Spatial-temporal dynamics of deforestation and its drivers indicate need for locally-adapted environmental governance in Colombia

Characterizing and managing deforestation are some of the most urgent environmental challenges in the tropics. Forest loss depends on multiple drivers that relate to ecological and socio-economic conditions, and varies significantly in time and space. Understanding the causes and potential consequences of forest loss in a region, requires detailed knowledge of the spatial and temporal behavior of deforestation drivers. This is especially important in Colombia, a country with an immense geographical, ecological and cultural diversity. In this work we use a spatially-explicit modeling platform to identify the influence of multiple deforestation drivers in different biogeographical and planning regions of Colombia. Models were used to produce different deforestation scenarios, resulting in contrasting trends. Results suggest that agricultural activities have a significant role in deforestation in Colombia, although its influence largely varies among regions. Notably, the long-lasting internal conflict in the country, indicated here by a proxy variable, produced contrasting effects on historical deforestation, such that in some areas it acted as an attractor and in some others as a deterrent of deforestation. Our scenarios for the mid-21st century indicate that current trends in deforestation and its drivers (except the internal conflict due to a recent peace agreement) can potentially lead to major ecological impacts, such as fragmentation and loss of connectivity among ecosystems. In the Governance scenario, however, enforcement of conservation in protected areas ensures partial forest conservation, highlighting the key relevance of these areas for biodiversity conservation in the country and for controlling the expansion of deforestation. Given the high spatial and temporal variability of deforestation, and that the same drivers can lead to different responses in different geographic areas, decision-making needs a more efficient approach for designing locally-adapted measurements to face current challenges of conservation and management of forests and their services.