Data and code from: Latitudinal gradients in position and climatology of upper tree limits across Europe
Data files
Jul 27, 2026 version files 1.27 MB
-
README.md
12.08 KB
-
UTL-DryadR2.zip
1.26 MB
Abstract
This dataset contains data and scripts associated with the manuscript “Latitudinal gradients in position and climatology of European upper tree limits”. Specifically, we provide shapefiles of positions of uppermost located trees in Europe, and grid cells where trees were mapped. Furthermore, we deliver R scripts created in order to extract bioclimatic variables for uppermost trees, and to assess latitudinal gradients in these variables. Using high-resolution imagery, we identified the current upper tree limits and calculated their bioclimatic metrics. Upper tree limits' elevation was modelled using latitude, diurnal temperature range, climatic water balance, and distance to summit. We analyzed trends in these bioclimatic variables along the upper tree limit latitudinal gradient and classified upper tree limit bioclimatic groups.
Dataset DOI: 10.5061/dryad.9w0vt4bvf
Description of the data and file structure
This repository is associated with Treml V., Kaczka R., Kalita J.Z., Mašek J., Tumajer J., Lange J., Romero E. (2026) Latitudinal gradients in position and climatology of upper tree limits across Europe, in Journal of Biogeography (https://doi.org/10.1111/jbi.70296).
In this study, we identified current upper tree limits across Europe and calculated their bioclimatic metrics, based on which upper tree limits reachingthe treeline isotherm were subset. Elevation of the upper tree limit and its offset from the treeline were modelled using latitude, diurnal temperature range, climatic water balance, distance to summit, and warming trends.
This repository contains the scripts and raw data necessary to perform all linear models, segmented regressions, and spatial regression models allowing prediction of the latitudinal gradient of the upper tree limit and its climatic metrics. Furthermore, scripts classifying European upper tree limits and scripts producing figures are presented. The software needed to run these scripts is R and RStudio. All software is open source and free to access. All data files and scripts are in the zipped file “UTL-DryadR2.zip”.
Data files
File: UTL-DryadR2.zip
Description:
Zipped archive contains all data files and scripts.
File: treel_F6-done.shp
Description:
Shapefile containing locations of the uppermost trees. All trees were identified and mapped based on aerial and satellite imagery through Google Earth Pro. Shapefile can be opened and edited in common GIS software or in R.
Variables:
- OBJECTID: automatic ID assigned to each tree
- Name: Unique tree code
- point_id: ID of tree within a gridcell
- tile: Grid cell; year of the image acquisition
- quality: Quality of image regarding the tree identification
- treeline: Elevation of tree based on Google Earth Pro elevation model
- conifer: Assignment of tree to conifers (1/0)
- broadleaf: Assignment of tree to conifers (1/0)
- forestline: Ebroaderleaf of nearest forestline based on Google Earth Pro elevation model
- grassland: Assignment of ground cover to grassland (1/0)
- bush: Assignment of ground cover to shrubland (1/0)
- screes: Assignment of ground cover to screes (1/0)
- Gradual: Treeline form is gradual (1/0)
- Abrupt: Treeline form is abrupt (1/0)
- Genus: Genus name of tree
- Species: Species name of tree
- dem_elev1: Elevation of tree identified from Copernicus DEM
- dif_tr_dem: Absolute value of difference treeline – dem_elev1
- _slopemean: Slope inclination in buffer around tree location
- _slopevari: Slope variability in buffer around tree location
- _curvmean: Curvature in buffer around tree location
File: gridcells_wgs3-done.shp
Description:
Shapefile containing gridcells in which uppermost trees were identified. Attribute table includes descriptive characteristics of each gridcell at a gridcell level. Shapefile can be opened and edited in common GIS software or in R.
Variables:
- OBJECTID_1: automatic ID assigned to each gridcell
- tile: Gridcell number
- _demmean: Mean elevation of gridcell
- _demmax: Maximum elevation of gridcell
File: kupdate2.csv
Description:
Table with tree characteristics determined from aerial imagery and elevation model – the primary data entering the analyses. Variables are separated by commas. The table isan extended attribute table of shapefile “treel_F6-done.shp”. The table can be opened and edited in any text editor or in R.
Variables:
- FID: automatic ID assigned to each tree
- OBJECTID: automatic ID assigned to each tree
- Name: Unique tree code
- point_id: ID of tree within a gridcell
- tile: Grid cell number; grid cell Year of the image acquisition
- quality: Quality of image regarding the tree identification
- treeline: Elevation of tree based on Google Earth Pro elevation model
- conifer: Assignment of tree to conifers (1/0)
- broadleaf: Assignment of tree to conifers (1/0)
- forestline: Elevation of nearest forestline based on Google Earth Pro elevation model
- grassland: Assignment of ground cover to grassland (1/0)
- bush: Assignment of ground cover to shrubland (1/0)
- screes: Assignment of ground cover to screes (1/0)
- Gradual: Treeline form is gradual (1/0)
- Abrupt: Treeline form is abrupt (1/0)
- Genus: Genus name of tree
- Species: Species name of tree
- dem_elev1: Elevation of tree identified from Copernicus DEM
- dif_tr_dem: Absolute value of difference treeline – dem_elev1
- X_slopemean: Slope inclination in buffer around tree location
- X_slopevari: Slope variability in buffer around tree location
- X_curvmean: Curvature in buffer around tree location
- lat: Latitude of tree (geographic coordinate, WGS84)
- long: Longitude of tree (geographic coordinate, WGS84)
- ecotone: Elevation difference treeline - forestline
File: TREELOC3.csv
Description:
Table with coordinates of all trees as derived from “treel_F6-done.shp”. Variables are separated by commas. The table can be opened and edited in any text editor or in R.
Variables:
- Name: Unique tree code
- Tile: Gridcell number
- dem_elevat: Grid cell number of tree identified from Copernicus DEM
- lat: tree latitude (WGS84)
- long: tree longitude (WGS84)
- Xcor: tree longitude (WGS84)
- Ycor: tree latitude (WGS84)
- Elev: Elevation of tree identified from Copernicus DEM (rounded)
- elev1: Elevation of tree identified from Copernicus DEM (rounded)
- CODE: Unique tree code
File: gridcells_wgs3-done.csv
Description:
Attribute table of “gridcells_wgs3-done.shp” converted to csv file. Variables are separated by commas. The table can be opened and edited in any text editor or in R.
Variables:
- OBJECTID_1: automatic ID assigned to each gridcell
- tile: Gridcell number
- _demmean: Mean elevation of gridcell
- _demmax: Maximum elevation of gridcell
File: cheldas.csv
Description:
Data table based on complete data set after step 7 of the description of code/software with reduced number of variables for the purpose of linear discriminant analysis and figure drawing. Variables are separated by commas. The table can be opened and edited in any text editor or in R.
Variables:
- ecotone: Elevation difference treeline – forestline (gridcell mean)
- X_slopemean: mean slope inclination at treeline (gridcell mean)
- X_slopevari: slope variability at treeline (gridcell mean)
- X_curvmean: curvature at upper tree limit (gridcell mean)
- lat: latitude of gridcell centroid
- long: longitude of gridcell centroid
- conifer: Assignment of tree to conifers (mean)
- broadleaf: Assignment of tree to broadleaves (mean)
- grassland: Assignment of ground cover to grassland (gridcell mean)
- bush: Assignment of ground cover to shrubs (gridcell mean)
- screes: Assignment of ground cover to screes (gridcell mean)
- sumitdist: Altitudinal distance between upper tree limit and the uppermost point in the gridcell (mean)
- clust1: Assignment of the given gridcell to cluster ID
File: Trends1.csv
Description:
Output table from script calculating temperature trends. Variables are separated by commas. Table can be opened and edited in any text editor or in R.
Variables:
- OBJECTID_1: automatic ID assigned to each grid cell
- tile: Gridcell number
- _demmean: mean elevation of gridcell
- _demmax: maximum elevation of gridcell
- Trend: slope of linear trend for growing season temperature
File: Chelsa-clim-gridcells.csv
Description:
File containing climatic variables characterizing upper tree limits at gridcell level (means from points within each gridcell), after extraction from CHELSA climatic surfaces.
Variables:
- tile: Gridcell number
- treeldem: mean elevation of upper tree limits in a gridcell derived from Copernicus DEM
- lat: latitude of gridcell centroid
- long: longitude of gridcell centroid
- gsl1: Growing season length (days)
- gst1: Mean growing season temperature (°C)
- dd51: Degree days exceeding 5°C
- ci1: Climatic water balance (mm)
- dtr: Diurnal temperature range over growing season (°C)
Additional data necessary to run the analyses
R scripts further needthe following data to properly run the analyses from the beginning; however, we also provide ready-to-use data files with variables extracted from the data sets listed below:
- CHELSA bioclimate datasets downloaded from https://chelsa-climate.org/
- GMTED2010 elevation model (USGS), which was used to derive CHELSA bioclimatology by Karger et al. 2017 (https://doi.org/10.1038/sdata.2017.122). GMTED2010 is released for public use without any restrictions and can be downloaded here: https://earthexplorer.usgs.gov/
- Era5 Land monthly temperature series were downloaded here: ttps://cds.climate.copernicus.eu/datasets/reanalysis-era5-land-monthly-means?tab=overview
Code/software
File: UTL-Europe_skript-dryad04.txt
Description:
Script is executable in R environment (R Core Team, https://www.R-project.org/, version 4.4.1 and newer). To run the scrip,t please copy all data files and downloaded climatic and elevation rasters (the names of variables are in the script) into a single folder and then set up the appropriate paths to files. Individual steps are described in detail in the script file.
The workflow is as follows:
1. The script first reads climatic data from the CHELSA bioclimate dataset;
2. We search for points with the smallest elevation difference between the GMTED elevation model (model used in CHELSA bioclimate data sets) and the COPERNICUS DEM (model used to identify the elevation of uppermost trees);
3. Climatic variables are extracted for points identified in step 2;
4. All climatic characteristics are aggregated at the grid cell level (resulting aggregates at grid cell level can also be uploaded separately from the file “Chelsa-clim-gridcells.csv”;
5. Site and tree characteristics (from kupdate2.csv) are also aggregated at the gridcell level and merged with climatic variables;
6. Clusters of grid cells are defined based on bioclimatic variables;
7. Topographic variables at grid cell level are added to the main dataset;
8. Temperature trends are calculated from the ERA5 land data set (provided also separately in “Trends1.csv”);
In further steps, we present scripts and analyses leading to the main figures presented in the manuscript.t
The following R packages have to be uploaded: terra, dplyr, tidyr, raster, ggplot2, cluster, NbClust, caret, factoextra, MASS, ggord, car, sf, spdep, spatialreg, relaimpo.
Access information
Other publicly accessible locations of the data:
- NA
Data was derived from the following sources:
- CHELSA bioclimate datasets downloaded from https://chelsa-climate.org/
- GMTED2010 elevation model (USGS), which was used to derive CHELSA bioclimatology by Karger et al. 2017 (https://doi.org/10.1038/sdata.2017.122). GMTED2010 is released for public use without any restrictions and can be downloaded here: https://earthexplorer.usgs.gov/
- Era5 Land monthly temperature series were downloaded here: ttps://cds.climate.copernicus.eu/datasets/reanalysis-era5-land-monthly-means?tab=overview