Effect of Grid Size of the Digital Terrain Model on Hydrologic Simulation

Digital Terrain Model (DTM) data are helpful for landscape representation, spatial analysis, and hydrologic modeling. In recent decades, DTM data have been widely applied in many fields, such as civil engineering, geography, and natural resources. Although DTM data are useful, their grid-based structures affect hydrologic simulation. Thus this paper attempts to understand the effects of grid sizes on the responses of the hydrologic model, TOPMODEL, including topographic indices, model parameter values, and streamflow simulation in Fushan watershed no. 2(94 ha). The relationship between the topographic index and optimized parameter values of hydraulic conductivity in TOPMODEL was also examined. Five different grid size DTM data, at 10, 20, 30, 40, and 50m, were applied to calculate topographic indices. The hydrologic data, including precipitation, streamflow, and air temperature from July to October 1995 in the study area, were selected. The Rosenbrock optimization algorithm was used to find optimal parameters within their reasonable ranges. Results show that the mean of topographic indices increases as grid size increases. When optimal parameters of TOPMODEL for 10-m DTM data are fixed and then applied to other grid sizes, model efficiencies slightly decrease with increase of grid size. However, model efficiencies can become virtually identical if parameters are properly optimized for different DTM data. The optimized parameter value of hydraulic conductivity increases as grid size increases. This is due to compensation for the effects of changing DTM grid size on topographic indices. In addition, there is a positive relationship between optimized parameter values of hydraulic conductivity and means of topographic indices. Based on the hydrologic simulation, landscape representation, and data handling requirements for these 5 different gird sizes, the 10-m grid size is recommended for DTM-based applications of geomorphologic and hydrologic modeling in the study area. However, we should check our results again if there are finer DTM grid sizes in Fushan watershed no. 2 in the future.