Efeito da seca na vegetação com suporte de sensoriamento remoto em biomas situados no nordeste e no sudeste do Brasil

• Ano de defesa: 2019
• Autor(a) principal: Rebello, Vitor Paiva Alcoforado
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Rio de Janeiro; Brasil; Instituto Alberto Luiz Coimbra de Pós-Graduação e Pesquisa de Engenharia; Programa de Pós-Graduação em Engenharia Civil; UFRJ
• Palavras-chave em Português: Sensoriamento remoto; Biomas do Brasil; Vegetação; Engenharia Civil
• Link de acesso: http://hdl.handle.net/11422/23163

The research work investigates the effect of drought on plants by monitoring vegetation based on remote sensing, involving aspects such as seasonality, productivity and phenology. The study region was the eastern portion of Brazil, where there was a severe drought from 2013 to 2015, mainly focusing on the biomes of atlantic forest, cerrado, caatinga and a portion of the Amazon forest. The vegetation index NDVI obtained with the MODIS sensor was employed. In addition, the study incorporated GLDAS surface model data, reanalysis data, drought indices and observed data from ground meteorological network of weather stations for improved comprehension with respect to climatic effect on vegetation. Evaluation by means of principal components and singular value decomposition pointed out to the impact of the behavior of hydrological variables such as precipitation, soil moisture and evapotranspiration, in that order, to predict vegetation anomalies. Caatinga revealed to be the most susceptible region to drought and with the highest potential for anomaly prediction. The NDVI seasonal parameters such as amplitude, integral curve, annual maximum and baseline profile curve of NDVI temporal variation were sensitive to drought severity, but also dependent on the total duration of drought and its temporal persistence, showing that these two last factors are relevant in considering the impact of water deficit on soil cover.