Análise de sensibilidade de parâmetros e de desempenho do modelo Soil and Water Assessment Tools (SWAT) na estimativa de vazões médias em bacias hidrográficas

• Autor(a) principal: Moreira, Luana Lavagnoli
• Data de Publicação: 2017
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da Universidade Federal do Espírito Santo (riUfes)
• Texto Completo: http://repositorio.ufes.br/handle/10/9445
• topic: Calibration; Flows; Sensitivity analysis; SWAT; Water production; Modelo SWAT; Produção de água; SWAT; Vazões; Engenharia Sanitária; Análise de sensibilidade; Calibração; Vazões naturais; Bacias hidrográficas; 628

Flows modelling is presents by uncertainties related to climatological input data and watershed physiographic characteristics. In order to get representative flows simulated from the study area it is necessary adjust these flows to observed data. A satisfactory calibration depends on some factors that interfere in this process. Thus, the efficacy of the calibration of the flows estimated by the Soil and Water Assessment Tool (SWAT) model in rural watershed in the southeast of Espírito Santo state was evaluated, by changing calibration parameters, fluviometric stations, historical series and number of sub-basins involved in the process. In additionally there were verified the process of sensitivity analysis of calibration parameters and the influence of input data in the estimation of the flow rates. The results of the sensitivity analysis identified that most sensitive parameters for the study area are the same found in other studies carried out in Brazil and in its Southeast region. There were differences in the most sensitive parameters found for each type of analysis performed. The mean flows obtained by the SWAT were overestimated when compared to the observed data. Hence, it was necessary to calibrate and analyze the influence of the input data on the results. The calibration tests were performed by the SWAT-CUP program and the efficacy of each test was statistically evaluated by Nash-Sutcliffe Efficiency Coefficient (COE), Trend Percentage (PBIAS) and Mean Square Error (EQM). Calibration tests for observed series flow (29 years) from two fluviometric stations did not generate satisfactory results. The best result was obtained by changing all the calibration parameters for all sub-basins in the study area. This test was applied to short flows series (3 years) and it was possible to obtain a satisfactory calibration with COE values equal to 0.53 and 0.50, PBIAS equal to 6.60 and 20.70 and EQM equal to 0.72 and 0.69 for each fluviometric station, respectively. It is verified that the calibration process is a fine adjustment, therefore, the high overestimation simulated data makes calibration difficult, this can be associated with the quality of the input data of soil, climate, pedology and spatialization of the climatic stations in the basin.