Previous studies have used separate regional models of species distribution and abundance to address shifting species–environment relationships. These models, based on hunting yield data, incorporate regional effects and improve reliability, but they often fail to capture continuous spatial gradients. This report evaluates a new approach for modelling gradual spatial changes in wild boar relative abundance using spatially explicit models. We compared downscaled hunting yield predictions from two models—an environmental-only model and a spatially explicit model—with estimates from the ENETWILD consortium (2022). Hunting yields were modelled using a Negative Binomial framework with covariates and a spatial random effect implemented via a Nearest Neighbour Gaussian Process (NNGP), generating predictions on a 10×10 km grid across the species’ range. Model performance was assessed by comparing total predicted harvests, uncertainty (credible vs. confidence intervals), spatial agreement (Spearman correlations), and country-level results for Spain, France, Italy, and Poland to account for differences in data resolution. Results indicate that spatial random-effect models produce smooth continental patterns but are limited by non-stationarity and heterogeneous data resolution, leading to unreliable predictions in areas with coarse or uneven spatial units. Environmental and spatial models capture different aspects of wild boar distribution, underscoring the need to explicitly account for regional variability when modelling at a European scale. The ENETWILD-consortium et al. (2024) model, which calibrates hunting bag data into wild boar density estimates, shows the strongest correlation with independent references and is currently the most reliable option for risk assessments that require spatially explicit wild boar density data. High-resolution hunting statistics are critical for producing realistic spatial patterns; where data resolution is low, models generate unreliable results, whereas areas with improved, high-resolution data yield spatial patterns consistent with ecological expectations.

EU, pdf, biohaw@efsa.europa.eu