Data and code for: Elevated CO2 advances vegetation phenology
Data files
Jun 23, 2026 version files 1 MB
-
AppendixS1_FE.xlsx
87.92 KB
-
FE_code.R
67.30 KB
-
FEpeng_zhangSA1.pdf
825.90 KB
-
README.md
8.54 KB
-
site_FE.xlsx
14.04 KB
Abstract
The continuous rise in atmospheric CO2 concentration has profoundly altered vegetation structure and function across terrestrial ecosystems. Among various indicators, vegetation phenology is conventionally considered an early warning indicator of changes in other ecosystem processes. However, the relationship between vegetation phenology and elevated CO2 (eCO2) remains highly elusive.
Here, we conducted a comprehensive meta-analysis of 624 experimental records from 39 published studies to evaluate eCO2 effects on vegetation phenology and quantify the relative importance of climate variables, soil factors, eCO2 magnitude, and experimental duration.
The findings showed that eCO2 significantly advanced overall vegetation phenology by 2.50 days (-1.99%), with leaf-out advancing by 3.00 days (-2.38%) and maturity by 4.09 days (-3.24%). Elevated CO2 exerted a negligible overall effect on plant growing season length. Importantly, model selection showed that plant functional groups were not essential predictors of phenological shifts. Instead, phenological shifts were primarily predicted by mean annual temperature (MAT) and eCO2 magnitude.
These results provide new insights into the mechanisms linking atmospheric change with plant developmental timing and highlight the need to explicitly incorporate CO2-driven phenological responses into ecosystem and Earth system models.
Dataset DOI: 10.5061/dryad.hqbzkh1zf
Description of the data and file structure
This repository contains the dataset and R code required to reproduce the meta-analysis and figures presented in the manuscript.
File List:
- AppendixS1_FE.xlsx: The main dataset containing all extracted phenological records, effect sizes, and environmental moderators.
- site_FE.xlsx: The dataset containing geographical coordinates and observation frequencies for the experimental sites, used to generate the site map (Figure 1).
- FE_code.R: The R script used for statistical analyses (meta-analysis, meta-regression) and figure generation.
- FEpeng_zhangSA1.pdf: The Supporting Information document containing supplementary figures, tables, and the PRISMA flow diagram.
Files and variables
File: AppendixS1_FE.xlsx
Description: The main dataset used for the meta-analysis. It contains all extracted experimental records on vegetation phenology responses to elevated CO2, including effect sizes, sampling variances, and associated environmental, climatic, and soil moderators.
Missing Data: null indicates that the specific ecological metric, parameter, or coordinate was not reported, monitored, or explicitly measured in the original primary publication. These represent genuine missing data points that could not be retrieved during historical data extraction.
Variables
- Study: Citation or identifier of the original study.
- Treatment: Type of experimental treatment.
- CO2_Method: The experimental facility or method used for CO2 enrichment (e.g. FACE, OTC, GC). FACE (Free-Air CO2 Enrichment): A fully open-air field experimental setup designed to study vegetation responses under natural ambient settings. OTC (Open-Top Chamber): A semi-open field enclosure system used as a cost-effective alternative to FACE. GC (Growth Chamber / Greenhouse): A completely enclosed, tightly monitored indoor experimental facility. WTC (Whole-Tree Chamber): A highly specialized, large-scale outdoor enclosure designed to track the responses of entire mature or semi-mature trees. Branch_bags: A localized, low-cost field experimental technique used to manipulate the microenvironment of individual tree branches rather than the entire plant.
- aCO2: Ambient CO2 concentration (ppm).
- eCO2: Elevated CO2 concentration (ppm).
- Magnitude: The magnitude of CO2 enrichment, calculated as the difference between elevated and ambient CO2 concentrations (ppm).
- Duration_years: The duration of the experiment in years.
- Start: The starting year of the experimental observation.
- End: The ending year of the experimental observation.
- Latitude: Latitude of the experimental site (decimal degrees).
- Longitude: Longitude of the experimental site (decimal degrees).
- Elevation: Elevation of the experimental site above sea level (meters).
- MAT: Mean annual temperature of the site (°C).
- MAP: Mean annual precipitation of the site (mm).
- Climate_zone: The classified climatic zone of the experimental site.
- Plant_type: Specific growth form or functional classification of the target plant species studied. Typical values include: trees, shrubs, forbs, grasses.
- Plant_subtype: Sub-classification indicating specific physiological traits.
- Plant_species: The binomial scientific name of the studied plant.
- Legume: Indicates whether the plant is a nitrogen-fixing legume (yes/no).
- Plant_growth_form: Plant functional group classified by growth form (Woody vs. Non-woody).
- Phenology: The specific phenological stage investigated (e.g., Leaf-out, Flowering, Maturity, Senescence).
- Sec: Standard error of the phenological event in the control group.
- Nc: Sample size (number of replicates) of the control group.
- Xt_Xc: Extracted mean values for the treatment group (Xt) and the control group (Xc).
- Set: Standard deviation of the phenological event in the control group.
- Nt: Sample size (number of replicates) of the treatment group.
- SOC: Soil organic carbon content at the control plot (g/kg).
- pH: Soil pH value.
- STN: Soil total nitrogen content (g/kg).
- SBD: Soil bulk density (g/cm³).
- Clay: Soil clay content (%).
- Sand: Soil sand content (%).
- Silt: Soil silt content (%).
- yi: The calculated effect size quantifying the phenological shift.
- vi: The sampling variance associated with the effect size.
File: site.csv
Description: A supplementary spatial dataset containing geographical coordinates, basic plant classifications, and observation frequencies for each experimental site. This file was primarily used to generate the global distribution map of the study sites (Figure 1).
Variables
- ID: Unique sequential identifier assigned to each experimental observation or row entry within the meta-analysis dataset.
- Study: Citation or identifier of the original study.
- CO2_Method: The experimental facility type used to deliver elevated carbon dioxide levels.
- Latitude: Latitude of the experimental site (decimal degrees).
- Longitude: Longitude of the experimental site (decimal degrees).
- Plant_type: Specific growth form or functional classification of the target plant species studied. Typical values include: trees, shrubs, forbs, grasses.
- Plant_growth_form: Plant functional group classified by growth form (Woody vs. Non-woody).
- Number: Sample size (n), number of independent experimental replicates, or specific observation counts extracted from the corresponding treatment group in the original study.
File: FE_code.R
Description: The R script containing all the code required to reproduce the statistical analyses and data visualizations presented in the manuscript. It includes data preprocessing, mixed-effects meta-analysis models, subgroup analyses, publication bias assessments (trim-and-fill), and the generation of main figures.
File: FEpeng_zhangSA1.pdf
Description: It contains supplementary methodological details, the PRISMA flow diagram detailing the literature search and study selection process, as well as additional supplementary figures and tables that support the main text findings.
Code/software
1. Software Requirements
- Data (
.csv): Universally accessible via Excel, text editors, or programming environments (R/Python). - Documents (
.docx): Accessible via MS Word, Google Docs, or LibreOffice. - Code (
.R): Requires the R programming language (RStudio recommended).
2. Key R Packages
metafor: For core meta-analysis models and publication bias tests.tidyverse/dplyr: For data cleaning and manipulation.ggplot2/maps: For generating scatter plots, forest plots, and spatial maps.
3. Code Workflow The FE_code.R script strictly follows the analytical flow of the manuscript:
- Step 1: Imports data and calculates effect sizes and variances.
- Step 2: Runs mixed-effects meta-analysis models to assess overall and subgroup (e.g., woody vs. non-woody) phenological shifts.
- Step 3: Conducts meta-regressions to evaluate the modulating effects of continuous predictors (e.g., MAT, MAP, eCO2 magnitude).
- Step 4: Performs publication bias tests (e.g., Trim-and-Fill) to ensure result robustness.
- Step 5: Automatically generates all visualizations (global maps, forest plots, and scatter plots) used in the paper.
Access information
Other publicly accessible locations of the data:
- Not applicable.
Data was derived from the following sources:
-
This dataset is a meta-analytical compilation of experimental records extracted from previously published scientific literature. The data were derived from the following sources:
1. Published Literature (Primary Source): Data were extracted from 39 peer-reviewed studies. Among these, 18 studies were retained from a previous meta-analysis (Zhou et al., 2023), and the remaining studies were newly identified through systematic literature searches across databases including Web of Science, Google Scholar, and China National Knowledge Infrastructure (CNKI).
2. Environmental Modulators: The climate and soil variables (e.g., MAT, MAP, SOC) matched to the experimental sites were either directly extracted from the original papers or, where missing, acquired from globally gridded climate/soil databases based on the geographical coordinates of the sites.