Deforestation drivers in Colombia Andes: a case study in Antioquia's province

Abstract:

Tropical Andes, one of the most diverse region of the world, is highly transformed due to anthropogenic activities such as agriculture and cattling. Antioquia´s province located in the northwest of Colombia have an extension of 63612 km² that conserves less than 52,4% natural forest cover. Under this scenario, monitoring land cover change drivers (LCCD) is essential to land management and conservation planning. Remote sensing data is useful for this purpose, allowing to assess global deforestation estimations. Nevertheless, these broad-scale assessments underestimate local LCCD. In this study, that is promoted by a regional ecological observation system (Observatorio de Bosques de Antioquia -OBA) to provide forest management decisions, we determined LCCD based on decision tree modeling and remote sensing data. First, we selected ~900 sites from Google Earth imagery across Antioquia, and these were classified according to LCCD categories (crops, livestock, degraded land by mining, forestry plantation, urban settlement, and fires). Second, we gathered 20 remote sensing-based land cover indexes from Lansat, MODIS, and ALOS PALSAR products. We ran decision trees employing data between 2010-2015 period and assess its performance employing data from 2015-2020. Currently, ca. 30000 ha y^-1 of forest loss are reported in Antioquia, it could be, principally, attributed to cattling and agricultural practices. This information is actively divulgate in OBA's platform and outreach campaigns to improve citizen awareness and political incidence (https://observatoriobosquesantioquia.org/) for decision-makers and the community.