Estimativa do escoamento superficial em uma bacia hidrográfica com base em modelagem dinâmica e distribuída

Abstract

Knowledge on the surface runoff in watersheds is very important for a good conservation of natural resources. However, surface runoff is a complex and dynamic process, especially in the context of spatial variability. An application of the Geographical Information System (GIS) tools in small grid cells is therefore adequate. This way, it is possible to consider the spatial behavior of variables associated to the origin of surface runoff. This study aimed to program the Hydrological Models SCS-CN (SCS-Curve Number) and CN - MMS (CN - Modified) based on the PC Raster Programming Language and a reduced dataset, in distributed and dynamic approaches, to estimate the direct surface runoff in a watershed of Oxissols, in the county of Nazareno, region of Campos das Vertentes, Minas Gerais State. SCS-CN model was applied structuring a CN-map linked to the soil infiltration capacity in the PCRaster GIS format. For the CN - MMS model, we used other maps: saturated volumetric soil moisture content, initial volumetric soil moisture content and soil depth. For the simulation and evaluation of both models, 18 rainfall events that produced surface runoff in the hydrologic year 2004-2005 were used, and their respective surface runoff depths observed. The model performance was evaluated by sensitivity analyses, based on the mean error and Root Square Error. Based on these accuracy statistics, the calibration of the CN - MMS model was better than in the original SCS-CN model, since the initial soil moisture was taken into account. Nevertheless, the adjustment of the models in the SIG PCRaster allowed the development of an effective and useful computer tool to simulate surface runoff, due to the possibility of establishing computer routines considering the problems elated to the spatial variability of data entries of the model.