Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Stability seeding : experimental testing of a new bank protection technique Tatham, Caitlin 2019

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

Item Metadata

Download

Media
24-ubc_2019_september_tatham_caitlin.pdf [ 72.48MB ]
Metadata
JSON: 24-1.0380598.json
JSON-LD: 24-1.0380598-ld.json
RDF/XML (Pretty): 24-1.0380598-rdf.xml
RDF/JSON: 24-1.0380598-rdf.json
Turtle: 24-1.0380598-turtle.txt
N-Triples: 24-1.0380598-rdf-ntriples.txt
Original Record: 24-1.0380598-source.json
Full Text
24-1.0380598-fulltext.txt
Citation
24-1.0380598.ris

Full Text

Stability Seeding:Experimental Testing of a New Bank Protection TechniquebyCaitlin TathamB.Sc., University of British Columbia, 2017A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMaster of ScienceinTHE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES(Geography)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)August 2019©Caitlin Tatham, 2019The following individuals certify that they have read, and recommend to the Faculty ofGraduate and Postdoctoral Studies for acceptance, a thesis/dissertation entitled:Stability Seeding: Experimental Testing of a New Bank Protection Techniquesubmitted by: Caitlin Tatham in partial fulfillment of the requirementsfor the degree of Master of Sciencein GeographyExamining Committee:Brett Eaton, GeographySupervisorMarwan Hassan, GeographySupervisory Committee MemberIan McKendry, GeographyAdditional ExamineriiAbstractThe demand for development near rivers, coupled with changes to the hydrologic regime haveresulted in an increase in the frequency of high magnitude flooding events. These floodingevents pose a threat to infrastructure. To protect infrastructure along channels, traditionalbank protection is installed. These structures prevent natural scour and limit bank migration,which can impair ecosystem function. As an alternative traditional approaches, we proposea new approach to bank protection. Instead of aiming to control channel shape and form,bank protection should work with the processes that govern channel stability.Recent research has shown that the coarse-tail of a channel’s grain size distribution hassurprisingly strong affects on a channel’s stability. This thesis aims to capitalize on thisfinding, by incorporating the coarsest grains on a channel’s bed into a bank protectiontechnique. We call this bank protection technique 'Stability Seeding'. Stability Seedingconsists of placing coarse grains onto a channel’s banks. When the channel widens, thesegrains fall into the bed and promote channel stability. To assess the feasibility of usingStability Seeding as bank protection, proof-of-concept experiments were run using scaledphysical models. Three experiments were run; the first used Stability Seeding, the secondused Riprap as a traditional method of bank protection, the third experiment was composedof a natural channel without any bank protection.It was found that Stability Seeding dampens bank migration, while still allowing a smallamount of natural channel adjustment. As a result, using Stability Seeding requires a largerminimum setback distance than traditional bank protection. However, as Stability Seedingallows for some level of bank migration, it results in less vertical degradation than Riprap.Therefore, Stability Seeding could lessen the chance of buried infrastructure being exposedcompared to traditional bank protection techniques. In addition, Stability Seeding allowsmore morphological variation than traditional bank protection. This variation could allowa reach protected by Stability Seeding to be more ecologically productive than reaches pro-tected by traditional bank protection. The findings of this thesis provide the basis for usingStability Seeding as an alternative to traditional bank protection.iiiLay SummaryLarge-scale flooding events have the potential to be devastating for infrastructure and com-munities located along channels. Despite an increase in these floods, techniques to mitigateflood-associated damages have not evolved to meet the new challenges. This thesis proposesa novel bank protection technique that incorporates large grains naturally found in the chan-nel. This technique was tested using three stream table experiments. It was found that usinglarge grains as bank protection reduces channel widening, while still allowing natural systemmovement. This technique has the potential to be less invasive and more ecologically-friendlythan traditional bank protection. This thesis provides the basis for this new technique to beincorporated into river engineering projects.ivPrefaceThis dissertation is the result of three experiments designed by Brett Eaton, Lucy Mackenzieand Caitlin Tatham. Caitlin Tatham conducted the experiments. All the analysis and writingpresented in this dissertation was done by Caitlin Tatham.vContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiLay Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ixList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiList of Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Hazards Associated with River Systems . . . . . . . . . . . . . . . . . . . . . 11.2 Traditional Bank Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.3 An Alternative to Traditional Bank Protection . . . . . . . . . . . . . . . . . 21.4 Research Objectives and Thesis Structure . . . . . . . . . . . . . . . . . . . 32 Experimental Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.1 Natural Channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.2 Stability Seeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.3 Riprap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93.1 Experimental Setup and Operation . . . . . . . . . . . . . . . . . . . . . . . 93.1.1 Scaling and A-BES . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93.1.2 Initial Channel Conditions and Natural Channel Design . . . . . . . . 93.1.3 Stability Seeding Design . . . . . . . . . . . . . . . . . . . . . . . . . 10vi3.1.4 Riprap Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123.1.5 Flow and Sediment Feed Conditions . . . . . . . . . . . . . . . . . . . 133.2 Data Collection and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 133.2.1 Channel Width and Depth . . . . . . . . . . . . . . . . . . . . . . . . 143.2.2 Surface Texture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143.2.3 Flow Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 Morphological Adjustments . . . . . . . . . . . . . . . . . . . . . . . . . . . 174.1 Natural Channel Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . 174.1.1 Channel Pattern and Bank Migration . . . . . . . . . . . . . . . . . . 174.1.2 Vertical Degradation and Channel Geometry . . . . . . . . . . . . . . 194.2 Riprap Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204.2.1 Channel Pattern and Bank Migration . . . . . . . . . . . . . . . . . . 204.2.2 Vertical Degradation and Channel Geometry . . . . . . . . . . . . . . 214.3 Stability Seeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244.3.1 Channel Pattern and Bank Migration . . . . . . . . . . . . . . . . . . 244.3.2 Vertical Degradation and Channel Geometry . . . . . . . . . . . . . . 264.4 Bank Migration and Vertical Degradation in the Protected Reaches . . . . . 274.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 Surface Texture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325.1 Validating Textural Classification . . . . . . . . . . . . . . . . . . . . . . . . 325.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356 Flow Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376.1 Evaluating Flow Model Methods . . . . . . . . . . . . . . . . . . . . . . . . 376.2 Flood Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386.2.1 Natural Channel Experiment Flood Conditions . . . . . . . . . . . . 406.2.2 Riprap Experiment Flood Conditions . . . . . . . . . . . . . . . . . . 426.2.3 Stability Seeding Experiment Flood Conditions . . . . . . . . . . . . 456.3 Mean Annual Flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 486.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507.1 Riprap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507.2 Stability Seeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51vii7.3 Implications for Land Management and Infrastructure . . . . . . . . . . . . . 527.4 Ecological Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557.5 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 567.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 578 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62A Surface GSDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67B Comparison of the constant Manning’s n and variable Manning’s n FlowModels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70C Specific discharges of mean annual flow . . . . . . . . . . . . . . . . . . . . 73viiiList of Tables3.1 Grain size statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103.2 Experimental flow and sediment feed conditions . . . . . . . . . . . . . . . . 143.3 Facies classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155.1 Surface texture after the 10-year flood . . . . . . . . . . . . . . . . . . . . . 346.1 D84 of the three experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . 40ixList of Figures2.1 A Natural Channel After a Flood . . . . . . . . . . . . . . . . . . . . . . . . 62.2 Stability Seeding After a Flood . . . . . . . . . . . . . . . . . . . . . . . . . 72.3 Riprap After a Flood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.1 Components of A-BES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103.2 Grain Size Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113.3 Image of Stability Seeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123.4 Image of Riprap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123.5 Flow regime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134.1 Natural Channel Experiment DEMs . . . . . . . . . . . . . . . . . . . . . . . 184.2 Natural Channel Experiment DEMs of difference . . . . . . . . . . . . . . . 184.3 Cross section of the Natural Channel Experiment . . . . . . . . . . . . . . . 194.4 Longitudinal profile of the Natural Channel Experiment . . . . . . . . . . . . 204.5 Riprap Experiment DEMs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214.6 Riprap Experiment DEMs of difference . . . . . . . . . . . . . . . . . . . . . 224.7 Cross section of the Riprap Experiment . . . . . . . . . . . . . . . . . . . . . 234.8 Longitudinal profile of the Riprap Experiment . . . . . . . . . . . . . . . . . 244.9 Stability Seeding Experiment DEMs . . . . . . . . . . . . . . . . . . . . . . . 254.10 Stability Seeding Experiment DEMs of difference . . . . . . . . . . . . . . . 254.11 Cross section of the Stability Seeding Experiment . . . . . . . . . . . . . . . 264.12 Longitudinal profile of the Stability Seeding Experiment . . . . . . . . . . . 274.13 Bank Migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294.14 Change in bed elevation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305.1 The standard deviation of bed elevation for the textural patches . . . . . . . 335.2 Facies maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356.1 Surface GSD of the Stability Seeding Experiment . . . . . . . . . . . . . . . 386.2 Comparison of the constant Manning’s n and variable Manning’s n StabilitySeeding Flow Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39x6.3 Specific discharge maps of the flood conditions for the Natural Channel Ex-periment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406.4 Comparisons of flood specific discharges, depths and velocities . . . . . . . . 416.5 Specific discharge maps of the flood conditions for the Riprap Experiment . . 426.6 Riprap Experiment shear stresses . . . . . . . . . . . . . . . . . . . . . . . . 436.7 Shear stresses in the reach protected by Riprap . . . . . . . . . . . . . . . . 446.8 Specific discharge maps of the flood conditions for the Stability Seeding Ex-periment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 456.9 Stability Seeding Experiment shear stresses . . . . . . . . . . . . . . . . . . . 466.10 Shear stresses in the reach protected by Stability Seeding . . . . . . . . . . . 476.11 Comparison of mean annual flow specific dischargse . . . . . . . . . . . . . . 487.1 Probability of lateral erosion . . . . . . . . . . . . . . . . . . . . . . . . . . . 547.2 Probability of vertical erosion . . . . . . . . . . . . . . . . . . . . . . . . . . 55A.1 Surface GSD of the Natural Channel Experiment . . . . . . . . . . . . . . . 68A.2 Surface GSD of the Riprap Experiment . . . . . . . . . . . . . . . . . . . . . 69B.1 Comparison of the constant Manning’s n and variable Manning’s n NaturalChannel Experiment flow models . . . . . . . . . . . . . . . . . . . . . . . . 71B.2 Comparison of the constant Manning’s n and variable Manning’s n RiprapExperiment flow models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72C.1 Specific discharge maps of the Natural Channel Experiment mean annual flows 74C.2 Specific discharge maps of the Riprap Experiment mean annual flows . . . . 74C.3 Specific discharge maps of the Stability Seeding Experiment mean annual flows 75xiList of SymbolsSymbol Definition Unitsd water depth LDi percentile of grain diameter Lg acceleration due to gravity L T−2n Manning’s Roughness Coefficent -Q discharge L3 T−1q specific discharge L2 T−1S slope -U mean flow velocity L T−1U∗ shear velocity L T−1w mean channel width Lρ density of water m L−3τ shear stress m L−1 T−2xiiAcknowledgementsThank you to my supervisor Brett Eaton, for pushing me to produce the best science Icould. I have learned so much over the last two-years working with you. Endless thanksgoes to my co-supervisor Marwan Hassan. Thank you for encouraging me and making funof me in equal parts. A huge thank you to Lucy MacKenzie, without whom I would nothave a thesis. Thank you for guiding me through this project and answering my millions ofquestions. I am very lucky to have you as a mentor. Thank you to all of my lab and officemates; Anya Leenman, Will C.D. Booker, Dave Adams, Lauren Vincent and Carina Helm.The feedback, R tech support and the millions of other things you have contributed to thisproject has been invaluable. More importantly, thank you for being the most supportive labmates and friends I could ask for. Working with you has been the highlight of my graduatestudies. This research would not have been possible without the laboratory assistance ofDeanna Shrimpton and Emily Ballon. Thank you for going above and beyond to make sureI had the best experiments possible. Thank you to Dan Moore, for the data for the FishTrap Creek Analysis. I would also like to thank Rick Ketler for his expertise on the workingsof the stream table. I am very grateful for the financial support of GeoMorphix. Withoutthem I would not have been able to undertake this project. In particular, thank you toPaul Villard for your valuable feedback. Thank you to all my friends who have supportedme throughout my graduate studies. You have provided so much encouragement as wellas endless opportunities to take a break from research. Finally, to my parents; thank youfor helping me navigate every bit of life. All of the homemade meals, long road trips andquick stops for ice cream have not gone unappreciated. I am so fortunate to have yourunconditional love and support, even though you still have no idea what I study.xiiiChapter 1Introduction1.1 Hazards Associated with River SystemsProximity to water makes land adjacent to rivers an attractive place to live. Historically,humans have chosen to live near rivers for access to domestic and agricultural water supplies(Fang and Jawitz , 2019). More recently, the recreational and aesthetic benefits of rivers(Wohl et al., 2015), has resulted in the development of homes and other essential infrastruc-ture along river banks (Heede, 1986; Schmetterling et al., 2001). Despite these advantages,building infrastructure alongside rivers makes humans vulnerable to flood risks. Extremeweather conditions, coupled with high run-off rates associated with urbanization can resultin high magnitude flooding events (e.g Pomeroy et al., 2016; Wayand et al., 2015). Suchevents can have devastating effects on riverside communities. For example, the 2013 floodsin Calgary Alberta resulted in the loss of 5 lives and an estimated $6 billion in damages(Biggs). Six years later, the City of Calgary is still working to repair infrastructure damagedby the flood (Biggs). Despite the danger associated with high magnitude flooding events,techniques to mitigate flood-associated damages have not yet evolved to meet the challenges.1.2 Traditional Bank ProtectionTo protect infrastructure located beside a channel, some type of bank protection is typicallyinstalled (e.g. Kondolf , 1994; Frothingham, 2008). Traditional approaches to bank protectionseek to minimize flood damage by controlling channel shape and form. This is typicallyachieved by installing immobile structures (such as riprap and cemented banks) along channelbanks (Baird et al., 2015; Downs and Gregory , 2004). By doing this, engineers impose adesigned form onto the channel and completely eliminate any natural channel widening or1bank migration (Bennett et al., 2013).Traditional bank protection, is generally designed to be resilient to flow conditions up toa specified design flood. For instance, the British Columbia Riprap Installation Guidelines(BCR, 2000) suggest that bank protection should be large enough to resist a design floodwith a return-period of 200-years. It is assumed that the bank protection will be effectivein flows equal to or less than this design flow. However, if bank protection is subject toflow rates greater than the 200-year return period, it is expected to fail (BCR, 2000). Thisleads to a threshold of channel response to differing flow rates. If the flow rate is less thanthe design flood, the channel will maintain its designed shape and have minimal channeladjustment (Bennett et al., 2013; Beechie et al., 2010). However, if the flow rate exceedsthe design flood, the bank protection could fail with catastrophic results (Reid and Church,2015).By engineering the shape and form of a channel, traditional bank protection often elim-inates natural channel adjustment. As aquatic organisms have evolved to survive and evenexploit the disturbances associated with some level of stream channel change (Ligon et al.,1994), bank protection that limits channel migration can impair ecological function. For ex-ample, many fish species rely on channel migration and influxes of sediment (Palmer et al.,2014; Reid and Church, 2015), both of which are essentially eliminated by the lack of bankerosion associated with traditional bank protection.The restriction of a channel’s lateral movement can cause other unexpected consequences,such as an increase in the average flow velocity (Downs and Gregory , 2004). Consequently,frequent and often expensive maintenance is needed to ensure the integrity of the design(Downs and Gregory , 2004; Piegay et al., 2008; Smith et al., 2014). For example, one bankprotection project cost an estimated $125 million to repair after a high magnitude flood event(Snodgrass and Sorel , 2015). While it has been shown that traditional bank protectioncan prevent bank migration and channel widening, its unintended effects on a channel’smorphodynamics and aquatic organisms may make it undesirable for many river engineeringprojects.1.3 An Alternative to Traditional Bank ProtectionRather than engineering a river to have a desired shape and form, this thesis proposes abank protection approach that engages with the processes that govern channel stability. Wedefine channel stability as the ability of a channel to resist erosion. In this approach, weexploit the ability of a river to engineer itself. By working with the processes of erosion anddeposition, it should be possible to develop bank protection that maintains elements of the2natural channel while preventing major channel reorganization and lateral migration, thatwould threaten along-channel infrastructure (Schmetterling et al., 2001).Recent research by Mackenzie and Eaton (2017) has provided insights into the stabilizingrole of the coarsest grains in a channel’s grain size distribution (GSD). When they increasedthe relative abundance of the coarsest grains that are naturally found on the channel bed,the channel had less bank migration, yet still maintained natural channel processes. This isbecause the coarsest grains on the bed only become fully mobilized at significantly highershear stresses than that required to mobilize finer grains (Wilcock and McArdell , 1993).Additionally, in the presence of coarse grains, small grains are 'hidden ' . from high stressesand are less likely to be entrained (Parker et al., 1983). The combination of which leads tothe protection of underlying fine material by the immobile coarse grains from entrainment,stabilizing the bed surface (MacKenzie and Eaton, 2017).The results of these experiments suggest that large grains provide a means of promotingchannel stability (and limiting bank migration), while still allowing natural channel function.Nevertheless, there are difficulties in translating these findings into realizable bank protectionsolutions. In the experiments of MacKenzie and Eaton (2017), the increase in coarse grainabundance was achieved by directly mixing coarse grains into the channel bed, but this isimpractical in natural rivers. Alternatively, we suggest placing coarse grains onto the channelbanks. If bank erosion occurs, the channel will widen, and the coarse grains will be seededinto the channel bed. Once on the bed, these coarse grains have the potential to promotestability and dampen bank migration rates. Thus the input of material utilizes the naturalerosive behaviour of the river channel to stabilize itself. We call this technique 'StabilitySeeding'.1.4 Research Objectives and Thesis StructureWhile the work of MacKenzie and Eaton (2017) showed that placing coarse grains directlyonto a riverbed can increase channel stability and dampen bank migration rates, it is un-known if other methods of delivering coarse grains to a channel will have similar affects.Therefore, the primary objective of this thesis is to determine if Stability Seeding can beused to provide bank protection, while still allowing natural channel adjustments. Thisobjective led us to ask the following research questions;ˆ How does Stability Seeding affect the morphodynamics of a channel?ˆ How does Stability Seeding affect the flow conditions of a channel?3ˆ What are the implications of using Stability Seeding as bank protections in terms ofinfrastructure development along a channel?ˆ What are the ecological implications of using Stability Seeding as a bank protectiontechnique?To investigate these questions, we performed proof-of-concept experiments using scaledphysical models. To do this, we designed a method of incorporating Stability Seeding intoa bank protection technique. To compare the effects of using coarse grains as bank protec-tion to a traditional bank protection technique, we conducted an initial experiment usingRiprap as traditional bank protection. The details of the Stability Seeding technique andthe Riprap design are described in Chapter 2. Chapter 3 details the experimental setup ofthe physical models, as well as the methods of data collection and analysis. Chapters 4-6present the experimental results. Chapter 7 discusses the experimental results and considersthe implications for both building development and the maintenance of ecological diversityin the channel. Chapter 8 synthesizes the previous chapter’s findings and suggest possiblefuture work. the findings of the previous chapters and suggests possible future work. Inaddition to these chapters, Appendix A reports the surface GSDs of the experiments andAppendix B compares two different flow modelling techniques. Finally, Appendix C, showsthe specific discharge maps of the mean annual flows.4Chapter 2Experimental DesignTo determine if Stability Seeding can provide bank protection to a river, three differentscenarios were tested using physical models. First, a Natural Channel Experiment (withoutany bank protection) was run. This allowed us to examine how the model would respondto floods of various magnitudes without the intervention of bank protection. The secondexperiment tested the use of Stability Seeding as bank protection. The final experimentused a traditional bank protection technique to prevent bank migration. The conceptualbasis behind the three experiments is presented below.2.1 Natural ChannelIn gravel-bed rivers with alluvial floodplains, channel degradation is typically dominated bybank migration as the excess shear stress is taken up by the transport of bank material andsubsequent decrease in shear stress (Church, 2006, 2010; Eaton et al., 2017). Therefore, whensubject to high magnitude flooding events, a natural channel (without bank protection) islikely to experience bank migration and channel widening. If infrastructure is located near achannel, it can be vulnerable to disturbance (Figure 2.1). This issue is especially prevalent inurban areas, where there is an increasing demand for development. To protect infrastructurelocated on a channel’s banks, interventions such as bank protection are often desirable.2.2 Stability SeedingThe use of Stability Seeding as bank protection has the potential to provide bank protection,while still allowing some level of natural channel adjustment. Stability Seeding consists ofplacing an abundance of 'Stability Seeds' (the coarsest grains found on the channel bed)5Figure 2.1: A figure depicting how a flood might affect a community located near a channelwithout any bank protection.6Stability seeding grainsChannel grainsFigure 2.2: A figure depicting how a flood might affect a community located near a channelthat is protected by Stability Seeding.on a channel’s banks. As the channel widens, the Stability Seeds fall into the channeland increase stability. This creates a negative-feedback process, where coarse-grains areincorporated into the channel as banks erode, causing the mean shear stress to decrease. Byincorporating Stability Seeds into the channel, the channel is provided with the ingredientsto stabilize. In allowing the channel to engineer itself, there is an expected level of naturalchannel adjustment and bank migration. This channel adjustment is desirable, as it canallow for greater ecological diversity (Schmetterling et al., 2001). However, allowing naturalchannel adjustment requires having a buffer-zone between the channel banks and criticalinfrastructure (Figure 2.2). This buffer-zone allows the channel to migrate without impactingcritical infrastructure.2.3 RiprapUnlike Stability Seeding, traditional bank protection aims to control channel shape andform (Bennett et al., 2013). One of the most common types of traditional bank protection7is Riprap (Heede, 1986). Riprap is composed of large angular blocks that are intentionallysized well beyond what a channel is capable of mobilizing (Reid and Church, 2015). Asa result, Riprap renders channel banks immobile. Because Riprap eliminates all bank mi-gration, infrastructure can be built much closer to the channel than the proposed StabilitySeeding design (Figure 2.3). While Riprap has been shown to be effective at reducing bankmigration (Heede, 1986; Reid and Church, 2015; Schmetterling et al., 2001), it can haveundesirable consequences for the channel. For instance, Riprap has been shown to limitecological variability of a channel as well as cause localized increases in channel velocities(Reid and Church, 2015; Heede, 1986; Schmetterling et al., 2001).Riprap grainsChannel grainsFigure 2.3: A figure depicting how a flood might affect a community located near a channelthat is protected by Riprap.8Chapter 3Methods3.1 Experimental Setup and OperationTo test the research objectives, three physical models were built using the channel designsdescribed in Chapter 2. The first experiment was a Natural Channel Experiment and had nobank protection. The second experiment used Stability Seeding as bank protection. Finally,an experiment was conducted where Riprap was used as bank protection.3.1.1 Scaling and A-BES’ Experiments were conducted on a 1:25 geometrically scaled model of a generic gravel bed.When using models, it is important to have similarity in process between the model andthe prototype. Our models ensured that gravity forces (Froude number, dimensionless shearstress, wave number) were kept similar to the prototype. However, it is impossible to keep allproperties similar when using water as a model fluid (Yalin, 1971). As a result, the Reynoldsnumber was relaxed, while maintaining a turbulent flow regime (Peakall et al., 1996). Themodels were created using the Adjustable-Boundary Experimental System (A-BES) at theUniversity of British Columbia. A-BES is comprised of a 1.5 m wide, 12 m long tilting streamtable, a recirculating water pump, a sediment conveyor belt and a computerized instrumentcart that contains a laser-scanning system capable of creating digital elevation models of thebed surface (Figure 3.1).3.1.2 Initial Channel Conditions and Natural Channel DesignEach of the experiments were started with identical bed conditions. The GSD used in theexperiments is similar to GSD2 used by MacKenzie and Eaton (2017) (Figure 3.2, Table9Figure 3.1: Components of the Adjustable-Boundary Experimental System (A-BES) at theUniversity of British Columbia, where a) is the stream table, b) is the computerized instru-ment cart and c) is a sediment conveyor belt.3.1). When scaled to prototype values, the GSD is similar to a gravel-cobble bed channel.In order to have a hydraulically rough boundary, sediment finer than 0.25 mm was notincluded in the initial bed mixture. This is equivalent to removing the sand componentof the prototype’s GSD. Before any bank protection was installed, a straight rectangularchannel was templated into the well-mixed bed. This initial channel had a width of 0.30 mand 0.02 m depth, allowing for 0.60 m of imposed flat floodplain on either side of the channel.During the experiments the slope was set at 0.02 m/m. The Natural Channel Experimentwas run from these initial channel conditions, without any bank protection.Table 3.1: The D50, D84, D90 and Dmax of the bed material as well as the grain sizes usedin the Stability Seeding and Riprap designs.D50 1.63 mmD84 3.71 mmD90 5.90 mmDmax 8.0 mmStability Seeding 5.6-7.0 mmRiprap 16-23 mm3.1.3 Stability Seeding DesignThe grains selected to be used in the Stability Seeding designs are 5.6-7 mm in size. This isapproximately larger than 90% of the initial bed (D90) (Figure 3.2). To distinguish StabilitySeeds from the initial bed material, the rocks were coloured blue using epoxy, however thedensity of the bed material and the Stability Seeds is comparable. Stability Seeds were100.5 1.0 2.0 5.0 10.0 20.00.00.20.40.60.81.0Particle size (mm)Proportion FinerGSDStability SeedingRiprapFigure 3.2: The Grain Size Distribution of the inital bed as well as the range of grain sizesused in the Stability Seeding and Riprap Designs.11Figure 3.3: A image of the Stability Seeding design used in the physical model.densely packed along the channel banks in a ’mat’ design that was 0.30 m wide and 0.005m deep (Figure 3.3). This allowed for the Stability Seeds to be distributed in a one grainsize thick layer. During the Stability Seeding Experiment, the Stability Seeding design wasinstalled along the middle four metres (4-8 m) of the channel.3.1.4 Riprap DesignDuring the Riprap Experiment, Riprap was installed in the middle 4 m of the channel. TheRiprap design was created as per the British Columbia Ministry of Environments guidelines(2000), and was composed of grains 16-23 mm in size (Figure 3.2). These grains were installedalong the channel banks on a 2:1 slope, with a minimum thickness of 30 mm (Figure 3.4).To ensure stability, the Riprap was toed 10 mm into the channel bed. In prototype values,the Riprap would be considered Class 4 Riprap (BCR, 2000).Figure 3.4: A image of the Riprap design used in the physical model.12Figure 3.5: The flow regime of the three experiments. The yellow lines indicate where flowwas stopped and the stream table was drained in order to generate DEMs of the bed andestimate the GSD of the bed surface.3.1.5 Flow and Sediment Feed ConditionsTo determine how each experimental design responds to floods of various magnitudes, eachexperiment was subject to the same 12 hr flow regime. To allow for the initial armouringof the channel, a flow rate of 1 L/s was run for 5 hrs. This was assumed to be equal tothe bankfull flow or a 2-year flood. The experiments were than subjected to increasing 1hr floods of 1.5 L/s, 2 L/s, and 3 L/s, respectively. Between each flood, the experimentsunderwent 2 hrs of 1 L/s (bankfull) flow (Figure 3.5). To better understand the magnitude ofthe model floods in prototype values, a Hydrograph of Fishtrap Creek was used to estimatethe return periods of the flood. Located 50km north of Kamloops, Fishtrap Creek has atypical British Columbia interior snow-melt flow regime (Andrews , 2010). Fishtrap creekwas selected to be used to estimate return periods as it has similar channel dimensions tothe prototype (S ∼ 0.02 m/m and w ∼ 13 m) and has recorded daily discharge values since1971. Using the Fishtrap Creek hydrograph, it was estimated that the 1.5 L/s flow has areturn period of 10-years, the 2 L/s flow has a return period of 50-years and the 3 L/s hasa return period of 500-years. To better replicate natural channel conditions, sediment wasinputted at the top of the stream table via a conveyor belt (Figure 3.1). Sediment feed wasestimated via the sediment output rates of a preliminary Natural Channel experiment thatwas run in zero-feed conditions. The sediment feed rates are presented in Table 3.2.3.2 Data Collection and AnalysisTwo primary types of data was collected during the experiments: bed elevation data andsurface texture. As mentioned above, A-BES is equipped with a computerized instrument13Table 3.2: The flow and sediment feed conditions during the experiments.Return Period(years)Discharge(L/s)Sediment Feed(g/min)2 (Bankfull) 1.0 11010 1.5 16550 2.0 220500 3.0 331cart that contains a laser scanning system. This scanning system can generate 2mm digitalelevation models (DEM). Between changes in flow, the stream table was drained in order toscan the bed. This resulted in the generation of 7 DEMs per experimental run (includinga DEM of the bed before it was subjected to flow) (Figure 3.5). The DEMs collectedby the laser scanning system were then smoothed using a 7x7 pixel averaging filter and a15x15 averaging filter was applied to fill in any missing data, using the Raster package inR programming. To estimate the bed surface of the channel, photos were taken using aSony Cyber-shot camera through a grid by sampling approach (Bunte and Abt , 2001) at fiveevenly-spaced locations along the stream table.3.2.1 Channel Width and DepthThe DEMs created via the laser scanner were used to assess changes in channel width anddepth. To assess these variables, the channel was delineated manually from the floodplainusing ArcMap 10.6.1 to create polygons of the channel. Using R programming, the poly-gons of the channel outline were used to determine the width of the channel every 0.10 mthroughout the length of the stream table.3.2.2 Surface TextureTo estimate the surface texture of the stream table, facies maps of the middle 4 m (4-8 m)were constructed using photos taken after the 10-year flood. The 10-year flood was selectedas it likely to occur at a relatively high frequency. Therefore, aquatic organisms are morelikely to spend time interacting with this bed surface than the bed surface after highermagnitude flood events.Facies mapping was completed using a variation of BuffingtonandMontgomery′s (1999)method of classifying textural facies in gravel-bed rivers. To create the facies maps, ortho-mosaics of the middle 4 m of the stream table were generated via Agisoft Photoscan (aphotogrammetric processing software). These orthomosaics were then imported into Ar-14Table 3.3: The dominant grain size of the three facies classes.Facies Type Dominant Grain SizeFine <0.71 mmMedium 0.71-4.0 mmCoarse 4.0-11.6 mmcMap 10.6.1, and textural patches were visually identified and classified.As the stream table is a 1:25 scale of a gravel-bed river, there was a high chance ofmisclassifying a textural patch (Buffington and Montgomery , 1999). To minimize misclas-sification, a simple three class system was used. Each textural patch was classified basedon its dominant grain size and was labelled as either coarse, medium or fine (Table 3.3).To determine the accuracy of the visual classification method, the standard deviation ofthe bed surface elevation was calculated for each textural patch. We selected bed elevationstandard deviation to validate the classification of textural patches as it has been correlatedwith the size of grains found on the surface of the bed in previous studies (e.g Aberle andNikora, 2006). The standard deviation of bed surface elevation was calculated by using highresolution DEMs (∼ 0.08 mm) of the channel created from photos of the bed surface inAgisoft Photoscan. The DEMs were then exported into R and the standard deviation ofeach textural patch was calculated.3.2.3 Flow ModellingDue to the shallow flows (d ∼ 5.0 mm) exhibited throughout the experiments, it was difficultto determine flow characteristics using physical measurements. Therefore, water depth, flowvelocity and shear stress were determined using a 2-D numerical flow modelling program,Nays2DH. Flow conditions were reconstructed for the first 5 hours of bankfull flow, as wellas for the three flood events. Flow models of the mean annual flow were also built. As themean annual flow is typically considered to be 1/10th the bankfull flow, the mean annualflow was estimated to be 0.1 L/s.Flow modelling was completed using a 2-step iterative process. For both iterations, theflow models were built by inputting DEMs created by the laser scanner into Nays2DH. Thefirst iteration of the flow models used a constant Manning’s roughness coefficient. While thesecond, more accurate, iteration of the flow models used a variable Manning’s RoughnessCoefficient. This was calculated using the flow depths estimated by the first iteration of theflow model as well as the D84 of each experiment.To estimate the D84, the bed surface of the channel was reconstructed using the cross-sectional photos taken during the experiments. These photos were inputted into Agisoft15Photoscan resulting in five, high resolution (∼ 0.08 mm) orthomosaics per change in flow foreach experiment. The orthomosaics were then exported to R, where digital pebble countswere performed. The GSD of the bed was estimated using the results of the pebble counts.Using a sample size of 500, preliminary analysis of the pebble counts revealed that the D84of the bed did not change significantly over the 12 hours of flow (see Chapter 5 for moredetails). Therefore, only pebble counts taken after the bankfull flow and 450-year flood, wereanalyzed. The D84 for both flow events was determined and an average of the two valueswas taken to find a single D84 for an entire experiment. To calculate the variable Manning’sRoughness Coefficient, the energy slope was calculated for each cell of the first iteration ofthe flow model by;S =τgρd(3.1)The velocity of each cell was then calculated using Ferguson (2007) Variable Power Equa-tion (VPE). This method was selected due its ability to accurately predict Manning’s nvalues, especially at low flows (Ferguson, 2007).U =[a1(d/D84)[(d/D84)5/3] + (a1/a2)2]1/2]U∗ (3.2)where a1=6.5, a2= 2.5, andU∗ = (gdS)1/2 (3.3)Once the velocity was calculated by Ferguson’s VPE, Manning’s n was then calculatedfor every cell that experienced significant flow (q > 0.005 m2/s) byn =d2/3S1/2U(3.4)Once the Manning’s n value was determined for each cell, the flow model was rerun tocalculate more accurate water depths, water elevation, flow velocities and shear stresses foreach cell.16Chapter 4Morphological AdjustmentsThis chapter presents the morphological channel adjustments associated with the NaturalChannel, Stability Seeding and Riprap Experiments. The differences in morphological be-haviour are analyzed to help determine the implications of the bank protection structuresfrom an engineering design standpoint.4.1 Natural Channel Experiment4.1.1 Channel Pattern and Bank MigrationThe Natural Channel Experiment demonstrates how a channel without bank protectionresponds to floods of various magnitudes. Figure 4.1 and Figure 4.2 show the DEMs andDEMs of Difference of the Natural Channel Experiment for bankfull flow, 10-year, 50-yearand 500-year floods. During the bankfull flow, the Natural Channel widened and developed ameandering pattern driven by the erosion of the banks, forming the outside of meander bends.The sediment transported from the banks due to erosion contributed to the development ofalternate bars throughout the channel. During bankfull flow, 11% of the initial floodplainwas impacted by channel erosion. The channel continued to widen throughout the 10-yearand 50-year floods. As the banks continued to erode, the amplitude of the meander bendsand alternate bars increased. Following the 50-year flood, 32% of the initial floodplain hadbeen subject to channel erosion. During the 500-year flood, a braided pattern formed in the8 m above the stream table outlet. The channel braiding resulted in the majority of theinitial floodplain being occupied by the channel and one of the channel threads migrated tothe stream table’s walls. After the 500-year flood, 52% of the initial floodplain had beenaffected by channel erosion.17a)b)c)d)0.060.080.10Figure 4.1: The DEMs of the Natural Channel Experiment after (a) bankfull flow, (b) 10-yearflood, (c) 50-year flood and (d) 500-year flood.a)b)c)d)−0.04−0.020.000.020.04Figure 4.2: The DEMs of Difference of the Natural Channel Experiment between (a) initialchannel and bankfull flow, (b) bankfull flow and 10-year flood, (c) 10-year flood and 50-yearflood and (d) 50-year flood and 500-year flood.184.1.2 Vertical Degradation and Channel GeometryAs presented above, the Natural Channel Experiment experienced extensive channel widen-ing throughout its experimental run. As a result, most of the degradation that occurredduring the Natural Channel Experiment resulted from the erosion of bank material (Fig-ure 4.2). This is demonstrated by Figure 4.3 which shows a cross-section of the NaturalChannel located 6 m from the stream table’s outlet. The channel widened and deepenedin the area outside of the initial rectangular channel. However, there was enough sedimentsupplied via bank erosion that the initial rectangular channel experienced net deposition. Tofurther analyze patterns of aggradation and degradation, longitudinal profiles of the meanbed elevations were created for the three experiments (Figures 4.4, 4.12, and 4.8). Through-out the experimental run, the Natural Channel did not experience any large changes in bedelevation (Figure 4.4). This suggests that overall the vertical degradation associated withthe evacuation of sediment from the channel banks was balanced by the deposition thatoccurred throughout the initial rectangular channel.smooth.spline(e.2$x, e.2$NoTreatmentb_1, spar = 0.3)$xsmooth.spline(e.2$x, e.2$NoTreatmentb_1, spar = 0.3)$yinitial channelbankfull flow0.060.1smooth.spline(e.2$x, e.2$NoTreatmentb_2, spar = 0.3)$xsmooth.spline(e.2$x, e.2$NoTreatmentb_2, spar = 0.3)$ybankfull flow10−year flood0.060.1smooth.spline(e.2$x, e.2$NoTreatmentb_3, spar = 0.3)$xsmooth.spline(e.2$x, e.2$NoTreatmentb_3, spar = 0.3)$y0.060.110−year flood50−year flood0.0 0.2 0.4 0.6 0.8 1.0 1.2smooth.spline(e.2$x, e.2$NoTreatmentb_5, spar = 0.3)$xsmooth.spline(e.2$x, e.2$NoTreatmentb_5, spar = 0.3)$y50−year flood500−year flood0.060.1Distanc  ( )Bed Elevation (m)Figure 4.3: A cross section located of the Natural Channel Experiment located 6m from theoutlet for the initial channel, bankfull flow, 10-year flood, 50-year flood and 500-year flood.190.070.10Distance from Outlet (m)initial channelbankfull flow0.070.10Distance from Outlet (m)bankfull flow10−year flood0.070.10Distance from Outlet (m)10−year flood50−year flood0 2 4 6 8 100.070.10Distance from Outlet (m)50−year flood500−year floodDistance fro  Outlet (m)Mean Bed Elevation (m)Figure 4.4: Longitudinal profiles of the Natual Channel Experiment for the initial channel,bankfull flow, 10-year flood, 50-year flood and 500-year flood. Mean bed elevation wascalculated at 0.10 m intervals.4.2 Riprap Experiment4.2.1 Channel Pattern and Bank MigrationCompared to the Natural Channel Experiment, the Riprap Experiment developed a frag-mented channel pattern. As seen in Figure 4.5, the reach protected by Riprap has a dra-matically different channel pattern to the upstream and downstream reachs. The channelmeandering and braiding experienced by the Natural Channel was virtually eliminated inthe reach where Riprap was installed. As a result, this reach retained its initial channelbanks throughout the experimental run. During the bankfull flow, alternate bars developedin the reach protected by Riprap. These lateral bars disappeared during the 10-year floodand the reach had no observable morphological features. A featureless bed was maintainedthroughout the rest of the experimental run. At the end of the experimental run, only 1%of the floodplain situated along the reach protected by Riprap had been subject to erosion.20The reaches upstream and downstream of the Riprap experienced bank migration anddeveloped more complex channel morphologies than the reach protected by Riprap. Duringbankfull flow, the banks eroded and the same channel planform observed in the NaturalChannel Experiment developed. Alternate bars were observed throughout the bankfull flow.Throughout the 10-year, 50-year and 500-year floods the reach upstream of the Riprapstraightened and widened. Similar to the upstream reach, the reach downstream of theRiprap straightened and widened during the 10-year and 50-year floods. However, duringthe 500-year flood, the reach downstream of the Riprap avulsed. This avulsion can beassociated with the large amount of sediment that was deposited in the downstream reachduring the 500-year flood (Figure 4.6). Due to the avulsion, the majority of the downstreamreach was impacted by channel erosion.a)b)c)d)0.040.060.080.10Figure 4.5: The DEMs of the Riprap Experiment after (a) bankfull flow, (b) 10-year flood,(c) 50-year flood and (d) 500-year flood. The red rectangles indicates the reach protectedby Riprap.4.2.2 Vertical Degradation and Channel GeometryAs the reach protected by Riprap had virtually no bank migration, almost all degradationwas a product of vertical erosion. This is demonstrated in Figure 4.7, which presents a cross-section of the reach protected by Riprap. The loss of bank migration caused a reductionin the amount of sediment being deposited in the reach protected by Riprap (Figure 4.6).Consequently, sediment was transported from the bed without replacement and the formationof depositional features was inhibited.21In contrast, the reaches upstream and downstream of the Riprap behaved similarly to theNatural Channel. These reaches experienced some level of bank erosion and had less verticaldegradation than the reach protected by Riprap (Figure 4.8). During the 500-year event thereach downstream of the Riprap experienced net aggradation. The differences in verticaldegradation rates between the reach protected by Riprap and the upstream and downstreamreaches, resulted in the reach protected by Riprap having a noticeably lower bed elevationthan the other two reaches at the end of the experimental run.a)b)c)d)−0.04−0.020.000.020.04Figure 4.6: The DEMs of Difference of the Riprap Experiment between (a) initial channeland bankfull flow, (b) bankfull flow and 10-year flood, (c) 10-year flood and 50-year floodand (d) 50-year flood and 500-year flood. The red rectangles indicates the reach protectedby Riprap.22smooth.spline(e.2$x, e.2$RiprapTreatment_1, spar = 0.3)$xsmooth.spline(e.2$x, e.2$RiprapTreatment_1, spar = 0.3)$yinitial channelbankfull flow0.040.08smooth.spline(e.2$x, e.2$RiprapTreatment_2, spar = 0.3)$xsmooth.spline(e.2$x, e.2$RiprapTreatment_2, spar = 0.3)$ybankfull flow10−year flood0.040.08smooth.spline(e.2$x, e.2$RiprapTreatment_3, spar = 0.3)$xsmooth.spline(e.2$x, e.2$RiprapTreatment_3, spar = 0.3)$y0.040.0810−year flood50−year flood0.0 0.5 1.0 1.5smooth.spline(e.2$x, e.2$RiprapTreatment_5, spar = 0.3)$xsmooth.spline(e.2$x, e.2$RiprapTreatment_5, spar = 0.3)$y50−year flood500−year flood0.040.08Distance (m)Bed Elevation (m)Figure 4.7: A cross section located of the Riprap Experiment located 4.5m from the outletfor the initial channel, bankfull flow, 10-year flood, 50-year flood and 500-year flood.230.080.12 initial channelbankfull flowDistance from Outlet (m)Mean Bed Elevation (m)0.070.11 bankfull flow10−year floodMean Bed Elevation (m)0.070.11 10−year flood50−year floodMean Bed Elevation (m)0 2 4 6 8 100.070.11Distance fro  Outlet (m)50−year flood500−year floodMean Bed Elevation (m)Figure 4.8: Longitudinal profiles of the Riprap Experiment for the initial channel, bankfullflow, 10-year flood, 50-year flood and 500-year flood. Mean bed elevation was calculated at0.10 m intervals.4.3 Stability Seeding4.3.1 Channel Pattern and Bank MigrationUnlike the reach protected by Riprap, the reach protected by Stability Seeding developeda complex channel morphology (Figures 4.9 and 4.10). Similar to the Natural ChannelExperiment, sediment was evacuated from the meander bends and alternate bars formedthroughout the reach. This channel pattern was maintained throughout the experimentalrun. However, the meander bends in the reach protected by Stability Seeding had a loweramplitude than the meander bends observed in the Natural Channel. At the end of theexperimental run, 20% of the floodplain had been subject to erosion.The reaches upstream and downstream of the Stability Seeding also developed complexchannel morphologies. During the bankfull flow, the upstream and downstream reachesdeveloped a meandering pattern. The upstream reach maintained its meandering patternthroughout its experimental run, while the reach downstream of the Stability Seeding exhib-ited braiding and rapidly widened starting in the 50-year flood. As a result, the reaches had60% more bank migration than the reach protected by Stability Seeding. Across all threereaches, 30% of the initial floodplain was impacted by bank erosion.24a)b)c)d)0.040.060.080.10Figure 4.9: The DEMs of the Stability Seeding after (a) bankfull flow, (b) 10-year flood,(c) 50-year flood and (d) 500-year flood. The red rectangles indicate the reach protected byStability Seeding.a)b)c)d)−0.04−0.020.000.02Figure 4.10: The DEMs of Difference of the Stability Seeding Experiment between (a) initialchannel and bankfull flow, (b) bankfull flow and 10-year flood, (c) 10-year flood and 50-year flood and (d) 50-year flood and 500-year flood. The red rectangles indicate the reachprotected by Stability Seeding.254.3.2 Vertical Degradation and Channel GeometryAll three reaches of the Stability Seeding Experiment had similar rates of vertical degreda-tion. The similarity between vertical degradation rates can be attributed to a similarity inprocesses between the reaches. Where, bank erosion occured throughout the channel andassisted in the development of alternate bars. After the 10-year flood, the longitudinal profilewas similar to the Natural Channel Experiment (Figure 4.12), and there was no significantaggradation or degradational differences in the three reaches. During the 10-year flood, therewas net aggradation in the reach downstream of the reach protected by Stability Seeding.During the 500-year flood, the aggradation experienced downstream of the reach protectedby Riprap disappeared, and the channel experienced net degradation.smooth.spline(e.2$x, e.2$Treatment1d_1, spar = 0.3)$xsmooth.spline(e.2$x, e.2$Treatment1d_1, spar = 0.3)$yinitial channelbankfull flow0.040.1smooth.spline(e.2$x, e.2$Treatment1d_2, spar = 0.3)$xsmooth.spline(e.2$x, e.2$Treatment1d_2, spar = 0.3)$ybankfull flow10−year flood0.040.1smooth.spline(e.2$x, e.2$Treatment1d_3, spar = 0.3)$xsmooth.spline(e.2$x, e.2$Treatment1d_3, spar = 0.3)$y0.040.110−year flood50−year flood0.0 0.5 1.0 1.5smooth.spline(e.2$x, e.2$Treatment1d_5, spar = 0.3)$xsmooth.spline(e.2$x, e.2$Treatment1d_5, spar = 0.3)$y50−year flood500−year flood0.040.1Distance (m)Bed Elevation (m)Figure 4.11: A cross section located of the Stability Seeding Experiment located 4.5m fromthe outlet for the initial channel, bankfull flow, 10-year flood, 50-year flood and 500-yearflood.260.080.11initial channelbankfull flow0.070.10bankfull flow10−year flood0.070.1010−year flood50−year flood0 2 4 6 8 100.070.10Distance from Outlet (m)50−year flood500−year floodDistance fro  Outlet (m)Mean Bed Elevation (m)Figure 4.12: Longitudinal profiles of the Stability Seeding Experiment for the initial channel,bankfull flow, 10-year flood, 50-year flood and 500-year flood. Mean bed elevation wascalculated at 0.10 m intervals.4.4 Bank Migration and Vertical Degradation in theProtected ReachesWhen designing a bank protection project, minimum setback distances for infrastructureare selected based off the position of the channel at the time of design. As a result, itis important to understand how a particular bank protection technique will affect bankmigration and vertical degradation relative to the initial channel. To help provide clarityon how Stability Seeding and Riprap affect channel morphodynamics, bank migration andvertical degradation rates were calculated cumulatively. That is, the results below presenthow much the reaches protected by Stability Seeding and Riprap change from their initialchannel location.From an engineering perspective it is important to determine how effective StabilitySeeding and Riprap are at preventing bank migration. By understanding how much bankmigration Stability Seeding and Riprap allow, engineers can make appropriate decisions27when selecting a bank protection technique. Figure 4.13 presents the cumulative amount ofbank migration that occurred in the reaches protected by Stability Seeding and Riprap. Toprovide context for an unaltered system, the Natural Channel’s bank migration rates arealso shown. As the Natural Channel Experiment was not subject to any bank protectionstructures, the channel had unconstrained bank migration. This resulted in the NaturalChannel Experiment having the highest average bank migration rates of the three experi-ments. In contrast, the reach protected by Riprap had no bank migration during the bankfullflow. The channel banks protected by Riprap remained virtually immobile throughout the10-year, 50-year and 500-year floods. Therefore, the area protected by Riprap had 92-95%less bank migration than the Natural Channel. While placing Stability Seeds on the banksof a reach dampened bank migration rates, it did not completely eliminate bank movement(Figure 4.13). Throughout its experimental run, the reach protected by Stability Seedingdeveloped a complex planform morphology, while still having less bank migration than theNatural Channel. After the 500-year flood, the reach protected by Stability Seeding still had62% less bank migration than the Natural Channel.While the goal of bank protection is not to reduce vertical degradation, it is important tounderstand the influence these bank protection techniques have on vertical degradation. It isespecially important for channels located in areas with buried infrastructure, where increasedvertical degradational rates could increase the probability of these structures being exposedor damaged. Figure 4.14 shows the minimum, mean and maximum change in bed elevationvalues of the reaches protected by Stability Seeding and Riprap. The change in bed elevationof the Natural Channel is also shown. As described above, most of the Natural Channeldegradation occurred through bank erosion. As a result, the Natural Channel Experimenthad the lowest vertical degradation rates of the three experiments. Altering a channel’sbank migration rates, affects the amount of vertical degradation that occurs, especiallyduring high magnitude flood events, where the channel needs to compensate for the highflow rates. During the bankfull flow and 10-year flood, the reaches protected by StabilitySeeding and Riprap had similar vertical degradation rates to the Natural Channel. BecauseRiprap eliminated virtually all bank migration, the reach protected by Riprap compensatedby increased vertical erosion in the larger flood events. This resulted in the reach protectedhaving the highest vertical degradation rates in the 50-year and 500-year floods. After the500-year flood, the reach protected by Riprap had 325% more vertical degradation thanthe Natural Channel. As the reach protected by Stability Seeding had some level of bankmigration, it had less vertical degradation than the reach protected by Riprap. After the500-year flood, the reach protected by Stability Seeding had 40% less vertical degradationthan the reach protected by Riprap.280510152025 a)0510152025 b)0510152025 c)0.0 0.1 0.2 0.3 0.4 0.5 0.60510152025 d)Bank Migration (m)DensityNaturalChannelStabilitySeeding RiprapFigure 4.13: The amount of bank migration from the initial channel after the (a) bankfullflow, (b) 10-year, (c) 50-year and (d) 500-year floods.29−0.050.00NTB_mean_bf$dist.a)−0.050.00NTB_mean_Hr06$dist.b)−0.050.00NTB_mean_Hr09$dist.c)4 5 6 7 8−0.050.00d)Distance from Outlet (m)Change in Bed Elevation (m)NaturalChannelStabiltySeeding RiprapFigure 4.14: The maximum, mean and minumum change in bed elevation of the channelafter the (a) bankfull flow, (b) 10-year, (c) 50-year and (d) 500-year floods.304.5 ConclusionsThe results from Chapter 4 show the morphological adjustments of the Natural Channel,Stability Seeding and Riprap experiments over the bankfull flow and the 10-year, 50-yearand 500-year floods. The Natural Channel Experiment shows how a channel without anybank protection responds to floods of various magnitudes. Throughout the experimentalrun the Natural Channel developed a complex planform morphology that included lateralbars and meander bends. As the Natural Channel had unconstrained bank migration, themajority of its channel degradation was in the form of bank erosion. This resulted in theNatural Channel having the lowest amount of vertical erosion of the three experiments.Installing Riprap virtually eliminates all bank migration from a reach. As the reachcan no longer adjust laterally, the reach compensates by having elevated levels of verticaldegradation. The Riprap experiment had the lowest bank migration rates but the greatestamount of vertical degradation throughout its experimental run. In addition to increasingrates of vertical degradation, eliminating bank migration and erosion resulted in the reachprotected by Riprap developing a uniform featureless bed.Placing Stability Seeds along a reach allows a channel to experience some level of naturalchannel adjustment while dampening bank migration rates. Throughout the experimentalrun, the Stability Seeding experiment developed lateral bars and meander bends, similar tothe Natural Channel Experiment. However the reach protected by Stability Seeding had62% less bank migration than the Natural Channel Experiment. As Stability Seeding doesreduce bank migration rates, it did have more vertical degradation than the Natural Channel.However, the reach protected by Stability Seeding still had less vertical degradation thanthe Riprap Experiment.31Chapter 5Surface TextureThis chapter presents the results of the surface texture analysis of the channel bed after the10-year flood event. These analyses were completed to compare the sediment texture in thereaches protected by Stability Seeding and Riprap to the Natural Channel. The results ofthis section are therefore not representative of the upstream and downstream reaches.5.1 Validating Textural ClassificationTo determine differences in the surface texture of the three experiments, facies maps werecreated for the middle 4 m of the stream table. These facies maps were made using ortho-mosaics of the bed surface after the 10-year flood. As described in Chapter 3, a 10-year floodis likely to occur at a relatively high frequency. Therefore, aquatic organisms are likely tospend more time interacting with this bed surface than the bed surface associated with thehigher magnitude flooding events.Textural patches were visually identified and classified as either a fine, medium or coarsefacies type. Classification was based on the dominant grain size of the textural patch (formore information, see Chapter 3). To assess the accuracy of the visual classifications, thestandard deviation of bed elevation was calculated for each textural patch. Figure 5.1 showsthat coarser facies types corresponded to a larger standard deviation of surface roughness.This result is in agreeance with recent research that shows the standard deviation of bed ele-vation can be correlated with surface roughness and grain size (Aberle and Nikora, 2006). Inaddition, each facies type had its own discrete range of standard deviations of bed elevations(Figure 5.1). This suggests that the visual classification of facies types not only provided anaccurate method of determining the relative roughness of a textural patch, it also allowedfor the differentiation of facies types.32fine medium coarse2e−046e−041e−03Facies TypeStandard deviation of bed elevation (m)Figure 5.1: The standard deviation of bed elevation for the fine, mediun and coarse texturalpatches.5.2 ResultsFigure 5.2 presents the facies maps of the Natural Channel, Stability Seeding and RiprapExperiments after the 10-year flood. From 4-8 m the Natural Channel had a relativelyheterogenous surface texture, and it did not exhibit a cross-sectional pattern of sedimenttexture. However, the majority of the coarse patches appear to be located on the outsideof meander bends (Figure 5.2). Approximately half of the Natural Channel had a coarsesurface texture (Table 5.1). The medium and fine facies types composed 32% and 20% ofthe channel and were located throughout the channel.As discussed in Chapter 4, the reach protected by Riprap maintained its initial rectan-gular shape throughout its experimental run. While there is no clear spatial relationshipbetween the surface texture and channel morphology, the reach protected by Riprap didexhibit a strong preference towards a homogenous medium surface texture as you movedownstream. This resulted in a reduction in the coarse and fine facies types; compared tothe Natural Channel, the reach protected by riprap had approximately 50% less coarse andfine facies.While there is a slight decrease in the abundance of coarse material in the channel (15%),33the reach protected by Stability Seeding had relatively similar proportions and pattern ofsurface texture to the Natural Channel. Unlike the reach protected by Riprap, the reachprotected by Stability does not experience any clear cross-sectional pattern in surface texture.Table 5.1: The proportion of each textural class after the 10-year flood.Fine Medium CoarseNaturalChannel0.21 0.32 0.47StabilitySeeding0.18 0.47 0.35Riprap 0.08 0.66 0.2634FlowfinecoarsemediumNatural Channel ExperimentStability Seeding ExperimentRiprap ExperimentFigure 5.2: Facies maps of the bed surfaces of the middle 4 m of the Natural Channel,Stability Seeding and Riprap Experiments after the 10-year flood5.3 ConclusionsThe results of Chapter 5 compare the bed surface texture of the reaches protected by Sta-bility Seeding and Riprap to the Natural Channel. To estimate the surface texture of eachexperiment, facies maps were created for the middle 4 m of the channel for each experiment.Textural patches were visually identified and classified as either a fine, medium or coarsefacies type. In order to validate the visual classification system, the standard deviation ofsurface roughness was calculated. Based on these standard deviations, it was found that thevisual classification system allows an accurate method of determining facies types.Examining the facies maps, it was found that both the Natural Channel and the reach35protected by Stability Seeding developed a heterogenous bed with coarse textural patchesalong the outside of the meander bends. Compared to the Natural Channel Experiment,the reach protected by Riprap had a reduction in both coarse and fine surface texture. Thisresulted in the channel having a relatively homogenous bed, where the majority of the fineand coarse sediment textures were located in the first metre of the Riprap installation.36Chapter 6Flow ModellingTo better understand how bank protection structures affect channels, the flow conditionsof the Natural Channel, Stability Seeding and Riprap Experiments need to be examined.Nays2DH, a 2-D numerical flow modelling program, was used to reconstruct the flow con-ditions of each experiments This chapter evaluates the methods used to create the flowmodels. The flow conditions of the bankfull flow, 10-year, 50-year and 500-year floods arethen reported. Finally, the results of the mean annual flow models are reported.6.1 Evaluating Flow Model MethodsAs the stream table experienced shallow flows and a rapidly evolving bed, it was difficultto determine flow characteristics using physical measurements. Therefore, a 2-D numericalflow modelling program (Nays2DH) was used to estimate water depth, velocity and shearstress. Flow modelling was completed using a 2-step iterative process. The first iterationof the flow models was created using DEMs generated by the laser scanning system and aconstant Manning’s roughness coefficient. The second iteration of the flow model used avariable Manning’s roughness coefficient. The variable Manning’s roughness coefficient wascalculated using the flow depths estimated by the first iteration of the flow model as wellas the D84 determined via digital pebble counts. As there was no significant change in thesurface D84 between the bankfull flow and the 500-year flood (Figure 6.1), each experimentused a single D84 to calculate the variable Manning’s roughness coefficient. The D84 usedto calculate the Manning’s roughness coefficient for each experiment was the average of thebankfull flow and 500-year flood D84 values (Table 6.1).Incorporating a variable Manning’s roughness coefficient into the Nays2DH flow modelsallowed for a more accurate estimation of water depth and velocity. Figure 6.2 presents therelationship between the specific discharge of the constant Manning’s roughness coefficient371 2 50.00.20.40.60.81.0Grain Size (mm)Percent Finerbankfull flow500−year floodFigure 6.1: The Grain Size Distribution of the bed surface of the Stability Seeding Experi-ment for the bankfull flow and 500-year flood.and the specific discharge of the variable Manning’s roughness coefficient flow model ofthe Stability Seeding Experiment (figures for the Natural Channel and Riprap Experimentsare displayed in Appendix B). Compared to the variable Manning’s roughness coefficientflow model, the constant Manning’s roughness coefficient flow model tends to overestimateareas of low specific discharge and underestimate areas of high specific discharge. As aresult, the variable Manning’s roughness coefficient flow model estimated a wider rangeof specific discharges than the constant Manning’s roughness coefficient flow model. Thevariable Manning’s roughness coefficient flow models were used to analyze the flow conditionsof the experiments.6.2 Flood ConditionsUsing the variable Manning’s roughness coefficient described above, flow models of the bank-full flow, 10-year, 50-year, and 500-year floods were created for the three experiments. Spe-cific discharge maps of the Natural Channel, Stability Seeding and Riprap Experiment arepresented in Figures 6.3, 6.5 and 6.8.380.000 0.002 0.004 0.006 0.008 0.010050100150200250300Specific Discharge (m2 s)DensityConstant Manning's nVariable Manning's na)0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.0140.0000.0040.0080.012Constant Manning's nVariable Manning's nb)Figure 6.2: Comparison of the specific discharges of the constant Manning’s roughness co-efficent and the variable Manning’s roughness coefficent flow models for the bankfull flowof the Stability Seeding Experiment. (a) is a density plot of specific discharges, while (b)compares the relation between the specific discharges of the two flow model methods. Theblack line indicates a 1:1 relation.39Table 6.1: The estimated D84 values inputted into the variable Manning’s roughness coeffi-cent Flow Models.D84 (mm)NaturalChannel2.23StabilitySeeding3.48Riprap 2.976.2.1 Natural Channel Experiment Flood ConditionsAs discussed in Chapter 4, the Natural Channel exhibits a meandering planform patternacross all flow events. These meanders resulted in the Natural Channel having a patternof specific discharges alternating on either side of the channel, where the outside of me-ander bends had a higher specific discharge than other areas of the channel. The patternand magnitude of specific discharges remained relatively consistent throughout the NaturalChannel Experiment (Figure 6.3). As the floods increased in magnitude the range of specificdischarges also increased (Figure 6.4).a)b)c)00.0050.010.0150.02d)Figure 6.3: Specific discharge maps of the Natural Channel Experiment for the (a) bankfullflow, (b) 10-year flood, (c) 50-year flood and (d) 500-year flood.400.0000.0050.0100.0150.020a) b) c) d)Spec. Discharge (m2s)0.000.010.020.030.04Depth (m)0.00.20.40.60.8Velocity (m/s)NaturalChannelStabilitySeeding RiprapFigure 6.4: Comparisons of specific discharges depths and velocities for the (a) bankfull flow,(b) 10-year flood, (c) 50-year flood and (d) 500-year flood.416.2.2 Riprap Experiment Flood ConditionsAt bankfull flow, the Riprap Experiment had alternating bars and exhibited a pattern ofvery slightly alternating specific discharges (Figure 6.5). During the 10-year flood the alter-nating bars disappeared, and the Riprap Experiment maintained a relatively straight anduniform channel the rest of the experimental run. As a result, there is minimal cross-sectionalvariation in specific discharge throughout the 10-year, 50-year and 500-year floods. Duringthe bankfull flow and 10-year flood, the Riprap Experiment experienced similar specific dis-charges as the Natural Channel Experiment (Figure 6.4). However, the Riprap experiencedhigher specific discharges than the Natural Channel Experiment in the 50-year and 500-yearfloods. This increase in specific discharge is concentrated in the reach where Riprap wasinstalled (Figure 6.5). As seen in Figure 6.6, the reach protected by Riprap has higher shearstresses than the reaches upstream and downstream of it. It also experiences shear stresseslarger than what was observed in the Natural Channel (Figure 6.7).a)b)c)00.0050.010.0150.02d)Figure 6.5: Specific discharge maps of the Riprap Experiment for the (a) bankfull flow, (b)10-year flood, (c) 50-year flood and (d) 500-year flood. The rectangle indicates the reachwhere Riprap was installed.420.00.20.40.60.8Shear StressDensitya)0.00.20.40.60.8Shear StressDensityb)0.00.20.40.60.8Shear StressDensityc)0 1 2 3 4 5 60.00.20.40.60.8Shear StressDensityd)Shear Stress (N m2)DensityUpstreamReachProtectedReachDownstreamReachFigure 6.6: Comparision of shear stresses between the reach protected by Riprap and thereaches the upstream and downstream reaches during the (a) bankfull flow, (b) 10-year flood,(c) 50-year flood and (d) 500-year flood.430.00.20.40.60.8Shear StressDensitya)0.00.20.40.60.8Shear StressDensityb)0.00.20.40.60.8Shear StressDensityc)0 2 4 6 80.00.20.40.60.8Shear StressDensityd)Shear Stress (N m2)DensityNaturalChannel RiprapFigure 6.7: Comparisons of shear stress in the reach protected by Riprap to the NaturalChannel during the (a) bankfull flow, (b) 10-year flood, (c) 50-year flood and (d) 500-yearflood.446.2.3 Stability Seeding Experiment Flood ConditionsThe Stability Seeding experiment maintained a pattern of specific discharges that alternateon either side of the channel throughout its experimental run (Figure 6.8). Despite having asimilar flow pattern to the Natural Channel, the Stability Seeding Experiment experiencedlarger water depths than the Natural Channel (Figure 6.4). These water depths contributedto the Stability Seeding Experiment having greater specific discharges than the NaturalChannel throughout the experimental run. While the Stability Seeding Experiment hasthe largest specific discharges in the bankfull flow and 10-year flood, it had lower specificdischarges than the Riprap Experiment during the 50-year and 500-year floods. Unlike theRiprap Experiment, placing Stability Seeding on a channel’s banks does not appear to havea localized effect on flow conditions (Figure 6.8). Throughout the experimental run, thethree reaches have relatively similar distributions of shear stress (Figure 6.9). In addition,it was found that the distribution of shear stress in the reach protected by Stability Seedingwere very similar to the Natural Channel Experiment (Figure 6.10).a)b)c)00.0050.010.0150.02d)Figure 6.8: Specific discharge maps of the Stability Seeding Experiment for the (a) bankfullflow, (b) 10-year flood, (c) 50-year flood and (d) 500-year flood. The rectangle indicates thereach where Stability Seeding was placed.450.00.20.40.60.8Shear StressDensitya)0.00.20.40.60.8Shear StressDensityb)0.00.20.40.60.8Shear StressDensityc)0 1 2 3 4 5 60.00.20.40.60.8Shear StressDensityd)Shear Stress (N m2)DensityUpstreamReachProtectedReachDownstreamReachFigure 6.9: Comparision of shear stresses between the reach protected by Stability Seedingand the reaches the upstream and downstream reaches for the (a) bankfull flow, (b) 10-yearflood, (c) 50-year flood and (d) 500-year flood.460.00.20.40.60.8N = 1252   Bandwidth = 0.1403Densitya)0.00.20.40.60.8N = 1719   Bandwidth = 0.1531Densityb)0.00.20.40.60.8N = 2908   Bandwidth = 0.1261Densityc)0 2 4 6 80.00.20.40.60.8N = 3016   Bandwidth = 0.1655Densityd)Shear Stress (N m2)DensityNaturalChannelStabilitySeedingFigure 6.10: Comparisons of shear stress in the reach protected by Stability Seeding to theNatural Channel during the (a) bankfull flow, (b) 10-year flood, (c) 50-year flood and (d)500-year flood.476.3 Mean Annual FlowsTo better understand how the use of bank protection would affect average flow conditions,flow models were created using an estimated mean annual flow (0.1 L/s) and the DEMs ofthe bankfull flow, 10-year, 50-year and 500-year floods. That is, the DEM associated with aflood was taken and used as the basis for a flow model run with the discharge of the estimatedmean annual flow. The specific discharge maps of the mean annual flows are displayed inAppendix C. When describing the results from the mean annual flow models, they arereferred to by the DEM that was inputted into the model. For instance the mean annualflow associated with the 500-year flood channel morphology will be described as the 500-yearmean annual flow. During the bankfull mean annual flow, all three experiments have a similardistribution of specific discharge values (Figure 6.11) . The Natural Channel and StabilitySeeding Experiments continued to have similar distributions of specific discharges throughthe 10-year, 50-year and 500-year mean annual flows. During these three mean annual flows,the Riprap Experiment experienced a narrower range of specific discharges. Compared tothe Natural Channel and Stability Seeding Experiments, the Riprap Experiment had a lowerabundance of areas with shallow, high velocity flows.0.00 0.10 0.200.0000.0040.0080.012NTB_chan.manflow_2$VelocityNTB_chan.manflow_2$Deptha)0.00 0.10 0.200.0000.0040.0080.012NTB_chan.manflow_3$VelocityNTB_chan.manflow_3$Depthb)0.00 0.10 0.200.0000.0040.0080.012NTB_chan.manflow_5$VelocityNTB_chan.manflow_5$Depthc)0.00 0.10 0.200.0000.0040.0080.012NTB_chan.manflow_7$VelocityNTB_chan.manflow_7$Depthd)V locity (m/s)Depth (m)NaturalChannelStabilitySeeding RiprapFigure 6.11: Comparisons of specfic discharge during the (a) bankfull mean annual flow, (b)10-year mean annual flow, (c) 50-year mean annual flow and (d) 500-year mean annual flow.486.4 ConclusionsIn order to estimate the flow conditions during the floods, a 2-D numerical modelling system(Nays2DH) was used. These models incorporated a variable Manning’s roughness coefficientin order to get a more robust distribution of specific discharges.Due to the development of a planform meandering pattern, the Natural Channel had apattern of alternating specific discharges throughout the floods. As discussed in Chapter 4,placing Stability Seeding along a channel bank still allows for the development of channelmeandering. Therefore, during the floods, the Stability Seeding Experiment had a similarpattern of specific discharges to the Natural Channel. In contrast, the reach protected byRiprap did not develop a complex channel morphology. As a result, the Riprap Experimentdid not experience alternating patterns of specific discharge. In addition, during the floodevents, the reach protected by Riprap had higher specific discharges and shear stresses thanits upstream and downstream reaches.To further understand the effect of bank protection on average flow conditions, models ofthe mean annual flow were created. Using these flow models it was found that the StabilitySeeding Experiment has a similar spatial distribution of water depth and velocity as theNatural Channel, while the Riprap Experiment may reduce the abundance of shallow, highvelocity flows.49Chapter 7DiscussionAs shown in the previous chapters, Stability Seeding and Riprap have two fundamentallydifferent approaches to preventing bank migration. Riprap aims to protect infrastructurelocated near a channel by controlling the channel shape and form. In contrast, StabilitySeeding works with the processes that govern channel stability to prevent bank migration.This chapter discusses the differences in how Stability Seeding and Riprap affect channelmorphodynamics and the implications of these two bank protection techniques on land man-agement and aquatic habitat. In addition, this chapter describes the limitations of theexperimental design.7.1 RiprapWhen using Riprap as bank protection, engineers control channel form by placing largeimmobile grains along a channel’s banks (Reid and Church, 2015). As a result, reaches pro-tected by Riprap have virtually no bank migration and retain their initial channel shape.While this may make Riprap effective at protecting infrastructure located along a chan-nel’s banks, it can have unintended consequences for a channel’s morphodynamics and flowcharacteristics.At the end of its experimental run, the reach protected by Riprap had 76% less bankmigration than the Natural Channel. The lack of bank migration was compensated forby vertical degradation and the reach protected by Riprap underwent 325% more verticaldegradation than the Natural Channel. The increase in vertical erosion can be partiallyexplained by the relationship between channel width and depth described in the continuityequation.Q = w · d · U (7.1)50As the reach protected by Riprap experienced very little widening throughout the ex-perimental run, the channel compensated for increasing discharges by increasing channeldepth.The restriction of bank migration and erosion associated with Riprap, reduces the amountof sediment supplied to a reach. This loss of sediment supply causes sediment to be trans-ported from the bed without replacement. As a result, there is an increase in verticaldegradation as well an absence of observable depositional features (such as bars). In naturalchannels excess energy is dissipated via variation in channel morphology (Kondolf , 1994).Unlike the Natural Channel Experiment, the reach protected by Riprap maintained a rela-tively straight channel shape with no obvious planform morphology. As a result, the reachprotected by Riprap had less energy dissipated than the Natural Channel, causing an in-creased average shear stress in the reach protected by Riprap. These increased shear stressescould possibly contribute to the increased vertical degradation. Furthermore, variation in achannel’s morphology can produce local variability in bed surface texture (Lisle and Madej ,1992). The absence of this variation could explain the relatively homogenous bed surfacetexture that was observed throughout the reach protected by Riprap.The results of the Riprap Experiment support field studies that examine the effects ofRiprap on a channel (e.g. Reid and Church, 2015; Schmetterling et al., 2001). As discussedin Chapter 1, by eliminating bank migration from a reach, Riprap reduces the amountof sediment supplied to the channel (Frothingham, 2008). By reducing bank migrationand sediment supply, Riprap fundamentally changes a channel’s shape and form. A reachprotected by Riprap is likely to have a simpler channel morphology and a more homogenoussurface texture than a natural channel.7.2 Stability SeedingUnlike Riprap, Stability Seeding does not prevent bank migration by controlling a channel’sshape and form. Instead, Stability Seeding aims to prevent bank migration by engagingwith the processes that govern channel stability. As described in Chapters 1 and 2, StabilitySeeding consists of placing coarse grains on a channel’s banks. As the channel widens theStability Seeds fall into the channel, increasing the abundance of coarse grains located onthe channel bed to promote channel stability.Placing Stability Seeding along a reach reduces bank migration by 50%, while still allow-ing bank erosion and natural channel development. As bank erosion still occurs in a reachprotected by Stability Seeding, there is unlikely to be a loss of sediment supply. By main-taining the processes of channel erosion and deposition, a reach protected by Stability can51still develop erosional and depositional features (such as meander bends and lateral bars).Stability Seeding allows for the development of a complex channel. Therefore, energy candissipate through changes in bedforms and channel pattern (Kondolf , 1994). As a result,Stability Seeding has similar flow characteristics and shear stresses as the natural channel.The variation in channel planform morphology, also allows a reach protected by StabilitySeeding to have a heterogenous surface texture.While Stability Seeding does allow for the development of a channel pattern similar towhat is seen in a natural channel, it does reduce some amount of lateral channel movement.As a result, a reach protected by Stability Seeding may compensate for its lack of lateralchannel adjustment with increased vertical degradation. The reach protected by StabilitySeeding had 200% more vertical degradation than the Natural Channel, which is still lessvertical degradation than the reach protected by Riprap experienced.The results of the Stability Seeding Experiment suggest that placing Stability Seedsalong a channel’s banks dampens bank migration, while still allowing for some level ofnatural channel adjustment. The Stability Seeding experiment demonstrates that while bankmigration rates were reduced in the reach protected by Stability Seeding, the reach still had asteady supply of sediment. This resulted in the reach protected by Stability Seeding having acomplex channel morphology, similar to what would be expected in a natural channel. Thususing Stability Seeding has a similar effect on channel stability as MacKenzie and Eaton’s(2017) method of placing coarse grains directly onto the channel bed.7.3 Implications for Land Management and Infrastruc-tureAs discussed above, Stability Seeding and Riprap have very different affects on a channel.Therefore both techniques have different implications for land management and infrastruc-ture development. Figure 7.1 presents the probability of the bank migrating from the initialchannel for the Natural Channel, Stability Seeding and Riprap Experiments. This can beused to determine the probability of infrastructure located a certain distance away from theinitial channel being disturbed during a flood event. By installing Riprap, virtually all bankmigration is eliminated. A reach protected by Riprap has a very low chance of disturbinginfrastructure along its banks. As a result, restoration guides often recommend using Riprapto protect critical infrastructure situated directly on the channel bank (i.e. bridges) (Reidand Church, 2015).The primary goal of bank protection is to prevent damage to infrastructure located on a52channel’s banks. As a result, the danger of infrastructure buried under the channel being ex-posed or damaged is often overlooked (Eaton et al., 2017). However, when developing alonga channel it is important to consider the risks of buried infrastructure (such as pipelines)being exposed during flooding events. The reach protected by Riprap has the highest prob-ability of buried infrastructure being exposed (Figure 7.2). The reach protected by StabilitySeeding had less vertical degradation than the reach protected by Riprap. Therefore, inareas where infrastructure is located both along a channel’s banks and under the channel,Stability Seeding may be a preferable alternative to traditional bank protection.Besides traditional bank protection projects, Stability Seeding could also be used tosupport other river restoration projects. For example, Stability Seeding could support theestablishment of riparian vegetation. When rehabilitating a riparian zone, trees and othervegetation are planted. Once established, this vegetation provides bank stability to a channelas well as valuable ecological habitat (Bancroft and Zielke, 2002). However, before thevegetation has significant root development, it is vulnerable to flooding events and is oftendestroyed. Incorporating Stability Seeding into a riparian restoration project could reducethe risk of bank migration impacting unestablished vegetation. Therefore, using StabilitySeeding in a riparian restoration project could increase the probability of a project beingsuccessful.There are trade-offs between the amount of lateral and vertical adjustment that occursin a channel. When selecting a bank protection technique, it is important to understand thegoals and constraints of the river engineering project. Stability Seeding could be a viablemethod of protecting infrastructure along a channel’s banks, while simultaneously limitingthe probability of buried infrastructure. The natural channel adjustment associated withStability Seeding can provide ecological benefits to the channel (discussed below). However,there are situations where using Stability Seeding as bank protection is simply not possible.Stability Seeding relies heavily on a buffer-zone that allows a channel to adjust withoutendangering critical infrastructure. If a buffer-zone cannot be created, then Riprap may bea better method of bank protection.530.0 0.2 0.4 0.60.00.20.40.60.81.0L. Distance from Initial Bank (m)Probability of Erosion10−year flood10 l od0.0 0.2 0.4 0.60.00.20.40.60.81.0L. Distance from Initial Bank (m)Probability of Erosion50−year flood50 l od0.0 0.2 0.4 0.60.00.20.40.60.81.0L. Distance from Initial Bank (m)Probability of ErosionControlStabilitySeedingRiprap500−year flood500 fl odFigure 7.1: The probability of lateral erosion from the initial bank during the 10-year, 50-year and 500-year floods. The blue line indicates the minimum setback distance needed whenusing Stability Seeding.540.000 0.010 0.020 0.0300.00.10.20.30.40.50.60.7V. Distance from Initial Bed (m)Probability of Erosion10−year flood10 l od0.000 0.010 0.020 0.0300.00.10.20.30.40.50.60.7V. Distance from Initial Bed (m)Probability of Erosion50−year flood50 l od0.000 0.010 0.020 0.0300.00.10.20.30.40.50.60.7V. Distance from Initial Bed (m)Probability of Erosion ControlStabilitySeedingRiprap500−year flood500 fl odFigure 7.2: The probability of vertical erosion during the 10-year, 50-year and 500-yearfloods.7.4 Ecological ImplicationsOver the past 20-years, engineers have recognized the importance of creating ecologicallydiverse channels and there has been a call for the development of more ecologically friendlyengineering techniques (Palmer et al., 2010; Wohl et al., 2015). As bank protection is ofteninstalled in channels that are vulnerable to ecological changes (Schmetterling et al., 2001),it is important to consider how Stability Seeding and Riprap impact a channel’s ecologicaldiversity.The abundance and diversity of organisms in an ecosystem is controlled by the number ofunique microhabitats available (Reid and Church, 2015). In river systems, microhabitats areproduced by variations in the physical channel state (channel morphology, surface texture),flow characteristics (water depth, water velocity), and biological processes (nutrient inputs,in-stream vegetation) (Beechie et al., 2010). By narrowing the range of one of these variables,the number of unique microhabitats decreases, which may result in a decrease in the diversityand abundance of aquatic organisms.55Compared to the Natural Channel, the reach protected by Riprap had a homogenoussurface texture. Variation in surface textures and associated sediments has been shown to beessential to the success of aquatic organisms, and many aquatic organisms require a specificrange of surface textures (Beechie et al., 2008; Kondolf et al., 2008). For example, spawningsalmonids requires sediment to be small enough that nests can be evacuated (Kondolf andWolman, 1993). However, sediment that is too fine may prevent the circulation of oxygenand endanger embryo survival (Kondolf , 2000). As Riprap reduces the variability in sedimenttexture, it could reduce the number of viable salmonid nesting sites.Additionally, it was found that Riprap decreases the abundance of shallow, high velocityflows. These flows play a critical role for many organisms. Returning to the salmonid exam-ple, shallow, high velocity flows are used to generate invertebrate drift that salmonids relyon as a source of nutrient input (Tamminga and Eaton, 2018). By decreasing the abundanceof shallow, high velocity flows, there may be a decrease in food supply for salmonids.Finally, the elimination of bank migration due to Riprap could reduce the volume ofwood input to channels (Schmetterling et al., 2001). In-stream wood is essential aquatichabitat and provides shelter for juvenile salmonids (Hafs et al., 2014). In the absence ofwood juvenile salmonids and other organisms are more vulnerable to predation.Overall, using Riprap as bank protection could decrease the ecological diversity of achannel. In contrast, the reach protected by Stability Seeding exhibited similar physicaland flow characteristics to the Natural Channel. The reach protected by Stability Seedinghad a wider range of mean annual flows, and a more variable sediment texture than thereach protected Riprap. Stability Seeding does allow for some bank migration and thereforeis expected to have a larger volume of in-stream wood than a reach protected by Riprap.The variation in sediment textures, flow characteristics and the abundance of in-stream woodcould allow a reach protected by Stability Seeding to have a large number of unique habitats.Using Stability Seeding therefore has more ecological benefits than Riprap.7.5 LimitationsWhile the results of these proof-of-concept experiments suggest that Stability Seeding couldbe used as an alternative to traditional bank protection, it is important to acknowledgethe limitations of using physical models. These experiments were conducted using the A-BES stream table which is a simplified experimental environment. As a result, much ofthe physical complexity found in natural river systems is ignored. For instance, vegetationlocated along a channel’s banks has been found to increase bank cohesion (Eaton et al.,2010). In additions, landslides and the clearing of land near a channel can cause an increase56in the sediment supply (Johnson, 2006). Changes to bank cohesion and sediment supply canhave significant affects on a channel’s morphology (Church, 2006; Palucis and Lamb, 2017).Therefore, using Stability Seeding in areas with increased bank cohesion or sediment supplymay not produce results congruent with the experiments detailed in this thesis.The three experiments were started from a narrow rectangular channel. During thefloods, Stability Seeds were able to disperse throughout the channel bed. The relativelyconsistent cross-sectional dispersal of Stability Seeds may be partially due to the relativelysmall initial channel width (∼ 0.30 m ∼). In wider channels, the Stability Seeds may notbe able to reach the middle of the channel and therefore the Stability Seeds may not beas affective at reducing bank migration. Further tests are required to determine the roleStability Seeding has on different channel sizes.The Natural Channel, Stability Seeding and Riprap Experiments were conducted usingmodels of a generic gravel bed river. Therefore, the results of these experiments do notprovide an explanation for how Stability Seeding would affect other channel types. In gravelbed channels, depositional features are formed around a coarse grained nucleus (Ashmore,1991). By adding Stability Seeds to a channel, bars may be more likely to form, and thechannel may become more stable. In contrast, labile channels form depositional featuresthrough the lateral accretion of fine grained sediment (Church, 2006; Peakall et al., 2007).As a result of this difference in process, Stability Seeding may not be as affective in labilechannels. While Stability Seeding may not be as affective at reducing bank migration in alabile channel than a gravel bed channel, labile channel’s typically have less lateral instabilitythan gravel bed channels (Church, 2006). Therefore, labile channels are less likely to needbank protection than gravel-bed channels.7.6 ConclusionsStability Seeding dampens migration rates, while still allowing natural channel adjustment.As a result, using Stability Seeding as bank protection provides benefits in terms of protectingburied infrastructure and ecological diversity. The findings of this chapter suggest that usingStability Seeding instead of Riprap reduces the probability of buried infrastructure beingexposed. In addition, the natural channel adjustment associated with Stability Seedinghas positive implications for ecological diversity. It was found that the reach protected byStability Seeding had more variation in its surface texture, channel morphology and flowsthan the reach protected by Riprap. This variation would likely allow a diverse range oforganisms to exist within the channel. If a buffer-zone can be developed, Stability Seedingcan protect infrastructure while still allowing natural channel and ecosystem function.57By installing Riprap along a reach, virtually all bank migration is eliminated. This allowsfor infrastructure built close to the banks to have a relatively low chance of being impactedby bank migration. While Riprap may maximize the amount of developable land along achannel, it has unintended consequences. It was found that the reach protected by Ripraphas the highest probability of exposing buried infrastructure. Installing Riprap in a reachalso eliminates natural channel variation, which can reduce the ecological diversity of thereach.While the results of these proof-of-concept experiments provide strong evidence that Sta-bility Seeding could be used as an alternative to traditional bank protection, it is importantto acknowledge the limitations of the experimental setup. When working with any experi-mental system, many natural components of the natural system are ignored. The additionof these components could result in fundamentally different results. In addition, this the-sis only investigated the affect Stability Seeding has on narrow gravel-bed channels. Otherchannel types have fundamentally different depositional and erosional processes. Therefore,the results of the experiments may not apply to other channel types. More experiments needto be launched to determine how Stability Seeding affects other channel types.58Chapter 8Concluding RemarksAn increase in the frequency of high magnitude flooding events coupled with the growingscarcity of developable land (Bernhardt and Palmer , 2007), has resulted in an increasedneed for the protection of infrastructure located along channels. Traditional bank protectionstructures aim to protect infrastructure located along a channel’s bank by controlling a river’sshape and form. By creating immobile banks, traditional bank protection eliminates all bankerosion and reduces the amount of sediment available to a channel (Reid and Church, 2015;Schmetterling et al., 2001). This approach ignores that rivers are dynamic systems thatnaturally adjust to changes in flow (Church, 2006; Leopold and Maddock , 1953). Therefore,eliminating these processes can have unintended consequences on a channel’s aquatic habitat(e.g Schmetterling et al., 2001; Reid and Church, 2015).This thesis proposes an alternative approach to bank protection. By engaging with theprocesses that govern channel stability, it is possible to reduce bank migration, while stillallowing elements of natural channel function. This proposed approach is termed 'StabilitySeeding' and consists of placing an abundance of grains that are approximately the samesize as the coarsest grains (Stability Seeds) on a channel’s banks. As the channel widens,the Stability Seeds are incorporated into the channel bed. Once on the bed, Stability Seedspromote channel stability.To test the effectiveness of this method, proof-of concept-experiments were conductedusing scaled physical models. Stability Seeding was placed along the banks of a rectangu-lar channel and subject floods of various magnitudes. To compare the effects of StabilitySeeding to a traditional bank protection technique, an experiment using Riprap was run.An experiment without any bank protection was also conducted. Chapters 4-6 present theresults of these experiments. Chapter 7 discusses the effects of the bank protections on thechannel as well as the implications of using these techniques in river engineering projects.We found that using Stability Seeding as bank protection reduces bank migration, while59still allowing natural channel adjustment. Unlike a reach protected by Riprap, a reachprotected by Stability Seeding has the ability to form a complex channel with a varied channelmorphology. This variation can increase the number of aquatic habitats available and allowfor a higher level of ecological function (Beechie et al., 2010). In addition, Stability Seedingallows the channel to compensate for high flow rates with a small amount of lateral movement.As a result, there is less vertical degradation associated with Stability Seeding, than iftraditional bank protection was used. Therefore, Stability could have the ability to protectburied infrastructure better than traditional bank protection. In addition to traditional bankprotection projects, Stability Seeding could be used to bolster riparian restoration projects,where limiting bank migration is essential to establishing riparian vegetation.While there are many benefits with using Stability Seeding as bank protection, Stabil-ity Seeding does result in less developable land than traditional bank protection. Due tothe small levels of bank migration associated with Stability Seeding, using Stability Seed-ing as bank protection requires a relatively large minimum setback distance between thechannel and critical infrastructure. This buffer-zone allows the channel to migrate withoutconsequences to infrastructure. While the required buffer-zone may be used for non-criticaldevelopment (such as recreational areas), it does reduce the area where critical infrastructurecan be built.While the results of this thesis provide evidence that Stability Seeding may be used as analternative to traditional bank protection, there are certain limitations to the study. First,all three experiments were conducted using models of a gravel bed channel. As differentchannel types have different mechanisms for channel erosion and deposition (Church, 2006),the results from these experiments may not be applicable to other channel types. Furtherexperimental testing is required to determine how Stability Seeding affects different channeltypes. Second, in physical modelling, much of the complexity found in natural river systemsis ignored. For example, the role of vegetation was omitted from the physical models. Vege-tation located along the banks and in the channel has been shown to have significant effectson channel morphology (Eaton and Giles , 2009). Finally, in order to maintain a hydraulicallyrough bed surface, sediment finer than 0.25 mm was not included in the initial bed mixture.Thus, the sand component of the protype’s GSD was omitted from the experiments.To test how effective Stability Seeding is in more complex channels, a pilot project shouldbe launched in an actual river. This pilot project would provide further evidence thatStability Seeding can be used as an alternative to traditional bank protection techniques.Overall the results of this thesis provide strong evidence that it is possible to work with theprocesses that govern channel stability to reduce bank migration and channel widening. Byincreasing the abundance of the coarsest grains found on a channel’s bed, Stability Seeding60has the potential to provide bank protection while still retaining natural channel function.61Bibliography(2000), Riprap Design and Construction Guide, Tech. Rep. March, Ministry of Environment,Province of British Columbia, doi:DOI 10.1016/j.neuron.2012.04.014.Aberle, J., and V. Nikora (2006), Statistical properties of armored gravel bed surfaces, WaterResources Research, 42 (11), 1–11, doi:10.1029/2005WR004674.Andrews, C. (2010), A stream in transition: short term morphodynamicd of Fishtrap Creekfollowing wildfire, Ph.D. thesis, University of British Columbia, doi:10.14288/1.0069407.Ashmore, P. E. (1991), How do gravel-bed rivers braid?, Canadian Journal of Earth Sciences,28 (3), 326–341, doi:10.1139/e91-030.Baird, D. C., L. Fotherby, C. C. Klumpp, and S. M. Scurlock (2015), Bank StabilizationDesign Guidelines, Tech. Rep. June, U.S. Department of the Interior, Albuquerque.Bancroft, B., and K. Zielke (2002), Guidelines for riparian restoration in British Columbia:Recommended riparian zone siliviculture treatments, October, 1 ed., Ministry of Forests,Victoria.Beechie, T. J., H. Moir, and G. R. Pess (2008), Hierarchical Physical Controls on Spawningand Incubation, in Salmoid Spawning Habitat in Rivers, chap. 5, pp. 83–102.Beechie, T. J., D. A. Sear, J. D. Olden, G. R. Pess, J. M. Buffington, H. Moir, P. Roni,and M. M. Pollock (2010), Process-based Principles for Restoring River Ecosystems, Bio-Science, 60 (3), 209–222, doi:10.1525/bio.2010.60.3.7.Bennett, S. J., A. Simon, J. M. Castro, J. F. Atkinson, C. E. Bronner, S. S. Blersch, and A. J.Rabideau (2013), The Evolving Science of Stream Restoration, in Geophysical MonographSeries, vol. 194, pp. 1–8, doi:10.1029/2011GM001099.Bernhardt, E. S., and M. A. Palmer (2007), Restoring streams in an urbanizing world,Freshwater Biology, 52 (4), 738–751, doi:10.1111/j.1365-2427.2006.01718.x.62Biggs, M. (), City of Calgary (2019), Flooding in Calgary.Buffington, J. M., and D. R. Montgomery (1999), A Procedure for classifying tex-tural facies in gravel-bed rivers, Water Resources Research, 35 (6), 1903–1914, doi:10.1029/1999WR900041.Bunte, K., and S. Abt (2001), Sampling surface and subsurface particle-size distributionsin wadable gravel- and cobble-bed streams for analyses in sediment transport, hydraulics,and streambed monitoring, Tech. rep., U.S. Department of Agriculture.Church, M. (2006), Bed Material Transport and the Morphology of Alluvial RiverChannels, Annual Review of Earth and Planetary Sciences, 34 (1), 325–354, doi:10.1146/annurev.earth.33.092203.122721.Church, M. (2010), Gravel-Bed Rivers, in Sediment Cascades, pp. 241–269, John Wiley &Sons, Ltd, Chichester, UK, doi:10.1002/9780470682876.ch9.Downs, P. W., and K. J. Gregory (2004), River Channel Management, 395 pp., OxfordUniversity Press, New York.Eaton, B., R. Moore, and T. Giles (2010), Forest fire, bank strength and channel instability:the ‘unusual’ response of Fishtrap Creek, British Columbia, Earth Surface Processes andLandforms, 35 (10), 1167–1183, doi:10.1002/esp.1946.Eaton, B., L. Mackenzie, M. Jakob, and H. Weatherly (2017), Assessing Erosion Hazardsdue to Floods on Fans: Physical Modeling and Application to Engineering Challenges, J.Hydraul. Eng, 143 (8), 1–14, doi:10.1061/(ASCE).Eaton, B. C., and T. R. Giles (2009), Assessing the effect of vegetation-related bank strengthon channel morphology and stability in gravel-bed streams using numerical models, EarthSurface Processes and Landforms, 34 (5), 712–724, doi:10.1002/esp.1768.Fang, Y., and J. W. Jawitz (2019), The evolution of human population distance to water inthe USA from 1790 to 2010, Nature communications, 10 (1), 430, doi:10.1038/s41467-019-08366-z.Ferguson, R. (2007), Flow resistance equations for gravel- and boulder-bed streams, WaterResources Research, 43 (January), 1–12, doi:10.1029/2006WR005422.Frothingham, K. M. (2008), Evaluation of stability threshold analysis as a cursory method ofscreening potential streambank stabilization techniques, Applied Geography, 28 (2), 124–133, doi:10.1016/j.apgeog.2007.07.006.63Hafs, A. W., L. R. Harrison, R. M. Utz, and T. Dunne (2014), Quantifying the role ofwoody debris in providing bioenergetically favorable habitat for juvenile salmon, EcologicalModelling, 285, 30–38, doi:10.1016/j.ecolmodel.2014.04.015.Heede, B. H. (1986), Desigining for Dynamic Equilibrium in Streams, Journalof the American Water Resources Association, 22 (3), 351–357, doi:10.1111/j.1752-1688.1986.tb01889.x.Johnson, P. (2006), Assessing stream channel stability at bridges in physiographic regions,Tech. Rep. July, Turner-Fairbank Highway Research Center, Mclean, Virginia.Kondolf, G. M. (1994), Hungry water: Effects of dams and gravel mining on river channels,Environmental Management, 28 (4), 533–551.Kondolf, G. M. (2000), Assessing Salmonid Spawning Gravel Quality, Trans-actions of the American Fisheries Society, 129 (1), 262–281, doi:10.1577/1548-8659(2000)129¡0262:ASSGQ¿2.0.CO;2.Kondolf, G. M., and G. Wolman (1993), The Sizes Of Salmonid Spawning Gravels, WaterResources Research, 29 (7), 2275–2285.Kondolf, G. M., J. Williams, T. Horner, and D. J. Milan (2008), Assessing Physical Qualityof Spawning Habitat, in Salmoid Spawning Habitat in Rivers, chap. 12, pp. 249–276.Leopold, L., and T. Maddock (1953), The Hydraulic Geometry of Stream Channels andSome Physiographic Implications, Tech. rep., doi:10.3133/pp252.Ligon, F. K., W. E. Dietrich, and W. J. Trush (1994), Downstream ecological effects of dams:A geomorphic perspective, Institute of Biological Sciences, 45 (3), 183–192.Lisle, T. E., and M. a. Madej (1992), Spatial Variation in Armouring in a Channel with HighSediment Supply, Dynamics of Gravel-bed Rivers, pp. 277–293.MacKenzie, L. G., and B. C. Eaton (2017), Large grains matter: contrasting bed stabilityand morphodynamics during two nearly identical experiments, Earth Surface Processesand Landforms, 42 (8), 1287–1295, doi:10.1002/esp.4122.Palmer, M. A., H. L. Menninger, and E. Bernhardt (2010), River restoration, habitat hetero-geneity and biodiversity: A failure of theory or practice?, Freshwater Biology, 55 (SUPPL.1), 205–222, doi:10.1111/j.1365-2427.2009.02372.x.64Palmer, M. A., S. Filoso, and R. M. Fanelli (2014), From ecosystems to ecosystem ser-vices: Stream restoration as ecological engineering, Ecological Engineering, 65, 62–70,doi:10.1016/j.ecoleng.2013.07.059.Palucis, M. C., and M. P. Lamb (2017), What controls channel form in steep mountainstreams?, Geophysical Research Letters, pp. 7245–7255, doi:10.1002/2017GL074198.Parker, G., P. C. Klingeman, and D. G. McLean (1983), Closure to “ Bedload and SizeDistribution in Paved Gravel-Bed Streams ” by Gary Parker, Peter C. Klingeman, andDavid G. McLean (April, 1982), Journal of Hydraulic Engineering, 109 (5), 793–793, doi:10.1061/(ASCE)0733-9429(1983)109:5(793).Peakall, J., P. Ashworth, and J. Best (1996), Physical Modelling in Fluvial Geomorphology :Principles , Applications and Unresolved Issues, in The Scientific Nature of Geomorphol-ogy: Proceedings of the 27th Binghamton Symposium in Geomorphology, September 1996,pp. 221–253.Peakall, J., P. J. Ashworth, and J. L. Best (2007), Meander-Bend Evolution, Alluvial Archi-tecture, and the Role of Cohesion in Sinuous River Channels: A Flume Study, Journal ofSedimentary Research, 77 (3), 197–212, doi:10.2110/jsr.2007.017.Piegay, H., L. A. Naylor, G. Haidvogl, J. Kail, L. Scmitt, and L. Bourdin (2008), IntegrativeRiver Science and Rehabilitation: European Experiences, in River Futures: An integrativescientific approach to river repair, edited by G. J. Brierly and K. A. Fryirs, chap. 11, pp.201–219, Island Press, Washington D.C.Pomeroy, J. W., X. Fang, and D. G. Marks (2016), The cold rain-on-snow event of June 2013in the Canadian Rockies - characteristics and diagnosis, Hydrological Processes, 30 (17),2899–2914, doi:10.1002/hyp.10905.Reid, D., and M. Church (2015), Geomorphic and Ecological Consequences of Riprap Place-ment in River Systems, Journal of the American Water Resources Association, 51 (4),1043–1059, doi:10.1111/jawr.12279.Schmetterling, D. A., C. G. Clancy, and T. M. Brandt (2001), Effects of Riprap BankReinforcement on Stream Salmonids in the Western United States, Fisheries, 26 (7), 6–13,doi:10.1577/1548-8446(2001)026¡0006:EORBRO¿2.0.CO;2.Smith, B., N. J. Clifford, and J. Mant (2014), The changing nature of river restoration,Wiley Interdisciplinary Reviews: Water, 1 (3), 249–261, doi:10.1002/wat2.1021.65Snodgrass, W., and D. Sorel (2015), Stream restoration design and management in the cityof Toronto, in Latornel Conference.Tamminga, A., and B. Eaton (2018), Linking geomorphic change due to floods to spatialhydraulic habitat dynamics, (July), 1–17, doi:10.1002/eco.2018.Wayand, N. E., J. D. Lundquist, and M. P. Clark (2015), Modeling the influence of hypsome-try, vegetation, and storm energy on snowmelt contributions to basins during rain-on-snowfloods, Water Resources Research, 51 (10), 8551–8569, doi:10.1002/2014WR016576.Wilcock, P. R., and B. W. McArdell (1993), Surface-based fractional transprot rates: partialtransport and mobilization thresholds, Water Resources Research, 29 (4), 1297–1312.Wohl, E., S. N. Lane, and A. C. Wilcox (2015), The science and practice of river restoration,Water Resources Research, 51 (8), 5974–5997, doi:10.1002/2014WR016874.Yalin, M. (1971), Theory of Hydraulic Models, Macmillan.66Appendix ASurface GSDs670.5 1.0 2.0 5.00.00.20.40.60.81.0Grain Size (mmPercent Finerbankfull flow500−year floodFigure A.1: The Grain Size Distribution of the bed surface of the Natural Channel Experi-ment for the bankfull flow and 500-year flood.681 2 5 100.00.20.40.60.81.0Grain Size (mm)Percent Finerbankfull flow500−year floodFigure A.2: The Grain Size Distribution of the bed surface of the Riprap Experiment forthe bankfull flow and 500-year flood.69Appendix BComparison of the constantManning’s n and variable Manning’sn Flow Models700.000 0.002 0.004 0.006 0.008 0.010050100150200250300Specific Discharge (m2 s)DensityConstant Manning's nVariable Manning's na)0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.0140.0000.0050.0100.015Constant Manning's nVariable Manning's n b)Figure B.1: Comparison of the specific discharges of the constant Manning’s roughnesscoefficent and the variable Manning’s roughness coefficent flow models for the bankfull flowof the Natural Channel Experiment. (a) is a density plot of specific discharges, while (b)compares the relation between the specific discharges of the two flow model methods. Theblack line indicates a 1:1 relation.710.000 0.002 0.004 0.006 0.008 0.010050100150200250300Specific Discharge (m2 s)DensityConstant Manning's nVariable Manning's na)0.000 0.005 0.010 0.0150.0000.0050.0100.015Constant Manning's nVariable Manning's nb)Figure B.2: Comparison of the specific discharges of the constant Manning’s roughnesscoefficent and the variable Manning’s roughness coefficent flow models for the bankfull flowof the Riprap Experiment. (a) is a density plot of specific discharges, while (b) comparesthe relation between the specific discharges of the two flow model methods. The black lineindicates a 1:1 relation.72Appendix CSpecific discharges of mean annualflow73a)b)c)04e−048e−040.00120.0016d)Figure C.1: Specific discharge maps of the Natural Channel Experiment for the (a) bankfullmean annual flow, (b) 10-year mean annual flow, (c) 50-year mean annual flow and (d)500-year mean annual flow.a)b)c)04e−048e−040.00120.0016d)Figure C.2: Specific discharge maps of the Riprap Experiment for the (a) bankfull meanannual flow, (b) 10-year mean annual flow, (c) 50-year mean annual flow and (d) 500-yearmean annual flow.74a)b)c)04e−048e−040.00120.0016d)Figure C.3: Specific discharge maps of the Stability Seeding Experiment for the (a) bankfullmean annual flow, (b) 10-year mean annual flow, (c) 50-year mean annual flow and (d)500-year mean annual flow.75

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0380598/manifest

Comment

Related Items