Application of Bureau of Reclamation's GSTARS To Lake Mescalero Dam Spillway

Abstract

When there is a significant change in discharge or sediment load, due to flood, for example, erosion and deposition cause the geometry of an ,alluvial channel to change rapidly in order to achieve a new equilibrium condition. Because the process is a highly complex multi-component, multi-dimensional, and time-dependent process, its complete mathematical modeling is not feasible at the present time. In order to predict the variations of channel width, as well as the river bed elevation, it is necessary to distinguish the sedimentation at the bank from that of the bed. The Bureau of Reclamation GST ARS model adopts a new concept which combines the conventional one-dimensional flow model with the theory of minimum energy dissipation rate such that the variation in channel shape can be calculated efficiently. A detailed description of the original GSTARS model is given in the User's Manual of the Bureau of Reclamation [1]. Because the GSTARS model is relatively new, the Bureau of Reclamation decided to carry out some case studies in order to verify and improve the model. The first case study, the erosion of the unprotected spillway of Willow Creek Dam in Montana due to hypothetical floods, was completed in April, 1987 [2]. Because there were no erosion data due to the hypothetical floods, the accuracy of the results could only be judged qualitatively and by comparing with other estimates. The results appeared quite reasonable and encouraging. The actual channel erosion of Mescalero Dam emergency spillway in New Mexico due to the 1984 flood has been selected as the subject of the second case study. The topographic data taken on June 12, 1979 was used as the initial condition. The water level data taken during the 1984 flood was used to calculate the flood hydrograph. In addition, sediment samples were taken and analyzed and used as input data. Four channel cross sections measured immediately after the 1984 floods were used to determine. the accuracy of the model output. During the course of the second case study; it was necessary to add three new features to the GSTARS model. The three new features are, in order of importance, (1) the bank slope stability constraint, (2) the erosion due to water fall at the upstream end, and (3) the lateral sediment transport due to secondary current in a bend. With these additional features, the modeled results agreed very well with field data, except at a station where a branch channel meets the main channel. The present GST ARS model is not capable of modeling channel junctions.