Avaliação espaço-temporal do processo de desertificação em sub-bacias hidrográficas do Rio Paraíba no Semiárido do Brasil

• Ano de defesa: 2018
• Autor(a) principal: Souza, Ramon Santos
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraíba; Brasil; Geografia; Programa de Pós-Graduação em Geografia; UFPB
• Palavras-chave em Português: Desertificação; RESTREND; SAVI; Google earth engine; Semiárido; Desertification; Semi-arid; CNPQ::CIENCIAS EXATAS E DA TERRA::GEOCIENCIAS
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/15372

Remote Sensing is an essential tool for the detection,monitoring and evaluation of desertification. In this context, this study analyzed the relationship between SAVI and precipitation to detect changes in vegetation cover caused by the desertification process in the period 1986-2017 in the sub-basins of the Paraíba River, Northeast Brazil. For this, a descriptive statistical analysis, Pearson correlation and an adaptation of the RESTREND method using SAVI (Landsat 30 m) andspatiallyexplicitraindata(CHIRPS)wereperformed. Theresultsshowthatduringthestudy period, the precipitation distribution varied between 350 and 400 mm, obtaining coefficients of variationgreaterthan30%. Theprecipitationbehaviorindicated5yearsofdryseason(DS),6dry years (D), 10 neutral years (N), 6 rainy years (R) and 5 rainier years (RA). The spatial variation of the mean SAVI was higher (> 0.7) mainly in the western portion and in some spots located in more pronounced topographies of the study area. In contrast, SAVI pixels <0.3 followed the major rivers in the study area. The RA years (SAVI = 0.85±0.14,min = 0.71,max = 0.87) and R (SAVI = 0.80±0.15,min = 0.64,max = 0.96) present the highest values of SAVI. It is observed that in D years (SAVI = 0.62±0.14,min = 0.53,max = 0.69) and DS (SAVI = 0.50±0.15,min = 0.37,max = 0.67) have the lowest values. SAVI and precipitation were significantly correlated (p≤0.05) to 83.17% of the pixels and mean correlation coefficient was 0.53. The SAVI trend indicated that 58.57% of the pixels presented significant increasing trends (p≤0.05)ofthepixelvaluesand34.04%ofthepixelsobtainedsignificantdecreasingtendencies (p≤0.05). Residual SAVI (adjusted by precipitation) had a negative residual tendency observed in 31.41% of the pixels and 26.75% of the pixels showed a positive residual trend. During 19862001, a negative residual trend was observed in 50.34% of the pixels and 38.48% of the pixels showed a positive residual trend. In the period 2002-2017, a negative trend was observed in 34% of the pixels and 16% of the pixels showed a positive trend. The SAVI intercept precipitation generally remained above zero, where in the western portion obtained values above 0.5 per pixel. In sum, precipitation impacts and human activities on vegetation dynamics varied in the studied area and specific local measures of environ mental protection and managements hould be adopted.