current understanding of landscape mechanics [Simpson, 1963]. Reconstruction of the geomorphic history is an integral part of constraining the mass-balance and thus process rates. This entails the collection of a broad set of field data in order to establish as much 178  Chapter 7. Conclusions information as possible about the modern and the antecedent (postglacial) configurations of the landscape. Rates of processes acting on the landscape (and their variability) are then inferred from accumulation zones in the landscape, such as alluvial fans, or voids that are evidently caused by erosion, such as the incised channel next to a valley terrace. A hypotheses advanced in the first chapter has guided the line of investigation in this thesis: the pattern of postglacial sediment yield is strongly conditioned by (1) the volume and geometry of aggraded glacial stratigraphy in the valley, (2) the rate and timing of sediment delivered from tributaries, (3) the rate of base-level fall in the mainstem, and (4) the size grading of material delivered to the mainstem channel. The following sections examine how some of the questions raised have been answered, or perhaps not, in the course of this study, and what further questions arise.  7.1  Process Domains  Network structure imposes characteristic behaviour at different scales in response to changing upstream area and local slope. Sediment transfer in steep hillslope catchments (out to 3th and 4th order, approximately 2-3 km2 ) follows a linear model of sediment delivery, whereas in larger catchments, the introduction of significant intermediate sediment stores along the network complicates the model. Process domains may be broadly divided into zones of diffusional processes (mass-wasting, creep) and concentrative/advective processes (the river network) [Howard , 1994]. There is an evident break in sediment yield patterns as the fluvial system gathers sufficient stream power to mobilize sediments from the hillslope domain to the fluvial system. Yield from fluvial basins generally display a different scaling behaviour, with rates of yield an order of magnitude lower than hillslopes, and better sorting of sediments. Brardinoni and Hassan [2006] have identified an important coupling between hillslope and fluvial domains, namely the “sink-colluvial channel”, which receives direct inputs of material discharged periodically from gullies. The pattern encountered in this thesis is one of significant deposition at breaks in slope and channel junctions, where catchments smaller than 2-3 km2 (up to fourth order channels) have mostly relaxed from the perturbation of Pleistocene glaciation. The relation between storage and transfer therefore changes across spatial scale. Long-term rates of hillslope erosion were estimated using a slope-based diffusion model [Roering et al., 1999]. Among the large tributary channels that empty into Chilliwack  179  Chapter 7. Conclusions Lake, roughly 10-15% of the sediment volume estimated to have eroded from hillslopes reaches the distal fan, while source-colluvial channels would appear to have rates of net fan yield closer to 60 or 70%. There is progressive sorting and sequestration of materials along the length of the system. Storage relations are different among similar domains in neighbouring basins, since configurational elements such as base level history, lithology, aspect and exposure, and precipitation gradients all introduce a measure of variability. Within different Chilliwack tributary basins, the catchment scale at which sediment transfer significantly outpaces deposition will change according to these factors. An interesting feature of a drainage network is that the great variability encountered at smaller spatial scales becomes averaged out as one traces the sediment mass downstream to the fluvial domain. The stochastic signal from the steeplands is progressively damped as the magnitude and frequency of the effective discharge changes, and sorting and storage processes within the main drainage conduit become more effective. Volumes of eroded sediments were calculated by reconstructing surfaces such as outwash plains and facets of gullied valley walls. Fans and cones that were built up during the Holocene were reconstructed using simple geometrical forms. The resultant isopach maps of difference constituted a volume surface that could then be routed along the drainage network in a GIS. The methodology is not entirely satisfactory, in that (1) it involves a spatial scale that is not very well resolved in a 25 m DEM, and (2) it entails a number of auxiliary hypotheses about the overall representation of smaller scale, episodic processes and slower ‘creep’ occurring at the Holocene time-scale. Our understanding of these processes has improved considerably [Dietrich et al., 2003]; however larger regional data sets are required to generate appropriately calibrated parameter values for large-scale application. Nevertheless, the postulated modelling framework does provide a spatially explicit picture of hillslope activity in the landscape, and permits a first-order assessment of hillslope yield for sediment routing. Figure 7-1 shows the estimated rates of long-term coarse sediment yield established in Chapter 2 overlaid upon the regional fluvial suspended sediment dataset of Church and Slaymaker [1989]. The spatial scale of the larger catchments of Chilliwack Valley overlap to some extent with the Church and Slaymaker dataset, reinforcing the pattern of increasing yield downstream between scales of 10 and 10 000 km2 . As the Chilliwack numbers are long term estimates, the data points plot somewhat higher in this zone. Contemporary yield is undoubtedly lower, and within the trend proposed by Church and Slaymaker. The trend  180  Chapter 7. Conclusions of increasing specific sediment yield relates to recruitment of material from the large glacial sediment stores along the length of the mainstem, contributing systematically greater proportions of sediment than headwater sources. The points from lower-order basins emphasize the disconnect in the landscape, resulting in large accumulations of material in the headwaters.  Specific Sediment Yield (Mg km -2 day-1)  10 glacial en velope  DISTURBED 1 lim per Up  no it:  ern rth  ta da  0.1  LACUSTRINE  0.01  0.001 0.01  0.1  1  10  100  1 000  10 000  100 000  1 000 000  Drainage area (km2)  Figure 7-1: Contemporary fluvial suspended sediment yield as a function of drainage area [Church and Slaymaker , 1989], overlaid with long-term estimates of coarse sediment delivery in Chilliwack Valley. Black points and open circles are from Church and Slaymaker [1989]; see their text for description of the different sediment yield regimes. Coloured points are from the present study (see Chap. 2).  7.1.1  Terminal Deposits of Chilliwack Valley  Chilliwack Valley provides an excellent opportunity for the study of changing yield patterns among scales, since there are two major ‘terminal’ deposits along the length of the drainage that have clearly bounded dates. Chilliwack Lake and Vedder Fan both provide a summary picture of sediment yield over the whole post-glacial period. Despite some uncertainty regarding the bounding geometry of the post-glacial sediment package, particularly within Vedder Fan, reasonable estimates can nevertheless be derived.  181  Chapter 7. Conclusions In the long-term, summary yield from Holocene subaerial bedrock weathering is relatively small in comparison to deliveries of till and outwash material in the postglacial period. Subsurface imaging of Chilliwack Lake and drillcore facies encountered in Vedder Fan emphasize the large volumetric contributions of reworked glacial material. Most of this material was redistributed in early post-glacial times and forms the core of these deposits. Rates of sediment transfer have been relatively stable throughout the Holocene, based on the paleomagnetic record in Chilliwack Lake. There is an indication of somewhat heightened rates of accumulation prior to 3 000 cal. years B.P., however more evidence is needed to constrain that trend, and to characterize rates prior to 5 000 cal. years B.P. Given the inferred thickness of the post-glacial lacustrine package, rates have not greatly exceeded those of the present. Information recovered on the pace of sedimentary disturbance indicates that past rates are not significantly different from those within the modern window of observation. Modern rates of specific fine sediment yield are 21 t/km2 /yr, a low rate for large Cordilleran Lakes, but consistent with other lakes with little or no glacial cover. Based on evidence presented throughout this thesis, it appears that much of the variability in sediment delivery following deglaciation is found within the coarser fractions. These fractions are subject to the most intermittency and sorting in their transit through the fluvial network, and are subject to the transient storage effects that characterize the paraglacial response.  7.1.2  Mainstem Deposits  The final segment of the sediment cascade consists of the distal reaches of Chilliwack River, a large meandering to braided river with broad floodplain reaches. Sampling was carried out in the major tributaries, at sediment sources and in the mainstem, recovering 62 surface samples and 35 subsurface samples in order to characterize the range of grain size distributions, clast provenance, and the longitudinal gradients of mixing and fining. Samples for geochemical characterization were taken from 86 fluvial, glacial and lacustrine environments in order to establish the provenance of materials or to track their dispersion through the drainage network. Observations of changes in clast lithology and geochemical character downstream allows some insight into the nature of coarse and fine sediment exchanges in the river mainstem. Both coarse and fine materials show a roughly similar longitudinal gradient, though the rates of mixing and the magnitude of tributary and relict glacial inputs are different. The pattern observed for the boulder fraction is relatively short transit distances over time, and 182  Chapter 7. Conclusions replenishment of supply from relict outwash and fan deposits. Finer cobbles to coarse gravels are more mobile, and tracing the fate of granitic material along the mainstem suggests that without recharge from lateral sources, this component of the bedload tends to wear down or disappear into storage over a relatively short distance, such as the 24 km interval between Chilliwack Lake and the Slesse Creek confluence. Geochemical tracing indicates a similar pattern of downstream dilution of minerals emanating from the headwaters. The composition of mainstem sediments in the reaches between Chilliwack Lake and Slesse Creek is highly influenced by lateral inputs, while the signal becomes much more stable downstream of Slesse Creek. The proposed reason for this is that the reworking of mainstem deposits such as gravel bars and relict glacial deposits constitutes a greater sediment source than lateral contributions in the downstream reach. There is an evident spike in geochemical concentration for indicator elements at the major tributary confluences, however the effect dies off less than 2 or 3 km downstream. The geochemical signature retains the character of the significantly larger mass of mainstem sediments.  7.2  Textural Evolution of the Mainstem  In Chapter 1 it was hypothesized that, in addition to the glacial valley fill, tributary additions, and mainstem base-level fall, the post-glacial evolution of Chilliwack River must be strongly influenced by the evolving texture of the active bed. A numerical model was employed as a means of linking the established hillslope, tributary and mainstem sediment budgets with downstream deposition at Vedder Fan, tracking grain size in the process. The ACRONYMbased [Parker , 1990] routing model is generally accepted as being among the best numerical representations of sediment transport principles available, particularly as they apply to the differential mobility of gravel mixtures. The transport relation of Wilcock and Crowe [2003] was used, as it provides good representation of the transport of gravel mixtures with a sand component. Conventional first-order steady flow hydraulics equations were used to simulate river flows. The model was then set up with the historical configuration of postglacial Chilliwack Valley, as best and as simply as they could reasonably be represented. Assessment of model performance was based on minimization of the RMS error between observed and modelled longitudinal gradients of channel slope, grain size and substrate sand composition. Some further qualitative assessments included the interruption of the fining gradient at tributary junctions and the total volume and stratigraphy deposited at Vedder  183  Chapter 7. Conclusions Fan. It emerged from these studies that in order for the bed profile to achieve a reasonable fit with the modern configuration, the bed texture was indeed extremely important. An essential feedback arises between the establishment of a fining gradient and the concavity of the long profile in a degrading river. Selective transport and abrasion of the substrate are key variables in achieving the optimum fit. As has previously been encountered with 1D river models [Ferguson et al., 2001], there were many tradeoffs between the criteria selected for assessment. In this case, a high abrasion coefficient produced a better model fit to the river long-profile form, but the fit for the longitudinal fining and subsurface sand content gradients became worse. Specification of a low model discharge had the outcome of enhancing selective transport, thus achieving a better fit to the profile. Coarsening the sediment feed could generate a better long profile, but the overall distribution of the bed would become highly skewed by a growing boulder population, degrading the grain-size fit (D50 or fining gradient). A model run that minimized error across all of these criteria was selected as the optimum representation of fluvial transport over the course of the Holocene.  7.3  Models of the Holocene Fluvial System  At longer time scales (centuries to millenia), resolution of individual, fine-scale processes becomes increasingly generalized, and thus appropriate rules must be devised to represent the essential physical processes and stochastic forcings that govern the evolution of river systems at the regional, drainage-network scale. Only a few of the essential processes have been adequately captured here. There are a number of qualitative aspects of the modelling work that are consistent with, but do not unambiguously constrain, some of the parameters that have been estimated to hold for the long-term evolution of the river. The thesis is a beginning point for developing a quantitative framework to explore this question further. Models allow us to refine our basic understanding of process interactions, testing whether we have a sufficient and consistent theoretical explanation of physical processes [Kirkby, 1996]. The Cordilleran sediment cascade is a complex system, with many elements interacting with varying intensity across both time and space. The task of the modeller is to determine the key processes that dominate the response at the scale of interest, and to represent them at a scale that is appropriate to the availability of data [Grayson et al., 2003]. Our ability to evaluate a model’s performance is only as good as the quality of the observations available. In the case  184  Chapter 7. Conclusions of the Chilliwack, we are mostly limited to measurement of contemporary conditions, with a number of important bounding topographic horizons that indicate the state of the landscape at deglaciation. It is anticipated that the development and refinement of dating tools such as cosmogenic nuclides and optically stimulated luminescence will assist in placing the necessary chronological bounds on Holocene fluvial landforms, helping to further constrain the model. The case of Chilliwack River offers a number of important generalizations for Cordilleran valleys, yet it is also unique in many respects. The variety of environments encountered in the upper hillslopes is broad enough that it encompasses most of the characteristic active sediment transport processes found throughout Southern British Columbia. Debris flow activity is likely somewhat more subdued than in regions with more erodible volcanic lithologies. The river mainstem exhibits many features in common with other Cordilleran Valleys, such as a wandering habit, a significant lag deposit, and abundant lateral glacial stores that are still being periodically recruited to the mainstem and reworked in the modern load. The volume of late-glacial deposition in the lower valley is the largest in the Lower Mainland [Clague et al., 1997], and the preservation of many significant glacial landforms throughout the valley make it an interesting and important site for the study of re-equilibration of the fluvial system following glacial disturbance. Understanding this essential component of river behaviour should be a central theme in geomorphology.  185  Bibliography  Adams, J., Wear of unsound pebbles in river headwaters, Science, 203, 171–172, 1979. Albani, M., and B. Klinkenberg, A spatial filter for the removal of striping artifacts in digital elevation models, Photogrammetric Engineering & Remote Sensing, 69 (7), 755–765, 2003. Armanini, A., and G. Di Silvio, One-Dimensional Model for the Transport of a Sediment Mixture in Non-Equilibrium Conditions, Journal of Hydraulic Research, 26 (3), 1988. Armstrong, J., Surficial Geology, Chilliwack (West Half ), British Columbia, Map 1487A, Geological Survey of Canada, Ottawa, Ottawa, 1980. Armstrong, J., Environmental and Engineering Applications of the Surficial Geology of the Fraser Lowland, British Columbia, Paper 83-23, 54 pp., Geological Survey of Canada, Ottawa, 1984. Armstrong, J., D. Crandell, D. Easterbrook, and J. Noble, Late Pleistocene Stratigraphy and chronology in south-western British Columbia and northwestern Washington, Geological Society of America Bulletin, 76 (3), 321–330, 1965. Baker, V., Late Pleistocene fluvial systems, in Late Quaternary Environments of the United States, edited by S. Porter, pp. 115–129, University of Minnesota Press, Minneapolis, 1983. Ballantyne, C., A general model of paraglacial landscape response, Holocene, 12 (3), 371– 376, 2002a. Ballantyne, C., Paraglacial geomorphology, Quaternary Science Reviews, 21 (18), 1935–2017, 2002b. Ballantyne, C., and D. Benn, Paraglacial slope adjustment and resedimentation following recent glacier retreat, Fabergstolsdalen, Norway, Arctic and Alpine Research, 26 (3), 255– 269, 1994. 186  Bibliography Ballantyne, C., and D. Benn, Paraglacial slope adjustment during recent deglaciation and its implications for slope evolution in formerly glaciated environments, in Advances in Hillslope Processes, Vol. 2., edited by M. Anderson and S. Brooks, pp. 1173–1195, John Wiley and Sons, Chichester, 1996. Begin, Z., Application of a diffusion-erosion model to alluvial channels which degrade due to base-level lowering, Earth Surface Processes and Landforms, 13, 487–500, 1988. Begin, Z., D. Meyer, and S. Schumm, Development of longitudinal profiles of alluvial channels in response to base-level lowering, Earth Surface Processes and Landforms, 6, 49–68, 1981. Benda, L., and T. Dunne, Stochastic forcing of sediment routing and storage in channel networks, Water Resources Research, 33, 2865–2880, 1997. Benda, L., K. Andras, D. Miller, and P. Bigelow, Confluence effects in rivers: Interactions of basin scale, network geometry, and disturbance regimes, Water Resources Research, 40, doi:10.1029/2003WR002583, 2004. Blair, T., and J. McPherson, Alluvial fans and their natural distinction from rivers based on morphology, hydraulic processes, sedimentary processes, and facies assemblages, Journal of Sedimentary Research, 64, 450–489, 1994. Bloomer, D., Sediment sorting in the gravel-sand transition along rivers: A field and modelling investigation, Unpublished PhD, University of Sheffield, 2000. Bluck, B., Texture of gravel bars in braided streams, in Gravel Bed Rivers, edited by J. Hey, J. Bathurst, and C. Thorne, pp. 339–355, John Wiley and Sons, Chichester, 1982. Booth, D., K. Troost, J. Clague, and R. Waitt, The Cordilleran Ice Sheet, in The Quaternary period in the United States, developments in Quaternary Science, edited by W. Gillespie, S. Porter, and B. Atwater, pp. 17–43, Elsevier, 2003. Bowman, J., The Chilliwack: The River that Couldn’t be Tamed, B.C. Historical News, Fall, 1992. Brardinoni, F., and M. Hassan, Glacial erosion, evolution of river long profiles, and the organization of process domains in mountain drainage basins of coastal British Columbia, Journal of Geophysical Research - Earth Surface, 111 (F1), doi:10.1029/2005JF000358, 2006. 187  Bibliography Braun, J., D. Zwartz, and J. Tomkin, A new surface process model combining glacial and fluvial erosion, Annals of Glaciology, 28, 282–290, 1999. Brooks, G. R., The fluvial reworking of late Pleistocene drift, Squamish River drainage basin, southwestern British Columbia, Geographie Physique et Quaternaire, 48, 51–68, 1994. Brown, E., Structural geology and accretionary history of the Northwest Cascade System of Washington and British Columbia, Geological Society of America Bulletin, 99, 201–214, 1987. Brune, G., Trap efficiency of reservoirs, Transactions of the American Geophysical Union, 34 (3), 407–418, 1953. Bunte, K., and S. Abt, Sampling Surface and subsurface particle-size distributions in wadable gravel- and cobble-bed streams for analyses in sediment transport, hydraulics, and streambed monitoring, General Technical Report RMRS-GTR-74, Fort Collins, CO, 2001. Cairnes, Hope, Yale, and New Westminster Districts, British Columbia, Geological Survey of Canada, Map 737A, scale 1:253,400, 1944. Cameron, V., Late Quaternary geomorphic history of the Sumas Valley, Unpublished M.Sc. thesis, Simon Fraser University, Burnaby, 1989. Campbell, D., and M. Church, Reconnaissance sediment budgets for Lynn Valley, British Columbia: Holocene and contemporary time scales, Canadian Journal of Earth Sciences, 40, 701–713, 2003. Cantelli, A., C. Paola, and G. Parker, Experiments on upstream-migrating erosional narrowing and widening of an incisional channel caused by dam removal, Water Resources Research, 40 (3), doi:10.1029/2003WR002940, 2004. Carpenter, M., The Church Mountain stuzstrom (mega-landslide) near Glacier, Washington, Unpublished M.Sc. thesis, Western Washington University, Bellingham, 1993. Carson, M., and G. Griffiths, Gravel transport in the braided Waimakariri River: mechanisms, measurements and predictions, Journal of Hydrology, 109, 201–220, 1989. Chatters, J., and K. Hoover, Response of the Columbia River fluvial system to Holocene climatic change, Quaternary Research, 37, 42–59, 1992. 188  Bibliography Chubb, P., A detailed study of two exposures of late Pleistocene sediments in the central Chilliwack Valley, British Columbia, Unpublished B.A.Sc Thesis, The University of British Columbia, Vancouver, 1966. Church, M., Space, time and the mountain: How do we order what we see?, in The Scientific Nature of Geomorphology, edited by B. Rhoads and C. Thorn, pp. 147–170, Wiley and Sons, London, 1996. Church, M., Fluvial sediment transfer in cold regions, in Landscapes of Transition, edited by K. Hewitt, pp. 93–117, Kluwer Academic Publishers, Netherlands, 2002. Church, M., Bed material transport and the morphology of alluvial river channels, Annual Review Of Earth and Planetary Sciences, 34, 325–354, 2006. Church, M., and J. Ryder, Paraglacial sedimentation: a consideration of fluvial processes conditioned by glaciation, Geological Society of America Bulletin, 83, 3059–3072, 1972. Church, M., and O. Slaymaker, Disequilibrium of Holocene sediment yield in glaciated British Columbia, Nature, 337, 452–454, 1989. Church, M., D. McLean, and J. Wolcott, River bed gravels: sampling and analysis, in Sediment Transport in Gravel-bed Rivers, edited by C. R. Thorne, J. C. Bathurst, and R. D. Hey, pp. 43–88, Wiley, Chichester, 1987. Church, M., R. Kellerhals, and T. Day, Regional clastic sediment yield in British Columbia, Canadian Journal of Earth Sciences, 26 (1), 31–45, 1989. Clague, J., The Quaternary stratigraphic record of British Columbia : Evidence for episodic sedimentation and erosion controlled by glaciation, Canadian Journal of Earth Sciences, 23, 885–894, 1986. Clague, J., and J. Luternauer, Excursion 30a: Late Quaternary sedimentary environments, southwestern British Columbia, in Field Excursion Guidebook, International Association of Sedimentologists, 11th International Congress on Sedimentology, pp. 75–83, 1982. Clague, J., R. Mathewes, J. Guilbault, I. Hutchinson, and B. Ricketts, Pre-Younger Dryas resurgence of the southwestern margin of the Cordilleran ice sheet, British Columbia, Canada, Boreas, 26 (3), 261–278, 1997. 189  Bibliography Collins, A., D. Walling, and G. Leeks, Use of the geochemical record preserved in floodplain deposits to reconstruct recent changes in river basin sediment sources, Geomorphology, 19, 151–167, 1997. Collins, B. D., D. Montgomery, and A. Sheikh, Reconstructing the historical riverine landscape of the Puget Lowland, in Restoration of Puget Sound Rivers, edited by D. Montgomery, S. Bolton, D. Booth, and L. Wall, pp. 79–128, University of Washington Press, Seattle, WA, 2003. Cooper, A., An examination of erosion in the Chilliwack-Vedder River system, International Pacific Salmon Fisheries Commission, New Westminster, B.C., 1977. Cui, Y., and G. Parker, Numerical model of sediment pulses and sediment-supply disturbances in mountain rivers, Journal of Hydraulic Engineering, ASCE, 131 (8), 646–656, 2005. Cui, Y., and A. C. Wilcox, Numerical modeling of sediment transport upon dam removal: Application to Marmot Dam in Sandy River, Oregon, in Sedimentation Engineering, ASCE Manual 54, Vol. 2, edited by M. Garcia, ASCE, Reston, VA, 2005. Cui, Y., G. Parker, and C. Paola, Numerical simulation of aggradation and downstream fining, Journal of Hydraulic Research, 32 (2), 185–204, 1996. Cui, Y., G. Parker, and C. Paola, A quasi-normal simulation of aggradation and downstream fining with shock fitting, International Journal of Sedimentary Research, 12 (2), 68–82, 1997. Cui, Y., G. Parker, C. Braudrick, W. Dietrich, and B. Cluer, Dam Removal Express Assessment Models (DREAM). Part 1: Model development and validation, Journal of Hydraulic Research, 44 (3), 291–307, 2006a. Cullers, R. L., T. Barrett, R. Carlson, and R. B., Rare earth element and mineralogic changes in Holocene soil and stream sediment: a case study in the Wet Mountains, Colorado, USA, Chemical Geology, 63 (3-4), 275–297, 1987. Cunge, J., F. Holly, and A. Verwey, Practical Aspects of Computational Hydraulics, 420 pp., Institute of Hydraulic Research, Iowa, 1980. Curry, A., Holocene reworking of drift-mantled hillslopes in the Scottish Highlands, Journal of Quaternary Science, 15, 529–541, 2000. 190  Bibliography Dadson, S., and M. Church, Postglacial topographic evolution of glaciated valleys: a stochastic landscape evolution model, Earth Surface Processes and Landforms, 30 (11), 1387–1403, 2005. Deigaard, R., Longitudinal and transverse sorting of grain sizes in alluvial rivers, Technical Report 26, Institute of Hydrodynamics and Hydraulic Engineering, Lyngby, Denmark, 1980. Desloges, J., and R. Gilbert, Sedimentary record of Harrison Lake: implications for deglaciation in southwestern British Columbia, Canadian Journal of Earth Sciences, 28, 800–815, 1991. Desloges, J., and R. Gilbert, Sediment source and hydroclimate inferences from glacial lake sediments: the post-glacial sedimentary record of Lillooet Lake, British Columbia, Journal of Hydrology, 159, 375–393, 1994. Desloges, J., and R. Gilbert, Sedimentation in Chilko Lake: a record of the geomorphic environment of the eastern Coast Mountains of British Columbia, Canada, Geomorphology, 25, 75–91, 1998. DeVries, P., Bedload layer thickness and disturbance depth in gravel bed streams, Journal of Hydraulic Engineering, 128, 983–998, 2002. Dietrich, W., D. Bellugi, L. Sklar, J. Stock, A. Heimsath, and J. Roering, Geomorphic transport laws for predicting landscape form and dynamics, in Prediction in Geomorphology: American Geophysical Union Geophysical Monograph No. 135, edited by P. Wilcock and R. Iverson, pp. 103–132, 2003. Dietrich, W. E., and T. Dunne, Sediment budget for a small catchment in mountainous terrain, Zeitschrift f¨ ur Geomorphologie, Suppl. 29, 191–206, 1978. Dirszowsky, R., Bed sediment sources and mixing in the glacierized upper Fraser River watershed, British Columbia, Earth Surface Processes and Landforms, 29, 533–552, 2004. Doyle, M., and J. Harbor, Modelling the effect of form and profile adjustments on channel equilibrium timescales, Earth Surface Processes and Landforms, 28, 1271–1287, 2003. Easterbrook, D., Geology and geomorphology of Whatcom County, 68 pp., Western Washington University Press, Bellingham, WA, 1971. 191  Bibliography Easterbrook, D., Advance and retreat of Cordilleran ice sheets in Washington, U.S.A., Geographie Physique et Quaternaire, 46 (1), 51–68, 1992. Easterbrook, D., D. Kovanen, and O. Slaymaker, New developments in Late Pleistocene and Holocene glaciation and volcanism in the Fraser Lowland and North Cascades, Washington, in Floods, Faults and Fire: Geological Field Trips in Washington State and Southwest British Columbia, GSA Field Guide 9, edited by P. Stelling and D. Tucker, p. 255, The Geological Society of America, Boulder, Colorado, 2007. EBA, Chilliwack River Watershed Landslide Inventory. Final Report, Project No. 0807 7875007, March 30, 2002, 30 pages + tables, 2002. Emerson, D., Sardis Aquifer Yield Assessment, Emerson Groundwater Consultants, submitted to City of Chilliwack, Chilliwack, BC, 2003. Evans, S., and J. Clague, Catastrophic rock avalanches in glacial environments, Landslides. Proceedings of the Fifth Symposium, Lausanne, Vol. 2., C. Bonnard, Ed., pp. 1153–1158, 1988. Eyles, N., and H. Mullins, Seismic-stratigraphy of Shuswap Lake, British Columbia, Sedimentary Geology, 109, 283–303, 1997. Eyles, N., C. Eyles, and A. McCabe, Late Pleistocene subaerial debris-flow facies of the Bow Valley, near Banff, Canadian Rocky Mountains, Sedimentology, 35, 465–480, 1988. Eyles, N., H. Mullins, and A. Hine, Thick and fast: sedimentation in a Pleistocene Fiord Lake of British Columbia, Canada, Geology, 18, 1153–1157, 1990. Eyles, N., H. Mullins, and A. Hine, The seismic stratigraphy of Okanagan Lake, British Columbia: a record of rapid deglaciation in a deep ‘fjord-lake’ basin, Sedimentary Geology, 73, 13–42, 1991. Fahnestock, R., and W. Haushild, Flume studies of the transport of pebbles and cobbles on a sand bed, Geological Society of America Bulletin, 73, 1431–1436, 1962. Ferguson, R., The missing dimension: effects of lateral variation on 1-d calculations of fluvial bedload transport, Geomorphology, 56, 1–14, 2003.  192  Bibliography Ferguson, R., and C. Paola, Bias and precision of percentiles of bulk grain size distributions, Earth Surface Processes and Landforms, 22, 1061–1077, 1997. Ferguson, R., M. Church, and H. Weatherly, Fluvial aggradation in Vedder River: testing a one-dimensional sedimentation model, Water Resources Research, 37, 3331–3348, 2001. Ferguson, R., J. Cudden, T. Hoey, and S. Rice, River system discontinuities due to lateral inputs: generic styles and controls, Earth Surface Processes and Landforms, 31, 1149–1166, 2006. Fisher, P., and N. Tate, Causes and consequences of error in digital elevation models, Progress in Physical Geography, 30 (4), 467–489, 2006. Friele, P., and J. Clague, Younger Dryas readvance in Squamish river valley, southern Coast Mountains, British Columbia, Quaternary Science Reviews, 21 (18-19), 1925–1933, 2002. Friele, P., J. J. Clague, K. Simpson, and M. Stasiuk, Impact of a Quaternary volcano on Holocene sedimentation in Lillooet River Valley, British Columbia, Sedimentary Geology, 176 (3-4), 305–322, 2005. Fripp, J., and P. Diplas, Surface sampling in gravel streams, Journal of Hydraulic Engineering, ASCE, 119, 473–490, 1993. Gasparini, N., G. Tucker, and R. Bras, Network-scale dynamics of grain-size sorting: implications for downstream fining, stream-profile concavity, and drainage basin morphology, Earth Surface Processes and Landforms, 29, 401–422, 2004. GDBC, Digital Baseline Mapping at 1:20 000, British Columbia Specifications and Guidelines for Geomatics, Content Series Volume 3, Geographic Data BC, Victoria, BC, 1992. George, G., L. Stad, and H. Gill, Sustainable Resource Management Plan: Biodiversity Chapter for the Chilliwack Landscape Unit, 2005. Gilbert, R., Sedimentation in Lillooet Lake, British Columbia, Canadian Journal of Earth Sciences, 12, 1697–1711, 1975. Gilbert, R., J. Desloges, and J. Clague, The glacilacustrine sedimentary environment of Bowser Lake in the northern Coast Mountains of British Columbia, Canada, Journal of Paleolimnology, 17 (3), 331–346, 1997. 193  Bibliography Gilbert, R., J. Desloges, S. Lamoureux, A. Serink, and K. Hodder, The geomorphic and paleoenvironmental record in the sediments of Atlin Lake, northern British Columbia, Geomorphology, 79, 130–142, 2006. Gill, M., Planning the useful life of a reservoir, International Water Power and Dam Construction IWPCDM, 40 (5), 46–47, 1988. Goff, J., and S. Hicock, Catastrophic floods during deglaciation in a small, mountain catchment, British Columbia, Geographical Review of Japan. Series B, 68 (2), 95–106, 1995. Gorler, K., and P. Giese, Aspects of the evolution of the Calabrian Arc, in Alps Apennines Hellenides, edited by T. Closs, G. Roeder, and S. Schmidt, pp. 374–388, IUCG Scientific Report 38, 1978. Grayson, R., G. Bloschl, A. Western, and T. McMahon, Advances in the Use of Observed Spatial Patterns of Catchment Hydrological Response, Elsevier Science BV Amsterdam, Netherlands, 2003. Gregory, K. J., and D. E. Walling, Drainage Basin Form and Process: A Geomorphological Approach, John Wiley & Sons, Hoboken, New Jersey, 1976. Groulx, B., Preliminary study of slope movement in the Chilliwack River Basin with considerations of sediment transport, Unpublished Honours Thesis, University of British Columbia, Vancouver, 1993. Hallett, D., D. Lepofsky, R. Mathewes, and K. Lertzman, 11 000 years of fire history and climate in the mountain hemlock rain forests of southwestern British Columbia based on sedimentary charcoal, Canadian Journal of Forest Research, 33 (2), 292–312, 2003. Halstead, E., Ground Water Supply - Fraser Lowland, British Columbia, Inland Waters Directorate, National Hydrology Research Institute, Saskatoon, iWD Scientific Series No. 145, 1986. Ham, D., Patterns of channel change on Chilliwack River, British Columbia, Unpublished Masters’. Thesis, University of British Columbia, Vancouver, 1996. Ham, D., and M. Church, Bed-material transport estimated from channel morphodynamics: Chilliwack River, British Columbia, Earth Surface Processes and Landforms, 25, 1123–1142, 2000. 194  Bibliography Harrison, S., Paraglacial or periglacial? The sedimentology of slope deposits in upland Northumberland, in Advances in Hillslope Processes, Vol. 2., edited by M. Anderson and S. Brooks, pp. 1197–1218, John Wiley and Sons, Chichester, 1996. Haschenburger, J., and P. Wilcock, Partial transport in a natural gravel bed channel, Water Resources Research, 39 (1), doi:doi:10.1029/2002WR001532,, 2003. Hay & Co. Consultants, Chilliwack River Hazard Management Study Interim Report, Prepared for Fraser Cheam Regional District, Vancouver, BC, 1992. Hicock, S., K. Hobson, and J. Armstrong, Late Pleistocene proboscideans and early Fraser glacial sedimentation in eastern Fraser Lowland, British Columbia, Canadian Journal of Earth Sciences, 19 (5), 899–906, 1982. Hirano, M., River bed degradation with armoring, Proceedings of the Japanese Society of Civil Engineers, 195, 55–65, 1971. Hodder, K., J. Desloges, and R. Gilbert, Pattern and timing of sediment infill at glacierfed Mud Lake: implications for lateglacial and Holocene environments in the Monashee Mountain region of British Columbia, Canada, The Holocene, 16 (5), 705–716, 2006. Hoey, T., and R. Ferguson, Numerical simulation of downstream fining by selective transport in gravel bed rivers: Model development and illustration, Water Resources Research, 30 (11), 2251–2260, 1994. Howard, A., A detachment-limited model of drainage basin evolution, Water Resources Research, 30 (7), 2261–2286, 1994. Hutchinson, I., T. James, and J. Clague, Reconstruction of late Quaternary sea-level change in southwestern British Columbia from sediments in isolation basins, Boreas, 33 (3), 183–194, 2004. Ibbeken, H., and R. Schleyer, Source and Sediment: A Case Study of Provenance and Mass Balance at an Active Plate Margin (Calabria, Southern Italy), Springer, Berlin, 1991. Jackson, L., G. MacDonald, and M. Wilson, Paraglacial origin for terraced river sediments in Bow Valley, Alberta, Canadian Journal of Earth Sciences, 19, 2219–2231, 1982.  195  Bibliography Jackson, L., O. Hungr, J. Gardiner, and C. Mackay, Cathedral Mountain debris flows, Bulletin of the International Association of Engineering Geology, 40, 35–54, 1989. Jakob, M., A size classification for debris flows, Engineering Geology, 79 (3-4), 151–161, 2005. Jakob, M., and M. Bovis, Morphometric and geotechnical controls of debris flow activity, southern Coast Mountains, British Columbia, Canada, Zeitschrift f¨ ur Geomorphologie, N.F. Suppl., 104, 13–26, 1996. James, L., Quartz concentration as an index of sediment mixing: hydraulic mine-tailings in the Sierra Nevada, California, Geomorphology, 4, 125–144, 1991. Johnsen, T., and T. Brennand, Late-glacial lakes in the Thompson Basin British Columbia: paleogeography and evolution, Canadian Journal of Earth Sciences, 41 (11), 1367–1383, 2004. Jones, L., and N. Humphrey, Weathering-controlled abrasion in a coarse-grained, meandering reach of the Rio Grande: implications for the rock record, Geological Society of America Bulletin, 109, 1080–1088, 1997. Jordan, P., Hydrology of the November, 1989, Chilliwack River flood, and some observations on the impacts of forest management, prepared for B.C. Ministry of Forests, Chilliwack District, Rosedale, B.C., 1990. Jordan, P., Debris flows in the Southern Coast Mountains, British Columbia: Dynamic behaviour and physical properties, Unpublished PhD Thesis, The University of British Columbia, Vancouver, 1994. Jordan, P., and O. Slaymaker, Holocene sediment production in Lillooet River Basin, British Columbia: a sediment budget approach, Geographie Physique et Quaternaire, 45, 45–57, 1991. Kellerhals, R., and M. Church, On the statistics of grain size variation along a gravel river, Canadian Journal of Earth Sciences, 15, 1151–1160, 1978. Kesel, R., E. Yodis, and D. McGraw, An approximation of the sediment budget of the lower Mississippi River prior to major human modification, Earth Surface Processes and Landforms, 17, 711–722, 1992. 196  Bibliography Kirchner, J., R. Finkel, C. Riebe, D. Grander, J. Clayton, J. King, and W. Megahan, Mountain erosion over 10 yr, 10 k.y, and 10 m.y. time scales, Geology, 29, 591–594, 2001. Kirkby, M., A role for theoretical models in geomorphology, in The Scientific Nature of Geomorphology, edited by B. Rhoads and C. Thorn, pp. 257–272, Wiley and Sons, London, 1996. Knox, J., Responses of river systems to Holocene climates, in Late Quaternary Environments of the United States, vol. 2, edited by J. Wright, H.E., pp. 26–41, University of Minnesota Press, Minneapolis, 1983. Kovanen, D., and D. Easterbrook, Late Pleistocene, post-Vashon, alpine glaciation of the Nooksack drainage, North Cascades, Washington, Bulletin of the Geological Society of America, 113 (2), 274–288, 2001. Kovanen, D., and D. Easterbrook, Morphologic and stratigraphic evidence for Allerød and Younger Dryas age glacier fluctuations of the Cordilleran Ice Sheet, British Columbia, Canada and Northwest Washington, USA, Boreas, 31, 163–184, 2002. Krumbein, W., Flood deposits of Arroyo Seco, Los Angeles County, California, Geological Society of America Bulletin, 16, 1355–1402, 1942. Lane, E., and V. Koelzer, Density of Sediments Deposited in Reservoirs, St. Paul US Engineer District Sub-Office, Hydraulic Laboratory, University of Iowa, Iowa City, 1943. Levson, V., P. Monahan, D. Meldrum, P. Matysek, R. Gerath, B. Watts, A. Sy, and L. Yan, Surficial geology and earthquake hazard mapping, Chilliwack, British Columbia, in Geological Fieldwork 1995, pp. 191–203, Geological Survey of Canada, Ottawa, 1996. Lewin, J., and P. Brewer, Laboratory simulation of clast abrasion, Earth Surface Processes and Landforms, 27 (2), 145–164, 2002. Liu, H., and K. Jezek, Investigating DEM error patterns by directional variograms and Fourier analysis, Geographical Analysis, 31, 249–266, 1999. Lowdon, J., and W. J. Blake, Geological Survey of Canada Radiocarbon Dates XVIII, Geological Survey of Canada, Ottawa, 1978.  197  Bibliography Luckman, B., M. Kearney, R. King, and A. Beaudoin, Revised 14 C age for St. Helens Y tephra at Tonquin Pass, British Columbia, Canadian Journal of Earth Sciences, 23, 734– 736, 1986. Malmon, D., T. Dunne, and S. Reneau, Stochastic theory of particle trajectories through alluvial valley floors, The Journal of Geology, 111, 525–542, 2003. Martin, Y., Sediment budget from morphology, Unpublished M.Sc. thesis, Department of Geography, University of British Columbia, Vancouver, 1992. Martin, Y., Modelling hillslope evolution: linear and nonlinear transport relations, Geomorphology, 34, 1–21, 2000. Martin, Y., and M. Church, Numerical modelling of landscape evolution: geomorphological perspectives, Progress in Physical Geography, 28 (3), 317, 2004. Mathewes, R., Paleobotanical evidence for climate change in southwestern British Columbia during Late-Glacial and Holocene time, Syllogeus, 55, 397–422, 1985. Mathewes, R., and L. Heusser, A 12,000 year palynological record of temperature and precipitation trends in Southwestern British Columbia, Canadian Journal of Botany, 59, 707–710, 1981. McLean, D., Flood control and sediment transport study of the Vedder River, Unpublished M.Sc. thesis, The University of British Columbia, Vancouver, 1980. Meidinger, D., and J. Pojar, Ecosystems of British Columbia, Research Branch, B.C. Ministry of Forests, Victoria, B.C., Spec. Rep. 6, 1991. Meigs, A., W. Krugh, K. Davis, and G. Bank, Ultra-rapid landscape response and sediment yield following glacier retreat, Icy Bay, southern Alaska, Geomorphology, 78, 207–221, 2006. Millar, R., Influence of bank vegetation on alluvial channel patterns, Water Resources Research, 36 (4), 1109–1118, 2000. Monahan, P., and V. Levson, Chilliwack seismic microzonation project - data collection and geological overview, Paper 2003-2 Earthquake Hazard Mapping for Landuse and Emergency Planning - Summary of Conference Presentations, 1996, 61–68, 2003.  198  Bibliography Monger, J., The stratigraphy and structure of the type area of the Chilliwack Group, southwestern British Columbia, Unpublished PhD, University of British Columbia, Vancouver, 1966. Montgomery, D., and M. Brandon, Topographic controls on erosion rates in tectonically active mountain ranges, Earth and Planetary Science Letters, 201 (3), 481–489, 2002. Montgomery, D., and J. Buffington, Channel-reach morphology in mountain drainage basins, Bulletin of the Geological Society of America, 109 (5), 596–611, 1997. Montgomery, D., and E. Foufoula-Georgiou, Channel network source representation using digital elevation models, Water Resources Research, 29 (12), 3925–3934, 1993. Mullineaux, D., Pre-1980 tephra-fall deposits erupted from Mount St. Helens, 99 pp., US Geological Survey Professional Paper 1563, 1996. Munshaw, E., The geomorphology of the Chilliwack Valley, Unpublished B.Sc. thesis, University of British Columbia, Vancouver, 1978. Nicholas, A., Modelling bedload yield in braided gravel bed rivers, Geomorphology, 36, 89–106, 2000. Ollier, C. D., and M. J. F. Brown, Erosion of a young volcano in New Guinea, Zeitschrift f¨ ur Geomorphologie, 15, 12–28, 1971. Orchard, I., Floodland and Forest: Memories of the Chilliwack Valley, 413 pp., Sound Heritage Series Number 37, Provincial Archives, Victoria, 1983. Oreskes, N., K. Shrader-Frechette, and K. Belitz, Verification, Validation, and Confirmation of Numerical Models in the Earth Sciences, Science, 263 (5147), 641–646, 1994. Paola, C., Chapter 33: Incoherent structure: Turbulence as a metaphor for braiding, in Coherent Flow Structures in Open Channels, edited by P. Ahsworth, S. Bennett, J. Best, and S. McLelland, pp. 705–723, John Wiley and Sons, Inc., New York, N.Y., 1996. Paola, C., and R. Seal, Grain size patchiness as a cause of selective deposition and downstream fining, Water Resources Research, 31 (5), 1395–1408, 1995. Paola, C., P. Heller, and C. Angevine, The large-scale dynamics of grainsize in alluvial basins. I: Theory, Basin Research, 4, 73–90, 1992. 199  Bibliography Paola, C., G. Parker, D. Mohrig, and K. Whipple, The influence of transport fluctuations on spatially averaged topography on a sandy, braided fluvial fan, in Numerical Experiments in Stratigraphy: Recent Advances in Stratigraphic and Sedimentologic Computer Simulations, SEPM Special Publications No. 62, pp. 211–218, Society for Sedimentary Geology, 1999. Parker, G., Surface-based bedload transport relation for gravel rivers, Journal of Hydraulic Research, 28 (4), 417–436, 1990. Parker, G., Selective sorting and abrasion of river gravel: I: Theory, Journal of Hydraulic Engineering, ASCE, 117 (2), 131–149, 1991. Parker, tions  G., to  1-D  Rivers  and  Sediment Turbidity  Transport Currents,  Morphodynamics World  Wide  with  applica-  Web  E-Book,  http://cee.uiuc.edu/people/parkerg/morphodynamics e-book.htm, last Accessed November 17 2006, 2006. Parker, G., C. Paola, K. Whipple, and D. Mohrig, Alluvial fans formed by channelized fluvial and sheet flow. I: Theory, Journal of Hydraulic Engineering, ASCE, 124, 985–995, 1998. Peart, M., and D. Walling, Fingerprinting sediment source: the example of a drainage basin in Devon, in Drainage Basin Sediment Delivery, IAHS Publication No. 159, edited by R. Hadly, pp. 41–55, International Association of Hydrological Sciences, 1986. Porter, S., and T. Swanson, Radiocarbon age constraints on rates of advance and retreat of the Puget Lobe of the Cordilleran Ice Sheet during the last glaciation, Quaternary Research, 50 (3), 205–213, 1998. Putkonen, J., and M. O’Neal, Degradation of unconsolidated Quaternary landforms in the western North America, Geomorphology, doi:10.1016/j.geomorph.2005.07.024, 2005. Rahuel, J., F. Holly, J. Chollet, J. Belleuyd, and G. Yang, Modeling of riverbed evolution for bedload sediment mixtures, Journal of Hydraulic Engineering, ASCE, 115 (11), 1521–1542, 1989. Reid, D., and T. Dunne, Rapid Evaluation of Sediment Budgets, 162 pp., Catena Verlag, Reiskirchen, 1991.  200  Bibliography Reneau, S., and W. Dietrich, Erosion rates in the Southern Oregon Coast Range: evidence for an equilibrium between hillslope erosion and sediment yield, Earth Surface Processes and Landforms, 16 (4), 307–322, 1991. Ribberink, J., Mathematical modeling of one-dimensional morphological changes in rivers with non-uniform sediment, technical report 87-2, Ph.D. thesis, Delft University of Technology, Delft, The Netherlands, 1987. Rice, S., Bed Material Texture Along Gravel Bed Rivers with Confluences, Unpublished PhD, University of British Columbia, Vancouver, 1996. Rice, S., Which tributaries disrupt downstream fining along gravel-bed rivers?, Geomorphology, 22, 39–56, 1998. Richards, T., Plutonic Rocks between Hope, BC, and the 49th Parallel, Unpublished PhD, University of British Columbia, Vancouver, 1971. Rivenaes, J., Impact of sediment transport efficiency on large-scale sequence architecture: results from stratigraphic computer simulation, Basin Research, 9, 91–105, 1997. Roering, J., J. Kirchner, and W. Dietrich, Evidence for nonlinear, diffusive sediment transport on hillslopes and implications for landscape morphology, Water Resources Research, 35 (9), 853–870, 1999. Rood, K., An aerial photograph inventory of the frequency and yield of mass wasting on the Queen Charlotte Islands, British Columbia, British Columbia Ministry of Forests, Land Management Report, Victoria, 34, 1984. Rood, K., Hydrology and water use for salmon streams in the Chiliwack/Lower Fraser Habitat Management Area, British Columbia, Fraser River Action Plan, Department of Fisheries and Oceans, Report No. 2288, 1995. Ryder, J., The stratigraphy and morphology of paraglacial alluvial fans in south-central British Columbia, Canadian Journal of Earth Sciences, 8, 279–298, 1971. Ryder, J.M. and Associates, Chilliwack River Watershed, terrain classification maps, 1995. Saunders, I. R., Late Quaternary Geology and Geomorphology of the Chilliwack River Valley, British Columbia, Unpublished M.Sc. Thesis, Simon Fraser University, Burnaby, 1985. 201  Bibliography Saunders, I. R., J. J. Clague, and M. C. Roberts, Deglaciation of the Chilliwack River Valley, British Columbia, Canadian Journal of Earth Sciences, 24, 915–923, 1987. Schaepe, D., The maps of K’hharlsertern, c.1918, in St´ o:lo - Coast Salish Historical Atlas, edited by K. Carlson, pp. 126–128, Chilliwack Heritage St´o:lo Trust, Chilliwack, 1999. Schieffer, E., O. Slaymaker, and B. Klinkenberg, Physiographically controlled allometry of specific sediment yield in the Canadian Cordillera: A lake sediment-based approach, Geografiska Annaler, 83A, 55–65, 1999. Schumm, S., Evolution of drainage systems and slopes in badlands at Perth Amboy, New Jersey, Bulletin of the Geological Society of America, 67 (5), 597–646, 1956. Schumm, S., Geomorphic thresholds and complex response of drainage systems, in Fluvial Geomorphology, edited by M. Morisawa, pp. 299–310, Binghampton Publications in Geomorphology, 1973. Schumm, S., The Fluvial System, 358 pp., Wiley and Sons, New York, 1977. Schumm, S., To Interpret the Earth: Ten Ways to be Wrong, 143 pp., Cambridge University Press, 1991. Schumm, S., and D. Rea, Sediment yield from disturbed earth systems, Geology, 23, 3–9, 1995. Schumm, S., M. Mosley, and C. Watson, Experimental Fluvial Geomorphology, 413 pp., Wiley and Sons, New York, 1987. Sevigny, J., and E. Brown, Geochemistry and tectonic interpretations of some metavolcanic rock units of the western North Cascades, Washington, Geological Society of America Bulletin, 101, 391–400, 1989. Shaw, J., and R. Kellerhals, The Composition of Recent Alluvial Gravels in Alberta River Beds, 151 pp., Alberta Research Council, Edmonton, 1982. Shilts, W., Geochemical partitioning in till, in Drift Exploration, edited by J. N. P.T. Bobrowsky, S.J. Sibbick and P. Matysek, pp. 149–166, B.C. Ministry of Energy, Mines and Petroleum Resources, Victoria, 1995.  202  Bibliography Simpson, G., Historical science, in The Fabric of Geology, edited by C. Albritton, pp. 24–48, Addison-Wesley, Reading MA, 1963. Sinha, S., and G. Parker, Causes of concavity in longitudinal profiles of rivers, Water Resources Research, 32 (5), 1417–1428, 1996. Stillwater Sciences, Dam removal express assessment models (DREAM), Tech. Rep. 2002, Stillwater Sciences, Berkeley, CA, 2002. Strahler, A., Hypsometric (area-altitude curve) analysis of erosional topography, Bulletin of the Geological Society of America, 63 (11), 1117–1141, 1952. Strahler, A., Quantitative analysis of watershed geomorphology, Transactions of the American Geophysical Union, 38 (6), 913–920, 1957. Strahler, A. N., Quantitative geomorphology of drainage basins and channel networks, in Handbook of Applied Hydrology, edited by V. T. Chow, pp. 39–76, McGraw Hill, New York, 1964. Sugden, D., M. Summerfield, and T. Burt, Editorial: Linking short-term geomorphic processes to landscape evolution, Earth Surface Processes and Landforms, 22 (3), 193–194, 1997. Summerfield, M., The changing landscape of geomorphology, Earth Surface Processes and Landforms, 30 (6), 779–781, 2005. Tabor, R., R. Haugerud, W. Hildreth, and E. Brown, Geologic map of the Mount Baker 30 60 minute quadrangle, Washington: U.S. Geological Survey Map I-2660, scale 1:100,000, 2 sheets., 2003. Talbot, T., and M. Lapointe, Numerical modeling of a gravel bed river’s response to meander straightening: the coupling between the evolution of the bed pavement and long profile, Water Resources Research, 38 (6), doi:10.1029/2001WR000330, 2002. Tarboton, D., R. Bras, and I. Rodriguez-Iturbe, On the extraction of channel networks from digital elevation data, Hydrological Processes, 5, 81–100, 1991. Taylor, J., An Introduction to Error Analysis, 327 pp., University Science Books, Mill Valley, CA, 1997.  203  Bibliography Tepper, J., and S. Kuehner, Geochemistry of mafic enclaves and host granitoids from the Chilliwack Batholith, Washington: Chemical exchange processes between coexisting mafic and felsic magmas and implications for the interpretation of enclave chemical traits, The Journal of Geology, 112 (3), 349–367, 2004. Tepper, J., B. Nelson, G. Bergantz, and A. Irving, Petrology of the Chilliwack Batholith, North Cascades, Washington: generation of calc-alkaline granitoids by melting of mafic lower crust with variable water fugacity, Contributions to Mineralogy and Petrology, 113, 331–351, 1993. Thom, B., and L. Cameron, Changing land use in S’´olh T´em´exw: Population, transportation, ecology and heritage, in You are Asked to Witness: The St´ o:lo- and Canadas Pacific Coast History, edited by K. Carlson, pp. 163–180, St´o:lo Heritage Trust, 1996. Thompson, M., and R. Howarth, A new approach to the estimation of analytical precision, Journal of Geochemical Exploration, 9 (1), 23–30, 1978. Thomson, B., Slope failures - Chilliwack River drainage basin: A mass movement classroom, in 8th Congress of the International Association for Engineering Geology and the Environment, edited by D. Moore and O. Hungr, pp. 1097–1101, Vancouver, Canada, 1999. Thurber Engineering, Geotechnical Assessment of Slesse Park Landslide, 122 pp., Report to Department of Fisheries and Oceans, Vancouver, Canada, 1997. TRIM, British Columbia Terrain Resource Inventory Management Program, Victoria, BC, sheets 092H001, 092H002, 092H003, 092H0011, 092H0012, 092H0013, 1996. Tucker, G., and R. Slingerland, Drainage basin responses to climate change, Water Resources Research, 33 (8), 2031–2047, 1997. Turner, G., A 5 000 year geomagnetic palaeosecular variation record from western Canada, Geophysical Journal of the Royal Astronomical Society, 91, 103–121, 1987. van Niekerk, A., K. Vogel, R. Slingerland, and J. Bridge, Routing of Heterogeneous Sediments over Movable Bed: Model Development, Journal of Hydraulic Engineering, 118 (2), 246–262, 1992.  204  Bibliography Verosub, K., P. Mehringer, and P. Waterstraat, Holocene secular variation in Western North America: Paleomagnetic record from Fish Lake, Harney County, Oregon, Journal of Geophysical Research, 91 (B3), 3609–3624, 1986. Vogel, J., W. Cornell, D. Nelson, and J. Southon, Vesuvius/Avellino, one possible source of seventeenth century BC climatic disturbances, Nature, 344 (6266), 534–537, 1990. Walker, I., and M. Pellatt, Climate change in coastal British Columbia: a paleoenvironmental perspective, Canadian Water Resources Journal, 28 (4), 531–566, 2003. Warburton, J., Relative rates of glacial and nonglacial erosion in alpine environments, Arctic and Alpine Research, 25 (1), 1–7, 1993. Ward, B., and B. Thomson, Late Pleistocene stratigraphy and chronology of lower Chehalis River valley, southwestern British Columbia: evidence for a restricted Coquitlam Stade, Canadian Journal of Earth Sciences, 41 (7), 881–895, 2004. Watson, A., Geological investigation and activity of the Slesse Park landslide, Chilliwack River Valley, B.C., Unpublished B.Sc., The University of British Columbia, Vancouver, 1999. Westgate, J., Identification and significance of late Holocene tephra from Otter Creek, southern British Columbia, and localities in west-central Alberta, Canadian Journal of Earth Sciences, 14, 2593–2600, 1977. Wilcock, P., and J. Crowe, Surface-based transport model for mixed-size sediment, Journal of Hydraulic Engineering, ASCE, 129 (2), 120–128, 2003. Wilcock, P., and B. McArdell, Surface-based fractional transport rates: Mobilization thresholds and partial transport of a sand-gravel sediment, Water Resources Research, 29 (4), 1297–1312, 1993. Wojdyr,  M.,  Fityk  [fi:tik],  software  for  peak  fitting,  SourceForge.Net,  http://www.unipress.waw.pl/fityk/, last Accessed March 17 2007, 2007. Wong, M., G. Parker, P. DeVries, T. Brown, and S. Burges, Experiments on vertical and streamwise dispersion of tracer stones under lower-regime plane-bed equilibrium bedload transport, Water Resources Research, doi:10.1029/2006WR005172, 2007. 205  Bibliography Wood, J., Landserf - GIS for the visualisation and analysis of surfaces, Ph.D. thesis, School of Informatics, City University, London, http://www.soi.city.ac.uk/ jwo/, 2004. Wooldridge, C., Channel bar radar architecture and evolution in the wandering gravel-bed Fraser and Squamish Rivers, British Columbia, Canada, Unpublished M.Sc. Thesis, Simon Fraser University, Burnaby, 1999.  206  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0066305/manifest

Comment

Related Items