Fig. 5

Comparison between GSSP and GBIF data. The upper panels show a visual comparison between GSSP data and GBIF data exemplified for a species with a match better than expected at random (panel A: Blumeria graminis, correlation = 0.50, p-value 0.04), and for a species with a match worse than expected at random (panel B: Phlebia subserialis, correlation = −0.39, p-value > 0.99). For GBIF data, all occurrence records are shown in green circles. For GSSP data, all sampling locations are indicated as a blue circle, including locations where the species was not observed. In locations where the species was observed, the size of the red circle shows the proportion of samples in which the species was observed. The lower panels (C) show a statistical comparison for all 1,251 species included in the analysis. The p-value shows the statistical significance of the comparison, with small p-values corresponding to cases where the GBIF data for the focal species was more consistent with the GSSP data for the focal species than with the GSSP data for a randomly selected non-focal species. The effect size shows the correlation between the GBIF data and GSSP data for each focal species. In both panels, the red line highlights the null expectation based on no consistency between the GBIF and GSSP dataset, indicating that for the majority of the species, the GBIF and GSSP datasets match much better in their spatial distributions than expected by random. The frequency bins into which the species exemplified in panels A and B fall are highlighted with letters A and B in panel C.