Avaliação da modelagem hidrológica via SMAP da Bacia do Rio Mundaú, utilizando dados do satélite TRMM e de pluviômetro

Abstract

Precipitation is one of the essential variables of the hydrological cycle and has a significant temporal space variability, requiring a dense network of equipment to monitor this variable, however the Brazilian basins are characterized by the sparse network of rain gauges. Another feature found in such basins is the length of unrepresentative flow series. Given the difficulties imposed for the management of water resources, hydrological studies appear, which aim to analyze alternatives for monitoring precipitation and flow in watersheds. Thus, this work was developed in order to analyze the efficiency of precipitation estimates of the TRMM satellite, product 3B42, and its application in a rain-flow hydrological model in the Mundaú River basin, located in the states of Alagoas and Pernambuco, which has significant importance in the economic and social development of the region. First, the study was carried out by comparing concentrated precipitation, the inverse square method of distance, between satellite data and rain gauges, through visual analysis and the statistical concepts of mean error (EM), root mean square error (REQM), correlation coefficients of Pearson (r) and efficiency of Nash-Sutcliffe (NS). Then, the hydrological modeling was done for both data sources, using the concentrated SMAP model, in its monthly version; this study was performed by visual analysis of the hydrographs, permanence curve and dispersion graph, and the statistical coefficients of Nash-Sutcliffe and Pearson. The satellite estimates showed a tendency to underestimate the surface data, the values obtained for the statistical coefficients were EM=-32.13, REQM=65.10, r=0.69 and NS=0.29, presenting low efficiency for the basin despite the good correlation found. Regarding the modeling performed both data sources presented acceptable efficiency, having the rain gauge model better performance for both coefficients analyzed: NS=0.65 and 0.69 for calibration and validation, respectively, and r=0.85 and 0.87 for both periods, respectively. For the TRMM NS=0.56 and 0.57 and r=0.75 and 0.79 for model calibration and validation, respectively. Analyzing the calibrated parameters of the model, it can be seen that the tendency of the estimates leads to significant differences in the characterization of the basin.