Causal attribution of temperature impacts on European breeding bird communities

Attributing drivers to ecological change and quantifying their effects is a central goal in ecology, yet remains challenging when relying on observational data. Temperature, for example, has been identified as a major driver of species’ range shifts, behaviour changes and population declines in many bird species, with consequences for community structure and diversity. However, temperature rarely varies in isolation and co-occurring factors - such vegetation changes, landuse shifts or habitat degradation - confound effect estimation if not correctly addressed. Accurately identifying the isolated contribution of temperature is therefore essential for conservation planning, forecasting models, and understanding ecological responses to global change. Causal inference is increasingly recognised as a valuable framework for addressing these challenges, combining promising statistical methods with causal reasoning. Moving beyond a focus on model choice, best practices emphasize explicit causal thinking, transparent assumptions, robustness checks, and the triangulation of evidence across complementary methods.

In this study, we use a Pan-European dataset of structured Breeding Bird Surveys to estimate the effect of temperature on bird community diversity over the past 15 years. We apply variants of two-way fixed-effects models to control for unobserved spatial and temporal confounding and quantify the responses of complementary diversity metrics towards both temperature mean conditions and measures of extremes. 

Preliminary results suggest that bird diversity responds most strongly - and generally positively - to changes in average breeding-season temperature. This effect was stronger concerning site-level anomalies rather than general temperature variation. Metrics of extreme temperatures show more heterogeneous effects and in most cases were linked to Bioregion-wide trends: higher maximum temperatures and greater interannual variability tend to reduce diversity, whereas increases in minimum temperatures seem to be associated with positive responses. By explicitly disentangling temperature effects from co-occurring environmental changes, our work demonstrates how causal inference can enhance the reliability in assessments of biodiversity change under a warming climate.