Use of Data from an Intensively-monitored Field Site to Determine Input Parameters for SWAT

Abstract

The effects of soil and water conservation practices on water quality are being simulated with hydrologic models. Input parameters for physically-based models are commonly taken from generalized property data. However, field-measured input parameters provide a higher level of assurance that model output uncertainty will be affected by model structure rather than by input parameter uncertainty. The objectives of this study were to calculate curve numbers for conventional and conservation strip tillage and to compare curve numbers for growing and dormant seasons. Precipitation - runoff data collected from 1999 - 2005 at a field study site in south Georgia were used to calculate curve numbers by an averaging method and by a log-normal method. For average antecedent moisture conditions, the curve numbers calculated by the average curve number method for conventional and strip tillages were 89 and 84, respectively, and by the log-normal method were 89 and 83, respectively. The curve numbers derived from the short-term field data are higher than standard table values, 81 and 75, respectively, which were developed from a long-term series of annual maximum runoff. Curve numbers for strip tillage were 81 and 86 for growing season and dormant season, respectively; however, there was no difference between growing and dormant season curve number, 89, for conventional tillage. Under wetter than normal soil conditions during the dormant season, the curve number for both conventional and strip tillage was 96. The results indicate that hydrologic modeling of Coastal Plain watersheds would be improved if curve number were varied seasonally.