Abstract: Forest loss and degradation have been identified as some of the primary drivers of declines in biodiversity. However, for many regions we lack clearly quantified species-specific responses to forest loss and degradation. This is especially true in SE Asia, a region of exceptionally high biodiversity, which has also experienced some of the most rapid rates of forest loss. Yet we know relatively little about species-level responses to the region's ongoing habitat loss. Recent developments in the availability of citizen science datasets (e.g. eBird), as well as user-friendly, cloud-based computing platforms, have opened up the possibilities for modeling species responses to habitat loss at a scale that has been logistically infeasible. Implementing all spatial processing and modeling in Google Earth Engine (GEE), we investigated distributional changes of forest-associated birds in association with forest loss and degradation throughout the past three decades in SE Asia. We hypothesized that the decrease in forest cover and increase in forest degradation would result in a steady loss of habitat for forest birds over the years. We obtained presence information from eBird on 20 species representing birds vulnerable to changes in forest habitat. We parameterized Species Distribution Models (SDMs) using eBird location data and predictor variables derived from remotely sensed imagery, including values extracted from unclassified Landsat reflectance bands and values generated using the Continuous Change Detection and Classification (CCDC) algorithm. Our predictor variables were selected to best detect changes in forest cover. We used Maxent within GEE to construct SDMs for each species and used AUC to evaluate the predictive performance of all projected distributions. Species distributions were back-cast for each year of available Landsat imagery (1987-2022), resulting in a 35 year record of distributional changes for each species. Our results show a contraction in available habitat for forest-associated birds. We also mapped shifts in the distribution of available habitat for each species, which can be used to identify areas of ongoing habitat loss and areas which should be prioritized for conservation. Future work will expand on our preliminary 20 species to include additional birds, which will provide a more in-depth understanding of species’ responses to habitat loss in SE Asia. We will use the final suite of SDMs to explore the roles of community managed land, non-community managed land, and protected areas in promoting biodiversity. Our results represent an important contribution to informing conservation practices and policies at both a regional and global level.
