GBIF filtered datasets used for defining the species climatic niche in Batllori et al. "Effects of niche marginality on hot drought tree mortality in angiosperms and gymnosperms" Each file (.Rdata file) is designated by the name of a species and contains two columns, longitude (long) and latitude (lat) in WGS1984 coordinate system. For most species, GBIF data was downloaded in July 2020 using the dismo R package and filtered following the steps described below. Data for the following species was manually downloaded Betula pendula, Fagus sylvatica, Olea europaea, Picea abies, Pinus halepensis, Pinus pinaster, Pinus sylvestris, Quercus robur.

Ocurrence records filtering: 1. removing fossil specimes 2. removing "duplicate" species on the basis of 'speciesKey' (if more than speciesKey one was present the one with more records was retained) 3. when available, filtering for coordinate uncertainty in meters (for a 1x1 km grid the diagonal value is 707). Note that NA values were retained and that when such filtering yielded very low sample (<100 records) this step was omitted. 4. filtering occurrences registered after 1950. Note that when such filtering yielded very low sample (<100 records) this step was omitted. 5. removing duplicates based on geographic coordinates 6. selecting only one observation in each 1x1km pixel (using the gridSample function from dismo package) 7. retaining only lon-lat coordinates

This derived dataset was created following GBIF instructions (https://www.gbif.org/derived-dataset/about).

The file "GBIF_citation_Derived_dataset.csv" contains the datasetKey and number of records retained in the filtered datasets for all species.