Análise das estimativas de umidade do solo dos satélites SMOS e SMAP com o NDVI e variáveis atmosféricas

• Ano de defesa: 2023
• Autor(a) principal: Leidiane de Paula Rezende
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; Brasil; IGC - DEPARTAMENTO DE GEOGRAFIA; Programa de Pós-Graduação em Geografia; UFMG
• Palavras-chave em Português: umidade do solo; vegetação; NDVI; SMAP; SMOS; Solos – Umidade; Vegetação e clima – Brasil; Satélites artificiais
• Link de acesso: http://hdl.handle.net/1843/56777; https://orcid.org/0000-0003-3083-3348

The study evaluates the soil moisture estimates of the passive sensors of the SMOS and SMAP satellites with the NDVI of the MODIS sensor, comparing their monthly averages with precipitation and evapotranspiration data, and analyzes the physical and environmental characteristics from a set of sample points between the Cerrado, Caatinga and Atlantic Forest biomes, considering herbaceous and shrubby areas, forests and cultivated areas. There is a direct relationship between vegetation vigor and soil moisture, having to consider the root zone and the delay time of vegetation response in relation to the availability of water in the soil. In summary, knowledge of the soil moisture variation regime provides essential information for the analysis and management of irrigation, the choice of planting time, and analysis of water consumption by plants, allowing better management of resources. Soil is seen as a potential reservoir of water and soil moisture as the fundamental variable of the soil-plant-atmosphere system. There are few soil moisture monitoring units in Brazil and soil moisture estimates through Remote Sensing are a possible alternative. Soil moisture products obtained by passive microwaves can be associated with optical sensors, which detect vegetation vigor which is closely linked to available soil moisture. In this study, the cross-correlation technique was applied to observe the vegetation response delay time in relation to soil moisture, but the results varied greatly to the detriment of the characteristics of each point analyzed. Soil moisture estimates and the NDVI were well correlated (0.5 to 0.75) in more than 85% of the dataset of the 36 points analyzed. The regression coefficients between soil moisture, precipitation and evapotranspiration showed values between 0.5 and 0.94 for the same set in general. It was observed that the texture of the soil, the characteristics of the relief and the vegetation are factors that can influence the relationship between the vigor of the vegetation and the soil moisture, and that the effects of anthropic action related to the management and use of the land can compromise the relationship of variables and analysis by time series. As for the regional aspect, soil moisture followed the seasonal effects of precipitation and evapotranspiration rates, according to the climatic gradient, without major oscillations. It was concluded that the soil moisture estimates from the satellites are very representative on a regional scale as they are well correlated with NDVI, precipitation and evapotranspiration, and represent well the dynamics of the soil-plant-water-atmosphere system.