By: Harry Jon Foord Owen, Clare Duncan and Nathalie Pettorelli
Abstract
Deserts are among the most poorly understood biomes in the world, currently experiencing among the highest rates of environmental change and biodiversity loss. A major controlling factor on the ecology and distribution of vegetation and animal populations in these harsh arid systems is the abundance and distribution of water sources. Accordingly, extraction and redistribution of water at artificial water points across desert landscapes can constitute a real threat to local ecosystem dynamics. A major challenge in tackling this potential threat is identifying changes in the distribution of artificial water points through space and time, due to the difficulties of collecting such information at relevant spatial and temporal scales. We here investigate the potential for freely available satellite imagery to provide reliable information about the distribution of artificial water points using the Ouadi Rime-Ouadi Achim Faunal Reserve (OROAFR), Chad, as a case study. We reveal that Landsat 8 data combined with texture analysis can accurately detect these artificial landmarks across the heterogeneous environment of OROAFR; 75.68% of artificial water points within the site were successfully identified and false positive detection rate was minimal at 7.69%. The methodological framework developed for this work, based on the treatment of freely available satellite data using open source software, adds to other works attempting to help monitor threats to biodiversity in desert ecosystems, enabling up-to-date information on the level of anthropogenic activities in these habitats to be easily and regularly collected.
