7853 - Laserfiche WebLink

Home Browse Search

552 Ecological Applications Vol. 4, No.3 Without the model, one might have predicted that di- version would cause a shift toward drier cover types, but the relative magnitude of the shift would have been unknown. More importantly, some of the predictions of the model are contrary to the conclusions a more cursory analysis might reach. A decrease in the vari- ability of flow results in an increase in the relative abundance of the extreme cover types, and an increase in the minimum allowable flow results in a decrease in the area occupied by the wettest vegetated cover type. Two general considerations are underscored by this analysis. First, it is possible to cause substantial changes in riparian vegetation without changing mean annual flow, as with the Moving-Average alternative. Second, special attention should be given to alternatives in- volving changes in flow boundaries. This includes both effectively implemented minimum flows and substan- tial reductions in peak flows. There are several reasons for being concerned about these hydrologic changes with respect to riparian vegetation. Changing the boundaries of the flow distribution creates new zones that are either always inundated or never inundated. Cover types in these zones (e.g., Open-Water) are gen- erally the most structurally distinct from other cover types and exhibit little temporal fluctuation because of the constant conditions. In addition, establishment and mortality of riparian vegetation are episodic. Changing the boundary flows may eliminate the infrequent es- tablishment opportunities in areas that are highly suit- able for survival. Thus conflicts can arise between dif- ferent goals of river management. The occasional high flows necessary for maintenance of some riparian plant species could damage property. The occasional low flows necessary for maintenance of other riparian spe- cies could decrease fish habitat. Models help make it possible to find a workable compromise. GREGOR T. AUBLE ET AL. Our study reach is a bedrock-controlled, steep-walled canyon cut into hard pre-Cambrian rock (Hansen 1987). The canyon walls preclude lateral migration or channel widening. The channel bed consists of cobbles and boulders over bedrock. The vegetated bars consist of cobbles and boulders sometimes covered by a layer of finer particles. The main sources of sediment to the Monument, both before and after dam construction, have been rock falls and debris flows within the canyon (Hansen 1987). Dam operation has moderated peak flows, preventing the stream from transporting the larg- er particles (Chase 1992). Therefore, the bed is rela- tively stable under current conditions and would re- main so under the conditions of decreased peak flows addressed in the alternatives. A decrease in peak flows could result in channel narrowing by deposition of a new surface adjacent to the channel. However, such a surface would necessarily be small. Finally, under any of these alternatives a bar could be obliterated by a rock fall or debris flow, but these events will not be influenced by the alternatives under consideration. Quasi-equilibrium vegetation The direct gradient method assumes that the sam- pled vegetation and defined cover types are an ade- quate quasi-equilibrium expression of the longer term hydrologic regime. It does not require that the vege- tation be constant, but it does require that variation over time does not obscure the relationship between vegetation and inundation duration. Conditions at the time of sampling on the Gunnison River were dry rel- ative to the overall range of conditions in the 1971- 1989 hydrologic record. Because calibration occurred during a period oflow flow, the model tends to place cover types at incorrectly high inundation durations. The magnitude of the error depends on the flow con- ditions prior to the time in question. This source of error does not affect a relative analysis of alternatives like the current study. However, caution should be exercised when using the model to predict the precise vegetation composition at a specific time. The predictions of our model have not yet been test- ed. The best way to validate the direct gradient method would be to repeat the plot sampling following a flow alteration. Sampling would have to be delayed long enough to allow the new flow regime to be expressed and to allow the vegetation to respond. The delay could be as short as 2 yr or as long as several decades, de- pending on the nature of the flow regime and the re- sponse time of the relevant vegetation. In the case of the three flow alterations considered here and their impacts on the largely herbaceous vegetation at our study site, a delay of roughly 10 yr might be necessary before validation could occur. IMPLICATIONS FOR MANAGEMENT The direct gradient method provides new, useful in- formation about the effects of alternative flow regimes. t ACKNOWLEDGMENTS We are grateful to C. Pettee, K. Chase, and J. Albright of the National Park Service for calibrating and running the hydraulic model. W. Weber identified some of the plant spe- cies. J. Roelle, J. Stromberg, J. Zedler, and an anonymous reviewer provided extremely helpful comments. This manu- script was prepared by employees of the National Biological Survey as part oftheir official duties and, therefore, may not be copyrighted. LITERATURE CITED Bedinger, M. S. 1971. Forest species as indicators of flooding in the lower White River valley, Arkansas. United States Geological Survey Professional Paper 750-C:248-253. -. 1979. Forests and flooding with special reference to the White River and Ouachita River basins, Arkansas. United States Geological Survey, Water-Resources Inves- tigations Open File Report 79-68. Bovee, K. D. 1982. A guide to stream habitat analysis using the Instream Flow Incremental Methodology. Instream Flow Information Paper 12. United States Department of the Interior, Fish and Wildlife Service, Office of Biological Ser- vices FWS/OBS-82126. Bovee, K. D., J. Gore, and A. Silverman. 1978. Field testing