H53B-1237 1340h
The Use of Fallout Radionuclides to Quantify Downstream Trends in Sediment Transport Below a Flood-control Dam
Flow regulation by dams has a major impact on the hydrology of river systems, altering the timing, magnitude, and frequency of natural flows. Dams both store water and capture sediment, so that the downstream geomorphic effect differs whether considering the dam's effect on sediment discharge or its effect on transport capacity. While the geomorphic effects of flow regulation have been well quantified, few studies have focused on the mechanism by which these changes occur. The purpose of this study is to investigate how flow regulation alters the sediment transport regime below the dam, specifically in relation to sediment residence time, and to use short-lived fallout radionuclides to quantify this effect. Sediment residence time varies as a function of landscape, position, climate, and anthropogenic disturbance. The 7Be activity of transitory bed sediment was measured bi-weekly from February 2004 until August 2004 at multiple sites below a flood-control dam on the Ompompanoosuc River, Vermont. Sites were selected to represent a progression of flow regulation; highly regulated below the dam and progressively less regulated with distance downstream. Samples were taken following large storm events and during dry periods in order to capture temporal trends in sediment flux, mobilization, and deposition. The transition from flood control to run-of-the-river operation in late spring allowed us to compare sediment flux under regulated and natural flows. Because of the short half-life of fallout 7Be (53.4 days), the sediment stored behind the dam quickly became depleted in 7Be activity during the winter. Thus the spring release provided a pulse of effectively "dead" sediment which could be tracked as it moves downstream. This pulse of "dead" sediment modified the normal recovery of downstream sites. In addition, sediment exposed by the draining of the reservoir behind the dam provided an easily mobilized source of "new" sediment that was transported downstream. Using the location of these sediment pulses at different points in time, we were able to calculate an average sediment transport velocity and flushing rate. These results indicate that changes occur over relatively short distances and short timeframes, especially during high flows. This study has major implications for river research and management, presenting a quantitative technique for assessing the effect of flow regulation and dam operation on sediment transport and storage.
H53B-1238 1340h
Source Determination and Residence Time Indications for a Large River System
Characterization of suspended particulates in large river systems is important and challenging. The world's 25 largest rivers, in terms of sediment discharge, account for approximately 40% of the fluvial sediments that enter the ocean. Particulates that enter the ocean from rivers are the products of integrated basin-wide processes. This study focuses on the development of a suite of proxies that give insight to the source and system-wide residence time of particulates discharged by rivers. To completely characterize the suspended sediments in a major river system, regular sampling over longer periods of time is necessary. A river's suspended load is highly variable over a range of time scales, therefore high temporal resolution sampling is required. Continuous, regular sampling incorporates not only seasonal and annual change, but it also allows interpretation for differences on a sub-seasonal scale. Biweekly sampling of the Mississippi River was performed from March 2002 through April 2004. Previous studies have used short-lived radionuclides (7Be, 137Cs, and 210Pb) to determine sediment source and residence time in small agricultural watersheds. Using the same approach in a large river system is more challenging because sources and sinks are more numerous and complex and residence times change from smaller to larger systems. As a result of these complexities, the use of short-lived radioisotopes alone was not completely diagnostic in the Mississippi River system. Trends in 7Be and 137Cs activities appear to be driven by mixing of old and new sediments. Therefore, both source and residence time effects control the observed radioisotope activities. Additional examinations of grain size, 234U/238U isotope activity ratios, and sediment mineralogy were done to better evaluate basin sources. Three up-basin sampling trips were also performed in February, April, and July of 2004. The upper Mississippi, Missouri, Ohio, and Arkansas rivers were sampled each time. These end-member samples are essential to determining primary source, residence time, and are used in a 7Be mass balance analysis for the Mississippi River system. When done in conjunction with basin source indicators, the mass balance approach sheds light on how man-made structures influence particle retention times, long term storage, and particle alterations.
H53B-1239 1340h
Flow and sediment-transport modeling of Kootenai River White Sturgeon Spawning Habitat.
The population of White Sturgeon in the Kootenai River downstream of Libby Dam in Montana and Idaho has declined since the construction of the dam in 1972. The White Sturgeon was listed as endangered in 1994 and an 11.2 mile reach of the river, downstream of Bonners Ferry, Idaho was designated as Critical Habitat in 2001. It is hypothesized that hydro-electric and flood control operations have contributed to poor spawning habitat and recruitment of juvenile fish. The successful incubation of eggs requires a stable and coarse bed material. Currently the sturgeon are spawning in a reach of poor substrate consisting of dunes up to 2 meters in amplitude and composed of fine sand while a short distance upstream there is suitable substrate of coarse gravel. We present here the preliminary results of a flow and sediment-transport modeling effort to aid in an understanding of both the current spawning habitat of the White Sturgeon and the potential to artificially enhance the current spawning habitat or to influence the sturgeon to move upstream to more suitable habitat. A 2.5 dimensional flow model was constructed for an 8-kilometer reach of the designated Critical Habitat. The modeled reach consists of several broad meanders and a mid channel island. The substrate is composed of fine sand with a median grain size of 0.22mm and has large dunes up to 2m in amplitude at relatively lows flows of 200 cms that wash out to a plane bed at around 600 cms. The model has been calibrated to a range of historical flow conditions from 170 cms to 1709 cms and verified against 16 ADCP velocity cross-section profiles collected during a period of steady flow at 554 cms. The model predicts well most of the salient features of the velocity field including the magnitude and location of the secondary flow, using a simple constant value for roughness. However for a few reaches of the river the bed forms and their spatial variability in size are shown to significantly affect the flow and the performance of the model can be enhanced by assigning a variable roughness. Over the range of modeled flows most of the bed is highly mobile. Comparison of the model results for velocity, bed shear stress and depth to the locations of known spawning activity show that spawning locations are found in and downstream of deep holes with high velocity.
H53B-1240 1340h
A Field Investigation to Assess the Effects of Riparian Vegetation Management on Alluvial Channel Morphology
The Platte River in central Nebraska intersects a principal migratory route for North American waterfowl including several endangered and threatened avian species. Upstream water resource development has altered the regime of water and sediment inputs to the central Platte River. The once wide, braided channel has narrowed significantly over the last century and a dense riparian forest has been established along some reaches. A recovery program proposed for the Platte River will use adaptive management to restore habitats lost through vegetation encroachment and channel narrowing. The Cottonwood Ranch, a 2,650-acre parcel of land owned by Nebraska Public Power District (NPPD), is a recovery program property that is being enhanced for avian habitat. Vegetation has been removed from river banks and islands to promote lateral erosion and widening of the channel with the objective of increasing unobstructed views for cranes. Because riparian management activities also have the potential to cause geomorphic change downstream, the U.S. Geological Survey designed a monitoring study to detect what effects, if any, the management might have upon the local channel morphology and sediment-transport processes. The study required the measurement of cross-section elevation profiles and bed-material size gradations upstream, within, and downstream of the managed area prior to, and following, the management activities. However, since the monitoring study began in 2000, Platte River flows have been relatively low. Hence, up to this point, fluvial action has not stimulated substantial lateral erosion and downstream redistribution of sediments from the managed area. The monitoring protocol, while specifically developed to study the effects of vegetation clearing, could be applied to evaluate the geomorphic response of additional reaches of the Platte River to other proposed management activities, including sediment augmentation from island leveling and controlled releases from upstream dams.
H53B-1241 1340h
The Use of Slope Creation as a Rehabilitation Tool on Regulated Rivers
Channel conditions below dams are degraded by bed armoring, channel narrowing, and channel incision. The last process yields a long-term decrease in slope, as a river's base level is fixed at sea level. Thus, hydrodynamic habitat conditions controlled by slope degrade over time, impacting populations naturally selected for pre-dam dynamics. In this experimental river rehabilitation, it was hypothesized that channel slope below the dam could be increased incrementally, and this increase could then be spread downstream in subsequent years, eventually yielding a reconstruction of the pre-dam longitudinal profile and increasing overall habitat quality. To test the hypothesis, competing experimental designs for the first stage of this plan were created and evaluated for the gravel-bed Mokelumne River in central CA below Camanche Dam using the Spawning Habitat Integrated Rehabilitation Approach. Gravel fill depth was constrained by the monitored 3 ft of incision over 40 years. Alternative designs aimed to maximize bed elevation gain, spawning habitat quality, and habitat heterogeneity, while avoiding bed scour at the test flow of 500 cfs. Using a 2D hydrodynamic model designs with a 3 ft fill were predicted to cause excessive scour regardless of design elements. Iteration yielded the key result that a maximum fill of 1.5 ft could be sustained by notching the upstream riffle to enable excess flow to bypass the spawning area thereby protecting it from scour and providing desirable habitat heterogeneity. Upon placing 2000 cu yds of gravel, riffle-to-riffle slope was raised from 0.001 to 0.0025. Due to project phasing the created slope yielded scour at the end of the site where the bed elevation dropped back to the original level, as predicted. Lower flows than originally expected resulted in gravel exposure. The second phase added another 2000 cu yds of gravel effectively spreading the slope downstream and backing water up to the first site to increase depth and decrease velocity to the desired levels. On-going post project monitoring and analysis is being used to assess slope creation as a spawning habitat rehabilitation tool.
http://shira.lawr.ucdavis.edu
H53B-1242 1340h
Relationships between River Discharge and Stream Bed Scour: Implications for Redd Scour on a Large Regulated River
A critical knowledge gap for implementing environmental flow releases on regulated rivers is understanding the effects of dam releases on the potential for scouring of spawning redds in downstream reaches. To address this question we need to understand the relationships between river discharge, bed mobility, and scour depth in river reaches heavily utilized by spawning salmon. Once these relationships are established it will be possible to calculate the probability that dam releases and/or tributary generated floods will result in a significant level of egg/embryo mortality. Our approach couples numerical modeling and empirical data to quantify spatially explicit predictions of bed mobility and identify specific areas where scour is deep enough to impact redd viability. Boundary shear stress values are predicted using the USGS's Multi-Dimensional Surface Water Modeling System (Nelson and Smith, 1989) for a segment of the Trinity River below Lewiston Dam. From model-generated shear stress and mapping of local particle size distributions, Shield stress parameters are determined to quantify areas of differential bed mobility. Field data acquired during high flow events is used to calibrate the model and to obtain estimates of scour depth. Scour depth is measured using a dense array of scour chains supplemented by motion-sensing radio tagged particles that record the specific time, and hence discharge, that particles in the surface and subsurface layer become mobile. Zones of differential mobility are compared with detailed maps of spawning sites for Chinook to quantify risk of significant redd scouring and to determine whether salmon are spawning in areas of inherently high or low mobility.
H53B-1243 1340h
Gravel Augmentation Below Dams: California Experiences
Most dams block all coarse sediment traveling downstream, such that reaches downstream are commonly typically depleted of gravel, causing a variety of effects such as incision, bank erosion, coarsening of the bed material, and reduction of salmonid spawning habitiat. To compensate for this reduction in coarse sediment supply, gravel has been artificially added below dams, using techniques such as high flow stock piling, high flow direct injection, artificial riffle construction, riffle supplementation, and construction of side channel or artificial spawning channels. In the Trinity and Sacramento-San Joaquin River systems of northern California, loss of suitable salmonid spawning gravels below dams has motivated augmentation of over 320,000 m3 of gravel in 73 separate projects on 19 rivers since 1978, mostly since 1990. Of the 67 projects for which adequate data were available, 48 involved adding less than 7,500 m3 each. Costs reported for 57 of the projects totaled nearly $8,753,000, but these figures generally did not include the cost of staff time involved in planning, design, and oversight. Despite the magnitude of this experimental intervention, fewer than half of the projects were monitored, and of those few had monitored sufficient parameters pre- and post- project to evaluate project performance. Performance of these projects to date has been mixed: in many cases the imported gravels have promptly washed out, some channel forms created have been unnatural and not heavily used by salmon. In all cases, the volumes of gravel artificially added have been only a small percentage of the annual coarse sediment deficit.
H53B-1244 1340h
Predictive Design Morphologies for Gravel Augmentation
Spawning habitat rehabilitation (SHR) is an interdisciplinary practice merging hydrology, geomorphology, aquatic ecology, and civil engineering to improve existing aquatic habitat and restoring fluvial complexity. Although SHR is widespread, it needs a science-based design process. The Spawning Habitat Integrated Rehabilitation Approach (SHIRA) is a scientifically peer-reviewed framework for doing SHR on regulated rivers. Although SHIRA has shown success with gravel augmentation on the Mokulmne River using hypothesis driven designs, the goal of this study was to evaluate several more natural processes for their potential in SHR, and to do so at the geomorphic-unit scale for the first time. Multiple design hypotheses were included in 6 SHR scenarios for rehabilitating the Lewiston Dam reach of the Trinity River, CA. Morphologies tested for their process mechanics included central bars, transverse-oblique bars, riffles, point bars, and bench-constricted pools. Varying longitudinal and lateral approach slopes for each feature were evaluated as well as feature sequencing. For each design scenario, a 2D model predicted local depth, velocity, shields stress, depth of scour, and habitat suitability for life stages of chinook and steelhead salmon at 300 and 6000 cfs. Data were analyzed to determine if conceptually expected geomorphic and ecological outcomes were in fact predicted by the 2D model. One design will be selected for actual construction in 2005 to evaluate 2D model predictions.
http://shira.lawr.ucdavis.edu
H53B-1245 1340h
Geomorphic Effects of Boulder Placement on Gravel Capture and Retention in a Regulated Reach of the North Umpqua River, OR.
Hydroelectric projects in the mountainous western Cascades often occur in steep, confined channels where salmonid spawning habitat is limited to gravel deposits forced by planform curvature, channel width changes, and flow separation associated with large bedrock and boulder obstructions. The paucity of gravel deposition in steepland channels may be exacerbated in regulated rivers where sediment trapping by impoundments reduces coarse sediment supply to downstream reaches. Placing boulders to capture and retain gravel may be an effective approach to enhancing spawning habitat in these settings. To better understand the potential use of boulders as a tool for enhancing spawning habitat, three experimental designs were tested in a 0.6-mile bypass reach of the North Umpqua River, OR. The bedrock-confined study reach has an average slope of 0.013 and plane-bed morphology with coarse cobble substrate, abundant marginal boulders, and small associated patches of sand and gravel. Experiments involved (1) placement of boulder clusters, (2) gravel augmentation and placement of boulder clusters, and (3) gravel augmentation alone. Boulder clusters were designed to promote scour and deposition during floods with a 5-10 year recurrence interval. Boulders were typically placed obliquely upstream at locations where existing hydraulics favored gravel deposition. Monitoring from 2002 to 2004 occurred prior to implementation, immediately following implementation, and following winter high flows. Sites were monitored using high-density topographic surveys, low-altitude aerial photography, facies mapping, pebble counts, scour cores and chains, and marked rocks. Stage heights were monitored using pressure transducers at the upstream and downstream ends of the study reach, and flood recurrence interval was assessed using a nearby USGS gauge. The arrangement of boulder clusters was modified after the first year of monitoring to improve gravel capture and retention. Peak flow during the two-year monitoring period had a recurrence interval of less than 1.5 years. Flows were insufficient to mobilize the bed as a whole, but did adjust bed surface texture and topography adjacent to boulder accumulations. Select sites captured and retained modest amounts of gravel even at the relatively low peaks experienced during 2003 and 2004. The effects of increasing coarse sediment supply will be tested in 2005 through the introduction of a large gravel pulse at the upstream end of the study reach.
H53B-1246 1340h
Predictive Tools for Evaluating Aeolian Sediment Redistribution After Experimental Floods: Monitoring Studies in the Colorado River Corridor, Grand Canyon, Arizona
The Colorado River through Grand Canyon is subject to complex river management protocols and multi-faceted geomorphic research through the Glen Canyon Dam Adaptive Management Program. Predicting aeolian redistribution of sediment following experimental floods is important for assessing the potential of controlled flooding to help preserve archaeological sites by replenishing sediment deposits above the flood-stage elevation. We present initial results of an ongoing instrumentation program supported by the Grand Canyon Monitoring and Research Center in which aeolian sediment transport rates, wind magnitude and direction, and precipitation are measured at multiple locations along the river corridor. These data allow resolution of seasonal and regional variability in wind intensity and direction, and resultant aeolian sediment transport, as well as precipitation patterns. Data collected since Fall 2003 indicate that wind velocities and sand transport were greatest during April and May 2004 at all locations studied (with winds locally $>$25 m/s, and transport rates locally $>$9 kg/m/day). Dominant wind direction during strong wind intervals varies with location, but during the April-May windy season the greatest transport potential was directed upstream in Marble Canyon (upper Grand Canyon). Such information can be used to evaluate the potential for aeolian reworking of new fluvial sand deposits, and restoration of higher-elevation aeolian deposits, following experimental beach/habitat building flows. These aeolian deposits, many of which contain and preserve archaeological material, comprise a critical part of the Grand Canyon ecosystem.
H53B-1247 1340h
A Comparison of Techniques for Mapping the Distribution of Sediment on the Bed of the Colorado River in Grand Canyon
The Grand Canyon Monitoring and Research Center is charged with establishing and implementing monitoring projects to provide scientific information to the Glen Canyon Dam Adaptive Management Program (GCDAMP) on the effects of operating Glen Canyon Dam on the downstream resources of the Colorado River ecosystem. One primary resource of concern to the GCDAMP is fine-grained sediment. Glen Canyon Dam traps approximately 94% of the pre-dam sand supply to the Colorado River in Grand Canyon, resulting in a decline in the size of eddy sand bars (25% decline in surface area over the past 15 years). Sand bars are an important resource because they provide habitat for endangered native fish, protect archeological sites, provide substrate for vegetation, are used as campsites and are a distinctive feature of the pre-dam environment. A combination of traditional survey techniques and multi-beam bathymetry has been used to determine the size and elevation of sandbars and to obtain topographic maps of the riverbed. These techniques have proven useful in evaluating the spatial changes and channel morphology along the Colorado River ecosystem. While previous studies have been very effective in measuring volumetric and spatial changes, a method is needed map the distribution of sediment along the submerged portion of the river channel. The distribution of fine-grained sediment is needed to evaluate the potential for deposition onto high elevation sand bars during proposed experimental high flows. This study used high-resolution multibeam bathymetry, acoustic backscatter and underwater video images collected on expeditions in 2002 and 2004 to evaluate the different methodologies. The purpose of this study was to evaluate possible technologies to be used in determining the distribution of sediment along the bed of the Colorado River in Grand Canyon. These technologies include: 1) visual interpretation of shaded relief images produced from multibeam bathymetry; 2) visual interpretation of acoustic backscatter images; 3) acoustic seabed classification using QTC Multiview. An evaluation of underwater video images was used to ground truth the various techniques. An evaluation of the results of each of the techniques, as well as a comparison of the efforts required will be presented.
H53B-1248 1340h
The Influence of Dams on Streams of the Mid-Atlantic Region, USA
Dams represent a significant perturbation on streams by influencing the supply of water and sediment to downstream reaches. We assessed the downstream effects of dams at fifteen sites of varying dam size in Pennsylvania and Maryland by comparing downstream reaches with an upstream control reach. We found that the bed material in the downstream reach is coarser than upstream. Specifically, downstream reaches have less mud, sand, and granules than upstream reaches, and more pebbles and cobbles. The fraction of boulders and exposed bedrock is statistically similar both upstream and downstream. Using a simple numerical model for the grain size distribution of bedrock influenced channels, we estimated that trap efficiencies for bed material are surprisingly high at our sites, ranging from around 10 percent to around 50 percent. We also found that dams had no consistent influence on channel width and slope, suggesting that dams do not significantly affect channel morphology at our sites. We attribute this to: 1) pervasive bedrock influence on streams in this region, 2) low regional sediment supply, so upstream reaches unaffected by dams are sediment starved, similar to downstream reaches and thus little sediment is available below dams to affect channel morphological change through deposition, 3) highly vegetated and cohesive banks that are difficult to erode, limiting possible width adjustment. Our results emphasize that geologic history and setting are primary factors controlling how dams influence stream channels.
H53B-1249 1340h
Dam-incuced Changes in Geomorphology and Vegetation Along a Stream in Northern California
Dams are well known for trapping sediment and altering natural flow regimes that affect downstream channel geometry and the distribution of riparian vegetation. While many studies have evaluated pre-dam and post-dam effects, and land-use activity adjacent to the channel on riparian vegetation and channel morphology, few have included GIS mapping and an undammed reference stream to serve as a control for studying responses in an alluvial system. This paper evaluated the effects of Warm Springs Dam (established in 1983) on the variation, magnitude, and directional changes of stream channel geometry and riparian vegetation distribution along Dry Creek and compared the changes to a nearby undammed stream with similar geomorphic and land-use characteristics. Six historical black and white aerial photographs were examined for both streams over a 34 year period prior to the dam's establishment (1942-1976), and a 13 year period after (1987-2000), after being scanned and georeferenced in a GIS. For each year, three stream and riparian features were manually digitized on-screen, including the center of the stream channel, bankfull width, and patches of riparian vegetation, as well as the distance land-use was to the channel. Multi-way statistical analyses evaluated variation in stream length and distance that land-use moved from the channel as well as variation and change in the rate and direction of bankfull area and riparian area. Rating curves and hydraulic geometry exponents used stream gauge measurements that analyzed changes in channel geometry (width, depth and velocity). While mean variation in the reference stream's length and bankfull area remained constant during the 58-year study period, Dry Creek's stream length varied 84% (P=0.02) less (shortened 550 m within the 10.5 km study reach), and bankfull area decreased by 52.5% (P=0.01) after the dam. Riparian vegetation decreased 28.5% (P$<$0.0001) from 1942-1986 on Dry Creek then increased by 2000 to levels similar to those in 1942. Areas of riparian vegetation in the reference stream consistently increased each year of the study by a total of 38.5% (P$<$0.0001). Spatial variation in the randomly selected study segments explained most of the variation (40%, P=0.04) and rate and directional changes (53%, P=0.04) in riparian area. Variation in the distance land-use activity moved from the channel significantly explained variation in the rate and directional changes of riparian area and bankfull area on both streams (P=0.006 and 0.004, respectively). Rating curves and hydraulic geometry exponents indicated that Dry Creek's channel bed just below the dam was fairly well armored by 2000; the mid-stream channel bed had lowered a mean 1.02 m (P=0.004) from 1987-2003, and a mean 1.41 m (P$<$0.0001) near the mouth (22 km downstream from the dam). Dry Creek's sediment starved channel incised and became entrenched allowing vegetation to colonize the less frequently flood prone banks and bars. Even though the dam caused major morphological changes to Dry Creek's channel and distribution of riparian vegetation, land-use activity played an important role in influencing channel characteristics and the presence of vegetation in both stream systems.
H53B-1250 1340h
Geomorphic Impacts of a Flood Control Reservoir on the Green River of Kentucky.
The Green River is a tributary of the Ohio River draining approximately 24,000 km$^{2}$ in south-central and western Kentucky. Green River is also one of the most biologically diverse waterways in the United States, supporting several threatened and endangered species. Green River Lake is a flood control reservoir receiving runoff from the upper 7% (1766 km$^{2}$ of the watershed. Reduction in flood peaks since the mid-1960's have eliminated overbank inundation and floodplain sediment deposition along much of the 197 km reach below the dam. The impact of flow regulation is greatest along the tailwater reach between Green River dam and Russell Creek, the first of four major surface tributaries entering along a reach extending from 47 to 90 km below the dam. In the tailwater reach, the reduction in flood peaks and fine sediment supply has altered patterns of in-channel sediment storage and mobilization. Widespread channel bed armoring has not occurred, presumably due to continued fine sediment input from small tributaries, floodplain gullying, and bank erosion. There is little evidence of encroachment of riparian vegetation and related channel narrowing in the tailwater reach. Inputs of water and sediment from the four large, unregulated tributaries modify the impacts of the reservoir on in-channel sediment dynamics downstream of Russell Creek. During intermediate magnitude events, particularly those associated with localized storms, flow and sediment dynamics on the Green R. are likely similar to those of the pre-regulated system. However, reduction in peak flows during large storm events has decreased sediment transport capacity and thereby altered seasonal patterns of fine sediment deposition and storage. In addition, the frequency and spatial extent of streambed gravel mobilization have been reduced. River islands provide most of the remaining areas for vertical accretion of fine sediment during flow events, although the dynamics of accretion and erosion of islands and island-bar complexes are also affected by flow regulation.
H53B-1251 1340h
Multi-scale Analysis of Sediment Yield in a Glaciated Environment
Global measurements of sediment yield in watersheds of different sizes have produced a general pattern. All other things being equal, large watersheds tend to have a smaller yield than small watersheds because there are more opportunities for sediment storage within a larger system. In the Little River valley near Stowe, VT, the influence of glacial geology and land use counteracts this trend. Some small watersheds, such as that of Ranch Brook (10 km2), West Branch (10 km2), and Stevenson Brook (15 km2), minor tributaries of the river, have sediment yields ranging from 5-30 tons/km2/year. In contrast, sediment yields measured through surveys of the Waterbury Reservoir (282 km2 watershed) on the Little River and the much smaller east branch of the reservoir (17 km2 watershed) have a yield of 200-220 tons/km2/year. (These numbers do not take reservoir trap efficiency into account and thus are minimum values.) The underlying causes of this bimodal distribution in sediment yield are the glacial history of the area and modern land use. During the Pleistocene, glaciers scoured away much of the regolith on the hillsides, leaving small modern streams with relatively little transportable sediment available. Thick deposits of glacial outwash were left in the valleys, so larger modern rivers such as the Little River, flow through poorly consolidated, highly erodible materials. Agriculture and development in the larger, relatively low-gradient valleys contribute to sediment production. The large sediment supply overwhelms the increased storage available in lower gradient areas and results in a substantially higher sediment yield compared to small, mountain streams in the area.
H53B-1252 1340h
Estimating Reservoir Sedimentation Rates: Long-Term Implications for California's Reservoirs
This paper presents a compilation and analysis of reservoir sedimentation rates applied to unsurveyed reservoirs in California. A number of reservoirs, primarily in the Coast Ranges, have already filled or are nearly filled with sediment and are being considered for removal, including the Matilija, San Clemente, Rindge and Searsville Dams. Of the over 1,400 dams in California listed in the National Inventory of Dams (NID), only 213 (15%) have been surveyed for reservoir sedimentation, presenting a significant gap in current knowledge. Using previously published reports on reservoir sedimentation rates (excluding flood control and debris flow basins), we determined the median reservoir sedimentation rates for the nine different geomorphic regions in California. For the geomorphic regions with sedimentation data, the median sedimentation rate +/- standard error in m3 km-2 yr-1, and (number of measured reservoirs) are: Siskiyou: 340 +/- 170 (3), Coast Range: 220 +/- 50 (49), Central Valley: 145 +/- 90 (5), Sierra Nevada: 95 +/- 25 (23), Transverse Range: 500 +/- 200 (18), and Peninsular Range: 260 +/- 185 (4). These reservoir sedimentation rates were then applied to the unsurveyed reservoirs. Of the approximately 42 million acre-feet of water storage in the state, approximately 5.14 million acre-feet may be currently occupied by sediment, representing a significant decrease of 12% of the state's water supply. Using the same rates to forecast the effects after 50-years, we found that the reservoirs with most risk of sedimentation are primarily small reservoirs ($<$2,500 acre-feet), such as municipal water-supply reservoirs, especially those operated by coastal towns and cities. Reservoirs in the Coast and Transverse Ranges are the most at risk, with high sedimentation rates, small reservoirs on large watersheds, and older reservoirs. There are over twenty reservoirs that will likely be full or nearly full in 50-years, including Tinemaha, Pleasant Valley, Lee Lake, Potrero, Hour House, Hansen, Century, and Benbow.