Autumn leaf phenology (i.e. leaf colouring or leaf senescence) marks the end of the growing season, during which trees assimilate atmospheric CO₂. Since autumn leaf phenology responds to climatic conditions, climate change affects the length of the growing season. Thus, autumn phenology is often modelled to assess possible climate change effects on future CO₂ mitigating capacities and species compositions of forests.

Here, we give access to the entire dataset of modelled autumn phenology analyzed in Meier and Bigler (2023). The data was derived from >2.3 million model calibration runs according to 21 such models, 5 optimization algorithms, ≥7 sampling procedures, and 26 climate model chains from two representative concentration pathways. Calibration and validation were based on >45 000 observations for common beech (Fagus sylvatica L.), pedunculate oak (Quercus robur L.), and European larch (Larix decidua Mill.) from 500 Central European sites each.

Cite as Meier, M., & Bigler, C. (2023). Process-oriented models of autumn leaf phenology: Ways to sound calibration and implications of uncertain projections. Geoscientific Model Development, 16(23), 7171–7201. https://doi.org/10.5194/gmd-16-7171-2023

Autumn leaf phenology was modelled with different process-oriented models that were calibrated as either site- or species-specific models. The calibrations differed in the choice of optimization algorithms and choice of sampling procedures to separate observations for the 5-fold cross-validation in site-specific calibration or to select the sites in species-specific calibration. Species-specific models were calibrated with 75% of the observations of the selected sites and validated with (the remaining) 25% of the observations of all 500 sites per species.

The data contains files of calibrated parameter sets and files of modelled dates of autumn leaf phenology.

The observed autumn leaf phenology data was downloaded from the PEP725 database on April 13, 2022 (http://www.pep725.eu/).

For more detailed information consult the original publication Meier and Bigler (2023) for which full reference is given in the abstract to this dataset.

The Rds files can be opened in the software R (R Core Team, 2022) with the base function readRDS().