Effects of grid cell size and time step length on simulation results of the Limburg soil erosion model (LISEM)

With increasing computer power, process-based models that use grids to discretize space have become increasingly popular. For such models, the simulation results might depend on both grid cell size and, in the case of dynamic models, on the time step length used in the model. In this study, the dynamic Limburg soil erosion model (LISEM) was applied to a small catchment on the Chinese Loess Plateau. To study the effect of grid cell size and time step length, simulations were performed for grid cell sizes ranging from 5 to 100 m for a single time step length, and for time step lengths ranging from 2 to 120 s for a single grid cell size. The results show that the LISEM results vary considerably as a function of both grid cell size and time step length. For both increase in cell size and increasing time step length, the trend was a decrease in predicted discharge and predicted soil loss. For discharge, the most important causes are likely to be a decrease in slope with increasing grid cell size, rainfall averaging for longer time step lengths, and numerical dispersion of the kinematic wave solution. For soil loss, the cause is less clear, reflecting the complexity of soil loss prediction, which depends on available water, transport capacity and sediment redistribution, all of which change in time and space. These results show that a choice for a certain grid cell size and a certain time step length should be made before calibration of the model. Similar erosion models are likely to have similar dependencies on grid size and time step length. Copyright (c) 2005 John Wiley & Sons, Ltd.

[1]  J. Poesen,et al.  The European Soil Erosion Model (EUROSEM): A dynamic approach for predicting sediment transport from fields and small catchments. , 1998 .

[2]  Lawrence W. Martz,et al.  Grid size dependency of parameters extracted from digital elevation models , 1994 .

[3]  Gerard Govers,et al.  Empirical relationships for the transport capacity of overland flow. , 1989 .

[4]  Zhi-Yong Yin,et al.  A cross-scale comparison of drainage basin characteristics derived from digital elevation models , 1999 .

[5]  V. Singh Is hydrology kinematic? , 2002 .

[6]  Jens Christian Refsgaard,et al.  Effect of grid size on effective parameters and model performance of the MIKE‐SHE code , 2002 .

[7]  Jeroen M. Schoorl,et al.  Three-dimensional landscape process modelling : the effect of DEM resolution , 2000 .

[8]  P. Braun,et al.  The problem of scaling in grid-related hydrological process modelling. , 1997 .

[9]  Gerard Govers,et al.  Erosion models: quality of spatial predictions , 2003 .

[10]  Derek Karssenberg,et al.  Integrating dynamic environmental models in GIS: The development of a Dynamic Modelling language , 1996, Trans. GIS.

[11]  A. D. Roo,et al.  Spatial Analysis of Erosion Conservation Measures with LISEM , 2001 .

[12]  Reinder A. Feddes,et al.  Simulation model of the water balance of a cropped soil: SWATRE , 1983 .

[13]  A. D. Roo,et al.  LISEM: a single-event physically based hydrological and soil erosion model for drainage basins; I: theory, input and output , 1996 .

[14]  K. Beven,et al.  Sensitivity to space and time resolution of a hydrological model using digital elevation data. , 1995 .

[15]  William W. Doe,et al.  Introduction to Soil Erosion and Landscape Evolution Modeling , 2001 .

[16]  Gerard Govers,et al.  The European soil erosion model (EUROSEM): documentation and user guide. , 1998 .

[17]  N. Farajalla,et al.  Capturing the essential spatial variability in distributed hydrological modelling: Infiltration parameters , 1995 .