Avaliação de um conjunto de índices na identificação do risco à desertificação na sub-bacia Taperoá - PB

Abstract

Finding a methodology that is able to identify areas susceptible to desertification in many regions of the world is a challenge that the scientific community has been facing since the 70’s. The desertification is a complex phenomenon that comprises large territorial and population extension, as it involves environmental, social and economic problems. An alternative for identification of desertification is the use of indicators that integrate human activities, environment and climate, along with satellite technology, because it can be applied in a simplified and inexpensive way. Among the methodologies found in literature, there is the Risk of Tendency to Desertification Index (RTDI). It comprises a group of soil and vegetation spectral indexes, climate indexes and human impact indexes, proving to be practical, adaptable, and cheap and with satisfactory results. Another approach that has been studied is the identification of degraded areas through multitemporal analysis so that the Change Vector Analysis (CVA) can be done, making it possible to identify desertification. In this sense, this thesis proposes to apply the CVA through the Soil Adjusted Vegetation Index (SAVI) and Bare Soil Index (BSI), and to evaluate the results, differentiating between rainy and dry periods. Besides that, the goal is to examine, validate and adapt the RTDI methodology in the Taperoá sub-basin, a semiarid area of Brazil located in the State of Paraíba, considering a temporal analysis between the years of 1995 and 2015. For the spectral data, the years selected to the study were 1995, 1999, 2005 and 2015, taking into account the seasonality. March, April and May form the rainy period, while the dry period is formed by the months September, October and November. The clouds of the images of each month were removed and the images were merged according to the respective seasonal period. The CVA presented the best results in the rainy period. The RTDI was also tested according to the seasonality and the class distribution range, using equal and quartile intervals, considering the variation of the SAVI L parameter (0.8, 0.9 and 1). In addition to that, the results of the collected field data were validated (43 points in 5 different campaigns carried out from 2016 to 2019), including the verification and categorization of the desertification degree according to experienced technical professionals. It was done comparative analysis pixel/point by performance indexes, identifying areas that are susceptible to desertification, according to the field survey and the estimates obtained within this thesis. The application of the RTDI with quartile interval class distribution, during the dry period and using the SAVI parameter as 1, presented better performance than the other method variations. The applied indexes’ driving force estimates showed that the parameters related to the soil and vegetation were decisive in terms of the classification level of the risk of tendency to desertification.