The conservation of threatened and rare species in remote areas often presents two challenges: there may be unknown populations that have not yet been documented and there is a need to identify suitable habitat to translocate individuals and help populations recover. This is the case of the reticulated giraffe (Giraffa reticulata), a species of high conservation priority for which: (a) there may be unknown populations in remote areas, and (b) detailed maps of suitable habitat available within its range are lacking. We implemented a species distribution modeling (SDM) workflow in Google Earth Engine, combining GPS telemetry data of 31 reticulated giraffe with Landsat 8 OLI, Advanced Land Observing Satellite Phased Arrayed Lband Synthetic Aperture Radar, and surface ruggedness layers to predict suitable habitat at 30-m spatial resolution across the potential range of the species. Models had high predictive power, with a mean AUC-PR of 0.88 (SD: 0.02; range: 0.86–0.91), mean sensitivity of 0.85 (SD: 0.04; range: 0.80–0.91), and mean precision was 0.81 (SD: 0.02; range: 0.79–0.83). Model predictions were also consistent with two independent validation datasets, with higher predicted suitable habitat values at known occurrence locations than at a random set of locations (P < 0.01). Our model predicted a total of 5519 km² of potentially suitable habitat in Kenya, 963 km² in Ethiopia, and 147 km² in Somalia. Our results indicate that is possible to combine moderate spatial resolution imagery with telemetry data to guide conservation programs of threatened terrestrial species. We provide a free web app where managers can visualize and interact with the 30 m resolution map to help guide future surveys to search for existing populations and to inform future reintroduction assessments. We present all analysis code as a framework that could be adapted for other species across the globe.
