Aim: Plant growth and phenology plastically respond to changing climatic conditions both in space and time. Species-specific levels of growth plasticity determine biogeographical patterns and the adaptive capacity of species to climate change. However, a direct assessment of spatial and temporal variability in radial-growth dynamics is complicated, as long records of cambial phenology do not exist.

Location: 16 sites across European distribution margins of Juniperus communis L. (the Mediterranean, the Arctic, the Alps and the Urals).

Time period: 1940-2016

Major taxa studied: Juniperus communis

Methods: We applied the Vaganov-Shashkin process-based model of wood formation to estimate trends in growing season duration and growth kinetics since 1940. We assumed that J. communis would exhibit spatially and temporally variable growth patterns reflecting local climatic conditions.

Results: Our simulations indicate regional differences in growth dynamics and plastic responses to climate warming. Mean growing season duration is the longest at Mediterranean sites and, recently, there is a significant trend towards its extension of up to 0.44 days per year. However, this stimulating effect of longer growing season is counteracted by declining summer growth rates caused by amplified drought stress. Consequently, overall trends in simulated ring-widths are marginal in the Mediterranean. By contrast, durations of growing seasons in the Arctic show lower and mostly non-significant trends. However, spring and summer growth rates follow increasing temperatures, leading to a growth increase of up to 0.32 % per year.

Main conclusions: This study highlights the plasticity in growth phenology of widely distributed shrubs to climate warming–an earlier onset of cambial activity that offsets the negative effects of summer droughts in the Mediterranean and, conversely, an intensification of growth rates during the short growing seasons in the Arctic. Such plastic growth responsiveness allows woody plants to adapt to the local pace of climate change.

This data package contains (i) input ring-width series of Juniperus communis used to calibrate the Vaganov-Shashkin model and (ii) outputs of the model important for the interpretations made in the above-referenced paper.

(i) were obtained using standard dendrochronological procedures (including field sampling, laboratory processing and ring-width measurement). Individual datasets are stored in RWL format which might be uploaded by most of dendrochronological softwares. Data represent annual ring widths of J. communis shrubs. If multiple discs/radii were measured at specific site, they were averaged into the shrub level. One RWL file is provided for each site of the network.

(ii) represent outputs of the Vaganov-Shashkin models' simulations applied on each site of the network.

For detailed description of each dataset, please, see enclosed ###README.txt file.

Site codes

The data are related to 16 analyzed sites spread across the European part of J. communis distribution. For the full spatial identification (coordinates), please, see Figure 1 in the above-referenced paper.

The site codes are used as follows:

ARCTIC SITES
ABI = Abisko (Sweden)
FIN = Finse (Norway)
KEV = Kevo (Finland)
KIR = Kirkenes (Norway)
KOB = Kobbefjord (Denmark, Greenland)

URAL SITES
PUR = Polar Ural
NUR = Northern Ural
SUR = Southern Ural (all Russia)

ALPINE SITES
CDL = Croda de Lago
SEL = Sella Nevea
RHE = Val di Rheme
VEN = Val Ventina (all Italy)

MEDITERRANEAN SITES
ALI = Aliaga (Spain)
PEN = Penarroya (Spain)
POL = Pollino (Italy)
VIL = Villarroya de los Pinares (Spain)

Ring-width series of J. communis

Folder 'OBSERVATION' contains raw (non-detrended) ring-width series for each sampled shrub. For sites, where multiple radii and/or cross-sections were measured, individual raw measurements were averaged to the level of individual shrubs. Data are stored in standard format RWL which can be loaded by most of dendrochronological programs (including the dplR package for R). One RWL file is presented for each site.

Simulated daily growth rates

Daily growth rates simulated by the VS-model are stored in 'SIMULATION-Growth Rates' subfolders. Simulated integral growth rates and partial growth rates to temperature have a format of a semicolon-separated matrix. The first row of the matrix represents a header. Each of the 365 rows indicates simulated growth rates for specific DOY (starting with DOY 1 in the first row and DOY 365 in the last row). Columns indicate individual calendar years of simulations (starting with the year 1940 in the first column).
Because partial growth rates to photoperiod are invariant from year to year, the respective matrix is simplified to 365 rows and 1 column.

Simulated number of differentiated cells

Numbers of differentiated cells since the beginning of the calendar year are for each day stored in the 'SIMULATION-Cell differentiation' folder. Individual files have a format of a semicolon-separated matrix. The first row of the matrix represents a header. Each of 365 rows indicates the numbers of differentiated cells since the beginning of the growing season for each DOY (starting with DOY 1 in the first row and DOY 365 in the last row). Columns indicate individual calendar years of simulations (starting with the year 1940 in the first column).

Simulated cambial phenology

Key dates of simulated cambial phenology are stored in the 'SIMULATION-Phenology' folder. Each file contains a semicolon-separated table with 3 columns. The first one indicates the calendar year, the second one simulated DOY of cambial activity onset and the last one simulated DOY of cambial activity cessation.

Simulated site chronologies

Simulated site chronologies are stored in the 'SIMULATION-Chronology' folder. Each file contains a semicolon-separated table with 2 columns. The first one indicates the calendar year, the second one value of the simulated chronology.