Monitoring tropical forest degradation using remote sensing. Challenges and opportunities in the Madre de Dios region, Peru

Abstract

This article focuses on the identification, categorization and discussion of the significant contribution of anthropogenic and climatic factors in forest degradation in a region south of Peru. A combination of existing dataset, spatial analysis, classification techniques, linear regression and logistic regression were used to identify, categorize and explore the key factors related to forest degradation. For the identification of areas in degradation we used Normalized Difference Vegetation Index (NDVI) for gradual changes and Land Use and Land Cover (LULC) for abrupt and accelerated changes in plant cover. The results demonstrated that the first indicator allowed to identify negative gradual trends, but also abrupt negative trends that in total represented 69.8% of the total degraded areas at a 95% reliability, while the second indicator allowed to identify the abrupt changes not identified by the first indicator, contributing with 30.2% to the identification of the degradation. On the other hand, our categorization resulted in three levels of degradation, i.e. low, moderate and high with 73.6%, 5.8% and 20.6% respectively of the total degraded areas. These different levels of degradation were mainly reflected by the contribution of anthropogenic factors such as gold mining, agriculture and human footprint. In addition, we developed a degradation risk map and showed that areas with a probability of degradation of less than 0.5 can become highly degraded regions in the medium and long term, causing serious repercussions on the environment and biodiversity. Therefore, the results of this work constitute a core of knowledge so that managers can make planning decisions properly.