Development of optimization module in the WHAT system for accurate hydrograph analysis and model application

Over the last years, significant areas in the watershed have been converted into impervious areas with human-induced development activities. These activities resulted in changes in the hydrologic response characteristics and pollutant generation trends. Thus, many hydrologic and water quality computer models have been developed and used to assess hydrologic and water quality impacts of land use changes. However, these models have to be first validated prior to the application in environment-friendly land use planning. The direct runoff estimation module of the Long-Term Hydrologic Impact Assessment (L-THIA) ArcView GIS system was validated using the Web GIS-based Hydrograph Analysis Tool, called WHAT (http://pasture.ecn.purdue.edu/-what). The L-THIA model was applied to the Little Eagle Creek (LEC) watershed and compared with the BFLOW filtered direct runoff values. The R2 value was 0.68, Nash-Sutcliffe coefficient value was 0.64. Also, the L-THIA estimates were compared with those separated using the Eckhardt digital filter (w/ default BFImax value of 0.80). The R2 value was 0.66, Nash-Sutcliffe coefficient value was 0.63. Although these results could indicate that the L-THIA model is good in estimating the direct runoff reasonably well, the filtered direct runoff values using the BFLOW and Eckhardt digital filter with the default BFImax value do not reflect hydrological and hydrogeological situations in the LEC watershed. Thus, the hydrograph recession curve analysis was automated using CGI programming and optimization module was developed. These two modules were integrated into the WHAT system for determinination of the optimum BFImax parameter of the Eckhardt digital filter. The recession curve analysis method was used to separate baseflow because the shape of hydrograph reflects hydrological and hydrogeological situations in the LEC watershed. With automated recession curve analysis method and optimization module of the WHAT system, the optimum BFImax value of 0.597 was determined. The comparison of the L-THIA estimates with filtered direct runoff using optimized BFImax value of 0.597 resulted in the R2 value of 0.66 and the Nash-Sutcliffe coefficient value of 0.63. However, the L-THIA estimates calibrated with optimized BFImax increased by 17% and the L-THIA estimated NPS pollutant loadings increased by more than 20%. This indicates the L-THIA model direct runoff estimates can be wrong by 17% and NPS pollutant loading estimation by more than 20% if accuracy of baseflow separation method is not validated for the study watershed. This study shows the importance in baseflow separation in hydrologic and water quality modeling using the L-THIA model.