THE GEOMORPHIC EFFECTS OF NATIVE AND INVASIVE RIPARIAN VEGETATION: SPRAGUE RIVER, OREGON by ADRIANA E. MARTINEZ A DISSERTATION Presented to the Department of Geography and the Graduate School of the University of Oregon in partial fulfillment of the requirements for the degree of Doctor of Philosophy June 2013 ii DISSERTATION APPROVAL PAGE Student: Adriana E. Martinez Title: The Geomorphic Effects of Native and Invasive Riparian Vegetation: Sprague River, Oregon This dissertation has been accepted and approved in partial fulfillment of the requirements for the Doctor of Philosophy degree in the Department of Geography by: Patricia F. McDowell Chair Andrew W. Marcus Member Daniel G. Gavin Member Bitty A. Roy Outside Member and Kimberly Andrews Espy Vice President for Research and Innovation Dean of the Graduate School Original approval signatures are on file with the University of Oregon Graduate School. Degree awarded June 2013 iii © 2013 Adriana E. Martinez iv DISSERTATION ABSTRACT Adriana E. Martinez Doctor of Philosophy Department of Geography June 2013 Title: The Geomorphic Effects of Native and Invasive Riparian Vegetation: Sprague River, Oregon Numerous studies have addressed the role of stream geomorphology on vegetation distribution. These studies have shown that channel morphology, including depositional and erosional processes, influence vegetation colonization. However, few studies have addressed the impact of vegetation on the geomorphic processes of streams. Vegetation has the ability to stabilize channel banks and alter stream hydrology and stream power. Little research has addressed the impact of invasive vegetation and its ability to change river channel processes. My research addresses the impact of the highly invasive Phalaris arundinacea and quantifies its influence on the stream channel form of the Sprague River, Oregon. I conducted field research that included root density and root strength surveys to determine the below ground influences of vegetation in terms of added bank cohesion provided by the invasive and two similar native species: Eleocharis palustris and Carex vesicaria. To ascertain differences between the species above ground characteristics and influences, I measured stem density and elasticity to calculate their roughness (Manning’s n) and determine their potential impact on stream velocity. Finally, I used these vegetation characteristics to model stream velocity, water depth, and bed shear stress within the 2-D model MD-SWMS. Differences in root size were significant with C. vesicaria having the v largest root diameters, largest root area ratio, and largest bank cohesion provided by roots. This was followed by the invasive and then E. palustris. E. palustris had the highest stem density, followed by C. vesicaria and P. arundinacea. The invasive had the highest stem stiffness. E. palustris was associated with the highest roughness value, closely followed by the invasive and C. vesicaria. Using modeling I found the presence of the invasive increased velocity compared to E. palustris and increased bed shear stress compared to C. vesicaria. Therefore, changes in species composition, such as a shift from either of the natives to the invasive, could affect channel morphology over time. By comparing the impact of this invasive to that of native grasses and sedges, this research provides insight into how further spread of the invasive may affect the Sprague River and other riparian ecotones. vi CURRICULUM VITAE NAME OF AUTHOR: Adriana E. Martinez GRADUATE AND UNDERGRADUATE SCHOOLS ATTENDED: University of Oregon, Eugene Texas A&M University, College Station DEGREES AWARDED: Doctor of Philosophy, Geography, 2013, University of Oregon Master of Science, Geography, 2008, Texas A&M University Bachelor of Science, Environmental Geosciences, 2005, Texas A&M University AREAS OF SPECIAL INTEREST: Fluvial Geomorphology Human Impacts on Rivers Biogeomorphology PROFESSIONAL EXPERIENCE: Course Instructor, University of Oregon, 2010-2012 Introduction to Physical Geography, Geomorphology Graduate Teaching Fellow, University of Oregon, 2009-2013 Climatology, Geomorphology, Maps and Geospatial Concepts, Introduction to GIS, Watershed Science and Policy, Global Environmental Change National Science Foundation Fellow, GK-12 Fellowship, 2006-2008 Teaching Assistant, Texas A&M University Department of Geography, 2005 Hydrology GRANTS, AWARDS, AND HONORS: Christopherson Geosystems Award for Best Applied Geography Paper, 2012 APCG Latin American Travel Scholarship, 2012 vii APCG Women’s Network Travel Grant 2012 APCG Student Travel Grant Cliff Matthew memorial Research Scholarship for Women in the Geosciences, 2011 Harry and Shirley Bailey Award for Outstanding Paper in Physical Geography, 2011 NCED Summer Institute on Rivers and Vegetation, 2010 Promising Scholar Award, University of Oregon, 2008 Graduate Diversity Fellowship, Texas A&M University, 2005 Hispanic Scholarship Fund Institute Environmental Management Scholarship, 2004 PUBLICATIONS: Martinez, A. E. and S. W. Hardwick (2010). "Building Fences: Undocumented Immigration and Identity in a Small Border Town." Focus on Geography 52(3-4): 48-55. Martinez, A. E., N. A. Williams, Metoyer, S., Berhane, S., Morris, J. (2009). "A Geospatial Scavenger Hunt." Science Scope 32(6): 48-53. Chin, A., L. Laurencio, Martinez, A.E. (2008). "The Hydrologic Impacts of Small- and Medium-sized Dams." The Professional Geographer 60(2): 238-251. viii ACKNOWLEDGMENTS I would like to thank my adviser Pat McDowell for her unwavering support throughout this process. You have been a great source of inspiration and great fun in the field. Special thanks also to my committee: Andrew Marcus who always inspires in me a love for rivers and geomorphology, Dan Gavin who made me realize there is a bit of a biogeographer inside me, and Bitty Roy who introduced me to all the little green friends I now have great fun identifying. Also, many thanks to Mark Fonstad who, even though he was not on my committee, gave me endless amounts of advice and support through this process in the last few years. Heartfelt thanks to my family for their support through this crazy experience called grad school. I could not have possibly accomplished my dreams without the support of my mom, dad, and sister. Thanks for pushing me forward and always believing I could do it. Ale, thanks for all of your practical advice and for being there for me when I was most nervous. In addition to family, I could not have possibly done this without the support of the many friends and colleagues I have amassed during this journey. Thanks to Kevin Merritt, who knew exactly what I was going through every step of the way. You made me laugh and encouraged me through the whole processes in a way only a best friend can. Without Susan Hardwick I would have never made it through my first year in Oregon. I thank you for your friendship and all our lunches. Thanks to Jenn and Ross Kusler, and Sonja and Rob Anthone, who became my family here in Eugene. Special thanks to all the River Rats past and present who helped me in the field and in the office: Jessica Phelps, Stephani Michelsen-Correa, Suzanne Walther, Pollyanna Lind, Chris ix Rasmussen, Paul Blanton, Jane Atha, James Dietrich, Amanda Reinholtz, Denise Tu, Swagata Goswami, Helen Beeson, and Sarah Praskievicz. Chris Rasmussen deserves special honor for putting up with my constant questions on the Sprague and always being a great sounding board for my ideas. You are a great mentor and colleague. Also, to Nicole Czarnomski for helping me brainstorm at the beginning of this project; my meetings with you were invaluable. Thanks to Lacie Holt, Patrick Miller, Mike Dawson, Jenn Kusler, Amanda Reinholtz, and Michael Hughes for your support in the field; I couldn’t have done it without you. Thanks to all of the GeoGrads who supported me, held me up when I was down, and celebrated my achievements with me. Special thanks to Leslie McLeese, Easther Chigumira (failure breeds success!), Lindsay Naylor, Innisfree McKinnon, Kenji Izumi, Erin Herring, Christine Carolan, Aquila Flower, Matt Landers, Alex Ginsburg, Meche Lu, Ingrid Nelson, and Laurie Trautman. The community we have created is like no other and it wouldn’t have been possible without each one of you. To Innisfree, thanks for all of your endlessly helpful support and encouragement. To Lindsay, I made it and you will too(!); thanks for growing through this graduate experience with me. Many thanks to landowners Greg Harris and Teeter Hess for access to their properties and Kris Fischer and Tony LaGreca at the Klamath Tribes research station for sediment data and equipment use. And last, but certainly not least, thanks to the Shields for hosting me in the field on their little piece of heaven. You made work on the Sprague an amazing experience. Your generosity and kindness floors me to this day. x To: My Mom, you are the wind beneath my wings, My Dad, I hope I’m just like you when I grow up, Ale, I don’t make sense without you, and Kevin, my Samwise Gamgee. You are a hero to me. xi TABLE OF CONTENTS Chapter Page I. INTRODUCTION .................................................................................................... 1 II. ROOT SIZE DISTRIBUTION AND ROOT TENSILE STRENGTH OF INVASIVE AND NATIVE RIPARIAN VEGETATION ........................................... 5 1. Introduction .................................................................................................. 5 2. Background .................................................................................................. 7 3. Study Site ..................................................................................................... 9 4. Plant Species ................................................................................................ 11 4.1. Reed Canarygrass (Phalaris arundinacea L.) .................................... 11 4.2. Creeping Spikerush (Eleocharis palustris L.) .................................... 12 4.3. Inflated Sedge (Carex vesicaria L.) .................................................... 13 5. Methods........................................................................................................ 14 5.1. Root Distribution and Root Area Ratio .............................................. 14 5.2. Root Strength ...................................................................................... 15 5.3. Cohesion Due to Roots ....................................................................... 16 6. Results .......................................................................................................... 16 6.1. Root Distribution and Root Area Ratio .............................................. 16 6.2. Root Strength and Cohesion ............................................................... 19 7. Discussion .................................................................................................... 20 7.1. Cohesion Due to Roots ....................................................................... 23 7.2. Restoration Implications ..................................................................... 26 xii Chapter Page 8. Conclusions .................................................................................................. 27 III. QUANTIFYING ROUGHNESS: FLEXIBLE RIPARIAN VEGETATION AND MANNING’S n ................................................................................................. 29 1. Introduction .................................................................................................. 29 2. Methods........................................................................................................ 31 2.1. Site Selection ...................................................................................... 31 2.2. Plant Characteristics and Other Variables .......................................... 32 2.3. Calculating Roughness........................................................................ 33 2.4. Measuring Deposition ......................................................................... 34 3. Results .......................................................................................................... 36 3.1. Plant Characteristics............................................................................ 36 3.2. Roughness ........................................................................................... 36 3.3. Deposition ........................................................................................... 40 4. Discussion .................................................................................................... 42 4.1. Plant Characteristics............................................................................ 42 4.2. Roughness ........................................................................................... 44 4.3. Deposition ........................................................................................... 46 5. Conclusion ................................................................................................... 48 xiii Chapter Page IV. INVASIVE AND NATIVE RIPARIAN VEGETATION: MODELING CHANNEL HYDRAULICS WITH CURRENT AND FUTURE VEGETATON CONDITIONS ............................................................................................................. 50 1. Introduction ........................................................................................................ 50 2. Background ........................................................................................................ 51 3. Methods.............................................................................................................. 53 3.1. MD-SWMS (Multi-Dimensional Surface-Water Modeling System) ....... 53 3.2. Modeling with Roughness ........................................................................ 57 3.3. Modeling with Vegetation Density and Height ........................................ 58 4. Results .............................................................................................................. 58 4.1. Current Conditions .................................................................................... 59 4.1.1. Modeling Based on Vegetation Density and Height .................. 59 4.1.2. Modeling Based on Vegetation Roughness .............................. 59 4.2. Future Conditions...................................................................................... 62 4.2.1. Modeling Based on Vegetation Density and Height .................. 62 4.2.2. Modeling Based on Vegetation Roughness ............................... 63 5. Discussion .......................................................................................................... 67 5.1. Model Performance ................................................................................... 67 5.2. Interpretation of Results ............................................................................ 68 5.2.1. Modeling Based on Vegetation Height and Density .................. 68 5.2.2. Modeling Based on Vegetation Roughness ............................... 69 xiv Chapter Page 6. Conclusion ................................................................................................... 71 V. SUMMARY ............................................................................................................ 72 REFERENCES CITED ................................................................................................ 76 xv LIST OF FIGURES Figure Page 2.1. Sprague River Basin and three study sites used in this study. .............................. 11 2.2. Root and stem systems of partial plants (a) P. arundinacea, (b) E. palustris, and (c) C. vesicaria. Squares on the scale are centimeters. .................................. 13 2.3. Root diameter histogram for three species of interest .......................................... 18 2.4. RAR distribution by depth for P. arundinacea, E. palustris, and C. vesicaria. Standard errors of RAR are presented by the extending lines. .............................. 18 2.5. Root diameter and root tensile strength power relationship for P. arundinacea, E. palustris, and C. vesicaria ................................................................................. 21 3.1. Deposition pin (yellow circles) and mat (green=turf mats, black=rubber mats) setup along the banks of the Sprague River (light blue polygon). ......................... 35 3.2. Stage heights for period of time deposition pins and turf mats were in place. Stage data obtained from Oregon Water Resources Department (OWRD) gage station 11497500 near Beatty, Oregon. .................................................................. 35 3.3. Box plots of stem density ...................................................................................... 37 3.4. Stem stiffness for three species of interest. ......................................................... 38 3.5. Vegetation roughness values for three species. .................................................... 40 3.6. Deposition (+) and erosion (-) depths for plant stands using pins.. ...................... 41 3.7. Deposition using turf mats for four plant stands. .................................................. 41 4.1. Sprague River Basin and associated study site. .................................................... 55 4.2. Current conditions modeled given vegetation height and density under 10-year (a1 and a2), and 25-year recurrence interval flows (b1 and b2) ............................ 60 4.3. Current conditions modeled given vegetation roughness under 10-year recurrence interval (a1, b1, c1) and 25-year recurrence interval (a2, b2, c2) .......................... 61 4.4. Future conditions modeled given vegetation height and density under 10-year recurrence interval flow: P. arundinacea – E. palustris (a1 and b1), P. arundinacea – C. vesicaria (a2 and b2). ................................................................ 63 xvi Figure Page 4.5. Future conditions modeled given vegetation height and density under 25-year recurrence interval flow: P. arundinacea – E. palustris (a1, and b1), P. arundinacea – C. vesicaria (a2 and b2) ............................................................ 64 4.6. Future conditions modeled given roughness under 10-year recurrence interval flow: P. arundinacea – E. palustris (a1, b1, c1), P. arundinacea – C. vesicaria (a2, b2, c2) ............................................................................................................. 65 4.7. Future conditions modeled given roughness given 25-year recurrence interval flow: P. arundinacea – E. palustris (a1, b1, c1), P. arundinacea – C. vesicaria (a2, b2, c2). ............................................................................................................ 66 4.8. P. arundinacea, C. vesicaria and E. palustris occupying typical channel locations (left bank), and conditions if the invasive outcompetes both of the natives (right bank) ................................................................................................ 71 xvii LIST OF TABLES Table Page 2.1. Roots per pit and root diameter size for three species. ......................................... 17 2.2. Root diameter and root tensile strength distributions for three species of interest .................................................................................................................... 20 2.3. Diameter and tensile strength power relationship variables for three species of interest ............................................................................................................... 20 2.4. RipRoot and Wu et al. value comparisons of added cohesion due to roots. ......... 20 2.5. Root diameter and strength of other species ......................................................... 24 2.6. Vegetation type and cohesion of other species ..................................................... 25 3.1. Variables for equations and associated descriptions ............................................. 33 3.2. Frontal and cross-sectional area characteristics for three species ......................... 39 3.3. Stem stiffness values for woody species published in previous work. ................. 42 3.4. Additive roughness values for the Sprague River................................................. 45 3.5. Previously established roughness values for other vegetation .............................. 46 3.6. Rivers similar to the Sprague River and their associated roughness values via photo reference guides. .......................................................................................... 47 4.1. Model grid parameters. ......................................................................................... 57 4.2. Discharge and associated boundary conditions. ................................................... 57 4.3. Roughness, vegetation height and density for each species ................................. 57 1 CHAPTER I INTRODUCTION Invasive species can cause major shifts in ecosystem dynamics, outcompete native species, and possibly produce other changes not recognized at present. With stream restoration efforts increasing across the United States, it is imperative that we understand the impact of invasive species. If an invasive species cannot be eliminated, for example, it may be impossible to restore a stream to its historical condition or even to reestablish some ecosystem functions. This doctoral research documents the impact of the invasive species Phalaris arundinacea (reed canarygrass) on the geomorphology of the low-energy, meandering Sprague River in eastern Oregon. Until recently, little research has focused on analyzing the influences of small, flexible vegetation on the geomorphic processes of fluvial systems. This project identifies and evaluates the vegetative and geomorphic processes taking place on the Sprague River. I use biogeomorphic techniques to answer three research questions (1) How does the stream bank stability provided by P. arundinacea compare to that provided by native vegetation Eleocharis palustris (creeping spikerush) and Carex vesicaria (inflated sedge)? (2) How does P. arundinacea impact stream velocity and deposition and does this differ from native vegetation? and (3) What is the potential impact of P. arundinacea on channel morphology? With restoration along the Sprague River underway and further projects planned, results from this study can affect how the invasive is addressed in these projects. In addition, with the pending removal of four Klamath dams downstream of the Sprague 2 River, historical salmon spawning habitat may return and the geomorphology and habitat characteristics of the stream are of extreme importance. In Chapter II I examine the below-ground influences of the invasive and native vegetation. Understanding of the role invasive species play in geomorphic processes is important because riparian vegetation provides added bank stability and cohesion through its root matrix. Here I examine the geomorphic effect of P. arundinacea along the Sprague River, Oregon, and compare its effect with two species it is outcompeting in the area: E palustris and C. vesicaria. Root distribution values and root tensile strengths were used to determine the role P. arundinacea plays in bank cohesion and stability via two methods: the traditional Wu et al. (1979) method and the fiber bundle model RipRoot (Pollen and Simon, 2005). Root size distribution analyses show the invasive plant has a large amount of small diameter roots. Tensile strength values for all three species decrease with increasing root diameter, which is consistent with previous studies. In addition, the Wu et al. (1979) method overestimates cohesion when compared to RipRoot for the three species. Modeled cohesion is greatest for C. vesicaria and therefore this species provides higher levels of bank stability and resistance to erosion compared to the invasive. In addition, the invasive provides greater bank cohesion when compared to E. palustris. In Chapter III I examine the above-ground influences of riparian vegetation along the Sprague River. Quantifying roughness in terms of riparian vegetation has largely been ignored or listed as a secondary characteristic on roughness reference tables. In methods used to calculate roughness mathematically, the added component of vegetation roughness is lacking. However, in instances where vegetation plays a dominant role in 3 the riparian landscape and forms an important mechanism to reduce water velocity at high flows, its roughness is critical. I calculated vegetation roughness based on measurements of plant stem elasticity, plant frontal area, stem density and stem area of three dominant herbaceous plants along the Sprague River, Oregon: the invasive P. arundinacea, and natives E. palustris, and C. vesicaria. Results show slightly lower roughness values than those predicted for vegetation using reference tables. In addition, E. palustris and P. arundinacea exhibit higher and similar roughness values whereas C. vesicaria exhibits values lower than the other two species. These findings are of particular importance where the invasive P. arundinacea is outcompeting native C. vesicaria because with such colonization, roughness is increasing in channel zones and therefore is likely changing channel processes. Direct depositional measurements show similar results within stands of the invasive and E. palustris. In Chapter IV I used the roughness values, stem density and plant height values calculated in Chapter III, as well as topography gathered along the Sprague using LiDAR and bathymetric methods, to model the channel changes associated with the three species of interest. Modeling is accomplished via two methods. First, current conditions are modeled by populating the channel banks with roughness, plant density and height of patches of vegetation mapped out during a raft trip down a representative reach of the Sprague. Next, modeling is carried out along the same reach with future conditions where the banks are populated by only one of the species. Modeling results for the two native species are then compared to results for the invasive to determine differences in water depth, velocity and bed shear stress. Current conditions show that the patches mapped have minimal interaction with the 10- and 25-year recurrence interval flow and 4 further work should be conducted regarding the influence of the placement of vegetation. Idealized condition model runs show that plant density and roughness causes the native E. palustris to have the highest impact on stream velocity and bed shear stress. This is of particular importance where the invasive is outcompeting E. palustris. Such changes could cause increases in velocity and less stable bank surfaces. Research of this nature has previously focused on large, woody species such as Tamarix (Graf, 1978). This research is one of the first studies to investigate the influence of a non-woody, but highly invasive vegetation species. Here I demonstrate that the invasive is significantly different in terms of above-ground and below-ground influences when compared to the natives. These differences, according to depositional measurements and modeling, can affect channel processes. With this study, we are only beginning to understand the impact of invasive species on stream geomorphology. To attempt to circumvent the negative impacts of invasive species, we must learn how to control them and reverse the changes already set in motion. This study begins to bridge the gap and discover how invasives may be altering geomorphic processes. 5 CHAPTER II ROOT SIZE DISTRIBUTION AND ROOT TENSILE STRENGTH OF INVASIVE AND NATIVE RIPARIAN VEGETATION 1. Introduction The encroachment of invasive species has become an increasingly apparent problem throughout ecosystems in North America. Invasives are often highly competitive and quickly take advantage of available habitat and resources. For this reason, invasive plant species are very well adapted for rapid dispersal and establishment in locations prone to disturbance (Sher et al. 2000; Tickner et al. 2001). For example, flood disturbances on river systems that remove native vegetation can accelerate the proliferation of competitive invasive vegetation (Cowell and Dyer 2002; Tickner et al. 2001; Sher et al. 2002). Invasive and native vegetation has the ability to stabilize channel banks and alter stream hydrology, erosion, and deposition. Above ground biomass intercepts rainfall and interrupts streamflow, decreasing stream velocities and thus shear stress (Pollen et al. 2004; Hickin 1984; Corenblit et al. 2007; Bennett et al. 2007). Bank stability is enhanced by the roots of riparian vegetation that directly strengthens bank sediment and thus prevents instances of mass failure (Murray et al. 2008; Tal and Paola 2007). Stability of bank material is, in part, a function of root density distribution, root depth, root diameter and root tensile strength (e.g. Pollen and Simon 2005); other factors include soil moisture, sediment size, and soil cohesion. Roots enhance the tensile strength and elasticity of soils, distributing stresses throughout, allowing for increased shear strength, and creating failure resistant zones (Gran and Paola 2001; Murray et al. 2008; Perucca et 6 al. 2007; Tal et al. 2004). Often, these zones form islands where flow is diverted in other, less resistant directions and thus they alter stream channel dimensions, sinuosity, and other characteristics (Nanson and Knighton 1996). Tamarisk was found to decrease channel widths by 27% along the Green River, Utah (Graf 1978) by stabilizing bar and bench surfaces. Pollen and Simon (2005) applied a fiber bundle approach to measure the tensile strength of 12, primarily woody, riparian species. Simon et al. (2006) found that Lemmons willow (Salix lemmonii) root reinforcement capabilities were an order of magnitude greater than that of lodgepole pine (Pinus contorta) along the Upper Truckee River, California, and therefore decreased the frequency of bank failure and delivery of fine grained sediment. In addition to bank stability, vegetation provides protection against erosion by fluid forces both through its aboveground biomass and anchoring by roots (Murray et al. 2008; Smith 1976). Along the Alexandra River, Alberta, the roots of plants on vegetated banks were found to enhance erosion resistance up to 20,000 times more than unvegetated banks. This caused a decrease in lateral channel migration (Smith 1976; Nanson and Knighton 1996; Graf 1978). In the Northern Plains of Australia, various grasses were found to provide higher levels of stability and erosion protection than trees and shrubs due to their substantial biomass and continuous root networks (Tooth and Nanson 1999). In addition, De Baets et al. (2007; 2008) found that 6 out of the 7 grasses they tested in gully environments had a “High” to “Very High” erosion reducing potential whereas the majority of shrubs and trees tested in the study ranged from “Very Low” to “High” erosion reducing potential. These studies show that riparian vegetation influences 7 geomorphic processes, and these influences can be different depending on the species colonizing the channel banks. Understanding the physical effects of invasive species in the riparian zone is important because changes in plant species composition as invasive species colonize riparian areas may result in changes in bank stability and resistance to erosion. Changes in these processes may lead to alterations in channel form and habitat characteristics. Therefore, objectives for this paper are to (1) determine the root characteristics of the invasive (Phalaris arundinacea) and two similar native species (Eleocharis palustris and Carex vesicaria) occupying similar areas along the Sprague River, Oregon and (2) determine the added bank cohesion the invasive and native species provide along the Sprague River using the Wu et al. (1979) method and the Pollen and Simon (2005) fiber bundle model RipRoot. 2. Background Root structure and architecture have long been recognized to influence bank and slope stability (Greenway 1987; Bethalahmy 1962; Bishop and Stevens 1964; Kaul 1965). Roots are strong in tension and weak in compression while soil is weak in tension and strong in compression. These combined factors allow roots to provide reinforcement within the soil matrix (De Baets et al. 2007; De Baets et al. 2008; Pollen et al. 2004; Simon and Collison 2002; Simon et al. 2006; Greenway 1987). Root strength is a function of root diameter with finer roots exhibiting higher tensile strengths per unit root area than those with larger diameters (De Baets et al. 2008; Gray and Barker 2004; Operstein and Frydman 2000). In addition to root strength, root distribution significantly 8 affects bank and slope stability. Abernethy and Rutherford (2001) found interspecies differences in root distribution to affect bank stability more than interspecies differences in root tensile strength. Several studies have suggested the use of Root Area Ratio (the ratio of root cross sectional area to soil cross sectional area; RAR) to estimate root contribution to soil strength (Abernathy and Rutherford 2001). Estimations of bank reinforcement consider root cohesion as a factor of soil shear strength and thus incorporate it into the Coulomb equation (Waldron 1977; Wu et al. 1979): (1) where S is soil shearing resistance (kPa), c is cohesion (kPa), is normal stress on the shear plane (kPa), and is soil friction angle (Pollen-Bankhead and Simon 2009). Waldron (1977) and others modified the equation to account for increased shear strength due to roots: (2) where is increased shear strength due to roots (kPa). can be further estimated by accounting for the tangential component resisting shear and the normal component increasing the confining pressure on the shear plane: ( )( ) (3) where Tr is average tensile strength of roots per unit area of soil (kPa), is the angle of shear distortion in the shear zone, and Ar/A is root area ratio. Work by Gray (1974) and Wu et al. (1979) further simplified this relationship by reporting that the effect of roots on the angle of internal friction is minor and the resulting shear strength due to roots is not sensitive to normal changes in and . Therefore, the equation simplifies to: 9 ( ) (4) where the added cohesion due to roots is dependent on the size, distribution and strength of roots. This type of model is considered “static” (Pollen and Simon 2005) because it captures root reinforcement for a single moment in time and assumes all roots reach their maximum tensile strength simultaneously (Greenway 1987). Subsequent studies have found the Wu et al. (1979) model overestimates soil shear strength by up to 91% (Pollen et al. 2004) and the values are often used as maximums (De Baets et al. 2008; Simon and Pollen 2006). More recently, Pollen et al. (2004) developed a fiber bundle model, RipRoot, that adopts the global load sharing (GLS) approach to account for the progressive breakage of roots within the soil matrix (Pollen and Simon 2005; Pollen et al. 2004; Simon and Pollen 2006). As one root breaks, the stress is redistributed to the remaining roots within the soil matrix. In addition, RipRoot accounts for two overlooked mechanisms within soil shearing and root strength, that of root breaking and root pullout. Therefore, recent versions of the RipRoot model now account for root pullout forces, a function of the bond between the roots and soil, and the soil moisture (Pollen-Bankhead and Simon 2009). To demonstrate this empirically, I chose the Sprague River, Oregon, which exhibits large stands of the invasive P. arundinacea and other similar native species. 3. Study Site The Sprague River, a tributary of the Upper Klamath River, (Figure 2.1.) lies at 1270 to 1350 m elevation on the volcanic plateau east of the Cascade Range in southern Oregon. The flow regime is snowmelt dominated with cold, wet winters and mild, dry 10 summers. High flow predominates from February to May and low flows from July to September (Hay et al. 2009; Risley et al. 2005). Precipitation averages approximately 50 cm per year (Loy et al. 2001). The area has shifted to a wetter climate over the last 500 years (likely leading to increases in discharge) with a warming and drying period occurring in the last 150 years (Briles et al. 2008). Historical land uses include irrigated agriculture, cattle grazing, ranching, and timber production (Friedrichsen 1997). Because the growing season is too short for most crops, riparian areas continue to be used mainly for summer pasture. The Sprague River has broad floodplains dominated by herbaceous meadows and wetlands (O'Connor et al. forthcoming 2013). Snowmelt floods regularly inundate parts of the floodplain during late spring. Shrub and tree cover on the floodplain is very low, except in a few narrow valley reaches, which support several willow species and quaking aspen (Populus tremuloides). The floodplains are dominated by native herbs, including Eleocharis palustris (spikerush), Mentha sp. (mint), Polygonum sp. (smartweed), P. arundinacea (reed canarygrass), Solidago sp. (goldenrod), Epilobium sp., and a variety of Juncaceae (rushes), and Cyperaceae (sedges, eg. Carex vesicaria) (Rasmussen 2011). The three dominant riparian species typically inundated during high flows are P. arundinacea, E. palustris, and C. vesicaria. P. arundinacea and E. palustris dominate Site 1 and the invasive and C. vesicaria dominate sites 2 and 3 (Figure 2.1). Work regarding the geomorphic influence of vegetation on this river is of particular importance due to the pending removal of four dams on the Klamath River downstream of the Sprague, which would reopen the river to salmon. Therefore, there is particular concern about river channel and habitat characteristics and any future alterations in geomorphic characteristics due to invasive plants. 11 Figure 2.1. Sprague River Basin and three study sites used in this study. 4. Plant Species 4.1. Reed Canarygrass (Phalaris arundinacea L.) P. arundinacea (Figure 2.2a.), recognized as one of the most invasive plants in North American wetlands (Lavoie et al. 2005), is a cool season, sod forming, perennial grass found in wetlands, riparian areas, and prairies throughout North America. P. arundinacea was first introduced by Euro-American settlers in the Pacific Northwest for stream channel stabilization, erosion control, wastewater treatment, and forage in 1918 (Anderson 1961; Marten and Heath 1973; Naglich 1994). Reproduction occurs both sexually and through rhizomatous propagation with the rhizomes accounting for over 70% of new shoot growth (Antieau, 2004). In the Sprague River basin, P. arundinacea is 12 taller (up to 1.5 m) than the native riparian herbaceous species (0.3 m). P. arundinacea invades vegetated areas after significant hydrological alterations (Galatowitsch et al. 1999). It outcompetes native vegetation by having a high tolerance for inundation, reproducing competitively, and developing a thick sod layer from the previous season leaf growth. This sod layer decreases available habitat for native species. Its ability to grow rapidly early in its life cycle and tolerance for anoxic soil conditions also provide an advantage for survival (Antieau 2004; Apfelbaum and Sams 1987; Foster and Wetzel 2005; Kercher and Zedler 2004) and its presence has been correlated with decreases in vegetation species richness (Fierke and Kauffman 2006). Studies have suggested P. arundinacea can reduce stream velocity, decrease suspended sediment, and increase sediment deposition (Tourbier and Westmacott 1981). Along the Sprague River, P. arundinacea occupies elevations up to 1m vertical from bankfull and up to 5m horizontally away from the bankfull line. Native species E. palustris and C. vesicaria occupy areas similar to the invasive. All three sites used in this study (Figure 2.1.) contain stands of the invasive species. 4.2. Creeping Spikerush (Eleocharis palustris L.) E. palustris (Figure 2.2b.) occurs in wet meadows, vernal pools, fields, pastures, brackish tidal marshes and shallow water globally (Pojar and MacKinnon 1994). Stems are typically 10-30 cm tall and leaves are primarily basal sheaths. Along the Sprague River, E. palustris inhabits areas continually inundated with water but can survive further from the channel in areas not inundated during low flows (Guard 1995; Pojar and MacKinnon 1994). Along the Sprague River, E. palustris inhabits elevations up to 0.3m 13 vertically from Bankfull and often immediately along the channel water line. All three sites (Figure 2.1.) contain stands of E. palustris, though it is more dominant of the natives at Site 1. 4.3. Inflated Sedge (Carex vesicaria L.) C. vesicaria (Figure 2.2c.) inhabits wetlands, ponds, lake shores, and river margins throughout eastern Oregon and Washington. As is typical of other Carex species, it has 1.8-6.5 mm wide leaves extending from the base of the plant and can be up to 60 cm tall. Along the Sprague River, C. vesicaria occupies shallow water areas directly adjacent to the channel and is often inundated during low flows. These areas are can be up to 0.3m higher in elevation than the bankfull line and within 1m horizontal from the bankfull line. P. arundinacea has been known to outcompete C. vesicaria in areas where water table lowering has occurred (Wilson et al. 2008). Sites 2 and 3 (Figure 2.1.) contain stands of C. vesicaria. (a) (b) (c) Figure 2.2. Root and stem systems of partial plants (a) P. arundinacea, (b) E. palustris, and (c) C. vesicaria. Squares on the scale are centimeters. 14 5. Methods 5.1. Root Distribution and Root Area Ratio Root distribution was assessed at three locations for each species using the wall profile method described in Bohm (1979). In this study, each wall profile (hereafter referred to as a pit) represents the root characteristics of one plant. I selected the sites located along the banks within plant stands likely to be inundated during winter high flows. In addition, sites had to contain plants stands that were dominated by the species of interest with few or no other species present. Three soil pits were dug for each species of interest. Soil pits were dispersed among the three sites based on species dominance and where the team had permission to dig. The three soil pits for each of the P. arundinacea and E. palustris species were located at Site 1. Two soil pits for C. vesicaria were located at Site 2 and one at Site 3. Soil pits measured 0.5 x 0.5 x 0.5 m, which allowed the recording of the location (depth) and size of all roots that intersected the vertical profile wall. Root diameters were measured and recorded according to their distance from the ground surface (depth). Then the root diameter and location for each species at 10cm depth increments were used to calculate RAR, or the ratio between the sum of the cross-sectional area of roots intersecting the profile wall and the area of the profile wall (Abernathy and Rutherford 2001; Gray and Leiser 1982). Root distribution, strength, and added cohesion due to roots can vary greatly according to moisture content, growing season, root bark characteristics, nutrient availability, and the angle of the root relative to the trench wall (Pollen and Simon 2005; Greenway 1987). For these reasons, sampling was carried out in the riparian environment where I am addressing the stabilizing effects of vegetation. Therefore, 15 moisture and nutrient values should be typical of other riparian locations along the Sprague River. In addition, sampling was conducted during the same period of the season for all plants and pits. Root bark characteristics likely vary by species and not by site or season and care was taken to pull the roots at the angle they were oriented to the pit wall. Therefore sampling is representative of the conditions I am attempting to capture in this study. 5.2. Root Strength Root strength was measured using the methods outlined in Abernethy and Rutherford (2001) and Pollen and Simon (2005). A 100 lb load cell was attached to a root exposed on the same profile wall of pits previously dug for root distribution surveys. Approximately 30-50 roots per profile wall were tested for each plant depending on the number of available roots and their condition after digging. Individual roots were pulled to their breaking point and the peak force required to break each root was recorded. In addition, the diameter of the root at the breaking point was recorded. I tested 98, 82, and 119 roots for P. arundinacea, E. palustris, and C. vesicaria, respectively. The peak force required to break each root of known diameter was then used to calculate tensile strength (Bischetti et al. 2005): (5) Where Fmax is the peak force (N) needed to break the root, and D is the root diameter (m) at the breaking point. These values were used to develop power relationships for each species relating root tensile strength to root diameter to input into the model RipRoot. 16 5.3. Cohesion Due to Roots Previous studies have observed that root strength per unit area of root decreases with increasing root diameter (e.g., Pollen and Simon 2005). This is due to higher cellulose content in smaller roots making them stronger per unit area than larger roots of the same species (Pollen and Simon 2005; De Baets et al. 2008; Greenway et al. 1984; Wu 1976; Burroughs and Thomas 1977; Bischetti et al. 2005). The relationship takes the form of a power law: Tr = a D -b (6) Where Tr is root tensile strength and D is root diameter. The power relationship coefficients for each of the three species, along with root diameter distributions (number and diameter of roots present), and soil type (moderate silt), were input into the model RipRoot (Pollen and Simon 2005; Pollen-Bankhead and Simon 2009) to quantify the added cohesion provided by each species tested. For comparative purposes, root tensile strength measurements were also used to calculate cohesion using the Wu et al. (1979) method (equation (4)). This method simplifies the tangential component resisting shear and the normal component increasing the confining pressure on the shear plane (equation (3)) and assumes all roots break simultaneously. 6. Results 6.1. Root Distribution and Root Area Ratio Root size distributions show variability within species (Figure 2.3.), but general trends are apparent. The residuals of the sizes of the three species were not normally 17 Table 2.1. Roots per pit and root diameter size for three species. # Roots/Pit Median Root Diameter (mm) P. arundinacea Pit 1 292 0.03 P. arundinacea Pit 2 268 0.12 P. arundinacea Pit 3 165 0.1 Average 241.67 0.08 E. palustris Pit 1 127 0.22 E. palustris Pit 2 113 0.11 E. palustris Pit 3 73 0.17 Average 104.33 0.17 C. vesicaria Pit 1 89 0.4 C. vesicaria Pit 2 146 0.49 C. vesicaria Pit 3 333 0.49 Average 189.33 0.46 distributed, therefore I used the Mann-Whitney-Wilcox statistical test to determine whether root sizes were significantly different between species. The three species have significantly different root sizes (p<0.001, Table 2.1). Root Area Ratio distribution with depth for the three species is shown in Figure 2.4. RAR values per soil depth class are average values over the 3 plants for that species. Overall, roots of all three species occupy less than 0.05% of the soil area. As expected, the RAR decreases with depth for each species. C. vesicaria has a higher RAR than the other two species. Standard errors for the measurements range from 1.05 x 10 -9 to 2.13 x 10 -4 . The three species have different RAR values in the upper 0.3m of soil where most root strength lies with little to no overlap among the species, given the standard error. C. vesicaria has the highest potential for cohesion based on RAR alone, whereas the invasive and native E. palustris have lower RAR. 18 Figure 2.3. Root diameter histogram for three species of interest Figure 2.4. RAR distribution by depth for P. arundinacea, E. palustris, and C. vesicaria. Standard errors of RAR are presented by the extending lines. 0 10 20 30 40 50 % F re q u en cy Root Diameter (mm) P. arudinacea E. palustris C. vesicaria 19 6.2. Root Strength and Cohesion Statistical tests (Mann-Whitney-Wilcox) show that P. arundinacea and E. palustris tensile strengths (Table 2.2) are significantly different (p<0.01). P. arundinacea and C. vesicaria tensile strengths are not significantly different (p=0.086). As found in previous studies (Pollen and Simon 2005; De Baets et al. 2007; De Baets et al. 2008), root tensile strength per unit root area decreases with increasing root diameter in the form of a power relationship for each of the three species. Table 2.3 lists the parameters (a and b), root diameter range, tensile strength range, R 2 values for each power relationship, and the number of roots tested for each species. Power relationship curves for each species are shown in Figure 2.5. The power relationships for each species, in addition to root size distribution for each of the three plants per species, were input into the model RipRoot to obtain cohesion values for each respective plant (Table 2.4). For comparison purposes, additional cohesion provided by roots was also calculated using the Wu et al. (1979) method. Using RipRoot, the results show that C. vesicaria, with an average of 0.88 kPa has the highest added cohesion due to roots. P. arundinacea, closely follows this with an average of 0.81 kPa. Both these values are much higher than E. palustris, which averages 0.30kPa. In addition, cohesion calculated using the Wu et al. model are greater for all excavated pits and species when compared to the RipRoot values. 20 Table 2.2. Root diameter and root tensile strength distributions for three species of interest Root Diameter (mm) Root Tensile Strength (Mpa) Species Median SD Max Min Median SD Max Min P. arundinacea 0.44 0.27 1.26 0.08 24.36 30.88 156.83 1.92 E. palustris 0.25 0.17 0.90 0.07 32.55 79.69 567.61 0.80 C. vesicaria 0.50 0.40 1.88 0.01 24.88 92.98 635.28 1.56 SD=Standard Deviation Table 2.3. Diameter and tensile strength power relationship variables for three species of interest Table 2.4. RipRoot and Wu et al. value comparisons of added cohesion due to roots. Species RipRoot, kPa Wu et al., kPa RipRoot/Wu et al. Wu et. al Overestimation (%) P. arundinacea Pit 1 0.59 2.99 0.20 408.67 P. arundinacea Pit 2 0.96 4.91 0.20 411.93 P. arundinacea Pit 3 0.88 4.54 0.19 417.49 E. palustris Pit 1 0.30 0.76 0.39 157.59 E. palustris Pit 2 0.36 0.93 0.38 160.79 E. palustris Pit 3 0.25 0.63 0.39 155.81 C. vesicaria Pit 1 1.14 7.70 0.15 577.46 C. vesicaria Pit 2 0.52 3.58 0.14 591.16 C. vesicaria Pit 3 0.99 6.67 0.15 571.93 7. Discussion Median and mean values of root diameter reported for similar grass species in other studies (Table 2.5) are larger than diameters reported in this study. In addition, previous work does not show significant differences in tree and herbaceous (grass) vegetation root diameters. The smaller root sizes reported here could be attributed to Plant Species Vegetation type a b n R 2 P. arundinacea Grass 0.0027 1.160 98 0.5352 E. palustris Sedge 0.0003 1.732 82 0.5346 C. vesicaria Sedge 0.0107 1.012 119 0.6536 21 Figure 2.5. Root diameter and root tensile strength power relationship for P. arundinacea, E. palustris, and C. vesicaria 0.1 1 10 100 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 R o o t T en si le S tr en g th ( M p a ) Root Diameter (mm) P. arundinacea E. palustris C. vesicaria Power (P. arundinacea) Power (E. palustris) Power (C. vesicaria) 22 both species differences and environmental differences like nutrient and water availability, and soil type. Previous studies have shown that erosion resistance is highly influenced by fine root density (<5mm). The overall trend in RAR is a general decrease with depth for the three species in question, with most of the RAR in the upper 30 cm of soil. Previous work shows a general decrease in RAR with depth, although woody species and some herbaceous species show more roots below 30 cm than the species in this study. For example, a number of tree and shrub species studied in DeBaets et al. (2007) showed roots to a depth of 0.4 to 0.8m. Similar to findings in this study, grass, herb, and reed species in De Baets et al. (2007) showed roots to depths of 0.15 to 0.35m. Young lodgepole pine showed roots to 0.85m deep (Simon et al. 2006) and Wynn et al. (2004) found herbaceous vegetation root volume ratios (a value comparable to RAR) up to 0.9m deep. Each of the species examined in this study shows a slight increase in RAR between 0.1-0.2m deep. Other studies have also shown a slight increase in RAR below the surface layer, followed by a decrease in RAR with increasing depth. For example, beech trees had a slight increase in RAR to 0.3m deep and then a gradually decrease to 1.10m deep. In addition, Norway spruce increased to 0.2m and then decreased to 1m below the surface (Bischetti et al. 2005). Other studies have found no patterns in RAR with depth (Abernathy and Rutherford 2001; Simon et al. 2006). As with root diameter, RAR with depth can depend on soil characteristics and nutrient availability. Previous studies regarding root tensile strength have shown similar results to those presented here for herbaceous plants (Table 2.5, switchgrass and gamma grass). 23 According to these other studies, tree species have higher root tensile strength values than herbaceous species and grasses. When compared to previous studies, the three species in this study show slightly lower cohesion values both when using the RipRoot model and the Wu et al. (1979) method (Table 2.6). The Wu et al. overestimation is consistent with previous literature (e.g. Pollen et al. 2004) and is due to the static model’s inability to account for the progressive breakage of roots. Many of the species previously studied are trees, though grass cohesion values reported in other studies are also higher than those found in this study. This may be due to the large number of small diameter roots found in this study. Also, unlike previous studies, roots in this study are not present in great numbers at depths greater than 0.35m, which likely deflates cohesion values (De Baets et al. 2007). Although small roots provide the most cohesion, the absence of larger roots could be significant and therefore affect overall cohesion. In addition, we used the pit excavation method and did not measure all roots present within a plant, which could underestimate total cohesion provided by roots. Further work should be conducted involving complete plant excavation. 7.1. Cohesion Due to Roots Cohesion is a function of three factors: root size distribution, RAR, and the tensile strength of individual roots. Root size distributions show the invasive P. arundinacea is dominated by the smallest root sizes. E. palustris is dominated by roots of medium-small size whereas C. vesicaria has the widest range of root diameters and, 24 unlike the other two species, is dominated by medium sized roots. RAR for the invasive P. arundinacea lies between that ofnative E. palustris and C. vesicaria with the latter exhibiting much higher RAR values at depths not exceeding 0.2 m. Therefore, in areas where the invasive is outcompeting native species, it is increasing RAR compared to E. palustris and decreasing RAR compared to C. vesicaria. Within the power law equation for root tensile strength the exponent b, relating tensile strength and root diameter, controls the rate of the decline of strength with diameter, and the constant a is a scale factor in the relationship (Bischetti et al. 2005). The species with the highest scale factor and lowest decay rate, C. vesicaria, is the species that provides most resistance to failure and erosion. The less resistant species, or that species with the lowest scale factor and highest exponent, is E. palustris. Therefore, the power equations show that for a given root diameter, C. vesicaria has the highest tensile strength. This is closely followed by the invasive P. arundinacea and the native E. palustris. Table 2.5. Root diameter and strength of other species Species Name Vegetation Type Diameter (mm) Root Tensile Strength (Mpa) Fagus sylvatica a Tree 1.33 57.47 Salix purpurea a Tree 1.28 51.47 Salix caprea a Tree 1.42 47.8 Fraxinus excelsa a Tree 1.95 36.86 Larix decidua a Tree 1.68 66.14 Switchgrass b Grass 1.3 19 Gamma grass b Grass 1.4 27.7 Swamp Paperbark c Tree 7.4 14.3 River Red Gum c Tree 5.7 19.4 a Bischetti et al. 2005, b Simon and collision 2002, c Abernethy and Rutherford 2001 25 Table 2.6. Vegetation type and cohesion of other species Species Name Vegetation Type Cohesion (RipRoot, kPa) Cohesion (Wu et al, kPa) Variety of Grasses a Grass 100 - 244 Juncus acutus a Rush 304 Anthyllis cytisoides a Shrub < 160 River birch b Tree 1.75 28.9 Black Willow b Tree 1.75 20 Sycamore b Tree 1.75 24.9 Switchgrass b Grass 1.85 50.5 Longleaf pine b Tree 1.75 14.2 Norway Spruce c Tree 14.3 - 58.6 European beech c Tree 40 - 100 European larch c Tree 15 - 60 Sweet Chestnut c Tree 6.4 - 19.0 European hop- hornbeam c Tree 5.4 - 30 Sweetgum d Tree 4 Eastern gamma grass d Grass 6 Switch grass d Grass 18 Tamarisk e Tree 0 - 6.9 Russian-olive e Tree 0 -14.2 a De Baets et al. 2008, b Pollen and Simon 2005, bank height is 1m, c Bischetti et al. 2009, conversion factor used to prevent Wu et al 1979 overestimation, d Simon and Collison 2002, e Pollen-Bankhead et al. 2009 Regarding cohesion obtained from the model RipRoot, the high cohesion values for C. vesicaria result from a combination of the medium density of large roots and the largest RAR. P. arundinacea exhibits a high cohesion value as well due to the presence of a large number of small roots. E. palustris plants provide the lowest amount of soil cohesion due to roots because, although its roots exhibit large tensile strengths, it has the lowest number of roots and they are small in diameter. 26 7.2. Restoration Implications Quantifying the root cohesion of riparian species is important because shifts in species dominance may change bank stability and bank erosion processes. Reaches occupied by the invasive that were previously inhabited by the native E. palustris are experiencing up to three times increased bank stability and resistance to erosion. Measurements conducted in this study primarily focus resistance to erosion via roots. However, the aboveground biomass of the vegetation also has the ability to cause geomorphological changes by altering roughness and trapping sediment. P. arundinacea, unlike the native species studied here, develops a sod layer that protects bank surfaces from erosion. In future work, the aboveground structures of the species should be examined to address this question. This may include quantification of sod layer properties and differences between the species using laboratory flume methods or field measurements. Changes in bank stability and bank erosion resulting from vegetation species shift can lead to changes in channel form. Expansion of P. arundinacea may reduce the ability of floods to uproot vegetation along the channel margin and therefore reduce the erosional effectiveness of floods. Increasing bank stability and decreasing erosion in these areas means channel width is more likely to remain static or even decrease (Myers and Swanson 1992, 1996; Duff 1977; Overton et al. 1994). With narrower channels, increased shear stress on the channel bed may result in incision, and increased stream depth that may ultimately affect fish habitat. Changes in channel form, either decreased channel width or increased channel depth, may help to decrease stream temperatures (Matthews 1996; White and Brynildson 27 1967). With the pending removal of dams downstream of the Sprague River and the potential for salmon presence in the future, changes in channel form that affect habitat, particularly water temperature, are important. Currently, the Sprague River exceeds summer water temperature standards, and decreasing water temperatures may be important for salmon survival (Beschta et al. 1987; Bjornn and Reiser 1991; Boyd and Kasper 2002). Ultimately, decreases in stream temperatures will result in more hospitable conditions for salmon in this area (Boyd and Kasper 2002). Values reported here for the Sprague River are likely applicable across different locations in some instances. Tensile strength measurements and power relationships are consistent within species; however, root distribution can be highly variable depending on substrate and local nutrient and water availability (see model RipRoot and Pollen and Simon 2005). Soil pits for this study were located on depositional surfaces and therefore close to the water table where roots were not required to extend to great depths. I have observed that when occupying cutbanks, the invasive has developed longer, deeper root systems to reach water levels and would therefore likely have different root distributions than when inhabiting depositional surfaces. Therefore, care should be taken when comparing results across habitats and watersheds. 8. Conclusions Few studies have examined the impact of small, non-woody riparian vegetation on stream bank stability and resistance to erosion. We have shown that differences in root size, root distribution, root strength, and root cohesion exist between the invasive P. arundinacea and natives E. palustris and C. vesicaria, the most significant of which is 28 the high bank stability and resistance to erosion properties of P. arundinacea compared to the native E. palustris. The increased level of cohesion in areas where the invasive is outcompeting the native species may be leading to zones of decreased erosion. The change in bank cohesion is likely to cause the most effect in areas where P. arundinacea is outcompeting E. palustris because the cohesion values of the invasive are almost three times greater. Such changes in bank stability could lead to channel narrowing and other channel form water temperature regime changes that would improve salmon habitat which is critical given the planned removal of four Klamath River dams Therefore, it is important to determine the past and predict the future effect of the invasive P. arundinacea on channel morphology. 29 CHAPTER III QUANTIFYING ROUGHNESS: FLEXIBLE RIPARIAN VEGETATION AND MANNING’S n 1. Introduction Manning’s n is used to quantify hydraulic resistance to flow, determine flow velocities, and calculate discharge. However, accurately estimating this parameter is a difficult task. Roughness can be influenced by vegetation, bedforms, sinuosity, obstructions (e.g. wood), bedrock outcrops, and skin friction (Simon and Castro, 2003; Gilley and Finkner, 1991; Roberson and Crowe, 1993). Researchers have continually modified and developed methods, the majority of which are based on statistical relationships with other hydraulic variables or based on the topographic characteristics of the channel, to combat the uncertainties in roughness calculations. However, few of these account for the added roughness provided by plant stands. Riparian vegetation passively engineers rivers (Viles, 1988) by increasing roughness through above ground biomass (Gurnell et al., 2010; McKenney et al., 1995; Perucca et al., 2007). The degree to which vegetation influences flow velocities is dependent on the density and height of plants and their stiffness, or resistance to flow (Gleason et al., 1979; Gran and Paola, 2001). Aquatic macrophyte shoot density and shape were found to influence flow velocity and therefore sediment deposition (Asaeda et al., 2010; Clarke et al., 2004; Schulz et al., 2003). Submerged bur-reed shoots with simple stems and no leaves increased fine sediment deposition within the vegetation stand. Within Callitriche cophacarpa stands, Sand-Jensen and Mebus (1996) found an eleven-fold decrease in stream velocity due to the establishment of dense stands and 30 shoots. Within tidal marsh landscapes, grasses were found to decrease velocity and near- bed turbulence (Leonard and Luther, 1995), cause sediment deposition and divert flow around vegetation patches (Reinhardt et al., 2010; Temmerman et al., 2006; Van Hulzen et al., 2007). Small changes in deposition and erosion brought on by vegetation can lead to changes in channel form (Corenblit et al., 2008; Gurnell, 1997; Perucca et al., 2007). Spartina anglica (common cordgrass) in a tidal marsh landscape both increased deposition by reducing velocities within the plant stand and caused scour outside the stand by diverting flow outside of the canopy (Van Hulzen et al., 2007). In rivers, channel width and wetted perimeter values are lower in channels with banks that contain a high percent of vegetation cover while depth tends to increase with increased levels of vegetation cover (Hey and Thorne, 1986; Millar and Quick, 1998). Studies show that channel widths in rivers lined with grass can be up to three times as wide for a given discharge as those covered with trees (Millar and Quick, 1998). In addition, vegetation can control channel planform and pattern (Millar, 2000; Millar, 2005; Nevins, 1969; Tal and Paola, 2007) and can shift braided channels to single thread channels (Murray et al., 2008; Tal and Paola, 2010). Therefore, quantifying vegetation roughness can be useful for predicting deposition, erosion, and channel planform changes. Changes in species composition and the subsequent change in roughness values is of particular importance where invasive species are rapidly outcompeting native species and establishing monocultures. Along the Sprague River in southeastern, Oregon, the invasive Phalaris arundinacea (reed canarygrass) occupies similar microsites as those occupied by dominant natives Eleocharis palustris (creeping spikerush), and Carex 31 vesicaria (inflated sedge) and is outcompeting these species within the riparian zone. If these species have different roughness values, changes in species composition can change deposition and erosion along the stream channel and thus alter channel morphology and aquatic habitat. Our goal is to determine differences in plant characteristics and roughness properties of P. arundinacea and native species C. vesicaria and E. palustris. Here I measure the stem stiffness and other plant characteristics of multiple individuals of each of the three species. These variables are used to calculate the associated roughness values for the three species to predict changes in roughness where the invasive is outcompeting native species. In addition, I directly measured deposition within plant stands in riparian areas using deposition pins and mats. 2. Methods 2.1. Site Selection I chose sites located on depositional surfaces below bankfull stage that contained representative plant stands dominated by the three species of interest. Plant stands were located approximately 1m above the channel bed (thalweg) and vegetation roughness, or nveg, was calculated for partially submerged vegetation. Twelve, 10, and 8 plants for P. arundinacea, E. palustris, and C. vesicaria, respectively were chosen at three main sites within the Sprague River Basin (Figure 2.1.). Sites were selected based on location access and presence of dominant plant stands that contained few or no other species within the stand. In addition, all sites were located on depositional surfaces at similar 32 elevations along the banks of the river where high flows would interact with the vegetation. 2.2. Plant Characteristics and Other Variables I measured the modulus of plant stiffness (Es, N/m 2 ), plant area (m 2 ), plant density (#plants/ m 2 ), plant frontal area (m 2 ), stem density (#stems/ m 2 ), and total cross-sectional area of all the stems of an individual plant, (m 2 ) for each of the three species. The modulus of plant stiffness, Es (Rahmeyer et al., 1999; Kouwen, 1988), or stem stiffness, was measured by securing a 100lb load cell at half the height of the vegetation stem to determine the force required to bend to an angle of 45°. Ten plants with a total of 137 stems were tested for E. palustris, eight plants with a total of 160 stems were tested for C. vesicaria, and 12 plants with a total of 181 stems were tested for P. arundinacea. I converted this load to Newtons and input it into the equation (Freeman et al., 2000; Rahmeyer et al., 1999: refer to Table 3.1 for variables): (1) to calculate the modulus of plant stiffness (Es, N/m 2 ). A high Es value would mean higher stiffness, or lower flexibility, of the plant stem. I estimated roughness based on the above discussed stem stiffness for each of these plant stands, and the total plant area occupied by each individual plant and determined plant density (number of plants/m 2 ). Plant frontal area was measured using a 50cm x 1.5m coverboard (Sage et al., 2004; Nudds, 1977; Griffith and Youtie, 1988). Plant frontal area was recorded as the percent cover of each 50cm x 10cm panel for each of the 10cm height increments from the ground up to the height of the plant. Percent was 33 then converted to area of coverage to determine total plant frontal area for the plant at 80% of its height (input required for partially submerged vegetation). Plant area was measured as total canopy cover. Table 3.1. Variables for equations and associated descriptions. Variable Description n Manning’s roughness term H Average undeflected plant height, m Ai * Net submerged frontal area of partially submerged plant, m2 As Total cross-sectional area of all the stems of an individual plant, m2 Ds Stem diameter, at a height of H/2, m Es Modulus of plant stiffness, N/m2 F45 Horizontal force to bend stem 45°, N Kn Unit conversion factor 1.0m1/3/sec M Relative plant density, m2 p Fluid density, 1000 kg/m3 Rh Hydraulic radius, m S Slope V* Shear velocity, m/s v Fluid dynamic viscosity, m2/s, Additionally, I measured stem density by counting the number of stems within a 0.1 x 0.1 m sampling plot. I measured stem diameter at ¼ the height from the ground for each stem, as is required according to equation protocol. Total cross sectional area of stems was measured within this plot by summing the stem diameters and extrapolating this to the total size of the plant, per equation protocol. 2.3. Calculating Roughness In addition to the above measured variables, standard variables were determined for this river system using accepted hydraulic values and measurements on topographic maps, sonar data, and digital elevation models. These include: hydraulic radius (Rh, 34 1m), fluid density ( 1000 kg/m3), slope (0.0002 m/m), shear velocity, ( , 0.0443 m/s), and fluid dynamic viscosity ( 1 m2/s). For each plant of each of the three species of interest, roughness with regard to vegetation alone (ignoring channel characteristics), nveg, for partially submerged vegetation (inundating the vegetation up to 80% of its height) was calculated using equation (2) (Rahmeyer et al., 1999) and the median Es value for that plant stand. This resulted in a unitless roughness value for each plant stand of interest (refer to section 6 for variables). ( ) ( ) ( ) ( ) (2) 2.4. Measuring Deposition To directly measure deposition with invasive and native plants stands, I set up four sampling stations for sediment deposition rates from mid October 2011 to late July 2012 using two methods: deposition pins and mats (Gurnell et al., 2006a; Gurnell and Petts, 2006). Each station consisted of 1 stand of the invasive species and a nearby stand of a native species at a similar bank elevation (Figure 3.1.). Two stations consisted of a pairing of the invasive P. arundinacea and E. palustris and two stations of P. arundinacea and C. vesicaria. Within each of the plant stands, I used two types of 9cm x 6cm mat pieces to minimize flow interruption within the plant stand: artificial turf and a rubber mat. In addition, at each stand, I inserted six deposition pins and the distance from the top of the pin to the current ground level was recorded for each pin. Both the mats and deposition pins were left in place through the flood season (Figure 3.2.) to determine sediment deposition after significant flows. Sites were inundated with up to 35 Figure 3.1. Deposition pin (yellow circles) and mat (green=turf mats, black=rubber mats) setup along the banks of the Sprague River (light blue polygon). Figure 3.2. Stage heights for period of time deposition pins and turf mats were in place. Stage data obtained from Oregon Water Resources Department (OWRD) gage station 11497500 near Beatty, Oregon. 1.5m of water given that they were placed approximately 1m above the channel bed. I collected the mats after the flood season and separated the sediment. Sediment was dried, weighed, and converted to volume using a bulk density of 1.3g/cm 3 . Deposition pins were measured and recorded with regard to the current ground level after the flood season. 0 0.5 1 1.5 2 2.5 S ta g e (m ) O N D J F M A M J J 2011 2012 Invasive Stand Native Stand 36 3. Results 3.1. Plant Characteristics I measured 43 stem density plots for P. arundinacea, 33 for E. palustris, and 36 for C. vesicaria. Median stem density values were 800, 4800, and 2850 stems per m 2 , for P. arundinacea, E. palustris, and C. vesicaria, respectively (Figure 3.3.). Overall, E. palustris showed the highest stem density and the invasive P. arundinacea showed consistently lower stem densities than the two natives. The invasive had significantly different stem density values compared to each of the native species (p<0.001). Stem stiffness values varied greatly for each species (Figure 3.4). I tested a total of 270, 214, and 230 stems for P. arundinacea, E. palustris, and C. vesicaria, respectively. Stem stiffness was highest for the invasive. This was followed by E. palustris and then C. vesicaria. P. arundinacea had significantly different stem stiffness compared to both of the native species (p<0.001). Table 3.2. shows the total plant area, plant density based on plant area, and plant frontal area for the three species. P. arundinacea, on average, had the largest frontal area. This is followed by C. vesicaria, and then E. palustris, which had the smallest frontal area. Additionally, cross sectional area of stems for all three species are displayed on Table 3.2. P. arundinacea and C. vesicaria averaged similar cross sectional areas while E. palustris averaged a much higher cross sectional area. 3.2. Roughness Vegetation roughness values for each plant within each species were calculated with the median stem stiffness values of that plant and their respective variables including plant 37 area, plant density, plant frontal area and stem cross sectional area, along with the standard variables established for hydraulic radius, slope, shear velocity and fluid dynamic viscosity, using equation (2). A boxplot of the associated vegetation roughness for each species is shown in Figure 3.5. E. palustris showed the highest vegetation roughness values with a median of 0.000323. It also had the largest range of values, which is due to its high range in stem stiffness. The invasive had a roughness median of 0.000304 and had a smaller range of values. E. palustris and P. arundinacea roughness values were not significantly different (Mann-Whitney-Wilcox, p=0.6277). C. vesicaria had the lowest roughness values with a median of 0.000132 and was significantly different from the invasive (p<0.001). Figure 3.3. Box plots of stem density. The bold line represents median values and the top and bottom box edges represent the 75 th and 25 th percentiles, respectively. Whiskers represent the range of data. P. arundinacea E. palustris C. vesicaria S te m D en si ty ( # S te m s/ m 2 ) 38 Figure 3.4. Stem stiffness for three species of interest. The bold line represents the median value and the top and bottom box edges represent the 75 th and 25 th percentiles, respectively. Whiskers represent the upper and lower quartiles to the nearest data point not beyond 1.5 times the inter-quartile range. Outliers are shown as open circles beyond the whiskers. S te m S ti ff n es s (E s, N /m 2 ) (S te m s/ m 2 ) 0 1 x 1 0 9 2 x 1 0 9 3 x 1 0 9 4 x 1 0 9 P. arundinacea E. palustris C. vesicaria 39 Table 3.2. Frontal and cross-sectional area characteristics for three species Total Plant Plant density Plant Frontal Area Cross Sectional Area Area (m 2 ) ( #plants/m 2 ) (m 2 , 80% height of plant) of Stems (m 2 , whole plant) P. arundinacea Plant A 3.46 0.29 0.62 0.03 Plant B 3.46 0.29 0.62 0.02 Plant C 5.04 0.20 1.60 0.03 Plant D 5.04 0.20 1.60 0.04 Plant E 9.00 0.11 0.83 0.04 Plant F 2.00 0.50 0.56 0.04 Plant G 9.00 0.11 0.81 0.04 Plant H 14.00 0.07 0.81 0.04 Plant I 6.00 0.17 0.96 0.04 Plant J 12.00 0.08 1.86 0.04 Plant K 2.50 0.40 0.71 0.04 Plant L 9.00 0.11 1.31 0.04 Median 5.52 0.18 0.82 0.04 E. palustris Plant A 4.35 0.23 0.49 0.16 Plant B 4.35 0.23 0.49 0.01 Plant C 0.06 16.67 0.04 0.0005 Plant D 0.03 31.25 0.02 0.0002 Plant E 3.90 0.26 0.33 0.05 Plant F 3.90 0.26 0.33 0.02 Plant G 0.75 1.33 0.11 0.21 Plant H 0.50 2.00 0.21 0.21 Plant I 10.00 0.10 0.36 0.21 Plant J 14.00 0.07 0.71 0.21 Median 3.90 0.26 0.33 0.13 C. vesicaria Plant A 2.34 0.43 0.78 0.08 Plant B 2.34 0.43 0.78 0.03 Plant C 3.65 0.27 0.77 0.04 Plant D 3.65 0.27 0.77 0.02 Plant E 1.38 0.72 0.77 0.02 Plant F 1.38 0.72 0.77 0.01 Plant G 6.00 0.17 0.37 0.03 Plant H 9.00 0.11 0.71 0.03 Median 3.00 0.35 0.77 0.03 40 R o u g h n es s (n ve g ) Figure 3.5. Vegetation roughness values for three species. The bold line represents median values and the top and bottom box edges represent the 75 th and 25 th percentiles, respectively. Whiskers represent the range of data. 3.3. Deposition Deposition likely occurred during the receding limb of the hydrograph from late April to late July (Figure 3.2.) and consisted of fine-grained sediment (silts and sands). Pin measurements showed a dominance in deposition within all plant stands with the exception of the C. vesicaria stand in Plant Stand D (Figure 3.6.). Deposition depths were similar within stands of the invasive and E. palustris and greater within the invasive when compared to C. vesicaria. Turf mat volumes were standardized by the area of the turf mat (volume of sediment divided by 54cm 2 ). Sediment deposited on the mats showed some patterns (Figure 3.7). P. arundinacea stands showed similar deposition values as those in E. palustris stands. In addition, P. arundinacea stands showed higher deposition values than C. vesicaria in Plant Stand C. Plant stand D seems to be P. arundinacea E. palustris C. vesicaria 2 x1 0 -4 4 x1 0 -4 6 x1 0 -4 41 anomalous given that the deposition values obtained from the pins and mats disagree. There did not appear to be a significant difference between mat types (turf mats are bars 1, 2, and 3 while rubber mats are bars 3, 4 and 5 in each cluster of species, Figure 3.7.) Figure 3.6. Deposition (+) and erosion (-) depths for plant stands using pins. Horizontal bars represent average for the species within that plant stand. Figure 3.7. Deposition using turf mats for four plant stands. Horizontal bars represent average for the species within that plant stand. -6 -4 -2 0 2 4 6 D ep th ( cm ) P. arundinacea E. palustris C. vesicaria Plant Stand A Plant Stand C Plant Stand B Plant Stand D 0 1 2 3 4 5 6 D ep o si te d S ed im en t (c m ) P. arundinacea E. palustris C. vesicaria Plant Stand A Plant Stand C Plant Stand B Plant Stand D 42 4. Discussion 4.1. Plant Characteristics P. arundinacea showed the lowest stem densities of the three species of interest. C. vesicaria stems exhibited low to medium stem densities while E. palustris exhibited the highest stem densities of all three species. Therefore, solely based on stem densities, E. palustris has a higher ability than the other species to interrupt flow, which may lead to higher roughness numbers and increased sediment deposition within plant stands. Stem stiffness values for E. palustris were the highest among the three species followed by P. arundinacea and C. vesicaria. P. arundinacea, though it exhibited mid- range stiffness values, also exhibited a large number of higher range values. Therefore, solely due to stem stiffness, P. arundinacea has a higher ability to interrupt flow. Stem stiffness values found in this study are an order of magnitude larger than those of woody species (Table 3.3). This is due to the smaller stem diameters present in the denominator of the Es equation (equation 1), which inflates stiffness values. Table 3.3. Stem stiffness values for woody species published in previous work. Species Stem Stiffness, Es (N/m 2 x10 8 ) Yellow Twig dogwood 1 3.210 Berried Elderberry 1 0.526 Purpleleaf Euonymus 1 4.140 Red Twig Dogwood 1 10.200 Yellow Twig Dogwood 1 29.900 Mulefat 1 5.950 Alder 1 17.000 Valley Elderberry 1 16.500 Willow (Salix exigua) 2 1.470 Willow (Salix lasiandra) 2 0.747 Composite Willow 2 1.390 1Values based on Freeman et al.(2000), 2Values based on Freeman et al (1998) 43 On average, P. arundinacea had the highest plant frontal area (m 2 ) values of the three species. This was followed by C. vesicaria and then E. palustris. Observations in the field coincide with this measurement. Often the invasive occupied the largest area and was the tallest. E. palustris on the other hand, was the smallest plant in terms of height and, unlike the invasive, has no leaves which would create higher plant frontal area measurements. C. vesicaria, had a large frontal area but is often no taller than 0.5m. In terms of its impact on roughness values based on frontal area, the invasive has a higher potential to interrupt flow and create larger obstructions to flow, thus causing deposition. In contrast to this, stem cross sectional area was highest in stands of E. palustris. This was followed by P. arundinacea and C. vesicaria which exhibited very similar values. Overall, E. palustris exhibited values for stem density, and cross sectional area of stems that will lead to the increased possibility of flow interruption and therefore the highest numbers in roughness values based on plant characteristics. P. arundinacea, on the other hand, exhibited the highest values for stem stiffness and plant frontal area because it was the tallest and occupied the largest area. Values presented here for the Sprague may be highly correlated with substrate characteristics, and nutrient and water availability, and therefore may not be applicable in other locations. In addition, measured variables, including stem stiffness, may differ depending on the season data is gathered. For example, new stems may be smaller with lower biomass and therefore lower plant frontal area. In addition, these stems may be more flexible than what would be encountered later in the season. Further work should take care, as this study did, to observe and record plant characteristics similar to those 44 that will be present during flows in which the vegetation is inundated. Additional work could be conducted that compared the variables in different growth stages. 4.2. Roughness P. arundinacea and E. palustris vegetation roughness values are not significantly different though in some cases E. palustris may exhibit higher roughness (Figure 3.5.). The invasive and C. vesicaria show significant differences in roughness. If P. arundinacea dominates and replaces E. palustris, roughness values will likely remain the same or only decrease slightly. However, if P. arundinacea is dominating C. vesicaria, roughness values will be significantly higher than previously. Differences in stem elasticity and stem cross-sectional area seem to have the largest influence on vegetation roughness. High stem densities within E. palustris (Figure 3.3.) likely account for its high roughness values while the low stem stiffness of C. vesicaria (Figure 3.4.) likely accounts for its low vegetation roughness. These roughness values account for only vegetation roughness. To fully predict channel changes, a composite roughness should be calculated. Values for portions of stream banks occupied by the invasive or native vegetation, the vegetation roughness much be treated as an added factor and combined with other roughness factors for the reach of interest. Previous work has illustrated the concept of an additive roughness value by establishing a roughness base value and adding roughness values for surface irregularities, channel shape and size, obstructions, vegetation, and meandering (Arcement and Schneider, 1989; Cowan, 1956; Chow, 1959; Aldridge and Garrett, 1973). Using this method, low gradient, sinuous reaches of the Sprague River in this study 45 would yield roughness values displayed in Table 3.4. The median roughness values for each species were added to this roughness value to obtain a composite channel roughness value which yields a roughness 0.0238 for areas occupied by P. arundinacea, 0.0240 for E. palustris, and 0.0235 for C. vesicaria. Table 3.4. Additive roughness values for the Sprague River. Roughness Component Sprague River Value Description nb: base roughness 0.012 smallest sand size, straight uniform channel n1: surface irregularities 0.002 carefully degraded channel with slightly eroded or scoured sideslopes (0.001-0.005) n2: variations in channel shape 0.002 large and small cross sections alternate occasionally, main flow shifts from side to side (0.001- 0.005) n3: obstructions 0.002 obstructions occupy <5% of cross section area (0-0.004) m: meandering correction 1.3 sinuosity >1.5 The roughness values established here are comparable to those established in previous studies (Table 3.5). With regard to nveg, all species are comparable to previously established nveg values (Gordon et al., 2004; Jarrett, 1984). The vegetation roughness for all three species falls below the “small” composite n value for vegetation ranging from 0.002-0.010 (Arcement and Schneider, 1989). The “small” level corresponds with dense growth of flexible turf grass and though our species are similar to turf grass, stem density is not as high and the three species would therefore have lower roughness values. In addition, additive roughness values estimated in this study are comparable to roughness values for streams similar to the Sprague River in published photo reference guides (Table 3.6). 46 4.3. Deposition Deposition depths measured with deposition pins within P. arundinacea and E. palustris plant stands shows similar patterns. P. arundinacea plant stands show higher deposition values than their comparable C. vesicaria plant stands. Deposition measured on turf mats within P. arundinacea and E. palustris plant stands are also similar and correlate with the deposition pin findings. In addition, they are comparable to previous studies on other rivers (Gurnell et al., 2006b). Deposition values within Plant Stand C comparing the invasive and C. vesicaria are comparable to the pin results showing more deposition within the invasive. However, in Plant Stand D, C. vesicaria shows a higher deposition value based on the mats being deposited than P. arundinacea and these findings disagree with the deposition pin findings at the same plant stand. According to previous studies, deposition mats can accurately capture deposition but not erosion occurring on a site because they provide surface roughness that Table 3.5. Previously established roughness values for other vegetation Description/Species Minimum Additive vegetation values 1 : Supple seedlings or dense grass/weeds 0.0005-0.010 Brushy growths, no growth in streambed; grass height of flow 0.01-.025 Young trees intergrown with weeds; grass twice depth of flow 0.025-0.05 Brushy growth on banks, dense growth in stream; trees with weeds; full foliage 0.05-0.1 Yellow Twig dogwood 2 0.043 Purpleleaf Euonymus 2 0.041 Red Twig Dogwood 2 0.07 Yellow Twig Dogwood 2 0.053 Mulefat 2 0.035 Alder 2 0.07 Valley Elderberry 2 0.072 Composite willow (Salix exigua & lasiandra) 3 0.138 51 cm Dogwood 4 0.034 (nveg) 71 cm Elderberry 4 0.033 (nveg) 97cm Dogwood 4 0.079 (nveg) 1Values based on Gordon et al. (2004) which establishes values based on Cowan (1956) and Jarrett (1984)., 2Freeman et al. (2000), 3Freeman et al. (Freeman et al., 1998), 4Rahmeyer et al. (1999) 47 Table 3.6. Rivers similar to the Sprague River and their associated roughness values via photo reference guides. River and Location Roughness Value Oakden Canal at Oakden Culvert (68521) 1 0.033 Loganburn at Gorge Downstream (74347) 1 0.020 Piako at Paeroa-Tahuna Bridge (9140) 1 0.030 Indian Bend Wash above Curry Road 2 0.036 Verge River near Paulden 2 0.029 Onondoga Creek at Dorwin Ave 3 0.027 Champlin Creek near Colorado City, Tx 4 0.027 Clearwater River at Kamiah, Idaho 4 0.033 1 values based on Hicks and Mason (1998), 2 values based on Phillips and Ingersoll (1998), 3 values based on Coon (1998), 4 values based on Barnes (1967) traps sediment (Steiger et al., 2003). It is possible that erosion, and not deposition is the dominant processes with C. vesicaria at this site. However, deposition pin results may be anomalous given that the field crew observed possible evidence of heave or disruption of the pins by cattle and the invasive and C. vesicaria stands are at significantly different elevations, unlike the other plant stands. Further study is required to ascertain a complete picture on the amount of deposition occurring within C. vesicaria stands. Deposition depths via the pin and mat method correlate adequately well with roughness calculations conducted for this study. P. arundinacea roughness values (nveg), although smaller than E. palustris, fall within the range of the native. Depositions are similarly comparable with one stand showing on average more deposited sediment in the invasive (Plant Stand A) and one stand showing less (Plant Stand B). C. vesicaria roughness values fall below the other two species as is reflected in Plant Stand C where deposition depths based on the mats are lower in the native stand. Therefore, the roughness values correlate with observed deposition and further emphasize the point that Plant Stand D may be anomalous due to hydraulic differences in stand placements. 48 Given these depositions and conditions over time, including flow levels, and sod layer conditions, these results would likely mean aggrading banks on these depositional surfaces on the order of 1cm per year for P. arundinacea and E. palustris. Over time, this may cause a narrowing of the channel in areas where the invasive has outcompeted C. vesicaria. 5. Conclusion Quantifying roughness is a complicated, multi-dimensional problem within river science. However, it is necessary for indirectly estimating discharge, channel capacity and water surface elevations, among many other variables of use. In addition, accounting for the influence of vegetation further complicates the determination of Manning’s n and estimations regarding this complication are almost nonexistent (Copeland, 2000). However, few are attempting to determine how vegetation affects channel flows and thus channel processes. Research regarding the effect of vegetation on channel processes has primarily considered riparian zones dominated by trees (e.g. Gurnell and Petts, 2006) and large woody shrubs such as Tamarix (e.g. Graf, 1978), rather than weaker and more flexible herbaceous vegetation. Previously, flume studies on groundcover plants (Rahmeyer et al., 1999), small trees and shrubs (Copeland, 2000; Freeman et al., 2000) have been conducted to determine resistance due to vegetation. Non-woody species, such as P. arundinacea, may have the ability to alter river channel form and process when present in extensive stands, particularly in a low energy river such as the Sprague. Furthermore, invasive vegetation may influence geomorphic processes differently than native vegetation because of its ability to outcompete other vegetation, its tendency to 49 develop monocultures, and the difficulty encountered in its removal. In this case, P. arundinacea has a significantly higher roughness value than the native C. vesicaria which means river channel processes are being altered in areas where it is outcompeting the native. Boyd and Kasper (2002) suggested that past channel widening along the Sprague River has been associated with negative effects on water temperature and fish habitat. However, the findings presented here reveal that areas with established invasive stands may begin to narrow. With the potential removal of Klamath dams downstream of the Sprague and the reopening of salmon habitat, it is important to determine the past and predict the future effect of the invasive on channel morphology. Establishing measurements of plant characteristics and roughness values is the first step in this endeavor. Future research includes modeling the effect of these roughness differences on channel morphology as well as long term monitoring of channel characteristics and channel changes. 50 CHAPTER IV INVASIVE AND NATIVE RIPARIAN VEGETATION: MODELING CHANNEL HYDRAULICS WITH CURRENT AND FUTURE VEGETATION CONDITIONS 1. Introduction Vegetation along stream banks has the ability to change channel morphology characteristics by altering velocity patterns and changing deposition and erosion. Recent research has focused on quantifying the effects of woody, rigid species (Petryk and Bosmajian, 1975; Pasche and Rouve, 1985; Musleh and Cruise, 2006), rather than flexible vegetation (Kouwen and Unny, 1973; Kouwen et al., 1969; Carollo et al., 2005) or a combination of the two (Freeman et al., 2000; Järvelä, 2004). However, in low gradient streams, dominant flexible vegetation can affect deposition when little to no rigid species are present. Here I examine the impact of the invasive and flexible Phalaris arundinacea L. (reed canarygrass) on channel morphology of the low-energy, meandering, Sprague River in eastern Oregon by modeling its effect on water depth, velocity, and bed shear stress within the 2-D model MD-SWMS (Multi-Dimensional Surface-Water Modeling System). In addition, I compare its geomorphic effect to that of native Eleocharis palustris (creeping spikerush) and Carex vesicaria (inflated sedge). In the near future, four dams along the Klamath River, downstream of the Sprague, will be removed and will open the area to salmon. Any trends in geomorphic processes that may develop or have developed in the past as a result of invasive species colonization are critical to understanding this system. 51 2. Background Investigating the effect of invasive riparian vegetation is critical because invasive species can cause major shifts in ecosystem dynamics, outcompete native species, and produce many other changes of which we are still not yet aware. This is of particular importance when the species has the potential to become an ecosystem engineer. Ecosystem engineers are organisms that control the availability of resources by causing changes in biotic or abiotic materials (Jones et al., 1994; Jones et al., 1997; Corenblit et al., 2008). Rather than directly supply these resources, ecosystem engineers are those organisms that adjust or transform them. Ecosystem engineers can control energy, materials, space, food organisms, or any combination of these resources. Within fluvial geomorphology, researchers study two aspects of vegetation ecosystem engineers: those vegetation characteristics that modify the environment either actively or passively (Corenblit et al., 2009). Riparian vegetation actively modifies the environment by establishing root systems that influence the cohesive properties of channel bed and bank materials (Prosser et al., 1995; De Baets et al., 2006). Vegetation passively modifies the environment by providing roughness within the channel bed or along the banks, decreasing flow velocity and increasing sediment deposition (Corenblit et al., 2009). Roughness can manifest itself and alter the river system in multiple ways. Here I provide some insight into studies regarding flexible vegetation though more work is available regarding rigid woody species (e.g., Gran and Paola, 2001; Edwards et al., 1999; Graf, 1978). Macrophyte shoot density and shape influence flow velocity and therefore, deposition (Schulz et al., 2003; Clarke et al., 2004; Asaeda et al., 2010). For 52 example, submerged Sparganium erectum shoots were found to increase fine sediment deposition within the vegetation stand. Following increased growth in the spring the shoots emerged from the water column which raised the water level and subsequently reduced water velocities (Asaeda et al., 2010). Gurnell et al. (2010) examined multiple macrophytes across 467 British rivers and found they have the ability to trap fine sediment and cause channel form changes. Within Callitriche cophacarpa stands, Sand- Jensen and Mebus (1996) found an eleven-fold decrease in stream velocity due to the establishment of dense stands and shoots. Various studies have also examined bank and island vegetation. On the Northern Plains of Australia, various grasses along channel banks and floodplains were found to provide higher levels of stability and erosion protection than trees and shrubs in the area due to their substantial biomass and expansive root networks (Tooth and Nanson, 1999). Carex nudata in California creates substrate upon which other species can colonize as well. Although Carex occupied only 33.05% of the channel area on the South Fork Eel River, California, the 13 most common species had an average of 84% of their individuals on Carex tussocks. Therefore, the substrate stabilization of Carex was vitally important to the survival of other species (Levine, 2000). The small changes in depositional and erosional processes brought on by the active and passive abilities of vegetative ecosystem engineers can lead to changes in channel form (Perucca et al., 2007; Corenblit et al., 2008; Gurnell, 1997). The deposition and erosion brought on by roughness and flow velocity around riparian trees and wood are critical in developing island dominated, anabranching rivers (Gurnell and Petts, 2006; Tooth and Nanson, 2000; Tooth and Nanson, 1999). Often, an initial deposit is formed 53 downstream of a tree or other roughness feature. Vegetation then begins to colonize this feature and additional deposition occurs enabling the formation of a ridge or island (Tooth and Nanson, 2000). Vegetation also has the ability to narrow stream channels. After a large flood, Plum Creek, Colorado, experienced significant widening. Multiple years of low flow following a major flood allowed vegetation to establish along stream banks creating stable, erosion resistant channels (Friedman et al., 1996). Using flume experiments, Tal et al. (2004) found that vegetation decreases channel width, braiding index and mobility. Despite these studies, when compared to woody riparian species, flexible channel bank and island vegetation and their effect on channel morphology and processes have been largely neglected. In particular, the geomorphic effect of an invasive species is important because they tend to dominate the landscape developing monocultures that outcompete native vegetation. One such species is P. arundinacea along the Sprague River, Oregon. 3. Methods 3.1. MD-SWMS (Multi-Dimensional Surface-Water Modeling System) MD-SWMS is a two dimensional depth-averaged hydraulic model developed by the USGS to model surface water hydraulics (McDonald et al., 2005). A two dimensional model was chosen because it can produce downstream velocity, depth, and shear stress at each point within a grid or mesh throughout the channel. In addition, two dimensional models are ideal for noting changes in water depth or velocity in and around two dimensional features like bars and islands (Nelson et al., 2003). I used Morpho 2D 54 within MD-SWMS because Morpho-2D accounts for vegetation characteristics, a variable that few models incorporate into their calculations. Vegetation can be represented in the model either as vegetation height and density inputs, or within the roughness values. When modeling with roughness inputs, the model produces output on bed shear stress in addition to water depth and velocity. Morpho-2D mimics the interaction between mixed grain sediment and vegetation (Nelson and McDonald, 2010). It calculates two-dimensional horizontal velocity and flow depth for grid nodes (built via a mesh) defined by the user (Nelson et al., 2003). The model does not alter bed morphology. The Klamath Tribes Research Station gathered Lidar and bathymetric topography data at a 0.5m resolution for the site of interest (Figure 4.1a.) along the Sprague River in the spring of 2005. For the purposes of modeling smaller portions of the channel, one straight reach was chosen (Figure 4.1.). This straight reach was chosen based on its close proximity to a gaging station and because it was located in an area with minimal interference with adjacent meanders. In addition, this reach contains stands of all three species of interest. For input into MD-SWMS, the topography data in the form of a DEM was converted to point data and imported into MD-SWMS. A mesh grid was created within MD-SWMS by digitizing a channel centerline and establishing grid parameters (Table 4.1.). Topography was mapped to the grid to create initial conditions upon which the model could operate. 55 Figure 4.1. Sprague River Basin and associated study site: (a) Basin and sample reach, (b) Lidar and bathymetry showing channel topography, (c) current vegetation patches mapped in the field, (d) future scenario vegetation patches. Arrow indicates flow direction. N a b c d 56 Polygons representing current vegetation condition in the reach were developed by mapping stands on the channel banks of 2 m 2 or more in area in the field using aerial photographs (Figure 4.1.). This mapping only encompassed the banks and did not extent onto the floodplain. Then I applied the measured vegetation height, density and roughness (Chapter 3) to the polygons for use in the model. Roughness values were calculated by first establishing a base roughness value for the channel without vegetation using the additive procedure outline in Arcement and Schneider (1989) and others. A roughness value for vegetation was then added to this value using the nveg equation in Rahmeyer et. al. (1999), which incorporates vegetation density, frontal area and stem flexibility (Table 4.3., Martinez, 2013). In addition, three future scenarios of vegetation occupation were created to simulate channel changes when channel banks were dominated completely by one of the three species (Figure 4.1d.). In these scenarios, vegetation polygons that extended approximately ¼ of the way into and out of the channel banks from the bank-full line were developed. In areas outside of the mapped patches and outside of the channel, a medium roughness was used along with a vegetation density and height for basic turf grass (Other value in Table 4.3.; Maurer et al., 2001). For the future scenarios, results for P. arundinacea were subtracted from each of the native species to determine differences between the invasive and each native species for each of the outputs: depth, velocity, and bed shear stress. The model was run at two different discharges, 10yr and 25yr recurrence intervals to capture circumstances under which the vegetation would be inundated. In addition, the model was run for current conditions and future conditions for each of the three species of interest: P. arundinacea, E. palustris, and C. vesicaria. Discharge data were obtained 57 from the closest gaging station to the site (Table 4.2.; Station 11497500; OWRD, 2013). This set of runs was carried out once to determine outputs (velocity and depth) with regard to each species’ roughness (section 4.1. below), and then again with regard to each species height and density (section 4.2. below), for a total of 16 runs. In addition, median sediment size from a nearby location was used as input into the model: d50= 1mm (T. LaGreca, Klamath Tribes Research Station, personal communication). Table 4.1. Model grid parameters Reach Size Grid Cell Spacing # Grid Cells Length (m) Width (m) Stream-wise (m) Stream- normal (m) Straight Reach 520 180 3 2 2400 Table 4.2. Discharge and associated boundary conditions. Recurrence Interval Discharge (cms) Downstream Channel Bottom Elevation (m) Stage (m) 10 yr 89.34 1313 2.9 25yr 113.00 1313 3.2 50yr 191.26 1313 3.85 Table 4.3. Roughness, vegetation height and density for each species. 3.2. Modeling with Roughness To model changes in hydraulic conditions due to the effects of each species , I varied roughness (Manning’s n), according to changing species cover. The model was run using current condition vegetation stands and future condition vegetation stands with the respective roughness for each of the three species of interest. Species Mannings n Vegetation Density (m2/m2) Vegetation Height (m) P. arundinacea 0.018328 0.04 0.92 E. palustris 0.018431 0.13 0.54 C. vesicaria 0.018135 0.03 0.74 Other 0.018300 0.90 0.30 58 3.3. Modeling with Vegetation Density and Height Vegetation height and density values for each species were obtained through previous field work (Martinez, 2013). Vegetation height is the average height (meters) for each plant and vegetation density is the percentage of area occupied by stems. For areas outside of the patches mapped in the field (Fig. 1c), height and density of turf grass was used (Maurer et al., 2001). As above, the model was run for current conditions and for future conditions for each species, at each of the two recurrence intervals. These outputs were differenced to obtain the marginal effect of the invasive relative to each native species. Validation data for 10- and 25-year flows were not available, so a validation could not be carried out. However, comparison among species and not validation of specific hydraulic conditions is the goal for this study. 4. Results Overall, MD-SWMS provided a useful platform upon which to model vegetation characteristic changes on hydraulic conditions. Given that the model is two dimensional, it provided the ability to model along a reach with various features and changes in channel width and depth. Results for depth and velocity were generally consistent and, for the most part, the model followed the general rules of at-a-station hydraulic relationships. However, some limits were reached when modeling high flows. At the 50 year recurrence interval, a ponding effect occurred where, despite increases in depth, and therefore discharge, velocity decreased. This problem with Morpho-2D has been previously noted (R. McDonald, USGS, personal communication). In addition, edges 59 were not adequately represented within the model. This seemed to be limited to 50 - 75m from the upstream and downstream edge of the modeled reach. Despite this, the middle portion of the reach followed the expected hydraulic relationships well, the results of which will be the bulk of the discussion here. 4.1. Current Conditions 4.1.1. Modeling based on vegetation density and height In the 10-year flow, velocities within vegetation patches were on the order of 0.6 m/s to 0.7m/s (Figure 4.2 a2.). High velocities on the order of 0.8m/s and 1.2 m/s were exhibited towards the downstream end of the channel within P. arundinacea and C. vesicaria patches, respectively, though these anomalously high results were likely due to edge effects. In the 25-year flow, velocities in the invasive were as high as 1.07 m/s within the center of the channel (Figure 4.2 b2.). 4.1.2. Modeling based on vegetation roughness Water depth values were on the order of 0.1 to 0.2 m lower than those depths obtained from the vegetation density and height runs (Figure 4.3 a1 and b1). A similar range in velocity was obtained in the roughness runs as those observed in the vegetation height and density runs. However, some limited areas experienced higher velocity values in the roughness runs. These were present within the center of the channel and away from vegetation patches (4.3 b1 and b2). In addition to velocity and depth outputs, the roughness runs allowed for bed shear stress mapping. Similar shear stresses are exhibited in the 10 and 25-year flows with the exception of slight increases in high shear stress 60 Figure 4.2. Current conditions modeled given vegetation height and density under 10 year (a1 and a2), and 25-year recurrence interval flows (b1 and b2). a1. b1. b2. a2. 61 Figure 4.3. Current conditions modeled given vegetation roughness under 10-year recurrence interval (a1, b1, c1) and 25-year recurrence interval (a2, b2, c2). a1. a2. b1. b2. c1. c2. 62 areas in the 25-year flow (4.3 c1 and c2). These were concentrated toward the center of the channel and ranged from 3.76 to 4.24 N/m 2 . 4.2. Future Conditions 4.2.1. Modeling based on vegetation density and height Hereafter, I will mainly focus on how flows within vegetation patches differ with regard to species. Velocity values in the 10-year flow were higher within P. arundinacea stands than E. palustris (Figure 4.4 b1.) by up to 0.09m/s. This value was large considering velocities along these areas at the channel margins were fairly low (0.6- 1m/s). Velocity under the invasive and C. vesicaria stands were similar (Fig. 4.4 b2). In the 25-year flow, velocity results were similar between the invasive and natives. Regarding water depth within the vegetation patches, P. arundinacea and E. palustris performed similarly in the 10-year flow (Figure 4.4 a1.). On the other hand, flow depths under C. vesicaria were slightly shallower (0.007 to 0.01m shallower) than when the invasive was modeled (Figure 4.4 a2.). This was most pronounced at the upstream portion of the reach, which may be anomalous due to edge effects, but continued through the center of the channel for much of the reach. When compared to C. vesicaria, the invasive affected the overall hydraulics of flow through the channel, but this is not evident when the invasive is compared to E. palustris. In the 25-year flow, results showed that the flow depth within the invasive was shallower than either of the two natives (Figure 4.5 a1. and a2.). Some differences existed between the invasive and C. vesicaria at the upstream end of the reach, but this may be due to edge effects. There is little difference in velocity between the invasive and either of the native species. 63 Figure 4.4. Future conditions modeled given vegetation height and density under 10-year recurrence interval flow: P. arundinacea – E. palustris (a1 and b1), P. arundinacea – C. vesicaria (a2 and b2). 4.2.2. Modeling based on vegetation roughness In the 10-year flow, velocity results showed generally faster values for the invasive than the two native species, though these results were patchy and not banded as they were in the vegetation height and density runs (Figure 4.6 b1. and b2.). Bed shear stresses in were higher in E. palustris stands than the invasive and were lower within C. a1. a2. b1. b2. 64 Figure 4.5. Future conditions modeled given vegetation height and density under 25- year recurrence interval flow: P. arundinacea – E. palustris (a1, and b1), P. arundinacea – C. vesicaria (a2 and b2). vesicaria stands than the invasive (Figure 4.6 c1. and c2.). Differences in depth were confined to edges and backwater areas (Figure 4.6 a1. and a2.). In the 25-year flow, the effect of the vegetation was less pronounced and depth and velocity results were similar between the invasive and natives (Figure 4.7 a1., a2., b1. and b2.). Bed shear stress was highest in E. palustris stands, intermediate in P. arundinacea, and lowest in C. vesicaria. a1. a2. b1. b2. 65 Figure 4.6. Future conditions modeled given roughness under 10-year recurrence interval flow: P. arundinacea – E. palustris (a1, b1, c1), P. arundinacea – C. vesicaria (a2, b2, c2). a1. a2. b1. b2. c1. c2. 66 Figure 4.7. Future conditions modeled given roughness under 25-year recurrence interval flow: P. arundinacea – E. palustris (a1, b1, c1), P. arundinacea – C. vesicaria (a2, b2, c2). a1. a2. b1. b2. c1. c2. 67 5. Discussion 5.1. Model Performance In general, the model produced realistic results regarding water depth, velocity and bed shear stress. Similarly realistic results were also produced within the vegetation patches. For example, the 10-year recurrence interval flow results agree, showing a decrease in bed shear stress (Figure 4.4.) and an increase in velocity (Figure 4.3.) with a switch in species composition from E. palustris to the invasive given roughness conditions (Figure 4.2.) and vegetation height and density (Figure 4.3.), respectively. The magnitude of the results in the vegetation runs were comparable to results obtained in the future condition model runs. According to sampling via ADV (Acoustic Doppler Velocimeter, M. Dawson, Klamath tribes Research Station, personal communication), velocities along the Sprague at these flows are similar to those modeled. In addition, shear stress results are consistent with what would be expected in this environment. In the shift from the 10-year to 25-year flow, an increase in discharge was accommodated by increases in both water depth and velocity, though depth increased more than velocity. The most realistic results were likely those run with the roughness calculations because these roughness values incorporate vegetation characteristics such as vegetation height and density as well as general channel characteristics. The roughness runs also provide the user with an additional output, bed shear stress, which can be useful when studying a particular system. However, anomalies were evident when comparing the vegetation height and density runs to the roughness runs. For example, vegetation height caused a significant difference in water depth when comparing the invasive and C. vesicaria in the 10-year flow. However, this difference was not present in the roughness 68 runs. Similarly, vegetation density altered velocity more significantly within E. palustris stands compared to P. arundinacea but this difference was not apparent in roughness runs. 5.2. Interpretation of Results 5.2.1. Modeling based on vegetation height and density When modeling hydraulic conditions under 10-year flows, water depths within C. vesicaria stands were lower than those found in the invasive. This was likely due to the invasive’s greater height which enhanced water flow interruption compared to the shorter C. vesicaria. Thus, the invasive caused the water to adjust vertically, creating higher water depths when it was occupying the channel. In this case, depths modeled with the invasive were on average 0.2m higher than those in the native. Both species were inundated within the main channel boundaries and less inundated with increasing distance from the thalweg and were likely to be partially bent down by the flow, though the model has no way to account for stem flexibility. Velocities were not significantly different. When comparing the invasive and E. palustris stands at the 10-year flow, the native experienced higher velocities, due to its lower density, but little to no difference in water depth. The model likely adjusted water depth given differences in vegetation height, thus accounting for higher vegetation stands and their interruption to the water flow by extending the water surface vertically. Velocity, which is primarily affected by the density of stems and plants within a stand, decreased more in regions where density was higher. I saw greater differences between the invasive and E. palustris stands than 69 between the invasive and C. vesicaria stands because vegetation density is more influential in the model than plant height. According to Gran and Paola (2001), the degree to which vegetation influences flow velocities is dependent on the density of plants and stems. Results from this model are consistent with this idea. In the future, a decrease in stem density resulting from a shift from E. palustris to the invasive may mean a decrease in flow interruption that will likely result in increases in velocity. Such changes in velocity could lead to decreased deposition and enhanced erosion. 5.2.2. Modeling based on vegetation roughness P. arundinacea had higher shear stress than C. vesicaria with differences on the order of 0.10 N/m 2 . This was due to a higher roughness within P. arundinacea stands. Differences in bed shear stress when comparing the invasive and E. palustris were small, at a maximum, 0.02 N/m 2 different. Therefore, according to changes in shear stress, if the invasive outcompetes E. palustris, shear stresses are likely to remain similar or decrease only slightly and geomorphic processes will continue to operate as they are currently. However, if the invasive outcompetes C. vesicaria, shear stress will increase, leading to lower velocities and possibly deposition within the invasive stand. Deposition could form higher banks that are less resistant to erosion, which may translate to increased channel bed erosion creating deeper, narrower channels in areas where C. vesicaria has been overtaken by the invasive. In the 25-year flow, differences in bed shear stress still existed but were less marked which was likely due to a dampening effect as discharge increased; as discharge and water depth increased, plants became fully submerged in the water column and had 70 less of an impact on velocity and shear stress within the channel (Wu et al., 1999). In effect, vegetation was felt less as flow conditions overcame the height of the vegetation patch. Depth was approximately 0.3m deeper in the 25-year flow which was fairly consistent throughout the channel. Velocity showed more scatter in difference and was anywhere from 5 to 30% greater in the 25-year flow. Finally, some attention should be placed on the location, rather than just the mere presence of particular species given that the invasive occupies specific areas along the channel banks (Figure 4.8.). This could be important given that the channel experiences higher velocities closer to the thalweg. If the invasive continues to occupy areas previously occupied by natives, it will begin encroaching further into the channel and decreasing velocity in these areas due to its higher roughness compared to C. vesicaria. Furthermore, the dominant native in the upper reaches of the Sprague (near the study site, Figure 4.1a) is C. vesicaria, while E. palustris is the primary dominant native in lower reaches of the river. In addition, previous work regarding the bank cohesion provided by roots for these three species has shown that P. arundinacea and C. vesicaria have similarly high cohesive properties. However, E. palustris has much lower cohesion abilities. Therefore, in areas where the invasive is outcompeting E. palustris, it is increasing bank stability and resisting erosion by fluid forces. This could mean more stable surfaces in these areas. Further work must be conducted to fully understand these processes. 71 Figure 4.8. P. arundinacea, C. vesicaria and E. palustris occupying typical channel locations (left bank), and conditions if the invasive outcompetes both of the natives (right bank). 6. Conclusion Differences in vegetation characteristics such as density, height, and roughness, can cause changes in stream channel velocity and bed shear stress that may lead to changes in channel morphology as a result of invasive species occupation. Along the Sprague River, Oregon, P. arundinacea is outcompeting native species E. palustris and C. vesicaria. Previous work regarding the effect of vegetation on channel processes has primarily considered woody vegetation. However, dominant, herbaceous, flexible vegetation, in low gradient streams such as the Sprague can influence roughness and therefore channel processes. In systems dominated by flexible vegetation, species composition changes from native plants to the invasive P. arundinacea, can affect channel morphology in the future. With a shift from native E. palustris to P. arundinacea a decrease in vegetation density will increase water velocities. A shift from C. vesicaria to the invasive will increase roughness. Based on the results, this shift will manifest itself as an increase in bed shear stress. These vegetation characteristics and hydraulic changes will cause changes in channel morphology that will affect habitat and therefore, must be taken into account as the invasive continues to expand in this environment. 10 year flow 25 year flow 72 CHAPTER V SUMMARY In this dissertation, I examined the above ground and below ground influences of native and riparian herbaceous vegetation on the geomorphology of the Sprague River, Oregon. Previous research of this nature has focused on large, woody species such as Tamarix. This research is one of the first studies to investigate the influence of a small, but highly invasive vegetation. Despite its similarity to the native species E. palustris and C. vesicaria, the invasive P. arundinacea has the potential to influence erosional and depositional processes differently than the native species. This could be causing an alteration in geomorphic processes where the invasive is outcompeting the natives. In Chapter II, I evaluated the below ground effects of the invasive and native species using traditional root distribution surveys, root tensile strength measurements, and the model RipRoot to estimate added bank cohesion due to roots. All three species in this study had smaller root diameters than comparable species in previous research. Root depths in this study, however, were comparable to other studies that have examined herbaceous vegetation. Root Area Ratio values showed general trends that have been observed in previous work as well. Among the three species, the native C. vesicaria exhibited the largest root sizes and largest RAR. C. vesicaria had the highest tensile strength associated with its roots. This was followed by P. arundinacea and E. palustris. Cohesion, which is highly influenced by root size distribution, was highest for C. vesicaria and was followed by P. arundinacea and E. palustris. The additional cohesion provided by the invasive when compared to E. palustris means that roots are enhancing 73 bank stability in areas where the invasive has outcompeted the native. This could mean a decrease in lateral channel migration, a decrease in the likelihood of channel widening, and may lead to increased shear stress on the channel bed that will result in incision and increased stream depth. This may enhance salmon habitat since water temperature, which is highly influenced by channel depth, is of concern. The above ground influences of the invasive and native species were addressed in Chapter III. In the field, I measured stem density, stem stiffness, plant density, height, and frontal area for input into an equation to determine roughness provided by the invasive and native species. Stem density was significantly higher in stands of the native E. palustris and was followed by C. vesicaria and P. arundinacea. Measured stem stiffness was highest for the invasive. The native E. palustris had the highest roughness which was likely due to its high stem density values. This was followed by the invasive which had high stem stiffness values. C. vesicaria had the lowest vegetation roughness value and was significantly different than the other two species. The roughness values obtained via this method were slightly smaller but still comparable to other roughness values calculated for similar vegetation. To obtain a complete channel roughness, I used the additive approach to determine base channel roughness and then added the calculated vegetation roughness. Deposition was also directly measured within the plant stands and correlates with calculated roughness values. Given these roughness values and observed deposition, these results would mean aggrading banks on depositional surfaces on the order of 1cm per year for the invasive and E. palustris. Over time, this may cause narrowing of the channel in areas where the invasive has outcompeted C. vesicaria. 74 In Chapter IV, I utilize the data gathered in Chapter III to model channel depth, water velocity and bed shear stress along two reaches of the Sprague River to simulate changes over time due to vegetation colonization. I modeled water depth and velocity by first occupying the channel with vegetation heights and densities currently present. Another set of runs was completed under future conditions involving complete colonization by each of the three species along the banks to determine changes in depth and velocity if one species completely colonizes the channel banks. I also preformed separate runs in which I varied roughness for the current and future conditions. E. palustris, given its high vegetation density and roughness, had the greatest ability to alter stream velocity and bed shear stress. Therefore, if the invasive continues to occupy and overtake areas previously inhabited by E. palustris, increases in velocity and decreases in bed shear stress will result. Overall, I have found that the invasive differs significantly in both above ground and below ground characteristics when compared to native species. Therefore, its geomorphic effects over time could be significant. These include a strengthening of channel banks against erosion when the invasive occupies areas previously occupied by E. palustris. In addition, channel roughness, and therefore likely deposition, will increase in areas where the invasive is outcompeting C. vesicaria. Finally, I showed via modeling that differences in vegetation density and roughness cause changes in water velocity and bed shear stress. With the number of restoration projects increasing throughout the United States and elsewhere, an understanding of the role invasive species play in our stream systems is important. Human modifications have changed streams so drastically that returning them 75 to their historical conditions is not possible. The introduction of invasive species is one such modification. In ecosystems like the Sprague River, it is no longer feasible to completely remove the invasive vegetation. Instead, its past and future effects on the channel must be understood to incorporate this knowledge into restoration planning. Only then can sustainable restoration practices have the potential to be successful. 76 REFERENCES CITED Chapter I Graf WL. (1978) Fluvial adjustments to the spread of tamarisk in the Colorado Plateau region. Geological Society of America Bulletin 89. Pollen N, Simon A (2005) Estimating the mechanical effects of riparian vegetation on stream bank stability using a fiber bundle model. Water Resource Research 41 (1- 11) Wu TH, McKinnel WP, Swanston DN (1979) Strength of tree roots and landslides on Prince of Wales Island, Alaska. Canadian Geotechnical Journal 16 (19-33) Chapter II Abernathy B, Rutherford ID (2001) The distribution and strength of riparian tree roots in relation to riverbank reinforcement. Hydrological Processes 15:63-79 Anderson DE (1961) Taxonomy and distribution of the genus Phalaris. Iowas State Journal of Science 36:1-96 Antieau CJ (2004) Biology and management of reed canarygrass, and implications for ecological restoration. Department of Transportation Seattle, Washington Apfelbaum SI, Sams CE (1987) Ecology and control of reed canarygrass. Natural Areas Journal 69-74 (2) Bennett SJ, Wu W, Alonso CV, Wang SSY (2007) Modeling fluvial response to in- stream woody vegetation: implications for stream corridor restoration. Earth Surface Processes and Landforms 33:890-909 Beschta RL, Bilby RE, Brown GW, Holtby LB, Hofstra TD (1987) Stream temperature and aquatic habitat: fisheries and forestry applications. In: Salo EO, Cundy TW (eds) Streamside managment: forestry and fishery interactions. University of Washington, Seattle, Washington, Bethalahmy N (1962) First year effects of timber removal on soil moisture. International Association of Scientific Hydrology 7 (2):34-38 Bischetti GB, Chiaradia EA, Simonato T, Speziali B, Vitali B, Vullo P, Zocco A (2005) Root strength and root area ratio of forest species in Lombardy (Northern Italy). Plant Soil 278:11-22 77 Bishop DM, Stevens ME (1964) Landslides on Logged Areas in Southeast Alaska. Research Paper NOR-1. Juneau Bjornn TC, Reiser DW (1991) Habitat Requirments for Salmonids in Streams. Influences of forest and rangeland management on salmonid fishes and their habitats 19:83- 138 Bohm W (1979) Methods of Studying Root Systems, vol 33. Springer-Verlag, Berlin Boyd M, Kasper B (2002) Upper Klamath Lake Drainage Stream Temperature Analysis: Attachment 1 Upper Klamath Lake Drainage TMDL. Upper Klamath Basing Temperature TMDL. Department of Environmental Quality, Portland, Oregon Briles CE, Whitlock C, Bartlein PJ, Higuera P (2008) Regional and local controls on postglacial vegetation and fire in the Siskiyou Mountains, northern California, USA. Palaeogeography, Palaeoclimatology, Palaeoecology 265:159-169 Burroughs ER, Thomas BR (1977) Declining root strength in Douglas fir after felling as a factor in slope stabiity. Res. Pap. INT-190. Intermtn. For. and Range Exp. Stn., Ogden, Utah Corenblit D, Tabacchi E, Steiger J, Gurnell AM (2007) Reciprocal interactions and adjustments between fluvial landforms and vegetation dynamics in river corridors: A review of complementary approaches. Earth-Science Reviews 84:56-86 Cowell CM, Dyer JM (2002) Vegetation Development in a Modified Riparian Environment: Human Imprints on an Allegheny River Wilderness. Annals of the Association of American Geographers 92 (2):189-202 De Baets S, Poesen J, Knapen A, Barbera GG, Navarro JA (2007) Root characteristics of representative Mediterranean plant species and their erosion-reducing potential during concentrated runoff. Plant Soil 294:169-183 De Baets S, Poesen J, Reubens B, Wemans K, De Baerdemaeker J, Muys B (2008) Root tensile strength and root distribution of typical Mediterranean plant species and their contribution to soil shear strength. Plant Soil 305:207-226 Duff DA Livestock grazing impacts on aquatic habitat in Big Creek, Utah. In: Proceedings of the workshop on wildlife-fisheries relationships in the Great Basin., Berkeley, CA, 1977. University of California Agriculture Station, pp 129- 142 Fierke M, Kauffman J (2006) Riverscape-level patterns of riparian plant diversity along a successional gradient, Willamette river, Oregon. Plant Ecology 185 (1):85-95. doi:10.1007/s11258-005-9086-z 78 Foster RD, Wetzel PR (2005) Invading Montoypic Stands of Phalaris arundinacea: A Test of Fire, Herbicide, and Woody and Herbaceous Native Plant Groups. Restoration Ecology 13 (2):318-324 Friedrichsen PT (1997) Summertime stream temperatures in the North and South Forks of the Sprague River, South Central Oregon. Oregon State University, Corvallis, Oregon Galatowitsch SM, Anderson NO, Ascher PD (1999) Invasiveness in wetland plants in temperate North America. Wetlands 19 (4):733-755 Graf WL (1978) Fluvial adjustments to the spread of tamarisk in the Colorado Plateau region. Geological Society of America Bulletin 89 (1491-1501) Gran K, Paola C (2001) Riparian vegetation controls on braided stream dynamics. Water Resource Research 37 (12):3275-3283 Gray DH (1974) Reinforcement and stabilization of soil by vegetation. Journal of Geotechnical Engineering ASCE 100 (GT6):695-699 Gray DH, Barker D (2004) Root-soil mechanics and interactions. In: Bennett JJ, Simon A (eds) Riparian vegetation and fluvial geomorphology. American Geophysical Union, New York, pp 113-123 Gray DH, Leiser A (1982) Biotechnical Slope Protection and Erosion Control. Van Nostrand Reinhold, New York Greenway DR (1987) Vegetation and Slope Stability. In: Anderson MG, Richards KS (eds) Slope Stability. John Wiley & Sons Ltd., New York, Greenway DR, Anderson MG, Brian-Boys KC (1984) Influence of vegetation on slope stability in Hong Kong. Paper presented at the 4th International Symposium on Landslides, Toronto, Ont, Guard BJ (1995) Wetland Plants of Oregon and Washington. Lone Pine Publishing, Auburn, Washington Hay LE, McCabe GJ, Clark MP, Risley JC (2009) Reducing streamflow forecast uncertainty: Application and qualitative assessment of the Upper Klamath River Basin, Oregon. Journal of American Water Resources Association 45 (3):580-596 Hickin EJ (1984) Vegetation and river channel dynamics. Canadian Geographer 28 (2):111-126 Kaul R (1965) The influence of roots on certain mechanical properties of an uncompacted soil. University of North Carolina, Raleigh 79 Kercher SM, Zedler JB (2004) Multiple disturbances accelerate invasion of reed canary grass (Phalaris arundinacea L.) in a mesocosm study. Oecologia 138:455-464 Lavoie C, Dufresne C, Delisle F (2005) The spread of reed canarygrass (Phalaris arundinacea) in Quebec: A spatio-temporal perspective. Ecoscience 12 (3):366- 375 Loy WG, Allan S, Buckley AR, Meacham JE, Andreas LJ, Martin GE, West R (2001) Atlas of Oregon. University of Oregon Press, Eugene Marten GC, Heath ME (1973) Reed Canarygrass. In: Heath ME, Metcalfe DS, Barnes RF (eds) Forages: The Science of Grassland Agriculture. The Iowa State University Press, Ames, Iowa, Matthews KR (1996) Diel movement and habitat use of California golden trout in the Golden Trout Wilderness. Transactions of the American Fisheries Society 125:78- 86 Murray AB, Knaapen MAF, Tal M, Kirwan ML (2008) Biomorphodynamics: Physical- biological feedbacks that shape landscapes. Water Resources Research 44:1-18 Myers TJ, Swanson S (1992) Variation of stream stability with stream type and livestock bank damage in northern Nevada. Water Resources Bulletin 28:743-754 Myers TJ, Swanson S (1996) Long-term aquatic habitat restoration: Mahogany Creek, Nevada, as a case study. Journal of the American Water Resources Association 32 (241-252) Naglich FG (1994) Reed Canarygrass (Phalaris arundinacea L.) in the Pacific Northwest: Growth Parameters, Economic Uses, and Control. The Evergreen State College, Olympia, Washington Nanson GC, Knighton AD (1996) Anabranching Rivers: Their cause, character and classification. Earth Surface Processes and Landforms 21:217-239 O'Connor JE, McDowell PF, Lind P, Rasmussen CG (forthcoming 2013) Geomorphology, channel and flood-plain processes, and flood-plain vegetation of the Sprague and Lower Sycan Rivers, Klamath Basin, Oregon. US Geological Survey Scientific Investigations Report. USGS, Operstein V, Frydman S (2000) The influence of vegetation on soil strength. Ground Improvement 4:81-89 80 Overton CK, Chandler GL, Pisano JA (1994) Northern/Intermountain regions; fish habitat inventory: grazed, rested, and ungrazed reference stream reaches, Silver King Creek, California. Forest Service General Technical Report vol INT-GTR- 311. USDA, Perucca E, Camporeale C, Ridolfi L (2007) Significance of the riparian vegetation dynamics on meandering river morphodynamics. Water Resource Research 43:1- 10 Pojar J, MacKinnon A (1994) Plants of the Pacific Northwest Coast. Lone Pine Publishing, Vancouver, BC Pollen-Bankhead N, Simon A (2009) Enhanced application of root-reinforcement algorithms for bank stability modeling. Earth Surface Processes and Landforms 34:471-480 Pollen N, Simon A (2005) Estimating the mechanical effects of riparian vegetation on stream bank stability using a fiber bundle model. Water Resource Research 41 (1- 11) Pollen N, Simon A, Collison A (2004) Advances in Assessing the Mechanical and Hydrologic Effects of Riparian Vegetation on Streambank Stability. In: Bennett SJ, Simon A (eds) Riparian Vegetation and Fluvial Geomorphology, vol WaterScience and Application 8. American Geophysical Union, Washington DC, Rasmussen CG (2011) Geomorphology, hydrology and biology of floodplain vegetation in the Sprague Basin, OR: History and potential for natural recovery. University of Oregon, Risley JC, Gannett MW, Lea JK, Roehl EA (2005) An Analysis of Statistical Methods for Seasonal Flow Forecasting in the Upper Klamath River Basin of Oregon and California. USGS and US Department of the Interior 2005-5177 Sher AA, Marshall DL, Gilbert SA (2000) Competition between Native Populus deltoides and Invasive Tamarix ramosissima and the Implications for Reestablishing Flooding Disturbance. Conservation Biology 14 (6):1744-1754 Sher AA, Marshall DL, Taylor JP (2002) Establishment patterns of native Populus and Salix in the presence of invasive nonative Tamarix. Ecological Applications 12 (3):760-772 Simon A, Collison AJC (2002) Quantifying the mechanical and hydrologic effects of riparian vegetation on streambank stability. Earth Surface Processes and Landforms 27:527-546 81 Simon A, Pollen N A Model of Streambank Stability Incorporating Hydraulic Erosion and the Effects of Riparian Vegetation. In: Federal Interagency Sedimentation Conference, Reno, NV, 2006. JFIC, Simon A, Pollen N, Langendoen E (2006) Influence of two woody riparian species on critical conditions for streambank stability: Upper Truckee River, California. Journal of the American Water Resources Association 42 (1):99-113 Smith DG (1976) Effect of vegetation on lateral migration of anastamosed channels of glacier meltwater river. Geological Society of America Bulletin 87:857-860. doi:doi:10.1130/0016-7606(1976)87<857:EOVOLM>2.0.CO Tal M, Gran K, Murray AB, Paola C, Hicks DM (2004) Riparian Vegetation as a Primary Control on Channel Characterisitcs in Multi-thread Rivers. In: Bennett SJ, Simon A (eds) Riparian Vegetation and Fluvial Geomorphology. Tal M, Paola C (2007) Dynamic single-thread channels maintained by the interaction of flow and vegetation. Geology 35:347-350 Tickner DP, Angold PG, Gurnell AM, Mountford JO (2001) Riparian plant invasions: hydrogeomorphological control and ecological impacts. Progress in Physical Geography 21 (1):22-52 Tooth S, Nanson GC (1999) Anabranching rivers on the Northern Plains of arid central Australia. Geomorphology 29:211-233 Tourbier JT, Westmacott R (1981) Water resources protection technology. Urban Land Institute, Washington D.C. Waldron L (1977) The shear resistance of root-permeated homogeoneous and stratified soil. Journal of the Soil Science Society of America 41 (843-849) White RJ, Brynildson OM (1967) Guideline for management of trout habitat in Wisconsin. Wisconsin Department of Natural Resources Technical Bulletin 39:65 Wilson BL, Brinerd RE, Lytjen D, Newhouse B, Otting N (2008) Field Guide to Sedges of the Pacific Northwest. Oregon State University Press, Corvallis Wu TH (1976) Investigations of landslides on Prince of Wales Island. Geotechnical Engineering Report 5. Ohio State University, Columbus Wu TH, McKinnel WP, Swanston DN (1979) Strength of tree roots and landslides on Prince of Wales Island, Alaska. Canadian Geotechnical Journal 16 (19-33) 82 Wynn T, Mostaghimi S, Burger JA, Harpold AA, Henderson MB, Henry L-A (2004) Ecosystem Restoration: Variation in Root Density along Stream Banks. Journal of Environmental Quality 33:2030-2039 Chapter III Aldridge BN and Garrett JM. (1973) Roughness coefficients for stream channels in Arizona. 87p. Anderson DE. (1961) Taxonomy and distribution of the genus Phalaris. Iowas State Journal of Science 36: 1-96. Antieau CJ. (2004) Biology and management of reed canarygrass, and implications for ecological restoration, Seattle, Washington: Department of Transportation Apfelbaum SI and Sams CE. (1987) Ecology and control of reed canarygrass. Natural Areas Journal 69-74. Arcement GJ and Schneider VR. (1989) Guide for selecting manning's roughness coefficients for natural channels and floodplains. United States Geological Survey Water-supply Paper 2339. Asaeda T, Rajapakse L and Kanoh M. (2010) Fine sediment retention as affected by annual shoot collapse: Sparganium Erectum as an ecosystem engineer in a lowland stream. River Research and Applications 26: 1153-1169. Barnes HH. (1967) Roughness characteristics of natural channels. U.S. Geological Survey Water-Supply Paper 1849. Boyd M and Kasper B. (2002) Upper Klamath Lake Drainage Stream Temperature Analysis: Attachment 1 Upper Klamath Lake Drainage TMDL. In: Quality ODoE (ed) Upper Klamath Basing Temperature TMDL. Portland, Oregon: Department of Environmental Quality. Chow VT. (1959) Open Channel Hydraulics, New York: McGraw-Hill. Clarke A, Lake PS and O'Dowd DJ. (2004) Ecological impacts on aquatic macroinvertebrates following upland stream invasion by a ponded pasture grass (Glyceria maxima) in southern Australia. Marine and Freshwater Research 55: 709-713. Coon WF. (1998) Estimation of Roughness Coefficients for Natural Stream Channels with Vegetated Banks. USGS Water-Supply Paper 2441. 83 Copeland RR. (2000) Determination of Flow Resistance Coefficients Due to Shrubs and Woody Vegetation. ERDC/CHL CHETN-VIII-3. Corenblit D, Gurnell AM, Steiger J, et al. (2008) Reciprocal adjustments between landforms and living organisms: Extended geomorphic evolutionary insights. Catena 73: 261-273. Cowan WL. (1956) Estimating hydraulic roughness coefficients. Agricultural Engineering 37: 473-475. Fierke M and Kauffman J. (2006) Riverscape-level patterns of riparian plant diversity along a successional gradient, Willamette river, Oregon. Plant Ecology 185: 85- 95. Foster RD and Wetzel PR. (2005) Invading Montoypic Stands of Phalaris arundinacea: A Test of Fire, Herbicide, and Woody and Herbaceous Native Plant Groups. Restoration Ecology 13: 318-324. Freeman GE, Copeland RR and Rahmeyer WJ. (1998) Field determination of Manning's n value for shrubs and woody vegetation. USCID Conference. Park City, UT. Freeman GE, Rahmeyer WH and Copeland RR. (2000) Determination of resistance due to shrubs and woody vegetation. U.S. Army Corps of Engineers ERDC/CHL TR- 00-25. Friedrichsen PT. (1997) Summertime stream temperatures in the North and South Forks of the Sprague River, South Central Oregon. Corvallis, Oregon: Oregon State University, 147. Galatowitsch SM, Anderson NO and Ascher PD. (1999) Invasiveness in wetland plants in temperate North America. Wetlands 19: 733-755. Gilley JE and Finkner SC. (1991) Hydraulic roughness coefficients as affected by random roughness. Transactions of the ASAE 34: 897-903. Gleason ML, Elmer DA, Pien NC, et al. (1979) Effects of stem density upon sediment retention by salt marsh cord grass, Spartina alterniflora Loisel. Estuaries 2: 271- 273. Gordon ND, McMahon TA, Finlayson BL, et al. (2004) Stream Hydrology: An Introduction for Ecologists, West Sussex, England: Wiley. Graf WL. (1978) Fluvial adjustments to the spread of tamarisk in the Colorado Plateau region. Geological Society of America Bulletin 89. Gran K and Paola C. (2001) Riparian vegetation controls on braided stream dynamics. Water Resource Research 37: 3275-3283. 84 Griffith B and Youtie BA. (1988) Two devices for estimating foliage density and deer hiding cover. Wildlife Society Bulletin 16: 206-210. Guard BJ. (1995) Wetland Plants of Oregon and Washington, Auburn, Washington: Lone Pine Publishing. Gurnell AM. (1997) The hydrological and geomorphological significance of forested floodplains. Global Ecology and Biogeography Letters 6: 219-229. Gurnell AM, Boitsidis AJ, Thompson K, et al. (2006a) Seed bank, seed dispersal and vegetation cover: Colonization along a newly-created river channel. Journal of Vegetation Science 17. Gurnell AM, Morrissey IP, Boitsidis AJ, et al. (2006b) Initial adjustments within a new river channel: Interactions between fluvial processes, colonizing vegetation, and bank profile development. Environmental Management 38. Gurnell AM, O'Hare JM, Dunbar MJ, et al. (2010) An exploration of associations between assemblages of aquatic plant morphotypes and channel geomorphological properties within British rivers. Geomorphology 116: 135-144. Gurnell AM and Petts G. (2006) Trees as riparian engineers: the Tagliamento River, Italy. Earth Surface Processes and Landforms 31: 1558-1574. Hay LE, McCabe GJ, Clark MP, et al. (2009) Reducing streamflow forecast uncertainty: Application and qualitative assessment of the Upper Klamath River Basin, Oregon. Journal of American Water Resources Association 45: 580-596. Hey RD and Thorne CR. (1986) Stable channels with mobile gravel beds. Journal of Hydraulic Engineering 112: 671-689. Hicks DM and Mason PD. (1998) Roughness Characteristics of New Zealand Rivers, Highlands Ranch, Colorado: Water Resources Publications. Jarrett RD. (1984) Hydraulics of high-gradient rivers. Journal of Hydraulic Engineering 110: 1519-1539. Kercher SM and Zedler JB. (2004) Multiple disturbances accelerate invasion of reed canary grass (Phalaris arundinacea L.) in a mesocosm study. Oecologia 138: 455- 464. Kouwen N. (1988) Field estimation of biomechanical properties of grass. Journal of Hydraulic Research 26: 559-568. Lavoie C, Dufresne C and Delisle F. (2005) The spread of reed canarygrass (Phalaris arundinacea) in Quebec: A spatio-temporal perspective. Ecoscience 12: 366-375. 85 Leonard LA and Luther ME. (1995) Flow hydrodynamics in tidal marsh canopies. Limnol. Oceanogr 40: 1474-1484. Marten GC and Heath ME. (1973) Reed Canarygrass. In: Heath ME, Metcalfe DS and Barnes RF (eds) Forages: The Science of Grassland Agriculture. Ames, Iowa: The Iowa State University Press. McKenney R, Jacobson RB and Wertheimer RC. (1995) Woody vegetation and channel morphogenesis in low gradient, gravel-bed streams in the Ozark Plateaus, Missouri and Arkansas. Geomorphology 13: 175-198. Millar RG. (2000) Influence of Bank Vegetation on Alluvial Channel Patterns. Water Resources Research 36: 1109-1118. Millar RG. (2005) Theoretical regime equations for mobile gravel-bed rivers with stable banks. Geomorphology 64: 207-220. Millar RG and Quick MC. (1998) Stable width and depth of gravel-bed rivers with cohesive banks. Journal of Hydraulic Engineering October: 1005-1013. Murray AB, Knaapen MAF, Tal M, et al. (2008) Biomorphodynamics: Physical- biological feedbacks that shape landscapes. Water Resources Research 44: 1-18. Naglich FG. (1994) Reed Canarygrass (Phalaris arundinacea L.) in the Pacific Northwest: Growth Parameters, Economic Uses, and Control. Environmental Studies. Olympia, Washington: The Evergreen State College. Nevins THF. (1969) River training-the single thread channel. New Zealand Engineering: 367-373. Nudds TD. (1977) Quantifying the vegetative structure of wild-life cover. Wildlife Society Bulletin 5: 113-117. Perucca E, Camporeale C and Ridolfi L. (2007) Significance of the riparian vegetation dynamics on meandering river morphodynamics. Water Resource Research 43: 1- 10. Phillips JV and Ingersoll TL. (1998) Verification of roughness coefficients for selected natural and constructed stream channels in Arizona. USGS Professional Paper 1584. Pojar J and MacKinnon A. (1994) Plants of the Pacific Northwest Coast, Vancouver, BC: Lone Pine Publishing. 86 Rahmeyer WH, Werth D and Freeman GE. (1999) Improved methods of determining vegetative resistance in floodplains and compound channels. Proceedings of the 1999 International Water Resources Engineering Conference. Seattle, WA: American Society of Civil Engineers. Rasmussen CG. (2011) Geomorphology, hydrology and biology of floodplain vegetation in the Sprague Basin, OR: History and potential for natural recovery. Geography. University of Oregon, 105. Reinhardt L, Jerolmack D, Cardinale BJ, et al. (2010) Dynamic interaction of life and its landscape: feedbacks at the interface of geomorphology and ecology. Earth Surface Processes and Landforms 35: 78-101. Risley JC, Gannett MW, Lea JK, et al. (2005) An Analysis of Statistical Methods for Seasonal Flow Forecasting in the Upper Klamath River Basin of Oregon and California. USGS and U.S. Department of the Interior 2005-5177. Roberson JA and Crowe JT. (1993) Engineering fluid mechanics, New York. Sage RB, Hallins K, Gregory CL, et al. (2004) Impact of roe deer Capreolus capreolus browsing on understorey vegetation in small farm woodlands. Wildlife Biology 10: 115-120. Sand-Jensen K and Mebus JR. (1996) Fine-scale patterns of water velocity within macrophyte patches in streams. Oikos 76: 169-180. Schulz M, Kozerski H-P, Pluntke T, et al. (2003) The influence of macrophytes on sedimentation and nutrient retention in the lower River Spree (Germany). Water Research 37: 569-578. Simon A and Castro J. (2003) Measurement and Analysis of Alluvial Channel Form. In: Kondolf GM and Piegay H (eds) Tools in Fluvial Geomorphology. West Sussex: John Wiley & Sons. Steiger J, Gurnell AM and Goodson JM. (2003) Quantifying and characterizing contemporary riparian sedimentation. River Research and Applications 19: 335- 352. Tal M and Paola C. (2007) Dynamic single-thread channels maintained by the interaction of flow and vegetation. Geology 35: 347-350. Tal M and Paola C. (2010) Effects of vegetation on channel morphodynamics: results and insights from laboratory experiments. Earth Surface Processes and Landforms. 87 Temmerman S, Bouma TJ, Van de Koppel J, et al. (2006) Plant growth initiates channel erosion in flat landscapes: evidence from tidal marshes. Geophysical Research Abstracts 8: 08871. Tourbier JT and Westmacott R. (1981) Water resources protection technology, Washington D.C.: Urban Land Institute. Van Hulzen JB, Van Soelen J and Bouma TJ. (2007) Morphological variation and habitat modification are strongly correlated for the autogenic ecosystem engineer Spartina anglica (Common cordgrass). Estuaries and Coasts 30: 3-11. Viles HA. (1988) Biogeomorphology, Oxford: Basil Blackwell. Wilson BL, Brinerd RE, Lytjen D, et al. (2008) Field Guide to Sedges of the Pacific Northwest, Corvallis: Oregon State University Press. Chapter IV Anderson DE. (1961) Taxonomy and distribution of the genus Phalaris. Iowas State Journal of Science 36: 1-96. Antieau CJ. (2004) Biology and management of reed canarygrass, and implications for ecological restoration, Seattle, Washington: Department of Transportation Apfelbaum SI and Sams CE. (1987) Ecology and control of reed canarygrass. Natural Areas Journal 69-74. Arcement GJ and Schneider VR. (1989) Guide for selecting manning's roughness coefficients for natural channels and floodplains. United States Geological Survey Water-supply Paper 2339. Asaeda T, Rajapakse L and Kanoh M. (2010) Fine sediment retention as affected by annual shoot collapse: Sparganium erectum as an ecosystem engineer in a lowland stream. River Research and Applications 26: 1153-1169. Carollo FG, Ferro V and Termini D. (2005) Flow resistance law in channels with flexible submerged vegetation. Journal of Hydraulic Engineering 131: 554-564. Clarke A, Lake PS and O'Dowd DJ. (2004) Ecological impacts on aquatic macroinvertebrates following upland stream invasion by a ponded pasture grass (Glyceria maxima) in southern Australia. Marine and Freshwater Research 55: 709-713. Corenblit D, Gurnell AM, Steiger J, et al. (2008) Reciprocal adjustments between landforms and living organisms: Extended geomorphic evolutionary insights. Catena 73: 261-273. 88 Corenblit D, Steiger J, Gurnell AM, et al. (2009) Plants intertwine fluvial landform dynamics with ecological succession and natural selection: a niche construction perspective for riparian systems. Global Ecology and Biogeography 18: 507-520. De Baets S, Poesen J, Gyssels G, et al. (2006) Effects of grass roots on the erodibility of topsoils during concentrated flow. Geomorphology 76. Edwards PJ, Kollmann J, Gurnell AM, et al. (1999) A conceptual model of vegetation dynamics on gravel bars of a large Alpine river. Wetlands Ecology and Managment 7: 141-153. Fierke M and Kauffman J. (2006) Riverscape-level patterns of riparian plant diversity along a successional gradient, Willamette river, Oregon. Plant Ecology 185: 85- 95. Foster RD and Wetzel PR. (2005) Invading Montoypic Stands of Phalaris arundinacea: A Test of Fire, Herbicide, and Woody and Herbaceous Native Plant Groups. Restoration Ecology 13: 318-324. Freeman GE, Rahmeyer WH and Copeland RR. (2000) Determination of resistance due to shrubs and woody vegetation. U.S. Army Corps of Engineers ERDC/CHL TR- 00-25. Friedman JM, Osterkamp WR and Lewis WMJ. (1996) Channel Narrowing and Vegetation Development Following a Great Plains Flood. Ecology 77: 2167-2181. Friedrichsen PT. (1997) Summertime stream temperatures in the North and South Forks of the Sprague River, South Central Oregon. Corvallis, Oregon: Oregon State University, 147. Galatowitsch SM, Anderson NO and Ascher PD. (1999) Invasiveness in wetland plants in temperate North America. Wetlands 19: 733-755. Graf WL. (1978) Fluvial adjustments to the spread of tamarisk in the Colorado Plateau region. Geological Society of America Bulletin 89. Gran K and Paola C. (2001) Riparian vegetation controls on braided stream dynamics. Water Resource Research 37: 3275-3283. Guard BJ. (1995) Wetland Plants of Oregon and Washington, Auburn, Washington: Lone Pine Publishing. Gurnell AM. (1997) The hydrological and geomorphological significance of forested floodplains. Global Ecology and Biogeography Letters 6: 219-229. 89 Gurnell AM, O'Hare JM, Dunbar MJ, et al. (2010) An exploration of associations between assemblages of aquatic plant morphotypes and channel geomorphological properties within British rivers. Geomorphology 116: 135-144. Gurnell AM and Petts G. (2006) Trees as riparian engineers: the Tagliamento River, Italy. Earth Surface Processes and Landforms 31: 1558-1574. Hay LE, McCabe GJ, Clark MP, et al. (2009) Reducing streamflow forecast uncertainty: Application and qualitative assessment of the Upper Klamath River Basin, Oregon. Journal of American Water Resources Association 45: 580-596. Järvelä J. (2004) Determination of flow resistance caused by non‐submerged woody vegetation. International Journal of River Basin Management 2: 61-70. Jones CG, Lawton JH and Shackak M. (1994) Organisms as ecosystem engineers. Oikos 69: 373-386. Jones, C.G., Lawton, J.H., Shachak, M., 1997. Positive and negative effects of organisms as physical ecosystem engineers. Ecology, 78(7), 1946-1957. Kercher SM and Zedler JB. (2004) Multiple disturbances accelerate invasion of reed canary grass (Phalaris arundinacea L.) in a mesocosm study. Oecologia 138: 455-464. Kouwen N and Unny TE. (1973) Flexible roughness in open channels. Journal of Hydraulics Division 99: 713-728. Kouwen N, Unny TE and Hill HM. (1969) Flow retardance in vegetated channels. Journal of the Irrigation and Drainage Division 95: 329-342. Lavoie C, Dufresne C and Delisle F. (2005) The spread of reed canarygrass (Phalaris arundinacea) in Quebec: A spatio-temporal perspective. Ecoscience 12: 366-375. Levine JM. (2000) Complex interactions in a streamside plant community. Ecology 81: 3431-3444. Loy WG, Allan S, Buckley AR, et al. (2001) Atlas of Oregon, Eugene: University of Oregon Press. Marten GC and Heath ME. (1973) Reed Canarygrass. In: Heath ME, Metcalfe DS and Barnes RF (eds) Forages: The Science of Grassland Agriculture. Ames, Iowa: The Iowa State University Press. Martinez AE. (2013) The Geomorphic Effects of Native and Invasive Riparian Vegetation: Sprague River, Oregon. Department of Geography. Eugene, Or: University of Oregon. 90 Maurer MA, Auld DA, Murphy CL, et al. (2001) Evaluation of Twenty-six Buffalograss Cultivars and Accessions for Use as Turfgrass on the High Plains of West Texas. Texas Journal of Agriculture and Natural Resources 14: 24-36. McDonald RR, Nelson JM and Bennett JP. (2005) Multi-dimensional surface-water modeling system user's guide. U.S. Geological Survey Techniques and Methods 6- B2: 136p. Musleh FA and Cruise JF. (2006) Functional relationships fo resistance in wide flood plains with rigid unsubmerged vegetation. Journal of Hydraulic Engineering 132: 163-171. Naglich FG. (1994) Reed Canarygrass (Phalaris arundinacea L.) in the Pacific Northwest: Growth Parameters, Economic Uses, and Control. Environmental Studies. Olympia, Washington: The Evergreen State College. Nelson JM, Bennett JP and Wiele SM. (2003) Flow and sediment transport modeling. In: Kondolph M and Piegay H (eds) Tools in Fluvial Geomorphology. Chichester, England: Wiley and Sons, 539-576. Nelson JM and McDonald RR. (2010) The International River Interface Cooperative: Public Domain Software for River Modeling. 2nd Joint Federal Interagency Conference. Las Vegas, NV. OWRD. (2013) OWRD Near Real Time Hydrographic Data, 11497500, SPRAGUE R NR BEATTY, OR. Pasche E and Rouve G. (1985) Overbank flow with vegetatively roughened floodplains. Journal of Hydraulic Engineering 111: 1262-1278. Perucca E, Camporeale C and Ridolfi L. (2007) Significance of the riparian vegetation dynamics on meandering river morphodynamics. Water Resource Research 43: 1- 10. Petryk S and Bosmajian G. (1975) Analysis of flow through vegetation. Journal of the Hydraulics Division 101: 871-884. Pojar J and MacKinnon A. (1994) Plants of the Pacific Northwest Coast, Vancouver, BC: Lone Pine Publishing. Prosser IP, Dietrich WE and Stevenson J. (1995) Flow resistance and sediment transport by concentrated overland flow in a grassland valley. Geomorphology 13. 91 Rahmeyer WH, Werth D and Freeman GE. (1999) Improved methods of determining vegetative resistance in floodplains and compound channels. Proceedings of the 1999 International Water Resources Engineering Conference. Seattle, WA: American Society of Civil Engineers. Rasmussen CG. (2011) Geomorphology, hydrology and biology of floodplain vegetation in the Sprague Basin, OR: History and potential for natural recovery. Geography. University of Oregon, 105. Risley JC, Gannett MW, Lea JK, et al. (2005) An Analysis of Statistical Methods for Seasonal Flow Forecasting in the Upper Klamath River Basin of Oregon and California. USGS and U.S. Department of the Interior 2005-5177. Sand-Jensen K and Mebus JR. (1996) Fine-scale patterns of water velocity within macrophyte patches in streams. Oikos 76: 169-180. Schulz M, Kozerski H-P, Pluntke T, et al. (2003) The influence of macrophytes on sedimentation and nutrient retention in the lower River Spree (Germany). Water Research 37: 569-578. Tal M, Gran K, Murray AB, et al. (2004) Riparian Vegetation as a Primary Control on Channel Characterisitcs in Multi-thread Rivers. In: Bennett SJ and Simon A (eds) Riparian Vegetation and Fluvial Geomorphology. Tooth S and Nanson GC. (1999) Anabranching rivers on the Northern Plains of arid central Australia. Geomorphology 29: 211-233. Tooth S and Nanson GC. (2000) The role of vegetation in the formation of anabranching channels in an ephemeral river, Northern plains, arid central Australia. Hydrological Processes 14: 3099-3117. Tourbier JT and Westmacott R. (1981) Water resources protection technology, Washington D.C.: Urban Land Institute. Wilson BL, Brinerd RE, Lytjen D, et al. (2008) Field Guide to Sedges of the Pacific Northwest, Corvallis: Oregon State University Press. Wu F-C, Shen HW and Chou Y-J. (1999) Variation of roughness coefficients for unsubmerged and submerged vegetation. Journal of Hydraulic Engineering September: 934-942.