Authors: Arenas-Castro, Salvador; Regos, Adrián;Martins, Ivone;Honrado, João;Alonso, Joaquim;

While data combinations from different sources proved to be a useful tool in species distribution models (SDMs), there is still much to explore on the usefulness and/or uncertainty of each input data source through the modelling process. We assessed the effects of uncertainty in SDMs using different data sources in both response (citizen-collected and standardized datasets) and predictor (macroclimate and remotely sensed) variables on SDM performance across a wide range of bird species (236) with contrasting distributional ranges in the Iberian Peninsula. We implemented a SDM ensemble-forecasting approach by using the occurrences of bird species grouped in four range size classes from the semi-structured eBird project and standardized Atlases, and by using three predictor types: climate, remotely sensed ecosystem functional attributes (EFAs), and their combination. Based on generalized linear mixed-effects model results, data source, size class and predictor showed significant effects on SDM performance. eBird-based models outperformed those built with Atlas data for less widespread (rare) species, and the combined climate-EFA predictors yielded models with the best performance. Our findings confirmed that less widespread species with clear conservation concerns benefited from the inclusion of citizen science data and other key environmental factors into SDMs, with strong implications for species conservation.