Research overview: predictive risk maps for human-carnivore conflict

This overview reviews the use of statistical and machine-learning models to build predictive risk maps that forecast where big-cat conflict is likely, rather than only describing past incidents. It synthesises approaches correlating depredation or attack records with environmental and social predictors, distance to forest, terrain, vegetation cover, livestock density, prey availability, and human settlement patterns, to estimate spatial probability of conflict. The literature discusses model validation, transferability, and the integration of risk maps into proactive planning and early warning. The survey frames predictive mapping as a maturing field that aims to anticipate conflict and pre-position prevention, while cautioning about data quality, changing conditions, and the gap between statistical risk and on-the-ground action.