Modelling saproxylic rove beetle abundance from forest inventory data: Habitat heterogeneity as the key recommendation

Abstract Saproxylic insects are among the most endangered species in Europe due to their habitat specialization, slow habitat formation and declining habitat availability. Understanding how environmental factors shape their abundance is essential for conservation. The objectives of this study are to rank the influence of local environmental factors on the abundance of dominant Staphylinidae species, to characterize species‐specific predictor–abundance relationships and to evaluate the performance of species abundance models. We built models using Multivariate Adaptive Regression Splines (MARS) and Least Absolute Shrinkage and Selection Operator (LASSO). Our analysis centred on the single‐species abundance of 21 dominant beetle species (2947 individuals) collected from a temperate lowland forest, using 12 predictors defined within study plots or derived from Principal Component Analysis. The models performed well for seven to nine species depending on the method used, enabling current abundance predictions. Other models described the data adequately but underperformed in cross‐validation and bootstrapping, likely due to small sample sizes and low species prevalence below the defined thresholds of 136 individuals and 62.7%. Modelling revealed species‐specific responses characterized by nonlinear relationships, with threshold effects between abundance and predictors and interactions among them. The important predictors are stump density and, to a lesser extent, dead wood composition and living tree diversity. Conservations should aim to create a mosaic of habitat features, aligning with the broader principles of the Habitat Heterogeneity Hypothesis. This study enhances understanding of factors affecting beetle abundance and offers insights for other beetles, along with guidance for abundance modelling.