Spatial, environmental and landscape variables
The 16 predictor variables included in the study were (see Appendix A1):
  1. Spatial : LAT: mid-gridcell latitude (°absolute); ALT: area of high altitude land per gridcell as % gridcell area >1000 m above sea level, a.s.l.;
  2. Environmental : ET0 (mm): potential evapotranspiration; AI (ratio between precipitation and ET0 × 10,000): aridity index; TYR (°C): average annual temperature; TMX (°C): maximum temperature of warmest month; TRG (°C): maximum temperature of warmest month – minimum temperature of coldest month; TDRY (°C): average temperature of driest quarter; PCP (mm): annual precipitation; PCPDR (mm): precipitation of driest quarter; PCPS: precipitation seasonality (coefficient of variation of monthly precipitation); CCV (m year-1): historic (Late Quaternary) climate change velocity;
  3. Landscape : grAH (km2): area of aquatic habitat present per gridcell; CROP (% agricultural land cover per gridcell);
  4. Biotic : Stot: total macrophyte species richness; Send: species richness of ecozone-endemic macrophytes, both as number of species per gridcell.
Climate variables were obtained from the Bioclim project source (www.worldclim.org/data/bioclim.html) and downloaded at 30 arc-seconds resolution, except for ET0 and AI, which were obtained from Trabucco and Zomer (2019), who derived them (also 30 arc-seconds resolution) from the Worldclim dataset; CCV spatial dataset was obtained from Sandel et al. (2019); water (grAH) and agricultural (CROP) cover data from the USGS Global Land Cover Characterization (GLCC, https://doi.org/10.5066/F7GB230D), and altitude from the USGS Global Multi-Resolution Terrain Elevation Data 2010 (https://topotools.cr.usgs.gov/GMTED_viewer/). Mean, median and standard deviation for climate variables per gridcell were obtained using a 10° gridcell shapefile and the Zonal statistics tool in QGIS (version 3.4.9-Madeira). The area occupied by each agricultural land use category (and subsequently total agricultural land cover) per gridcell was obtained using the above-mentioned shapefile and the Zonal Histogram tool in the ArcMap™ Spatial Analyst toolbox using ArcMap™ v. 9.3.1, see Murphy et al. (2019, 2020) for further details. All-species richness and ecozone-endemic species richness values were obtained from Murphy et al. (2019).