An Investigation on the Effects of DEM Resolution and Flow Tracing Algorithms on the Topo Index and TOPMODEL Performance (Case Study: Kasilian and Karde Catchments)

Topography is an important factor in watershed response to input rainfall. The important role of flow tracing algorithms for deriving geomorphologic parameters of watersheds to feed distributed/semi distributed rainfall runoff models is undeniable. In present research, the effects of some widely used algorithms like D8, MFD, and Lea on the performance of a semi distributed model, TOPMODEL, in different cell sizes were investigated. The model was applied in two study watersheds, Kasilian and Karde, with two different areas and climates. It was illustrated that while for all algorithms the Topo Index values increased with the increase of the DEM cell size, in a fixed cell size MFD and D8 algorithms gave the highest and lowest values of the Topo Index, respectively. The most important result drawn from this observation is the fact that the drainage area of each cell varies by changing the flow tracing algorithm. The model calibration, focusing on its important parameter “m” which controls the amount of runoff generated by the model, compensated the effects of DEM cell size’s increase and improved the model performance in terms of Nash-Sutcliffe index. It was concluded that using automatic calibration for enhancing the model performance could be misleading through giving unrealistic contribution of simulated surface and subsurface flow components. For example, in Kasilian watershed using D8 as the flow tracing algorithm the surface flow and subsurface flow showed the minimum and maximum contribution to the generated flow with respect to the other two algorithms, respectively. Moreover, MFD algorithm because of giving low values of “m” in the Kasilian watershed as a humid region is not a suitable algorithm in this type of climate and is in conflict with the basics of TOPMODEL.