Impact of Time Step On the Results of a Fully-Distributed Watershed Model for Scenario Analysis

Abstract

Major storms are significant drivers of runoff from agricultural areas, which implies that modeling watershed processes at smaller time steps, such as hourly versus daily, will allow such effects to be expressed in a watershed model. Thus, the results of these models at this level would be effectively used in scenario analysis to determine the impacts of certain agricultural practices, which affects decisions that can have major economic and environmental consequences. This topic has been investigated for semi-distributed models, but not yet for fully-distributed models. To help fill this research gap this study investigated whether different time steps in a fully-distributed watershed model resulted in significantly different surface runoff and streamflow results. Surface runoff and streamflow results were generated for different time steps (5-minutes, 30-minutes, 1-hour, and 1-day) for a case study area in Nebraska, USA, using the Spatially Explicit Integrated Modeling System (SEIMS). These results were then compared for significant difference using ANOVA. The major findings of this study included that the streamflow and surface runoff results were generally significantly different when generated using different time steps. However, when SEIMS was calibrated, the difference between the results of different time steps reduced such that, in many cases, they were non-significantly different (p > 0.05). These trends were seen across multiple grazing management scenarios implemented within the model. These findings imply that time step should be a significant consideration when modeling agricultural practices in ungauged watersheds.