USING THE SACRAMENTO SOIL MOISTURE ACCOUNTING MODEL TO IMPROVE FLOOD FREQUENCY ESTIMATES FOR DAM SAFETY

摘要

The Bureau of Reclamation uses multiple methods to develop hydrologic loadings fordam safety risk analysis. Hydrologic hazard estimates typically consist of peak flowfrequency, volume frequency, and flood hydrographs for a full range of AnnualExceedance Probabilities (AEPs). The methods used to estimate hydrologic hazardinclude deterministic and stochastic modeling approaches often requiring rainfall-runoffmodels. The use of rainfall-runoff modeling introduces potential for error and increasedproject cost in model development and calibration. Reclamation has begun to use theSacramento Soil Moisture Accounting (SAC-SMA) model in drainage basins alreadybeing modeled by the National Weather Service (NWS) to improve hydrologic hazardestimates while reducing model calibration, setup, and project cost.SAC-SMA is a continuous soil moisture accounting model with spatially lumpedparameters. The model is ideal for large drainage basins and uses multiple years ofrecords for calibration. There are many advantages in using the SAC-SMA model forhydrologic hazard estimates, including: the model is calibrated to actual basinparameters, snowmelt is also modeled, precipitation is a unique input and can be easilymanipulated, and the model can be automated to run multiple iterations.A case study of East Park Dam, California is presented to demonstrate how the SACSMAmodel was used to make hydrologic hazard estimates for dam safety. A detailedhydrologic hazard study was conducted for East Park Dam, California using multiplemethods. One of the methods used was a custom rainfall-runoff modeling method withL-moments precipitation and the SAC-SMA model; the second method was peak-flowfrequency with site-specific paleoflood data collection. The study used the SAC-SMAmodel in a quasi Monte Carlo type approach to calculate runoff for a large range ofhydrologic conditions and frequency rainfall. Code written in the Python programminglanguage was used to automate the model runs allowing several thousand iterations of themodel. This approach resulted in improved estimates for the hydrologic hazard at EastPark Dam from preliminary studies, and increased potential for use of the SAC-SMAmodel in future studies.