Data and code from: A functional-trophic pathway links riparian land use intensification to riverine community stability
Data files
Jul 20, 2026 version files 351.02 MB
-
Custom-curated_reference_databases.zip
1.78 MB
-
Fig.1.zip
348.71 MB
-
Fig.2_S2-4.zip
7.24 KB
-
Fig.3-4_S5-13.zip
234.70 KB
-
Fig.5_S14-16.zip
211.76 KB
-
Fig.6.zip
70.47 KB
-
README.md
7.21 KB
Abstract
Intensive land use is transforming river corridors worldwide, with the agricultural and urban land use intensification being most prominent. Understanding how these pervasive pressures affect functional and trophic community structure – and ultimately ecosystem stability – is thus critically needed. Here, we disentangle the contribution of agricultural and urban land use changes on aquatic ecosystems by combining high-resolution environmental DNA (eDNA) surveys across 213 spatially and temporally stratified sampling sites with functional trait profiling and metaweb-based trophic networks in a major subtropical river. We found that the functional trait composition of fish, aquatic insects and algae was strongly structured and exhibited pronounced temporal reorganization. Riparian land use intensification was consistently associated with lower functional diversity and structural complexity of metaweb-derived local trophic networks, including lower modularity, nestedness and robustness, along agricultural and urban impervious land use gradients. Functional diversity and trophic network structure were strong positive predictors of community stability, and piecewise structural equation models indicated that increasing land use intensity destabilizes multitrophic communities primarily by suppressing functional and structural attributes, rather than through direct pathways. Overall, by uniting these two pathways within a single empirical framework, our study provides a mechanistic basis for predicting how human land conversion undermines ecosystem stability and offers a foundation for safeguarding biodiversity in rapidly changing environments.
Dataset DOI: 10.5061/dryad.k98sf7mp6
Description of the data and file structure
This dataset contains the spatial data, curated taxonomic reference databases, processed ecological data, and R scripts used to reproduce the analyses and figures associated with the manuscript “A functional–trophic pathway links riparian land-use intensification to riverine community stability.”
The study was conducted in the Dongjiang River basin in southern China, a subtropical river system within the Pearl River basin. The dataset supports analyses of riverine community composition, functional diversity, trophic-network structure, temporal community stability, and their relationships with agricultural land cover and urban impervious cover.
The ecological analyses include fish, aquatic insects, algae, and other aquatic organism groups identified using eDNA approach. Functional analyses focus on fish, aquatic insects, and algae. All dataset documentation, file names, script annotations, variable descriptions, and figure-specific instructions are provided in English.
Files and variables
Files are organized into compressed archives according to the corresponding main and supplementary figures.
The statistical analyses and figures were produced in R. The required R packages are identified in the package-loading section at the beginning of each script. Users should install the indicated packages before running the analyses.
The spatial data associated with Figure 1 were prepared using ArcGIS and are supplied in standard GIS-compatible formats. These files may be opened using ArcGIS, QGIS, or other software capable of reading the supplied geospatial formats.
The custom taxonomic reference databases are provided as compressed sequence-database files and may be inspected or used with standard bioinformatic software compatible with FASTA-formatted nucleotide sequences and their associated taxonomy information.
File: Custom-curated_reference_databases.zip
Description: Custom-curated mitochondrial 12S ribosomal RNA and cytochrome c oxidase subunit I reference databases used for the taxonomic assignment of fish and aquatic insects detected through environmental DNA metabarcoding in the Pearl River basin.
The archive contains reference nucleotide sequences and their associated taxonomic annotations. The 12S database was used primarily for fish taxonomic assignment, whereas the COI database was used primarily for aquatic-insect taxonomic assignment.
File: Fig.1.zip
Description: Spatial data used to prepare Figure 1 in the associated manuscript, which presents the study area and sampling-site distribution in the Dongjiang River basin.
The archive contains the GIS data describing river topology, elevation, subcatchment boundaries, sampling-site locations, and land-cover distribution. These spatial files correspond to the study-region materials used to produce the Figure 1 map and therefore partially duplicate the study-area ArcGIS files deposited with the dataset.
River topology, elevation, and subcatchment boundaries were derived from digital elevation data. Land-cover classification was derived from Landsat-8 Operational Land Imager imagery at a spatial resolution of 30 m and represents mean annual land-cover conditions from 2021 to 2024. The principal land-cover categories include forest, agricultural land, and urban impervious surfaces.
File: Fig.2_S2-4.zip
Description: Processed R scripts used to reproduce Figure 2 and Supplementary Figures S2–S4.
Figure 2 evaluates temporal changes in the functional composition of fish, aquatic-insect, and algal communities across six sampling occasions. The analyses include principal coordinates analysis of community-weighted mean functional traits, trait-vector fitting, permutational multivariate analysis of variance, pairwise comparisons among sampling occasions, and heatmaps of pairwise centroid distances in multivariate functional-trait space.
File: Fig.3-4_S5-13.zip
Description: Processed data and R scripts used to reproduce Figures 3 and 4 and Supplementary Figures S5–S13.
Figure 3 examines relationships between temporal changes in four components of functional diversity and agricultural or urban-impervious land-use gradients. The four functional-diversity metrics are functional richness, functional evenness, functional divergence, and functional dispersion. Analyses are provided separately for fish, aquatic-insect, and algal communities.
Figure 4 examines relationships between temporal changes in local metaweb-derived food-web properties and the same land-use gradients. The principal network properties include node density, nestedness, modularity, and robustness.
File: Fig.5_S14-16.zip
Description: Processed data and R scripts used to reproduce Figure 5 and Supplementary Figures S14–S16.
Figure 5 quantifies associations between temporal changes in community stability and changes in functional-diversity or metaweb structural metrics. Functional analyses include functional richness, functional evenness, functional divergence, and functional dispersion for fish, aquatic-insect, and algal communities. Network analyses include node-related properties, nestedness, modularity, and robustness.
All functional-diversity, network, and stability variables should be interpreted according to the calculations implemented in the corresponding R scripts. Dimensionless indices have no physical measurement unit.
File: Fig.6.zip
Description: Processed data and R scripts used to reproduce Figure 6.
Figure 6 presents structural equation models evaluating associations among riparian land-use intensification, functional diversity, metaweb structural properties, and riverine community stability. Separate models were fitted for the functional-diversity pathway and the trophic-network structural pathway.
The functional model summarizes functional richness, evenness, divergence, and dispersion using the first axis of a principal component analysis. The network model summarizes node-related properties, link density, connectivity, nestedness, modularity, and robustness using the first principal component axis. Agricultural land cover and urban impervious cover are included as land-use predictors, and community stability is the response variable.
The archive contains the standardized model-input data, principal-component scores or the data required to calculate them, structural-equation-model specifications, model-fit outputs, path coefficients, and plotting scripts.
Contact and citation
Users of this dataset are recommended to cite both the Dryad dataset and the associated research article. Questions regarding the data structure, analytical workflow, or reuse of the custom-curated reference databases should be directed to the corresponding author of the associated manuscript. Readers who require additional data or have other data-related requests are encouraged to contact the corresponding author by email.
