Drivers of tick distributions in Latvia: a multi‑species assessment of joint and single‑species distribution models (2017-2019).

The ecological factors shaping the distribution of key tick vector species remain only partially understood. We analysed the spatial patterns of three medically important ticks, Dermacentor reticulatus, Ixodes persulcatus, and Ixodes ricinus in Latvia, using extensive field records from a single survey combined with environmental predictors harmonised to a 1 km spatial resolution. We pursued two complementary objectives: (i) to model and map the spatial distribution of three epidemiologically important tick species in Latvia, and (ii) to evaluate whether a joint species distribution modelling framework provides additional inferential or predictive benefit compared with single‑species models. The joint species distribution modelling analysis did not yield stable residual association estimates across fitting approaches, so subsequent inference and spatial prediction focused on single-species distribution models using Random Forest, Boosted Regression Trees, and Support Vector Machines under spatial cross-validation. Model performance differed among species, with highest predictive accuracy for I. persulcatus (mean AUC ≈ 0.93), moderate performance for I. ricinus (≈ 0.81), and lower but above-random performance for D. reticulatus (≈ 0.69). Temperature- and moisture-related predictors, including satellite-derived indicators of seasonal land surface temperature dynamics, were consistently influential. Response curves revealed strong non-linear and threshold relationships, with D. reticulatus showing relatively narrow climatic optima, I. persulcatus associated with cooler and more humid conditions, and I. ricinus exhibiting broader environmental tolerances. Final habitat suitability maps showed clear spatial differentiation: I. ricinus had the broadest area of high suitability, I. persulcatus displayed a more regionally structured pattern, and D. reticulatus was restricted to patchier areas. Our results suggest that, at the spatial scale considered, strong environmental filtering dominated species distributions, limiting the added value of joint modelling relative to more stable single‑species approaches. Together, these results provide a spatially robust and ecologically grounded distribution assessment for ticks in Latvia. The findings support public-health surveillance, identify regions of elevated potential exposure, and highlight the need to investigate additional ecological mechanisms, beyond species interactions, that shape tick communities across the Baltic region.