Satellite-based tree cover as a surrogate for forest bird diversity & abundance in the South American Dry Chaco

LEANDRO MACCHI; MATIAS BAUMANN; CHRISTIAN LEVERS; HENDRIK BLUHM; GASPARRI, IGNACIO; GRAU, H.RICARDO; KUEMMERLE, TOBIAS

Congreso; GEOBON; 2016

Land use change has drastic effects on biodiversity by reducing, fragmenting and degrading habitats. Assessing the impact of land use change on biodiversity is therefore important yet challenging. Linking local scale biodiversity data with satellite-based information is a powerful tool to map Essential Biodiversity Variables (EBVs) across large areas. Most studies that describe the spatio-temporal patterns of land use change are based on categorical maps of land use/cover, which fail to account for heterogeneity within class, ecotones between classes, non-linearity and thresholds that often characterize the response of biodiversity. We focus on the South American Dry Chaco, one of the world deforestation and forest degradation hotspots where the rapid expansion and intensification of agriculture for cattle and crop production threatens biodiversity. Intensification along the land use intensity gradient is characterized by a progressive loss of tree cover, and should ideally be assessed using continuous tree/shrub covers indicators. Here, we combine high-resolution imagery (QuickBird) and all available medium resolution satellite imagery from Landsat 8 to model percent of tree and shrub covers at the regional scale using a Boosted Regression Tree framework. Then, we relate both covers to bird species field data to test whether satellite-derived tree/shrub covers are good surrogates for birds. We identify a positive tipping point of avian species at 20.5% and X% of tree and shrub covers respectively. Open habitat species responded negatively at X% and X% of tree/shrub cover. Our results show the value of satellite-derived continuous vegetation measures as EBVs and thus for broad-scale monitoring of biodiversity.