Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Application of a 2-dimensional hydrodynamic model for the assessment and design of instream channel restoration… Lacey, Ralph William Jay 2001

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

Item Metadata

Download

Media
831-ubc_2001-0052.pdf [ 4.81MB ]
Metadata
JSON: 831-1.0063717.json
JSON-LD: 831-1.0063717-ld.json
RDF/XML (Pretty): 831-1.0063717-rdf.xml
RDF/JSON: 831-1.0063717-rdf.json
Turtle: 831-1.0063717-turtle.txt
N-Triples: 831-1.0063717-rdf-ntriples.txt
Original Record: 831-1.0063717-source.json
Full Text
831-1.0063717-fulltext.txt
Citation
831-1.0063717.ris

Full Text

Application of a 2-dimensional hydrodynamic model for the assessment and design of instream channel restoration works by  Ralph William Jay Lacey B.Sc. Environmental Engineering, University of Guelph, 1996  A THESIS SUBMITTED IN PARTIAL F U L F I L M E N T OF T H E REQUIREMENTS FOR T H E D E G R E E OF MASTER  OF APPLIED SCIENCE in  T H E F A C U L T Y OF G R A D U A T E STUDIES D E P A R T M E N T OF CIVIL ENGINEERING HYDROTECHNICAL  We accept this thesis as conforming to the required s^c/aqd  The University of British Columbia January, 2001 © Ralph William Jay Lacey, 2001  In  presenting  degree  at  this  the  thesis  in  partial  University of  fulfilment  of  of  department  this or  publication of  thesis for by  his  of  or  her  this thesis for  C'^'A  £^»  DE-6 (2/88)  Z<  200/  agree that  representatives.  It  financial gain shall not  ^g<LrV>«r,-  The University of British Columbia Vancouver, Canada  Date y L ^  that the  scholarly purposes may be  permission.  Department  requirements  British Columbia, I agree  freely available for reference and study. I further copying  the  is  for  an advanced  Library shall make  it  permission for extensive  granted  by the  understood  that  head  of  copying  my or  be allowed without my written  Application  of a hydrodynamic model for the assessment and design of instream works  Page ii  ABSTRACT  Currently, stream restoration activities in British Columbia emphasise the use of large woody debris (LWD), boulder, and other habitat structures (Slaney and Zaldokas 1997). The purpose of these works is to create locally varied hydraulic conditions; promote scour and pool formation; and create low velocity refuges. Little work, however, has been done to characterize the changes to local flow hydraulics, morphology and in stream habitat that results from these efforts.  This study provides a methodology, through the use of field surveys and 2-dimensional hydrodynamic modelling, to assess morphological and hydraulic effects of instream L W D and boulder structures. Changes in available fish habitat were quantified through the use of Bovee (1978) probability-of-use curves. The selected field study site is a side channel of the Chilliwack River. Instream structures, installed in the summer of 1999, were subjected to a bankfull flow event that caused significant scour and bed morphology change. Results indicate that pool area increased by 50% due to the hydraulic effects of the instream structures. Two-dimensional flow model velocity and depth predictions compare favourably to recorded field values, while the predicted shear stresses, derived from the model's output, coincide reasonably with the newly formed pool locations. Pre- and post-restoration differences in fish habitat were quantified for a range of discharges using the weighted usable area (WUA) method (Bovee 1982). This analysis determined that the greatest benefit offered by the instream restoration structures was in supplying low velocity refuge areas during high flow events.  Application of a hydrodynamic model for the assessment and design of instream works Page i T A B L E OF CONTENTS  ABSTRACT  .  1 1  T A B L E OF C O N T E N T S _  1 1 1  L I S T OF T A B L E S  .  V  L I S T OF FIGURES  V  1  INTRODUCTION  1  2  BACKGROUND  5  2.1  Channel Hydraulics 2.1.1  2.2  7  1 0  2.2.1  Sedimentation  2.2.2 2.2.3  Orientation Pool type, spacing and formation  ^  2.3.1  Flow resistance  2.3.2  Scour  L  L J  1 4  ,  15  LWD Structures 2.4.1  2.5  .  Durability  :  Fish Habitat 2.5.1  Cover  2.5.2  Coho Salmon habitat  2.5.3  Steelhead trout habitat  2.5.4  Pool use  2.5.5  Logging  Habitat Requirements TAMIHI SIDE C H A N N E L FIELD SITE  3.1  19 ^  .—_  Hydraulics of LWD  2.4  3  Flow Modelling  Natural L WD  2.3  2.6  5  Site History  1  6  2  2  2  3  2 5  —  2  5  _  ^  2  ^  2  — .  ^' 2  7  ^° 3  1  3 1  Application of a hydrodynamic model for the assessment and design of instream works Page i  4  3.2  Watershed History  3.3  Flood Events  3.3.1  Flood frequency analysis of the Chilliwack River  3.3.2  November 1999flood event  RESULTS AND DISCUSSION  4.1  Field Surveys _ _  34 35 36 37 39  39  4.1.1  Structures  4.1.2  Pebble counts  39  4.1.3  Site bathymetry  42  4.1.4  Measured scour  42  4.2  5  . _  Two-Dimensional Flow Model - River2D  39  50  4.2.1  Data manipulation  50  4.2.2  Model verification  51  4.2.3  Discharge scenarios  54  4.2.4  Pool Distributions  61  4.2.5  Shear predictions  61  4.2.6  Fish habitat  69  SUMMARY AND RECOMMENDATIONS  73  REFERENCES  77  APPENDIX A  81  Application of a hydrodynamic model for the assessment and design of instream works  Page v  LIST OF T A B L E S  Table 2.1: Average habitat utilization coefficients for coho salmon and steelhead trout  28  Table 3.1: C F A - Log Pearson Type III flood predictions  37  Table 4.1: Tamihi side channel grain size analysis  39  Table 4.2: Model verification - mean absolute error (MAE) and mean error (ME)  51  Table 4.3: Modelled %WUA for coho salmon at selected discharges  69  Table 4.3: Modelled %WUA for steelhead trout at selected discharges  69  LIST OF FIGURES  Figure 2.1:  Illustration of the flow and scour patterns at an abutment  18  Figure 2.2:  Photograph of flow around an abutment  18  Figure 2.3:  Juvenile steelhead preference curves  29  Figure 3.1:  Site map and site location  32  Figure 3.2  a: Aerial photograph of the Chilliwack River above Tamihi Bridge  33  b: Aerial photograph of the Tamihi side channel  33  Figure 4.1:  Multiple-LWD triangular structure (Strl 1)  40  Figure 4.2:  Single-LWD rootwad structure (Strl4)  40  Figure 4.3:  Single-rock groyne structure (Strl2)  41  Figure 4.4:  Double-rock groyne structure (Str9)  41  Figure 4.5  a: Upper Tamihi side channel pre- and post-flood bathymetry b: Middle Tamihi side channel pre- and post-flood bathymetry  43 44  c: Lower Tamihi side channel pre- and post-flood bathymetry  45  Figure 4.6:  Tamihi side channel longitudinal thalweg pre- and post-flood profiles  46  Figure 4.7  a: Upper Tamihi side channel net scour plot b: Middle Tamihi side channel net scour plot  47 48  c: Lower Tamihi side channel net scour plot  49  Figure 4.8:  Tamihi side channel velocity verification plot (Q = 4.7 m /s)  52  Figure 4.9:  Tamihi side channel water depth verification plot (Q = 4.7 m /s)  3  3  Figure 4.10 a: Upper Tamihi side channel modelled velocity magnitude plots (m/s) b: Middle Tamihi side channel modelled velocity magnitude plots (m/s) c: Lower Tamihi side channel modelled velocity magnitude plots (m/s)  53 55 56 57  Application of a hydrodynamic model for the assessment and design of instream works Figure 4.11  a: Upper Tamihi side channel modelled velocity vector plots b: Middle Tamihi side channel modelled velocity vector plots c: Lower Tamihi side channel modelled velocity vector plots  Page vi 58 59 60  Figure 4.12 a: Upper Tamihi side channel pre- and post-flood pool distributions b: Middle Tamihi side channel pre- and post-flood pool distributions c: Lower Tamihi side channel pre- and post-flood pool distributions  62 63 64  Figure 4.13  65 66 67  a: Upper Tamihi side channel net scour and shear stress intensity plots b: Middle Tamihi side channel net scour and shear stress intensity plots c: Lower Tamihi side channel net scour and shear stress intensity plots  Figure 4.14: Tamihi side channel percent weighted usable area for coho salmon  71  Figure 4.15: Tamihi side channel percent weighted usable area for steelhead trout  72  Figure A. 1 a: Upper Tamihi side channel grain size distribution b: Middle Tamihi side channel grain size distribution c: Lower Tamihi side channel grain size distribution  82 83 84  Application  1  of a hydrodynamic model for the assessment and design of instream works  Page 1  INTRODUCTION  There has been a long history of large woody debris (LWD) removal from the world's river systems. Removal has been undertaken for a variety of purposes including navigation, flood control, and agriculture. In the Pacific Northwest, it was common practice up until the mid-1970's to remove all merchantable timber and L W D from riparian zones and channels (Ralph et al 1994). During this time, many fisheries managers believed that L W D impaired fish production by creating impassable jams which blocked anadromous fish migration (Cherry and Beschta 1989) or scouring channels (Slaney and Martin 1997). Furthermorefromthe 1960's to 1970's a previous section of the Canadian Fisheries Act promoted the removal of L W D from streams. Most streams in coastal British Columbia were logged to the stream banks up until 1988 (Slaney and Zaldokas 1997) when the Government of British Columbia introduced the Coastal Fisheries-Forestry Guidelines (superseded by the Forest Practices Code in 1995) that included instream and riparian zone protection. The long history of disturbance in the Pacific Northwest has left many natural streams and rivers severely impacted and unstable (resulting from increased sediment loads and bank destabilization).  Past forest practices have played an important role in river disturbance where the impacts are not solely restricted to the streams, but are indicative of basin wide problems (Frissell and Nawa 1992). Studies have focused on old logging methods (including riparian vegetation removal, cross-stream felling and instream yarding) that, for the most part, have been replaced by less disruptive methods. Yet recently, declines in coho and steelhead stocks in British Columbia have been related to logging activities (Slaney et al 1996).  The dominant channel impacts due to logging are caused by increases in sediment loads (not peak flow) (Beschta 1984). Excess loads of coarse material promote bed aggradation leading to expanded bars and riffles, infilled pools (particularly where gradients start to decrease (Allan and Lowe 1997) and bank erosion (Hogan et al 1998). Logging roads have been shown to increase coarse and suspended sediment supply to streams through accelerated landslide activity and improper drainage systems. When the bedload transport capacity is insufficient to carry the increased sediment supply (coarse fraction), sediment storage bars become more prominent.  Application of a hydrodynamic model for the assessment and design of instream works  Page 2  Bank destabilization and subsequent channel widening is caused by lateral or medial bars that redirect and concentrate flow towards the banks causing erosion (Madej 1978). Hartman and Scrivener (1990) and Millar (2000) observed significant channel width increases in study reaches where stream banks had been logged. Bank stability declines with riparian vegetation removal (Millar 2000) and the dominance of sand and gravel in alluvial valley streambanks renders them highly susceptible to erosion - particularly when riparian vegetation has been removed by logging (Frissell and Nawa 1992). Intensive timber harvesting, including riparian removal, significantly decreases pool area (Beschta 1984) and depth (Ralph et al 1994). Main channel sections are straightened and pool-riffle diversity is reduced causing a loss of fish habitat (Keller and Tally 1979; Hartman and Scrivener 1990). L W D in forest streams has been found to influence channel morphology [and stability], the movement of sediment, the retention of organic mater, and the composition of the biological community (Bilby and Ward 1989). While the number of small woody debris (< 3m in length) increases after logging, the amount, size, and stability of L W D decreases resulting in reduced cover and complexity of stream habitat (Hartman and Scrivener 1990). The trend of increasing L W D size with increasing stream order (Bilby and Ward 1989; Ralph et al 1994) is not apparent in logged basins. As well, the interaction of L W D with the low-flow channel decreases in logged systems (Ralph et al 1994). Rehabilitation of damaged streams for the purpose of fish habitat restoration has gained much interest in recent years. Fish habitat is directly improved by the addition of instream structures through the creation of locally varied hydraulic condition and low velocity refuges; and the promotion of scour and pool formation.  Slaney and Zaldokas (1997) emphasise the use of cabled-LWD structures, boulder complexes, and other structures as intermediate means of preserving structural diversity as streams and their riparian zones slowly recover from past logging disturbances. Between 1983 through 1998, U.S.A. federal agencies have reintroduced thousands of instream structures to mitigate the effects of timber harvesting (Roper et al 1998).  Little work has been done to characterize the morphological and habitat changes imparted by instream restoration efforts. Many instream restoration activities are performed assuming the  Application of a hydrodynamic model for the assessment and design of instream works  Page 3  problem is a lack of wood, and the simple addition of L W D structures will create incremental fish habitat improvement. Large and costly projects continue to be planned and implemented by federal and state agencies with little or no analysis of their effectiveness (Frissell and Nawa 1992). Instream L W D structures are assumed to act similarly to natural L W D , yet specific knowledge of local flow hydraulics in relation to optimizing fish habitat is unavailable. Engineering interest in fluvial morphology is relatively new and has come about with the realization that design practices are much improved with the understanding and interpretation of morphological features (Kellerhals and Church 1989). Methods of determining structure locations have included determining areas of instability and bank erosion, determining pre-disturbance pool-riffle spacing and investigating existing pool locations. The selected design is usually left to the discretion of the engineer or field biologist and may vary from site to site. As well restoration approaches should consider the interconnectiveness of the physical and biological processes within a watershed (Slaney and Martin 1997). The instream habitat structure design process may be improved through the use of advancements in computational flow modeling programmes.  The ability to perform complex river analysis using either one- or two-dimensional flow models has increased along with the computational ability of desktop computers. One-dimensional (ID) models are generally used to determine flood stage, while two-dimensional (2D) models can be used to determine local hydraulic conditions in terms of local depth (y), and depth-averaged velocity (u ). Knowledge of local hydraulics prior to instream construction would allow for adjustments and potential optimization of design. The hydraulic outputs from 2D models can be used to identify habitat characteristics for a system - thus allowing both physical and biological contributions in the designed restoration works.  The purpose of this study is to provide a methodology, using a 2D flow model, to assess the morphological and hydraulic effects of instream habitat structures (including LWD). The study investigated the use of a 2D hydraulic model, River2D, to determine local values of y and u at a side channel of the Chilliwack River, British Columbia. The future application of this work is to provide a means of assessing and optimizing planned instream habitat structures. Chapter 2 contains a review of previous literature on channel hydraulics, flow modelling, natural LWD,  Application of a hydrodynamic model for the assessment and design of instream works  Page 4  instream structures, fish habitat and habitat assessment procedures. In Chapter 3 the history, morphology, hydrology and fish biology of the selected field site, the Tamihi side channel, are discussed. Chapter 4 contains the field bathymetric survey results, the model verification results, and the hydraulic and habitat results calculated by the River2D model. Chapter 5 contains a summary of the study results and the methodology for the studies generalized application.  Application of a hydrodynamic model for the assessment and design of instream works  2  BACKGROUND  2.1  Channel Hydraulics  Page 5  Channel hydraulics can be simplified by investigating only one or two dimensions (instead of three dimensions). In one-dimensional (ID) analysis cross-sectional parameters are averaged producing representative values in ID (usually the longitudinal direction). The Manning equation is a widely used ID empirical resistance equation. The equation relates the cross-sectional average depth (y), cross-sectional area (A), flow (Q), and bed slope (S ) in the 0  following relationship (wide channel form) (SI units):  Q = -Ay"sJ n  (2.1)  where, n  = Manning roughness value  The Manning equation, like other ID equations, is used to give simple parameter approximations, which include flow, velocity, flow depth and flow resistance. Equation 2.1 was simplified by assuming the channel is wide enough that flow can be considered the same as a rectangular channel of infinite width. The wide channel approximation can be used when the bankfull width (w ) exceeds 10 y (Chow 1959:p. 27). b  While the information obtained through ID analysis is restricted to average cross-sectional values, two-dimensional (2D) analysis allows for the calculation of local flow parameters across the channel. Local hydraulics such as bed shear stress (t ) and depth-averaged velocity (u ) 0  (which are ignored under ID analysis) play important rolls in determining channel morphology. The following equations allow for the determination of local hydraulics and were used in this study for determining local T . Shear velocity («*) can be calculated from the Keulegan equation 0  (SI units):  ^ r = 5.75 log u  I2.2y  (2.2)  Application of a hydrodynamic model for the assessment and design of instream works  Page 6  where, k p  = equivalent roughness height = density of water (1000 kg/m )  s  3  (2-3)  "* =J^  Bray (1980) derived the following empirical relationship between the equivalent roughness height and the mean grain size (D ): 50  *,=6.8D  (2.4)  50  Substituting Equation 2.3 and Equation 2.4 into Equation 2.2 results in a definition of x in terms 0  of y and u (SI units):  ^  r =  (2.5)  T  5.75 log \2.2y 6.5D  50  In the 1930's, Shields investigated the threshold of motion through the application of shear stress. Shields experimental works developed values of critical shear stress ( T  c r j t  )  in relation to  the particle Reynolds number (Re*). When Re* values are in the turbulent flow range, as is the case in most gravel bed rivers (Carling 1996), Shields found that dimensionless critical shear stress ( T ,  c r i t  )  values were constant at 0.056. The value of  T.  c r i t  can be dimensionalised by the  following equation (SI units):  ^cnt=WAo(^-l)  (2-6)  where, y S  = specific weight of water (9800 N/m ) = specific gravity of particle (2.65) 3  s  Some studies on gravel bed rivers have indicated that the value of Shields  T»  c r i t  ,  which was based  on uniform grain experiments, can be greater than 0.1 due to cluster formation and imbrication  Application of a hydrodynamic model for the assessment and design of instream works  Page 7  between stones which would resist entrainment (Church et al 1998). In non-uniform bed material x.  crit  is also dependent on: local hydraulic conditions, grain size, grain shape and relative  protrusion. The mean value reported by Church et al (1998) is T,  crit  = 0.079 (for widely graded  surfaces). A study by Shvidchenko and Pender (2000) indicated the relative depth of flow also influences  T»  c r i t  .  Reorganization of bed material takes place after a fully mobile bed event has  occurred. Experiments indicated that stone clusters and stone lines form during periods lower than T  c r i t  and add greatly to bed stability during periods of higher than T  Once the t  crit  c r j t  (Church et al 1998).  threshold has been exceeded, scour occurs and increases with applied stress  (Carling 1996). A shear stress intensity (x. ), defined as the ratio of x /T , > 1.0 indicates the int  0  crit  threshold for entrainment has been exceeded. 2.1.1  Flow Modelling  Shear stress and scour processes can be determined at varying degrees of complexity through the use of numerical models. In general, one-dimensional (ID) models are appropriate for use in calculating general aggradation or degradation using available scour equations (Melville and Coleman 2000), while two-dimensional (2D) models can be used to address localised scour and morphological issues. Melville and Coleman (2000) investigated previous work on ID and 2D models and found that the models used, were heavily data dependent - where data was essential for each of the model development, calibration, verification, and implementation stages.  ID models simulate flow conditions in one direction (generally longitudinally) while averaging velocity and depth values over the entire cross-section (producing and u ). These models are c  generally used for determining flood stage at various reach locations and can only be applied when changes in cross-sectional area and hydraulic roughness are gradual. ID models are effective for large-scale flow variation calculations (Melville and Coleman 2000) but do not offer the resolution required for detailed, localised applications.  Two-dimensional models are much more useful for characterizing river hydraulic behaviour. The majority of 2D models are depth-averaged and compute lateral variations in depth and velocity. While 2D models have become relatively sophisticated, the models have not been extensively generalised to the case of an erodible bed (Melville and Coleman 2000).  Application  of a hydrodynamic model for the assessment and design of instream works  Page 8  Depth-averaged models assume uniform vertical velocity distribution and thus are unable to simulate spiralling flow and flow curvature. All 2D models are based on solving the mass (continuity) and momentum equations (Steffler 1999). The governing differential equations can be solved using finite difference, finite element or finite volume methods: finite element methods offer the best geometric flexibility and finite volume methods offer the best stability and efficiency (Steffler 1999). The finite element method offers the flexibility of allowing complex boundaries to be traced by changing element shape and size. Jia and Alonso (1994) used a case study investigated the difference between ID and 2D models. The 2D model used in their study was CCHE-2D, a finite element model based on the Saint-Venant continuity and momentum equations. Jia and Alonso (1994) noticed that the qualitative aspects of stream flow were similar in the ID and 2D model results, yet significant quantitative differences existed. The 2D model was able to calculate lateral flow currents and local hydraulics around every grid point and was therefore more applicable for simulation of intense cross flows, scour and deposition around structures.  Enggrob and von Lany (1994) investigated the use of a 2D curvilinear model to describe scour around river bends and at river confluences. The 2D model used was System21 developed by the Danish Hydraulic Institute. The model solves vertically integrated (depth-averaged) Saint-Venant equations of continuity and conservation of momentum. Comparison of bend and confluence scour with field data showed a great deal of scatter, yet the mean scour depth simulated in the model corresponded closely to the mean maximum scour depth measured in the field. Sensitivity analysis indicated significant sensitivity to lateral distribution of discharge, grain size distribution, eddy viscosity, and upstream boundary conditions.  A quasi 3D model was used by Lisle et al (2000) to examine the relationship between modeled shear stress and measured sediment size. The flow model calculated depth-average velocities (as a 2D model) and used a subsequent subroutine to calculate the vertical structure. While local values of x and D 0  50  were poorly correlated, the approach lends credibility to the use of 2D or 3D  flow models to predict local hydraulics.  Application of a hydrodynamic model for the assessment and design of instream works  Page 9  River2D is a two-dimensional, depth-averaged, finite element model intended for use on natural streams due to its accommodation of supercritical/subcritical flow transitions and variable wetted area. Input data for River2D is channel bed bathymetry, roughness and transverse eddy viscosity distributions, boundary conditions and initial flow conditions. The model creates a finite element mesh composed of triangular polygons (of varying shapes and sizes) for the hydrodynamic component input. River2D is based on the Saint-Venant equations for conservation of mass and momentum, which are used along with the Petrov-Galerkin implicit method to solve depth and discharge intensities in the lateral (x) and longitudinal (y) directions. The conservation of mass formula is:  M  +  ^  dt  ^  +  L  =  dx  dy  PL t q q  7  (2.7)  0  where, flow depth = time = x-discharge intensity =j/-discharge intensity  x  y  The conservation of momentum in the x-direction:  —— + — (Uq ) + (Vqf) + ——H dt dx dy 2 dx  1  K  = H(S -SJ g  m  P  +  (2.8)  1f dx  +—  P  The conservation of momentum in the y-direction:  dt  dx  dy  y  = gH(S -S ) oy  {y  y  +-  2 dy i +—  (2.9)  r  p where,  g  gravitational acceleration (9.81 m /s) 2  Application of a hydrodynamic model for the assessment and design of instream works Page 10  U, V S, S S , S m  {K  oy  {y  = x- andjy-depth averaged velocity -x- and^-bed slope = x- and >>-friction slope  Txx' T xy? w  r , x = horizontal turbulent stress tensors n  M  River2D has 3 basic assumptions: the pressure distribution in the vertical direction is hydrostatic; the distribution of horizontal velocities over the depth is constant; and coriolis and wind forces are neglected. Momentum dispersion closure is made by weighted eddy viscosity coefficient. The habitat component of the River2D uses modelled velocity, and depth values to determine the weighted usable area (WUA) for a specific fish species. Species preference is determined using the Bovee (1978) preference curves. Both the W U A and the preference curves are discussed more thoroughly in Section 2.6. 2.2  Natural L W D  L W D hydraulically controls the downstream movement of debris, sediments and organic nutrients causing local scour and deposition. The definition of L W D varies between studies, but generally pieces are < 0.10 m - 0.15 m and < 1.0 - 3.0 m, in diameter and length, respectively (Keller and Tally 1979; Bilby and Ward 1989; Cherry and Beschta 1989; Ralph et al 1994; Montgomery et al 1995). Specifying a minimum diameter proves problematic, as the significance of L W D is dependent on stream size (Gippel et al 1992). Small L W D becomes less important as river size increases (Gippel et al 1992). L W D size increases with increasing stream order (Bilby and Ward 1989; Ralph et al 1994) while the frequency of L W D decreases due to increased river transport capacity (Keller and Tally 1979; Bilby and Ward 1989; Montgomery et al 1995)(contrary to Robison and Beschta (1990)).  The interactions of L W D with the stream channel vary significantly with stream order. L W D supplied to 1 and 2 st  nd  order streams is suspended above the channel and only through breakage  does interaction with the stream take place (Nakamura and Swanson 1993). L W D either forms effective cross-channel structures (depending on length) or are swung to the side by flow and offer bank protection. L W D in higher order systems mainly deflects the thalweg, which results in lateral scouring and deposition of sediments (Nakamura and Swanson 1993). Higher order  Application of a hydrodynamic model for the assessment and design of instream works Page 11  streams conglomerate L W D pieces together and form jams (Keller and Tally 1979) that are unevenly distributed (Gippel et al 1992). L W D is predominantly delivered to low-order stream via debris slides, debris flows and windthrow, and deposits in unevenly distributed clumps (Gippel et al 1992) rather than alone. L W D is supplied to 3 order and higher streams mainly rd  through treefall and occasionally pieces floating down from upstream. 2.2.1 Sedimentation Investigations of sediment storage by L W D indicate an inverse relationship to stream size (Bilby and Ward 1989) and a direct relationship with L W D loading (Montgomery et al 1995). In headward systems, coarse sediments remain in storage for long periods of time behind L W D and are remobilized only during extreme events (after degradation or collapse of L W D pieces) (Mosley 1981; Nakamura and Swanson 1993). Remobilized sediments travel only short distances before being re-deposited behind downstream L W D (Mosley 1981; Nakamura and Swanson 1993). Rates of transport were found to be entirely dependent on sediment availability. L W D offers important storage sites for, bed sediments, which help regulate bed material load transport (Keller and Tally 1979). Keller and Tally (1979) found debris-stored sediment accounted for 40% of total channel area in higher order streams. Sediment accumulations in 3  rd  and 4 order streams by individual L W D are small relative to LWD-jams (Nakamura and th  Swanson 1993). LWD-jams in these systems cause widening (upstream) and steepening (downstream); yet at higher stream orders LWD-jams (and subsequent sediment storage) are rare due to increased stream competence. Klein (1984) investigated the effects of removing road fill from stream crossings and found where sufficient L W D was not present downstream degradation and a paved bed occurred. The paved bed was not observed in streams that had sufficient LWD. Structural L W D steps developed early after initial disturbance and remained relatively stable throughout the study period (Klein 1984). The field investigations conducted by Lisle (1986) found that stream obstructions caused the deposition of bars 3 w to 4 w downstream and 1 w upstream. The bars formed on the b  b  h  obstruction side of the channel or in the case of bends - on the outside bank of the bend. Upstream deposition of material occurred due to backwater reductions in stream power, while downstream deposition was caused by flow expansion. Heede (1972) observed the average  Application of a hydrodynamic model for the assessment and design of instream works Page 12  spacing in headward systems between bars and L W D to be 1 w to 4 w (results were not given as h  b  ratios to bankfull width: estimations were made using the data provided). Heede (1972) suggests that when frequency of L W D recruitment is insufficient for the required energy reduction, gravel bars form instead. 2.2.2  Orientation  Perpendicularly oriented (or channel spanning) L W D are more frequent in smaller streams, while orientation in larger streams is more commonly in the downstream direction (Bilby and Ward 1989; Robison and Beschta 1990). Gippel et al (1992), as well, found L W D orientation was downstream resulting in small projected cross-sectional area. Bilby and Ward (1989) observed orientation parallel to flow to be similar in all stream orders, while upstream oriented L W D was thought to be unstable and was found infrequently in all stream orders. L W D oriented either oblique or perpendicular to flow has the greatest influence on pool formation (Montgomery et al 1995), while Bilby and Ward (1989) did not find orientation to have much effect on sediment entrapment.  Gippel et al (1992) observed the median L W D angle of tilt from horizontal to be 5° in higher order streams. Blockage ratios were low with the average L W D occupying only 0.4% of the channel while the largest pieces occupied 10% of the channel. As with Nakamura and Swanson (1993), Gippel et al (1992) observed only large accumulations consisting of multiple-LWD caused significant backwater effects. Furthermore, most L W D at angles of less than 40° (with the downstream bank) were hydraulically benign. Cherry and Beschta (1989) conducted flume experiments investigating how the orientation of L W D affects scour. Wooden dowels were used as scale representations of L W D . When dowels were positioned flat on the bed, scour usually occurred along the full length of the dowels with maximum scour depth occurring near the middle. Upstream oriented dowels were found to induce large scour holes and conduct water towards the stream bank, while the scour holes produced by downstream oriented dowels were less extensive and projected water away from the bank. When one end of the dowel was elevated off the bed (assuming it rested on a bank),the maximum scour areas moved out towards the tip. Extrapolation of the flume study indicates that  Application of a hydrodynamic model for the assessment and design of instream works Page 13  partial elevation of the L W D produces more localised scour than positioning the L W D on the channel bed. Upstream orientations caused major flow disturbances; produced relatively large scour holes; and appeared to increase the potential for stream bank erosion, while perpendicular and downstream orientations produced lesser disturbances and were thought to be more stable (Cherry and Beschta 1989). The flow depths in the Cherry and Beschta (1989) experiments were only slightly greater than the diameter of the dowels; and therefore, flow patterns that develop as a result of significant structure burial were not investigated (and may differ considerably). Upstream oriented dowels can be likened to one-sided upstream V-weirs, which are used in restoration practices to scour out holes in the middle of the channel (Allan and Lowe 1997). Flow under the design conditions (high stage) is not directed towards the bank, but towards the middle of the channel instead. By this comparison, L W D oriented upstream in a river may serve to protect banks at higher flow rather than erode them.  2.2.3  Pool type, spacing and formation  Pool types were investigated and classified by Bisson et«/ (1981). Backwater pools are formed by eddies of large obstructions such as rootwads or boulders (they characteristically have low velocities and accumulate fine-grained sediments). Plunge pools occur when flow passes over a complete or nearly complete channel obstruction (water is plunged vertically into the streambed below, scouring out a depression that often results in a large deep pool with variable substrate size). Lateral scour pools occur when flow is directed to one side of the steam by a partial channel obstruction and often result in eroded banks. Trench pools are characteristically U-shaped scour holes with bedrock controls. Dammed pools consist of impounded water upstream of complete or nearly complete channel obstructions (accumulations of fine gravels and sand occur in the pools with characteristically low water velocities).  Free-formed pool-riffle channels are rare in forested environments as L W D forced pools dominate (Montgomery et al 1995). L W D loading is directly proportional to pool frequency (Keller and Tally 1979; Robison and Beschta 1990; Montgomery et al 1995) and L W D size is proportional to scour pool surface area (Bilby and Ward 1989). In forest environments, in  Application of a hydrodynamic model for the assessment and design of instream works Page 14  contrast to free-formed pools in alluvial reaches, local scour caused by L W D can form more than one pool in a single cross-section (Montgomery et al 1995). Montgomery et al (1995) found that L W D forced the creation of 73% of channel pools (n = 471), while Keller and Tally (1979) found 50% of pools were associated with L W D . Typically, up to 40% of the in-channel debris exerts a dominant influence on pool formation (Montgomery et al 1995). Mosley's (1981) investigations of headward streams indicate L W D is responsible for the location of riffles, pools and gravel bars in 40% of the sample reach. Robison and Beschta (1990) observed pool types to vary with stream size: deflector and under-scour pools comprised 10% and 17%, respectively, in smaller streams; and increased to 22% and 27%, respectively, in the largest stream investigated. As stream size increases pool morphology becomes more complex (Robison and Beschta 1990). L W D in smaller streams is more likely to span the entire channel and cause plunge pools (also local widening). In medium to larger streams (although cross channel spanning is still possible) L W D is more easily moved by flow and is therefore pushed diagonally or jammed up against other pieces forming more complex structures and lateral scour pools (Bilby and Ward 1989; Robison and Beschta 1990). Abbe and Montgomery (1996) observed that LWD-jams strongly influence the formation of scour pools and bars in large alluvial channels and were associated with 70% of all pools.  Pool frequency is also directly related to w (Robison and Beschta 1990; Montgomery et al b  1995). Pool frequency in 1 and 2 order channels are characteristically 2 w to 4 w while st  nd  h  b  lengths in 3 and 4 order channels are often 4 w to 7 w (Church 1992; Keller and Tally 1979). rd  th  b  b  Robison and Beschta (1990) found the percentage channel length comprised of pools increased from 45% to 65% from small to larger streams (n = 5). Heede (1972) observed the average spacing between individual L W D to be 2 w to 10 w (results were not given as ratios to bankfull b  b  width, estimations were made using the data provided). Pool-riffle sequences in intermediate streams decreased from greater than 13 w to less than 1.0 w with increasing L W D supply b  b  (Montgomery et al 1995). Montgomery et al (1995) suggest that the 4 w to 7 w spacing b  b  suggested in the literature may represent a morphology without obstructions such as LWD. They found even very low debris loadings (< 0.03 pieces/m ) resulted in pool spacing of 2 w to 4 w . 2  b  b  Application of a hydrodynamic model for the assessment and design of instream works Page 15  2.3  Hydraulics of L W D  Obstructions such as L W D are of interest to geomorphologists because they are commonly formed or introduced by non-fluvial processes and can independently affect stream behaviour. L W D removal eliminates or reduces zones of near zero velocity and causes zones of highest velocity to shift towards the centre of the channel. This reduces the capacity of the channel to dissipate the hydraulic energy (Hartman and Scrivener 1990) and causes bank erosion and headward-progressing degradation (Shields and Gippel 1995). 2.3.1  Flow resistance  Shields and Gippel (1995) investigated the effects of L W D on flow resistance by analyzing streams before and after L W D removal. The method used by the study split the Darcy-Weisbach friction factor into four roughness components: grain; bar and bed form; meander bend; and L W D . In their procedure, the cumulative effects of L W D were distributed evenly throughout the reach (a gross simplification of local hydraulic characteristics). At high flows, friction factors for cleared and uncleared reaches converged (indicating the effects of L W D are drowned out (or buried) with increasing stage). Energy dissipation through flow-contraction and pool-formation processes was gradually eliminated as flows increased in L W D streams (Shields and Gippel 1995). The removal of L W D caused and overall friction factor decreased of 20% to 30%, while flow capacity increased by 5% to 20%. Gippel et al (1992) suggested that increased channel capacity due to L W D removal may not be due to decreased roughness, but rather enlargement of the channel due to bed scour.  Gippel et al (1992) conducted laboratory experiments on horizontal cylinders (representing LWD) and found that blockage ratio, orientation to flow and shielding effect were the most important factors affecting flow resistance. Flow resistance caused by the objects was determined by calculating the drag coefficient (C ) from the measured drag force (F ) given the D  D  following equation (SI units):  (2.10)  Application of a hydrodynamic model for the assessment and design of instream works Page 16  where, l <j> 0  0  = object projected length = object diameter  The calculated C value was high when oriented parallel to flow due to the flow disturbance D  caused by the blunt end of the cylinder. As the cylinder was rotated from parallel to perpendicular an initial sharp decrease followed by a subsequent increase in C was calculated. D  Less variation with rotation was calculated when branches were attached to the cylinder. A n empirical relationship (SI units) for calculating C for a single cylinder of known length was D  presented: /  C  D  n 0.062  =0.81  (2.11)  When a cylinder was placed close to the bed surface it created the highest C values, which D  decreased as the cylinder was raised towards the water surface. This variation occurred since, close to the bed, a zone of low velocity develops upstream of the cylinder disturbing the turbulent wake. Gippel et al (1992) found that the blockage ratio (defined as the cross-sectional area obstructed over the total cross-sectional area) was the most important parameter affecting C . Blockage ratios were related to C by an exponential relationship and measurable effects D  D  were observed when blockage ratios were above 0.1.  2.3.2  Scour  Often hydraulic studies conducted on other types of structures like piers or rock groynes offer useful incite into the potential effects of L W D (Cherry and Beschta 1989), but with respect to bridge piers results have to be extrapolated to the case of complex horizontal cylinders (Gippel et al 1992). General scour occurs as a result of upstream conditions such as increased flow or decreased sediment load, whereas local scour occurs below the equilibrium bed elevation due to some localised obstruction. Local scour in rivers is generally the result of flow concentration and secondary currents or vortices that occur as a result of river features such as bends, impingements, constrictions, confluences and local obstructions (Galay et al 1987). The extent  Application of a hydrodynamic model for the assessment and design of instream works Page 17  of the local scour hole formed is dependent on the strength of the secondary currents developed due to the upstream obstruction or impingement.  Bend scour occurs as flow curves around the outside of a bend. Secondary currents are formed which are forced down towards the bed accelerating flow and causing higher shear stresses. Flow is in the form of a spiral or corkscrew and decelerates as it approaches the inside bank where deposition occurs. Contraction scour occurs when the channel is restricted either through width or depth and flow is concentrated and forced through a smaller area. Confluence scour occurs at the junction of two converging channels where flow is directed towards the channel bed and returns to the water surface along the sides of the confluence (Melville and Coleman 2000). The resulting helicoidal secondary currents create a deep scour hole.  Bridge pier scour has been extensively studied over the past few decades (Laursen 1963; Galay et al 1987; Melville and Coleman 2000) for the purpose of preventing bridge failures. Flow patterns undergo significant changes as a result of piers (Melville and Coleman 2000). Upstream flow undergoes separation and rolls down to form a vortex system in front of the pier caused by flow deceleration and pressure gradients. The sides of the vortex are swept downstream by flow giving the system a horseshoe appearance. The depth of scour has been found to be strongly dependent on the width of the pier (Melville and Coleman 2000).  Abutments, spurs and rock groynes have similar scour processes as piers. The principal features comprised of surface roller, downflow, principal vortex and wake vortices are similarly present (Melville and Coleman 2000). In addition there is a large reversed eddy generated in front of the obstruction near the bank (Melville and Coleman 2000). Figure 2.1 and Figure 2.2 developed by Kwan (1984) and reproduced in Melville and Coleman (2000) illustrate the local hydraulic effects of an abutment. The deepest scour hole seen in Figure 2.2 is located at the apex of the abutment. This is consistent with Galay et al (1987) who observed the deepest scour hole to be associated with strong secondary currents located at the nose or apex of the obstruction.  Application  of a hydrodynamic model for the assessment and design of instream works  Page 18  Figure 2.1: Illustration of the flow and scour patterns at an abutment (Kwan 1984 cited by Melville and Coleman 2000).  Application of a hydrodynamic model for the assessment and design of instream works Page 19  Garde et al (1961) conducted sand flume experiments of scour around spur-dikes perpendicular to flow for a range of opening ratios. Spur-dikes restrict channel width, thereby altering flow patterns and inducing scour. Maximum scour was significantly affected by sediment size. Scour tended to extend downstream from the tip of the dike toward the opposite bank. Maximum d  s  occurred at the outer tip (or apex) of the dike. Garde et al (1961) developed an empirical equation to predict d given y , u , w and spur-dike length. The equation, developed for sand, is s  c  h  very sensitive to small grain size differences (0.00029 m to 0.00245 m) and is most likely not applicable to gravel bed rivers. Melville and Coleman (2000) present a mean scour depth equation (SI units) for large gravel rivers (D > 0.002 m) originally developed by Blench (1969): 50  q  1  rf.=1.23  (2.12)  where, discharge intensity The above equation was based on general regime theory and was originally presented by Blench (1969) as two separate equations: a general regime equation, and a "bed-factor" equation. The Blench (1969) equations were not meant to predict local scour, but were instead intended to be "crude tentative formula" for general regime depth calculations. The empirical scour equations developed for bridge piers, groins and abutments, are based on average depth (y) and cross-sectional average velocity ( M ) and are therefore not generally c  applicable for local scour predictions. These equations give maximum scour for the purpose of determining the design depths of pier footings or buried pipes. Local scour, for which predictive equations are not found in the literature, is dependent on local velocity and depth values. Field investigations conducted by Gippel et al (1992) indicate that L W D exerts considerable influence on local hydraulics: creating a wide variety of channel velocities. Erosional zones are associated with large turbulent eddies rather than high velocities. Similarly to abutment flow  Application  of a hydrodynamic model for the assessment and design of instream works  Page 20  patterns stated above, flow against the upstream face of an obstruction produces high velocities near the base forming a large persistent vortex that scours the bed and deflects bedload particles across the channel away from the obstruction (Melville 1975 cited by Lisle 1986). The vortex and associated secondary currents are thereby capable of radically altering sediment transport patterns and completely changing downstream channel morphology (Lisle 1986). In a summary study of previous lab and field work, Smith and Beschta (1994) investigating the hydraulics responsible for L W D pool formation and maintenance. In their flume experiments with cylinders, they found average pool velocities within 0.05 m of the bed were only half the average velocities outside the pool. The time derivative of velocity (used as an index of turbulence) was 1.4 to 1.8 times greater inside the pools than outside. Gravel particles (D = 0.0054 m) were observed to lift vertically and subsequently saltate downstream. Field 50  results indicated near-bed velocity increased linearly with discharge, and velocities measured in the deepest parts of the pool did not exceed the pool head or tail velocities. Pool cross-sectional area remained relatively constant over the range of subjected flows. Obstruction related pools anchor the location of scouring vortices and turbulent velocity fluctuations, creating more stable pool morphology than in non-obstruction related pools (Smith and Beschta 1994). Smith and Beschta (1994) found that scour and deposition in their study pool occurred in pulses at discharges well below bankfull on both the rising and falling hydrograph limbs. Although mean shear stress in the pool was lower than in the surrounding areas, bedload rate, competence, and grain-size distribution were not significantly different from those at the pool head or tail. Smith and Beschta (1994) believe that turbulent lift and drag forces may be supplementing the mid-pool shear stress. Velocity reversal theory (where velocities in pools are lower than riffles at low flows and become greater than in riffles at high flows - scouring out any deposited sediment) does not account for pool formation or maintenance in obstruction-related pools (Smith and Beschta 1994).  Lisle (1986) observed scour holes with steep upstream slip-faces located adjacent to all obstructions (the deepest scour was located just downstream of the farthest obstruction projection). Scour holes gradually shoaled and widened downstream. The total volume of material mobilized from the bars accounted for a large proportion of the total bedload transported  Application of a hydrodynamic model for the assessment and design of instream works Page 21  during storm events. Although sediment transport rates were determined to be sufficient enough to cause major bar changes, observed changes were solely associated with movement of L W D and flow approach angel. The bars therefore appeared to be in dynamic equilibrium with bedload transport and the obstructions that formed them. Results indicate that obstructions wider than 1/3 w formed channel spanning pools while smaller obstructions formed scour holes within b  a single bar. When obstructions were spaced too close (less than 2 w ) pools were less likely to b  form. When spacing between obstructions was greater than 5 w , diagonal bars formed b  downstream with their location and length determined by the obstruction. Obstructions that occur at a minimum frequency ensure channel stability and arrest migrating bar movement (Lisle 1986). Lisle (1986) observed water surface superelevation against some of the obstructions with transverse water surface gradients increasing with discharge. These gradients in some cases were greater than the channel slope and generated strong secondary currents. Scour was associated with these large-scale circulation patterns. Approach angles appeared to be less important than the blockage ratio in establishing pools and scour holes. Lisle (1986) developed a series of empirical (dimensionless) equations predicting scour hole width, length and angle: \ 0.606  w. w„ ^ = 0.616 V  ^2-=  w  1.41 V  h  sin P  W„ W  = 1.05  b  w  (2.13) J  b  v 0.548  (2.14) J str  0.0279  (2.15)  sin a where,  ^scr P a  = = = = =  structure or bend width scour hole width (perpendicular axis at widest structure point) scour hole length (widest point of structure to downstream terminus) deflection angle of scour hole from parallel with banks angle of upstream side of structure with upstream flow lines  Christner and Custer (1998) investigated the Lisle (1986) scour prediction equations and found  Application of a hydrodynamic model for the assessment and design of instream works Page 22  them site dependent. For the most part, Christner and Custer (1998) found poor correlations between the Lisle (1986) relationships.  Abbe and Montgomery (1996) investigated flow around bar apex jams (BAJ) (which are essentially the same as single-LWD structures with rootwads (where the rootwad faces upstream) with additional L W D members caught up on the upstream side). Several alluvial bedforms were associated with these structures: a crescentric pool (directly upstream the LWD) due to vortex development; a semi-parabolic bar (upstream of the vortex) formed by flow deceleration; and an elliptical depositional zone (around the L W D trunk) formed by flow separation and re-circulation. The upstream vortex (which is similar to pier vortices) is downward circulating and develops as a result of the translation of predominantly horizontal flow to downward directed vertical flow. Accumulation of static pressure directly upstream of the L W D leads to flow reversal near the bed and formation of a point of zero velocity some distance directly upstream from the L W D . Flow is constricted adjacent to the L W D rootwad causing flow to accelerate around the sides which results in flow separation along the sides. Downstream drag occurs accounting for energy losses due to pressure gradients, skin friction and flow separation.  Abbe and Montgomery (1996) investigated an empirical constriction scour equation (Imperial units) developed by Laursen (1963) and found total predicted scour was within 3% and 17% of observed pool depth for two structures investigated. The Laursen (1963) equation is as follows: 0.429  V crt J T  ( l co„ J W  where, w  con  = constricted width of channel  As with the empirical equations devised for bridge piers and abutments, the scour equations developed by Laursen (1963) and Abbe and Montgomery (1996) are based on u and c  y conditions upstream and/or downstream of the obstruction and do not adequately represent local hydraulic effects.  Application of a hydrodynamic model for the assessment and design of instream works Page 23  2.4  L W D Structures  Deep-water habitat can be simply created by excavating holes in the streambed, but these pools would fill in over time (Allan and Lowe 1997). Instream structures attempt to duplicate the natural process of hydraulically controlled pool formation by promoting and maintaining scour (Allan and Lowe 1997). Allan and Lowe (1997) recommend opposing rock deflectors (double rock groynes) as the most suitable alternative to create deep-water runs and pools. Paired opposing deflectors have the advantage of directing flow to the centre of the channel and reducing erosion risk to the adjacent banks (Allan and Lowe 1997). Tripp (1986) investigated the morphological effect of introducing 13 L W D structures (primarily cross-stream placements) to a debris torrented stream and found the structures greatly increased morphological variability in the channel. Within the first year, many different types of pools were created including under scour pools and backwater pools. After the second year, most of the pool types were either lateral scour or plunge (which would be predicted by cross-stream placements). The change in pool types occurred as sediments and debris deposited upstream of the structures. Total pools area increased from 2.5% to 15.4% two years after the instillation.  2.4.1  Durability  Durability is an important criterion in assessing the effectiveness of L W D structures. Instream restoration structures that fail (moved, buried or broken apart such that they no longer influence flow) within 5 years of their installation may have minimal stream benefits (Roper et al 1998). Keeping structures in place and useful after high flows has proven to be difficult (even in stable reaches). Often the hydraulic stability and function of these structures cannot be ascertained until they have undergone one or more periods of high flow (Cherry and Beschta 1989). In general, success rates have been found to decreased with increasing return periods and stream order (Roper et al 1998; Frissell and Nawa 1992).  Roper et al (1998) analyzed the durability of almost 4000 structures in 94 streams located in the Pacific Northwest. A l l structures had undergone floods with return periods exceeding 5 years. The overall success rate was over 80%, yet this included structures that had shifted but remained on site. Specific types of structures were not discussed. If shifted structures are not included as  Application of a hydrodynamic model for the assessment and design of instream works Page 24  successes, the rate reduces to 63%. Roper et al (1998) found that structures in large streams were up to 20 times more likely to fail than structures in small streams. Furthermore 63% of structures in 6 order streams failed, while only 3% failed in 1 order streams. th  st  Frissell and Nawa (1992) investigated 161 structures from 16 streams (mostly low order) from the Pacific Northwest. Overall success rate was 68%, which included damaged (or impaired) structures. Success rate was reduced to 38% if damaged structures were discounted. The Frissell and Nawa (1992) study found much lower success rates than Roper et al (1998). This may be a result of definitional and flood return period differences, which makes strict comparisons between the two studies difficult. The Roper et al (1998) study investigated a much broader range of flows; and the rivers investigated by Frissell and Nawa (1992) were highly unstable. The highest rates of structure failure observed by Frissell and Nawa (1992) occurred in watersheds severely damaged by roads, logging, and landslides that delivered very high sediment loads to the channel. Structure failure was found to occur due to cables and anchoring device failures, bank erosion, lateral channel movement, and burial (especially in low gradient reaches). Lateral L W D deflectors and cross-stream L W D weirs had high rates of failure (>50%). Structural failure (cabling or ballasting failure) appears to be far less prevalent than failures resulting from watershed driven channel changes. Frissell and Nawa (1992) believe it is unrealistic to expect L W D structures to stabilize channels with basin wide instabilities.  An extreme event during the first year of the Tripp (1986) study resulted in the partial or entire filling of many of the pools with sediments. Most of the pools were re-scoured, but L W D structures at some locations were isolated by channel shift. Three years after the study was initiated, a debris torrent occurred which destroyed all but 3 of the L W D structures - bringing into question the effectiveness of restoration under such unstable conditions. Yet Tripp (1986) suggests that waiting until the stream channel settles before initiating restoration measures increases the risk of loosing fish populations altogether.  Only 3 out of the 13 structures installed by Tripp (1986) had some amount of cabling, and these were the only structures left after the last debris torrent. Anchored structures were found to be more durable (Tripp 1986; Roper et al 1998). Roper et al (1998) found that structures not  Application of a hydrodynamic model for the assessment and design of instream works Page 25  connected to the channel edge were 50% more likely to fail. In low order streams, structures that spanned the entire channel (perpendicular to flow) were found to be more effective (Roper et al 1998, Tripp 1986) as they were less likely to become isolated through channel meanderings. Guidelines for ballasting L W D structures, using hydrodynamic force balances, were developed by D'Aoust and Millar (1999). Two types of structures were investigated: single-LWD structures which consist of a L W D (with or without rootwad) ballasted at both ends by boulders using steel cables; and multiple-LWD structures which are triangular structures (V-shaped with the apex directed out into the channel) ballasted on the bank and at the apex with boulders and steel cables). The observed stability of the single-LWD structures (with and without rootwads) agreed well with predicted stability analysis where all structures predicted to not-move remained in place, while predicted movement occurred in 11 of the 13 unstable structures. The multiple-LWD structure results were less predictable. It was assumed that sliding would not occur as the structures were sufficiently braced; therefore stability of these structures was only dependent on buoyancy. Movement occurred in 6 of the 21 triangular structures predicted to be stable, while predicted movement occurred in 8 of the 13 unstable triangular structures.  2.5  Fish Habitat  Numerous factors influence site productivity - these include: pool depth, local velocity, food availability, structural complexity, riparian vegetation, and substrate size (Blackwell et al 1999). L W D provides important habitat for direct use by aquatic and terrestrial organisms. Uses include shelter form fast flows, shade, feeding sites, spawning sites, nursery areas for larvae and juvenile fish, territory markers and refuge from predators (Rutherfurd et al 1999).  L W D influences pool frequency and type; and therefore has a large effect on fish habitat. Deep pools often form around these obstructions and the combination of deep, slow water and good cover offers important wintering areas, particularly for older salmonids (Bustard and Narver 1975). Direct relationships have been found between near-natural L W D structures and salmonid smolt production (Slaney and Martin 1997). Stable sediment sites between pools, created by L W D , represent desirable spawning areas for adult fish (Tripp 1986).  Application of a hydrodynamic model for the assessment and design of instream works Page 26  2.5.1  Cover  Cover refers to instream and overhead features that: protect fish from high current velocities, provide concealment for both predators and prey, and provide cold water refuges by interfering with incoming light (Bovee 1982). Bisson et al (1981) identified eight distinct types of cover. These included three types of woody debris: rootwads, large debris (tree trunks), and small debris (branches and twigs). Also included were: overhanging terrestrial vegetation, undercut banks, water turbulence (where bubbles prevented a clear view through the water column), rocks (which provided overhead ledges and crevices for hiding); and deep water zones (which provided cover from non-diving predators). L W D is the most abundant and diverse cover type in pools (Bisson et al 1981) and provides complex cover and nutrient trapping (Abbe and Montgomery 1996). Bisson et al (1981) found it difficult to correlate fish preference with cover type, as there are many other factors that were not specifically investigated in their study such as: the presence of competitor and predator species, population densities, and food availability. 2.5.2  Coho Salmon habitat  Coho fry migrate to off-channel areas such as side channels, beaver ponds, wetlands, protected alcoves and tributaries for summer and overwintering rearing. Migration into most off-channel habitats in coastal streams and rivers occurs during the fall, winter, and spring when mainchannel discharges are high (Swales and Levings 1989). Yet McMahon and Hartman's (1989) experiments indicate significantly more coho remain in the channel and use L W D when it is provided under simulated freshet conditions. Increased off-channel structural complexity such as woody debris has been shown to increase overwinter survival of coho (Nickelson et al 1991). In general overwintering coho abundance increased with increased cover complexity and overwintering fish are more dependent on cover than those during the summer (McMahon and Hartman 1989). Tripp (1986) showed that the installation of L W D structures in a debris torrented stream resulted in a four-fold increase in coho overwinter survival.  Coho fry seemed to prefer rootwads and L W D abundance in pools; while in riffles they prefer overhanging riparian vegetation and undercut banks (Bisson et al 1981). Coho use of structures  Application of a hydrodynamic model for the assessment and design of instream works Page 27  increases significantly when overhead shade is provided (McMahon and Hartman 1989). The presence of deep pools maximises survival of spring smolt emigration since the smolts tend to conglomerate in deeper areas before migrating in groups downstream (Nickelson et al 1991). 2.5.3  Steelhead trout habitat  Juvenile steelhead are able to occupy a much wider variety of microhabitats while steelhead fry strongly prefer shallow water (less than 0.15 m) close to shore (Hartman 1965; Bustard and Narver 1975). Steelhead are more commonly found in mainstem and tributary streams where they have adapted to surface temperature regimes with relatively high water velocities such as riffles (Swales and Levings 1989). Off-channel habitats are important refuge areas during flood events, yet steelhead are reported to use them much less than coho (Swales and Levings 1989). Yearling steelhead had a mild preference to bank cover in pools, and preferred L W D in riffles. With decreasing water temperatures juvenile steelhead, moved closer to objects providing cover preferring rootwads and L W D (Bisson et al 1981). Rocks and boulder were the principal source of cover for steelhead fry (Bustard and Narver 1975). 2.5.4  Pool use  Bisson et al (1981) found very few coho or steelhead occupied secondary channel pools during the summer. It was thought that these habitats, being isolated from the main channel, had high temperatures and dense algal growth. Coho fry utilized backwater, plunge, and dammed pools (Bisson et al 1981; Nickelson et al 1991). Side pools (which dried up in the summer) provided good coho refuge from severe winterfreshets(Bustard and Narver 1975). Juvenile steelhead preferred riffles, rapids, cascades, plunge pools, lateral scour pools, and dammed pools (Bisson et al 1981).  A summary of the average habitat utilization coefficients for coho salmon and steelhead trout obtained by Bisson et al (1981) is presented in Table 2.1. From the table, it is apparent that coho prefer pools while steelhead fry prefer riffles. Steelhead juveniles appear to prefer all habitat types investigated.  Application of a hydrodynamic model for the assessment and design of instream works Page 28  Habitat type  Coho 0+  Steelhead 0+  Steelhead 1+  Pools Riffles Glides  1.46 -0.90 -0.91  -0.23 0.60 0.34  0.70 0.29 0.86  Table 2.1: Average habitat utilization coefficients for coho salmon and steelhead trout. Negative values indicate avoidance, while positive values indicate preference. 2.5.5  Logging  The removal of L W D within the low-flow, active channel generally decreases habitat quantity and variability required to sustain aquatic species (Ralph et al 1994). Hartman and Scrivener (1990) study of Carnation Creek found survival to emergence of coho salmon decreased from 29% to 16% following logging. Survival declined due to bed aggradation, which prevented fry emergence. Steelhead numbers also declined following logging due to loss of main channel summer and winter habitats (Hartman and Scrivener 1990). Juvenile steelhead preference for main stem habitats makes them more vulnerable to logging impacts than coho (Hartman and Scrivener 1990).  2.6  Habitat Requirements  Methodologies have been developed to quantify instream fish habitat. One widely used method devised by Bovee (1982), is the weighted usable area (WUA) concept used in the development of the physical habitat simulation system (PHABSIM). The method quantifies physical habitat requirements on a unit area basis using habitat indices. Bovee (1978) developed preference curves (probability of use curves) to describe the habitat conditions (depth, velocity, substrate size and temperature) preferred by selected fish species. The Bovee (1978) preference curves for juvenile steelhead trout are presented in Figure 2.3. Preference curves are a simple way of representing very complex or discontinuous mathematical functions describing fish preference (Bovee 1982). Criticism of the preference curves has included inadequate representation of 'nose level' water velocities (0.2 m above the bed for a spawning salmon) and the validity of transferring habitat suitability from the rivers in which they were derived to other systems (Mosley 1982). However, the PHABSIM model is widely used in practice for assessing instream habitat.  Application of a hydrodynamic model for the assessment and design of instream works Page 2  \  \  7T ill  \  i  ••  *,/ •  \  Figure 2.3: Juvenile steelhead preference curves (Bovee 1978). Velocity and depth refer to local (net cross-sectional average) values. The W U A is discharge and species dependent and is calculated as follows:  WUA^=£(cv4)  (2.17)  i=i  where, C  i,s  - suitability index fu fy  Ai  /. fy fc  fsb  fx  = = = = = =  fc  fsb  ft  (5.18)  unit surface area velocity factor depth of flow factor cover factor substrate factor temperature factor  The velocity, depth, cover, substrate and temperature factors range between 0.0 -1.0 and are determined for each unit area. Velocity, depth, substrate and temperature factors are extrapolated from the Bovee (1978) preference curves; while the cover factor is determined subjectively for each given unit area investigated.  Application  of a hydrodynamic model for the assessment and design of instream works  Page 30  Hogan and Church (1989) used probability of use curves and the W U A method in a study of coho salmon habitat in the Queen Charlotte Islands. Different flow scenarios were determined through at-a-station hydraulic geometry. W U A increased initially and subsequently declined significantly as flows increased towards the Q  ]Q  event. Channel characteristics changed from  relatively deep, slow moving water at low flows, through to fast flowing water at high flows (unsuitable for fish). The initial rise in W U A was due to an increase in the total water surface area as stage increased. As the discharge increased still further, there was a drop in available habitat (since velocity increased at a faster rate than depth). Although substrate and temperature preference curves are available (Bovee 1978), for simplicity Hogan and Church (1989) used only velocity and depth curves to demonstrate their technique. They also chose not to include habitat cover. Even though the W U A method may not reflect actual use by fish it remains a useful measure of area potentially available under appropriate management (Hogan and Church 1989).  Mosley (1982) used velocity and depth measurements to investigate the W U A method for braided streams. In the analysis, water temperature, cover and sediment size were assumed to be optimum (as with the Hogan and Church (1989) study) so that habitat changes could be more accurately observed. Contrary to the Hogan and Church (1989) study, W U A remained relatively stable for discharges below g  2 33  . The apparent stability is due to the nature of braided systems.  New-flow channels are created as the channel widens (with increasing discharge and stage) replacing the pre-existing channels, which are drowned out by the increased discharge (Mosley 1982). When discharges rose above Q  232  W U A decreased dramatically.  Both Mosley (1982) and Hogan and Church (1989) studies calculated W U A by extrapolating velocity and depth measurements (taken at a few distinct cross-sections) over the entire reach. This limits the accuracy of predicting W U A since many local reach irregularities, which cause velocity and depth changes, would not be taken into account. The use of local flow parameters with the Bovee (1978) preference curves should improve the accuracy of the calculated WUA, even though the original curves were designed for reach average values. For these reasons, 2D models are better suited for habitat assessment as they provide local u andy values throughout the channel.  Application of a hydrodynamic model for the assessment and design of instream works Page 31  3  T A M I H I SIDE C H A N N E L FIELD SITE  3.1  Site History  Restoration activities within the Chilliwack watershed have been ongoing for the past 15 years (Blackwell et al 1999). Works have been completed by the Department of Fisheries (DFO) and through the Watershed Restoration Program (WRP). During the 1998-1999 budget period, seven WRP projects were in progress. Such projects included 'Bulbeard side channel and ponds' and 'Foley Creek side channel fish habitat restoration'. In the summer of 1999, instream structures were installed at the Tamihi side channel through the WRP. The purpose of the instream structures was to create channel diversity and to provide additional habitat for coho fry and steelhead fry and juveniles. Additional habitat for salmonid spawners was also desired. The Tamihi side channel was selected as a field site based on its density and age of instream structures, as well as, its relative ease of access. The side channel is located (49°04'32" latitude and 121°46'25" longitude) on the Chilliwack River approximately 13.5 km east of Vedder Crossing, halfway between the confluences of Tamihi Creek and Slesse Creek. A site location map is presented on Figure 3.1. Figure 3.2a presents an aerial photograph, taken on May 23 1999, of a section of the Chilliwack River with the Tamihi side channel included at the top of the photo. The Tamihi Bridge located just above the confluence of Tamihi Creek with Chilliwack River can be seen at the bottom of Figure 3.2a. Large scale meanders and wandering braided stretches with frequent gravel bars and islands can be seen in Figure 3.2a. The Chilliwack River is approximately 45 m in width along this reach (Hay and Company Consultants Limited 1992).  An enlarged view of the Tamihi side channel is presented in Figure 3.2b. The side channel has an average w of approximately 15 m and is located on the south side of a stable medial gravel h  bar, which is approximately 400 m in length. The gravel bar forms a sill at the side channel's inlet and thereby controls inflow to the channel. During low flow seasons (late winter and late summer) the water level in the Chilliwack River is below the inlet sill and the resulting flow in the channel is limited to groundwater seepage. The right bank (looking upstream) of the side channel is heavily rip rapped for the full length of the reach to protect the adjacent road from erosion. Historical aerial photos of the area were investigated for the years 1963, 1979, 1983 and  Application of a hydrodynamic model for the assessment and design of instream works Page 32  1993. The side channel and medial bar are not present prior to and including 1983, but are present in 1993 indicating that the side channel has only been in existence for 8 to 18 years.  Figure 3.1: Site map and site location.  Application  of a hydrodynamic model for the assessment and design of instream works  Page 33  b)  Figure 3.2: a) Aerial photograph of the Chilliwack River above Tamihi Bridge. Photo was taken on May 23,1999 and is approximately 1:30 000. The Tamihi side channel is located adjacent to a large medial bar of approximately 400 m in length - seen in the upper portion of the photo, b) Aerial photograph of the Tamihi side channel (enlarged view approximate scale 1:7000).  Application of a hydrodynamic model for the assessment and design of instream works Page 34 3.2  Watershed History  The Chilliwack River is a 5 order system with a drainage area of approximately 1230 km . The th  2  headwaters are located in the Cascade Mountain Range in northwestern Washington and drain north across the border into Chilliwack Lake (southern British Columbia). At the outlet of the lake, the river makes an abrupt change in direction and flows west towards the City of Chilliwack. As it enters the outskirts of the city it passes under Vedder Crossing and is re-named the Vedder River, which joins the Sumas River and flows into the Fraser River approximately 10 km to the west of Vedder Crossing. The Chilliwack use to flow north at Vedder Crossing, but during a late 1800's flood, it avulsed and flowed down Vedder Creek instead. Pressures from urbanization and agriculture have resulted in the dyking of 84% of the Vedder River and the subsequent loss of most of the lower channel habitat (Blackwell et al 1999). The Chilliwack watershed is a dry maritime Coastal Western Hemlock ecosystem. The mainstem of the river is narrow in the upstream reaches and widens downstream of Slesse Creek where it becomes well braided with channel bars and islands (Hay and Company Consultants Limited 1992). The Chilliwack River meanders along the valley floor degrading through glacial and glaciolacustrine sediments exposing bedrock locally (Hay and Company Consultants Limited 1992). Soils and bedrock types within the area are diverse ranging from plutonic to sedimentary and metamorphic rock. Soils within the watershed are composed of deposits of till, glaciofluvial and fluvial materials and colluvium (Zaldokas 1999). The Chilliwack watershed is prone to natural landslides due to glacial clay deposits and studies have indicated that landslide activity has increased as a result of logging roads. The bulk of the landslide materials are wash load sized silts and sands and as a result do not significantly effect channel morphology until downstream of Vedder Crossing (Hay and Company Consultants Limited 1992). Conversely, sediment inputs from tributary streams do have morphological significance (Hay and Company Consultants Limited 1992).  Logging in the Chilliwack River watershed began in the 1800's and by the late 1930's most of the easily accessible timber had been harvested (Hay and Company Consultants Limited 1992). Logging then shifted into the steeper sloped major tributaries where it continues to the present day. The upper Chilliwack River, as well as, the headwaters of Slesse Creek and Tamihi Creek  Application of a hydrodynamic model for the assessment and design of instream works Page 35  are located in the United States in areas classified as either national parks or wilderness areas: where logging is not permitted (Hay and Company Consultants Limited 1992). The total cumulative percent of watershed area logged by 1992 was 20% (Hay and Company Consultants Limited 1992). Many of the side channels below Chilliwack Lake were formed as a result of redirected flow due to semi-permanent log-jams. Streamside logging eliminated most of the L W D recruitment, which maintained the log-jams and as a result many of the historical side channels only receive flow during extreme events (Zaldokas 1999). Roads and dyking have cut across off-channel habitats rendering them impassable to fish through poorly designed culverts. The present populations of wild salmonids in the Chilliwack River are below historical levels due to urban development and logging related habitat changes (Zaldokas 1999). Natural instabilities such as sediment input due to landslides are compounded by the disturbances associated with logging, dyking and roadways (Blackwell et al 1999). Most species of anadromous salmonids are present in the Chilliwack watershed. These include coho, chum, chinook, and pink salmon, as well as, steelhead trout (Zaldokas 1999). Resident rainbow trout, anadromous and resident char and mountain whitefish are also present (Zaldokas 1999). The freshwater residency of coho is typically 1 year, while the majority of steelhead rear for 2 years and some may remain up to 3 years (Hartman 1965; Bustard and Narver 1975; Hay and Company Consultants Limited 1992). Coho salmon spawn in the main stem and tributaries of the Chilliwack River during November to January with the heaviest spawning occurring in the upper reaches near Chilliwack Lake (Hay and Company Consultants Limited 1992). Blackwell et al (1999) found coho outmigration on the Chilliwack River occurred in peaks between May 13 and June 10 while steelhead outmigrated in early May.  3.2  Flood Events  In most intermediate and large systems, channel forming events recur every 1.5 to 2.5 years (Church 1992). The mean annual flood (Q ^, 2  which recurs every 2.33 years, is generally  accepted and often used for river characterization. The bankfull flood (Q ) is defined as the b  discharge at which the channel is completely full. In many cases the mean annual flood is equivalent to a bankfull event. In headward systems, channel forming events occur much more  Application of a hydrodynamic model for the assessment and design of instream works Page 36  infrequently and the Q  23 3  3.2.1  is generally insufficient in size.  Flood frequency analysis of the Chilliwack River  Three gauging stations in the Chilliwack watershed were investigated. Station 08MH101 is located downstream of the site at the Vedder Bridge, station 08MH103 is located upstream of the site (above the confluence with Slesse Creek) and station 08MH056 is located on Slesse Creek just upstream of its confluence with the Chilliwack River. The drainage area of the Chilliwack River adjacent to the Tamihi side channel is 880 km . The area was determined using a 2  1:125 000 topographic map and recorded gauging station areas. Hydrologic maximum daily and instantaneous peak flows from all three gauging stations were downloaded from the Water Survey of Canada's H Y D A T C D - R O M . The maximum monthly and yearly flows were extracted from each data set. The downloaded maximum monthly flows produced a bimodal distribution (two distinct high flows periods): snowmelt events occurred between April and August (spring/summer), while rainfall events usually occurred between October and March (fall/winter). Log Pearson Type III frequency analysis was performed on the maximum monthly flow records from the lower and the combined upper gauging stations using the Consolidated Frequency Analysis (CFA) statistical package (originally distributed by the Surface Water and Sediment Data Atmospheric Environment Service). The Vedder Crossing predicted flows {Q , Q o> and Q ) were adjusted 233  10  200  to the site's drainage area using the following equation presented by Coulson and Obedkoff (1998) (Vedder Crossing values were chosen since the records date back much further than those of the upper two gauging stations): 10.785  e„=e  (3.1)  g  where,  Gu g  4u  Ag  = = = =  ungauged flow gauged flow ungauged watershed area gauged watershed area  The exponent value 0.785 is a constant for the whole Province of British Columbia and was  Application of a hydrodynamic model for the assessment and design of instream works Page 37  determined by plotting and comparing peak flow against drainage area for each hydrologic zone (Coulson and Obedkoff 1998). The daily maximum flood frequency analysis for the Chilliwack River adjacent to the site is presented in Table 3.1.  Chilliwack River (adjacent to site)  02.33 (m /s) 3  £?ioo (m /s) 3  02oo (m /s) 3  spring/summer 180 300 315 fall/winter 205 500 540 Table 3.1: C F A - Log Pearson Type III flood predictions. The predicted fall/winter period peak flows (Q  100  and Q ) were almost double the predicted 200  values of the spring/summer period. The instantaneous daily discharges were evaluated and are related to the maximum daily flows by a peaking factor of 1.46. The instantaneous Q  200  for the  fall/winter period is 790 m /s. Rainfall events were flashy lasting only 1 to 5 days, while 3  snowmelt events were greater in volume and duration (lasting for greater then two weeks). At lower return periods the difference between the two time periods shrunk (especially higher up in the watershed where Q  233  for the combined upper gauging stations showed little difference  between fall/winter and spring/summer periods). Hay and Company Consultants Limited (1992) determined that most extreme floods in the Chilliwack basin occur during the fall/winter period with heavy rain and the subsequent melting of some or all of the early snow pack.  3.2.2 November 1999flood event On November 12, 1999, the Chilliwack River experienced an estimated 5 year return period flood (Q ). The estimated peak discharge adjacent to the site (based on peak flows recorded at 5  the Vedder Bridge gauging station 08MH103) for this event was approximately 290 m /s. As 3  previously mentioned in Section 3.1, flow into the Tamihi side channel is restricted at its entrance by a sill formed by the adjacent medial gravel bar. As a result, only a portion of the predicted fall/winter mean annual flood or Q would pass through the side channel. During this 5  event, flows in the side channel were likely bankfull. The side channel is not gauged; therefore Q had to be predicted. The side channel's Q was calculated using two different methods: the b  b  Manning equation (Equation 2.1) and the two-dimensional flow model - River2D. The Manning's roughness (n) was calculated using measured low-flow u , y , S, and A. The c  c  Application of a hydrodynamic model for the assessment and design of instream works Page 38  bankfull average cross-sectional depth (y ) was determined by examining the cross-sectional b  profile at-a-station and predicting the over bank flow depth. This value was compared to the observed high flow mark elevation (where flood-debris was caught up on the sides of the channel) (the two values were within 0.15 m of each other). The bankfull discharge was then calculated using Equation 2.1. The second method of calculating Q modelled different discharge scenarios using River2D and b  compared output water surface elevations with the observed bankfull depths. Many different flow scenarios were attempted in order to predict a reasonable bankfull flow event. The high flow mark elevation was compared to the model output water surface values. Analysis of the two methods led to the acceptance of a bankfull discharge of approximately 40 m /s in the 3  Tamihi side channel.  Application of a hydrodynamic model for the assessment and design of instream works Page 39  4  R E S U L T S A N D DISCUSSION  4.1  Field Surveys  4.1.1 Structures The current study investigated eleven structures (Str) installed at the Tamihi side channel in August 1999 (during minimum flow). The type of structures installed consisted of: 4 multipleL W D triangular, 4 single-LWD rootwad, 2 double-rock groynes, and 1 single-rock groyne (an example of each type of structure is presented in Figure 4.1 through Figure 4.4). As previously discussed single-LWD structures consist of a L W D (with or without rootwad) ballasted at both ends by boulders; multiple-LWD triangular structures are V-shaped with the apex directed out into the channel, ballasted on the bank and at the apex with boulders; double- and single-rock groyne structures are V-shaped (apex directed into flow) and are constructed using stones of classes large enough to withstand the shear stress caused by the river.  4.1.2  Pebble counts  A Wolman (1954) pebble count was performed at upstream, middle, and downstream locations of the study channel. The methodology involved measuring the intermediate or "B" axis of 100 pebbles randomly picked off the channel bed. The pebble counts were completed prior to the November 1999 flood. The D , D , geometric mean (D ), geometric standard deviation (a ), 50  90  g  g  and equivalent roughness height (k ) (from Equation 2.4) are presented in Table 4.1. Grain size s  distribution curves for all three pebble counts are presented in Appendix A .  Sample Location Upstream middle downstream  Ao  Coverage (m) -160- -10 -9 - -62 - 6 3 - -160  0.095 0.070 0.060  (m) 0.240 0.170 0.170  A(m)  a„(-)  0.070 0.061 0.059  2.793 2.449 2.366  £ (m) 0.65 0.48 0.41 s  Table 4.1: Tamihi side channel grain size analysis. The results indicate a trend towards downstream fining of bed sediments in the side channel. Observations of varying D and a values indicate a progressive downstream sorting of the bed %  g  material. Field observations indicated large classed material at the toe of the adjacent gravel bar (at the entrance to the side channel).  Application of a hydrodynamic model for the assessment and design of instream works  Figure 4.1: Multiple-LWD triangular structure (Strl 1).  Figure 4.2: Single-LWD rootwad structure (Strl4).  Page 40  Application of a hydrodynamic model for the assessment and design of instream works Page 41  Figure 4.3: Single-rock groyne structure (Strl2).  Figure 4.4: Double-rock groyne structure (Str9).  Application of a hydrodynamic model for the assessment and design of instream works Page 42 4.1.3 Site bathymetry The November 12 1999 bankfull flood was large enough to mobilize the bed of the Tamihi side th  channel. Bathymetric pre- and post-flood surveys were conducted in late-September 1999 and mid-February 2000, respectively, using Nikon 500 and Leika 700 series total station equipment. The channel was surveyed subjectively, with a high density of measurements (approximately 0.5 - 1.0 m apart) around and below L W D structures. The survey attempted to pick up any irregularities or inconsistencies in the bathymetry, which could cause local hydraulic changes. The survey data was downloaded to a desktop computer and a surface mapping programme (Surfer) was used to create a digital elevation model (DEM) with a 0.5 m spacing for both the pre- and the post-flood surveys. The side channels pre- and post-flood D E M bathymetries are presented on Figure 4.5a, Figure 4.5b, and Figure 4.5c. Instream structures are represented by brown square symbols. Greater bed variation around instream structures can be seen in the post-flood bathymetries. The variation is observed to be dependent on instream structure placement - with scour occurring at or around structure nick points (furthest points projected into flow). Figure 4.6 presents the longitudinal pre- and post-flood thalweg profile for the side channel. The local slope around instream structures in the post-flood profile has changed significantly, while the total slope has increased only slightly from 0.0041 m/m to 0.0050 m/m, pre- and post-flood, respectively. Aggradation, as seen in Figure 4.6, occurred at the top of the reach (upstream of Str6), while local degradation around individual structures was observed downstream. The upstream aggradation is most likely not a result of instream structure hydraulic effects, but rather a result of increased sediment load (due to the Q flood) which mobilized sediments from the toe of the 5  adjacent medial bar and deposited them into the side channel.  4.1.4 Measured scour The net scour plots for the side channel, obtained by subtracting the post -flood bathymetry from the pre -flood bathymetry, are presented on Figure 4.7a, Figure 4.7b, and Figure 4.7c. Positive values indicate degradation, while the contoured area < 0.0 represents aggradation. Up to 0.43 m of scour was measured in the lower portion of the side channel (at the apex of Strl 1) while maximum scour in the middle portion was up to 0.7 m (adjacent to Str9).  Application  of a hydrodynamic model for the assessment and design of instream works  Page 43  160-  150-^  99.5  (m)  99.4 (m) Str1  99.3 (m) 99.2 (m) 99.1 (m)  Str5  99.0  (m)  98.9  (m)  98.8 (m) 98.7 (m) 98.6  (m)  98.5  (m)  98.4  (m)  98.3 (m) Str6  98.2 (m) 98.1 (m)  -30  -20  -10  Lateral Distance (m) Pre-flood bathymetry  -30  -20  -10  98.0  (m)  97.9  (m)  Lateral Distance (m) Post-flood bathymetry  Figure 4.5a: Upper Tamihi side channel pre- and post-flood bathymetry. Contours represent elevation levels (in). Instream structures are represented by brown square symbols.  Application  of a hydrodynamic model for the assessment and design of instream works  50H  Page 44  Str7  Str8  Str9  98.7  (m)  98.6  (m)  98.5 (m)  98.4  (m)  98.3  (m)  98.2  (m)  98.1 (m) 98.0 (m)  stno  97.9  (m)  97.8  (m)  97.7  (m)  97.6  (m)  97.5  (m)  97.4 (m) 97.3 -10  0  Lateral Distance (m) Pre-flood bathymetry  -10  0  10  (m)  97.2 (m)  Lateral Distance (m) Post-flood bathymetry  Figure 4.5b: Middle Tamihi side channel pre- and post-flood bathymetry. Contours represent elevation levels (m). Instream structures are represented by brown square symbols.  Application  of a hydrodynamic model for the assessment and design of instream works  Page 45  98.5 (m) Str11  9  98.4 (m) 98.3 98.2  -100H  98.1  8 c  3TO-11CH  Str12  b  (m) (m) (m)  98.0 (m) 97.9 97.8  (m) (m)  97.7 (m) 97.6  C  o  97.5  -130 Str13  97.4  (m) (m) (m)  97.3 (m)  -140  97.2 (m) Strl 4 -150  97.1 (m) 97.0 (m) 96.9 (m)  -160- 1  0  10  20  0  10  20  Lateral Distance (m)  Lateral Distance (m)  Pre-flood bathymetry  Post-flood bathymetry  96.8  (m)  Figure 4.5c: Lower Tamihi side channel pre- and post-flood bathymetry. Contours represent elevation levels (m). Instream structures are represented by brown square symbols.  Application  of a hydrodynamic model for the assessment and design of instream works  Page 46  5.5  Pre-flood bathymetry Post-flood bathymetry Pre-flood slope = 0.0041 Post-flood slope = 0.0050  99.0  98.5  c o  S  98.0  LU  > 97.5  97.0  96.5 50  100 150 200 Cumulative Distance Downstream (m)  250  300  Figure 4.6: Tamihi side channel longitudinal thalweg pre- and post-flood profiles. The instream structure locations indicated represent the apex or furthest point projected into flow.  Application of a hydrodynamic model for the assessment and design of instream works Page Al  Figure 4.7a: Upper Tamihi side channel net scour plot. Positive contours represent degradation (scour depth relative to pre-flood bathymetry). Yellow shading indicates aggradation. Instream structures are represented by brown square symbols.  Application  of a hydrodynamic model for the assessment and design of instream works  Page 48  Figure 4.7b: Middle Tamihi side channel net scour plot. Positive contours represent degradation (scour depth relative to pre-flood bathymetry). Yellow shading indicates aggradation. Instream structures are represented by brown square symbols.  Application  of a hydrodynamic model for the assessment and design of instream works  Page 49  Figure 4.7c: Lower Tamihi side channel net scour plot. Positive contours represent degradation (scour depth relative to pre-flood bathymetry). Yellow shading indicates aggradation. Instream structures are represented by brown square symbols.  Application  of a hydrodynamic model for the assessment and design of instream works  Page 50  Degradation zones, in Figure 4.7a through Figure 4.7c, are observed to extended up to 40 m downstream of their related structures (Str9 and Strl 1). The deepest scour holes are located adjacent the apex of the multiple-LWD triangular and rock-groin structures. This finding is consistent with Lisle (1986) who observed the deepest scour to occur just downstream of the farthest obstruction projection. Backwater and sheltering effects of the triangular structures led to the deposition of fine material directly downstream and upstream along the banks of the channel. Scour and deposition around the single-LWD rootwad structures (specifically StrlO and Strl2) appears to be consistent with the alluvial bedforms described by Abbe and Montgomery (1996) - a crescent pool can be seen just upstream of the rootwad, a depositional bar formed further upstream and a second depositional bar formed downstream along the L W D trunk. Field observations indicated that sediments that deposited along the L W D trunks were of a finer calibre than the sediment composition of the upstream bar (or in the channel). 4.2  Two-Dimensional Flow Model - River2D  4.2.1  Data  manipulation  The pre- and post-flood D E M bathymetries produced by Surfer were converted into .CSV format with a 1.0 m spacing. A programme was written in Visual Basic to filter out the extraneous values outside the area of interest. Due to the computational restrictions of River2D the resulting CSV files (representing the channel bathymetry) were subdivided into upper and lower reaches that were modeled in separate runs. Four different .CSV (.bed) files representing pre- and post-flood bathymetries with and without instream structure representation (plain bed and bed with structures) were developed for River2D input. The instream structure representation was removed from the pre-flood bathymetry (plain bed) in order to give a representation of the channel prior to structure installation. Instream structures were assumed to be shelf like solid obstructions where water was restricted to flowing overtop and around - not through. Shields and Gippel (1995) used a similar assumption when considering natural LWD: field observations indicate that the interstitial spaces between branches and/or roots are often filled with leaves and sediment and serve as solid objects rather than permeable clusters. River2D was used to create finite element meshes for the four bed files and were then modeled at different discharge scenarios.  Application of a hydrodynamic model for the assessment and design of instream works Page 51 4.2.2 Model verification Field velocity and water surface surveys were completed on 29/05/00 and 15/06/00 (velocity measurements on 15/06/00 were limited to a single cross-section) using flow metering and Leika 700 series total station equipment. Velocity, depth and water surface values were measured at 1.0 m spacing intervals along channel cross-sections from which specific discharge was calculated. Random velocity, depth and water surface values were also measured throughout the channel for model verification purposes. Flows determined from the cross-sectional velocity and area measurements were 4.7 m /s and 11.3 m /s for the first and second surveys, respectively. 3  3  The 4.7 m /s flow scenario was modelled using River2D and the post-flood bathymetry (with 3  structures) - using the observed downstream boundary conditions. The modelled water depth (y ) and velocity values (w ) were compared to observed values and are presented on mod  mod  comparison plots Figure 4.8, Figure 4.9 (comparison plots for Q = 11.3 m /s were not prepared 3  due to insufficient data points). A line representing a 1:1 relationship (w  mod  = w or y obs  mod  = y ) is obs  presented on the two plots. The mean absolute error (MAE) and the mean error (ME) are presented in Table 4.2 and were calculated given the following equations (velocity forms presented):  moi  obs  U  . n  ME =  ^  ^ ^V  n  U  (4.1)  *obs mod  M  obs  U  obs  (4.2) J  u  where, w w n  = modelled velocity = observed velocity = number of samples  mod  obs  Unit  0(m7s) 4.7 4.7  velocity verification water depth  M E A (%)  M E (%)  24 6.1  15 -0.65  Table 4.2: Model verification - mean absolute error (MAE) and mean error (ME). Table 4.2 indicates that River2D predicts y  moi  predicted u . Yet w moi  mod  well, while more scatter is observed with the  values predicted by River2D are still within 24% of w . The M E for obs  Application of a hydrodynamic model for the assessment and design of instream works Page 52 both verifications is less than the M E A indicating that the spread tends to be centred on the "mod= "obs (or^  0  mod  =.y ) line. obs  0.4  0.8  1.2  Observed Velocity (m/s)  Figure 4.8: Tamihi side channel velocity verification plot (Q = 4.7 m /s). 3  1.6  Application of a hydrodynamic model for the assessment and design of instream works Page 53  Application of a hydrodynamic model for the assessment and design of instream works Page 54 4.2.3 Discharge scenarios Four different discharge scenarios were modelled using River2D: 40 m /s (Q determined in 3  b  Section 3.3.2), 5.0 m /s, 0.5 m /s, and 0.05 m /s. The downstream boundary condition (water 3  3  3  surface elevation) for each case was obtained through an iterative process: by changing downstream conditions until observed output water depths were approximately uniform. Modelled velocity contour plots of the 40 m /s discharge are presented on Figure 4.10a, 3  Figure 4.10b, and Figure 4.10c. Greater variations in local velocity magnitudes are observed when instream structures are included. Plain bed velocities, over the entire reach, tend to be uniform over long stretches (up to 110 m) with thalweg values ranging from 2.0 m/s to 3.0 m/s. With the inclusion of structures, the long stretches of constant velocity are reduced to distinct structure dependent zones (less then 35 m) with velocity values ranging from 1.5 m/s to 4.0 m/s. Maximum modelled velocities tend to occur close to the apex on the downstream side of the instream structures. Local flow hydraulics are affected by structure placement and spacing. The spacing between Strl3 and Sri4 is small enough (approximately 4 m) so that Strl4 is somewhat sheltered from flow and has less hydraulic significance. Lisle (1986) suggests that obstructions spaced less than 2 w from each h  other restrict pool formation. While this is true in the above case with Strl3 and Strl4, other structures that are within 2 w from each other still cause local flow variation (such as between b  Str7 and Str8 (see Figure 4.10b), and Strl land Strl2 (see Figure 4.10c)). It may therefore be more appropriate to state that flow disturbance and variation would be maximized at a spacing greater than 2 w . b  Modelled velocity vector plots are presented on Figure 4.1 la, Figure 4.1 lb, and Figure 4.1 lc and demonstrate output velocities calculated by River2D. Blue lines represent the channel bed boundary, which was subjectively determined by analysing the bathymetric contour plots. The vectors are scaled with velocity magnitude and greater variation can again be seen around structures, especially Str9, Strl 1 and Strl 2. The vector plots give a good indication of where flow acceleration and slack water occurs throughout the reach. Side currents and backwater eddies can be seen along the bank on the downstream side of Str7, Strl 1 and Strl2 (similarly to those observed in Figure 2.1 and Figure 2.2).  Application  of a hydrodynamic model for the assessment and design of instream works  Page 55  4.0 (m/s)  Strl 3.5 (m/s) 3.0 (m/s) 2.5 (m/s) Str5 2.0 (m/s) 1.5 (m/s) 1.0 (m/s) Stro  0.5 (m/s)  -30  -20  -10  -30  -20  -10  Lateral Distance (m)  Lateral Distance (m)  Plain bed  Bed with structures  0.0 (m/s)  Figure 4.10a: Upper Tamihi side channel modelled velocity magnitude plots (m/s). Two pre-flood bathymetry modelling scenarios are presented (Q = 40 m 3/s): plain bed and bed with structure representation. Instream structures are represented by brown square symbols. A  Application of a hydrodynamic model for the assessment and design of instream works Page 56  Lateral Distance (m)  Lateral Distance (m)  Plain bed  Bed with structures  Figure 4.10b: Middle Tamihi side channel modelled velocity magnitude plots (m/s). Two pre-flood bathymetry modelling scenarios are presented (Q = 40 m 3/s): plain bed and bed with structure representation. Instream structures are represented by brown square symbols. A  Application of a hydrodynamic model for the assessment and design of instream works Page 57  Lateral Distance (m) Plain bed  Lateral Distance (m) Bed with structures  Figure 4.10c: Lower Tamihi side channel modelled velocity magnitude plots (m/s). Two pre-flood bathymetry modelling scenarios are presented (Q = 40 m 3/s): plain bed and bed with structure representation. Instream structures are represented by brown square symbols. A  Application of a hydrodynamic model for the assessment and design of instream works Page 58  Strl  Str5  Stro  -30  -20  -10  Lateral Distance (m) Plain bed  -30  -20  -10  Lateral Distance (m) Bed with structures  Figure 4.11a: Upper Tamihi side channel modelled velocity vector plots. Two pre-flood bathymetry modelling scenarios are presented (Q = 40 m3/s): plain bed and bed with structure representation. Plotted vectors are scaled to modelled velocity magnitudes. Blue lines represent channel bed boundary. Instream structures are represented by brown square symbols.  Application  of a hydrodynamic model for the assessment and design of instream works  Page 59  Str7  Str8  Str9  Str10  -10  0  10  Lateral Distance (m) Plain bed  -10  0  10  Lateral Distance (m) Bed with structures  Figure 4.11b: Middle Tamihi side channel modelled velocity vector plots. Two pre-flood bathymetry modelling scenarios are presented (Q = 40 m3/s): plain bed and bed with structure representation. Plotted vectors are scaled to modelled velocity magnitudes. Blue lines represent channel bed boundary. Instream structures are represented by brown square symbols.  Application of a hydrodynamic model for the assessment and design of instream works Page 60  Str11  8  c co  Str12  b c T3 -*—.  D) C  o  Str13 -14f> Str14  -150  -160 0  10  20  Lateral Distance (m) Plain bed  10  20  Lateral Distance (m) Bed with structures  Figure 4.11c: Lower Tamihi side channel modelled velocity vector plots. Two pre-flood bathymetry modelling scenarios are presented (Q = 40 m3/s): plain bed and bed with structure representation. Plotted vectors are scaled to modelled velocity magnitudes. Blue lines represent channel bed boundary. Instream structures are represented by brown square symbols.  Application of a hydrodynamic model for the assessment and design of instream works Page 61 4.2.4 Pool Distributions River2D was used to determine pre- and post-flood pool distributions for the Tamihi side channel by modelling low discharges (Q = 0.5 m /s) and retrieving the water depth (y) from the 3  output files. The retrieved depths were contoured and plotted using Surfer and are presented on Figure 4.12a, Figure 4.12b, and Figure 4.12c. Contours of pool depths less than 0.2 m were intentionally omitted as they may represent survey inconsistencies and not pool formation. The pre-flood pool distributions indicate that pools were already in existence around Strl, Str5, Str6, Str7 and Strl 1. The formation of these initial pools is uncertain. The structures may have been installed in pre-existing pools or the pools may have been created at these locations during instream structure placement. The post-flood distributions indicate that pool formation and shape adjustment are a direct result of structure placement. Multiple-LWD triangular and rock-groyne structures appear to cause larger pools than single-LWD rootwad structures. This is most likely a reflection of constriction width (blockage ratio), which Lisle (1986) had found to be instrumental in causing channel-spanning pools. Likewise, Shields and Gippel (1995) found blockage ratio to be the most important parameter influencing flow resistance. With the exception of Strl, pool depth increased around all structures (up to a maximum of 0.6 m). In most cases, large increases in pool area were measured. Pool area deeper than 0.3 m increased from 320 m to 470 m in the pre- and post-bathymetric conditions, respectively. While pool 2  2  area deeper than 0.4 m increased from 55 m to 165 m , respectively. The changes in pool area 2  2  correspond to a 50% and 150% increase (deeper than 0.3 m and 0.4 m cases, respectively).  4.2.5 Shear predictions The u and y values produced by River2D using the pre-flood bathymetry (with structure representation) were used to calculate x values given Equation 2.5. This bed scenario was used 0  for the shear stress calculations since it represents the in situ conditions encountered by the flood (Qb  =  40 m /s). Shear stress intensity (x. ) is equal to the ratio of the calculated local shear stress 3  jnt  over the dimensionalised critical shear stress (T /T ) where x„ 0  crit  crit  = 0.056 (from Shields) and was  dimensionalised using Equation 2.6. The x» plots are presented (along with the previously int  presented net scour plots) on Figure 4.13a, Figure 4.13b, and Figure 4.13c.  Application of a hydrodynamic model for the assessment and design of instream works Page 62  Lateral Distance (m) Pre-flood bathymetry  Lateral Distance (m) Post-flood bathymetry  Figure 4.12a: Upper Tamihi side channel pre- and post-flood pool distributions. Contours represent pool depths (m). Blue lines represent channel bed boundary. Instream structures are represented by brown square symbols. Depths less than 0.2 m were omitted as they may represent survey inconsistencies and not pool formation.  Application  of a hydrodynamic  Lateral Distance (m) Pre-flood bathymetry  model for the assessment  and design of instream  works  Page 63  Lateral Distance (m) Post-flood bathymetry  Figure 4.12b: Middle Tamihi side channel pre- and post-flood pool distributions. Contours represent pool depths (m). Blue lines represent channel bed boundary. Instream structures are represented by brown square symbols. Depths less than 0.2 m were omitted as they may represent survey inconsistencies and not pool formation.  Application  of a hydrodynamic model for the assessment and design of instream works  Lateral Distance (m)  Lateral Distance (m)  Pre-flood bathymetry  Post-flood bathymetry  Page 64  Figure 4.12c: Lower Tamihi side channel pre- and post-flood pool distributions. Contours represent pool depths (m). Blue lines represent channel bed boundary. Instream structures are represented by brown square symbols. Depths less than 0.2 m were omitted as they may represent survey inconsistencies and not pool formation.  Application of a hydrodynamic model for the assessment and design of instream works Page 6 5 160  150H  14(H 130-  0.6 (m)  3.0 Strl  0.5 (m)  |  120 2.5  CO  tfl  0.4 (m)  0.3 (m)  % 110 co 'c»  o  1  0  (  2.0 Str5  H  1.5  -30  -20  -10  Lateral Distance (m) Net scour plot  0.2 (m)  90  0.1 (m)  80  0.0 (m)  70  1.0  Str6  60 -30  -20  -10  0.5  0.0  Lateral Distance (m) Shear stress intensity  Figure 4.13a: Upper Tamihi side channel net scour and shear stress intensity plots. The net scour plot was previously presented as Figure 4.7a. Positive contours represent degradation (scour depth relative to pre-flood bathymetry), while yellow shading indicates aggradation. The shear stress intensity plot is based on the model results using the pre-flood bathymetry with structure representation. Plot contours represent the ratio of bed shear/critical shear. Shear stress intensity contours greater than 1.0 indicate areas of potential scour. Instream structures are represented by brown square symbols.  Application of a hydrodynamic model for the assessment and design of instream works Page 66  Lateral Distance (m) Net scour plot  Lateral Distance (m) Shear stress intensity  Figure 4.13b: Middle Tamihi side channel net scour and shear stress intensity plots. The net scour plot was previously presented as Figure 4.7b. Positive contours represent degradation (scour depth relative to pre-flood bathymetry), while yellow shading indicates aggradation. The shear stress intensity plot is based on the model results using the pre-flood bathymetry with structure representation. Plot contours represent the ratio of bed shear/critical shear. Shear stress intensity contours greater than 1.0 indicate areas of potential scour. Instream structures are represented by brown square symbols.  Application of a hydrodynamic model for the assessment and design of instream works Page 67  Lateral Distance (m) Net scour plot  Lateral Distance (m) Shear stress intensity  Figure 4.13c: Lower Tamihi side channel net scour and shear stress intensity plots. The net scour plot was previously presented as Figure 4.7c. Positive contours represent degradation (scour depth relative to pre-flood bathymetry), while yellow shading indicates aggradation. The shear stress intensity plot is based on the model results using the pre-flood bathymetry with structure representation. Plot contours represent the ratio of bed shear/critical shear. Shear stress intensity contours greater than 1.0 indicate areas of potential scour. Instream structures are represented by brown square symbols.  Application of a hydrodynamic model for the assessment and design of instream works Page 68 It is difficult to draw any specific conclusions from Figure 4.13a, Figure 4.13b, and Figure 4.13c as they compare two separate parameters (scour depth (d ) and x„ ). Nevertheless, the locations s  int  of deep scour (in the middle and lower portions of the channel) appear to be in close proximity to the locations of predicted x. > 1.0. The predicted x» area is much more uniform than the int  int  measured extent of d , however the overall area between the two plots (in the middle and lower s  portions, Figure 4.13b, and Figure 4.13c) is similar. Discrepancies occur due to small irregularities in the channel that may have been smoothed over in the process of transferring the bathymetric survey data to the D E M and .bed files.  The predicted x» values at Str5 through Str8 seem low considering the scour that that can be int  seen around these structures (see Figure 4.13a and Figure 4.13b). This may be a result of structure burial. The reach narrows and deepens in its upper portion, and, as a result, instream structures (such as Str6 and Str8) are likely to be buried during high flow events. As water depth increases overtop of the structures, the hydraulic influence (flow separation losses and local high velocities) would be reduced. This reduction in hydraulic influence was investigated by calculating x. using a lower discharge (20 m /s). Under this scenario, the model predicted a 3  int  larger area of high x» values around Str7 and Str8 than was predicted using Q - indicating int  b  burial was an issue at high flows. Likewise, Shields and Gippel (1995) found pool-forming processes around L W D were reduced with increased discharges; furthermore, Smith and Beschta (1994) found scour and deposition occurred at discharges well below bankfull.  The model predicted very high shear stresses at the upper most region of the side channel (see Figure 4.13a). Contrarily, the net scour plot indicates that aggradation has taken place in this region. As discussed in Section 4.1.3, aggradation occurred in the upper channel most likely because of large classed sediments, which were transported from the adjacent medial bar. This occurrence could not be predicted by River2D. As well, model output close to the upper boundary may be erroneous and produce higher shear stress than realistic.  For completeness the selected Shields value of x. by Church et al (1998) and x, x,  crit  crit  crit  = 0.056 was compared to the value suggested  = 0.079. The model results appear to be quite sensitive to the  selected. With increased x„  crit  values from 0.056 to 0.079, the area of predicted shear is  Application of a hydrodynamic model for the assessment and design of instream works Page 69 substantially reduced. The reduction in shear is even more apparent in the upper portion of the channel.  4.2.6 Fish habitat The Bovee (1978) preference curves, for the site's target species (coho and steelhead), were used to develop velocity and depth lookup tables required for the habitat component of River2D. The depth and velocity preference tables are used by River2D to determine the W U A for each flow scenario (Q = 40 m /s, Q = 5.0 m /s, Q = 0.5 m7s, Q = 0.05 m7s). 3  3  As with Mosley (1982), and  Hogan and Church (1989), cover, temperature and substrate preferences were assumed to be optimum to simplify results. Table 4.3, Table 4.4, Figure 4.14 and Figure 4.15 summarize the W U A of the target species computed by River2D for the discharges investigated. (9(m /s)  spawners  pre-  fry post-  % change  pre-  post-  % change  40  0.9  2.5  275  2.0  5.5  275  5.0 0.5 0.05  1.0 1.3 0.2  1.5 2.5 0.8  50 190 400  4.6 11.9 0.2  4.1 8.9 0.3  90 75 150  3  Table 4.3: Modelled %WUA for coho salmon at selected discharges. e(m /s)  juveniles  pre-  fry post-  % change  pre-  post-  % change  40  4.3  13.4  310  4.9  12.1  245  5.0 0.5 0.05  6.1 20.5 11.0  5.9 17.0 12.0  95 80 110  10.8 15.3 3.7  10.1 13.6 4.8  95 90 130  3  Table 4.4: Modelled %WUA for steelhead trout at selected discharges. The addition of instream structures to the side channel resulted in an increase (factor > 2.4) %WUA for all ages (both species) during Q . The %WUA for steelhead at high flow can be seen b  to be much higher (factor of 2 - 5) than the %WUA used by coho. The differences between the two species were not unexpected. Juvenile steelhead prefer fast flowing water, while coho being not as well adapted for high velocity zones prefer off-channel areas (such as ponds) with lower velocities and shallower depths (Swales and Levings 1989). During low flow seasons in the Chilliwack River (late summer and late winter (before spring  Application of a hydrodynamic model for the assessment and design of instream works Page 70 freshet)), flow into the Tamihi side channel is restricted to groundwater seepage due to the entrance sill. Table 4.3 and Table 4.4 indicate that the %WUA for Q = 0.05 m /s increases 3  slightly from pre- to post-flood due to instream morphologic and hydraulic changes. The hydraulic effects of instream structures at low flows are minimal since flow is restricted to the centre portion of the channel, yet the increased pool area (created by the structures during high flows) results in more refuge areas being available. The %WUAs obtained during the Q = 0.5 m /s decreased slightly in the post-flood scenario for 3  coho spawners, steelhead fry and steelhead juveniles (factors of 1.3, 1.2 and 1.1, respectively), while coho fry incurred an increase (factor of 1.9 increase). The decrease in coho spawners is understandable since the formation of pools would reduce available spawning area. The decrease in steelhead %WUA in the post-flood run may be a result of their preference for faster, shallower water (Bisson et al 1981). Figure 4.14 and 4.15 indicate that the most beneficial aspect of the instream structures is to offer refuge areas during high-flow events for steelhead and coho of both ages. As stage increases there is a trend towards faster moving water. Fish must find refuge areas behind boulders or in pools to avoid being swept downstream by the current. The instream structures create low velocity zones on their downstream sides where fish can hold and rest. At lower flows these refuge areas become less important. The results seem to indicate that the importance of the pools formed by the structures only become important when flows reach critically low levels. Yet as was discussed in Section 2.5, deep pool areas (such as those created by the instream structures) are desirable for older salmonids (Bustard and Narver 1975) and may not be well represented by the Bovee (1978) preference curves.  Application  of a hydrodynamic  model for the assessment  and design of instream  works  Page 71  16 Post-flood fry Pre-flood fry Post-flood spawners Pre-flood spawners  • •  -  —e— e -  u  | 0.01  I I I I I I llj 0.1  I  I! I ! I 1  11  I I I I I I I l| 10  1—! I I I I I|j 100  Discharge (m /s) 3  Figure 4.14: Tamihi side channel percent weighted usable area for coho salmon. Post-flood values were modelled using the bed with structure representation bathymetry, while the pre-flood values were modelled using the plain bed bathymetry.  Application of a hydrodynamic model for the assessment and design of instream works Page 72  24  Post-flood fry — —  Pre-flood fry  o  Post-flood juveniles  e -  Pre-flood juveniles  20  16 CO  < to  .o ro tn  3  TJ  12  ~1  0.01  I I I ! II I |  1  M i l l  1  0.1  I I I II  11  10  II  I I I  100  Discharge (m^/s)  Figure 4.15: Tamihi side channel percent weighted usable area for steelhead trout. Post-flood values were modelled using the bed with structure representation bathymetry, while the pre-flood values were modelled using the plain bed bathymetry.  Application of a hydrodynamic model for the assessment and design of instream works Page 73 5  SUMMARY AND RECOMMENDATIONS  This study assesses the morphological and hydraulic effects, as well as, the habitat benefits of instream L W D and boulder structures using a two-dimensional hydrodynamic model (River2D). Eleven instream structures, installed on the Tamihi side channel in August 1999, were investigated. Pre- and post-flood bathymetric surveys revealed bed scour to have occurred, to a maximum of 0.7 m, adjacent to and downstream of instream structures. Bed variation and complexity was also seen to increase. The observed scour was qualitatively seen to be a result of structure placement. Shape and location of scour holes around instream structure types were consistent with the literature findings of Lisle (1986), and Abbe and Montgomery (1996). Instream structures had a direct effect on local thalweg slope, yet the overall channel slope increased only marginally and was not attributed to instream structures.  River2D was used to predict channel hydraulic variables (local velocities and depths) for both the pre- and post-structure scenarios. Comparison between pre- and post-structure velocities indicated a trend towards greater velocity variation and higher local velocity values resulting form instream structures. Comparisons between modelled and observed u andy indicate River2D's predictions of u to be satisfactory (MAE = 24%), while y is well predicted (MAE = 6.1%). The verification results lend confidence in the use of River2D in predicting local channel hydraulics.  River2D was used as a tool to assess pool changes resulting from instream structure placement. Pool formation and shape adjustment were directly related to instream structures. Blockage ratio appeared to be an important factor in determining pool size. Reach length pool area deeper than 0.3 m increased by 50%, while pool area deeper than 0.4 m increased by 150%. The hydraulic variables predicted by River2D were used to calculate local shear stress throughout the reach. Areas of possible mobility were determined by dividing the predicted local bed shear stress by the critical shear stress suggested by Shields. Predicted bed mobility coincided with areas of scour in the middle and lower portions of the channel, while the upper portion of the channel, which underwent substantial aggradation during the November 12 1999 th  flood, was poorly related. Structure position and spacing affected predicted shear. Reduced  Application of a hydrodynamic model for the assessment and design of instream works Page 74 shear values were predicted when instream structures were positioned directly downstream from each other with < 2 w spacing. This sheltering effect was also observed in the net scour plots. b  Fish habitat analysis was performed on the pre- and post-flood bathymetry of the Tamihi side channel using River2D at various discharge scenarios. The results for the bankfull condition (Q ) h  indicated that the %WUA of coho salmon and steelhead trout would increase by a factor > 2.5 with the addition of instream structures. The effect of instream structures was reduced at lower discharges and in some cases resulted in a decrease in available %WUA. The W U A method was used to rate the potential habitat benefits of instream structures in the Tamihi side channel. The Bovee (1978) preference curves offer a simple indices for determining site specific fish habitat preference; yet the analysis is dependent on the correctness and applicability of transferring the preference curves from the regions and specific conditions under which they were developed to other regions.  The methods used in this study for scour and habitat analysis can be followed and used for other systems where instream works are planned. Planned instream works could be optimized to produce greater hydraulic variation and fish habitat. Furthermore, optimization would target specific species. The following steps, derived from this study, provide an iterative methodology for the optimization of instream works: 1.  Conduct a bathymetric survey of the reach under investigation. Spacing between sample points can be variable but should be sufficient to pick up bed and bank elevation changes.  2.  Conduct pebble counts and grain size analyses to characterize channel roughness.  3.  Investigate watershed hydrology to determine the bankfull flow (Q ) (the event which would likely induce the most scour and hydraulic variance). Calculate the bankfull depth (y ) for the downstream boundary using Equation 2.1. b  b  4.  Convert survey data into a digital elevation model (DEM) and add in preferred instream structure design representation. Spacing between nodes should be around 1.0 m or less.  5.  Convert D E M to mesh files (with equivalent roughness height k values) for input into River2D. s  Application of a hydrodynamic model for the assessment and design of instream works Page 75 6.  Run River2D under the bankfull boundary conditions and adjust downstream condition to delineate for backwater or draw-down effects.  7.  Input target species preference curves and run the habitat component River2D to determine the W U A for the particular species of interest. Cover information can also be added if desired.  8.  Using River2D's output file produce a velocity contour plot and a shear stress intensity plot using Equation 2.5 and Equation 2.6. The plots can be used to determine areas of influence, and optimal positioning and spacing of instream structures.  9.  Adjust the design of the instream structures (given the information learned in the above steps) and then repeat steps 5 through 8, until an optimum design is reached.  Fish habitat assessment is site, discharge and species specific and is related to local hydraulic and morphologic conditions. The habitat component River2D uses the predicted local hydraulic characteristics for evaluating the W U A and as such offers a powerful tool for assessing instream structure designs before they are installed. The methodology described above could also be used as a post-installation assessment to verify the value of previous restoration works. Without the use of River2D (or other 2D flow models), detailed local high-flow hydraulic information could only otherwise be gained through rigorous field sampling during high discharges (which is in many cases not possible).  The current study offers a methodology for standardising instream restoration practices such that some of the subjectivity and randomness of structure placement is removed. Unfortunately, the lack of a direct relationship between measured scour and predicted shear stress intensity, and the simplification of fish preferences into two hydraulic parameters (u and y) make it difficult to obtain a truly optimal design. But given the lack of methodologies currently available, the above technique, although not optimal, is a few steps closer towards that goal. Furthermore the information gleaned on structure spacing and size, through the use of local hydraulic analyses, is valuable and has previously been unavailable. As the study's methodology is highly field data dependent, the practical application of this method should be reserved for sites where bathymetric information is already available or where the most optimal solution is a design requirement.  Application of a hydrodynamic model for the assessment and design of instream works Page 76 Further areas of interest and possible research would be in investigating the relationship between observed scour and the shear stress intensity predicted by 2D hydrodynamic models. Much work in the past has been spent on trying to predict scour through the use of general or averaged channel hydraulic parameters and perhaps advances could be gained by investigating local hydraulics with the aid of 2D models. The fish preference curves are a simple method of identifying stream habitat availability, but research needs to be carried out to verify that the curves are indeed representative of fish preference and that the curves are applicable to river systems deferent from those they were originally developed from. Furthermore, in this study, as with Mosley (1982) and Hogan and Church (1989), the Bovee (1978) preference curves for streamside cover, substrate size and water temperature were assumed to be optimum. This assumption should be tested through the use of a 2D model and biological fish numeration. On a much more general level, it is important to remember that instream structure restoration activities should be viewed as interim management tools to improve habitat until restoration of riparian zones or upslope conditions occurs (House 1996; Roper et al 1998). River restoration projects should use instream structures to hasten ecological recovery, yet it may take up to 200 years before the rehabilitated riparian vegetation makes a significant contribution of L W D to rivers (Gippel et al 1992). Two-dimensional modelling offers the opportunity to improve interim habitat restoration practices, but watershed wide restoration must be addressed or instabilities caused by upstream and/or upslope effects will prematurely affect the durability and life of instream restoration works.  Application of a hydrodynamic model for the assessment and design of instream works Page 11 REFERENCES Abbe, T.B., and D.R. Mongomery. 1996. Large woody debris jams, channel hydraulics and habitat formation in large rivers. Regulated Rivers Research and Management 12:201-221. Allan, J.H., and S. Lowe. 1997. Chapter 11: Rehabilitating mainstem holding and rearing habitat. In Slaney, P.A., and D. Zaldokas. [ed.] Fish Habitat Rehabilitation Procedures. Province of B.C., Ministry of Environment, Lands and Parks, and Ministry ofForests. Watershed Restoration Technical Circular No. 9:34 lp. Beschta, R.L. 1984. River channel responses to accelerated mass soil erosion. Presented at Symposium on the Effects of Forest Land Use on Erosion and Slope Stability, May 7-11, Honolulu, Hawaii: 155-163 Bilby, R.E., and J.W. Ward. 1989. Changes in characteristics and function of woody debris with increasing size of streams in western Washington. Transactions of the American Fisheries Society 118:368-378. Bisson P.A., J.L. Nielsen, R.A. Palmason, and L.E.Grove. 1981. A system of naming habitat types in small streams, with examples of habitat utilization by salmonids during low streamflow. In N.B. Armantrout [ed.] Acquisition and Utilization of Aquatic Habitat Inventory Information, American Fisheries Society Western Division:62-73. Blackwell, C.N., C R . Picard, and M. Foy. 1999. Smolt productivity of off-channel habitat in the Chilliwack River watershed. Province ofB. C, Ministry of Environment, Lands and Parks, and Ministry of Forests. Watershed Restoration Project Report No. 14, 42 p. Bovee, K.D. 1978. Probability of use criteria for the family salmonidae. U.S.D.I., Fish and Wildlife Service, Office of Biological Services. Instream Flow Information Paper No. 4, 80p. Bovee, K.D. 1982. A guide to stream habitat analysis using the Instream Flow Incremental Methodology. U.S.D.I., Fish and Wildlife Service, Office of Biological Services. Instream Flow Information Paper No. 12, 248p. Bray, D.I. 1980. Evaluation of effective boundary roughness for gravel-bed rivers. Canadian Journal of Civil Engineering 7:392-397. Bustard, D.R., and D.W. Narver. 1975. Aspects of the winter ecology of juvenile coho salmon (Oncohynchus kisutch) and steelhead trout (Salmo gairdneri). Journal of the Fisheries Research Board of Canada 32:667-680. Carling, P.A. 1996. In-stream hydraulics and sediment transport. In Petts, G.E., and P. Calow [ed.] River flows and channel forms. Oxford, Blackwell Science: 160-184. Cherry, j. and R.L. Beschta. 1989 Course woody debris and channel morphology: a flume study. Water Resources Bulletin 25 (5): 1031-1036. Christner, W.T.Jr., and S.G.Custer. 1998. In-stream obstructions and the formation of pools and scour holes. Presented at Wetlands Engineering and River Restoration Conference, ASCE, March 22-27, Denver, Colorado. Chow, V.T. 1959. Open -Channel Hydraulics, McGraw-Hill Book Company, Inc., Boston, Massachusetts, 680 p. Church, M . 1992. Channel morphology and typology. In Calow, P. and Petts, G.E., [ed.] The Rivers Handbook. Oxford, Blackwell, 1:126-143.  Application of a hydrodynamic model for the assessment and design of instream works Page 78 Church, M., M.A. Hassan, and J.F. Wolcott. 1998. Stabilizing self-organized structures in gravel-bed stream channels: Field and experimental observations. Water Resources Research 34(11):3169-1379. Coulson, C.H., and W. Obedkoff. 1998. British Columbia Streamflow Inventory. Province of B.C., Ministry of Environment, Lands and Parks, Resources Inventory Branch: 49 p. D'Aoust, S.G., and R.G. Millar. 1999 Large woody debris fish habitat structure performance and ballasting requirements. Province of B.C., Ministry of Environment, Lands and Parks, and Ministry of Forests. Watershed Restoration Management Report No. 8. 119 p. Enggrob, H.G., and P.H. von Lany. 1994. An application of 2-D mathematical modelling on the Brahmaputra River, p. 1-14. Proceedings of the 2 International Conference on River Flood Hydraulics 22-25 March, York, England. nd  Frissell, C A . and R.K. Nawa. 1996. Incidence and causes of physical failure of artificial structures in streams of western Oregon and Washington. North American Journal ofFisheries Management 12:182-197. Galay, V.J., E.K. Yaremko, and M.E. Quazi. 1987. River bed scour and construction of stone riprap protection. In Thorne, C.R., J.C.Bathurst, and R.D. Hey [ed.] Sediment Transport in Gravel-bed Rivers. John Wiley and Sons Ltd., New York, USA:353-379 Garde, R.J., K. Subramanya, and K.D. Nambudripad. 1961. Study of scour around spur-dikes. Journal of the Hydraulics Division, ASCE 87(HY6):23-37. Gippel, C.J., I.C O'Neill, and B.L. Finlayson. 1992. The hydraulic basis of snag management. Report, Centre for Environmental Applied Hydrology, Department of Civil and Agricultural Engineering, University of Melbourne, Australia. 116 p. Hartman, G.G., and J.C. Scrivener. 1990. Impacts of forestry practices on a coastal stream ecosystem, Carnation Creek. Department of Fisheries and Oceans, Ottawa, ON. Canadian Bulletin of Fisheries and Aquatic Sciences 223. 148p. Hartman, G.F. 1965. The role of behaviour in the ecology and interaction of underyearling coho salmon and steelhead trout. Journal of the Fisheries Research Board of Canada 22:1035-1081. Hay and Company Consultants Limited. 1992. Chilliwack River hazard management study interim report. Fraser Cheam Regional District, Volume 1. Heede, B.H. 1972 Influences of a forest on the hydraulic geometry of two mountain streams. Water Resources Bulletin 8(3):523-530. Hogan, D.L., and M. Church. 1989. Hydraulic geometry in small, coastal streams: progress toward quantification of salmonid habitat. Canadian Journal of Fisheries and Aquatic Science 46:844-852. Hogan, D.L., S.A. Bird, and M.A. Hassan. 1998. Spatial and temporal evolution of small coastal gravel-bed streams: influenced of forest management on channel morphology and fish habitats. In P.C. Klingeman, R.L. Beschta, P.D. Komar, and J.B. Bradley [ed.] Gravel-bed rivers in the environment. Highlands Ranch, CO., Water Resources Publications: 365-392. House, R. 1996. An evaluation of stream restoration structures in a coastal Oregon stream, 1984-1993. North American Journal of Fisheries Management 16:272-281.  Application of a hydrodynamic model for the assessment and design of instream works Page 79 Jia, Y., and C.V. Alonso. 1994. One- and two-dimensional analysis of flow in Hotophia Creek, MS. Proceeding of the A.S.C.E Hydraulic Engineering '94 conference, August 1-5, Buffalo, New York (l):589-594. Keller, E.A., and T. Tally. 1979. Effects of large organic debris on channel form and fluvial processes in the coastal redwood environment. In D.D. Rhodes, and G.P. Williams [ed.] Adjustments of the fluvial system. Dubuque, Iowa, Kendall-Hunt: 169-197. Kellerhals, R., and M. Church. 1989. The morphology of large rivers: characterization and management, p. 31-48. In D.P. Dodge [ed.] Proceedings of the International Large River Symposium. Canadian Special Publication of Fisheries and Aquatic Sciences 106. Klein, R.D. 1984. Channel adjustments following logging road removal in small steep land drainages. Presented at Symposium on the Effects of Forest Land Use on Erosion and Slope Stability, 7-11 May, Honolulu, Hawaii: 187-195. Laursen, E.M. 1963. An analysis of relief bridge scour. Journal of the Hydraulics Division, ASCE 89(HY3):93-118. Lisle, T.E. 1986. Stabilization of a gravel channel by large streamside obstructions and bedrock bends, Jacoby Creek, northwestern California. Geological Society ofAmerica Bulletin 97:999-1011. Lisle, T.E., J.M. Nelson, J. Pitlick, M.A. Madej, and B.L. Barkett. 2000. Variability of bed mobility in natural, gravel-bed channels and adjustments to sediment load at local and reach scales. Water Resources Research 36(12):3743-3755. Madej, M.A. 1978. Sediment transport and channel changes in an aggrading stream in the Puget Lowland, Washington. U.S.D.A. Forest Service. General Technical Report PNW-14L97-108. McMahon T.E., and G.F. Hartman. 1989. Influence of cover complexity and current velocity on winter habitat use by juvenile coho salmon {Oncorhynchus kisutch). Canadian Journal of Fisheries and Aquatic Science 46:15511557. Melville, B.W. and S.E. Coleman. 2000. Bridge Scour. Water Resources Publications, LLC. Colorado, USA, 550p. Millar, R.G. 2000. Influence of bank vegetation on alluvial channel patterns. Water Resources Research 36(4):1109-1118. Montgomery, D.R., J.M. Buffington, R.D. Smith, K.M. Schmidt, and G. Pess. 1995. Pool spacing in forest channels. Water Resources Research 31 (4): 1097-1105. Mosley, M.P. 1981. The influence of organic debris on channel morphology and bedload transport in a New Zealand forest stream. Earth Surface Processes and Landforms 6:571-579. Mosley, M.P. 1982. Analysis of the effect of changing discharge on channel morphology and instream uses in a braided river, Ohau River, New Zealand. Water Resources Research 18(4):800-812 Nakamura, F., and F.J. Swanson. 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 Nickelson, T.E, M.F. Solazzi, S.L. Johnson, and J.D. Rodgers. 1991. Effectiveness of selected stream improvement techniques to create suitable summer and winter rearing habitat for juvenile coho salmon (Oncorhynchus kisutch) in Oregon coastal streams. Canadian Journal of Fisheries and Aquatic Science 49:790-794.  Application  of a hydrodynamic model for the assessment and design of instream works  Page 8 0  Robison, E.G., and R.L. Beschta. 1990. Coarse woody debris and channel morphology interactions for undisturbed streams in southeast Alaska, U.S.A. Earth Surface Processes and Landforms 15:149-156. Ralph, S.C., G.C. Poole, L.L. Conquest, and RJ. Naiman. 1994. Stream channel morphology and woody debris in logged and unlogged basins of western Washington. Canadian Journal of Fisheries and Aquatic Science 51:37-51. Roper, B.B., D. Konnoff, D. Heller and K. Wieman. 1998. Durability of pacific northwest in stream structures following floods of a five-year or greater return interval. North American Journal of Fisheries Management. 18:686-693. Rutherfurd, I.D., K. Jerie, andN. Marsh. 1999. A Rehabilitation Manual for Australian Streams Volume 2.  LWRRDC and CRCCH, World Wide Web, 400 p. Shields, F.D.Jr., and C J . Gippel. 1995. Prediction of effects of woody debris removal on flow resistance. Journal of Hydraulic Engineering\2\{A):ZA\-2>SA.  Shvidchenko, A.B., and G. Pender. 2000. Flume study of the effect of relative depth on the incipient motion of coarse uniform sediments. Water Resources Research 36(2):619-628. Slaney, T.L., K.D. Hyatt, T.G. Northcote, and R.J. Fielden. 1996. Status of anadromous salmon and trout in British Columbia and Yukon. Fisheries 21(10):20-35. Slaney, P.A., and D. Zaldokas. 1997. Fish Habitat Rehabilitation Procedures. Province of B.C., Ministry of Environment, Lands and Parks, and Ministry of Forests. Watershed Restoration Technical Circular No. 9:341p. Slaney, P.A., and A.D. Martin. 1997. Chapter 1, Planning fish habitat rehabilitation: linking to habitat protection. In Slaney, P.A., and D. Zaldokas. [ed.] Fish Habitat Rehabilitation Procedures. Province of B. C, Ministry of Environment, Lands and Parks, and Ministry of Forests. Watershed Restoration Technical Circular No. 9:  Smith, R.D., and R.L. Beschta. 1994. A mechanism of pool formation and maintenance in forest streams. Proceeding of the A.S. CE Hydraulic Engineering '94 conference, August 1-5, Buffalo, New York (2):824-828.  Steffler, P. 1999. River2D two-dimensional depth average model of river hydrodynamics and fish habitat. Unpublished user's manual, University of Alberta.45$.  Swales, S., and C D . Levings. 1989. Role of off-channel ponds in the life cycle of coho salmon and other juvenile salmonids in the Coldwater River, British Columbia. Canadian Journal of Fisheries and Aquatic Sciences 46:232242. Tripp, D. 1996. Using large organic debris to restore fish habitat in debris-torrented streams. Province of B.C., Ministry of Forests and Lands, Land management report no. 47, ISSN 0702-9861, 26 p. Wolman, M.G. 1954. A method of sampling coarseriver-bedmaterial. Transactions, American Geophysical Union 35 (6):951-956. Zaldokas, D.O. [editor] 1999. Annual compendium of aquatic rehabilitation projects for the Watershed Restoration Program 1998-1999. Province of B.C., Ministry of Environment, Lands and Parks, and Ministry of Forests.  Watershed Restoration Project Report No. 13.  Application of a hydrodynamic model for the assessment and design of instream works Page 81  Appendix A  Application  of a hydrodynamic  model for the assessment  and design of instream  works  100 o o o o o o o o o o o o o o o o o o o o o o  80  o o o  o o o o o o o o o o o o o o o o o o o o o  60 CD  c CD CL  o o  Page  40  20  1  0.001  I  I  ? I I 11  1  I  0.01  M i l l  0.1 Grain Size (m)  Figure A.la: Upper Tamihi side channel grain size distribution.  i  rr  TTT  82  Application of a hydrodynamic model for the assessment and design of instream works Page 83 100  —i  80  _  60  0)  2  cu D.  40  20  1  0.001  I  I I I II  1  I  0.01  M i l l  0.1  Grain Size (m)  Figure A.lb: Middle Tamihi side channel grain size distribution.  TTT  Application of a hydrodynamic model for the assessment and design of instream works 100 o o o o o o o o o o o o o o o o  80  o o o o o  60 CD  c CD CD Q.  40  20  1 0.001  I  1  I II  I  I I I I  0.01  0.1  Grain Size (m) Figure A.lc: Lower Tamihi side channel grain size distribution.  i  r~rr  Page 84  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0063717/manifest

Comment

Related Items