The Effects of Deforestation and Forest Degradation on Climate Variables Have Not Scaled Up to the Watershed Level in the Southwestern Amazon, yet

Abstract

This research focuses on the effects of deforestation and forest degradation on climate variables in one of the most preserved and ecologically important yet least studied regions of the Amazon basin, the Southwestern Amazon. Indigenous and non-indigenous groups in this region rely on the provisioning and regulating services provided by humid, tropical forest. To map forest cover changes, we implemented and compared the performance of three Landsat-based forest disturbance algorithms that rely on pixel segmentation strategies and machine learning methods. Based on the mapped results of the algorithm that performed best, we evaluated the effects of forest disturbances on three climate variables of most concern to local people: evapotranspiration, land-surface-temperature, and precipitation. We estimated annual and monthly differences in climate variables over areas classified as intact forest, degraded forest, or deforested and determined the influence of forest disturbances on climate variables at the local and watershed scale over the last two decades through a Partial Mann-Kendall approach. The results show that although the climate implications of deforestation and forest degradation haven not scaled up to the watershed level yet, even the local impacts we detected can have important implications for areas outside of the Southwestern Amaxon, which emphasizes the importance of continuing to conserve this remote region. Moreover, the overall approach of this research addresses how the integration of geospatial modeling with indigenous ecological knowledge is essential to understand and cope with environmental change.