Influência da resolução espacial do pixel na aplicação do modelo de simulação SWAT na bacia hidrografica do rio Guaporé, RS.

Abstract

Today we know the need to preserve environmental resources, which include soils and hydrology. Estimates of water loss and soil erosion are held in different countries, using empirical or conceptual models such as SWAT (Soil and Water Assessment Tool), used to predict the impact of changes in the use and soil management, among others on soil loss and the flow of the watercourse. This model, through their routines, is able to simulate and quantify in a watershed as the dynamics of water flow and sediment occurs, the supply of information on the maps and tabular data format is required. This thesis set up two goals: first to identify the most appropriate spatial resolution for simulations, comparing the measured field data with simulated data flow and sediment yield. To this end, we analyzed how the product image affected the performance of the SWAT simulation application without being calibrated by changing their spatial resolutions of 10, 30, 50, 100 and 200 meters. After there were three analysis processes to determine which resolution gave better results of simulation, called analysis of percentage, and statistical regression, concluding that the image with 200 meter resolution showed the best results of simulation of liquid flow in the channel and without sediment transport model calibration. The second objective is to perform a simulation for basin Guapore validate and calibrate the model for liquid flow, analyzing the results for the sediment delivery with a period of 10 years of data. Performed the simulation is performed sensitivity analysis indicated that seven parameters as the most sensitive. Then ensued for calibration, effecting to change these parameters manually for flow and later to sediment production. Was used to analyze the efficiency calibration, the coefficient of efficiency. According to bibliography, this should keep values above 0.50 to be considered calibrated. For flow data were obtained for a range of average six years analyzed 0, 578 After calibration values, executed to validate the model using a seven-year period, reaching up to data an average flow of 0.68 COE. The model presented below one year with values determined by the bibliography, being satisfactory calibration and validation. As for the transport of sediment was not possible to calibrate and validate the model because the coefficient of efficiency showed high values, inferring that the main causes have been simplifying the soil loss equation uses the simulator to make some calculations, especially related to the length and steepness of slope, and basin size. Furthermore, by deducting the measured sediment data showed high variability within months of the years, making it difficult adjusting the parameters. It was concluded that the simulation process is valid for analysis and establishment of public conservation measures and preventive use of environmental resources, especially those related to the cultivation of land, and that the most appropriate spatial resolution of 200 meters for large basins. This resolution is not, perhaps, the reality for small basins, further studies should be performed.