Characterization of SWAT hydrologic parameter sensitivity and behavior across spatial and temporal gradient
Abstract
Hydrologic models reflect our understanding of factors that regulate stream discharge and range from simple empirical models to highly complex process-based models. Sensitivity analysis is a commonly used tool to detect the parameters that significantly impact model results. In this study, we hypothesized that (1) analysis of patterns in parameter sensitivity could be used to better understand variation in controls on hydrologic behavior within and across a mountain-piedmont-coastal gradient, and (2) shifts in parameter significance among wet, dry, and average precipitation years could reveal differing sensitivities to variation in precipitation. To test our hypotheses, we applied the Soil and Water Assessment Tool (SWAT) to several small headwater sub-watersheds in the Yadkin-Pee Dee basin, located in North and South Carolina, USA. In global sensitivity analyses, we found that main channel routing and curve number for surface runoff parameters were the most significant parameters across all of the study watersheds. Parameters influence on hydrologic behavior varied across the physiographic gradient as well. Soil parameters were more sensitive in the Mountains, while the surface runoff lag coefficient and the plant uptake compensation factor were significant in the Piedmont and Coastal Plain watersheds. The groundwater revap coefficient was significant only in the Piedmont watersheds. We also found noticeable shifts in the behavioral ranges of parameters along the geographical gradient, including surface runoff, main channel routing and soil related parameters. There were also inter-annual variations across the dry, wet, and normal water yield years at the study watersheds. Mountain watersheds exhibited noticeable temporal variation in the behavior of parameters driving evapotranspiration, main channel routing, and soil properties. Two Piedmont watersheds had different temporal variations in parameter behavior, which might be due to the difference in landuse/landcover patterns. The inter-annual variation in Coastal watersheds was not clear. Overall, better understanding of SWAT parameter behavior across geographical locations and variation in inter-annual precipitation will improve the efficiency and accuracy of SWAT modeling efforts.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2016
- Bibcode:
- 2016AGUFM.H41K..08K
- Keywords:
-
- 1804 Catchment;
- HYDROLOGYDE: 1846 Model calibration;
- HYDROLOGYDE: 1855 Remote sensing;
- HYDROLOGY