Bird Conservation Regions (BCRs) are ecologically distinct areas with similar bird populations, ecological characteristics, and natural resource management concerns. Understanding vegetation distributions and dynamics may guide land management and conservation decisions, but current information available in the Sonoran and Mojave Deserts Bird Conservation Region (BCR 33) are insufficient due to the area’s low vegetative cover and complex vegetation communities, and the discontinuity in data availability across the United States-Mexico border. This research produced the first high-resolution, continuous United States-Mexico land cover map of BCR 33 by prototyping new methods for desert vegetation classification using a random forest (RF) machine learning (ML) approach. The RF classification model was based on multitemporal Landsat 8 Operational Land Imager spectral and vegetation index data from 2013–2020, augmented with phenology metrics that capture the unique growing seasons of desert vegetation. This RF model can classify 31 different land cover class types within the Sonoran and Mojave Deserts BCR.
The model achieved an overall classification f-score of 0.80 and an overall accuracy of 91.68%. Results portrayed the vegetation cover at a much finer resolution than existing land cover maps from the United States and Mexican portions of the study area. This approach helped identify small habitat pockets, including riparian communities, which are critically important for desert wildlife and are often misclassified or nonexistent in current maps. Results supported the assumption that land cover labels provided within the United States portion of the study area were representative of vegetation south of the border, allowing for the extension of this classification system into the Mexican region of BCR 33. While these results were mostly consistent with respect to existing classification of the region, they did highlight the need for a unified land cover classification scheme across the United States and Mexican portions of BCR 33. This application and tools serve as a resource for the stakeholders, regional agencies, and non-governmental organizations involved in this research in support of habitat conservation and land management efforts within the sensitive Sonoran and Mojave Deserts BCR. The associated manuscript is available here.
Land cover map of the Bird Conservation Region 33 subregion study area in the Sonoran and Mojave deserts. The associated manuscript is available here.
Disclaimer: Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.