Accuracy versus realism: static and dynamic species distribution models reveal complementary dimensions of biodiversity under climate change

Aim: Biodiversity is inherently dynamic, yet classical Species Distribution Models (SDMs) assume climate stability (static approach) and frequently overlook temporal variability, especially under climate change. Species, however, respond in diverse ways, often constrained by their dispersal capabilities. In this study, we compared the performance of static and dynamic SDMs, to assess the efficiency of the protected areas for conserving Romanian herpetofauna. Location: Romania. Time Period: Contemporary climate conditions. Major Taxa Studied: Amphibians and reptiles. Methods: We developed optimized Maxent-based static and dynamic SDMs to predict the potential distributions of Romanian herpetofauna. Model performance was evaluated using standard accuracy metrics, while differences in niche structure were assessed using Principal Component Analysis and niche equivalency tests across the available climate space. We further quantified differences in niche breadth and overlap using hypervolume approaches. Conservation relevance was assessed through gap analysis comparing predicted suitable areas with existing protected areas. Results: Static SDMs achieved higher predictive accuracy and produced smaller, more compact niches nested within the dynamic niches. In contrast, dynamic SDMs captured spatio-temporal variability and extreme climatic conditions, resulting in expanded niche estimates and greater representation of range-edge environments. However, dynamic models relied more on extrapolation, leading to increased variability, overprediction of suitable habitats, and reduced performance scores. These results reveal a clear trade-off between predictive accuracy and ecological realism. Furthermore, gap analysis based on static models indicated moderate to high coverage of species-rich areas and threatened reptile. Main Conclusion: Static and dynamic SDMs provide complementary perspectives on species distributions under climate variability. Static models offer conservative and precise estimates of core habitats, whereas dynamic models better capture environmentally variable and marginal conditions. Integrating both approaches can improve the identification of core and range-edge habitats, supporting more robust and ecologically realistic conservation strategies in a changing climate.