Bayesian spatial prediction of three medically important tick species in Illinois

Tick-borne diseases are reported across much of Illinois, yet tick surveillance remains spatially uneven, leaving many counties with limited species-specific observations. We developed Bayesian hierarchical spatial models to predict county-level relative abundance of Amblyomma americanum, Dermacentor variabilis, and Ixodes scapularis using surveillance data from 2019-2022. Relative abundance was defined as the species-specific share of the collected focal ticks and modeled using beta-binomial models that incorporated climate, land cover, forest fragmentation, deer habitat suitability, and BYM2 spatial effects in INLA. Spatial five-fold cross-validation evaluated whether environmental information improved prediction beyond spatial structure alone. A. americanum showed the highest predicted relative abundance in southern Illinois, with additional elevated areas in central-western counties, and deer habitat suitability was its clearest environmental predictor. D. variabilis was broadly distributed and best represented by the spatial-only model. I. scapularis was concentrated mainly in northern Illinois, with the climate providing the strongest environmental signal. Statewide predictions extended inference to counties lacking species-specific observations and identified where uncertainty remained greatest, providing a basis for prioritizing additional surveillance across Illinois.