Large Rivers II: Processes and Methods
Presiding: J Syvitski, University of Colorado; W Dietrich, University of California, Berkeley
H24A-01 15:30h
River Response to Post-Glacial Sea Level Rise: The Fly-Strickland River System, Papua New Guinea
The most recent deglaciation resulted in a global sea level rise of some 120 m over approximately 12000 years. A moving boundary numerical model is developed to predict the response of rivers to this rise. The model was motivated by experiments at small-scale, which have identified two modes describing the transgression of a river mouth: autoretreat without abandonment of the river delta (no sediment starvation at the topset-foreset break) and sediment-starved autoretreat with abandonment of the delta. In the latter case transgression is far more rapid, and its effects are felt much farther upstream of the river mouth. The moving boundary numerical model is checked against experiments, and then adapted to describe the response of the Fly-Strickland River system, Papua New Guinea. In the absence of better information, the model is applied to the case of sea level rise without local climate change in New Guinea. The model suggests that a) sea level rise forced the river mouth to transgress over 700 km since the last glacial maximum, b) sediment-starved autoretreat forced enough bed aggradation to block a tributary with a low sediment load and create the present-day Lake Murray, c) the resulting aggradation was sufficient to move the gravel-sand transition on the Strickland River upstream, d) the present-day Fly Estuary is in part a relict river valley drowned by sea level rise and partially filled by tidal effects, and e) the Fly River is presently reforming its bankfull geometry and prograding into the Fly Estuary. A parametric study with the model indicates that sediment concentration during floods plays a key role in determining whether or not, and to what extent, transgression is expressed in terms of sediment-starved autoretreat. A sufficiently high sediment concentration can prevent sediment-starved autoretreat during the entire sea level cycle. This observation may explain why some present-day river mouths are expressed in terms of deltas protruding into the sea, and others are wholly contained within embayments or estuaries in which water has invaded landward.
H24A-02 15:45h
Late Cenozoic Response of the Susquehanna River to Climatic and Base Level Forcing
The Susquehanna, the largest river basin draining the Appalachians, has been studied geomorphically since the 19th century and has been the subject of major geomorphic models. New cosmogenic 10Be exposure data and model erosion rates raise important challenges to both the geographic cycle and dynamic equilibrium models, and provide constraints important to future modeling efforts. The Susquehanna River basin exhibits rapid erosion rates at a variety of timescales. Cosmogenic 10Be exposure ages and model erosion rates provide evidence that: Bedrock gorge incision approached 1m/ky during the last glacial maximum, probably due to periods of high discharge and lowered base- level relative to the Holocene, and total incision of >20m occurred in a geologically short interval; hill slope response to channel incision following Miocene uplift, base level fall and/or stream capture is a continuing process in all three major physiographic provinces (the Plateau, Valley and Ridge and Piedmont), and erosion delivers sediment from slopes irrespective of lithology at rates proportional to relief; and climate-related slope deposits, such as fans and debris flows, do not store sediment in sufficient quantities at the 100 ky timescale to mask the isotopic signal of the delivery from slopes to channels. It is unlikely that the bedrock gorge incision in the Piedmont during the most recent glacial maximum (~30 ka to 10 ka) was a unique Pleistocene event. Thus the disturbance of stream profiles by headcutting has probably proceeded continuously or in climate-related intervals over at least the last 2 Ma. The disturbance and on-going adjustment of stream profiles may have a much longer history in this basin as shown by Pazzaglia and Gardner's (1994) flexure model. Data from terrace deposits and offshore stratigraphy show that the river has responded to forcing at a range of timescales. Sediment discharge increased dramatically during the Miocene, possibly in association with major stream capture and divide migration north and west of the Piedmont. Taken together, these results show that various disequilibrium processes have affected this passive margin river over the past 20 Ma. Relief has probably increased over this time. Thus, dynamic equilibrium is not a valid explanation for the geomorphic relations of valleys and ridges over this time interval. While it is possible that the pre-Miocene landscape had more subdued relief due to a long decay following the late Paleozoic orogenic climax, the rapid response of the present river to Pleistocene forcing raises doubts about the concept of a geomorphic decay cycle lasting tens of millions of years. As shown by the Susquehanna data, geomorphic investigations using cosmogenic isotopes can provide important constraints on processes operating at the millanial to million year timescale.
H24A-03 16:00h
Minimizing the Error Associated With Measurements of Migration-Related Sediment Exchange on Meandering Rivers
The floodplains of meandering rivers represent reservoirs that both store and release sediment. Bed material is generally released from cut banks and replaced in nearby point bars wherever migration occurs. Measuring the associated bed material flux is important for tracing the movement of contaminants that may be mixed with the bed material. Approximations of this flux can be made using a representative channel depth and sequences of aerial photography to estimate average absolute migration rates (or reworked areas) between photographs. Error in the aerial photographs leads to a positive bias in computed release rates. A method for removing this bias is introduced that uses the apparent offset of fixed linear features such as roads (along smaller rivers) or abandoned channel courses (along larger rivers). Measuring the rate of release of fine sediment is important both for predicting the long term morphodynamic evolution of the channel/floodplain system and for tracing the movement of contaminants that may be adsorbed to the fine sediment. While fine sediment can be mixed throughout the depth of the floodplain, it is most concentrated in the upper portion of older parts of the floodplain where it has had time to accumulate through overbank deposition. Its release rate can be estimated using migration rates computed from aerial photography in combination with local measurements of bank topography, both of which are highly variable even within a given reach. Where detailed bank topography is available for an entire reach, estimating the release of fine sediment is relatively straightforward. However, detailed topography is often unavailable along the banks of large lowland rivers, forcing estimates of the fine material flux to be made using a relatively small number of physically surveyed cross-sections. It is not immediately clear how many cross sections are required for a good estimate. This study performs Monte Carlo simulations on a detailed topographic dataset obtained using LiDAR along a 91 km reach of the Pearl River, near New Orleans, Louisiana, to determine the error introduced by sampling only a small number of cross sections. Estimates of both the bed material flux and fine material flux associated with migration are presented for this reach.
H24A-04 16:15h
Sediment accumulation determined with 210Pb geochronology and geochemical tracers for Strickland River flood plains, Papua New Guinea
The Strickland River is the primary sediment source for the Fly River system, a large tropical river that ranks in the global top 20 for both water and sediment discharge. As part of a "Source to Sink" NSF Margins program, the patterns and rates of floodplain sedimentation are being investigated. Previous research on the Middle Fly has documented an exponential decrease in sedimentation rates with distance from channel bank and a large influence of distributary floodplain channels in directing sediment to the floodplain environment. In the Strickland, a mine has discharged waste into the river since 1992, and though the total load increase is small for the lowland Strickland, elevated Ag and Pb levels occur in the river sediment, providing a clear environmental tracer across the floodplain. Work on other flood plain environments has demonstrated that 210Pb can be used to map the spatial and temporal patterns of sedimentation. Here we present geochronological results from an intensive floodplain coring campaign conducted in 2003 on the lower Strickland, which employed both 210Pb geochronology and Ag and Pb penetration depths to quantify sedimentation rates. We will first outline our procedure for dating Strickland sediment with 210Pb geochronology and summarize some early results from 36 cores. Flood plain accumulation rates appear to be highest upstream near the gravel-sand transition, low in the middle portion of the river, and higher again in the lower reaches of the Strickland near to its confluence with the Fly River. Overall patterns of sedimentation from 210Pb geochronology seem to be spatially consistent, for series of cores collected along single flood plain transects. We will next compare these results to accumulation rates determined from duplicate cores that were measured for the concentration of heavy metals from the upstream mine. These two techniques are independent and cover different temporal and spatial (in the vertical dimension) scales, so we will outline how they agree and key scientific questions that are posed by any differences. We will conclude with a summary of the rates and spatial patterns of sediment accumulation across Strickland River flood plains.
H24A-05 16:30h
Quantitative Measurements of Bedform Transport Rates and Sand Sheet Character in the Lower Mississippi River
Channel sand volume and downstream flux in the Mississippi River have important implications for proposed mitigation projects (dredging and pipelines) that seek to utilize this resource for replenishing neighboring barrier islands and restoring Louisiana's deteriorating wetlands. This study quantifies bedform migration-induced sand flux through the lower river on daily and seasonal timescales, and evaluates the sedimentary character of the bedload component. Observations and measurements were conducted along three study grids (Audubon Park, English Turn and Venice) over a range of river discharges between April 2003 and January 2005. Two multibeam bathymetric profiles of the study grids were conducted 24 h apart to document bedform migration, and stratigraphy and thickness of the sand layer were confirmed by CHIRP seismic profiling. Downstream transport is evaluated from bed elevation changes for a 1 m grid after correcting for river stage, and utilized to calculate bedload sand fluxes for larger, averaged grid cells after visual examination confirmed dunes had migrated <1 wavelength. Algorithms were formulated to remove spurious grid cells created by vessel motion, navigation and swath-matching errors. Initial data analysis indicates flux rates conform to expected trends: values are proportional to river discharge and are higher in the channel thalweg of straight reaches relative to shallower water. Bedform size also increases with river discharge and spatial changes in flux rates; height ranges from <1 m to 10 m, and wavelength from 10 m to 100 m. Seasonal trends in sand sheet thickness are evident, particularly in deeper meander reaches, where aggradation occurs at low flow and scour is observed during high flow. At highest discharges observed (35,000 m3/sec), bedform troughs bottom out on exposed relict fluvio-deltaic strata that the river has incised (i.e., sediment starved). A spatially uniform grab sampling effort (250 samples) provided grain size data of the active sand sheet for the lower 135 km of the river. A downstream trend of decreasing grain size (2.6 to 2.0 φ) is likely due to the finest sand fraction leaving suspension and becoming bedload as river gradient and velocity progressively decrease. Locally, grain size increases in the deepest portions of the channel as a result of enhanced flow scouring the fine and medium sand fraction.
H24A-06 16:45h
Up-scaling Dune Morphodynamic Models for the Study of Bedload Transport in Large Rivers
This paper presents preliminary results from a first attempt to give a channel-scale perspective to models of sediment transport over dunes, thereafter integrating effects at all bedform scales. Recent morphodynamic models have demonstrated the relationships between composition and rate of sediment transport, and the probability distribution of bed elevation relative to mean bed level (Ps) over a train of dunes. These models are yet essentially theoretical and there exists only limited data to implement them. Lately, the analysis of Ps curves for new data from the Mississippi River revealed the potential interest of this approach for assessing, although mostly qualitatively for the moment, the sediment transport stage and the erosion-deposition pattern. Just as bed elevations vary over dune backs, crests, and troughs, bed elevations over a meander planform vary in the alongstream direction from bend scours to point-bar tops. The range and shape of channel-scale Ps curves, and their time variation, thus can provide information on bedload-transport conditions, especially in large rivers where it is difficult to obtain. The survey line in this study is located just upstream of Baton Rouge in Louisiana and is nearly 50-km long, hence comprising several meander bends and low-sinuosity reaches of the Mississippi River. Two surveys were conducted during different high-stage events. The bathymetry was measured with a high-precision echo sounder and a 1MHz transducer. Data on flow conditions are those reported by the USACE. The Ps curves compare well with those from experimental data sets at smaller scale. This study shows that depth-continuous models for sediment continuity are promising for scale-integrated analysis of sediment transport in large rivers.