Effect of Spatial Variability on the Simulation of Overland and Channel Flow

Abstract

ABSTRACT A finite element storm hydrograph model (FESHM) was used to simulate overland and channel flow. A hypothetical watershed was used to demonstrate spatial variability of such factors as soils, landuse or cover con-ditions, topography and rainfall. Several graphical displays were created to illustrate the effect of the above factors which vary spatiotemporally. Other examples were presented to illustrate potential errors emanating from ignoring spatial variability in selected variables. Rainfall is the most sensitive variable and possibly the most difficult to obtain good areal measurement. A natural watershed was used to illustrate the effect of aggregation of soils, landuse and topographic descrip-tors. In general, the error in prediction became pro-gressively worse as the level of resolution of the data was decreased. Gross errors were noted when the area was sub-divided into one HRU and different rainfall distribu-tions were used to simulate discharge. Peak discharge varied from zero to approximately 28 mm/h.