High-resolution prediction and risk mapping of Haemaphysalis longicornis habitat suitability in the Jiaodong Peninsula, China.

Ticks are vectors of numerous human pathogens, and accurate predictions of their spatial distribution are essential for the effective prevention and control of tick-borne diseases. High-resolution predictions remain scarce in China. To address this gap, a tick survey was conducted across the Jiaodong Peninsula at 2,647 points during the summer of 2024, integrating high-resolution environmental variables. Seven machine-learning algorithms were used, and ensemble learning was applied, resulting in a habitat-suitability map for Haemaphysalis longicornis at a 10-meter resolution. The ensemble model showed robust predictive accuracy (AUC = 0.8355). Summer normalized difference vegetation index was the primary environmental driver, followed by land use and land cover and winter normalized difference vegetation index. High risk villages accounted for 70.0%, 40.2%, and 16.9% of rural, urban rural fringe, and urban areas, respectively. Ensemble machine-learning provides a robust approach, revealing how ecological factors and human activities shape distribution patterns and risk heterogeneity.