A Spatial Probability Model for the Arrival and Establishment of Coconut Rhinoceros Beetle

Coconut rhinoceros beetle (CRB, Oryctes Rhinoceros , L.) is a destructive pest of coconut and oil palms globally. Native to Southeast Asia, the pest has progressively invaded the Pacific and more recently non-Pacific regions, causing severe ecological and economic damage. After decades of suppression through biological control with a viral pathogen, a new wave of invasions began in 2007, largely driven by human-mediated dispersal. We describe a spatial probability model for the introduction and establishment of CRB by integrating ecological suitability and human-mediated pathways that address critical gaps in existing CRB risk assessments. The model was first developed and tested using data from Vanuatu and subsequently validated with independent datasets from Solomon Islands and Mexico. The model’s performance was evaluated using multiple approaches, including correlation analysis, risk distribution metrics, precision-recall curves (PR AUC) and spatial risk maps. Predictor variables showed significant correlations with CRB invasion score, with population and port connectivity being dominant drivers in Vanuatu and the Solomon Islands, while elevation suitability was more influential in Mexico. Probability score distribution revealed that most Global Biodiversity Information Facility (GBIF) presence points fell above the 90th percentile, and none had zero CRB scores, aligning with model expectations. Similarly, known historical invasion areas for all three countries closely aligned with high-risk sites predicted by the current model. The model achieved high predictive performance, with PR AUC values exceeding 0.99 across all regions.