H51F-01 INVITED 08:05h
An Analytical Framework for Predicting the Downstream Geomorphic Effects of Dams on Rivers
Despite decades of research and abundant case studies on downstream effects of dams on rivers, we have few general models predicting how any particular river is likely to adjust following impoundment. We present a conceptual and analytical framework for predicting geomorphic response of rivers to dams, emphasizing the role of geologic setting and history as first-order controls on the trajectory of change. Basin geology influences watershed and channel processes through a hierarchical set of linkages, extending from the drainage basin to the valley and channel, which determine the sediment transport and discharge regimes. Geology also directly shapes the suite of hillslope processes, landforms, and geomorphic disturbances impinging on the channel and valley floor. These factors, in turn, affect the "lability" or capacity for adjustment of the downstream channel, determining the type, direction, and extent of channel adjustments that occur, including incision, widening, and textural changes. We develop an analytical framework, based on two dimensionless variables, which predicts geomorphic responses to dams depending on the ratio of sediment supply below to that above the dam (S*) and the fractional change in frequency of sediment-transporting flows (T*). Drawing on examples from rivers in the western United States and globally, we explore how trajectories of geomorphic change, as defined by these two variables, are predicted by this analytical framework and influenced by the geological setting and history of the river. This approach holds promise for predicting the magnitude and trend of downstream response to other dammed rivers, and can identify river systems where geological controls are likely to dominate.
http://www.fsl.orst.edu/wpg
H51F-02 08:25h
Understanding Geology; Understanding Rivers
Rivers reflect the geologic setting and history of their watersheds. This reflection results from two related aspects: (1) The strong controls that watershed geology exerts on the present-day environment or "regime" of a river, including basic factors such as slope, sediment, and water. (2) The effects of specific past events, such as landslides, floods, and glaciation, which also affect rivers, sometimes overwhelming environmental gradients established by regional geologic conditions. Deciphering the relative role of present day geologic conditions vis-a-vis the cumulative effects of a river's history is often key to understanding a particular river, especially at time scales relevant to understanding the affects of major perturbations to river systems, such as emplacing or removing dams. For western North American rivers (at least all the rivers we have worked on), which mostly flow through diverse and geologically active environments, determining the balance between the past and present is a complicated endeavor requiring capabilities beyond GIS mapping analysis and channel classification. Despite the required investments, such in-depth case studies are warranted, mostly because of the unique geologic history of every watershed makes each river a unique large-scale experiment, but also because of the significant natural and cultural resources at stake.
H51F-03 08:45h
Characterizing The Hydro-Geomorphic Effects of Impoundment: Reference Conditions, Approaches, and Long-Term Fluvial Adjustments
Predicting, or even understanding, the effects of impoundment requires knowledge of the pre-dam boundary conditions, including water and sediment discharge regimes, longitudinal and cross-sectional channel characteristics, and watershed disturbance history. To characterize pre-dam hydrologic regimes, we identified 21 gages across the U.S. from the National Inventory of Dams that met length-of-record criteria. Using a hydrologic model, the Indicators of Hydrologic Alteration, and GEV flood frequency analysis, we determined the pre- and post-dam hydrologic regimes for these sites. Results highlight changes in the extremes for high and low flows, and the importance of evaluating these effects over various flow durations. Other significant adjustments included changes in the annual hydrograph, primarily in the increased number of hydrograph reversals and significant shifts in the timing of high and low flows. Post-dam reductions in the bankfull discharge ranged from 20-95%, averaging 60%. To capture long-term geomorphic adjustments, we focus on the Upper Connecticut River watershed where numerous dams exist. For the at-a-station morphological adjustments, we used long-term field measured USGS data of channel hydraulic geometry spanning pre- and post-dam conditions. Significant adjustments occurred at cross-sectional scales, although not all of the streams responded similarly. The magnitude of hydrologic adjustment, dam type, and initial conditions control the magnitude and direction of change. The range of hydraulic and morphologic adjustments limits prescriptive management approaches as not all streams behaved similarly to impoundment, signaling the importance of disturbance history and boundary conditions. To identify the contemporary sediment discharge regime and impoundment's impact on embeddedness and sediment residence time, we use short-lived fallout radionuclides, especially, 7Be, to express event-scale sediment discharge. With this technique we are able to capture the effect of dam operation on the longitudinal transport and storage of sediment.
H51F-04 09:05h
The Regulation of Peace River: a Large-scale Experiment on Fluvial Governing Conditions
In 1967, British Columbia Hydro and Power Authority closed W.A.C.Bennett Dam, creating what was then the sixth largest hydropower project in the world. The dam is located in the Rocky Mountain front range so that, although it controls about half the runoff of the 293 000 sq.km basin, almost all of the sediment load originates downstream from the dam in the Alberta Plateau. Hence, the effects of these two principal governing conditions of fluvial systems can be separated. The 378 km immediately downstream to the Smoky River confluence are a wandering, cobble-gravel reach It has effectively ceased to be alluvial and the channel pattern has been simplified. Aggradation is occurring at major tributary junctions, whilst the tributaries themselves have degraded in their lowermost reaches. Smoky River, the principal tributary, delivers a large sand load. The 250 km reach to Carcajou is sandy gravel and the final 600 km to the Peace-Athabasca delta is sand-bed. Aggradation, with a change in fluvial style toward low-order braiding, appears to be underway in the proximal sand-bed reach. More generally, channel shrinkage in response to the regulated flow regime is controlled by the rate of progradation of riparian vegetation onto former bar surfaces In 1996, after 29 years of regulated flow, reservoir drawdown for dam repairs led to full spillway flows for 8 consecutive weeks, creating an effectively bankfull condition in the proximal post-regulation channel. Significant degradation was observed for the first time in many cross-sections but overall changes were surprisingly modest, reflecting the refractory bed and the degree to which riparian vegetation has become firmly established in former channel areas. Overall, sediment supply and flow competence are the principal controls of fluvial response in the system. The experimental aspect of this study of a large, northward flowing, boreal river can be controlled by before-after comparison. However, this strategy must take into account a changing hydroclimate which has seen increased precipitation but decreased winter snowfall, the latter being the chief source of runoff. However, it can also be pursued by comparison with Liard River, of comparable scale and morphology, located to the north. In both rivers, winter ice regime represents a significant additional dimension for study.
H51F-05 09:20h
The Effects of Urbanization and Flood Control on Sediment Discharge of a Southern California River, Evidence of a Dilution Effect
The southern California landscape has undergone dramatic urbanization and population growth during the past 60 years and currently supports almost 20 million inhabitants. During this time, rivers of the region have been altered with damming, channel straightening and hardening, and water transfer engineering. These changes have drastically altered water and sediment discharge from most of the region's drainage basins. Here we focus on changes in sediment discharge from the largest watershed of southern California, the Santa Ana River. Order-of-magnitude drops in the suspended sediment rating curves (the relationship between suspended sediment concentration and instantaneous river discharge) are observed between 1967 and 2001, long after the construction of a major flood control dam in 1941. These sediment concentration decreases do not, however, represent alteration of the total sediment flux from the basin (a common interpretation of sediment rating curves), but rather a dilution of suspended sediment by increases (approx. 4x) in stormwater discharge associated with urbanization. Increases in peak and total stormwater discharge are consistent with runoff patterns from urbanizing landscapes, supporting our hypothesis that the diluting water originated from stormwater runoff generated in urban areas both up- and downstream of dams. Our dilution hypothesis is further supported with water and sediment budgets, dilution calculations, and suspended and bed grain size information.
H51F-06 09:35h
High Suspended Sediment Yields of the Conestoga River Watershed to the Susquehanna River and Chesapeake Bay are the Result of Ubiquitous Post-Settlement Mill Dams
The Conestoga River watershed of Lancaster County, Pennsylvania constitutes 1.7% of the total area of the Susquehanna River basin and drains mostly Piedmont Province with very low relief and hillslope gradients, yet it contributes the largest annual suspended sediment yield of any tributary in the basin. Nearly one-third of the Susquehanna's sediment is deposited in the northern Chesapeake Bay, causing significant impairment to water quality. Poor farming practices (60% agricultural land) and late 20th Century suburban sprawl have been blamed for the anomalously high sediment yields observed in the Conestoga watershed. Our study indicates, however, that the main cause of these high sediment yields is sediment trapped behind ~500 post-settlement mill dams. On average, there was one mill dam every 2.5 km of stream length. We calculate that a minimum of 27 x 10$^{6}$ m$^{3}$ of post-settlement legacy sediment was stored in Conestoga watershed valleys. This estimate is based on field mapping and coring and on analysis of historical records, air photos, and maps. We compiled a GIS database for all known mill dams and constructed stream profiles showing the location and height of each dam. Average dam height was 2.4 m (range 1-9 m) and average stream gradient is 0.001, so the average reservoir extended ~2.4 km upstream. Field measurements show that the average stream valley bottom is 100-m wide and coring reveals a broad, planar bedrock valley floor beneath post-settlement alluvium. The legacy sediments overly a thin veneer of organic-rich sediments (peats and leaf mats) and tree stumps (some cut), which yield late 17th century and older $^{14}$C ages. These organic layers overly ca. 20 cm of pebbly-sands, which together represent the pre-settlement to early-settlement valley floor. From these findings, we calculate that the average mill reservoir could trap 0.3 x 10$^{6}$ m$^{3}$ of sediment. Field surveys and air photographs taken in the 1930s and 40s reveal that all reservoirs were filled to capacity with sediment. Numerous factors contributed to the storage of such large volumes of sand, silt, and clay along Conestoga stream corridors: 1) parent material of thick saprolites developed on Paleozoic silty limestone; 2) widespread soil erosion with land clearing for agriculture from ~1700 to 1800 AD; 3) intense deforestation and industrial charcoaling from ~1850-1910; and 4) the advent of the mechanized plow in the late 1800s that initiated widespread gullying and associated downslope deposition of sediment. Pennsylvania leads the nation in the removal of low-head dams. Once these dams are removed, however, the streams quickly incise through the stored mill pond sediments to the level of their former valley floors, reaching gravels that are eroded more easily and undercutting steep banks of finer-grained, slightly cohesive legacy sediments. Bank erosion occurs along at least 80% of the 1036 km of stream channels in the watershed. We estimate that 10% of the sediment stored along valley floors since 1710 has been removed in the past several decades by channel incision and widening that closely resemble arroyo-cutting (lateral bank erosion rates of $>$0.5 m/yr measured at multiple sites). The large volume of legacy sediment has become a major source of suspended sediment load to the Susquehanna River and Chesapeake Bay in the past 35 yrs, and will remain so unless substantial remediation efforts are made.
H51F-07 09:50h
Reconstructing Watershed History from Reservoir Stratigraphy: Englebright Lake, Yuba River, Northern California
Reservoirs provide the opportunity to study fluvial processes and rates in a controlled setting because they are effective traps of sediment and are often well monitored. An extensive sediment coring and sampling campaign was done in Englebright Lake on the Yuba River in northern California as part of a fish-habitat restoration study. The Yuba watershed (particularly the southern part) was the site of intensive hydraulic gold mining in the 19th and early 20th century, and Englebright Dam was built in 1940 to trap mining debris. Results of a bathymetric survey in 2001 indicate that the reservoir was 26% full (22x10$^{6}$ m$^{3}$ of material). The physical properties of the entire deposit were extrapolated from $\sim$300 m of cores collected at 7 sites along the longitudinal axis of the reservoir in 2002. The mass of the deposit is 26x10$^{6}$ metric tons, of which 3.2% is organic. The sediment is $\sim$65% sand and gravel, and distinct layers of differing grain size (sand-gravel, silt-clay, organics) are well preserved in the cores. The depositional chronology of the reservoir was established using $^{137}$Cs analysis and the relations between the cored stratigraphy and the hydrologic and impoundment history of the watershed. Deposits from three major flood events (1955, 1964, 1997; each with discharge $>$3,400 m$^{3}$/s) were identified in the stratigraphy of most of the coring sites. Observations of recent (post-1997) depositional patterns are guiding the development of a conceptual model of reservoir-sedimentation processes during floods, drawdowns, and intraflood periods. Enlargement of an upstream dam on the North Yuba River in 1970 caused a decrease in flood frequency in the Yuba River and changed management of Englebright Lake (ending annual drawdowns). A relict topset-foreset-bottomset sequence observed in the cored stratigraphy is interpreted to correlate with this change in watershed management; a second deltaic sequence was deposited on top of the first after 1970. Post-1970 average annual deposition rates appear to be $\sim$25% lower than pre-1970 rates. Although this change is small, it does appear to be robust (based on two methods for extrapolating cored sediment properties throughout the reservoir deposit: assumptions of constant and variable layer thickness). The decline in sediment deposition rates is expected for several reasons including (1) winnowing of Gold Rush-era mining debris; (2) fewer large flood events during the period; and (3) changed watershed management including construction of dams on tributaries.