ABSTRACT Land-use changes are considered one of the most important factors that affect the water resources. The objective of this study was to investigate the potential impacts of land-use changes on the hydrological behavior of the Upper Grande River Basin, southern Minas Gerais state, Brazil, based on different land-use scenarios using the SWAT model. For this purpose, daily streamflow records from the Macaia gauge station were calibrated and validated under the current land-use. To assess land-use change impacts four land-use scenarios were developed following official environmental planning reports: S1 and S2 - conversion of forest into pasture of 20 and 50%, respectively; S3 and S4 - conversion of pasture into forest of 20 and 50%, respectively. The results have showed that, in general, the deforestation scenarios (S1 and S2) presented an increase in total runoff and peak flow and a decrease in the baseflow and evapotranspiration, whereas the reforestation scenarios (S3 and S4) have showed the opposite. The results showed that the land-use changes can generate positive impacts, such as reduction of surface runoff and increase in the baseflow, as well as negative ones, like the increase of soil erosion and flood risks.

RESUMO Informações confiáveis sobre a cobertura vegetal e sua dinâmica ao longo do tempo são essenciais para a gestão de vidas, recursos, e apoio a políticas. Embora muitas iniciativas tenham sido realizadas no Brasil desde a década de 1980, o conhecimento sobre suas florestas secas ainda é limitado. Neste estudo, abordamos os moduladores do desmatamento e sua ameaça à floresta seca brasileira chamada Caatinga. Avaliamos a dinâmica da cobertura da terra e métricas de paisagem na bacia do rio Pontal, a qual se tornou uma das áreas mais socialmente e ambientalmente importantes do Brasil. O processo em geral consiste em uma amostragem sistemática baseada em características hidrológicas, onde imagens Landsat (1975 a 2013) combinadas com o software FRAGSTATS foram utilizadas para avaliar índices de métricas de paisagem para a Caatinga. Os dados mostraram que a área relativa (RA) diminuiu de 90,25% para 60,98% ao longo do período de 38 anos, enquanto que a fragmentação (PD) apresentou uma tendência crescente. Além disso, a distribuição espacial de ambos os índices se tornou mais heterogênea e agrupada ao norte. Sugerimos que a aptidão da terra para a agricultura foi a principal causa do desmatamento. A chuva é 112% mais intensa no norte da bacia do que no sul. Portanto a fauna e flora ao norte foram significativamente alteradas e reduzidas. A relação entre desmatamento e tempo mostrou uma ameaça crescente. Outros estudos que avaliem essas causas são necessários para melhorar a compreensão da dinâmica da vegetação nesta região.

ABSTRACT Reliable information on vegetation cover and its dynamics over time is essential for managing life, resources, and supporting policies. Although many related initiatives have been undertaken in Brazil since the 1980s, knowledge about its dry forests is still limited. In this study, we address the drivers of deforestation and their threat to the Brazilian dry forest called Caatinga. We evaluated the land cover dynamics and landscape metrics in the watershed of the Pontal River, as it has become one of the most socially and environmentally important areas in Brazil. The overall process consists of a systematic sampling grid of hydrological samples, where Landsat images (1975 to 2013) combined with the FRAGSTATS package were used to evaluate landscape metrics indices for the Caatinga. Data showed that the relative area (RA) decreased from 90.25% to 60.98% over the 38-year period, while fragmentation (PD) presented an increasing bias. In addition, the spatial distribution of both indices became more heterogeneous and clustered in the north. We suggest that appropriateness of land for farming was the leading cause of deforestation; rainfall is 112% more intense in the north of the watershed than in the south. Therefore, the northern fauna and flora have been significantly altered and reduced. The relationship between deforestation and time shows an increasing threat. Further studies evaluating these causes are needed to improve our understanding of the vegetation dynamics in this region.