Data from: Large-scale variation in biodiversity–ecosystem functioning (BEF) relationships in aquatic metacommunities on terrestrial islands
Data files
Aug 20, 2025 version files 1.27 MB
-
00-prepare-raw-data.qmd
24.08 KB
-
01-data-documentation.qmd
5.16 KB
-
02-exploratory-data-analysis.qmd
2.88 KB
-
03-clean-data.qmd
3.63 KB
-
04-all-taxa-h1.qmd
12.55 KB
-
05-active-dispersers-h1.qmd
19.53 KB
-
06-passive-dispersers-h1.qmd
19.24 KB
-
07-combine-h1-plots.qmd
1.18 KB
-
08-all-taxa-h2.qmd
15.92 KB
-
09-active-dispersers-h2.qmd
20.39 KB
-
10-passive-dispersers-h2.qmd
18.69 KB
-
11-combine-h2-plots.qmd
1.23 KB
-
12-active-vs-passive-dispersers-h3.qmd
11.33 KB
-
13-bioclim-pca.qmd
3.01 KB
-
14-species-accumulation-curve.qmd
3.65 KB
-
15-sediment-depth-comparison.qmd
2.85 KB
-
active-passive-split-taxa.csv
5.97 KB
-
alpha-active-raw.csv
24.98 KB
-
alpha-all-raw.csv
26.82 KB
-
alpha-passive-raw.csv
24.68 KB
-
BEF_rock_pools.Rproj
267 B
-
bioclim-data.csv
2.74 KB
-
bioclim-pca-scores.csv
452 B
-
gamma-active-raw.csv
1.30 KB
-
gamma-all-raw.csv
1.29 KB
-
gamma-passive-raw.csv
1.30 KB
-
helper-ci-lmm.R
3.68 KB
-
helper-compare-lmms.R
2.63 KB
-
helper-dfbetas-lmm.R
2.86 KB
-
helper-fisher-test.R
1.11 KB
-
helper-get-dags.R
1.54 KB
-
helper-lmm-path-table.R
1.22 KB
-
helper-plot-functions.R
2.52 KB
-
helper-plotting-theme.R
334 B
-
pool-community-abundance-data.csv
220.86 KB
-
pool-community-biomass-data.csv
260.67 KB
-
pool-sediment-data.csv
6.05 KB
-
prep-env.R
1.06 KB
-
raw-data-alpha.csv
3.47 KB
-
raw-data-gamma.csv
306 B
-
README.md
7.27 KB
-
renv.lock
501.87 KB
Abstract
Recent work has shown that the biodiversity of potential colonists in a landscape (the local species pool) may be more important for ecosystem functioning than the biodiversity in local habitat patches. However, it is unknown how such biodiversity-ecosystem functioning (BEF) relationships may change across different biomes. To explore such patterns, nested insular ecosystems where variation in local biodiversity and local species pool biodiversity can be reliably quantified can provide important insights. Study locations were rock pool metacommunities on isolated rocky outcrops (i.e., inselbergs) in Africa, Australia, Europe, and North America. The sampling time period was 2011-2019. Major taxa studied were freshwater invertebrates. We assembled a large-scale dataset of invertebrate metacommunities from replicated rock pool clusters on inselbergs as a model system to test the ability of local biodiversity and local species pool biodiversity to explain community biomass in organisms with different survival strategies (active or passive dispersers). To test our hypotheses, we used a combination of directed acrylic graph based path analyses and general linear mixed-effects models. The biodiversity of the local species pool was influenced by climate but did not significantly impact community biomass. Instead, local environmental gradients seem to override any species pool effects on community biomass. However, in line with expectations, the relationship between local biodiversity and biomass varied across inselbergs. Contrary to expectations, inselberg prominence did not influence the BEF slope. However, in drier conditions, the BEF relationship weakened for active dispersers, likely reflecting environmental limits on recolonisation. Thus, climate and dispersal strategy jointly shaped how biodiversity influenced community biomass. This study illustrates that even in a simple ecosystem there can be substantial geographical variation in the relationship between biodiversity and ecosystem functioning that may be partially explained by environmental conditions and by the survival strategy of the organisms considered.
https://doi.org/10.5061/dryad.wwpzgmsv0
Description of the data and file structure
All code and data to reproduce the analysis reported in the following publication can be found in this repository: "Large-scale variation in biodiversity–ecosystem functioning (BEF) relationships in aquatic metacommunities on terrestrial islands."
We assembled a global dataset of invertebrate metacommunities from replicated rock pool clusters on inselbergs as a convenient natural model system to test the ability of local biodiversity and local species pool biodiversity to explain community biomass in organisms with different survival strategies (active or passive dispersers). To test our macro-ecological hypotheses, we used a combination of hierarchical piecewise structural equation models (SEM) and general linear models.
Files and variables
File: active-passive-split-taxa.csv
Description: The list of taxa names split into two columns, one for active dispersers and one for passive dispersers.
Files: alpha-all-raw.csv, alpha-active-raw.csv, alpha-passive-raw.csv
Description: Alpha diversities for each rock pool in three datasets - a full dataset that contains all taxa and two subsets of the data that contain only actively or passively dispersing taxa, respectively.
Variables
- Inselberg: three letter code specifying the inselberg from which the data were gathered.
- Pool: identity of the specific rock pool sampled.
- Alpha: alpha diversity for each rock pool as estimated using the Chao-1 estimator based on abundance.
- ASE: standard error corresponding to the alpha Chao diversity.
- Biomass: community-level invertebrate biomass per litre.
- Gamma: gamma diversity for each inselberg as estimated using the Chao-1 estimator based on incidence frequency.
- GSE: standard error corresponding to the gamma Chao diversity.
- Depth: estimated depth for each rock pool (cm).
- PC1: PC1 scores of the principal component analysis (PCA) of 19 different bioclimatic variables for each inselberg.
- PC2: PC2 scores of the principal component analysis (PCA) of 19 different bioclimatic variables for each inselberg.
File: bioclim-data.csv
Description: For each inselberg, 19 different bioclimatic variables were extracted from the WorldClim database (Fick & Hijmans, 2017) at a resolution of 0.04167°. More information about the bioclimatic variables can be found on the WorldClim website (www.worldclim.org).
File: bioclim-pca-scores.csv
Description: PC1 and PC2 scores for the principal component analysis (PCA) conducted on the 19 bioclimatic variables at each inselberg location.
Files: gamma-all-raw.csv, gamma-active-raw.csv, gamma-passive-raw.csv
Description: Gamma diversities for each inselberg in three datasets - a full dataset that contains all taxa and two subsets of the data that contain only actively or passively dispersing taxa, respectively.
Variables
- Inselberg : three letter code specifying the inselberg from which the data were gathered.
- Gamma: gamma diversity for each inselberg as estimated using the Chao-1 estimator based on incidence frequency.
- Biomass (median): meadian value of the community-level invertebrate biomass (per litre) value of each of the pools on an inselberg.
- PC1: PC1 scores of the Principal component analysis (PCA) of 19 different bioclimatic variables for each inselberg.
- PC2: PC2 scores of the Principal component analysis (PCA) of 19 different bioclimatic variables for each inselberg.
- Depth (median): median value of the depths of all rock pools on an inselberg
- SSE: the standard error corresponding each slope coefficient.
- GSE: standard error corresponding to the gamma Chao diversity.
- Prominence: height of an inselberg above the surrounding landscape.
File: pool-community-abundance-data.csv, pool-community-biomass-data.csv
Description: Abundances values and biomass per litre for each taxa in each of the rock pools. Rows correspond to individual rock pools. Column headings correspond to individual taxon.
File: pool-sediment-data.csv
Description: Sediment cover and depth for individual rock pools.
Variables
- Location: three letter code specifying the inselberg from which the data were gathered.
- Poolname: identity of the specific rock pool sampled.
- Lndepth: Depth of the rock pools. The values are ln-transformed to improve linearity.
- LnSedPerVRatio: Sediment cover measured as the percentage of the basin covered with sediment and is standardised by the pool volume. The values are ln-transformed to improve linearity.
File: raw-data-alpha.csv
Description: Raw data at individual rock pool scale with depth values for each rock pool.
Variables
- Inselberg: three letter code specifying the inselberg from which the data were gathered.
- Pool: identity of the specific rock pool sampled.
- Depth: estimated depth for each rock pool (cm).
File: raw-data-gamma.csv
Description: Raw data at inselberg scale with data on inselberg prominence (m) and location.
Variables
- Inselberg: three letter code specifying the inselberg from which the data were gathered.
- Prominence: height of an inselberg above the surrounding landscape.
- Longitude: Longitude value for the inselberg expressed in decimal degrees.
- Latitude: Latitude value for the inselberg expressed in decimal degrees.
Code/software
To run the files you need the software R. Download all the data and scripts and save it locally in a folder - this will be your working directory. Within the working directory there should be three sepparate folders labelled exactly as follows (without the invered commas) (1) "Data-raw", (2) "Code" and (3) "Figures". Store all the downloaded data files in the 'Data' folder and all the R scripts in the 'Code' folder.
To run the code correctly, it is important to create a R-Project. In R-Studio go to File > New Project... > Existing Directory > Choose the extracted directory.
Download the relevant packages
For package version control, we used renv (Ushey and Wickham 2023). Therefore, in the repository, there is a renv lockfile (renv.lock) that contains information about the versions of all packages used in the analysis.
To download the correct versions of all packages that we used in the analysis, you need to install the renv package:
+ install.packages("renv")
Once renv is installed, in the R-console, run the following command:
+ renv::restore()
The restore() command will download the correct versions of all packages used in the analysis into a local folder in the repository called 'renv'.
Run the scripts
In the 'Code' folder, there are 15 scripts numbered 01-15 that need to be run in order. If you correctly downloaded and stored the files as described above, then the scripts should run smoothly. File '00-prepare-RAW-data' is the file used to prepare the data used for analysis. This is already uploaded into this repository and thus, this script doesn't need to be run.
The dataset consists of invertebrate communities sampled in 238 rock pools from ten different inselbergs across four continents. On each inselberg site, between 12 and 34 pools were sampled. After sampling, the invertebrates from each rock pool were identified to the lowest possible taxonomic resolution in the laboratory (genus or species, when possible), and the number of individuals of each group was counted. The community-level invertebrate biomass per litre was calculated for each pool by multiplying species-specific biomass estimates by density. Alpha diversity was calculated as the local-scale species richness of each rock pool, and gamma diversity was calculated as the inselberg-scale species richness representing the diversity of the local species pool.
