A study predicts increasing human-elephant conflict in Southern Africa. A growing number of farmers and 290,000 African savanna elephants ( Loxodonta africana ) share space in Southern Africa, with conflicts arising from elephants raiding cropland. Crop raids by elephants can be financially devastating for farmers.
Evan Patrick and colleagues used both causal inference statistical methods and machine learning models to analyze a dataset of crop raiding events across Namibia’s communal conservancies from 2004 to 2020 to determine the predictors of human-elephant conflict. The authors used these event data to identify trends across a wider area, including northern Botswana and portions of Angola and Zambia in addition to Namibia, to evaluate the drivers of conflict.
The analysis identifies human population growth, cropland expansion, and climate-driven aridity as major drivers of increasing rates of crop raiding. The authors also mapped the probability of conflict throughout the study area. Key variables for these maps include tree cover, distance to roads, distance to fences, distance to rivers, human population density, and productivity of vegetation. The models predict a general increase in the probability of crop raiding toward the end of the century under all climate change scenarios in both wet and dry seasons, with the area at risk of crop raiding doubling under the change climate scenarios. Increasing human land use will continue to place pressure on elephants even as climate change reduces their wild food supply. According to the authors, the model’s predictions can inform the proactive land use planning and mitigation measures that will be essential for long-term coexistence between humans and elephants.
PNAS Nexus
An expanding human footprint drives escalating human–elephant conflict across a transboundary African landscape through 2085
7-Jul-2026
E.P. has a consulting relationship with Carbon Direct Inc., a company combining science, technology, and capital to deliver quality carbon dioxide management at scale. The other authors declare no competing interests.