Assessment and methodological proposal for the restoration of Brazilian Atlantic forest ecosystems

Abstract

This research proposes a methodology based on fuzzy logic and spectral indices to identify priority areas for forest recovery in the Environmental Protection Area (EPA) of Lagoa Encantada and Rio Almada (Bahia-Brazil), integrating temporal analyses from 2000 to 2023 of environmental and anthropogenic variables. Time series of Landsat 7 images processed in Google Earth Engine were used, with analyses in RStudio and QGIS. Vegetation and water indices were calculated, correlated with land cover and land use data and forest loss. Fuzzy modeling is considered favorable soil. The correlation revealed a strong negative relationship between forest formation and pastures, evidencing the conversion of natural areas to agricultural use. The NDWI, with values between −0.60 and −0.30, indicated a progressive reduction in soil moisture, associated with variables, such as SAVI and Permanent Preservation Areas, and unfavorable variables, such as pastures and urban infrastructure, to map forest recovery potential. The results indicated an annual reduction of 0.73 km² in forest formation, accompanied by the advance of pastures (+1 km²/year) and urbanization. The NDVI remained stable (0.60–0.72), but the SAVI showed a downward trend, identifying an increase in exposed the decline in water bodies (0.021 km²/year). In addition, 18.42 km² with moderate potential for recovery were identified, concentrated mainly in the western zone of the EPA. It is estimated that there are areas with compacted soil that require specific interventions, such as decompaction and nucleation. The methodology proved effective in prioritizing areas for restoration and as an accessible tool for sustainable environmental management.