Species data and Environmental variables

We have submitted table of studied ectomycorrhizal (ECM) fungi including their scientific names and host specificity (Fungal_traits.csv), environmental variables in ECM fungal and host trees distribution modelling (nitrogen.tif, ocs.tif, ph.tif, Current_Bioclim.zip, 2050_SSP126_Bioclim.zip, 2050_SSP370_Bioclim.zip, 2050_SSP585_Bioclim.zip, 2080_SSP126_Bioclim.zip, 2080_SSP370_Bioclim.zip, 2080_SSP585_Bioclim.zip), and R scripts (SDM_of_ECM_fungi.zip and SDM_of_hosts.zip).

Descriptions:

1. ECM fungal data

   ECM fungal data includes fruitbody and root occurrence data.

   Fruitbody data (including human observations and preserved specimens) were obtained from the Global Biodiversity Information Facility (GBIF) and PlutoF API by using provided species names. GBIF occurrence data were accessed using the rgbif R package. These data were filtered and cleaned by following the recommendations of Zizka et al. (2019, 2020). PlutoF (https://plutof.ut.ee/) is a biodiversity data platform providing curated fungal occurrence records. Data from PlutoF API were cleaned by removing records without coordinate information, collected before 1971 or collected outside Europe.

   Root data were obtained from van der Linde et al., 2018, based on samples collected from 136 plots within ICP Forests. 

   Fungal_traits.csv:

   • Column 1: Species - Scientific name of ECM fungal species (as used in GBIF/PlutoF).
   • Column 2: Host_specificity - Host association category of each fungal species. Categories include: 1) 'broadleaved', means ECM fungi only associated with broadleaved trees; 2) 'coniferous', means ECM fungi only associated with coniferous trees; and 3) 'generalist' means ECM fungi associated with both broadleaved and coniferous trees.

2. Host trees data

   Five host tree species were used in this study: Fagus sylvatica, Picea abies, Pinus sylvestris, Quercus petraea, and Quercus robur. 

   Occurrence data for these species were downloaded GBIF using the species names listed above. Data were filtered by retaining records from 1971 onwards, excluding records without geographic coordinates, and restricting records to Europe. The same data cleaning procedures as described for the ECM fungal GBIF data were then applied.

   Note: The host tree occurrences data are not included in this Dryad dataset due to licensing restrictions associated with GBIF data. Users can reproduce this part of the analysis by downloading the data directly from GBIF using the species names and filtering steps described above.

3. Abiotic environmental variables in ECM fungal and host trees distribution modelling

   All .tif files represent raster layers used as environmental predictors in species distribution models (SDMs). Each pixel contains the value of the corresponding environmental variable.

   These raster datasets can be opened and explored using open-source GIS software such as QGIS. They can also be imported and processed in R (e.g., with the raster or terra packages).

   Spatial properties:

   Projection: WGS84 (EPSG:4326)

   Resolution: 2.5 arc-minutes (~5 km)

   Extent: Western Europe (W-11, E41, S35, N72)

   Soil variables:

   • nitrogen.tif: Soil total nitrogen (g/kg) at 5–15 cm depth (original resolution: 250 m in SoilGrids, resampled to 2.5 arc-minutes).
   • ocs.tif: Soil organic carbon stock (kg/m²). Data were downloaded from an earlier version of SoilGrids where a 5–15 cm depth layer was available. This dataset was retained to ensure consistency with other soil variables used in the analysis.
   • ph.tif: Soil pH (H₂O) at 5–15 cm depth (original resolution: 250 m in SoilGrids, resampled to 2.5 arc-minutes).

   Climate variables (CHELSA Bioclim v2.1):

   Current and future bioclimatic variables were obtained from the CHELSA dataset (version 2.1) and resampled to the same spatial properties of soil variables.

   Current_Bioclim.zip: 

   • bio1.tif: Mean Annual Near-Surface Air Temperature (℃).
   • bio2.tif: Mean diurnal temperature range computed as the average of monthly (tasmax - tasmin) (℃).
   • bio3.tif: Isothermality (℃): .
   • bio4.tif: Temperature seasonality given by the standard deviation of mean monthly temperatures (℃/100).
   • bio8.tif: Average monthly mean temperature over the wettest 3-month period of the year (℃).
   • bio12.tif: Annual Precipitation (kg m-2 month-1),  downloaded from CHELSA-bioclim (Version 2.1).
   • bio15.tif: Coefficient of variation (100 × SD ÷ mean) of monthly precipitation totals (kg m-2).
   • bio18.tif: Average monthly precipitation during the warmest 3-month period of the year (kg m-2 month-1).
   • bio19.tif: Average monthly precipitation during the coldest 3-month period of the year (kg m-2 month-1).

   Future climate datasets:

   Future projections were derived from the GFDL-ESM4 climate model for two periods:

   • 2041–2070
   • 2071–2100

   Under three shared socioeconomic pathways (SSPs):

   • SSP126: sustainability (low challenges to mitigation and adaptation)
   • SSP370: regional rivalry (high challenges)
   • SSP585: fossil-fuelled development (high emissions scenario)

   Each future dataset includes the same variables as Current_Bioclim.zip.

4. R scripts

   SDM_of_ECM_fungi.zip: 

   Contains R scripts for modelling the current (baseline) and future distributions of ECM fungi under different host association scenarios (single host, two hosts, and generalists). Future projections include three shared socioeconomic pathways (SSP126, SSP370, SSP585) for two time periods (2041–2070 and 2071–2100).

   SDM_of_hosts.zip:

   • Baseline_of_hosts.R: current distribution modelling of host tree species.
   • Future_of_hosts.R: future projections (2041–2070 and 2071–2100 under SSP126, SSP370, SSP585).

Software and packages:

All analyses were conducted in R (version 4.3.3). RStudio is recommended for running the scripts. 

The CoordinateCleaner R package was used for GBIF data cleaning.

The biomod2 R package was used to model the current and future distributions of ECM fungi and host tree species.

Additional R packages used in this study are documented within the provided R scripts.