Uso de dados do sensor Modis/Agua e do algoritmo Sebal para estimativa da evapotranspiração real na bacia do Rio Paranaíba

• Ano de defesa: 2015
• Autor(a) principal: Martins, Alécio Perini
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Uberlândia; BR; Programa de Pós-graduação em Geografia; Ciências Humanas; UFU
• Palavras-chave em Português: Algoritmo SEBAL; Evapotranspiração real; Sensoriamento remoto; Ciclo hidrológico; Evapotranspiração - Rio Paranaíba (MG); Sensoriamento remoto - Rio Paranaíba (MG); SEBAL algorithm; Real evapotranspiration; Remote sensing; CNPQ::CIENCIAS HUMANAS::GEOGRAFIA
• Link de acesso: https://repositorio.ufu.br/handle/123456789/16014; https://doi.org/10.14393/ufu.te.2015.70

Due to intense changes in the landscape over the last century, as well as the improvement of study techniques and prevision models of environmental scenarios changes, geography has played an important part for being a science that presents an integrated vision of phenomena, allowing planning and management activities, as well as the impacts mitigation efficiency. Paranaíba River, whose drainage basin was chosen as the area of the method application, presents an area of about 230.000 km², with a diversity of extremely interesting environments for environmental studies. Therefore, this research had as the main objective evaluating the data use of MODIS/AQUA sensor to estimate the real evapotranspiration in the river basin of Rio Paranaíba. The estimates are important to understand processes that compose the hydrological cycle and, consequently, the water availability; therefore we accomplished these estimates using SEBAL (Surface Energy Balance Algorithms for Land) algorithm, developed by Bastiaanssen (1995). This model uses physical principles involved in the reflection, transmission and absorption of energy by the surface, requiring few data collected in the field and, therefore, allowing the study of extensive areas or to regions with few data available or of difficult access. The model applies as to images of TM/Landsat sensor, as to images of MODIS (TERRA or AQUA) sensor, since the last ones are the most recommended to extensive areas due to covering areas of 10° x 10°. From the modeling results, it s possible to notice that as the radiation balance to the surface, as well as the estimated heat fluxes and the real evapotranspiration present a strong relation towards the land use and vegetation coverage and, secondarily, with the seasonal rainfall variation, with identified higher levels in images between the months of November and March, which correspond to the rainy period of the region. Regarding the land use, areas with a dense vegetation coverage (silviculture and forests), as well as the reflecting pools of hydroelectric power plants present higher levels of evapotranspiration in every month, while cerrado and pasture areas don t present significant differences between them. It wasn t possible to notice great differences in the registered values in areas of irrigated and not irrigated agriculture, mainly due to the low spatial resolution of the products generated by MODIS/AQUA images, which is 500 meters. One of the failures presented by the model is in the fact that it doesn t consider specific data of the soils in the evapotranspiration estimate, since the soils present distinguish patterns of retention and loss of water in different processes. Another failure can be found in the estimate of sensitive heat flux, which needs climatic data, specially wind speed, to its calibration. From the analysis of the results and, considering the bibliography about the theme, we can consider a reliability of the estimated data over 90% for the used scale, which is 1:2.000.000.