Balancing structural complexity with ecological insight in Spatio‐temporal species distribution models

The potential for statistical complexity in species distribution models (SDMs) has greatly increased with advances computational power. Structurally complex provide the flexibility to analyse intricate ecological systems and realistically messy data, but can be difficult interpret, reducing their practical impact. Founding model theory improve insight gained from SDMs. Here, we evaluate a marked point process approach, which uses multiple Gaussian random fields represent population dynamics of Eurasian crane Grus grus spatio-temporal model. We discuss role components impacts on predictions, comparison simpler binomial presence/absence approach. Inference is carried out using Integrated Nested Laplace Approximation (INLA) inlabru, an accessible computationally efficient approach Bayesian hierarchical modelling, not yet widely used Using was predicted dependent density suitable habitat patches, as well close observations existing population. This demonstrates advantage accounting (such preferences dispersal limitations) that are explained by environmental variables. However, including AR1 temporal correlation structure resulted unrealistic predictions distribution; highlighting need careful consideration when determining level complexity. Increasing complexity, evaluation effects additional components, more realistic representation system, particular importance impact-focused discipline such ecology (though these methods extend applications wide range systems). contextual only fundamental maintaining interpretability useful improving outputs.