Tree regeneration is a key demographic process influencing long-term forest dynamics. It is driven by climate, disturbances, biotic factors, and their interactions. Thus, predictions of tree regeneration are challenging due to complex feedbacks along the wide climatic gradients covered by most tree species. We developed statistical models based on >50,000 recruitment events observed for 24 tree species in an extensive permanent plot network (6,540 plots from 299 unmanaged European temperate, boreal, and subalpine forests) covering a wide climatic gradient.

This repository stores csv files with response and explanatory variables and R-Scripts that are needed to run the analysis and create the figures that are presented in the accompanying article titled "Sheltered or suppressed? Tree regeneration in unmanaged European forests" which was published in Journal of Ecology in 2023.

Our study shows that competition between trees towards climatic stress decreases systematically but depends on species stress tolerance to climate and shade. These findings explain within- and between-species differences in tree recruitment patterns in European temperate forests. Moreover, our findings imply that projections of forest dynamics along wide climatic gradients and under climate change must accommodate both competition and positive interactions, as they strongly affect rates of community turnover.

These data are the processed model inputs for the article "Sheltered or suppressed? Tree regeneration in unmanaged European forests" (Journal of Ecology).

Recruitment (the number of trees surpassing a 7 cm or 10 cm diameter-at-breast-height threshold) and stand variables (e.g. total leaf area index) were derived per forest sampling plot and inventory period from an extensive permanent-plot network of unmanaged European forests. The environmental predictors are:

• dds — seasonal degree-day sum (Allen 1976; Fischlin et al. 1995)
• swb — site water balance (Speich 2019), calculated as described in Appendix A1 of the article

Plot, site, institute and inventory-period identifiers were anonymised (replaced by integer IDs) for publication. Full variable definitions, software requirements, and reproduction steps are given in the README.

Note (this version): the processed data were corrected for a plot-identifier misalignment present in the first deposit, which attached the climate predictors to the wrong plots in the DBH 7 cm analysis. See the README and the change log for details. The analysis code and the published results are unaffected.

These data are the processed model inputs for the article "Sheltered or suppressed?
Tree regeneration in unmanaged European forests" (Journal of Ecology).

Recruitment (the number of trees surpassing a 7 cm or 10 cm diameter-at-breast-height
threshold) and stand variables (e.g. total leaf area index) were derived per forest
sampling plot and inventory period from an extensive permanent-plot network of
unmanaged European forests. The environmental predictors are:

• dds — seasonal degree-day sum (Allen 1976; Fischlin et al. 1995)
• swb — site water balance (Speich 2019), calculated as described in Appendix A1 of the article

Plot, site, institute and inventory-period identifiers were anonymised (replaced by
integer IDs) for publication. Full variable definitions, software requirements, and
reproduction steps are given in the README.

Note (this version): the processed data were corrected for a plot-identifier
misalignment present in the first deposit, which attached the climate predictors to
the wrong plots in the DBH 7 cm analysis. See the README and the change log for
details. The analysis code and the published results are unaffected.

Full file descriptions, variable definitions, software requirements, and reproduction steps are provided in the README.

Use the corrected files:

data_corrected.zip — processed model inputs (data/processed_data/) and the raw inputs used to create the figures (data/raw_data/).

scripts_corrected.zip — the analysis code (scripts/fit_models.R and supporting scripts).

The first-version archives (data.zip, scripts.zip) have been replaced by the corrected files above and remain available under the first version of this dataset on Dryad.

To reproduce the analysis: extract data_corrected.zip and scripts_corrected.zip into the same working directory and run scripts/fit_models.R.

Full file descriptions, variable definitions, software requirements, and reproduction
steps are provided in the README.

Use the corrected files:

• data_corrected.zip — processed model inputs (data/processed_data/) and the
  raw inputs used to create the figures (data/raw_data/).
• scripts_corrected.zip — the analysis code (scripts/fit_models.R and
  supporting scripts).

The first-version archives (data.zip, scripts.zip) have been replaced by the
corrected files above and remain available under the first version of this dataset
on Dryad.

To reproduce the analysis: extract data_corrected.zip and scripts_corrected.zip
into the same working directory and run scripts/fit_models.R.

Changes after Aug 7, 2023:

29.06.2026 Version update: correction of a plot-identifier misalignment in the processed data. This version replaces the processed data files (data/processed_data/) to correct an anonymisation error in the original deposit. When the full data were anonymised for publication, the plot identifiers (unique_plot_id2) in site_data_unique_plot_id2.csv were renumbered using a different indexing than the stand-variable tables. Because the DBH 7 cm and DBH 10 cm analyses are based on different plot subsets (701 and 865 plots, respectively), the shared site_data file was keyed to the DBH 10 cm plot numbering, so the seasonal climate predictors (dds, swb) were matched to the wrong plots in the DBH 7 cm recruitment analysis. As a consequence, re-fitting the models from the originally deposited data did not reproduce the published recruitment count-model coefficients (Appendix Table A4) or the response-curve figures (Figs 3–4). The analysis code and the published results are unaffected — only the deposited public data were incorrect. The corrected files (data_corrected.zip) use a single, consistent plot/site/institute identifier mapping across both DBH thresholds and the climate data. Re-running scripts/fit_models.R on the corrected data exactly reproduces the published coefficients, selected model types and figures for both DBH thresholds (verified for all 23 species). scripts_corrected.zip additionally includes the corrected data-anonymisation script (data_preparation/make-public-data.R).