The Driftless Area of southeast Minnesota, southwest Wisconsin, northwest Illinois, and northeast Iowa presents many challenges to the numerical modeler of the precipitation to streamflow runoff process. The North Central River Forecast Center (NCRFC) models river watersheds of 50 to 250 square miles in this area. Lumped parameters that run on a 6-hour time step are used in the model. Improved model performance for the Kickapoo River basin of southwest Wisconsin is of particular interest to the NCRFC. The watershed comprises 687 square miles and is broken into 7 sub-watersheds. The Kickapoo River has been plagued by record breaking floods including both flash floods and multi-day events in the past decade. Experience has shown that modeling the Kickapoo River at a 6-hour time step is not adequate to capture the quick response of these small scale sub-watersheds. A reduced time step is required to better resolve the response hydrograph from the various input forcings ranging from snowmelt to intense summer thunderstorms.

The long-term goal of this study is to improve flood forecast accuracy for the Kickapoo River using the Community Hydrologic Prediction System (CHPS). The first step is to characterize the time scales of flood response for the Kickapoo River watershed. Flash floods with durations less than 6 hours and multi-day events were considered. Response times of each sub-watershed, the duration of rainfall, duration of streamflow above flood stage, season, and the pattern of rainfall were quantified. Spatial and temporal scaling behaviors of the rainfall were also analyzed for flood events with NEXRAD rainfall data available. Post-event analysis was used to compare observed streamflow with archived operational forecast products to better understand the deficiencies of the current 6-hour time step, lumped model. The next phase of the project will be to use the characterization of the flood response time scales of the Kickapoo River under differing seasons, types of storms and rainfall patterns to explore possible model modifications using CHPS that will improve flood forecast skill for the sub-watersheds.