Data and code from: Pre-industrial land-use limits contemporary shrub encroachment in the French Alps
Data files
Mar 23, 2026 version files 961.12 KB
-
data_Ecography.RData
899.21 KB
-
Data_preparation.R
18.31 KB
-
FIGURE_01.R
3.28 KB
-
FIGURE_02.R
4.70 KB
-
FIGURE_03.R
9.57 KB
-
FIGURE_04.R
12.48 KB
-
LIBRARIES.R
1.82 KB
-
README.md
5.37 KB
-
SUPP_FIGURES.R
6.39 KB
Abstract
Shrub encroachment has become a global phenomenon in recent decades. While global warming in the Arctic is often cited as the primary cause, human-managed mountain regions have experienced intense historical land-use that may also play a considerable role. Shrub encroachment has significant implications for biodiversity, carbon storage, and transformation of mountain landscapes. However, disentangling between the effects of present-day environmental conditions and historical land-use trajectories in explaining present-day shrubland distribution remains challenging. We use early 19th-century land registry records — rare historical sources documenting past land-use and vegetation cover — to assess whether these legacy conditions influence present-day shrub distribution. By combining these archive data with dendrochronological sampling and a diachronic analysis of shrub expansion based on historical aerial photographs, we provide a comprehensive analysis of shrub colonisation. Our results reveal an increase in shrub cover over the past two centuries, primarily driven by the expansion of common juniper (Juniperus communis). Dendrochronological analyses from never mown or ploughed plots indicate that shrub recruitment began around 1870, intensified until the 1970s, and then slowed slightly thereafter. While shrubs preferentially establish on steep, south-facing slopes and topographically rough terrain, their distribution remains highly heterogeneous and is strongly shaped by pre-industrial land-use patterns. Parcels that were ploughed or mown 200 years ago remain largely devoid of shrubs, even decades after mowing ceased. In contrast, surrounding areas — whether grazed or unused — have experienced shrub encroachment. We hypothesize that the low shrub density on former ploughed lands and hay meadows may reflect persistent legacy effects of dense, competitive herbaceous communities, which hinder the establishment of woody species. This study demonstrates that pre-industrial land-use continues to constrain both the distribution and recent dynamics of shrub expansion, highlighting the importance of accounting for land-use legacies when interpreting contemporary shrub encroachment.
Dataset DOI: 10.5061/dryad.wpzgmsc3t
Description of the data and file structure
Data and codes for producing the figures in the article ‘Nicoud et al. 2026 - Pre-industrial land-use limits contemporary shrub encroachment in the French Alps’. There is an .Rdata file containing the data, an .R file for preparing the data, one for the libraries, one for each figure in the main manuscript, and one for the supplementary figures.
First, run the data preparation script that will process the data from the .Rdata file. Then you can run the scripts for each figure independently. The code is fully annotated to explain each step of the script. The scripts required for the different stages of the analysis can be run directly from the main file using the source() function.
Files and variables
File: Data_preparation.R
Description: script needed to process data of the .Rdata file. It produces data that are ready to create the figures. It uses the 5 datasets provided in the .Rdata file to prepare additional data needed for the following figures.
File: FIGURE_01.R
Description: script to produce the figure 1. This figure shows a transition matrix between land use and land cover in 1819 and 2022.
File: data_Ecography.RData
Description: file containing all the needed data to produce the figures.
This file contains five datasets:
-
‘cadastre’: provides, for each plot in the study municipality, the parcel identifier (“NPARC”), the historical local name (“toponyme”), the parcel area (“Area_ha”, in hectares), the minimum elevation (“Alti_Min”, m), centroid elevation (“Alti_Centr”, m), and maximum elevation (“Alti_Max”, m). It also includes the land-use matrix in 1819 (“MOS_1819”), land-cover classifications for 1819 and 2022 (“LC_1819” and “LC_2022”; Building, Forest, Grazed, Mowed, Ploughed, Unused), land use in 1819 (“LU_1819”), two land-use variables for 2022 (“OCS_2022” and “OCS_SIMPL”), the cadastral register (“RPG”), the date of the last mowing (“Dern_Fauch”, year), and the land-use trend (“TRAJ”).
Some variables are provided for completeness and documentation purposes but are not used in the analyses. The data preparation stage mainly consists of formatting the dataset to generate the figures.
-
‘shrub.2022’: provides, for all plots where juniper density was photo-interpreted in 2022, the juniper density (“Junip_dens”, individuals ha⁻¹), the land-use classification in 1819 (“LULC_1819”), elevation (“Elevation”, m a.s.l.), the land-use trajectory (“LULC_Traj”), slope (degrees), terrain roughness (“Roughness”, unitless), percentage of bare soil (“Baresoil”, %), diurnal anisotropic heating (“DAH”, unitless, ranging from −1 to 1), soil fertility in 1819 (“Fertility”, classes from 1 = low fertility to 5 = high fertility), and the parcel identifier (“NPARC”).
-
‘shrub.65_22’: provides, for each site where juniper density was photo-interpreted in both 1965 and 2022, the site identifier (“Site”), the number of shrubs recorded in 1965 and 2022 (“Eff_65” and “Eff_22”), the difference between the two (“delta”), a difference index (“index”, unitless), the elevation (“Altitude”, m a.s.l.), and the land-use trajectory (“TRAJ”).
-
‘shrub.age’: provides, for each juniper individual sampled in the field, the site identifier (“Site”), the age of the individual (“Age”), the individual identifier (“Individus”), the year of establishment (“birth”), the decade of establishment (“decade”), and the land-use trajectory (“Classe”, corresponding to “TRAJ” in the shrub.65_22 dataset).
-
‘surface.lulc’: provides a summary of the different LULC trajectories between 1819 and 2022 and the areas associated with each trajectory (hectare).
File: FIGURE_02.R
Description: script to produce the figure 2. This figure shows the age structure of juniper individuals recruited over the past 200 years, as determined from dendrochronological sampling. It also illustrates the change in juniper density in the plots that were photo-interpreted in 1965 and 2022.
File: FIGURE_03.R
Description: script to produce the figure 3. This figure summarises the different land-use (LULC) patterns by showing the area they cover within the municipality, their topographical characteristics (percentage of bare soil, surface roughness, elevation, and DAH), and the current density of juniper shrubs.
File: FIGURE_04.R
Description: script to produce the figure 4. This figure presents the results of modelling shrub density in 2022 using a random forest model. It shows the relative importance of the predictor variables as well as the partial dependence plots for LULC1819, date of last mowing, elevation, DAH, percentage of bare soil, and terrain roughness.
File: LIBRARIES.R
Description: script to run the libraries.
File: SUPP_FIGURES.R
Description: script to produce the supplementary figures.
Code/software
The R and Rstudio software packages are required to produce and visualise the figures.
