8256 - Laserfiche WebLink

Home Browse Search

I I 1 '1 fi 418 BIOLOGICAL REroRr 19 finer than that found on the bed-to enter the river. Such material will generally remain in sus- pension, traveling downstream at the water veloc- ity, until it reaches a point where the prevailing shear stress is low enough that it settles persist- ently to the bed and begins to interact with it. This transport mechanism, called suspended-load transport, can also occur for sediments eroded from the bed in an area of high shear stress; the significance of it is that sediment particles travel downstream essentially at the same speed as the water, that is, at the order of 1 m/s. By contrast, bedload transport, by which material travelljl downstream very near the bed by rolling, sliding, and making short hops in response to turbulent flow fluctuations, carries sediment downstream very slowly in comparison with the speed of water movement, at speeds of the order of several kilo- _ meters per year. Indeed, the essential mechanism for downstream movement of bedload is intermit- tent local deposition, whereby increased velocity causes the shear stress to increase sufficiently to re-entrain the sediment and move it on down a bit farther, and so on. Anyone who has walked along exposed portions of a river bed knows that there can be a remarkable heterogeneity of bed sediment in a particular loca- tion. Yet Fig. 1 indicates that there is a single slope, a single sediment load and size, and so forth. A key feature of a river's nonequilibrium response to per- turbations can be spatial redistribution of bed sedi- ment, bedforms, and even subchannels within the overall waterway in a braided system. The non- equilibrium response of meandering channels is typically characterized by spatial and temporal re- distribution of geomorphic features. Thus, while the essential features of a river's response to perturbed equilibrium are conceptu- ally summarized in Fig. 1, there are many other subtle features of the response that are poorly understood at best, and in any case not amenable to direct analysis. One example is the dependence of viscosity on temperature, which leads to major changes in suspended-sediment transport, as well as bedform change and consequent dramatic rat- ing-curve shifts on the Missouri River. The goal of numerical simulation of mobile-bed processes is to provide effective prediction of nonequilibrium river response by bringing together the best pos- sible conceptual mathematical models of the im- portant processes, and solving the resulting sys- tems of equations using appropriate numerical methods. Basis for Computer Simulation Many investigators are involved in developing mobile-bed computational models. Such models can differ at several levels, from. the gross to the subtle. The grossest distinction is in dimensional- ity. Nature is three-dimensional (3-D), and ideally one should construct mobile-bed simulation mod- els on the basis of a full 3- D description of the water and sediment processes. Until recently, however, the sheer computer-resource demands have pre- cluded serious development of 3-D mobile-bed simulation codes. At present, the Iowa Institute of Hydraulic Research is working with the U.S. Army Corps of Engineers Waterways Experiment Sta- tion to implement mobile-bed processes in the ex- isting Computational Hydraulics 3- Dimensional (CH3D) 3-D fixed-bed hydrodynamic code fornatu- ral waterways. Two-dimensional (2-D) simulation can be fo- cused on width-averaged processes, or depth-av- eraged ones. Width-averaged analysis is often used to study mechanisms of sediment exchange between the bed and the water column, see for example van Rijn (1984). The depth-averaged (plan-view) simulation generally used for analysis of natural waterways and reservoirs is often jus- tified when stratification effects are weak. A depth-averaged representation allows the spatial heterogeneity of bed sediment and geometry to be recognized, but at the price of significant demands on computer resources. The U.S. Army Corps of Engineers TABS-2 system (Thomas and McAnally 1985) has been in active use in the last decade. This model is limited to a single sediment size class and only the suspended-load mode of sedi- ment transport. The MOBED2 system, recently developed at the Iowa Institute of Hydraulic Re- search by Spasojevic (Spasojevic and Holly 1990), treats suspended-load and bedload transport of sediment mixtures, and has been successfully used for simulation of reservoir sedimentation. One-dimensional (I-D) models deal with cross- sectional average channel and flow properties, in- cluding bed sediment. They are incapable ofrecog- nizing the spatial heterogeneity of bed sediments, local channelization, and so forth. But to the extent that they enable the river engineer to study the response of a simple 1-D system having certain macroscopic properties in common with the real river under study, they can be useful tools within their limits of applicability, especially because