Data and code from: Resource stoichiometry and acidification, not resource availability, drive ecosystem responses to nutrient enrichment
Data files
Sep 28, 2026 version files 557.51 KB
-
Experiment_Invert_Data_Vial_metadata.xlsx
10.59 KB
-
Experiment_Invert_Data_Vial.csv
472.29 KB
-
Experiment_Plant_Data_metadata.xlsx
12.54 KB
-
Experiment_Plant_Data.csv
39.07 KB
-
README.md
3.47 KB
-
Sem_bromeliads.R
19.55 KB
Abstract
This dataset was generated from a 90-day field experiment conducted in Parque Estadual da Ilha do Cardoso, Brazil. We used tank bromeliads (Neoregelia cruenta) as natural microcosms to investigate how nitrogen (N) and phosphorus (P) enrichment affect aquatic ecosystems. The study specifically evaluates three mechanistic pathways: (1) basal resource quality (microbial stoichiometry), (2) basal resource quantity (biomass), and (3) abiotic conditions (water pH). The data includes water chemistry, chlorophyll-a concentrations, and the taxonomic identification, biomass, and elemental composition (C:N:P) of the macroinvertebrate communities. Results indicate that nutrient enrichment restructures these ecosystems primarily through changes in microbial stoichiometry and acidification rather than increased resource biomass.
Description of the data and file structure
The data is organized into four primary CSV files (two data files and two corresponding metadata files) and one R script for analysis.
Primary Data Files
File: Experiment_Plant_Data.csv
Description: Contains plant-level environmental and response data. Each row represents one of the 30 individual bromeliads used in the experiment. This file includes the nutrient treatment levels (TN and TP), water chemistry (pH, DO, conductivity), and community-level metrics like Shannon diversity and total biomass of different invertebrate orders (e.g., Diptera, Odonata).
File: Experiment_Invert_Data_Vial.csv
Description: Contains high-resolution data for the macroinvertebrates collected. Each row represents a specific "vial" or sample containing individuals of a particular morphospecies from a specific plant. This file includes taxonomic hierarchy (Phylum down to morphospecies), feeding guilds (e.g., Predator, Scraper), and elemental analysis results (Percent N, C, and P).
Metadata Files
File: Experiment_Plant_Data_metadata.xlsx
Description: A dictionary for the plant data file. It defines every column header and provides measurement units (e.g., mg/L for nutrient concentrations, cm for plant diameter).
File: Experiment_Invert_Data_Vial_metadata.xlsx
Description: A dictionary for the invertebrate data file. It defines taxonomic coding and explains the methodology for elemental processing (e.g., "Represented" vs "Individual" samples).
Missing Data Codes: NA is used in all files to indicate missing values. In the invertebrate data, this typically occurs when the biomass of a sample was too low (below the detection limit of 0.05 mg) to perform elemental analysis for Nitrogen, Carbon, or Phosphorus.
Relationships between files: The files are linked by the unique identifier 'Plant_Number' (1–30). Users can aggregate the invertebrate data in Experiment_Invert_Data_Vial.csv by Plant_Number to join it with the environmental variables found in Experiment_Plant_Data.csv.
Sharing/Access information
Links to other publicly accessible locations of the data:
The data is currently only hosted on the Dryad Digital Repository.
Data was derived from the following sources:
This is original primary research data and was not derived from other sources.
Code/Software
File: R Script: Sem_bromeliads.R
Description: This script replicates the statistical workflow described in the manuscript. It performs data cleaning, calculates community-level stoichiometric means, and constructs the piecewise Structural Equation Models (pSEM).
Software Requirements:
R version 4.2.0 or higher.
Required R packages:
piecewiseSEM (for the primary path analysis)
lavaan (for latent variable modeling)
dplyr and plyr (for data manipulation)
ggplot2 and viridis (for data visualization)
vegan (for ecological diversity metrics)
Workflow:
The script is designed to be run in its entirety. It first reads the two raw CSV data files, merges them based on Plant_Number, filters out baseline (T0) data to focus on the final experimental state, and then fits the SEMs for the three hypothesized pathways (Quality, Quantity, and Acidification).