Deep Learning Prediction of Human–Primate Conflict During Seasonal Resource Scarcity

Abstract

Human encroachment has continued to increase in the agricultural and forest-fringe landscapes, human–primate conflict has grown. The availability of natural food and water sources plays a significant role in this conflict, and during dry seasons, transitional seasons, or harvest seasons, primates have been known to migrate to crop fields and orchards or even to human settlements when natural food and water sources are limited. This paper proposes a conceptual framework on the use of deep learning as an analytical tool to proactively predict HPC events due to seasonal resource scarcity rather than a reactive approach. The framework combines environmental factors like natural food and water shortages, vegetation status, and forest and water body proximity with social and agricultural factors like the maturity stage of the crop, the conflict rates in the past, and the intensity of human activity. A deep learning classification technique is used to conceptually fuse these indicators to classify time periods and locations into low-, moderate-, and high-conflict-risk classes. The paper presents expected seasonal conflict patterns, which indicate that risk is expected to be highest during times of high natural resource shortages and during the peak crop maturity (late dry season and during harvest), and lowest during the times of high natural resource availability (wet season). A conceptual risk classification table, and an illustrative seasonal conflict-frequency graph have been provided to facilitate interpretation of these patterns. The consequences of this predictive strategy for conservation and community management are explored, from seasonal early-warning systems to resource-based habitat interventions, and community-centre non-harmful mitigation strategies. The framework aims to move conservation authorities and communities away from using reactive control measures toward taking action that is preventative, seasonally appropriate, and humane to avoid and manage human–primate conflict.