TOPMODEL Based On DEM With Different Spatial Resolutions

Digital Elevation Drainage Network Model (DEDNM) was used for the aggregation of DEM data from a spatial resolution of 3 s of latitude or longitude to six kinds of spatial resolutions, i.e. 9s, 18 s, 27 s, 36 s, 45 s, 54 s. Also DEDNM was applied to generating the digital basin from DEM data with seven kinds of different spatial resolutions.Due to the heterogeneity of precipitation, evaporation and topographical index within a catchment, TOPMODEL was developed over each subcatchment in the Baohe catchment with routing algorithm modified. The parameters in TOPMODEL were calibrated by the SCE-UA method. It is seen from the analysis on 24 flood events that TOPMODEL performs better in terms of model efficiency coefficient and relative error of runoff depth than that in terms of relative error of flood peak discharge and difference of peak appearing time between actual and computed ones.It is found from the analysis on different spatial resolutions that topographical parameters are affected by DEM resolution remarkably. The coarser the resolution of DEM, the more the loss of information in topographical parameters. Also TOPMODEL behaves badly as the resolution of DEM becomes coarse.As the resolution factor is introduced to upslope area, and the fractal to land surface slope, the scale-invariant TOPMODEL can be developed because of scale-invariant topographical index being obtained. It is seen from the case study that the scale-invariant TOPMODEL has the similar performance to TOPMODEL in the case of higher resolution of DEM. Namely it could remove partial effect of DEM resolution on TOPMODEL to some extent.The static parameters of TOPMODEL refer to those calibrated by SCE-UA based on DEM data with 3 s resolution. The dynamic parameters are re-calibrated by SCE-UA method in each specific situation with different resolutions, respectively. It is seen that TOPMODEL with dynamic parameters performs better than that with static parameters, except for flood peak appearing time. And the parameter of TOPMODEL, SZM changes much more in the case of dynamic parameters in comparison with the situation of static parameters.