UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effects of glacier retreat on proglacial streams and riparian zones in the Coast and North Cascade Mountains Cowie, Natasha Michel 2011

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

Item Metadata

Download

Media
24-ubc_2012_spring_cowie_natasha.pdf [ 35.32MB ]
Metadata
JSON: 24-1.0072381.json
JSON-LD: 24-1.0072381-ld.json
RDF/XML (Pretty): 24-1.0072381-rdf.xml
RDF/JSON: 24-1.0072381-rdf.json
Turtle: 24-1.0072381-turtle.txt
N-Triples: 24-1.0072381-rdf-ntriples.txt
Original Record: 24-1.0072381-source.json
Full Text
24-1.0072381-fulltext.txt
Citation
24-1.0072381.ris

Full Text

Effects of Glacier Retreat on Proglacial Streams and Riparian Zones in the Coast and North Cascade Mountains by Natasha Michel Cowie B.Sc., Sewanee: The University of the South, 2007 a thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in the faculty of graduate studies (Geography) The University of British Columbia (Vancouver) November 2011 c© Natasha Michel Cowie, 2011 Abstract This study examined the trajectories of proglacial channel evolution in coastal British Columbia and Washington. Reach morphologies were identi- fied by field surveys of 70 headwater reaches in ten catchments in the Coast and North Cascade Mountains. Riparian vegetation development and poten- tial fish habitat were characterized in an additional 22 catchments using GIS analysis and satellite image interpretation. The study focused on reaches exposed by post-Little Ice Age (LIA) glacier retreat. Channel morphologies were predominantly governed by slope. The pre- dominance of slope as a morphological control is a reflection of the landscape template imposed by the Quaternary glaciation, which appears to override most of the modern effects of the LIA. However, high sediment yield in- duced by post-LIA deglaciation influences channel form by providing large amounts of readily mobile sediment. Logistic regression models of riparian vegetation and forest development indicate that vegetation presence and maturity are positively correlated with reach age and negatively correlated with reach elevation. The timeline for ri- parian development is longer than that reported for other proglacial streams, suggesting that post-LIA instability and sediment inputs delay the estab- lishment of riparian vegetation. A gradient-based classification tree model of potential fish presence in post-LIA channels suggests that fish may be able to access the majority of recently deglaciated headwaters, and that low-gradient, glacially carved hanging valleys may present habitat opportunities. The ability of fish to colonize proglacial streams will become increasingly important as climatic ii warming shifts thermal barriers for cold water species further upstream. Es- timates of maximum weekly average stream temperature (MWAT) indicate that, at present, the majority of proglacial streams are thermally suitable for cold water fish. However, future projections of MWAT without basin ice cover show a 30% decline in cold water species habitat within the study basins. The work presented here contributes to the understanding of recently deglaciated headwaters by identifying first-order controls on proglacial stream morphology and riparian vegetation, which influence habitat and govern channel change in new streams. iii Table of Contents Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . iv List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . xii 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 Motivation for the study . . . . . . . . . . . . . . . . . . . . . 1 1.2 Literature review . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.2.1 Landscape context for proglacial channel change . . . 3 1.2.2 Proglacial stream hydrology and biotic succession . . . 4 1.3 Research questions and thesis structure . . . . . . . . . . . . 12 2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2.1 Site selection . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2.2 Field surveys . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.2.1 Site characteristics and survey methods . . . . . . . . 16 2.2.2 Reach morphology . . . . . . . . . . . . . . . . . . . . 19 2.2.3 Topographic analysis . . . . . . . . . . . . . . . . . . . 21 2.2.4 Shear stress, stream power, and discharge calculations 23 2.3 Riparian forest analysis . . . . . . . . . . . . . . . . . . . . . 26 2.3.1 Data collection . . . . . . . . . . . . . . . . . . . . . . 26 iv 2.3.2 Channel pattern and vegetation classification . . . . . 27 2.3.3 Data analysis . . . . . . . . . . . . . . . . . . . . . . . 30 2.4 Fish habitat analysis . . . . . . . . . . . . . . . . . . . . . . . 32 2.4.1 Habitat characterization . . . . . . . . . . . . . . . . . 32 2.4.2 Data collection . . . . . . . . . . . . . . . . . . . . . . 32 2.4.3 Modelling approach . . . . . . . . . . . . . . . . . . . 35 3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 3.1 Channel morphology . . . . . . . . . . . . . . . . . . . . . . . 37 3.1.1 Effects of slope and contributing area . . . . . . . . . 37 3.1.2 Shear stress and substrate resistance . . . . . . . . . . 43 3.1.3 Channel size and grain roughness characteristics . . . 46 3.2 Riparian vegetation analysis . . . . . . . . . . . . . . . . . . . 49 3.2.1 Correlations among numerical variables . . . . . . . . 49 3.2.2 Correlations among categorical variables . . . . . . . . 53 3.3 Fish habitat analysis . . . . . . . . . . . . . . . . . . . . . . . 54 3.3.1 Gradient-based assessment . . . . . . . . . . . . . . . 54 3.3.2 Temperature-based assessment . . . . . . . . . . . . . 58 4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 4.1 Channel morphology . . . . . . . . . . . . . . . . . . . . . . . 60 4.2 Riparian vegetation . . . . . . . . . . . . . . . . . . . . . . . . 64 4.3 Fish habitat . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 4.4 Implications for models of proglacial stream evolution . . . . 69 5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 5.1 Key findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 5.2 Directions for future research . . . . . . . . . . . . . . . . . . 72 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 A Study sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 B Field site descriptions . . . . . . . . . . . . . . . . . . . . . . 98 B.1 Data sources and overview map . . . . . . . . . . . . . . . . . 98 v B.2 British Columbia Coast Mountains: Garibaldi and Joffre Lakes Provincial Parks, Pemberton . . . . . . . . . . . . . . . . . . 100 B.2.1 Helm Glacier . . . . . . . . . . . . . . . . . . . . . . . 102 B.2.2 Overlord Glacier . . . . . . . . . . . . . . . . . . . . . 105 B.2.3 Russet Lake . . . . . . . . . . . . . . . . . . . . . . . . 108 B.2.4 Wedgemount Glacier . . . . . . . . . . . . . . . . . . . 111 B.2.5 Matier Glacier . . . . . . . . . . . . . . . . . . . . . . 114 B.2.6 Miller Glacier . . . . . . . . . . . . . . . . . . . . . . . 117 B.3 British Columbia Insular Mountains: Strathcona Provincial Park . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 B.3.1 Septimus Glacier . . . . . . . . . . . . . . . . . . . . . 121 B.3.2 Colonel Foster Glacier . . . . . . . . . . . . . . . . . . 124 B.4 Washington North Cascade Mountains: Mount Baker-Snoqualmie National Forest . . . . . . . . . . . . . . . . . . . . . . . . . . 127 B.4.1 Boulder Glacier . . . . . . . . . . . . . . . . . . . . . . 128 B.4.2 Easton Glacier . . . . . . . . . . . . . . . . . . . . . . 131 C Channel morphology types . . . . . . . . . . . . . . . . . . . 134 D Discharge estimates . . . . . . . . . . . . . . . . . . . . . . . 139 E Data sources for riparian forest analysis . . . . . . . . . . . 141 F Additional reach morphology plots . . . . . . . . . . . . . . 144 vi List of Tables Table 1.1 Summary of previous proglacial stream studies . . . . . . . 6 Table 1.2 Timeline of proglacial stream development . . . . . . . . . 7 Table 2.1 Characteristics of catchments included in field survey . . . 16 Table 2.2 Summary of mountain channel morphology characteristics 20 Table 2.3 Comparison of discharge estimates . . . . . . . . . . . . . . 26 Table 2.4 Data sources for fish habitat analysis . . . . . . . . . . . . 33 Table 3.1 Coefficients and p-values for binary logistic regression model to predict the presence of riparian vegetation . . . . . . . . 51 Table 3.2 Coefficients and p-values for binary logistic regression model to predict the presence of riparian forest . . . . . . . . . . 52 Table 3.3 Results of Fisher’s exact test for signficance of categorical variables in riparian development . . . . . . . . . . . . . . 53 Table 3.4 Summary of results from fish habitat analysis for channels deglaciated post-Little Ice Age . . . . . . . . . . . . . . . . 55 Table 3.5 Summary of projected maximum weekly average temper- atures for basins with and without ice . . . . . . . . . . . . 58 Table A.1 Catchments and reaches included in study . . . . . . . . . 91 Table B.1 Data sources for field sites . . . . . . . . . . . . . . . . . . 98 Table B.2 Characteristics of Helm Glacier and catchment . . . . . . . 103 Table B.3 Characteristics of Overlord Glacier and catchment . . . . . 106 Table B.4 Characteristics of Russet Lake catchment . . . . . . . . . . 109 Table B.5 Characteristics of Wedgemount Glacier and catchment . . 112 vii Table B.6 Characteristics of Matier Glacier and catchment . . . . . . 115 Table B.7 Characteristics of Miller Glacier and catchment . . . . . . 118 Table B.8 Characteristics of Septimus Glacier and catchment . . . . 122 Table B.9 Characteristics of Colonel Foster Glacier and catchment . 125 Table B.10 Characteristics of Boulder Glacier and catchment . . . . . 129 Table B.11 Characteristics of Easton Glacier and catchment . . . . . . 132 Table D.1 Summary of k -factors used to calculate discharge . . . . . 140 Table E.1 Data sources for riparian forest analysis . . . . . . . . . . . 142 viii List of Figures Figure 1.1 Schematic representation of area-based process domains in unglaciated and glaciated landscapes . . . . . . . . . . 5 Figure 2.1 Locations of study areas . . . . . . . . . . . . . . . . . . . 15 Figure 2.2 Definition of bankfull width . . . . . . . . . . . . . . . . . 19 Figure 2.3 Illustration of a stream profile . . . . . . . . . . . . . . . . 22 Figure 2.4 Channel types classified from satellite images . . . . . . . 29 Figure 2.5 Riparian vegetation classified from satellite images . . . . 29 Figure 2.6 Example of stream segment delineation . . . . . . . . . . 34 Figure 3.1 Slope-area plot . . . . . . . . . . . . . . . . . . . . . . . . 39 Figure 3.2 Slope-discharge plot . . . . . . . . . . . . . . . . . . . . . 40 Figure 3.3 Bankfull channel dimensions and grain size plotted as functions of drainage area . . . . . . . . . . . . . . . . . . 42 Figure 3.4 Boxplot of slope distributions . . . . . . . . . . . . . . . . 43 Figure 3.5 Bankfull channel dimensions and grain size plotted as functions of shear stress . . . . . . . . . . . . . . . . . . . 45 Figure 3.6 Shear stress-D95 ratio plotted as a function of drainage area 46 Figure 3.7 Bankfull channel dimensions and grain size plotted as functions of slope . . . . . . . . . . . . . . . . . . . . . . . 48 Figure 3.8 Boxplots of riparian forest development stage versus reach age and elevation . . . . . . . . . . . . . . . . . . . . . . . 50 Figure 3.9 Logistic regression model of riparian vegetation establish- ment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Figure 3.10 Logistic regression model of riparian forest establishment 52 ix Figure 3.11 Classification tree model of fish presence or absence in proglacial streams . . . . . . . . . . . . . . . . . . . . . . 57 Figure 3.12 Plots of projected maximum weekly average temperatures for basins with and without ice . . . . . . . . . . . . . . . 59 Figure B.1 Map of field survey region . . . . . . . . . . . . . . . . . . 99 Figure B.2 Field study sites in the BC Coast Mountains . . . . . . . 101 Figure B.3 Map of Helm Glacier and catchment . . . . . . . . . . . . 102 Figure B.4 Helm channel profile . . . . . . . . . . . . . . . . . . . . . 103 Figure B.5 Photographs of Helm Glacier channel . . . . . . . . . . . 104 Figure B.6 Map of Overlord Glacier, Russet Lake, and catchments . 105 Figure B.7 Overlord channel profile . . . . . . . . . . . . . . . . . . . 106 Figure B.8 Photographs of Overlord Glacier channel . . . . . . . . . 107 Figure B.9 Russet Creek profile . . . . . . . . . . . . . . . . . . . . . 109 Figure B.10 Photographs of Russet Creek . . . . . . . . . . . . . . . . 110 Figure B.11 Map of Wedgemount Glacier and catchment . . . . . . . . 111 Figure B.12 Wedgemount channel profile . . . . . . . . . . . . . . . . . 112 Figure B.13 Photographs of Wedgemount Glacier channel . . . . . . . 113 Figure B.14 Map of Matier Glacier and catchment . . . . . . . . . . . 114 Figure B.15 Matier channel profile . . . . . . . . . . . . . . . . . . . . 115 Figure B.16 Photographs of Matier Glacier channel . . . . . . . . . . . 116 Figure B.17 Map of Miller Glacier and catchment . . . . . . . . . . . . 117 Figure B.18 Miller channel profile . . . . . . . . . . . . . . . . . . . . . 118 Figure B.19 Photographs of Miller Glacier channel . . . . . . . . . . . 119 Figure B.20 Field study sites in the BC Insular Mountains . . . . . . . 120 Figure B.21 Map of Septimus Glacier and catchment . . . . . . . . . . 121 Figure B.22 Septimus channel profile . . . . . . . . . . . . . . . . . . . 122 Figure B.23 Photographs of Septimus Glacier channel . . . . . . . . . 123 Figure B.24 Map of Colonel Foster Glacier and catchment . . . . . . . 124 Figure B.25 Colonel Foster channel profile . . . . . . . . . . . . . . . . 125 Figure B.26 Photographs of Colonel Foster Glacier channel . . . . . . 126 Figure B.27 Field study sites in the WA North Cascade Mountains . . 127 Figure B.28 Map of Boulder Glacier and catchment . . . . . . . . . . 128 x Figure B.29 Boulder channel profile . . . . . . . . . . . . . . . . . . . 129 Figure B.30 Photographs of Boulder Glacier channel . . . . . . . . . . 130 Figure B.31 Map of Easton Glacier and catchment . . . . . . . . . . . 131 Figure B.32 Easton channel profile . . . . . . . . . . . . . . . . . . . . 132 Figure B.33 Photographs of Easton Glacier channel . . . . . . . . . . 133 Figure C.1 Channel morphology types . . . . . . . . . . . . . . . . . 138 Figure F.1 Bankfull channel dimensions and grain size plotted as functions of stream power . . . . . . . . . . . . . . . . . . 145 Figure F.2 Stream power-D95 ratio plotted as a function of drainage area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 xi Acknowledgments I am deeply grateful for the support and contributions from many people who made this work possible. First, thanks to my supervisors Marwan Has- san and Dan Moore. Ideas, guidance, and encouragement from Dan and Marwan helped me enormously. Throughout this project I received a great deal of assistance from other members of the UBC Department of Geogra- phy, specifically Rich McCleary, Brett Eaton, Michael Church, Jason Leach, Joel Trubilowicz, Dori Kovanen, and David Reid. Thanks also to Stephen Mitchell of the UBC Department of Forest Sciences, Jordan Rosenfeld of the BC Ministry of Environment, and Mauri Pelto of Nichols College for sugges- tions and advice. Many thanks to my extraordinary research assistant Luisa Muenter. Also, thanks to Pascal Szeftel and Sarah Davidson for field assis- tance. Despite often-rigorous field conditions, Luisa, Pascal, and Sarah kept up the hard work and enjoyable companionship. I appreciate the inspiration I received from the faculty at the University of the South. Special thanks to Robert Fuller at North Georgia College and State University, who first introduced me to geography. My family has always given me tremendous support and encouragement, for which I am grateful beyond words. Finally, thanks to Julius for accompanying me on this and many other adventures. Funding was provided by the Canadian Foundation for Climate and At- mospheric Sciences through its support of the Western Canadian Cryospheric Network and by a Natural Sciences and Engineering Research Council Dis- covery Grant. Additional funding was provided by the UBC Department of Geography in the form of a Teaching Assistantship. xii Chapter 1 Introduction 1.1 Motivation for the study Glaciers worldwide have been in sustained retreat since the end of the Little Ice Age (LIA) in the 19th century. Over the last several decades glacier retreat has accelerated, with pronounced loss in coastal regions with mar- itime climates (Harper, 1993). Recent rates of glacier loss in the British Columbia Coast Mountains are approximately double that of the last two decades, most likely due to increased temperatures and decreased snowfall throughout the late 1980s and 1990s (Schiefer et al., 2007; VanLooy and Forster, 2008). Between 1985 and 2005, British Columbia glaciers lost al- most 11% of their area, a regional annual shrinkage rate of 0.55% per year (Bolch et al., 2010). As glaciers retreat, new ice-marginal proglacial landscapes emerge that are directly conditioned by glaciation. New streams forming in the absence of glaciers are profoundly influenced by the glacial legacy and, if any ice remains, the present-day glacial activity. On a template set by previous glaciation, complex interactions among a wide range of processes and con- trols shape the evolution of proglacial streams. Glacier decline causes a shift in water source contributions and altered discharge patterns. Debuttressed, unconsolidated glacial deposits feed large amounts of sediment into stream systems. Pioneer vegetation stabilizes streambanks and provides organic 1 matter to steep, cold proglacial streams. Ongoing glacial recession and associated landscape changes will impact physical, biogeochemical, and ecological conditions in streams, with impli- cations for aquatic ecosystem community composition and productivity, fish spawning habitat, water supply, and downstream fluxes of nutrients and sediment. As headwaters, proglacial streams are of critical ecological impor- tance as habitat and in providing sediment, nutrients, and organic matter to downstream reaches. Variations in disturbance regimes, channel stability, and riparian succession in headwaters alter physical and biological condi- tions along the entire channel. Understanding the development of proglacial streams is crucial to informing management decisions as glaciers continue to retreat. Although proglacial systems have received much research attention in recent years, most work has focused on hydrology. Relatively few studies have evaluated proglacial streams in the context of habitat, channel mor- phology, and landscape recovery. The two existing conceptual models of proglacial stream development by Sidle and Milner (1989) and Gurnell et al. (1999) provide a basis for future research, but they are respectively region- specific and highly generalized, making them unsuitable for application to other areas. This study addresses the gap in knowledge by developing a conceptual model of proglacial stream development in the Pacific North- west, using a combination of field- and GIS-derived data and focusing on streams deglaciated since the end of the LIA. Previous studies of mountain stream morphology in unglaciated settings (Montgomery and Buffington, 1997) and in basins that were glaciated during the Pleistocene, but not the LIA (Brardinoni and Hassan, 2006, 2007) provide a regional context for the results presented here. The overall objective of this study was to contribute to the understanding of controls on proglacial stream development. The remainder of this chapter consists of a literature review (Section 1.2) which provides context for the specific research questions outlined in Section 1.3. 2 1.2 Literature review 1.2.1 Landscape context for proglacial channel change Present-day proglacial processes are superimposed on a landscape template formed by the Pleistocene glaciation 18,000 years ago (Smith et al., 2001; Brardinoni and Hassan, 2006; Collins and Montgomery, 2011). After ice disappears, a transitional paraglacial landscape state may last for tens of thousands of years (Ryder, 1971a,b; Church and Ryder, 1972; Hewitt, 1972; Ballantyne, 2002). A high degree of instability and geomorphic heterogene- ity characterizes these landscapes in incomplete recovery from the distur- bance of glaciation. Subsystems may reach interim states of adjustment while the landscape remains in long-term disequilibrium (Church and Slay- maker, 1989; Church and Hassan, 2002; Slaymaker, 2009). The classic conceptual model of paraglacial sediment yield developed by Church and Slaymaker (1989) predicts a prolonged sediment pulse through stream systems after deglaciation. Sediment yield peaks in upland basins at the time of deglaciation, then gradually declines over thousands of years. As sediment flushes from upland basins, it initiates an extended cycle of movement through larger basins. The entire sediment pulse appears to op- erate over timescales of 101-104 years (Church and Slaymaker, 1989; Ballan- tyne, 2002), indicating that landscapes may not completely recover between glaciations. The persistent effects of glaciation give mountain landscapes a high degree of geomorphic diversity and impose a metastable state, mak- ing them prone to disturbance and accelerated erosion (Barsch and Caine, 1984). Headwater channels that have not been glaciated since the Pleis- tocene are presently undergoing degradation after the long-term aggrada- tion of the last glacial maximum (Brardinoni et al., 2009). In contrast, neoglaciated channels are presently at or near peak sediment yield as dis- integration of LIA moraines contributes enormous volumes of material to modern-day proglacial systems (Church et al., 1989; Hasholt et al., 2008). The long-term glacial landscape legacy is reflected by process domain se- quences in glaciated basins (Brardinoni and Hassan, 2006). In unglaciated 3 fluvial landscapes, process domains typically follow a regular progression from small, high gradient areas dominated by mass movement to large, lower gradient areas dominated by fluvial processes (Figure 1.1a). Glacial land- forms introduce discontinuities in the typical sequence. Glaciation imposes a sequence of hanging valleys, relatively low gradient zones separated by steeper bedrock steps. A ‘hanging fluvial’ process domain develops within the hanging valleys, giving rise to channel-reach morphologies such as riffle- pool that are generally more characteristic of unglaciated landscapes (Figure 1.1b). Stream reaches separated by glacial landscape features may individu- ally achieve a state of quasi-equilibrium, but persistent glacial features pre- vent the reaches from adjusting to each other, thus preventing equilibrium within the entire channel (e.g. Ponton, 1972; Ballantyne, 2002; Slaymaker, 2009). 1.2.2 Proglacial stream hydrology and biotic succession Many elements of proglacial landscape development have received detailed research attention. Studies with particular relevance to proglacial stream development are summarized in Table 1.1 and discussed further in the re- maining review. However, to date only two conceptual models have been proposed to synthesize the body of knowledge about proglacial stream evo- lution. Sidle and Milner (1989) developed a timeline of proglacial fluvial succes- sion based on chronosequence studies of streams flowing from valley glaciers in Glacier Bay, Alaska, linking sediment supply, hydrology, and vegetation succession to channel change (Table 1.2). Drawing on data from long-term studies in the European Alps and Pyrenees, Gurnell et al. (1999) intro- duced altitude and glacier activity as additional controls on proglacial chan- nels. Their model proposes a steady downstream progression from sediment- controlled wide braided channels to more stable single- or multi-threaded channels controlled by riparian vegetation. Altitude and ice front behaviour (advancing, stationary, or retreating) govern the temporal scale of channel development. Both models encompass the timescale of 101-102 years. 4 (a) (b) Figure 1.1: Schematic representation of area-based process domains (dashed lines) and topographic signatures (solid lines) in (a) unglaciated (adapted from Brardinoni and Hassan, 2006 after Montgomery and Foufoula-Georgiou, 1993) and (b) glaciated (adapted from Brardinoni and Hassan, 2006) landscapes. Rel- evant macro-forms are noted in brackets. Images reproduced by permission of the American Geophysical Union. 5 Table 1.1: Summary of previous proglacial stream studies. Geomorphology, Hydrology, and Sedimentation Riparian and Fluvial Biotic Succession Process Linkages and Conceptual Models Church and Ryder (1972) Milner (1987) Sidle and Milner (1989) Paraglacial geomorphology Stream colonization Timeline of stream development Ponton (1972) Chapin et al. (1994) Ward (1994) Hydraulic geometry of glacial streams Primary terrestrial succession Alpine stream ecology Smith (1976) Milner and Petts (1994) McGregor et al. (1995) Morphological role of vegetation Stream habitat development Alpine stream ecology Maizels (1983) Flory and Milner (1999, 2000) Tockner et al. (1997) Channel change over varying timescales Macroinvertebrate communities Physicochemical characterization Church and Slaymaker (1989) Robertson and Milner (1999, 2006) Gurnell et al. (1999) Paraglacial sediment cycles Meiofaunal communities Model of proglacial river evolution Ballantyne (2002) Milner et al. (2000) Smith et al. (2001) Paraglacial geomorphology Macroinvertebrate communities Proglacial hydrogeomorphology Chew and Ashmore (2001) Milner et al. (2001) Brown et al. (2003, 2007) Channel adjustment Macroinvertebrate communities Stream habitat classification Brardinoni and Hassan (2006, 2007) Malard et al. (2003) Hannah et al. (2007) Glacial influence on stream morphology Hyporheic invertebrates Climate-hydrology-ecology linkages Hasholt et al. (2008) Brown et al. (2005, 2006) Hood and Berner (2009) Landscape and sediment processes Physicochemical habitat variables Stream biogeochemistry Klaar et al. (2009, 2011) Milner and Gloyne-Phillips (2005) Milner et al. (2009) Geomorphic role of instream wood Macroinvertebrate communities Hydroecology and glacier decline Collins and Montgomery (2011) Milner et al. (2008) Moore et al. (2009) Pleistocene glacial legacy Biotic community development Implications of climate change 6 Table 1.2: Conceptual model of proglacial stream development modified from Sidle and Milner (1989). Early Mid Late Hydrology Somewhat attenuated peak flows; highly depen- dent on glacial melting Peakflows more depen- dent on autumn storms; varying snowmelt peak- flows Peakflow magnitudes di- minish over time as veg- etation establishes; lakes may attenuate rainfall peakflows Sediment supply Major inputs from glacier; minor inputs from bank erosion, surface ero- sion, and talus creep Initially active bank ero- sion becoming more mod- erate with time Minor erosion from banks and upstream sources Channel condition and adjustment Wide channel and braided floodplain; most silt transported through system Vegetation initiates bar formation; minor scour associated with bedload transport; channel nar- rows Sediment supply and transport reaching equi- librium; thalweg deepens; instream wood generates pools Role of riparian vegetation Insignificant Streambank beginning to stabilize; alder and willow clones establish in channel and form matrix for woody debris dams Major factor influencing channel form and sediment transport; large trees sta- bilize bars and influence pool development 7 Both the Sidle and Milner (1989) and Gurnell et al. (1999) models in- corporate the effects of proglacial sediment supply on channel form in terms of present-day glacier activity, relating channel sediment regime to stream age, stage of vegetation development, and proximity to the ice front. How- ever, neither model frames proglacial channel change in a broader context of time and space. The model timescales fall short of the longer scope of glacial landscape influence, which may span 104 years. Additionally, both models make the major simplifying assumption that no downstream water or sediment inputs significantly impact the channel. The Sidle and Milner (1989) and Gurnell et al. (1999) models both in- dicate an initially transport-limited regime as glaciers retreat. In the time period immediately following deglaciation, landscape sediment yield is at a maximum and streams are choked with more sediment than they can carry, giving rise to wide, braided channels. Ice front stabilization, or the disap- pearance of the glacier, may eventually stabilize sediment inputs, allowing formation of bars and islands in a more supply-limited regime. If the channel is below tree line, establishment of riparian vegetation is likely to accelerate channel stabilization by binding sediment with roots to reinforce banks. Hydrologic regime is a primary component of the Sidle and Milner (1989) model. It is not explicitly included by Gurnell et al. (1999), although later work by Smith et al. (2001) linked the Gurnell model to climatic and hydro- logical data to provide a context for ecological research in proglacial streams. Glacier retreat changes water source contributions and discharge patterns (e.g. Fleming, 2005; Moore et al., 2009). The behaviour and relative size of the glacier substantially affect streamflow, thus influencing local channel morphology. Additionally, the age and stability of the stream itself dictate its response to glacier changes. Glaciers modulate runoff by temporarily storing water as ice. This tends to result in a delay of maximum seasonal flow and a decrease in annual and monthly runoff variation (Fountain and Tangborn, 1985). With glacier de- cline, meltwater contributions initially increase streamflow, but over time flow decreases as glaciers shrink. Flows fluctuate more as basin ice cover decreases and the hydrological regime becomes predominantly driven by 8 spring meltwater and autumn storms. Groundwater contributions may pro- portionally increase. Stahl and Moore (2006) suggested that most glaciers in British Columbia have already passed the initial phase of warming-induced increased late summer runoff and are now facing continual declines, as ev- idenced by significant negative trends in August streamflow in glacier-fed streams. Changing water source contributions influence biogeochemistry and, in turn, benthic communities (McGregor et al., 1995; Smith et al., 2001; Brown et al., 2003, 2007; Hannah et al., 2007). Geomorphically, a reduction of melt- water inputs would significantly impact off-channel habitats (side-channels and sloughs) that depend on glacial runoff to sustain habitat availability and connectivity, particularly for fish (Milner et al., 2009). Vegetation colonization of glacial forefields has been extensively stud- ied (e.g. Cooper, 1923a,b,c; Lawrence et al., 1967; Chapin et al., 1994; Fastie, 1995; Dolezal et al., 2008). However, patterns of proglacial terres- trial vegetation colonization cannot be directly applied to riparian zones. Establishment and succession of vegetation tends to be slower in riparian zones than in adjacent uplands due to the instability of shifting stream channels, high sediment loads, and topographic position (Sidle and Milner, 1989). In contrast to terrestrial vegetation studies relating terrain age to vegetation development (Matthews, 1979; Matthews and Whitaker, 1987; Mizuno, 1998; Caccianiga et al., 2001), Milner and Gloyne-Phillips (2005) found that riparian vegetation was independent of stream age, but positively correlated with bank stability. Altitude appears to be a major controlling in- fluence on riparian woody vegetation and the rate of vegetation succession (Gurnell et al., 1999). Previous work has emphasized the buffering influ- ence of proglacial lakes on channel development by stabilizing sediment and discharge fluctuations, thus allowing riparian vegetation to establish more quickly (Milner, 1987; Sidle and Milner, 1989; Milner et al., 2000; Chew and Ashmore, 2001; Milner and Gloyne-Phillips, 2005; Milner et al., 2008). As riparian succession proceeds, vegetation has an increasingly impor- tant role in channel development. Riparian vegetation communities interact with their adjacent channels; they are well-adjusted to local flow conditions 9 and provide an active control on local channel change (Gurnell, 1995). Soil stabilization begins at an early stage of succession, as pioneer herbaceous communities retain fine sediment and intercept runoff (Milner et al., 2000; Corenblit et al., 2009). Even small roots of grass and willow assist in re- inforcing bank sediment, creating a riprap-like protection of channel banks from erosion (Smith, 1976). At later stages of succession, the amount of instream wood increases, contributing substantially to channel stabilization. Large debris dissipates stream energy. If positioned in a flow-diverting orientation, it may protect stream banks from erosion. The introduction of woody debris to a stream diversifies flow velocity and facilitates the formation of stable channel struc- tures such as steps and pools, which trap sediment and buffer downstream areas from sediment and flow variations (Swanson and Lienkaemper, 1978; Fetherston et al., 1995; Buffington and Montgomery, 1999; Hassan et al., 2005b). Even relatively small pieces are potentially of morphological impor- tance in proglacial streams. Klaar et al. (2011) found that alder boles play an important role in creating stable structures such as mid-channel bars due to their ability to re-root in the streambed. In addition to their role as a stabilizing component, instream wood struc- tures generate habitat heterogeneity. A high level of habitat diversity is important in recently deglaciated areas for the establishment of many inver- tebrate taxa (Tockner et al., 1997; Milner et al., 2000; Milner and Gloyne- Phillips, 2005; Milner et al., 2008) and in providing potential habitat and spawning areas for fish (Kondolf et al., 1991; Bryant et al., 2007). De- bris structures in streams force patterns of scour and deposition, generating alterations in stream depth, velocity, and shear stress surrounding the struc- tures. This creates a high diversity of geomorphic units and thus increased habitat heterogeneity (Milner et al., 2000; Hassan et al., 2005b; Milner et al., 2008; Klaar et al., 2009, 2011). The hydraulic roughness provided by wood generates textural fining, potentially creating salmonid spawning gravels in channels that otherwise would be too coarse (Kondolf et al., 1991; Buffington and Montgomery, 1999). Once riparian forest is established, it serves as a ‘diversity reservoir’ 10 which can support regeneration after destructive floods (Corenblit et al., 2009). Many riverine invertebrate taxa are dependent on established ripar- ian vegetation (Flory and Milner, 1999; Milner et al., 2008). Roots and branches of riparian vegetation trailing into streams (trailing riparian habi- tat) support a number of macroinvertebrate taxa and provide cover for ju- venile and spawning fish (Milner et al., 2000; Milner and Gloyne-Phillips, 2005). Invertebrate colonization of glacial streams has received a large amount of research attention (e.g. Milner, 1994; Milner and Petts, 1994; Flory and Milner, 1999, 2000; Milner et al., 2001; Malard et al., 2003; Brown et al., 2003, 2005, 2006, 2007; Robertson and Milner, 1999, 2006; Milner et al., 2008). The development of macroinvertebrate and meiofaunal assemblages is important to development of the fish community, even if the reach does not contain fish - fishless habitat has potential to provide important subsidies to downstream communities (Milner et al., 2000, 2001; Wipfli and Gregovich, 2002; Wipfli et al., 2007; Richardson et al., 2010). Much work has focused on temperature gradients as drivers of downstream changes in glacial stream biotic communities (e.g. Ward, 1994; Milner et al., 2008) but flow regime, sedimentation, and time since deglaciation (elements of channel stability) are also important variables in biotic colonization and development (Milner, 1987; Sidle and Milner, 1989; Milner, 1994; Milner and Petts, 1994; Flory and Milner, 2000). Fish colonization of proglacial streams has not been extensively inves- tigated, with the exception of several long-term studies in Glacier Bay Na- tional Park, Alaska. Milner (1987) found that establishment and production of Glacier Bay salmonid populations is principally related to streamflow and sediment characteristics. Later studies in the same area suggest that colo- nization and succession of fish communities are strongly related to stream age, because habitat complexity and stability (as indicated by variables such as instream wood, pools, and fish cover) increase with age (Sidle and Milner, 1989; Milner et al., 2000; Robertson and Milner, 2006). The Glacier Bay fish colonization studies were conducted on streams originating from valley glaciers and flowing at relatively low gradients di- 11 rectly into the ocean, so physical barriers to migration were not a factor. The physical limitations to longitudinal fish migration in streams flowing from high altitude glaciers are less clearly understood and may involve complex interactions between barriers and temperature. At higher altitudes, physical barriers to longitudinal migration often prevent fish from exploiting habi- tat that would otherwise be available to them upstream (Hari et al., 2006). However, the diversity of stream morphologies left by the glacial influence on the landscape may present a habitat opportunity for fish, particularly in lower gradient hanging valleys. 1.3 Research questions and thesis structure The literature review in Section 1.2 has identified a number of knowledge gaps in the understanding of proglacial stream evolution, which provide the context for the study presented here. The overall objective of this study was to contribute to the understanding of proglacial stream development, focusing on streams originating from alpine glaciers in the Coast and North Cascade Mountains. The specific research questions addressed by the thesis are: 1. What are the most important controls on morphology in streams deglaciated since the end of the LIA? How do the morphological char- acteristics of these recently deglaciated streams compare to other streams in the same region that have never been glaciated, or have remained unglaciated since the end of the Pleistocene glacial maximum? 2. What are the primary controls on the establishment and growth of ri- parian vegetation alongside proglacial streams? How many years does it take for riparian vegetation to become a potential morphological influence on proglacial channels? 3. Is glacier retreat exposing new areas for potential fish habitat? Are proglacial stream gradients and temperatures generally amenable to fish colonization? 12 The remainder of the thesis is organized as follows. Chapter 2 describes the field study sites and survey methods, additional data gathered from remote sensing sources, and methods of data analysis. Chapter 3 presents the results of field surveys and GIS analysis. Chapter 4 discusses how the results address the research questions outlined above. Chapter 5 summarizes the main conclusions of the study and identifies topics for future research. 13 Chapter 2 Methods 2.1 Site selection The primary criterion for site selection was the availability of previously mapped and dated historic glacier extents. However, three sites were in- cluded for which moraine ages are not available. Clearly defined, recently deposited moraines were observed at all of the undated locations, indicat- ing that the surveyed channel extent has been exposed since the Little Ice Age (LIA) peak. Accessibility of channels by hiking was another important constraint on site selection for field surveys, as most proglacial streams are in remote, high elevation locations that are not easily reached. Topographic maps, air photos, and satellite images were used to determine accessibility. In total, study sites included 117 headwater reaches in 32 catchments in the Coast and Insular Mountains of British Columbia and the North Cascades of Washington (Figure 2.1, Appendix A). Field surveys were con- ducted in ten of these catchments. Information collected from maps, air photos, satellite images, and fish distribution datasets was used to extend the analysis to 22 additional catchments. 14 Figure 2.1: Locations of study areas. Data for areas marked in red were obtained through field surveys and remote image inter- pretation, and data for areas marked in blue were obtained through remote image interpretation only. 15 2.2 Field surveys 2.2.1 Site characteristics and survey methods Of the ten field sites, nine channels (60 reaches) experienced glaciation dur- ing the LIA, with Russet Lake the exception (Table 2.1). Field sites were located in or near Garibaldi Provincial Park, Strathcona Provincial Park, Joffre Lakes Provincial Park, Pemberton, and the Mount Baker Wilderness Area. Detailed field site descriptions are in Appendix B. Table 2.1: Characteristics of catchments included in field survey. Catchment area is determined at the downstream end of the surveyed channel. Glacier Coordinates Geology Catchment Area (km2) Survey Length (m) Easton 48◦45’N 121◦45’W basaltic & andesitic volcanic 3.35 413 Boulder 48◦43’N 121◦50’W basaltic & andesitic volcanic 5.45 733 Matier 50◦20’N 122◦27’W dioritic intrusive 3.21 202 Miller 50◦22’N 122◦59’W quartz dioritic in- trusive 3.25 1315 Wedgemount 50◦09’N 122◦48’W quartz dioritic in- trusive 5.87 317 Helm 49◦57’N 123◦00’W basaltic volcanic, marine sedimentary 8.23 1774 Overlord 50◦02’N 122◦50’W andesitic volcanic 9.68 1525 Russet Lake 50◦01’N 122◦51’W andesitic volcanic 1.86 475 Colonel Foster 49◦45’N 125◦51’W basaltic volcanic 2.45 330 Septimus 49◦29’N 125◦31’W basaltic & rhyolitic volcanic 6.78 258 16 Field survey methods were adapted from the BC Channel Assessment Procedure, hereinafter referred to as the CAP (Province of British Columbia, 1996). Surveys began as close as possible to the base of the ice, although in most cases steep terrain or a lake prevented access to the ice terminus. Surveys generally began 130-750 m from the ice, although steep forefields on the two Mount Baker glaciers forced those surveys to start about 900-1000 m downvalley. In several reaches, flow originated from multiple points beneath the ice or split into many channels; in these cases only the main channel was surveyed in accordance with the CAP, which recommends following the thalweg. Channel reaches were identified in the field. Reach limits were defined as major changes in channel form, sediment supply, riparian vegetation, streambed or bank materials, confinement, relative coupling, tributary con- fluence, or gradient (Province of British Columbia, 1996; Wohl, 2000). In the strictest sense, reaches are morphologically homogeneous sections of channel under the influence of uniform governing conditions (Church, 1992). How- ever, applying this definition in the field is somewhat subjective and reach lengths vary in the literature. Previous studies have used reach lengths of approximately 30 bankfull widths (w) (Hogan, 1986), 10-20 w (Mont- gomery and Buffington, 1997), and roughly 20 w (Wood-Smith and Buffin- gton, 1996). The CAP uses a minimum of only 3 w, which in some cases is the scale of individual channel units. The original goal for this study was a minimum of 20-30 w, designated as an appropriate length to capture the range of channel variability. However, this occasionally had to be compro- mised due to terrain and time limitations. Reach lengths in this study range from 10-60 w. Reach measurements included gradient (S ), bankfull width (w), bankfull depth (d), and b-axis diameter of the largest stone on the bed moved by flowing water (to approximate the 95th percentile, D95). Where terrain allowed access to the channel, measurements were taken at 3-5 points at approximately equal intervals along the reach and averaged. Frequently, however, steep and slippery banks prohibited multiple measurements, and in several cases the channel could not be safely entered. Locations for which 17 measurements are less reliable are flagged in Appendix A. Channel gradient was measured to within ± 1% using a hand-held Su- unto clinometer over the longest length of channel possible (greatest distance visible between field workers). Bankfull width was measured to within ± 0.1 m using a fibre tape. Standard criteria were used to identify bankfull width (Leopold, 1994; Province of British Columbia, 1996), including the following: • Change in vegetation (i.e. from no moss to moss-covered ground, or from bare ground to grass-covered ground; Figure 2.2a) • Change in texture of deposited sediment (i.e. from boulders to cobbles, or boulders to pebbles; Figure 2.2b) • Topographic break from steep bank to more gentle slope (occasionally useful, but more difficult to apply due to low, unconsolidated banks; not pictured) Bankfull depth was measured to within ± 10% with a marked snow depth probe (used instead of a stadia rod for portability). Channels were too shallow to sight off the bank with a clinometer, so height of bankfull stage was estimated by extending the fibre tape from the snow depth probe to the bank. When step and riffle forms were present, depth was measured at the step-pool or riffle-pool break. In poorly sorted mixtures, bed roughness is usually expressed by the D84, D90, or D95 of the bed surface material (e.g. Knighton, 1998; Wohl, 2000). The most practical method of approximating D95 is to take the average b- axis of the five largest stones in the active bed, measured at intervals twice the local channel width (Brardinoni and Hassan, 2006). However, difficulty of accessing the stream bed restricted the number of possible measurements to three. Inclusion in the active bed was determined by evidence of regular movement by flowing water (e.g. stones that were not covered in moss or algae and were not isolated boulders). Measurements were to the nearest centimetre. 18 (a) (b) Figure 2.2: Definition of bankfull width by (a) vegetation and (b) sediment texture. Longitudinal profile surveys were conducted using a surveyor’s hip chain and a Garmin GPSmap 60CSx to measure reach lengths and record points of significant change in vegetation and morphological characteristics. Major channel features were recorded and photographed, including tributaries, sed- iment sources, and instream wood. Where bankside vegetation was present, dominant species were identified and any potential geomorphic function of the vegetation was noted (e.g. sediment retention, bank stabilization). 2.2.2 Reach morphology Reach morphology types were designated using criteria from Montgomery and Buffington (1997) and Hassan et al. (2005a). This classification frame- work incorporates the fundamental relations between bed surface roughness, sediment supply, and sediment transport (Table 2.2). Detailed descriptions and photographs of the channel morphologies used in this study are included in Appendix C. 19 Table 2.2: Summary of mountain channel morphologies and characteristics. Modified from Montgomery and Buffington (1997) and Hassan et al. (2005a). Riffle-pool Plane bed Step-pool Cascade Bedrock Colluvial Typical bed material Gravel Gravel-cobble Cobble- boulder Boulder Rock Variable Bedform pat- tern Laterally oscillatory Featureless Vertically oscillatory Random Irregular Variable Dominant roughness ele- ments Bedforms (bars, pools), grains, sinu- osity, banks Grains, banks Bedforms (steps, pools), grains, banks Grains, banks Boundaries (bed & banks) Grains Dominant sed- iment sources Fluvial, bank failure Fluvial, bank failure, debris flows Fluvial, hill- slope, debris flows Fluvial, hill- slope, debris flows Fluvial, hill- slope, debris flows Hillslope, de- bris flows Sediment stor- age elements Overbank, bedforms Overbank Bedforms Lee & stoss sides of flow obstructions Pockets Bed Confinement Unconfined Variable Confined Confined Confined Confined Pool spac- ing (channel widths) 5-7 None 1-4 <1 Variable Unknown Gradient <2% 2-10% 3-30% 4-45% Typically high Typically high Relative roughness <1 0.5-1 ∼ 1 >1 Typically low Typically high 20 2.2.3 Topographic analysis In order to place each channel within its larger valley-scale context, field data were supplemented with GIS-based topographical analysis. Using Ar- cMap 10, catchments and channels were hand-digitized from Digital Eleva- tion Models (DEMs) with resolutions of 25 m (British Columbia) and 10 m (Washington). The expanded GIS dataset was used in conjunction with field survey data to examine field-measured channel characteristics in slope-area space, generate longitudinal profiles and SL/k index curves, and estimate discharge. Although field surveys covered only the uppermost portion of the channels, longitudinal profiles and SL/k index curves were extended down- stream to the nearest major confluence or lake in order to place recently deglaciated reaches within a broader context. Slope and area are indices of two physical controls on the activity of different morphological processes, specifically the availability of potential energy and the magnitude of streamflow. Plots of reach slope versus con- tributing drainage area can be used to delineate process domains (e.g. Mont- gomery and Foufoula-Georgiou, 1993; Tucker and Bras, 1998; Brardinoni and Hassan, 2006). To investigate the effects of glaciation on process domains and channel longitudinal profile, slope and area data were extracted in Ar- cMap 10 following the methods described by Stock and Dietrich (2003). Us- ing ArcMap 10, 25 m elevation contours were generated for each catchment. Slope and drainage area were measured at a point equidistant between con- tours for every contour crossing of the stream. Slope was calculated as the contour interval divided by the stream distance between adjacent contours. In the basic form of longitudinal river profile, the equilibrium (graded) profile is represented by a straight line on a plot of elevation versus logarith- mic distance (Hack, 1957), referred to as the HACK profile. Natural streams frequently deviate from the HACK profile, especially in dynamic landscapes. To capture detail within stream longitudinal profiles, curved HACK profiles are created with elevation and distance measures for successive reaches (e.g. Figure 2.3). Under- and over-steepened reaches are identified by comparing the curved HACK profile to the straight-line graded profile. 21 200 500 1000 2000 0 20 0 40 0 60 0 0 20 40 60 80 10 0 H ei gh t (m ) SL /k  In de x Distance from stream source (m) Figure 2.3: Illustration of a stream profile using data for Boulder Glacier. The actual profile is shown as the solid black line; the dashed line represents a hypothetical graded profile for comparison. The stepped red line shows an SL/k index curve. The SL/k index (where S = reach slope, L = distance from reach to source, and k = graded river gradient) is a way to quantify anomalies in the longitudinal profile. It indicates the deviation of an individual reach from its graded river profile (Hack, 1973). The calculation of k used in this study is a modification by McCleary et al. (2011) of the equation used by Pérez-Peña et al. (2008): k = (hs − hf ) ln(Lt) (2.1) where hs is the river head elevation (m), hf is the river mouth elevation (m), and Lt is the total length of the river (m). In this study, the point of interest is the downstream end of each reach rather than the stream mouth, so the formula for k was adjusted using the McCleary et al. (2011) methodology to compute the SL/k index as: SL/k = Sxrln(xr) (hs − hr) (2.2) where xr is the distance from the stream source to the middle of the reach (m) and hr is the reach mid-point elevation (m). The curved HACK profile 22 is displayed with height on the primary Y axis and a step curve of the SL/k index using a secondary Y axis (e.g. Figure 2.3). The SL/k index is held constant over the length of each reach to create the step-like appearance (Chen et al., 2003). Longitudinal profiles and SL/k index curves for field- surveyed channels are included in Appendix B. 2.2.4 Shear stress, stream power, and discharge calculations Flow resistance was approximated by assigning an estimated Manning’s roughness coefficient (n) to every channel reach. A lower limit for the range of potential n values was established with a modification of the Strickler- Manning equation: ng = 0.04(D50) 1/6 (2.3) for D50 in metres, where ng represents a particle roughness coefficient. Al- though the original Strickler-Manning equation is based on D50, in this analysis D95 was used because the typically large D95 in proglacial streams supplies a substantial portion of the flow resistance. Because roughness can also originate from other channel elements, such as sinuosity, vegetation, and structuring, Manning’s n value was established by visual comparison with representative reaches (Barnes, 1967; Hicks and Mason, 1998). The value of ng from Strickler-Manning was used as a lower limit to check the val- ues assigned by visual comparison. Manning’s coefficient assigned to study reaches ranged across channel types as follows: bedrock 0.03-0.06; colluvial 0.15-0.2; cascade 0.05-0.29; step-pool 0.06-0.09; plane bed 0.05-0.29; riffle- pool 0.04-0.07. Total shear stress (τ0) was calculated at the reach scale using average values of bankfull depth and channel slope in the equation: τ0 = ρgRSf (2.4) where ρ is the water density (1000 g/m3), g is the gravitational acceleration (9.81 m/s2), R is the hydraulic radius (m), and Sf is the hydraulic gradient (for steady flow Sf = S0, the slope of the river bed). 23 Mean water velocity (V ) was calculated using Manning’s equation: V = d2/3S1/2 n (2.5) where d is bankfull depth (m), S is reach gradient (m/m), and n is Manning’s roughness coefficient. Unit stream power (ω) was derived from the product of mean water velocity and total shear stress. Total stream power (Ω) was obtained by multiplying unit power by bankfull width. Shear stress was calculated exclusively from field data, making it more reliable than stream power, which incorporates values of mean velocity cal- culated using Manning’s equation. Manning’s equation is best suited for purely alluvial environments; thus reach values of unit power and total power should be considered approximations. Discharge data were not available for the study basins, so three different approaches were taken to its estimation and results were compared to de- termine the most suitable method. Basic discharge calculations were made using: Q = AV (2.6) where Q is discharge (m3/s), A is cross-sectional channel area (m2), and V is mean velocity (m/s). Areas were taken from field measurements and velocities were derived from Manning’s equation as described above. In a second approach, drainage area was used as a proxy for bankfull discharge. It was assumed that bankfull discharge (Q) varies as a power function of drainage area as: Q = k(Ad) β (2.7) where k is a measure of baseflow (as mean annual flow per unit area), Ad is drainage area (km2), and β is a scaling exponent sensitive to lithology, climate, and land use. There is a lack of regional relations between discharge and drainage area for basins smaller than 50 km2 (Eaton et al., 2002), so k -values were calculated for nearby gauged basins, or larger gauged basins containing the study basin, and scaled to the study basins. The k -factor 24 can be computed for a gauged basin as follows: k = Qpeak Aβd (2.8) where Qpeak is the mean annual peak flow (m 3/s). In British Columbia, β is about 0.75 (Eaton and Moore, 2010). Three types of gauged reference catchment were used to calculate k -factors to apply to the study areas: (1) large catchments containing the study catchment, (2) large catchments nearby the study catchment but not containing it, and (3) small nearby catchments of a similar size to the study catchment. A third discharge estimate was obtained by extracting k -factors from a digital provincial map developed by Eaton et al. (2002). However, these values were not available for the two catchments in Washington. The map- derived k -factors are presented alongside reference catchment-derived k - factors in Appendix D. A comparison of the methods of discharge calculation is given in Table 2.3. Because channels were often difficult to access, there is a considerable level of uncertainty in the channel depth measurements used to calculate velocity. Additionally, visual estimation of Manning’s roughness coefficient is subjective and introduces more uncertainty to the velocity calculation. Given the potential sources of error in Manning’s equation, drainage area- based discharge calculations were assumed to be a more reliable estimate. Reference basin-derived k -factors are likely to be more reliable for some catchments than others, due to the wide variation in period of record in the reference basins (Appendix D). Large discrepancies in record length (e.g. only seven years of data for stations 08HC004 and 12190718) and time period (e.g. spanning only the first half of the 20th century at sta- tion 08MG003) of hydrometric records suggest a potentially large source of uncertainty. Map-derived k -factors were selected as the most reliable ap- proach to estimating discharge in British Columbia catchments. For Easton and Boulder catchments in Washington, the North Fork Nooksack hydro- metric station (12205000) was used to calculate a reference k -factor because of its long period of record (1938-2010). 25 Table 2.3: Comparison of discharge estimates. Figures presented are for the estimated maximum daily discharge within the surveyed portion of the channel. Q values are m3/s. Catchment Q = AV Q from reference basin- derived k -factor Q from map-derived k -factor Easton 27.95 3.61, 7.482, 4.563 * Boulder 32.47 10.672, 14.983, 6.53 * Matier 11.09 2.512, 1.853 2.69 Miller 2.29 4.491 4.66 Wedgemount 14.85 5.081 8.14 Helm 7.68 6.961 12.27 Overlord 3.27 3.641 12.08 Russet Lake 1.79 1.061 3.51 Colonel Foster 7.47 10.782 6.89 Septimus 0.84 3.862 3.07 1,2,3 Denotes reference catchment type * Map-derived k -factor values not available for Washington 2.3 Riparian forest analysis 2.3.1 Data collection In order to evaluate the development of riparian forest in proglacial zones, data were compiled from glacial chronologies established by other researchers, air photographs, satellite images, DEMs, and climatic programs (Appendix E). Selection criteria included established histories of past glacial extents and sufficiently high image resolution to allow interpretation of riparian for- est presence. Sites were selected within the same region as the field surveys, with the sampling area expanded to include locations in the Mount Wadding- ton and Bella Coola areas of British Columbia and on Mount Rainier in Washington (Figure 2.1). A total of 54 reaches in 25 channels were included (Appendix A). Variables of potential influence to riparian vegetation development in- cluded latitude and longitude, glacier area, glacial forefield aspect, reach 26 length, minimum and maximum reach elevation, minimum and maximum reach age, reach gradient, presence of a proglacial lake, channel form, vege- tation class, growing degree-days, average July temperatures, and precipita- tion as snow (the latter three variables given as 1901-2000 normals). Reach delineation was based on general criteria in order to accommodate the rel- atively coarse resolution of remote images. Reach divisions were identi- fied at points where channel pattern, riparian vegetation, or both changed. Moraine ages established by previous researchers and historic ice front posi- tions identified from air photos allowed approximate age dating of reaches. Reach lengths ranged from 0.15 to 2.8 km and reach ages ranged from 0 to 860 years. Climatic data were obtained using ClimateBC 3.21 and ClimateWNA 4.62 (Wang et al., 2006, 2007, 2010). Glacier areas were measured from satellite images dated from 2003-2010 and hand-digitized in ArcMap 10. Reach length and gradient were calculated in ArcMap 10. Reach elevations were taken from DEMs with resolutions of 25 m (British Columbia) and 10 m (Washington). 2.3.2 Channel pattern and vegetation classification The approach to channel pattern classification was adapted from Kellerhals et al. (1976), Schumm (1977), and Church (1983). Channel type categories included braided, wandering, or delta. Braided channels (Figure 2.4a) are bedload dominated and transport limited, with continually shifting bars and thalweg in an unstable channel. Sediment load and clast size are large. Gravel bars and islands form and migrate through the channel (Schumm, 1985). Wandering channels (Figure 2.4b) are transitional between braided and meandering. They exhibit channel division that is less continuous and less intense than in braided rivers and irregular sinuosity that is generally lower than in meandering rivers (Church, 1983). Delta channels (Figure 2.4c) are wide sediment fans with multiple shifting channels flowing into lakes. Riparian forest classification was based on an adaptation of the approach 27 proposed by Kellerhals et al. (1976), which was developed for high resolution air photo interpretation and could not be directly applied to the relatively coarse resolution satellite images. However, satellite image resolution was sufficient to identify forest in three broad categories, defined using Kellerhals et al. (1976) as a guide: (1) nonfunctional (scattered or no apparent vegeta- tion, Figure 2.5a), (2) possibly functional (present in valley but of uncertain morphological function in the channel, Figure 2.5b), and (3) functional (for- est in immediate proximity to the channel, Figure 2.5c). Resolution was not fine enough to identify species on any images. 28 (a) (b) (c) Figure 2.4: Examples of channel types classified from satellite im- ages. (a) Braided, (b) wandering, (c) delta. (a) (b) (c) Figure 2.5: Examples of riparian vegetation classified from satellite images. (a) Nonfunctional, (b) possibly functional, (c) func- tional. 29 2.3.3 Data analysis The goal of analysis was to identify the relative strength of relations among possible controls on riparian forest development. Analysis was completed using the statistics program R 2.10.1 (R Development Core Team, 2011). Descriptive statistical assessments and distribution plots identified possible relations between vegetation class and independent variables. A Kruskal- Wallis nonparametric ANOVA test was used to examine potential correla- tions between variables and vegetation class. The Kruskal-Wallis test eval- uates whether independent population distributions are identical without assuming normality (Neter et al., 1996). In order to improve the strength of the regression and compensate for the uncertainty associated with remote image interpretation, the somewhat subjective “nonfunctional forest” category was reclassified for two separate regression analyses. One binary reclassification grouped vegetation into ab- sence or presence categories, with vegetation presence including any stream bank vegetation regardless of its apparent functional role in shaping the channel. The other binary reclassification grouped vegetation into “nonfunc- tional” (including reaches with no vegetation, reaches containing vegetation of uncertain maturity, and reaches where mature forest was present on the valley bottom but not near the channel) and “functional” (mature forest in direct proximity to the channel) categories. The influence of numerical predictor variables (reach-scale variables of age, elevation, length, gradient, mean July temperature, growing degree- days, and precipitation as snow, and the catchment-scale variable glacier area) was evaluated using backward stepwise regression. All potential nu- merical variables were included in the initial model. Variable retention in the model was based on the significance of the F statistic from analysis of covari- ance between the groups, with the variables remaining in the model serving as covariates. Variables with p > 0.05 were dropped from the model; only those contributing to the explanatory power of the model were retained. An automatic search was performed on the full models to detect any significant 30 two-factor interactions. The general model form is y = 1 1 + e−z (2.9) where z is defined as z = b0 + n∑ i=1 bixi (2.10) and n is the number of predictor variables included in the specific model. The ClimateBC and ClimateWNA programs use latitude-longitude co- ordinates and elevation to interpolate climate. Therefore, temperature vari- ables are directly correlated with elevation and reflect any biases and uncer- tainties present in the climate program algorithms. Due to the correlation between elevation and climate predictor variables, two versions of the lo- gistic regression were conducted on the riparian forest establishment data set: one with elevation, and one without elevation. The regression out- comes were then compared using leave-one-out cross-validation (LOOCV), selected because of its suitability for small datasets. In LOOCV, a single ob- servation from the original sample is used for validation, and the remaining observations are used as training data. The procedure is repeated for each observation in the sample. The true model classification error is estimated as the average error rate on test examples. The model that minimized clas- sification error rate was selected as the best in each of the two regression analyses. The relations between categorical variables (reach aspect, proglacial lake presence, and channel form) and vegetation class were evaluated using Fisher’s exact test (Agresti, 1992). Fisher’s test was selected because it is particu- larly well-suited to a small sample size. 31 2.4 Fish habitat analysis 2.4.1 Habitat characterization To determine the potential for fish habitat in recently deglaciated streams, habitat was characterized in 17 catchments using map-derived data (Ap- pendix A). Gradient and downstream barriers were used as primary predic- tors of potential fish absence or presence. These two variables were selected on the basis of BC government guidelines (Province of British Columbia, 1998; Norris and Mount, 2009; Mount et al., 2011) and previous field- and map-based studies indicating that gradient and downstream barriers are first-order indicators of fish habitat (Kruse et al., 1997; Latterell et al., 2003; McCleary and Hassan, 2008). Site-specific environmental attributes can be used in conjunction with first-order predictors of fish presence (gradient and downstream barriers) to provide an indication of fish abundance and productivity. Water tem- perature is a particularly important habitat variable because fish distribu- tion and productivity are often strongly correlated with stream temperature (e.g. Eaton and Scheller, 1996; US Environmental Protection Agency, 2003). Daily, seasonal, and interannual temperature fluctuations give rise to many possible indicators of stream thermal regime. Maximum weekly average temperature (MWAT) was selected as an index for this study because it is commonly used to set stream temperature standards (e.g. State of Washing- ton, 2003; State of Oregon, 2004) and because it can be easily and reliably computed from air temperature (Nelitz et al., 2008). Additionally, previous work has shown MWAT to be closely related to biological processes of rain- bow trout Oncorhynchus mykiss (Nelitz et al., 2007), the most commonly observed fish species in the study area. 2.4.2 Data collection Data for fish observations and obstacles come from a wide variety of sources (Table 2.4). Although these sources are the best available for the study area, measurement uncertainty imposes some constraints on data reliability. 32 Table 2.4: Stream attributes, data resolution, and data sources used to identify potential fish-bearing streams and to calculate water temperature. Attribute Resolution Data Source Fish obstacles and observations Varies Obstacles to Fish Passage & Fish Observations Datasets (BC) SalmonScape Program (WA) Stream networks 1:20,000 (BC) 25 m DEMs (BC) 1:24,000 (WA) 10 m DEMs (WA) Drainage area (A) 1:20,000 (BC) 25 m DEMs (BC) 1:20,000 (WA) 10 m DEMs (WA) Fractional glacier (fg) and lake (fl) coverage 1:20,000 (BC) Freshwater Atlas (BC) 1:24,000 (WA) USGS National Hydrography Dataset (WA) Stream slope (S) 1:20,000 (BC) 25 m DEMs (BC) 1:24,000 (WA) 10 m DEMs (WA) Mean basin elevation (Zm) 1:20,000 (BC) 25 m DEMs (BC) 1:24,000 10 m DEMs (WA) 2-year flood frequency parameter (K2) Scale-independent Provincial K2 layer (Eaton et al., 2002) (BC) k = Qpeak/(Ad) 0.75 (WA) Average July-August air temperatures (Ta) 2.5 x 2.5 arcmin ClimateBC version 3.21 ClimateWNA version 4.6 Stream networks were hand-digitized from 25 m-interval contour maps created from 25 m (BC) and 10 m (WA) DEMs. Network paths were checked by the construction of a flow accumulation layer in ArcMap 10 and, in BC, the overlay of the Freshwater Atlas dataset. Because many of the typical characteristics that define uniform reaches (e.g. width, morphology, channel unit aggregations) cannot be derived from DEMs, segments were generated based strictly on channel gradient. Therefore, they are referred to as ‘stream gradient segments’ rather than reaches. Gradient- and barrier-based habitat characterization followed BC Min- istry of Environment methodology (Norris and Mount, 2009; Mount et al., 2011). Stream channels were split at each intersection with a 25 m contour line to create a segmented stream line (Figure 2.6). Stream gradient segment end points were defined at significant gradient changes, locations where the 33 Figure 2.6: Example of stream segment delineation based on chan- nel gradient. difference in the length of two adjacent stream segments was greater than the standard deviation of the lengths of all the segments making up the stream. The gradient of each stream segment was then calculated as the elevation difference divided by the stream line length. Evaluation began at the uppermost known fish observation on a stream and proceeded upstream. Generally, gradients ≥20% are barriers to fish. However, this threshold is species specific. Cutthroat trout, bull trout, Dolly Varden char, and rainbow trout have been observed in reaches with >20% gradient, especially in step-pool morphologies or channels with a lake at the head of the drainage (Watson and Hillman, 1997; Province of British Columbia, 1998; Latterell et al., 2003). For this study, fish barriers were defined as gradients >25% or falls >5 m high, in accordance with provincial fish-stream identification 34 guidelines (Province of British Columbia, 1998; Norris and Mount, 2009; Mount et al., 2011). MWAT values were estimated for each gradient segment using the em- pirical model developed by Nelitz et al. (2008): MWAT = 7.996 + (0.5083 Ta) + (1.016 log A) − (0.003192 Zm)− (16.19 fg) + (17.39 √ fl)− (0.05882 S)− (0.7788 K2) (2.11) where Ta is mean July-August air temperature ( ◦C), A is drainage area (km2), Zm is mean basin elevation (m), fg is fractional glacier coverage, fl is fractional lake coverage, S is gradient segment slope (m/m), and K2 is the 2-year flood frequency parameter. 2.4.3 Modelling approach Fish presence is often modelled by logistic regression, but regression was not used for this study because it could not be tested against actual data. Clas- sification tree analysis was selected as a more suitable model given the data resolution and reliability. Previous studies have indicated that classification trees are a robust approach to modelling fish presence (Olden and Jackson, 2002; Rosenfeld, 2003). Stream segments were divided into four classes on the basis of gradient and downstream barriers: • 1a definite potential for fish habitat (local gradient ≤25%, no down- stream barriers to fish movement) • 1b possible fish habitat if an adjacent lake is stocked (local gradient ≤25% but fish migration is prevented by downstream barrier) • 0a no fish habitat (local gradient >25%) • 0b no fish habitat (migration prevented by downstream barrier, no adjacent lake) 35 The results of the gradient- and barrier-based analysis were organized into a classification tree, and MWAT values were added to the assessment to predict fish communities within the accessible channels. Salmonids (trout and salmon) are the dominant cold water species in northern or high elevation aquatic systems throughout North America. Wa- ter temperatures ≤20◦C are physiologically optimal for cold water species, and they are generally not found where summer water temperatures ex- ceed 20-24◦C (Eaton et al., 1995). Although there is minimal risk of direct temperature mortality while the annual maximum temperature is less than 26◦C, prolonged exposure to unfavourably high or low temperatures can significantly impact fish health and growth (e.g. Sullivan et al., 2000). Specific temperature optima are species-dependent. Therefore, rather than establishing temperature ranges for a specific species, this study used the temperature parameters established by Nelitz et al. (2008) to delineate community thresholds: • MWAT <12.5◦C favourable for cold water species such as bull trout (Salvelinus confluentus) • MWAT 12.5-20◦C favourable for cool water species such as rainbow trout (O. mykiss) • MWAT >20◦C favourable for warm water species such as minnows and suckers MWAT values were estimated for each stream gradient segment under two scenarios: (1) present basin ice coverage and (2) future ice-free condi- tions. MWAT was calculated without the influence of basin ice coverage by setting fractional glacier coverage (fg) to 0. 36 Chapter 3 Results The chapter begins with a characterization of reach-scale channel morphol- ogy using results from field surveys and a comparison to other studies of mountain channel morphology in the Pacific Northwest (Section 3.1). Stud- ies used for comparison include channels that have never been subject to glaciation (Montgomery and Buffington, 1997) and channels that underwent Pleistocene and early-Holocene glaciation, but remained ice-free during the neoglacial Little Ice Age (LIA) period (Brardinoni and Hassan, 2006, 2007). Models of riparian vegetation establishment are presented in Section 3.2, and an evaluation of fish habitat potential is presented in Section 3.3. 3.1 Channel morphology 3.1.1 Effects of slope and contributing area In order to examine the degree of discrimination between channel types in slope-area space, all study reaches are plotted together in Figure 3.1. The slope-area plot exhibits considerable scatter, but a distinct slope-induced division is evident between cascade and plane bed reaches at slopes of 7- 8%. Step-pool, cascade, and colluvial reaches plot at the highest slopes, and riffle-pool and plane bed reaches plot at the lowest slopes, regardless of drainage basin size. Colluvial and fluvial process domains are not clearly 37 differentiated by the slope-area plot. Cascade reaches, characterized by a transitional semifluvial domain, plot within the same slope range as colluvial and high-gradient fluvial (step-pool) reaches. However, the results from this study must be interpreted with caution, as there are minimal data from basins with areas <1 km2. The ten reaches from Russet Lake, which remained unglaciated during the LIA, plot within the same range as recently glaciated reaches (Figure 3.1, denoted in red). The similarity of the Russet Lake reaches to those exposed post-LIA suggests that the Russet Lake channel still reflects the long-term effects of the Pleistocene glaciation. Because of this degree of similarity, the Russet Lake data are included in this analysis and are not assigned unique symbology in the other plots in this section. Potential limitations of the data arise from the narrow range of drainage basin areas (10-1 to 101 km2) and the low number of observations of col- luvial, step-pool, riffle-pool, and bedrock morphologies. Interpretation is for the most part restricted to comparisons between cascade and plane bed morphologies, the most commonly observed. The slope-discharge plot (Figure 3.2) closely resembles the slope-area plot, an expected result due to the nearly linear relation between discharge and area in small basins and the use of area in discharge calculations. Drainage area is used as a proxy for discharge in the rest of this analysis due to the higher degree of error inherent in calculating discharge estimates. 38 ll ll l l l l l l l l l l l l l l Drainage Area (km2) Sl op e (m /m ) 0.1 1 10 0. 01 0. 1 1 l bedrock colluvial riffle−pool plane bed step−pool cascade Figure 3.1: Reach slope versus contributing drainage area for each field-surveyed site. Markers indicate channel types. Unglaciated reaches from Russet Lake are in red. 39 ll ll l l l l l l l l l l l l Discharge (m3/s) Sl op e (m /m ) 1 10 100 0. 01 0. 1 1 l bedrock colluvial riffle−pool plane bed step−pool cascade Figure 3.2: Reach slope versus discharge for each field-surveyed site. 40 To examine trends in morphology with respect to drainage basin size, bankfull depth (d), bankfull width (w), and coarse grain fraction (D95) were plotted against area (Figure 3.3). Lines indicating the empirical downstream hydraulic geometry (DHG) relations between channel dimensions and dis- charge (here, drainage area is used as a surrogate for discharge) have been added to plots of depth (Figure 3.3a) and width (Figure 3.3b). No differen- tiation of reach morphologies is evident in the area-based plots, which could be due to the narrow range of data. Slight increases in depth with increasing area are apparent in both cascade and plane bed morphologies, but there is no separation between cascade and plane bed across spatial scales (Figure 3.3a). The depth-area relation appears to roughly parallel the empirical rela- tion d ∝ Q0.4, but when the four reaches with drainage areas <1 km2 are excluded, the depth-area relation is close to vertical. Width appears to be relatively unchanged with increasing area (Figure 3.3b). The trend of the width-area relation is nearly horizontal when the four reaches <1km2 are in- cluded, but when these reaches are removed the relation appears to parallel the w ∝ Q0.5 empirical DHG line. Although the area-D95 plot does not discriminate between reach mor- phologies, D95 for both colluvial and alluvial domains shows an increasing trend and wider range at larger contributing areas (Figure 3.3c). A similar lack of spatial dependence within the same range of areas has been reported elsewhere for landscapes last glaciated in the Pleistocene (Brardinoni and Hassan, 2007). The absence of clear downstream trends observed here may be a con- sequence of the narrow range of drainage areas. Generally, lack of spatial correlation also suggests poorly developed DHG, but no conclusions can reasonably be drawn about DHG from this data given the limited spatial range. To compare these results with previous studies of channel-reach mor- phology (Montgomery and Buffington, 1997; Brardinoni and Hassan, 2007), reach slope distributions categorized by morphology types were plotted (Fig- ure 3.4). Glaciated reaches (red and yellow) plot at consistently higher 41 ll l l l l l l l l l l ll lll l l Drainage Area (km2) D ep th  (m ) d ∝ Q 0.4(a) 0.1 1 10 0. 1 1 l l l l l l l ll l l l l ll l l l l Drainage Area (km2) W id th  (m ) w ∝ Q 0.5 (b) 0.1 1 10 1 10 10 0 l l ll l ll l l l l ll l l l l Drainage Area (km2) D 95  (cm ) (c) 0.1 1 10 1 10 10 0 l bedrock colluvial riffle−pool plane bed step−pool cascade Figure 3.3: (a) Bankfull depth, (b) bankfull width, and (c) coarse grain size fraction, plotted as functions of drainage area. Lines indicating empirical downstream hydraulic geometry relations are added to (a) and (b). 42 0. 0 0. 1 0. 2 0. 3 0. 4 Sl op e (m /m ) −− riffle−pool −− −− plane bed −− −− step−pool −− −− cascade −− Montgomery and Buffington (1997) Field Surveys Brardinoni and Hassan (2007) Figure 3.4: Boxplot of reach slopes categorized by morphology type. Blue indicates data from unglaciated reaches (Montgomery and Buffington, 1997), red indicates data from this study, and yellow indicates data from Pleistocene-glaciated reaches (Brardinoni and Hassan, 2007). slopes than unglaciated reaches (blue) for all morphologies. The similarity of slopes for glaciated reaches regardless of time since glaciation suggests long-term topographic effects of glaciation. The high degree of overlap be- tween glaciated step-pool and cascade reaches reflects the transitional nature of cascade morphologies, which occupy a semifluvial domain and are gov- erned by mass movement and fluvial processes (Hassan et al., 2005a). 3.1.2 Shear stress and substrate resistance To investigate the interactions between driving force and substrate resis- tance, total shear stress (τ) and total stream power (Ω) were examined relative to D95, reach slope, and contributing area. Stream power-based plots were similar to shear stress-based plots, an expected result given the similar variables used to calculate both. Shear stress was selected as the primary expression of driving force because in this study it is a more re- 43 liable estimate than stream power. Although the shear stress calculations used in this study contain uncertainty from field measurements of slope and depth, stream power calculations have additional uncertainty introduced by discharge estimation. Stream power-based plots are included in Appendix F but are not discussed here. Figure 3.5 presents plots of shear stress as a function of slope, drainage area, and D95. Shear stress is most closely correlated with slope (Figure 3.5a), because shear stress is calculated directly from slope. The slope-shear stress relation differentiates well between cascade and plane bed morpholo- gies, with the division at slopes of 7-8% and shear stresses of 200-250 N/m2. Shear stress does not show any clear trends across spatial scales (Figure 3.5b). Fluvial reaches plot at lower shear stress than colluvial and semiflu- vial (cascade) reaches regardless of contributing area. The range of shear stress values is generally similar to that observed across the same spatial ex- tent for previous studies in unglaciated (blue) and Pleistocene-glaciated (yel- low) landscapes (Montgomery and Buffington, 1997; Brardinoni and Hassan, 2007). The wide range of D95 values at higher shear stresses (Figure 3.5c) is most likely inherited from glacial deposits. The range of particle sizes is reflected in the τ/D95-area plot (Figure 3.6). The steady increase in D95 with increasing shear stress observed here contrasts with results from basins glaciated during the Pleistocene, which exhibit a slight decrease in particle size within a similar shear stress range (Brardinoni and Hassan, 2007). 44 ll l l l l l l l l l l l ll l l Slope (m/m) Sh ea r S tre ss  (N /m 2 ) (a) 0.01 0.1 1 10 0 10 1 10 2 10 3 10 4 l l l l l l l l l l l l l l l l l l Drainage Area (km2) Sh ea r S tre ss  (N /m 2 ) (b) 0.1 1 10 10 0 10 1 10 2 10 3 10 4 l l l l l l l l l l l l l l l l l l D95 (cm) Sh ea r S tre ss  (N /m 2 ) (c) 1 10 100 10 0 10 1 10 2 10 3 10 4 l bedrock colluvial riffle−pool plane bed step−pool cascade Figure 3.5: Reach morphology plotted by bankfull shear stress as a function of (a) slope, (b) drainage area, and (c) D95. Lines in (b) denote shear stress ranges from previous studies in unglaciated (blue; Montgomery and Buffington, 1997) and glaciated (yellow; Brardinoni and Hassan, 2007) landscapes. 45 ll l l ll l l ll l l lll l l l l l Drainage Area (km2) τ D 95 0.1 1 10 10 2 10 3 10 4 10 5 l bedrock colluvial riffle−pool plane bed step−pool cascade Figure 3.6: Reach morphology plotted by τ/D95 as a function of drainage area. 3.1.3 Channel size and grain roughness characteristics To examine potential reach-scale controls on channel morphology, the vari- ables bankfull depth (d), bankfull width (w), coarse grain fraction (D95), relative roughness (D95/d) and width-to-depth ratio (w/d) were evaluated against slope (Figure 3.7). The slope-induced division between cascade and plane bed reaches is evident in all slope-based plots, but other reach mor- phology types are poorly differentiated, possibly due to the limited number of observations. Width-to-depth ratio shows a decrease with increasing slope (Figure 3.7a), driven by slightly increasing depth (Figure 3.7b) rather than width, which shows no clear trend in regard to slope (Figure 3.7c). The decreasing w/d trend mirrors that previously reported for Pleistocene-glaciated reaches (Brardinoni and Hassan, 2007), but with a causal difference - the w/d de- crease observed in Pleistocene-glaciated reaches is primarily driven by de- 46 creasing width, reflective of a transition from partially coupled to fully cou- pled hillslopes at higher channel slopes. In this study, evidence of hillslope- channel coupling was rarely observed because recent glaciation has carved valley flats that are large relative to their occupying channels. Potential or partial coupling was only observed at a few sites, generally where channels incised LIA end moraines. A notable example is the Colonel Foster Glacier channel, which is deeply incised into a series of nested LIA end moraines (photographs of the Colonel Foster channel are included in Appendix B). Cascade reaches are characterized by a wide range of D95 sizes (Figure 3.7d), a relic of glacial deposits. The steady increase in D95 with slope contrasts with results for Pleistocene-glaciated reaches, which show an ini- tial D95 increase with a peak at about 7% slope, then a gradual decrease (Brardinoni and Hassan, 2007). The D95 values in this study encompass a similar range as reported for unglaciated reaches (Montgomery and Buffin- gton, 1997), but are an order of magnitude lower than values observed in Pleistocene-glaciated reaches (Brardinoni and Hassan, 2007). The positive trend and wide range of D95 in cascade reaches is reflected in the slope-relative roughness relation (Figure 3.7e). Relative roughness values for this study range from 0.1-2, lower than the range of approximately 1-2 reported for other glaciated reaches (Brardinoni and Hassan, 2007). 47 ll l l l l l l l l l ll l l l l l Slope (m/m) w /d  (m /m ) (a) 0.01 0.1 1 10 0 10 1 10 2 10 3 l l l l l l l l l l l l ll l l l Slope (m/m) D ep th  (m ) (b) 0.01 0.1 1 0. 1 1 l l l l l l l ll l l l l l l l l l l Slope (m/m) W id th  (m ) (c) 0.01 0.1 1 1 10 10 0 l l l l l l l l l l l l l l l ll Slope (m/m) D 95  (cm ) (d) 0.01 0.1 1 1 10 10 0 l l l l ll l l l l l ll l l l ll Slope (m/m) D 95 d (m m ) (e) 0.01 0.1 1 0. 1 1 10 l bedrock colluvial riffle−pool plane bed step−pool cascade Figure 3.7: Channel types plotted by slope versus (a) bankfull width-to-depth ratio, (b) bankfull depth, (c) bankfull width, (d) D95, and (e) relative roughness. 48 3.2 Riparian vegetation analysis 3.2.1 Correlations among numerical variables Although there is considerable overlap, boxplots indicate a positive corre- lation between riparian forest development and reach age and a negative correlation between forest development and reach elevation (Figure 3.8). Results of a Kruskal-Wallis nonparametric ANOVA test confirm that forest type distributions are nonidentical as a function of reach age and elevation (p-values <0.001). Significant Kruskal-Wallis results also indicated trends between riparian forest and climatic variables (not pictured). Forest devel- opment was positively associated with mean July temperature (p-value = 0.02) and growing degree-days (p-value = 0.02), and negatively associated with precipitation as snow (p-value = 0.01). Results of the logistic regression using absence or presence of vegetation (the binary response “absent” or “present”) are summarized in Table 3.1 and Figure 3.9. Leave-one-out cross-validation (LOOCV) resulted in a mean squared error of prediction of 0.138, indicating that this model misclassified vegetation absence or presence in approximately 13.8% of the dataset. Al- though the reach elevation is not significant alone at α = 0.05, it is included because of the interaction effect between reach elevation and age. Due to the correlation between climate variables and elevation, the back- wards stepwise regression was repeated starting with all variables except ele- vation. This yielded a model with significant effects from reach age, growing degree-days, and mean July temperature, and an interaction effect between reach age and growing degree-days. However, LOOCV gave a generalization error of 0.162, higher than the generalization error for the original model, so the elevation model was selected as the best fit. The results of the logistic regression using absence or presence of mature forest in direct proximity to the channel (the binary response “nonfunc- tional” or “functional”) are summarized in Table 3.2 and Figure 3.10. The generalization error is 0.108. A second backwards stepwise regression beginning with all variables ex- 49 ll l 50 10 0 20 0 50 0 Riparian Forest M ax im u m  R ea ch  A ge  (y e a rs ) None (21) Nonfunctional (21) Functional (12) (a) Riparian Forest M ax im u m  R ea ch  E le va tio n (m  as l) None (21) Nonfunctional (21) Functional (12) 10 00 15 00 20 00 (b) Figure 3.8: Boxplots of riparian forest class versus (a) maximum reach age and (b) maximum reach elevation. Numbers in brackets are sample sizes. 50 Table 3.1: Coefficients and p-values for binary logistic regression model to predict the presence of riparian vegetation. Full model contained all variables including elevation. Variable Coefficient p-value Intercept -6.246 0.4291 Maximum reach age 0.1605 0.0298 Maximum reach elevation 0.002371 0.609 Age x elevation -0.00008672 0.0361 0 0 1 0 1 1 11 1 1 1 0 1 1 0 1 1 0 100 0 1 1 1 1 0 1 1 1 0 1 1 1 0 1 1 1 1 0 0 0 0 1 0 1 0 0 1 0 1 1 1 50 100 200 500 Maximum Reach Age (years) M ax im u m  R ea ch  E le va tio n (m  as l) 10 00 15 00 20 00 0 1 Absent Present Figure 3.9: Riparian vegetation absence (red) or presence (blue) plotted as a function of maximum reach elevation and maxi- mum reach age. 51 Table 3.2: Coefficients and p-values for binary logistic regression model to predict the presence of riparian forest. Full model contained all variables including elevation. Variable Coefficient p-value Intercept 5.4767 0.03198 Maximum reach elevation -0.006173 0.0032 Maximum reach age 0.006686 0.00672 0 0 0 0 0 1 01 1 1 1 0 0 1 0 0 1 0 100 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 50 100 200 500 Maximum Reach Age (years) M ax im u m  R ea ch  E le va tio n (m  as l) 10 00 15 00 20 00 0 1 Nonfunctional Functional Figure 3.10: Riparian forest absence (red) or presence (blue) plot- ted as a function of maximum reach elevation and maximum reach age. 52 cept elevation indicated significant effects from mean July temperature and reach age. LOOCV for this model resulted in a generalization error of 0.119, slightly higher than the error for the model including elevation, so the orig- inal model was selected as the best. 3.2.2 Correlations among categorical variables Among categorical variables, lake presence and reach aspect are not sig- nificantly correlated with higher stages of vegetation development (Table 3.3). However, more stable channel form (wandering rather than braided) is correlated with more extensive stages of vegetation development. This re- lationship is most likely a reflection of the control of vegetation on channel form, rather than channel form as a control on vegetation. Table 3.3: Fisher’s exact test p-values for the relationship between stages of riparian vegetation and categorical variables. Proglacial lake Channel form Aspect Binary Multinomial Multinomial (yes or no) (braided, wan- dering, or delta) (N, S, E, W) Classification 1 0.70 0.0005* 0.56 Multinomial (none, possibly functional, or functional) Classification 2 0.46 0.0014* 0.60 Binomial (vegetation absence or pres- ence) Classification 3 1 0.0006* 0.73 Binomial (forest absence or presence) * Statistically significant (α = 0.05) 53 3.3 Fish habitat analysis 3.3.1 Gradient-based assessment A total of 92 stream gradient segments in 17 catchments were included in the fish habitat assessment. The results are summarized in Table 3.4, grouped by stream and gradient class for a more concise presentation. Of all stream gradient segments included in this study, 80% of the to- tal length is of favourable gradient for fish (≤25%) and physically could permit upstream fish migration (Figure 3.11). Although 14% of the total stream gradient segment length is unavailable to fish due to downstream barriers, a portion of these streams (4% of the total length) could support fish populations if adjacent lakes are stocked. Only 16% of the total length is completely unavailable to fish. In recently deglaciated (post-LIA) streams, 75% of the total length is presently physically available to fish (Figure 3.11). An additional 15% could support fish with lake stocking. Within these new streams, 3% of the length is inaccessible to fish due to downstream barriers (and no adjacent lakes), and 7% is too steep to support fish. Rainbow trout (Oncorhynchus mykiss) and Dolly Varden char (Salveli- nus malma malma) were the most common fish sightings (Table 3.4). Sul- phur Creek and Boulder Creek, the two channels fed by Easton and Boulder Glaciers on Mount Baker, are unusual among the channels in this study because they are accessible by a diverse population of fish from Baker Lake, which supports coho (O. kisutch) and kokanee (O. nerka) salmon, rainbow, bull (S. confluentus), and cutthroat (O. clarkii) trout, and Dolly Varden char. Cutthroat trout were also reported in the Elk River below Colonel Foster Glacier, and coho salmon were reported in the Homathko River at its confluence with an unnamed stream from Cathedral Glacier. However, fish movement into the Cathedral Glacier channel is impeded by a sustained 39% gradient immediately upstream of its convergence with the Homathko River. 54 Table 3.4: Summary of results from fish habitat analysis for channels deglaciated post-LIA. Gradient seg- ments are listed beginning at the LIA moraine and proceeding upstream. Fish observations refer to species noted downstream of surveyed segments in previous inventories. Catchment Stream Segment length (m) Gradient class MWAT with ice (◦C) MWAT without ice (◦C) Nearest fish observation Anniversary Unnamed 974 1a 8.62 9.83 Rainbow Black Tusk Unnamed 1510 1a 10.64 10.92 Rainbow Boulder Boulder Creek 1627 1a 11.19 14.50 Coho, Steelhead, Rainbow, Dolly Varden, Cutthroat, Bull, Kokanee Cathedral Unnamed 294 0b 5.30 7.34 Coho, Dolly Varden Cathedral Unnamed 474 0a 3.82 7.00 Coho, Dolly Varden Cathedral Unnamed 286 0b 2.22 6.45 Coho, Dolly Varden Cathedral Unnamed 495 0a 0.02 5.90 Coho, Dolly Varden Colonel Foster Unnamed 1374 1a 12.23 13.77 Cutthroat, Dolly Varden Easton Unnamed 2167 1a 0 8.19 Coho, Steelhead, Rainbow, Dolly Varden, Cutthroat, Bull, Kokanee Fleur des Neiges Unnamed 450 0a 9.51 14.47 Rainbow Helm Helm Creek 848 1a 4.78 10.57 Rainbow Helm Unnamed 154 0a 0.09 8.48 Rainbow 55 Catchment Stream Segment length (m) Gradient class MWAT with ice (◦C) MWAT without ice (◦C) Nearest fish observation Matier Joffre Creek 900 0a 0 7.10 Rainbow Miller Unnamed 1455 1a 5.38 10.36 Rainbow Moving Unnamed 167 0a 1.93 7.26 Rainbow Overlord Unnamed 1943 1a 4.95 10.14 Rainbow, Dolly Varden Razor Creek Unnamed 2518 1a 2.62 5.43 Dolly Varden Septimus Unnamed 1302 1b 7.85 9.51 Rainbow Staircase Ayceecee Creek 812 1a 0.74 8.27 Rainbow Warren Culliton Creek 2023 1b 8.07 10.95 Rainbow Wedgemount Wedgemount Creek 1687 1b 9.51 12.28 Rainbow Wedgemount Wedgemount Creek 332 0a 4.09 8.46 Rainbow 56 Local gradient supports fish? Yes No Class 0a 6% total, 7% LIA No presence due to high gradient Downstream barrier? Yes No Class 1a 80% total, 75% LIA Highly possible presence; favourable gradient Lake upstream of barrier? Yes No Class 1b 4% total, 15% LIA Potential presence if adjacent lake is stocked Class 0b 10% total, 3% LIA No presence; downstream barrier prevents movement Figure 3.11: Classification tree model of fish presence or absence in proglacial streams, based on stream gradient segment GIS evaluation. ‘LIA’ refers to stream segments deglaciated post-LIA. 57 3.3.2 Temperature-based assessment A summary of maximum weekly average temperatures (MWAT) calculated from the regression model developed by Nelitz et al. (2008) is reported in Table 3.5. Under present ice conditions, the majority of channel length is within temperature class T1 (<12.5◦C), favourable to cold water species. Only 3% of channel length exposed post-LIA is presently within class T2 (12.5-20◦C), indicating that thermal barriers may prevent upstream migra- tion into proglacial streams by cool water species such as rainbow trout. However, post-LIA channels are thermally ideal for cold water species such as bull trout. In order to predict future water temperature increases due to glacier loss, MWAT was calculated without basin ice coverage by setting fractional glacier coverage (fg) to 0. This increased the water temperature estimates for all streams, suggesting a potential species shift in many streams and an overall decrease in suitable habitat for cold water species. The impact of glacier loss on water temperature is likely to be much more drastic in headwaters with small contributing areas (Figure 3.12). Water temperatures are not projected to increase as much in downstream reaches with large contributing areas, as they have less glacier cover at present. Table 3.5: Summary of projected MWAT values for basins with and without ice coverage, and percentage of stream segment length within each temperature class. Min Max Mean T1 T2 T3 <12.5◦C 12.5-20◦C >20◦C Total 0◦C 15.2◦C 8.9◦C 79% 21% 0% with ice Total 5.4◦C 15.7◦C 11.8◦C 45% 55% 0% without ice Post-LIA 0◦C 13.6◦C 5.0◦C 97% 3% 0% with ice Post-LIA 5.4◦C 14.9◦C 9.5◦C 66% 34% 0% without ice 58 ll l l l l l l l l l ll l l l l ll ll l l l llll l l l l l l l l l l l l l l l l l l ll l l l l ll l l l l l l l ll ll l ll l l 0 5 10 15 20 Drainage Area (km2) M W AT  (° C) 10−2 10−1 100 101 102 103 (a)l l 1a 1b 0a 0b l l l l lll l l l ll l l l l ll ll l l l llll l l l l l l l l ll l l l l l l l ll ll l l l lllll ll l ll ll l lll l 0 5 10 15 20 Drainage Area (km2) M W AT  (° C) 10−2 10−1 100 101 102 103 (b)l l 1a 1b 0a 0b Figure 3.12: Stream gradient segment maximum weekly average temperatures (MWAT) versus drainage area for basins under current ice conditions (a) and future projection without ice (b). 59 Chapter 4 Discussion This study has identified channel morphology, riparian vegetation, and fish habitat development as key elements of proglacial stream evolution in coastal British Columbia and Washington. This chapter first discusses channel mor- phology, which serves as the physical template for channel change (section 4.1). Next described are the establishment of riparian vegetation (section 4.2) and potential for fish habitat (section 4.3) in recently deglaciated chan- nels. Finally, the suitability of previous conceptual models of proglacial stream development is evaluated in the context of the Coast Mountains and North Cascades (section 4.4). 4.1 Channel morphology Quaternary glaciation in British Columbia and Washington State has left a glacial landscape legacy that appears to override contemporary landscape changes. Similarity between reach morphologies observed in this study and previous work by Brardinoni and Hassan (2006, 2007) suggests that effects of the Little Ice Age (LIA) glaciation are superimposed on a larger landscape template carved by Pleistocene ice. The resemblance of the Russet Lake reaches (unglaciated since the Pleistocene) to reaches exposed post-LIA re- inforces the significance of underlying topography in dictating present-day reach morphology. Most of the sites included in this study are experiencing 60 increasing sediment yield after recent deglaciation, and are likely at or near the peak of the sediment yield curve proposed by Church and Slaymaker (1989). However, the Church and Slaymaker (1989) model does not include a specific timescale, making it difficult to know how close the post-LIA sites are to peak sediment yield. Results from Brardinoni and Hassan (2006) in- dicate that Coast Mountain landscapes have not recovered from glaciation even after 14,000 years, implying that channels exposed since the end of the LIA have had little time to develop structured morphology. Patterns of channel morphology reflect the complex geomorphic signa- ture imposed by glaciation. Inherited Quaternary geomorphology introduces a pattern of steep, rocky valley steps juxtaposed with relatively wide, low- gradient hanging valleys. This stepped topography dictates local channel gradient, which in turn controls reach-scale shear stress, stream power, chan- nel morphology, and habitat characteristics. Transport capacity decreases downvalley, then resets in supply limited conditions at the inception of each valley step (Brardinoni and Hassan, 2006). In contrast, steep unglaciated mountain channels are characterized by steady downstream decreases in cou- pling and sediment supply, and steady downstream increases in transport capacity (Montgomery and Buffington, 1997). Most channel types observed in this study were cascade and plane bed. Previous work by Brardinoni and Hassan (2006, 2007) has shown cascade and plane bed reaches to be respectively associated with valley steps and hanging valleys in Pleistocene-glaciated settings. Steep, confined channels on valley steps are characterized by bedrock, colluvial or cascade morphol- ogy. Lower gradient, decoupled channels on hanging valleys tend to exhibit plane bed morphology. In most of the field-surveyed locations, the length of channel exposed post-LIA is contained within a valley step or hanging valley. For example, the entire post-LIA extent of the channel below Colonel Foster Glacier occupies a single valley step, while the entire post-LIA Helm Glacier channel flows through a large hanging valley. Only one field-surveyed post- LIA channel, at Miller Glacier, encompasses a succession of valley steps and hanging valleys. A key difference between this study and Brardinoni and Hassan (2007) 61 is the lack of coupling observed in post-LIA reaches. Even typically coupled cascade reaches on valley steps were in most cases isolated from surround- ing hillslopes by a broad valley flat. Nevertheless, the channel morphologies observed in this study are similar to those reported by Brardinoni and Has- san (2007). A possible explanation for this similarity is the large amount of mobile sediment stored within or close to the active channel in post-LIA reaches. Over time, cascade and plane bed reaches may evolve into more struc- tured forms. Large-diameter sediment supplied by glacial deposits will even- tually create jamming and lead to pool formation, causing a transition from cascade to step-pool morphology on valley steps. Low-gradient hanging val- leys may develop riffle-pool rather than plane bed morphology. However, results from Pleistocene-glaciated channels in the same region (Brardinoni and Hassan, 2007) suggest that substantial morphological transition may not take place for thousands of years post-glaciation, governed by landscape recovery. The dominance of slope as a control on reach-scale morphology is sim- ilar to results reported previously for glaciated and unglaciated reaches, although slope ranges for each morphology type are consistently higher in glaciated terrain (Montgomery and Buffington, 1997; Wohl et al., 2004; Wohl and Merritt, 2005; Flores et al., 2006; Brardinoni and Hassan, 2007). The significance of slope as a first-order control regardless of underlying land- scape structure strengthens evidence supporting the channel morphology type criteria developed by Montgomery and Buffington (1997) and Hassan et al. (2005a) and indicates that this classification scheme can be successfully applied to glaciated streams. Previous studies of glaciated and unglaciated mountain streams have noted the power of slope and contributing drainage area in differentiating landscape process domains and hence channel morphology (Montgomery and Foufoula-Georgiou, 1993; Montgomery and Buffington, 1997; Brardi- noni and Hassan, 2006, 2007). The results presented here do not support drainage area as an effective discriminator of channel morphology. This lack of correlation is likely to be a consequence of the limited range of drainage 62 areas. Most contributing areas were on the order of 100-101 km2. Only four reaches were outside this range, with areas on the order of 10-1 km2. Although the bankfull depth-area and bankfull width-area relations suggest similarity to the empirical downstream hydraulic geometry (DHG) relations d ∝ Q0.4 and w ∝ Q0.5, it is difficult to draw any information about DHG from this study due to the limited spatial range. The apparent lack of correlation between channel geometry and drainage area found by this study and by Brardinoni and Hassan (2007) is possibly ex- plained by the unique topography of the glaciated Pacific Northwest moun- tains, which gives rise to multiple channel type sequences at substantially different drainage areas. Time is an additional factor governing develop- ment of channel morphology. The general disorganization of morphology types with respect to area suggests that post-LIA reaches have not had sufficient time to develop structured morphology. Landscape morphometry partially dictates shear stress, which performed well in differentiating plane bed and cascade reaches. The similar chan- nel type discrimination achieved by slope and shear stress is an expected result, because shear stress is related to slope. Higher shear stress is cor- related with steeper morphologies (cascade and step-pool) found on valley steps. Lower-gradient plane bed and riffle-pool reaches on hanging valleys are characterized by lower shear stress. Similar results have been observed for unglaciated landscapes (Montgomery and Buffington, 1997). The driving force supplied by local flow hydraulics is countered by re- sisting force from channel substrate, and the ratio between driving and re- sisting force (expressed in this study by τ/D95 and Ω/D95) is often used to explain morphological variations (e.g. Montgomery and Buffington, 1997). In unglaciated mountain streams, the ratio of hydraulic driving force to sub- strate resistance is expected to steadily increase downstream. Results from this study and other work in glaciated settings (Brardinoni and Hassan, 2007) indicate that this trend does not apply to glacially induced hanging valley sequences. This reflects not only the influence of slope variations on hydraulic force, but also variations in substrate resistance originating from glacial deposits. Trends of downstream coarsening observed in this study 63 and by Brardinoni and Hassan (2007) suggest contemporary remobilization of glacial sediment deposits. Higher values of shear stress and D95 are asso- ciated with cascade reaches on valley steps, while lower shear stress and D95 characterize lower-gradient plane bed reaches in hanging valleys. Trends in τ/D95 appear to be driven by shear stress, with cascade reaches showing a larger τ/D95 than plane bed reaches. However, neither D95 nor shear stress shows clear trends with respect to drainage area, suggesting that down- stream fining across each valley step-hanging valley transition is reset at the inception of the next valley step. 4.2 Riparian vegetation Results from this study indicate that altitude and time since deglaciation are primary controls on vegetation establishment and growth alongside proglacial streams. Previous studies of vegetation colonization on glacial forefields have also identified altitude and age as first-order controls (Matthews, 1979; Matthews and Whitaker, 1987; Mizuno, 1998; Caccianiga et al., 2001), al- though these studies did not focus specifically on riparian zones. At 1430 m above sea level, the mean reach elevation in this analysis, vegetation es- tablishment is predicted to take about 80 years (Figure 3.9). Although a degree of bank stabilization from roots may be initiated at the earliest stages of vegetation colonization, riparian forest may take 500 years to reach a size of potentially significant influence on channel morphology (Figure 3.10). Although riparian vegetation contributes to bank strength and sediment stabilization before fully mature forest develops (e.g. Smith, 1976), the timeline for mature forest presence does not represent the establishment of pool-forming instream wood. The lag time between riparian forest devel- opment and inputs of functional instream wood may be significant and is affected by several factors. Wood recruitment is dependent upon physical processes, such as landslides and avalanches, and tree mortality. Although several studies have shown landslides to be an important source of wood to headwater streams (e.g. Keller and Swanson, 1979; May and Gresswell, 2003; Reeves et al., 2003), the decoupled nature of post-LIA channels may 64 limit lateral delivery of wood. In addition, the rate of chronic mortality (often represented by maximum tree age) may extend the timeline for wood inputs. Maximum tree age of mountain hemlock (Tsuga mertensiana), one of the most commonly observed tree species in this study, may exceed 800 years, even at high elevations (Means, 1990). Functional instream wood was only observed in two reaches. At Over- lord Glacier, it clearly originated from a moraine, because the reach had only been ice-free for about 180 years and there was no mature forest near the channel. The only reach containing functional wood that appeared to be of riparian rather than moraine origin was near the LIA maximum of Colonel Foster Glacier, over 600 years since deglaciation. Elevation of the Colonel Foster channel (894-967 m asl) is the lowest of any in this study, which possibly contributes to the higher degree of riparian forest develop- ment. Younger (200-300 years) reaches in the Colonel Foster channel were also bordered by mature forest, with numerous fallen trees caused by bank failure. However, the channel was so narrow that the trees spanned rather than entered it. This observation suggests a potentially significant time lag between mature forest establishment and instream wood contributions, as the channel-spanning fallen trees must decay and break before entering the channel (Nakamura and Swanson, 1993). Decomposition rates of wood in subalpine zones can be long. Cool air temperatures may limit microbial ac- tivity and wood decay (Kueppers et al., 2004; Hassan et al., 2005b). Terres- trial decay rates (as fractional mass loss) for coarse woody debris in northern coniferous forests range from 0.0025 to 0.071 yr-1 (Laiho and Prescott, 2004) and are typically less than 0.015 yr-1 (Harmon et al., 1986). Decay rates for northern forests are low relative to those reported for warmer mountainous areas such as the Appalachians (e.g. Harmon, 1982; Mattson et al., 1987). Sediment inputs from disintegrating LIA moraines are likely to extend the timescale for riparian forest development. Results from this study indi- cate a correlation between more advanced stages of vegetation development and more stable channel form, but the long recovery period of glaciated land- scapes indicates that this is likely a reflection of the control of vegetation on channel form, rather than channel form as a control on vegetation. Previous 65 studies have also found that the establishment and succession of vegetation in riparian zones is likely to be slower than in the surrounding valley flat and uplands due to the effects of bank instability and high sediment loads (Sidle and Milner, 1989; Milner and Gloyne-Phillips, 2005). Delayed riparian forest development has implications for stream biota. Canopy closure was not observed for any post-LIA reaches in this study. Channel shading, which regulates water temperature (e.g. Moore et al., 2005) and solar ultraviolet radiation (Kelly and Bothwell, 2002; Kelly et al., 2003), is unlikely to occur in these reaches for hundreds of years. Although pioneer herbaceous vegetation and shrubs will supply minor inputs of organic material to the channels, reaches are likely to be starved of allochthonous inputs until larger and more abundant vegetation is present. The models proposed by this study are useful to establish a general timeline of riparian forest development and maturity, but they can only be applied at relatively coarse resolution. Only six of the 25 catchments in- cluded in the vegetation analysis were visited in the field, and no systematic vegetation surveys were performed. Precise riparian zone delineation was not possible using remote imagery. Although climatic variables were not direct measurements, elevation is a useful first-order indicator of local climate. Increasing elevation coincides with decreasing air temperature and local increases in both total precipi- tation and the fraction of precipitation that falls as snow, which are major controls on treeline (e.g. Körner, 1998). 4.3 Fish habitat The majority of recently deglaciated streams included in this study were found to be potentially suitable for fish based on a coarse analysis of channel gradients. Glacially imposed topography gives rise to lower gradient reaches on hanging valleys, which provide habitat opportunities. Gradient segment analysis suggests that fish may be able to travel up many of the high-gradient valley steps below hanging valleys to access more amenable reaches. Only 25% of the post-LIA reach length included in this analysis is inaccessible 66 to fish due to gradient. Additional habitat could be provided if proglacial lakes are stocked, although the ecological consequences of lake stocking with nonnative species are subject to much debate (e.g. Pilliod and Peterson, 2001; McGarvie Hirner and Cox, 2007). Under present ice conditions, most of the recently deglaciated streams included in this study are ideally suited to cold water species such as bull trout. However, most fish observations in downstream reaches are of rainbow trout, a cool water species. This suggests that there is currently a thermal barrier preventing rainbow trout migration further upstream. Future water temperature projections for basins without ice cover indicate that there may be a shift in thermal habitat suitability from cold water to cool water species in almost half of the surveyed stream length. Climatic warming is likely to lead to additional loss of habitat for cold water species, some of which may disappear altogether if their thermal limit moves above the zone of streams that are suitable in terms of gradient, morphology, and flow (e.g. Meisner, 1990; Rahel et al., 1996; Finstad et al., 2011). As cold water fish are thermally forced further upstream, the total amount of suitable habitat will decline and the habitat will become increasingly fragmented. The results of this study represent an initial, generalized assessment of fish habitat in recently deglaciated streams. Uncertainties in the fish habitat analysis arise from data quality and resolution, as well as the generalized designation of fish barriers. Data on fish observations and obstructions originate from many different sources, so there is no way to assess data quality. Provincial and state fish survey databases do not indicate the full extent of stream surveys. Therefore, the absence of recorded barriers on a stream does not necessarily mean that the stream has been surveyed and found not to contain barriers. Likewise, the absence of fish observations for a stream does not mean that there are no fish present in the stream. The resolution of the analysis is determined by the resolution of the Digital Elevation Models (DEMs) from which elevation was derived. Stream gradient segments are not surrogates for reaches, as gradient segments are defined at a broader scale. Another limitation is that a substantial degree of site-specific information is lost with the use of remote sensing data. 67 Although additional uncertainty is introduced by the use of estimated rather than measured water temperature, the maximum weekly average tem- perature (MWAT) regression equation used to calculate water temperature from air temperature appears to provide a reasonable estimate of the sen- sitivity of MWAT to catchment and climate characteristics. Nelitz et al. (2008) report a standard deviation of prediction errors of 2.1◦C and no ob- vious regional pattern in prediction error. The general designation of barriers as sustained slopes >25% or falls >5 m high is useful for a map-based analysis, but it is not an entirely accurate representation of field conditions. For example, cascades were not considered barriers in this study, which may underrepresent actual barriers, as fish cannot traverse some cascade forms. However, if the roughness elements in a cascade are sufficiently large relative to the channel (such as boulders from glacial lag deposits), they might create periodic resting areas, allowing fish to move through the reach (Powers and Orsborn, 1985). A more accurate determination of fish barriers would require field surveys. Resource availability and pool habitat are important elements of fish habitat quality, and may be important controls on fish establishment and abundance (e.g. Milner et al., 2000; Latterell et al., 2003). Water temper- ature, channel stability, and allochthonous inputs are important determi- nants of macroinvertebrate colonization and community composition (Mil- ner, 1994; Milner and Petts, 1994; Milner et al., 2001; Brown et al., 2003; Milner et al., 2008), with implications for fish populations because macroin- vertebrates are an important food source for fish. The estimated MWAT values for streams in this study indicate that they are thermally suitable for numerous benthic macroinvertebrate taxa (Milner and Petts, 1994; Mil- ner et al., 2001). However, unstable channels may limit macroinvertebrate colonization, and results from the riparian vegetation analysis described in section 4.2 suggest that decades or even centuries may pass before reaches receive substantial allochthonous inputs and instream wood begins to form pool habitat. 68 4.4 Implications for models of proglacial stream evolution The geomorphic effects of Quaternary glaciation, time scale for riparian for- est establishment, and fish habitat characteristics described by this study un- derscore the importance of regionally specific data in the study of proglacial streams. Neither of the two existing conceptual models of proglacial stream development can be directly applied to mountain streams in British Columbia and Washington. The timeline proposed by Sidle and Milner (1989) is based on a region with substantially different topography and climate. The concep- tual model developed by Gurnell et al. (1999) is highly generalized, focusing only on relatively wide, low-gradient fluvial channels with no major lateral inputs of water or sediment. Both the Sidle and Milner (1989) and the Gurnell et al. (1999) mod- els predict proglacial channel changes in the absence of slope-based con- trols. Sediment delivered from the glacier, rather than channel gradient, is incorporated into both models as a first-order morphological control. In contrast, results from this study and other regional work (Church and Slay- maker, 1989; Brardinoni and Hassan, 2006, 2007; Collins and Montgomery, 2011) indicate that large-scale Pleistocene glaciation has imposed an indeli- ble landscape template and sediment legacy that govern present-day channel changes. Glacially carved rock steps are not present within the low-gradient valleys in Glacier Bay, Alaska, where the Sidle and Milner (1989) model applies. The Gurnell et al. (1999) model assumes that the presence of valley steps will only generate minor perturbations in a steady downstream pro- gression from a wide, braided channel to one that is more stable and single- or multi-threaded. Given the short time since post-LIA deglaciation, it is likely that re- cently deglaciated channels are located on the rising limb of the sediment yield curve proposed by Church and Slaymaker (1989). This implies a high degree of landscape instability, large amounts of available sediment, and high sediment yield. Both of the existing conceptual models are based on the assumption that sediment inputs are primarily from the glacier itself, 69 and landscape sediment yield is insignificant. This assumption does not ap- pear to hold true for post-LIA mountain streams, where a large amount of mobile sediment is stored within or close to the active channel. Sediment supply and transport in post-LIA channels vary with location on the valley step-hanging valley sequence, rather than exhibiting the downstream contin- uum predicted by Gurnell et al. (1999) or the temporal continuum described by Sidle and Milner (1989). The results of this study suggest that proglacial riparian development follows a longer timeline than that indicated by previous conceptual models. In Glacier Bay, morphologically functional instream wood is generally estab- lished within 200 years post-glaciation (Sidle and Milner, 1989; Milner and Gloyne-Phillips, 2005; Klaar et al., 2011). Observations from studies in the European Alps, which form the basis of the Gurnell et al. (1999) model, in- dicate that closed herbaceous riparian vegetation develops less than 30 years after glaciation. The relatively long timescale of vegetation establishment found in this study may be indicative of altitude and climate differences, channel instability as a result of high post-LIA sediment inputs, or differ- ences in available pioneer species. Additional delays in wood recruitment and decay are likely to substantially extend the timescale for wood to be- come morphologically functional in the channel, possibly on the order of hundreds of years. Additionally, both the Sidle and Milner (1989) and Gurnell et al. (1999) models are based on catchments that still contain remnant ice, in which glaciers are still large enough to be a major controlling influence on hydro- logic regime. Both models assume that the discharge and sediment regimes of the proglacial rivers are controlled primarily by flows from the glacier itself. This is generally not the case in coastal British Columbia and Wash- ington, where remaining alpine glaciers are in rapid decline. Many glaciers in the Pacific Northwest are already so small relative to their catchments that the hydrologic regime is primarily pluvial and nival, with peak flows driven by autumn rainstorms and snowmelt in late spring and early summer (Stahl and Moore, 2006; Moore et al., 2009). With further glacier shrinkage, the influence of ice will continue to decrease in these basins. 70 Chapter 5 Conclusions This chapter presents the key findings of the research, which address the re- search questions posed in Section 1.3. Possible directions for future research are discussed in the final section. 5.1 Key findings The results of this study emphasize the significance of the Quaternary glacial legacy on modern-day geomorphological processes in the coastal mountains of the Pacific Northwest. Upon this landscape template, post-Little Ice Age (LIA) deglaciation has introduced high sediment yield and instability in headwater catchments. Consistent with the findings of Brardinoni and Hassan (2006, 2007), local slope, imposed by the Pleistocene glaciation, appears to be a primary control on channel morphology. This study demonstrates that riparian vegetation is slow to establish alongside recently-deglaciated channels in British Columbia and Washing- ton, controlled by reach altitude and age. Qualitative observations and the results of previous work (e.g. Nakamura and Swanson, 1993; Hassan et al., 2005b) suggest that substantial inputs of large wood may not occur for many years after mature forests develop. Bank instability and high sediment loads from degrading LIA moraines may contribute to the long timescale of vege- tation establishment. 71 An initial assessment of fish habitat suggests that in many channels, fish may be able to move up steep valley steps to reach more suitable habi- tat in hanging valleys. This is likely to have significant consequences for cold water fish populations. As continued warming shifts thermal limits for cold water species further upstream, their ability to populate recently deglaciated reaches becomes increasingly important. However, suitability of these reaches for fish in terms of allochthonous inputs and abundance of pool habitat is unclear and requires field work to confirm. Generalized, map-based models such as those presented here are useful as exploratory studies, but they cannot be applied to a finer-scale assessment. Map-derived data must be supplemented with field data to refine the models and make them suitable for more specific application. 5.2 Directions for future research This study presents an initial examination of the trajectories of proglacial stream development in the mountains of coastal British Columbia and Wash- ington. The results presented here highlight the importance of understand- ing the evolution of these new headwater reaches as glaciers continue to retreat. Previous conceptual studies are regionally specific (Sidle and Mil- ner, 1989) or highly generalized (Gurnell et al., 1999) and cannot be directly applied to the unique environment of the Coast and North Cascade Moun- tains. An important goal for future work is to add more field data that can be used to predict channel change and habitat formation. Results from this study and by Brardinoni and Hassan (2006, 2007) in- dicate that the topographic legacy of glaciation is a dominant influence on present-day stream morphology development in the Coast and North Cas- cade Mountains, regardless of time since glaciation. However, further work is needed to assess the effects of other environmental conditions, such as hydroclimatic and rock uplift variability. Different colluvial-alluvial transi- tion trends may be expected in different glaciated landscapes (Brardinoni and Hassan, 2006), depending on pre-glacial litho-topographic properties and the nature of the glaciation. One potential approach is extension of 72 field-based channel assessments to other mountain ranges in the region that are subject to different geological characteristics and patterns of glaciation, such as additional channels in the Cascade and Olympic Mountains. A large body of previous work suggests that proglacial lakes are a ma- jor control on stream morphology and temperature (Milner, 1987; Sidle and Milner, 1989; Milner et al., 2000; Chew and Ashmore, 2001; Milner and Gloyne-Phillips, 2005; Milner et al., 2008; Richards et al., in press). An in- triguing avenue for future research is offered by paired catchments of similar size, one with and one without a lake. Paired catchment studies could in- vestigate whether morphological and riparian development are accelerated downstream of a proglacial lake. Two ideal locations for this work were identified in the process of selecting field sites for this study: Miller Creek, near Pemberton, BC (the glacier visited in this study abuts a similar glacier to the north with a large proglacial lake), and Lorna Lake, near Gold Bridge, BC. A potential focus for future research is the effect of hanging valleys on material transport dynamics through streams. The impact of the valley step- hanging valley landscape sequence on downstream cascades of materials and sediment is not clearly understood, and the mountains of British Columbia and Washington present opportunities to study transport dynamics over a wide range of temporal and spatial scales (e.g. a comparison study of catchments that were glaciated during the LIA and catchments that have not been glaciated since the Pleistocene). Refinement of the forest and fish habitat assessments presented here can be achieved by incorporating field measurements of reach-scale gradient, water temperature, channel morphology, and riparian vegetation. Although extensive inventories of proglacial stream biotic communities have been con- ducted in other regions (e.g. Milner and Petts, 1994; Brown et al., 2006; Milner et al., 2008), there is a lack of data for streams in coastal British Columbia and Washington. Fish and macroinvertebrate inventories in post- LIA streams in the Pacific Northwest would strengthen future models of biotic community development. 73 Bibliography Agresti, A. 1992. A survey of exact inference for contingency tables. Statistical Science, 7(1):131–153. Ballantyne, C. K. 2002. Paraglacial geomorphology. Quaternary Science Reviews, 21(18-19):1935–2017. Barnes, H. H. 1967. Roughness Characteristics of Natural Channels. US Geological Survey, Washington, DC. USGS Water-Supply Paper 1849. Barsch, D. and Caine, N. 1984. The nature of mountain geomorphology. Mountain Research and Development, 4(4):287–298. Bolch, T., Menounos, B., and Wheate, R. 2010. Landsat-based inventory of glaciers in western Canada, 1985-2005. Remote Sensing of Environment, 114(1):127–137. Brardinoni, F. and Hassan, M. A. 2006. 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):F01013. Brardinoni, F. and Hassan, M. A. 2007. Glacially induced organization of channel-reach morphology in mountain streams. Journal of Geophysical Research-Earth Surface, 112(F3):F03013. Brardinoni, F., Hassan, M. A., Rollerson, T., and Maynard, D. 2009. Colluvial sediment dynamics in mountain drainage basins. Earth and Planetary Science Letters, 284(3-4):310–319. Brown, L. E., Hannah, D. M., and Milner, A. M. 2003. Alpine stream habitat classification: an alternative approach incorporating the role of dynamic water source contributions. Arctic, Antarctic, and Alpine Research, 35(3):313–322. 74 Brown, L. E., Hannah, D. M., and Milner, A. M. 2005. Spatial and temporal water column and streambed temperature dynamics within an alpine catchment: implications for benthic communities. Hydrological Processes, 19(8):1585–1610. Brown, L. E., Hannah, D. M., and Milner, A. M. 2007. Vulnerability of alpine stream biodiversity to shrinking glaciers and snowpacks. Global Change Biology, 13(5):958–966. Brown, L. E., Milner, A. M., and Hannah, D. M. 2006. Stability and persistence of alpine stream macroinvertebrate communities and the role of physicochemical habitat variables. Hydrobiologia, 560:159–173. Bryant, M. D., Gomi, T., and Piccolo, J. J. 2007. Structures linking physical and biological processes in headwater streams of the Maybeso watershed, Southeast Alaska. Forest Science, 53:371–383. Buffington, J. M. and Montgomery, D. R. 1999. Effects of hydraulic roughness on surface textures of gravel-bed rivers. Water Resources Research, 35(11):3507–3521. Caccianiga, M., Andreis, C., and Cerabolini, B. 2001. Vegetation and environmental factors during primary succession on glacier forelands: some outlines from the Italian Alps. Plant Biosystems, 135(3):295–310. Chapin, F. S., Walker, L. R., Fastie, C. L., and Sharman, L. C. 1994. Mechanisms of primary succession following deglaciation at Glacier Bay, Alaska. Ecological Monographs, 64(2):149–175. Chen, Y.-C., Sung, Q., and Cheng, K. Y. 2003. Along-strike variations of morphotectonic features in the Western Foothills of Taiwan: tectonic implications based on stream-gradient and hypsometric analysis. Geomorphology, 56(1-2):106–137. Chew, L. C. and Ashmore, P. E. 2001. Channel adjustment and a test of rational regime theory in a proglacial braided stream. Geomorphology, 37(1-2):43–63. Church, M. 1983. Pattern of instability in a wandering gravel bed channel. In Collinson, J. D. and Lewin, J., editors, Modern and Ancient Fluvial Systems, pages 169–180. Blackwell Scientific, Oxford. 75 Church, M. 1992. Channel morphology and typology. In Calow, C. and Petts, G., editors, The Rivers Handbook, pages 126–143. Blackwell Scientific, Oxford. Church, M. and Hassan, M. A. 2002. Mobility of bed material in Harris Creek. Water Resources Research, 38(11):1237. Church, M., Kellerhals, R., and Day, T. J. 1989. Regional clastic sediment yield in British Columbia. Canadian Journal of Earth Sciences, 26:31–45. Church, M. and Ryder, J. M. 1972. Paraglacial sedimentation: a consideration of fluvial processes conditioned by glaciation. Geological Society of America Bulletin, 83(10):3059–3071. Church, M. and Slaymaker, O. 1989. Disequilibrium of Holocene sediment yield in glaciated British Columbia. Nature, 337(6206):452–454. Collins, B. D. and Montgomery, D. R. 2011. The legacy of Pleistocene glaciation and the organization of lowland alluvial process domains in the Puget Sound region. Geomorphology, 126(1-2):174–185. Cooper, W. S. 1923a. The recent ecological history of Glacier Bay, Alaska: I. The interglacial forests of Glacier Bay. Ecology, 4:93–128. Cooper, W. S. 1923b. The recent ecological history of Glacier Bay, Alaska: II. The present vegetation cycle. Ecology, 4:223–246. Cooper, W. S. 1923c. The recent ecological history of Glacier Bay, Alaska: III. Permanent quadrats at Glacier Bay: An initial report upon a long-period study. Ecology, 4:355–365. Corenblit, D., Steiger, J., Gurnell, A. M., Tabacchi, E., and Roques, L. 2009. Control of sediment dynamics by vegetation as a key function driving biogeomorphic succession within fluvial corridors. Earth Surface Processes and Landforms, 34(13):1790–1810. Dolezal, J., Homma, K., Takahashi, K., Vyatkina, M. P., Yakubov, V., Vetrova, V. P., and Hara, T. 2008. Primary succession following deglaciation at Koryto Glacier Valley, Kamchatka. Arctic, Antarctic, and Alpine Research, 40(2):309–322. Eaton, B., Church, M., and Ham, D. 2002. Scaling and regionalization of flood flows in British Columbia, Canada. Hydrological Processes, 16(16):3245–3263. 76 Eaton, B. and Moore, R. D. 2010. Regional Hydrology. In Pike, R. G., Redding, T. E., Moore, R. D., Winkler, R. D., and Bladon, K. D., editors, FORREX Compendium of Forest Hydrology and Geomorphology in British Columbia, pages 85–110. BC Ministry of Forests and Range Forest Science Program and FORREX Forum for Research and Extension in Natural Resources. BC Land Management Handbook 66. Eaton, J. G., McCormick, J. H., Goodno, B. E., O’Brien, D. G., Stefany, H. G., Hondzo, M., and Scheller, R. M. 1995. A field information-based system for estimating fish temperature tolerances. Fisheries, 20(4):10–18. Eaton, J. G. and Scheller, R. M. 1996. Effects of climate warming on fish thermal habitat in streams of the United States. Limnology and Oceanography, 41(5):1109–1115. Fastie, C. L. 1995. Causes and ecosystem consequences of multiple pathways of primary succession at Glacier Bay, Alaska. Ecology, 76(6):1899–1916. Fetherston, K. L., Naiman, R. J., and Bilby, R. E. 1995. Large woody debris, physical processes, and riparian forest development in montane river networks of the Pacific Northwest. Geomorphology, 13:133–144. Finstad, A. G., Forseth, T. J., Jonsson, B., Bellier, E., Hesthagen, T., Jensen, A. J., Hessen, D. O., and Foldvik, A. 2011. Competitive exclusion along climate gradients: energy efficiency influences the distribution of two salmonid fishes. Global Change Biology, 17(4):1703–1711. Fleming, S. W. 2005. Comparative analysis of glacial and nival streamflow regimes with implications for lotic habitat quantity and fish species richness. River Research and Applications, 21(4):363–379. Flores, A. N., Bledsoe, B. P., Cuhaciyan, C. O., and Wohl, E. E. 2006. Channel-reach morphology dependence on energy, scale, and hydroclimatic processes with implications for prediction using geospatial data. Water Resources Research, 42:W06412. Flory, E. A. and Milner, A. M. 1999. Influence of riparian vegetation on invertebrate assemblages in a recently formed stream in Glacier Bay National Park, Alaska. Journal of the North American Benthological Society, 18(2):261–273. 77 Flory, E. A. and Milner, A. M. 2000. Macroinvertebrate community succession in Wolf Point Creek, Glacier Bay National Park, Alaska. Freshwater Biology, 44(3):465–480. Fountain, A. G. and Tangborn, W. V. 1985. The effect of glaciers on streamflow variations. Water Resources Research, 21(4):579–586. Gurnell, A. M. 1995. Vegetation along river corridors: hydrogeomorphological interactions. In Gurnell, A. and Petts, G., editors, Changing River Channels, pages 237–260. Wiley, New York, NY. Gurnell, A. M., Edwards, P. J., Petts, G. E., and Ward, J. V. 1999. A conceptual model for alpine proglacial river channel evolution under changing climatic conditions. Catena, 38(3):223–242. Hack, J. T. 1957. Studies of longitudinal profiles in Virginia and Maryland. US Geological Survey Professional Paper, 294-B:45–97. Hack, J. T. 1973. Stream-profile analysis and stream-gradient index. US Geological Survey Journal of Research, 1:421–429. Hannah, D. M., Brown, L. E., Milner, A. M., Gurnell, A. M., McGregor, G. R., Petts, G. E., Smith, B. P. G., and Snook, D. L. 2007. Integrating climate-hydrology-ecology for alpine river systems. Aquatic Conservation-Marine and Freshwater Ecosystems, 17(6):636–656. Hari, R. E., Livingstone, D. M., Siber, R., Burkhardt-Holm, P., and Guttinger, H. 2006. Consequences of climatic change for water temperature and brown trout populations in alpine rivers and streams. Global Change Biology, 12(1):10–26. Harmon, M. E. 1982. Decomposition of standing dead trees in the southern Appalachian mountains. Oecologia, 52(2):214–215. Harmon, M. E., Franklin, J. F., Swanson, F. J., Sollins, P., Gregory, S. V., Lattin, J. D., Anderson, N. H., Cline, S. P., Aumen, N. G., Sedell, J. R., Lienkaemper, G. W., Cromack Jr., K., and Cummins, K. W. 1986. Ecology of coarse woody debris in temperate ecosystems. Advances in Ecological Research, 15:133–302. Harper, J. T. 1993. Glacier terminus fluctuations on Mount Baker, Washington, USA, 1940-1990, and climatic variations. Arctic and Alpine Research, 25(4):332–340. 78 Hasholt, B., Krueger, J., and Skjernaa, L. 2008. Landscape and sediment processes in a proglacial valley, the Mittivakkat Glacier area, Southeast Greenland. Geografisk Tidsskrift-Danish Journal of Geography, 108(1):97–110. Hassan, M. A., Church, M., Lisle, T. E., Brardinoni, F., Benda, L., and Grant, G. E. 2005a. Sediment transport and channel morphology of small, forested streams. Journal of the American Water Resources Association, 41(4):853–876. Hassan, M. A., Hogan, D. L., Bird, S. A., May, C. L., Gomi, T., and Campbell, D. 2005b. Spatial and temporal dynamics of wood in headwater streams of the Pacific Northwest. Journal of the American Water Resources Association, 41(4):899–919. Hewitt, K. 1972. The mountain environment and geomorphic processes. In Slaymaker, H. and McPherson, H., editors, Mountain Geomorphology: Geomorphological Processes of the Canadian Cordillera, pages 17–34. Tantalus Research, Vancouver, BC. BC Geographical Series 14. Hicks, D. M. and Mason, P. D. 1998. Roughness Characteristics of New Zealand Rivers. National Institute of Water and Atmospheric Research, Christchurch. Hogan, D. L. 1986. Channel Morphology of Unlogged, Logged and Torrented Streams in the Queen Charlotte Islands. Land Management Report 49, BC Ministry of Forests, Victoria, BC. Hood, E. and Berner, L. 2009. Effects of changing glacial coverage on the physical and biogeochemical properties of coastal streams in southeastern Alaska. Journal of Geophysical Research-Biogeosciences, 114:G03001. Keller, E. A. and Swanson, F. J. 1979. Effects of large woody organic material on channel form and fluvial processes. Earth Surface Processes and Landforms, 4:361–380. Kellerhals, R., Church, M., and Bray, D. I. 1976. Classification and analysis of river processes. Journal of the Hydraulics Division, American Society of Civil Engineers, 17(HY7):813–829. Kelly, D. J. and Bothwell, M. L. 2002. Avoidance of solar ultraviolet radiation by juvenile coho salmon (Oncorhynchus kisutch). Canadian Journal of Fisheries and Aquatic Sciences, 59(3):474–482. 79 Kelly, D. J., Bothwell, M. L., and Schindler, D. W. 2003. Effects of solar ultraviolet radiation on stream benthic communities: an intersite comparison. Ecology, 84(10):2724–2740. Klaar, M. J., Hill, D. F., Maddock, I., and Milner, A. M. 2011. Interactions between instream wood and hydrogeomorphic development within recently deglaciated streams in Glacier Bay National Park, Alaska. Geomorphology, 130:208–220. Klaar, M. J., Maddock, I., and Milner, A. M. 2009. The development of hydraulic and geomorphic complexity in recently formed streams in Glacier Bay National Park, Alaska. River Research and Applications, 25(10):1331–1338. Knighton, A. D. 1998. Fluvial Forms and Processes: A New Perspective. Edward Arnold, London. Koch, J., Clague, J. J., and Osborn, G. D. 2007. Glacier fluctuations during the past millennium in Garibaldi Provincial Park, southern Coast Mountains, British Columbia. Canadian Journal of Earth Sciences, 44(9):1215–1233. Koch, J., Menounos, B., and Clague, J. J. 2009. Glacier change in Garibaldi Provincial Park, southern Coast Mountains, British Columbia, since the Little Ice Age. Global and Planetary Change, 66(3-4):161–178. Koch, J., Menounos, B., Clague, J. J., and Osborn, G. D. 2004. Environmental change in Garibaldi Park, southern Coast Mountains, British Columbia. Geoscience Canada, 31(3):127–135. Kondolf, G. M., Cada, G. F., Sale, M. J., and Felando, T. 1991. Distribution and stability of potential salmonid spawning gravels in steep boulder-bed streams of the eastern Sierra Nevada. Transactions of the American Fisheries Society, 120(2):177–186. Körner, C. 1998. A re-assessment of high elevation treeline positions and their explanation. Oecologia, 115:445–459. Kovanen, D. J. and Slaymaker, O. 2005. Fluctuations of the Deming Glacier and theoretical equilibrium line altitudes during the Late Pleistocene and Early Holocene on Mount Baker, Washington, USA. Boreas, 34(2):157–175. 80 Kruse, C. G., Hubert, W. A., and Rahel, F. J. 1997. Geomorphic influences on the distribution of Yellowstone cutthroat trout in the Absaroka Mountains, Wyoming. Transactions of the American Fisheries Society, 126:418–427. Kueppers, L. M., Southon, J., Baer, P., and Harte, J. 2004. Dead wood biomass and turnover time, measured by radiocarbon, along a subalpine elevation gradient. Oecologia, 141:641–651. Laiho, R. and Prescott, C. E. 2004. Decay and nutrient dynamics of coarse woody debris in northern coniferous forests: a synthesis. Canadian Journal of Forest Research, 34:763–777. Larocque, S. J. and Smith, D. J. 2003. Little Ice Age glacial activity in the Mt. Waddington area, British Columbia Coast Mountains, Canada. Canadian Journal of Earth Sciences, 40(10):1413–1436. Latterell, J. J., Naiman, R. J., Fransen, B. R., and Bisson, P. A. 2003. Physical contstraints on trout (Oncorhynchus spp.) distribution in the Cascade Mountains: a comparison of logged and unlogged streams. Canadian Journal of Fisheries and Aquatic Sciences, 60:1007–1017. Lawrence, D. B., Schoenike, R. E., Quispel, A., and Bond, G. 1967. Role of Dryas drummondii in vegetation development following ice recession at Glacier Bay, Alaska, with special reference to its nitrogen fixation by root nodules. Journal of Ecology, 55(3):793–&. Leopold, L. B. 1994. A View of the River. Harvard University Press, Cambridge, MA. Lewis, D. and Smith, D. 2004. Dendrochronological mass balance reconstruction, Strathcona Provincial Park, Vancouver Island, British Columbia, Canada. Arctic, Antarctic, and Alpine Research, 36(4):598–606. Maizels, J. K. 1983. Proglacial channel systems: change and thresholds for change over long, intermediate and short time-scales. In Collinson, J. D. and Lewin, J., editors, Modern and Ancient Fluvial Systems, pages 251–266. Blackwell Scientific, Oxford. Malard, F., Galassi, D., Lafont, M., Doledec, S., and Ward, J. V. 2003. Longitudinal patterns of invertebrates in the hyporheic zone of a glacial river. Freshwater Biology, 48(10):1709–1725. 81 Matthews, J. A. 1979. The vegetation of the Storbreen gletschervorfeld, Jotunheim, Norway: I. Introduction and approaches involving classification. Journal of Biogeography, 6:17–47. Matthews, J. A. and Whitaker, R. J. 1987. Vegetation succession on the Storbreen glacier foreland, Jotunheim, Norway: a review. Arctic and Alpine Research, 19:385–395. Mattson, K. G., Swank, W. T., and Waide, J. B. 1987. Decomposition of woody debris in a regenerating, clear-cut forest in the southern Appalachians. Canadian Journal of Forest Research, 17:712–721. May, C. L. and Gresswell, R. E. 2003. Large wood recruitment and redistribution in headwater streams in the southern Oregon Coast Range, USA. Canadian Journal of Forest Research, 33:1352–1362. McCleary, R. J. and Hassan, M. A. 2008. Predictive modeling and spatial mapping of fish distributions in small streams of the Canadian Rocky Mountain foothills. Canadian Journal of Fisheries and Aquatic Sciences, 65(2):319–333. McCleary, R. J., Hassan, M. A., Miller, D., and Moore, R. D. 2011. Spatial organization of process domains in headwater drainage basins of a glaciated foothills region with complex longitudinal profiles. Water Resources Research, 47:W05505. McGarvie Hirner, J. and Cox, S. P. 2007. Effects of rainbow trout (Oncorhynchus mykiss) on amphibians in productive recreational fishing lakes of British Columbia. Canadian Journal of Fisheries and Aquatic Sciences, 64:1770–1780. McGregor, G. R., Petts, G. E., Gurnell, A. M., and Milner, A. M. 1995. Sensitivity of alpine stream ecosystems to climate change and human impacts. Aquatic Conservation-Marine and Freshwater Ecosystems, 5(3):233–247. Means, J. 1990. Tsuga mertensiana (Mountain hemlock). In Silvics of North America Volume 1. US Department of Agriculture Forest Service, Washington, DC. Meisner, J. D. 1990. Effect of climate warming on the southern margins of the native range of brook trout, Salvelinus fontinalis. Canadian Journal of Fisheries and Aquatic Sciences, 47(6):1065–1070. 82 Milner, A. M. 1987. Colonization and ecological development of new streams in Glacier Bay National Park, Alaska. Freshwater Biology, 18(1):53–70. Milner, A. M. 1994. Colonization and succession of invertebrate communities in a new stream in Glacier Bay National Park, Alaska. Freshwater Biology, 32(2):387–400. Milner, A. M., Brittain, J. E., Castella, E., and Petts, G. E. 2001. Trends of macroinvertebrate community structure in glacier-fed rivers in relation to environmental conditions: a synthesis. Freshwater Biology, 46(12):1833–1847. Milner, A. M., Brown, L. E., and Hannah, D. M. 2009. Hydroecological response of river systems to shrinking glaciers. Hydrological Processes, 23(1):62–77. Milner, A. M. and Gloyne-Phillips, I. T. 2005. The role of riparian vegetation and woody debris in the development of macroinvertebrate assemblages in streams. River Research and Applications, 21(4):403–420. Milner, A. M., Knudsen, E. E., Soiseth, C., Robertson, A. L., Schell, D., Phillips, I. T., and Magnusson, K. 2000. Colonization and development of stream communities across a 200-year gradient in Glacier Bay National Park, Alaska, USA. Canadian Journal of Fisheries and Aquatic Sciences, 57(11):2319–2335. Milner, A. M. and Petts, G. E. 1994. Glacial rivers: physical habitat and ecology. Freshwater Biology, 32(2):295–307. Milner, A. M., Robertson, A. L., Monaghan, K. A., Veal, A. J., and Flory, E. A. 2008. Colonization and development of an Alaskan stream community over 28 years. Frontiers in Ecology and the Environment, 6(8):413–419. Mizuno, K. 1998. Succession processes of alpine vegetation in response to glacial fluctuations of Tyndall Glacier, Mt. Kenya, Kenya. Arctic and Alpine Research, 30(4):340–348. Montgomery, D. R. and Buffington, J. M. 1997. Channel-reach morphology in mountain drainage basins. Geological Society of America Bulletin, 109(5):596–611. 83 Montgomery, D. R. and Foufoula-Georgiou, E. 1993. Channel network source representation using digital elevation models. Water Resources Research, 29(12):3925–3934. Moore, R. D., Fleming, S. W., Menounos, B., Wheate, R., Fountain, A., Stahl, K., Holm, K., and Jakob, M. 2009. Glacier change in western North America: influences on hydrology, geomorphic hazards and water quality. Hydrological Processes, 23(1):42–61. Moore, R. D., Spittlehouse, D. L., and Story, A. 2005. Riparian microclimate and stream temperature response to forest harvesting: a review. Journal of the American Water Resources Association, 41(4):813–834. Mount, C., Norris, S., Thompson, R., and Tesch, D. 2011. GIS modelling of fish habitat and road crossings for the prioritization of culvert assessment and remediation. Streamline Watershed Management Bulletin, 14(2):7–13. Published by FORREX Forum for Research and Extension in Natural Resources. Nakamura, F. and Swanson, F. J. 1993. Effects of coarse woody debris on morphology and sediment storage of a mountain stream system in western Oregon. Earth Surface Processes and Landforms, 18:43–61. Nelitz, M. A., MacIsaac, E. A., and Peterman, R. M. 2007. A science-based approach for identifying temperature-sensitive streams for rainbow trout. North American Journal of Fisheries Management, 27(2):405–424. Nelitz, M. A., Moore, R. D., and Parkinson, E. 2008. Developing a Framework to Designate ‘Temperature Sensitive Streams’ in the BC Interior. Technical report, ESSA Technologies, University of British Columbia, and BC Ministry of Environment, Victoria, BC. Neter, J., Kunter, M. H., Nachtsheim, C. J., and Wasserman, W. 1996. Applied Linear Statistical Models. McGraw-Hill, Boston, MA, 4th edition. Norris, S. and Mount, C. 2009. Fish Passage GIS Analysis Methodology and Output Data Specifications. Technical report, BC Ministry of Environment, Victoria, BC. Olden, J. D. and Jackson, D. A. 2002. A comparison of statistical approaches for modelling fish species distributions. Freshwater Biology, 47(10):1976–1995. 84 Osborn, G., Menounos, B., Koch, J., Clague, J. J., and Vallis, V. 2007. Multi-proxy record of Holocene glacial history of the Spearhead and Fitzsimmons ranges, southern Coast Mountains, British Columbia. Quaternary Science Reviews, 26(3-4):479–493. Pelto, M. S. and Hartzell, P. 2004. Change in longitudinal profile on three North Cascades glaciers during the last 100 years. Hydrological Processes, 18(6):1139–1146. Pelto, M. S. and Hedlund, C. 2001. Terminus behavior and response time of North Cascade glaciers, Washington, USA. Journal of Glaciology, 47(158):497–506. Pérez-Peña, J. V., Azañón, J. M., Azor, A., Delgado, J., and González-Loderio, F. 2008. Spatial analysis of stream power using GIS: SLk anomaly maps. Earth Surface Processes and Landforms, 34(1):16–25. Pilliod, D. S. and Peterson, C. R. 2001. Local and landscape effects of introduced trout on amphibians in historically fishless watersheds. Ecosystems, 4:322–333. Ponton, J. R. 1972. Hydraulic geometry in the Green and Birkenhead basins, British Columbia. In Slaymaker, H. and McPherson, H., editors, Mountain Geomorphology: Geomorphological Processes in the Canadian Cordillera, pages 151–160. Tantalus Research, Vancouver, BC. BC Geographical Series 14. Porter, S. C. 1981. Lichenometric studies in the Cascade Range of Washington: establishment of Rhizocarpon geographicum growth curves at Mount Rainier. Arctic and Alpine Research, 13(1):11–23. Powers, P. D. and Orsborn, J. F. 1985. Analysis of Barriers to Upstream Fish Migration: An Investigation of the Physical and Biological Conditions Affecting Fish Passage Success at Culverts and Waterfalls. Technical report, Albrook Hydraulics Laboratory, Department of Civil and Environmental Engineering, University of Washington, Seattle, WA. Province of British Columbia 1996. Channel Assessment Procedure Field Guidebook. BC Ministry of Forests, Lands, and Natural Resource Operations. Province of British Columbia 1998. Fish-stream Identification Guidebook. BC Ministry of Forests, Lands, and Natural Resource Operations. 85 R Development Core Team 2011. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Rahel, F. J., Keleher, C. J., and Anderson, J. L. 1996. Potential habitat loss and population fragmentation for cold water fish in the North Platte River Drainage of the Rocky Mountains: response to climatic warming. Limnology and Oceanography, 41(5):1116–1123. Reeves, G. H., Burnett, K. M., and McGarry, E. V. 2003. Sources of large wood in the main stem of a fourth-order watershed in coastal Oregon. Canadian Journal of Forest Research, 33:1363–1370. Richards, J., Moore, R. D., and Forrest, A. L. In press. Late-summer thermal regime of a small proglacial lake. Hydrological Processes. Richardson, J. S., Zhang, Y., and Marczak, L. B. 2010. Resource subsidies across the land-freshwater interface and responses in recipient communities. River Research and Applications, 26:55–66. Robertson, A. L. and Milner, A. M. 1999. Meiobenthic arthropod communities in new streams in Glacier Bay National Park, Alaska. Hydrobiologia, 397:197–209. Robertson, A. L. and Milner, A. M. 2006. The influence of stream age and environmental variables in structuring meiofaunal assemblages in recently deglaciated streams. Limnology and Oceanography, 51:1454–1465. Rosenfeld, J. 2003. Assessing the habitat requirements of stream fishes: an overview and evaluation of different approaches. Transactions of the American Fisheries Society, 132(5):953–968. Ryder, J. M. 1971a. Some aspects of the morphometry of paraglacial alluvial fans in south-central British Columbia. Canadian Journal of Earth Sciences, 8:1252–1264. Ryder, J. M. 1971b. The stratigraphy and morphology of paraglacial alluvial fans in south-central British Columbia. Canadian Journal of Earth Sciences, 8:279–298. Schiefer, E., Menounos, B., and Wheate, R. 2007. Recent volume loss of British Columbian glaciers, Canada. Geophysical Research Letters, 34(16):L16503. 86 Schumm, S. A. 1977. The Fluvial System. Wiley-Interscience, New York, NY. Schumm, S. A. 1985. Patterns of alluvial rivers. Annual Review of Earth and Planetary Sciences, 13:5–27. Sidle, R. C. and Milner, A. M. 1989. Stream development in Glacier Bay National Park, Alaska, USA. Arctic and Alpine Research, 21(4):350–363. Slaymaker, O. 2009. Proglacial, periglacial or paraglacial? Periglacial and Paraglacial Processes and Environments, 320:71–84. Smith, B. P. G., Hannah, D. M., Gurnell, A. M., and Petts, G. E. 2001. A hydrogeomorphological context for ecological research on alpine glacial rivers. Freshwater Biology, 46(12):1579–1596. Smith, D. G. 1976. Effect of vegetation on lateral migration of anastomosed channels of a glacier meltwater river. Geological Society of America Bulletin, 87(6):857–860. Smith, D. J. and Desloges, J. R. 2000. Little Ice Age history of Tzeetsaytsul Glacier, Tweedsmuir Provincial Park, British Columbia. Géographie physique et Quaternaire, 54(2):135–141. Stahl, K. and Moore, R. D. 2006. Influence of watershed glacier coverage on summer streamflow in British Columbia, Canada. Water Resources Research, 42(6):W06201. State of Oregon 2004. Oregon Administrative Rules, Chapter 340, Division 41: Water Quality Standards-Beneficial Uses, Policies, and Criteria for Oregon. OR Department of Environmental Quality. State of Washington 2003. Water Quality Standards for Surface Waters of the State of Washington, Chapters 173-201A WAC. WA Department of Ecology. Stock, J. and Dietrich, W. E. 2003. Valley incision by debris flows: evidence of a topographic signature. Water Resources Research, 39(4):1089. Sullivan, K., Martin, D. J., Cardwell, R. D., Toll, J. E., and Duke, S. 2000. An Analysis of the Effects of Temperature on Salmonids of the Pacific Northwest with Implications for Selecting Temperature Criteria. Technical report, Sustainable Ecosystems Institute, Portland, OR. 87 Swanson, F. J. and Lienkaemper, G. W. 1978. Physical Consequences of Large Organic Debris in Pacific Northwest Streams. General Technical Report PNW-69, USDA Forest Service. Tockner, K., Malard, F., Burgherr, P., Robinson, C. T., Uehlinger, U., Zah, R., and Ward, J. V. 1997. Physico-chemical characterization of channel types in a glacial floodplain ecosystem (Val Roseg, Switzerland). Archiv für Hydrobiologie, 140(4):433–463. Tucker, G. E. and Bras, R. L. 1998. Hillslope processes, drainage density, and landscape morphology. Water Resources Research, 34:2751–2764. US Environmental Protection Agency 2003. EPA Region 10 Guidance for Pacific Northwest State and Tribal Temperature Water Quality Standards. Region 10 Office of Water, Seattle, WA. EPA 910-B-03-002. VanLooy, J. A. and Forster, R. R. 2008. Glacial changes of five southwest British Columbia icefields, Canada, mid-1980s to 1999. Journal of Glaciology, 54(186):469–478. Wang, T., Hamann, A., and Aitken, S. N. 2007. ClimateBC v3.2: a program to generate climate normal, decade, annual, seasonal and monthly data for genecology and climate change studies in British Columbia. University of British Columbia Department of Forest Science Centre for Forest Conservation Genetics. Wang, T., Hamann, A., and Spittlehouse, D. 2010. ClimateWNA v4.60: a program to generate climate normal, annual, seasonal and monthly data for genecology and climate change studies in Western North America (WNA) region. University of British Columbia Department of Forest Science Centre for Forest Conservation Genetics. Wang, T., Hamann, A., Spittlehouse, D., and Aitken, S. N. 2006. Development of scale-free climate data for western Canada for use in resource management. International Journal of Climatology, 26(3):383–397. Ward, J. V. 1994. Ecology of alpine streams. Freshwater Biology, 32:277–294. Watson, G. and Hillman, T. W. 1997. Factors affecting the distribution and abundance of bull trout: an investigation at hierarchical scales. North American Journal of Fisheries Management, 17(2):237–252. 88 Wipfli, M. S. and Gregovich, D. P. 2002. Export of invertebrates and detritus from fishless headwater streams in southeastern Alaska: implications for downstream salmonid production. Freshwater Biology, 47(5):957–969. Wipfli, M. S., Richardson, J. S., and Naiman, R. J. 2007. Ecological linkages between headwaters and downstream ecosystems: transport of organic matter, invertebrates, and wood down headwater channels. Journal of the American Water Resources Association, 43(1):72–85. Wohl, E. 2000. Mountain Rivers. American Geophysical Union, Washington, DC. Water Resources Monograph 14. Wohl, E., Kuzma, J. N., and Brown, N. E. 2004. Reach-scale channel geometry of a mountain river. Earth Surface Processes and Landforms, 29(8):969–981. Wohl, E. and Merritt, D. 2005. Prediction of mountain stream morphology. Water Resources Research, 41:W08419. Wood-Smith, R. D. and Buffington, J. M. 1996. Multivariate geomorphic analysis of forest streams: implications for assessment of land use impacts on channel condition. Earth Surface Processes and Landforms, 21:377–393. Zimmerman, A. and Church, M. 2001. Channel morphology, gradient profiles and bed stresses during spring runoff in a step-pool channel. Geomorphology, 40:311–328. 89 Appendix A Study sites 90 Table A.1: Catchments and reaches included in study. Letters are assigned to each reach for identifica- tion. ‘Field’ indicates that the reach was surveyed in the field, ‘riparian’ indicates that the reach was included in the riparian forest analysis, and ‘fish’ indicates that the reach was included in the fish habitat analysis. Catchment Region Coordinates Reach Field Riparian Fish Tzeetsaytsul Bella Coola 52◦34’N 126◦21’W A x Black Tusk Garibaldi 49◦58’N 123◦01’W A x x Fleur des Neiges Garibaldi 49◦50’N 122◦35’W A x x Garibaldi Garibaldi 49◦50’N 122◦58’W A x Garibaldi Garibaldi 49◦50’N 122◦58’W B x Griffin Garibaldi 49◦50’N 122◦37’W A x Helm Garibaldi 49◦58’N 123◦00’W A x x x Helm Garibaldi 49◦58’N 123◦00’W D x x x Helm Garibaldi 49◦58’N 123◦00’W E x x x Helm Garibaldi 49◦58’N 123◦00’W F x x x Helm Garibaldi 49◦58’N 123◦00’W H x x x Helm Garibaldi 49◦58’N 123◦01’W I x x x Helm Garibaldi 49◦58’N 123◦01’W J x x x Helm Garibaldi 49◦58’N 123◦01’W K x x x 91 Catchment Region Coordinates Reach Field Riparian Fish Helm Garibaldi 49◦58’N 123◦01’W L x x x Helm Garibaldi 49◦59’N 123◦01’W M x x Helm Garibaldi 49◦59’N 123◦01’W N x x Helm Garibaldi 49◦59’N 123◦01’W O x x Lava Garibaldi 49◦48’N 122◦57’W A x Lava Garibaldi 49◦48’N 122◦57’W B x Lava Garibaldi 49◦48’N 122◦57’W C x Overlord Garibaldi 50◦02’N 122◦51’W A x1 x x Overlord Garibaldi 50◦02’N 122◦51’W B x1 x x Overlord Garibaldi 50◦02’N 122◦51’W C x1 x x Russet Lake Garibaldi 50◦01’N 122◦51’W A x Russet Lake Garibaldi 50◦01’N 122◦51’W B x Russet Lake Garibaldi 50◦01’N 122◦51’W C x Russet Lake Garibaldi 50◦01’N 122◦51’W D x Russet Lake Garibaldi 50◦01’N 122◦51’W E x Russet Lake Garibaldi 50◦01’N 122◦51’W F x Russet Lake Garibaldi 50◦01’N 122◦51’W G x Russet Lake Garibaldi 50◦01’N 122◦51’W H x Russet Lake Garibaldi 50◦01’N 122◦51’W I x 92 Catchment Region Coordinates Reach Field Riparian Fish Russet Lake Garibaldi 50◦01’N 122◦51’W J x1 Staircase Garibaldi 49◦50’N 122◦40’W A x x Warren Garibaldi 49◦53’N 123◦00’W A x x Warren Garibaldi 49◦53’N 123◦00’W B x x Wedgemount Garibaldi 50◦09’N 122◦48’W A x x Wedgemount Garibaldi 50◦09’N 122◦48’W B x x Wedgemount Garibaldi 50◦09’N 122◦48’W C1 x x Wedgemount Garibaldi 50◦09’N 122◦48’W D x x Wedgemount Garibaldi 50◦09’N 122◦48’W E x x Wedgemount Garibaldi 50◦09’N 122◦48’W F x x Wedgemount Garibaldi 50◦09’N 122◦48’W 2A x x Wedgemount Garibaldi 50◦09’N 122◦48’W 2B x x Anniversary Joffre Lakes 50◦20’N 122◦25’W A x Matier Joffre Lakes 50◦20’N 122◦28’W A x x Matier Joffre Lakes 50◦20’N 122◦28’W B x x Matier Joffre Lakes 50◦20’N 122◦28’W C x x Matier Joffre Lakes 50◦20’N 122◦28’W D x x Boulder Mount Baker 48◦45’N 121◦45’W A x1 x Boulder Mount Baker 48◦45’N 121◦45’W B x1 x 93 Catchment Region Coordinates Reach Field Riparian Fish Boulder Mount Baker 48◦45’N 121◦45’W C x1 x Boulder Mount Baker 48◦45’N 121◦45’W D x1 x Boulder Mount Baker 48◦45’N 121◦45’W E x1 x Coleman Mount Baker 48◦48’N 121◦51’W A x Coleman Mount Baker 48◦48’N 121◦52’W B x Coleman Mount Baker 48◦48’N 121◦52’W C x Coleman Mount Baker 48◦48’N 121◦52’W D x Coleman Mount Baker 48◦48’N 121◦52’W E x Coleman Mount Baker 48◦48’N 121◦53’W F x Coleman Mount Baker 48◦48’N 121◦53’W G x Coleman Mount Baker 48◦48’N 121◦53’W H x Deming Mount Baker 48◦44’N 121◦52’W A x Deming Mount Baker 48◦42’N 121◦52’W B x Deming Mount Baker 48◦43’N 121◦53’W C x Easton Mount Baker 48◦43’N 121◦50’W A x x x Easton Mount Baker 48◦43’N 121◦50’W B x x x Easton Mount Baker 48◦43’N 121◦50’W C x x x Easton Mount Baker 48◦43’N 121◦50’W D x x x Easton Mount Baker 48◦43’N 121◦50’W E x x x 94 Catchment Region Coordinates Reach Field Riparian Fish Easton Mount Baker 48◦43’N 121◦50’W F x x x Easton Mount Baker 48◦43’N 121◦50’W G x x x Easton Mount Baker 48◦43’N 121◦50’W H x x x Easton Mount Baker 48◦43’N 121◦50’W I x x x Easton Mount Baker 48◦43’N 121◦50’W J x x x Easton Mount Baker 48◦43’N 121◦49’W K x x Nisqually Mount Rainier 46◦47’N 121◦44’W A x Nisqually Mount Rainier 46◦47’N 121◦45’W B x Astarte Mt Waddington 51◦38’N 125◦12’W A x Byamee Mt Waddington 51◦38’N 125◦11’W A x Cathedral Mt Waddington 51◦14’N 124◦52’W A x x Escape Mt Waddington 51◦37’N 125◦07’W A x Escape Mt Waddington 51◦36’N 125◦07’W B x Liberty Mt Waddington 51◦35’N 124◦50’W A x Liberty Mt Waddington 51◦35’N 124◦50’W B x Liberty Mt Waddington 51◦35’N 124◦50’W C x Oval Mt Waddington 51◦29’N 125◦15’W A x Ragnarok Mt Waddington 51◦29’N 125◦15’W A x Ragnarok Mt Waddington 51◦36’N 125◦18’W B x 95 Catchment Region Coordinates Reach Field Riparian Fish Ragnarok Mt Waddington 51◦36’N 125◦18’W C x Ragnarok Mt Waddington 51◦36’N 125◦18’W D x Razor Creek Mt Waddington 51◦34’N 124◦46’W A x x Razor Creek Mt Waddington 51◦34’N 124◦46’W B x x Siva Mt Waddington 51◦39’N 125◦10’W A x Siva Mt Waddington 51◦39’N 125◦10’W B x Siva Mt Waddington 51◦39’N 125◦10’W C x Siva Mt Waddington 51◦39’N 125◦10’W D x Miller Creek Pemberton 50◦22’N 122◦59’W A x x Miller Creek Pemberton 50◦22’N 122◦59’W B x x Miller Creek Pemberton 50◦22’N 122◦59’W C x x Miller Creek Pemberton 50◦22’N 122◦59’W D x x Miller Creek Pemberton 50◦22’N 122◦59’W E x x Miller Creek Pemberton 50◦22’N 122◦59’W F x x Miller Creek Pemberton 50◦22’N 122◦59’W G x1 x Colonel Foster Strathcona 49◦45’N 125◦51’W A x x x Colonel Foster Strathcona 49◦45’N 125◦51’W B x1 x x Colonel Foster Strathcona 49◦45’N 125◦51’W C x1 x x Colonel Foster Strathcona 49◦45’N 125◦51’W D x x 96 Catchment Region Coordinates Reach Field Riparian Fish Colonel Foster Strathcona 49◦45’N 125◦51’W E x x Colonel Foster Strathcona 49◦45’N 125◦51’W F x x Moving Strathcona 49◦32’N 125◦23’W A x Septimus Strathcona 49◦29’N 125◦31’W A x x x Septimus Strathcona 49◦29’N 125◦31’W B x x x Septimus Strathcona 49◦29’N 125◦31’W C x x x Septimus Strathcona 49◦29’N 125◦31’W D x x x 1uncertain field data due to difficult channel access 97 Appendix B Field site descriptions B.1 Data sources and overview map Table B.1: Data sources for field sites. Data Resolution Source BC geology 1:250,000 BC Ministry of Energy and Mines WA geology 1:100,000 WA Division of Geology and Earth Resources BC ecozones 1:250,000 BC Forest Service Bio- geoclimatic Ecosystem Classification Program WA ecozones 1:100,000 US Geological Survey GAP Analysis Program BC climatic data 2.5 x 2.5 arcmin ClimateBC WA climatic data 2.5 x 2.5 arcmin ClimateWNA 98 Figure B.1: Map of field survey region. Study sites in Canada are in Garibaldi and Strathcona Provincial Parks and near the town of Pemberton (in Crown land and Joffre Lakes Provin- cial Park, not pictured). Study sites in the United States are in the Mount Baker-Snoqualmie National Forest. 99 B.2 British Columbia Coast Mountains: Garibaldi and Joffre Lakes Provincial Parks, Pemberton Gariabldi Provincial Park is located in the southern Coast Mountains about 70 km north of Vancouver. The 1950 km2 park has a strong east-west climate gradient, with moist maritime air masses strongly influencing the southwest part of the park, but increasingly continental climate to the north and east. Joffre Lakes Provincial Park, containing Matier Glacier, is approximately 15 km northeast of Garibaldi Park. The 14.6 km2 park centers around a chain of three glacially-carved lakes. Matier Glacier is the easternmost Canadian glacier included in this study, characterized by an interior, rather than coastal, biogeoclimatic zone. Miller Glacier, approximately 20 km northwest of Garibaldi Park and 15 km northwest of the city of Pemberton, is located in Crown land. Dendrological studies have established approximate historic ice front po- sitions for Helm and Overlord Glaciers (Koch et al., 2007; Osborn et al., 2007; Koch et al., 2009). Approximate Little Ice Age (LIA) maxima for Wedgemount, Matier, and Miller Glaciers were inferred from moraines and regional studies on glacier fluctuations (Koch et al., 2004; Osborn et al., 2007). 100 Figure B.2: Field study sites in the British Columbia Coast Moun- tains. Garibaldi and Joffre Lakes Provincial Parks are delin- eated in black, and study catchments are delineated in red. 101 B.2.1 Helm Glacier The post-LIA Helm Glacier channel flows westward through a chain of proglacial lakes on a large hanging valley. Scattered riparian vegetation is present in the upper channel, but no riparian forest was observed in post-LIA reaches. The catchment is underlain by basaltic volcanic rocks. The biogeoclimatic zone of the post-LIA portion of the channel is Coastal Mountain-heather Alpine. Figure B.3: Map of Helm Glacier indicating surveyed portion of channel. Catchment is delineated at the downstream point of the field survey. 102 Table B.2: Characteristics of Helm Glacier and catchment. Eleva- tion refers to the channel head elevation where field surveys began. MAT is mean annual temperature and MAP is mean annual precipitation. Coordinates Glacier Area (km2) Drainage Area (km2) Elevation (m asl) Survey Length (m) MAT (◦C) MAP (mm) 49◦58’N 123◦00’W 0.74 8.23 1838 1774 2.6 3833 500 1000 2000 5000 10000 0 20 0 40 0 60 0 80 0 10 00 0 20 40 60 80 10 0 H ei gh t (m ) SL /k  In de x Distance from stream source (m) Figure B.4: Helm channel HACK profile (black solid line) with SL/k index curve (red solid line). Dotted purple line is boundary between geological formations. Dashed black line is fault boundary. 103 (a) (b) (c) (d) (e) (f) Figure B.5: Helm Glacier channel. (a) Channel age 30 years; (b) first appearance of herbaceous vegetation, channel age 30-40 years; (c) instream wood from moraine, channel age 65 years; (d) LIA maximum, channel age 80 years; (e-f) mature forest and step-pool morphology in undated channel approximately 2.3 km downstream from LIA maximum. 104 B.2.2 Overlord Glacier The channel flowing from Overlord Glacier cascades over a steep bedrock outcrop before flowing through the steep forefield exposed since the LIA. The post-LIA channel is high gradient cascade morphology. After the LIA end moraine, however, the gradient abruptly decreases and the channel be- comes riffle-pool. Scattered trees and shrubs are present within the post-LIA zone, but there is no riparian forest until after the LIA end moraine. The catchment is underlain by andesitic volcanic rocks. The post-LIA portion of the Overlord channel is in the Mountain Hemlock biogeoclimatic zone. Figure B.6: Map of Overlord Glacier and Russet Lake indicating surveyed portion of channels. Catchments are delineated at the downstream point of the field survey. 105 Table B.3: Characteristics of Overlord Glacier and catchment. Ele- vation refers to the channel head elevation where field surveys began. MAT is mean annual temperature and MAP is mean annual precipitation. Coordinates Glacier Area (km2) Drainage Area (km2) Elevation (m asl) Survey Length (m) MAT (◦C) MAP (mm) 50◦02’N 122◦50’W 1.9 9.68 1690 1525 1.8 1419 100 200 500 1000 2000 5000 10000 0 20 0 40 0 60 0 80 0 10 00 0 20 40 60 80 10 0 H ei gh t (m ) SL /k  In de x Distance from stream source (m) Figure B.7: Overlord channel HACK profile (black solid line) with SL/k index curve (red solid line). Dotted purple line is boundary between geological formations. Dashed black line is fault boundary. 106 (a) (b) (c) (d) Figure B.8: Overlord Glacier channel. (a) Bedrock cascade emerg- ing from the base of the glacier; (b) channel age 80 years; (c) instream wood from moraine, channel age 180 years; (d) LIA maximum, channel age 310 years. 107 B.2.3 Russet Lake The Russet Lake catchment is adjacent to Overlord Glacier (Figure B.6) but remained unglaciated during the LIA. Russet Creek, flowing from a small lake, borders the Overlord LIA lateral moraine for much of its extent before converging with the Overlord channel near the LIA end moraine. The Russet Creek channel flowing from the lake has alternating sections of high and low gradient. Herbaceous vegetation surrounds the channel, and there are mature trees on surrounding hillslopes, but riparian forest is not present until near the confluence with the Overlord channel. Like the Overlord catchment, the Russet Lake catchment is underlain by andesitic volcanic rocks. The upper half of Russet Creek is in the Coastal Mountain-heather Alpine biogeoclimatic zone and the lower half, near its confluence with the Overlord channel, is in the Mountain Hemlock zone. 108 Table B.4: Characteristics of Russet Lake catchment. Elevation refers to the channel head elevation where field surveys began. MAT is mean annual temperature and MAP is mean annual precipitation. Coordinates Glacier Area (km2) Drainage Area (km2) Elevation (m asl) Survey Length (m) MAT (◦C) MAP (mm) 50◦01’N 122◦51’W no glacier 1.86 1884 475 0.7 1373 200 500 1000 2000 0 10 0 20 0 30 0 40 0 0 20 40 60 80 10 0 H ei gh t (m ) SL /k  In de x Distance from stream source (m) Figure B.9: Russet Creek HACK profile (black solid line) with SL/k index curve (red solid line). 109 (a) (b) (c) (d) Figure B.10: Russet Creek. (a) Cascade near lake; (b) herbaceous vegetation surrounding channel; (c) channel flowing towards Overlord lateral moraine (glacier is visible in upper right); (d) lower gradient plane bed reach. 110 B.2.4 Wedgemount Glacier The Wedgemount channel flows over the steep, rocky glacier forefield before entering a wide delta into Wedgemount Lake. Scattered herbaceous veg- etation and seedlings <1 m tall were observed, but no mature forest was present in the riparian zone or on surrounding hillslopes. The channel and lake are in the Coastal Mountain-heather Alpine biogeoclimatic zone. The catchment is underlain by quartz diorite. Figure B.11: Map of Wedgemount Glacier indicating surveyed por- tion of channel. Catchment is delineated at the downstream point of the field survey. 111 Table B.5: Characteristics of Wedgemount Glacier and catchment. Elevation refers to the channel head elevation where field sur- veys began. MAT is mean annual temperature and MAP is mean annual precipitation. Coordinates Glacier Area (km2) Drainage Area (km2) Elevation (m asl) Survey Length (m) MAT (◦C) MAP (mm) 50◦09’N 122◦48’W 1.49 5.87 1960 317 1.8 2049 100 200 500 1000 2000 5000 0 20 0 60 0 10 00 14 00 0 20 40 60 80 10 0 H ei gh t (m ) SL /k  In de x Distance from stream source (m) Figure B.12: Wedgemount channel HACK profile (black solid line) with SL/k index curve (red solid line). Dotted purple line is boundary between geological formations. 112 (a) (b) (c) (d) Figure B.13: Wedgemount Glacier channel. (a) Glacier and sur- veyed portion of channel; (b) cascade near base of glacier; (c) channel entering delta at head of Wedgemount Lake; (d) herbaceous vegetation on delta. 113 B.2.5 Matier Glacier The channel flowing from Matier Glacier is initally colluvial. About 100 m downstream from the ice front, flow paths converge and the channel takes on a cascade morphology. Abundant herbaceous vegetation and shrubs were observed around the channel, and scattered trees about 1-2 m tall were present on the surrounding hillslopes, but there were no mature trees in the riparian zone. The channel flows through discontinuous exposures of bedrock and steep boulder cascades before entering Upper Joffre Lake. The catchment is underlain by diorite and is in the Interior Mountain-heather Alpine biogeoclimatic zone. Figure B.14: Map of Matier Glacier indicating surveyed portion of channel. Catchment is delineated at the downstream point of the field survey. 114 Table B.6: Characteristics of Matier Glacier and catchment. Ele- vation refers to the channel head elevation where field surveys began. MAT is mean annual temperature and MAP is mean annual precipitation. Coordinates Glacier Area (km2) Drainage Area (km2) Elevation (m asl) Survey Length (m) MAT (◦C) MAP (mm) 50◦20’N 122◦27’W 2.58 3.21 1622 202 2.1 2220 200 500 1000 2000 0 10 0 30 0 50 0 0 20 40 60 80 10 0 H ei gh t (m ) SL /k  In de x Distance from stream source (m) Figure B.15: Matier channel HACK profile (black solid line) with SL/k index curve (red solid line). Dotted purple line is boundary between geological formations. 115 (a) (b) (c) (d) (e) Figure B.16: Matier Glacier channel. (a) Glacier and surveyed por- tion of channel; (b) colluvial channel near base of glacier; (c) bedrock exposure; (d) shrubs surrounding channel; (e) channel flowing into Upper Joffre Lake. 116 B.2.6 Miller Glacier At the base of Miller Glacier, the channel is low gradient plane bed, flowing from a small proglacial lake. About 200 m downstream from the ice front the channel cascades over a valley step, then enters another low gradient hanging valley for several hundred metres before flowing over a second valley step. The base of the second step is the approximate LIA maximum for the glacier. Abundant herbaceous vegetation and shrubs surround the upper channel, but mature forest was only observed near the LIA maximum. The catchment is underlain by quartz diorite. The upper portion of the channel is in the Coastal Mountain-heather Alpine biogeoclimatic zone, and the lower portion near the LIA maximum is in the Mountain Hemlock zone. Figure B.17: Map of Miller Glacier indicating surveyed portion of channel. Catchment is delineated at the downstream point of the field survey. 117 Table B.7: Characteristics of Miller Glacier and catchment. Eleva- tion refers to the channel head elevation where field surveys began. MAT is mean annual temperature and MAP is mean annual precipitation. Coordinates Glacier Area (km2) Drainage Area (km2) Elevation (m asl) Survey Length (m) MAT (◦C) MAP (mm) 50◦22’N 122◦59’W 1.86 3.25 1573 1315 3.6 1895 200 500 1000 2000 5000 10000 0 20 0 40 0 60 0 80 0 10 00 0 20 40 60 80 10 0 H ei gh t (m ) SL /k  In de x Distance from stream source (m) Figure B.18: Miller channel HACK profile (black solid line) with SL/k index curve (red solid line). Dotted purple line is boundary between geological formations. 118 (a) (b) (c) (d) (e) (f) Figure B.19: Miller Glacier channel. (a) Channel between LIA lat- eral moraines; (b) proglacial lake outlet; (c) herbaceous veg- etation along channel near glacier; (d) cascade on first valley step; (e) low gradient channel on hanging valley; (f) second valley step near LIA maximum. 119 B.3 British Columbia Insular Mountains: Strathcona Provincial Park Strathcona Provincial Park, located on central Vancouver Island, has a cool and wet climate with deep late-melting snowpacks. Glaciers are not un- common in the 2500 km2 park, but they are generally very small. Colonel Foster and Septimus Glaciers are approximately 40 km apart and are both located in the center of the island, roughly equidistant from the coastlines. Glacial histories of Colonel Foster and Septimus are well-established (Lewis and Smith, 2004). Figure B.20: Field study sites in the British Columbia Insular Mountains. Strathcona Provincial Park is delineated in black, and study catchments are delineated in red. 120 B.3.1 Septimus Glacier The Septimus post-LIA channel flows through a series of low, nested LIA moraines before entering Cream Lake. Scattered herbaceous vegetation is abundant along the channel, and mature trees are present on hillslopes, but no mature forest was observed in direct proximity to the channel. The catchment is underlain by basaltic and rhyolitic volcanic rocks, and is in the Mountain Hemlock biogeoclimatic zone. Figure B.21: Map of Septimus Glacier indicating surveyed por- tion of channel. Catchment is delineated at the downstream point of the field survey. 121 Table B.8: Characteristics of Septimus Glacier and catchment. Ele- vation refers to the channel head elevation where field surveys began. MAT is mean annual temperature and MAP is mean annual precipitation. Coordinates Glacier Area (km2) Drainage Area (km2) Elevation (m asl) Survey Length (m) MAT (◦C) MAP (mm) 49◦29’N 125◦31’W 0.08 6.78 1340 258 3.7 2805 100 200 500 1000 2000 5000 0 20 0 40 0 60 0 80 0 0 20 40 60 80 10 0 H ei gh t (m ) SL /k  In de x Distance from stream source (m) Figure B.22: Septimus channel HACK profile (black solid line) with SL/k index curve (red solid line). Dotted purple line is boundary between geological formations. 122 (a) (b) (c) (d) Figure B.23: Septimus Glacier channel. (a) First appearance of vegetation near base of ice, channel age 75 years; (b) near LIA maximum, channel age 314 years; (c) near LIA maxi- mum; (d) channel entering Cream Lake approximately 100 m downstream from LIA maximum. 123 B.3.2 Colonel Foster Glacier The Colonel Foster Glacier channel flows from a small proglacial lake, then cascades through a series of steep, unconsolidated LIA moraines. Near the LIA maximum, the gradient decreases and the channel develops riffle-pool morphology. Mature riparian forest is present near the LIA maximum, and one piece of functional instream wood was observed. The catchment is underlain by basaltic volcanic rocks. The biogeoclimatic zone of the upper surveyed channel is Mountain Hemlock and the zone of the lower surveyed channel is Coastal Western Hemlock. Figure B.24: Map of Colonel Foster Glacier indicating surveyed portion of channel. Catchment is delineated at the down- stream point of the field survey. 124 Table B.9: Characteristics of Colonel Foster Glacier and catchment. Elevation refers to the channel head elevation where field sur- veys began. MAT is mean annual temperature and MAP is mean annual precipitation. Coordinates Glacier Area (km2) Drainage Area (km2) Elevation (m asl) Survey Length (m) MAT (◦C) MAP (mm) 49◦45’N 125◦51’W 0.35 2.45 967 330 6.0 3535 200 500 1000 2000 5000 0 10 0 20 0 30 0 40 0 0 20 40 60 80 10 0 H ei gh t (m ) SL /k  In de x Distance from stream source (m) Figure B.25: Colonel Foster channel HACK profile (black solid line) with SL/k index curve (red solid line). 125 (a) (b) (c) (d) (e) Figure B.26: Colonel Foster Glacier channel. (a) Channel at proglacial lake outlet; (b) boulder cascade, channel age 240 years; (c) mature forest and channel-spanning wood, channel age 300 years; (d) riffle-pool morphology near LIA maxi- mum, channel age 600 years; (e) instream wood near LIA maximum. 126 B.4 Washington North Cascade Mountains: Mount Baker-Snoqualmie National Forest Mount Baker-Snoqualmie National Forest extends over 10,357 km2 in the North Cascades of Washington State (the 1450 km2 Mount Baker Ranger District is pictured). Boulder and Easton Glaciers are located on the south- ern slopes of Mount Baker in the northernmost portion of the national forest, near the Canada-US border. Mount Baker is the second-most glaciated of the Cascade Range volcanoes, after Mount Rainier. The mountain receives very heavy snowfall. Approximate historic ice front positions of Boulder and Easton were obtained from studies by Pelto and Hedlund (2001) and Pelto and Hartzell (2004). Figure B.27: Field study sites in the Washington North Cascade Mountains. The Mount Baker Ranger District is delineated in black, and study catchments are delineated in red. 127 B.4.1 Boulder Glacier The Boulder Glacier channel flows over a steep bedrock outcrop at the base of the glacier into a broad valley flat bordered by steep LIA moraines. The channel is predominantly boulder cascade morphology, with large grain size. Scattered herbaceous vegetation, trees, and shrubs are present within the valley flat and near the channel, but vegetation was not observed to be a controlling influence on channel form. Catchment geology is basaltic and andesitic volcanic, and the post-LIA channel is within the Moraine Mixed Conifer Forest biogeoclimatic zone. Figure B.28: Map of Boulder Glacier indicating surveyed portion of channel. Catchment is delineated at the downstream point of the field survey. 128 Table B.10: Characteristics of Boulder Glacier and catchment. Ele- vation refers to the channel head elevation where field surveys began. MAT is mean annual temperature and MAP is mean annual precipitation. Coordinates Glacier Area (km2) Drainage Area (km2) Elevation (m asl) Survey Length (m) MAT (◦C) MAP (mm) 48◦43’N 121◦50’W 3.4 5.45 1170 733 7.8 2176 200 500 1000 2000 0 20 0 40 0 60 0 0 20 40 60 80 10 0 H ei gh t (m ) SL /k  In de x Distance from stream source (m) Figure B.29: Boulder channel HACK profile (black solid line) with SL/k index curve (red solid line). Dotted purple line is boundary between geological formations. 129 (a) (b) (c) (d) Figure B.30: Boulder Glacier channel. (a) Upper channel, bedrock cascade from base of glacier visible in the background; (b) bedrock exposure; (c) channel and LIA lateral moraines; (d) channel near LIA maximum surrounded by mature forest. 130 B.4.2 Easton Glacier The post-LIA Easton Glacier channel flows through a wide forefield at a fairly consistent slope. Channel morphology is predominantly boulder cas- cade. Scattered vegetation was observed near the channel, including herba- ceous plants, shrubs, and trees 1-3 m tall. The catchment is underlain by basaltic and andesitic volcanic rocks. The upper portion of the surveyed channel is in the Alpine Grasslands and Shrublands biogeoclimatic zone, and the lower portion near the LIA maximum is in the Moraine Mixed Conifer Forest zone. Figure B.31: Map of Easton Glacier indicating surveyed portion of channel. Catchment is delineated at the downstream point of the field survey. 131 Table B.11: Characteristics of Easton Glacier and catchment. Ele- vation refers to the channel head elevation where field surveys began. MAT is mean annual temperature and MAP is mean annual precipitation. Coordinates Glacier Area (km2) Drainage Area (km2) Elevation (m asl) Survey Length (m) MAT (◦C) MAP (mm) 48◦45’N 121◦45’W 2.07 3.35 1452 413 6.3 2116 200 500 1000 2000 0 10 0 20 0 30 0 40 0 0 20 40 60 80 10 0 H ei gh t (m ) SL /k  In de x Distance from stream source (m) Figure B.32: Easton channel HACK profile (black solid line) with SL/k index curve (red solid line). 132 (a) (b) (c) (d) Figure B.33: Easton Glacier channel. (a) Cascade morphology in upper channel near ice front; (b) channel and LIA lateral moraines; (c) bedrock cascade 5-6 m high; (d) channel near LIA maximum, mature forest present on hillslopes but not in riparian zone. 133 Appendix C Channel morphology types This section provides definitions of the channel morphology types and clas- sification approach used in this study. Bedrock. Characteristically steep and sediment supply limited, bedrock reaches are usually located in the uppermost part of drainage basins (Figure C.1a). They lack alluvial deposits, although episodic events may generate temporary storage of material. Catastrophic scouring from debris flows may periodically sweep stored sediment away to expose the bedrock. The domi- nant roughness elements in bedrock reaches are the bed and banks. Colluvial (chaotic). Colluvial (referred to as chaotic by Brardinoni and Hassan 2006, 2007) channels are small, transport limited headwater streams with sediment inputs from adjacent hillslopes (Figure C.1b, C.1c). These channels generally exhibit weak or ephemeral flow and have insuffi- cient stream power to transport sediment inputs. In recently deglaciated landscapes, hillslope inputs are usually large clasts, which serve as the pri- mary roughness elements. The majority of sediment transport in these steep channels is from episodic debris flows, which drive the long-term sediment flux. In paraglacial landscapes, lag sediment from glaciers supplies a large amount of material to colluvial channels. Residence time of sediment in colluvial channels may be on the order of hundreds of years. 134 Cascade. The term “cascade” can be applied to both channel units and reaches (e.g. Montgomery and Buffington, 1997; Hassan et al., 2005a). In this context it refers to reaches characterized by tumbling flow around large, nonstructured clasts (Figure C.1d, C.1e). Small, non-channel span- ning pools form in flow wakes among disorganized cobbles and boulders. Cascade reaches are most accurately considered a transitional form between colluvial and fluvial, rather than an entirely fluvial morphology (Hassan et al., 2005a). The large particle size to flow depth ratio makes the largest bed-forming material of cascade reaches essentially immobile during typi- cal flows, although fine gravels move during low to moderate flow events. The long residence times of these large clasts result in substantial trapping of finer sediment under and around them, causing high sediment transport rates during the major (i.e. 50-100 year) hydrologic events that move them. Cascade reaches are very steep, occurring on gradients of as much as 45% (Hassan et al., 2005a). Step-Pool. In step-pool morphologies, longitudinal steps formed by large clasts (Figure C.1f) or large wood (Figure C.1g) alternate with pools containing finer material. This form generally occurs at moderate to high gradients of about 3 to 30%, with low width-to-depth ratios and a relative roughness at high flow close to 1 (Hassan et al., 2005a). Most of the sedi- ment storage is in the bedforms, and the steps and banks are the dominant roughness elements. Both step-pool and cascade channels primarily receive sediment inputs from hillslopes and debris flows, but step-pool morphologies are fluvially organized, while cascade morphologies are controlled by both fluvial and nonfluvial processes. Plane bed (rapid). Plane bed channels (alternately termed “rapid” channels by Zimmerman and Church (2001) to distinguish them from the plane bed regime of sand bed hydraulics) lack discrete bars due to low width- to-depth ratios and large relative roughness values (Figure C.1h). They typically occur at moderate to high slopes in relatively straight channels and are dominantly gravel to cobble bedded (although grain sizes may range from sand to small boulders). Plane bed channels are different from both step- pool and riffle-pool channels because they typically lack rhythmic bedforms 135 and pools. Instead, they have long stretches of relatively featureless bed. Bed surfaces are often armoured, giving plane bed channels a high mobility threshold. They appear to represent a transition between supply limited and transport limited regimes (Montgomery and Buffington, 1997). Riffle-pool. These channels have an undulating gravel bed alternating between shallow riffles and deep pools (Figure C.1i). Bedform oscillation is lateral, rather than vertical (as in step-pool channels). Riffle-pool morpholo- gies occur at moderate to low gradients, usually at less than 2% with relative roughness <1 (Hassan et al., 2005a). The primary forms of flow resistance are bedforms and grain roughness. Like plane bed channels, riffle-pool chan- nels exhibit both supply- and transport-limited characteristics, depending on the degree of bed surface armouring and consequent mobility thresholds. 136 (a) Bedrock (b) Colluvial (c) Colluvial (d) Cascade (e) Cascade (f) Step-pool boulder 137 (g) Step-pool wood (h) Plane bed (i) Riffle-pool Figure C.1: Channel morphology types. (a) Bedrock, (b-c) colluvial, (d-e) cascade, (f) step-pool without wood, (g) step-pool with wood, (h) plane bed, (i) riffle-pool. 138 Appendix D Discharge estimates 139 Table D.1: Summary of k -factors used to calculate discharge. Includes gauged catchments used to cal- culate k -factors and k -factors from map by Eaton et al. (2002). Gauged catchment types are: (1) large catchments containing the study catchment, (2) large catchments nearby the study catch- ment but not containing it, and (3) nearby catchments of a similar size to the study catchment. Glacier Hydro station coordinates Hydro station ID Reference catchment area (km2) Reference catchment type Period of record Reference catchment k -factor Map- derived k -factor Overlord & Russet Lake 50◦07’N 122◦57’W 08MG026 90.9 1 1994-2010 0.66428 2.2024 Helm 50◦05’N 123◦02’W 08GA072 285 1 1982-2010 1.4326 2.5254 Wedgemount 50◦17’N 122◦51’W 08MG003 855 1 1914-1951 1.3463 2.1588 Miller 50◦20’N 122◦48’W 08MG005 2160 1 1914-2010 1.8576 1.9281 Matier 50◦26’N 122◦37’W 08MG019 7.25 3 1970-1989 0.7705 1.123 Matier 50◦20’N 122◦43’W 08MG008 596 2 1947-1970 1.0488 1.123 Colonel Foster 49◦42’N 126◦06’W 08HC002 187 2 1957-2010 5.5055 3.5153 Septimus 49◦24’N 125◦45’W 08HC004 114 2 1991-1998 5.1593 4.1075 Easton 48◦41’N 122◦45’W 12191800 21.65 1 1964-1976 1.3064 * Boulder 48◦44’N 122◦41’W 12190718 27.19 3 1983-1990 3.8033 * Easton & Boulder 48◦54’N 121◦50’W 12205000 272 2 1938-2010 2.7095 * Easton & Boulder 48◦46’N 121◦58’W 12207750 10.7 3 1999-2009 1.6513 * * Map-derived k -factor values not available for Washington 140 Appendix E Data sources for riparian forest analysis 141 Table E.1: Riparian forest data sources. ‘Field’ refers to reaches included in the field surveys, with the exception of Coleman Glacier, which was photographed but not surveyed in the field due to steep terrain preventing access to the channel. Glacier Region Image source Image date Historic extent Tzeetsaytsul Bella Coola DigitalGlobe 2003 Smith and Desloges (2000) Black Tusk Garibaldi GeoEye 2009 Koch et al. (2009) Fleur des Neiges Garibaldi DigitalGlobe 2006 Koch et al. (2007, 2009) Garibaldi Garibaldi Air photo, GeoEye 1964, 2009 Koch et al. (2007, 2009) Griffin Garibaldi DigitalGlobe 2006 Koch et al. (2007, 2009) Helm Garibaldi GeoEye, field 2009, 2010 Koch et al. (2007, 2009) Lava Garibaldi Air photo, GeoEye 1964, 2009 Koch et al. (2007, 2009) Overlord Garibaldi Province of BC, field 2003, 2010 Osborn et al. (2007) Staircase Garibaldi DigitalGlobe 2006 Koch et al. (2007, 2009) Warren Garibaldi Air photo, GeoEye 2009 Koch et al. (2007, 2009) Coleman Mt Baker US Geological Survey, field 2006, 2010 Pelto and Hedlund (2001) 142 Glacier Region Image source Image date Historic extent Deming Mt Baker US Geological Survey 2006 Kovanen and Slaymaker (2005) Easton Mt Baker US Geological Survey, field 2006, 2010 Pelto and Hedlund (2001); Pelto and Hartzell (2004) Nisqually Mt Rainier US Dept of Agri- culture Farm Ser- vice Agency 2009 Porter (1981) Astarte Mt Waddington Province of BC 2006 Larocque and Smith (2003) Byamee Mt Waddington Province of BC 2006 Larocque and Smith (2003) Cathedral Mt Waddington Province of BC 2006 Larocque and Smith (2003) Escape Mt Waddington Province of BC 2006 Larocque and Smith (2003) Liberty Mt Waddington Province of BC 2006 Larocque and Smith (2003) Oval Mt Waddington Province of BC 2006 Larocque and Smith (2003) Ragnarok Mt Waddington Province of BC 2006 Larocque and Smith (2003) Razor Creek Mt Waddington Province of BC 2006 Larocque and Smith (2003) Siva Mt Waddington Province of BC 2006 Larocque and Smith (2003) Colonel Foster Strathcona DigitalGlobe, field 2007, 2010 Lewis and Smith (2004) Septimus Strathcona Field 2010 Lewis and Smith (2004) 143 Appendix F Additional reach morphology plots 144 ll l ll l l l l l l ll l l l l l l Slope (m/m) To ta l S tre am  P ow e r (N /m 2 ) (a) 0.01 0.1 1 10 0 10 1 10 2 10 3 10 4 10 5 l l l l l l l l l l l l l ll l l l l l Drainage Area (km2) To ta l S tre am  P ow e r (N /m 2 ) (b) 0.1 1 10 10 0 10 1 10 2 10 3 10 4 10 5 l l l ll l l l l l l l l ll l l l l l D95 (cm) To ta l S tre am  P ow e r (N /m 2 ) (c) 1 10 100 10 0 10 1 10 2 10 3 10 4 10 5 l bedrock colluvial riffle−pool plane bed step−pool cascade Figure F.1: Reach morphology plotted by total stream power as a function of (a) slope, (b) drainage area, and (c) D95. 145 ll l l l l l l l l l l ll l l l l l l Drainage Area (km2) Ω D 95 0.1 1 10 10 2 10 3 10 4 10 5 10 6 l bedrock colluvial riffle−pool plane bed step−pool cascade Figure F.2: Reach morphology plotted by total stream power/D95 as a function of drainage area. 146

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:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0072381/manifest

Comment

Related Items