Process parameterization and temporal scaling in surface runoff and erosion modelling

It is common for runoff and erosion models to be run at coarse time steps (e.g. daily) owing to limited data availability. However, such models are unable to capture adequately the small-scale surface runoff and erosion processes that are dominated by, for example, precipitation characteristics at time-scales of minutes. This dilemma calls for the development of approaches that appropriately capture short time-scale processes when running the models at daily scales. In this paper, we assess two scaling approaches that are simple and are intended to preserve prediction quality as time scales become coarser: (i) use of effective rates of precipitation and runoff, and (ii) use of a rainfall disaggregation scheme. Runoff and erosion are simulated using a six-parameter surface runoff model and a one-parameter erosion model. A downhill simplex optimization algorithm is applied for parameterization at 2 min and daily time steps, using measured 2 min and daily precipitation data respectively, against measured daily runoff and erosion. The 2 min simulation results are aggregated to daily values to make comparisons with the measured plot-scale daily runoff and erosion data for 2 years (1997-98) from six sites in the Jhikhu Khola catchment in Nepal. The runoff and erosion simulation results calibrated from the 2 min precipitation data are good (E = 0.92 for runoff and 0.80 for erosion, all sites collectively), whereas other results, particularly the erosion simulation from the calibrated daily model, are very poor. These results are taken as the reference to compare with the two different scaling approaches. The effective rate approach is poor compared with the 2 min estimates but is a considerable improvement over daily erosion estimates. The rainfall disaggregation scheme is comparable to the 2 min estimates and produces better results than the effective rate approach. This study illustrates that rainfall scaling has significant scope for temporal scaling of hydrologic processes, particularly where continuous simulation is preferred. Copyright (C) 2004 John Wiley Sons, Ltd.