Automated Calibration of the GSSHA Watershed Model: A Look at Accuracy and Viability for Routine Hydrologic Modeling
The goal of hydrologic models is to accurately predict a future event of a given magnitude. Historic data are often used to calibrate models to increase their ability to forecast accurately. The GSSHA model is a distributed model that uses physical parameters and physics based computations to compute water flow from cell to cell based on a 2 dimensional grid. The goal of calibration is to obtain good estimates for the actual parameters of the watershed. These parameters should then transfer to other storm events of different magnitudes more easily than an empirical model. In conducting this research three watersheds were selected in different parts of the United States and the required data were collected to develop and run single event hydrologic models. The WMS software was used to preprocess digital spatial data for model creation before calibrating them with the GSSHA model. A calibrated HEC-HMS model was also developed for each watershed for comparative purposes. Establishing GSSHA's usability in routine hydrologic modeling is the primary objective of this research. This has been accomplished by developing guidelines for GSSHA calibrations, assisted by WMS, testing model accuracy in the calibration and verification phases, and comparing results with HEC-HMS, a model widely accepted for routine hydrologic modeling. As a result of this research, the WMS interface has become well equipped to set up and run GSSHA model calibrations. The focus has been on single event, or routine hydrologic model simulations, but continuous simulation calibrations, an important strength of GSSHA, can also be developed. Each of the model simulations in the study calibrated well in terms of matching peak and volume. However, the verification for two out of the three watersheds used in the study was less than ideal. The results of this research indicate that the physical factors, which GSSHA should represent well, are particularly sensitive for single event storms. The use of calibration of single events is therefore difficult in some cases and may not be recommended. Further research could be done to establish guidelines for situations (e.g. watershed conditions, storm type, etc.) where single event calibration is plausible.
College and Department
Ira A. Fulton College of Engineering and Technology; Civil and Environmental Engineering
BYU ScholarsArchive Citation
Shurtz, Kayson M., "Automated Calibration of the GSSHA Watershed Model: A Look at Accuracy and Viability for Routine Hydrologic Modeling" (2009). Theses and Dissertations. 2013.
WMS, GSSHA, HEC-HMS, calibration, watershed modeling, distributed model, routine hydrologic modeling