Fig 2 . Boxplots show the AUC (area under the curve of a receiver operating characteristic of extinction events) of models for three European breeding birds using different dimensions of climate change. The results are based on four machine-learning techniques and 50 replications.
The association between potential range loss and climate change metrics reveals that different dimensions of climate change could be linked with species losses across Europe. The importance of metrics varies among species. For Milvus milvus (Red kite), for example, the average temperature in spring (TempSpring), trend in precipitation (TrendPrec) and Trend in temperature (TrendTemp) and extreme events (ExtremeEvents) are the four most important variables linked with to climate suitability loss (Fig3.a). For Merops apiaster (European bee-eater) the top four most important variables are: Extreme events (ExtremeEvents); trend in temperature (TrendTemp); Velocity of climate change (Velocity); and temperature in spring (TempSpring) (Fig3.b). For Cettia cetti(Cetti’s warbler), four variables rank on top: Changes in probability of local extremes (ExtremeEvents); standardized local anomalies (sTLocA); trend in precipitation (TrendPrec); and average precipitation in spring (Prec.Spring) and (Fig3.c).
These findings emphasize that different bird species may respond differently to climate change, and certain climate change metrics are more relevant for predicting climate change susceptibility for each species. Understanding species-climate associations is vital for conservation efforts and for tailoring strategies to mitigate the impact of climate change on different species.