# Drivers of plant community composition and species richness in Western Greenland Dataset DOI: [10.5061/dryad.8sf7m0d17](10.5061/dryad.8sf7m0d17) **Principal Investigator Contact Information** ``` Name: Jacob Nabe-Nielsen Institution: Aarhus University Email: jnn@ecos.au.dk ``` ## Description of the data and file structure This dataset contains the data required to replicate analyses in Nabe-Nielsen et al. (2025), studying the potential drivers of vegetation composition and species richness along coast to inland and altitudinal gradients by the Nuuk fjord in western Greenland using Hierarchical Modelling of Species Communities (HMSC) and linear mixed models. Data originate from five study sites along the Nuuk fjord in western Greenland. At each site three groups of six plots were selected for every 100 m increase in altitude. Plot centres were placed exactly 10 m apart along the isoclines, and plot groups were placed exactly 500 m apart. The permanently marked plots were circular, with a diameter of 2 m. For each plot a complete inventory of all species of vascular plants was conducted in the period 2011–2013. Positions of individual plots are provided in the UTM zone 22 N projection. **Nuuk_plant_data_231114.xlsx** (located in the "data" folder in the file "HMSC_models_and_data.zip") is an Excel-file with four sheets containing all the field data used in the publication. The sheet **Species data** contains a list of all plant species recorded in the 414 vegetation plots. Species: Scientific name of the studied plant species N plots: Number of plots where the species was found P001–P414: Presence of the species in each of the 414 study plots (presence recorded with an 'x') The sheet **Positions and inclin:** plot: Plot numbers (P001–P414), corresponding to the numbers on the metal tags used in the field site: Study site number (1–5) alt: Altitude of the plot (in m) grp.no: Name of the plot group number within the specific site and altitude band (a, b, or c) plot.grp: Name of the plot group (combination of site, altitude and grp.no) long.garmin: Longitude, measured using Garmin GPS (blank means missing) lat.garmin: Latitude, measured using Garmin GPS (blank means missing) x.garmin: East-west position, measured using Garmin GPS (UTM zone 22N; blank means missing) y.garmin: North-south position, measured using Garmin GPS (UTM zone 22N; blank means missing) x.trimble: East-west position, measured using Trimble differential GPS (UTM zone 22N; blank means missing) y.garmin: North-south position, measured using Trimble differential GPS (UTM zone 22N; blank = missing) inclin.down: Average plot steepness (degrees) inclin.dir: Direction of slope (i.e., direction downhill), in degrees soil.water.x: Soil water, with x taking the values 1-3 (pct. water; missing values shown as NA) date1 and date.2: the dates where the plot was visited / variables measured The sheet **Hgt and cover 2021-22**: plot: Plot numbers (P001-P414) site: Study site (1-5) Taxon: Name of the taxonomic group (or 'Rock') Max hit: Maximum height of the taxon observed within the 2-m plot (cm) Max dia: Maximum diameter (mm) of different taxa within the 2-m plots Cover: How much of the 2-m plot that was covered by the taxon (in percent; average of two estimates). ## Code/Software The file "HMSC_models_and_data.zip" is a compressed archive contaning the field data and the code required to run the HMSC analyses presented in the publication, along with directories required for handling input and output from the analyses. The archive contains the folders "data", "models", and "results", and the R code files "S1_define_and_fit_models.R", "S2_evaluate_convergence.R", "S3_show_parameter_estimates.R". The first file contains code for reading in and rearranging the species data stored in "data/Nuuk plant data 231114.xlsx". It also reads in the file "data/b for Hmsc 231114.RData" **b for Hmsc 231114.RData**: stores covariates used in the analysis (rearranged from data stored in the tabs "Species data" and "Hgt and inclination 2021-22" in the Excel data sheet, and summer temperatures and summer precipitation data from the Chelsa database version 2.1 (Karger et al. 2017)). One line per plot: plot.grp: Plot group number (see above) plot: Plot number (one line per plot, numbered P001 to P414) lon: Longitude of the plot lat: Latitude of the plot site: Study site (1-5) alt: Altitude (m) grp.no: Name of the plot group number within the specific site and altitude band (a, b, or c) date.1: Date of first visit to the plot; collection of vegetation composition data (yyyy-mm-dd) date.2: Date of second visit to the plot; collection of plant height and cover (yyyy-mm-dd) inclin.down: Plot steepness (degrees) inclin.dir.true: Direction of slope (degrees) soil.water.mean: Average of three soil water measures (percent soil water) Sri: Solar radiation index: Calculated from slope, direction and position (see paper) cover.p: Cover values from pin-point data; not used in the present data analyses (range 0–1) cover.c: Cover values estimated within 2-m circle (percent); used in the paper n.spp.per.plot: Number of species observed in the plot group (response variable used in the linear analyses) max.hgt: Maximum plant height (cm) summer.temp: Mean summer temperature (degrees Censius) summer.prec: Mean summer precipitation (mm) cont.idx: Continentality index; fraction of distance moved between outer coast and inland glacier **S1_define_and_fit_models.R**: Code used to define the spatial nesting used in the analysis, with plots nested within plot groups nested within sites, and specifies the number of replicates to use. Finally the models are fitted using the "sampleMcmc" function from the Hmsc package, and the model objects are stored in "models/unfitted_models.RData". **S2_evaluate_convergence.R:** code for plotting the fitted model objects to evaluate model convergence (outputs stored in the "results" directory). **S3_show_parameter_estimates.R:** code for producing the plots presented in the paper based on the fitted model objects. The **data** folder contains the file **unfitted_models.RData**; a list of model outputs used in code found in the files S2_evaluate_convergence.R and S3_show_parameter_estimates.R. The results folder contains a number of produced automatically when running the code in the files S2_evaluate_convergence.R and S3_show_parameter_estimates.R. The files are: Fig 4 - Hmsc parameter_estimates.csv, MCMC_convergence.pdf, MCMC_convergence.txt, parameter_estimates_Beta_plant model.xlsx, parameter_estimates_Omega_plant model_group.xlsx, parameter_estimates_Omega_plant model_plot.xlsx, parameter_estimates_Omega_plant model_site.xlsx, parameter_estimates_VP_plant model.csv, parameter_estimates_VP_R2T_Betaplant model.csv, parameter_estimates_VP_R2T_Yplant model.csv, parameter_estimates.pdf, parameter_estimates.txt, Prop raw var - VPr.csv. The values stored in the csv files and the plots in the pdf files make it possible to inspect how well the model fits and provides various summaries of the model outputs. Contents of all outputs are described in detail in the paper (Nabe-Nielsen et al. 2025) and the help files accompanying the Hmsc-package, which is available on CRAN ([https://cran.r-project.org](https://cran.r-project.org)). All statistical analyses were performed using R version 4.4.1 **Dates of Data Collection** ``` 2011-2013. ``` **Data Spatial Scope** The Nuuk fjord (Godthåbsfjorden) in western Greenland. **Sharing/Access** This work is licensed under a CC0 1.0 Universal (CC0 1.0) Public Domain Dedication license. **Recommended Citation** Nabe-Nielsen, J., Nabe-Nielsen, L.I. & Ovaskainen, O. (2025). Drivers of plant community composition and species richness in Western Greenland. Ecography.

The Arctic experiences rapid climate change, but our ability to predict how this will influence plant communities is hampered by a lack of data on the extent to which different species are associated with particular environmental conditions, how these conditions are interlinked, and how they will change in coming years. Increasing temperatures may negatively affect plants associated with cold areas due to increased competition with warm-adapted species, but less so if local temperature variability is larger than the expected increase. Here we studied the potential drivers of vegetation composition and species richness along coast to inland and altitudinal gradients by the Nuuk fjord in western Greenland using Hierarchical Modelling of Species Communities (HMSC) and linear mixed models. Community composition was more strongly associated with random variability at intermediate spatial scales (among plot groups 500 m apart) than with large-scale variability in summer temperature, altitude or soil moisture, and the variation in community composition along the fjord was small. Species richness was related to plant cover, altitude and slope steepness, which explained 42% of the variation, but not to summer temperature. Jointly, this suggests that the direct effect of climate change will be weak, and that many species are associated with microhabitat variability. However, species richness peaked at intermediate cover, suggesting that an increase in plant cover under warming climatic conditions may lead to decreasing plant diversity.

The study was conducted at five different sites along the Nuuk fjord (Godthåbsfjorden) in western Greenland. In 2011–2013, 414 permanent plots were established across the five study sites. The plots were circular, with a diameter of 2 m. Three groups of six plots were selected for every 100 m increase in altitude. The position of the first plot at each altitude was selected by walking uphill straight towards a pre-selected point until the isocline was reached (measured using a hand-held GPS). Plot centres were placed exactly 10 m apart along the isoclines, or slightly more if needed to prevent plots from being entirely in water or having an average slope >45°. Plot groups were placed exactly 500 m apart. The lowest plots were moved to 20 m a.s.l. to avoid exposure to salt. The highest altitudes used in sites 1–5 were 200, 200, 400, 500 and 500 m a.s.l., respectively, reflecting that the mountains are higher further east. The study design was the exact same as previously used in Young Sund (Nabe-Nielsen et al. 2017). The vegetation survey included a complete inventory of all species of vascular plants in each plot. This inventory was conducted in 2011–2013. Maximum height and diameter were measured for the woody species in 2021–2022. The cover of woody species, graminoids, and herbaceous plants was assessed for each plot by two independent observers as percent cover <2 m from plot centres. Subsequently the two estimates were averaged. Plants were named following Böcher et al.(1978).