8256 - Laserfiche WebLink

Home Browse Search

about 30 cm less if the channel were 60 m wider, and 60 cm less if it were 120 m wider. A model also can provide valuable guidance for field sampling programs in indicating what pa- rameters are of greatest significance for long-tenn response, and in indicating the particular river reaches in which relatively rapid or severe re- Whenever river banks are erodible, the river's sponse justifies focused data-collection efforls. response to perturbations can include both bed and For the river engineer charged with proposing a bank adjustments. Most models, such as the ones small number of alternative development scenar- described here, consider only bed response, implic- ios to meet flXed objectives, a numerical model can itly assuming that the banks are stabilized or oth- be used to eliminate the obviously infeasible solu- erwise prevented from movement (this is often the tions, so that further field and physical-modelling case in developed rivers.) Although some of the , efforls can be focused on those alternatives that modelling techniques outlined by Fan (1988) in- seem to be the most viable ones. clude explicit attempts to allocate river response But given the simulation deficiencies described between planform (bank-erosion, meandering) and above, does mobile-bed modelling have an impor- bed (aggradation, degradation) changes, most do tantroletoplaYinriver-developmentplanningand not. Yet this is clearly an important element of design? In our view, the answer is a strong but channel response to perturbations. qualified yes. Numerical mobile-bed modelling must not be viewed as being a complete or definitive tool'on which development should be based. It is, rather, to be included with field experience, field observations, analytic calculations, study of pub- lished experiences, and possibly physical modelling to support the engineer in what is ultimately his or her responsibility alone-to make the best possible design decisions with regard to present and future river and ecosystem responses. The river engineer must not abdicate decision responsibility to anyone of these tools. The processes are so complex, and the time scale of river response is so long, that it is incumbent upon the river engineer to base deci- sions on all possible input. The unique contribution of numerical mobile- bed modelling is that it enables the river engineer to acquire an intuitive feel for and understanding of the way a river system is apt to respond to changes over the long tenn. As an inexpensive and readily available tool, numerical modelling gives the engineer the opportunity to test a range of "what if' scenarios on the data set that emulates the real river. I I r 424 BIOLOGICAL REPoRT 19 degradation in nonequilibrium. river response, as is the case for the Missouri River example dis- cussed earlier. Planform Geometry Adjustments Simulation Opportunities In addition to overall degradation and aggrada- tion, mobile-bed models provide insight into per- turbed-equilibrium effects on many habitat fea- tures of importance to the ecosystem, such as: water depths and velocities, water turbidity and solids content, water temperature (when thennal processes are included), ice-cover fonnation, ac- tual bedload and suspended-load transport rates, and bed-surface sediment composition. Using models that include planfonn adjustment (bank erosion and bar deposition) in their re- sponse, it is also possible to study the habitat effects of releasing stabilized banks in certain ar- eas to achieve a natural widening, and to reconcile this response with navigation and flood-control needs. The absolute predictions of bed-level change, turbidity increase, bed coarsening, and so forth must be viewed as suspect because of all the un- certainties described earlier. But the sensitivity of river response to uncertain factors, such as the sediment transport rate, grain-size distribution, bank erodibility, and so forth, can be studied with considerable confidence in such models. For exam- ple, in Fig. 4 one might be reluctant to bank on the Sioux City degradation in the year 2000 being within a few centimeters of the predicted value for run S1. But one might have considerably more confidence in saying that the degradation would be Summary and Conclusions It has been our purpose in this paper to temper enthusiasm for the predictive potential of mobile- bed numerical models with a sober realization of the fragile understanding of mobile-bed processes on which they are based. Issues of spatial heteroge- neity, stochastic properties of turbulence, cohesive