Off-highway vehicle route density maps for the Mojave Desert, USA

These datasets are published as a complement to:Robillard, A.J., M. Standen, N. Giebink, M. Spangler, A.C. Collins, B. Folt, A. Maguire, E.M. Olimpi, B.G. Dickson (2025). Application of computer vision for off-highway vehicle route detection: A case study in Mojave desert tortoise habitat. Remote Sensing in Ecology and Conservation. https:/doi.org/10.1002/rse2.70004.Our study leverages an advanced computer vision model to detect off‐highway vehicle (OHV) routes in the Mojave Desert, a critical habitat for the threatened Mojave desert tortoise (Gopherus agassizii). By analyzing historical (1970–1989) and contemporary (2010–2022) aerial imagery, we generated map-based outputs of OHV route density with a model accuracy of 77%.These datasets include raster maps of OHV route density for the 1970s, 1980s, 2010s, and 2020s time steps. We applied a focal window with a 1km radius and used the mean value to smooth the raw data (raw data values: 0=no routes, 1 = low route density, 2 = medium route density, 4 = high route density). OHV route density is presented as a continuous variable ranging from 0 (no routes) to 4 (high route density). Our methods and inferences are described in Robillard et al. (2025).We have also included a sample of images used to train our computer vision model. These files are raw historical aerial imagery for the 1970s and 1980s from the BLM National Operations Center (NOC) and the Nationwide Environmental Title Research, LLC (NETR) and contemporary aerial imagery for 2003–2022 from the U.S. Department of Agriculture National Agriculture Imagery Program (NAIP).