A globally consistent scaling relationship reveals stabilizing effects of dominant species in plant communities
Data files
Feb 03, 2026 version files 9.56 MB
-
comm_1.txt
13.14 KB
-
coms20.txt
1.42 KB
-
original_data.txt
9.51 MB
-
R_code.Rmd
27.64 KB
-
README.md
10.06 KB
Abstract
Despite extensive research, the mechanisms stabilizing ecosystems remain uncertain. Taylor’s Power Law (TPL), which describes how variance scales with mean abundance (σ² = aμᵇ), is a pervasive ecological pattern. While TPL has been widely examined within populations, its role across species within communities and its implications for stability remain largely unexplored. A TPL scaling exponent (b) < 2 implies a stabilizing influence of dominant species—hereafter referred to as the dominance effect—where community stability emerges because dominant species are relatively more stable than subordinate species. Using data from over 9,000 permanent vegetation plots worldwide, we quantified within-community TPL, linked variation in the exponent b to dominance effects on temporal stability, and identified the biotic and abiotic drivers shaping b. We found a ubiquitous within-community TPL (mode R² = 0.92) with consistently b < 2, indicating widespread dominance effects. Variation in b, together with species evenness, strongly contributed to dominance-driven stability. Lower b values were associated with resource-conservative strategies and greater climatic seasonality, highlighting the role of environmental filtering in shaping community stability. Overall, these results demonstrate that dominance effects on temporal stability are widespread, particularly in communities dominated by woody, large-seeded species in cold and seasonal climates, and identify the TPL exponent b as a powerful indicator of the stabilizing role of dominant species, complementing the well-established effects of species diversity.
https://doi.org/10.5061/dryad.msbcc2g7f
Description of the data and file structure
This repository contains the data and R Markdown code required to reproduce the analyses and figures presented in the manuscript “A Globally Consistent Relationship Reveals Stabilizing Effects of Dominant Species in Plant Communities.”
The dataset provided here includes the results of calculating Taylor’s Power Law (TPL) relationship for 11,676 plant communities derived from the LOTVS database (https://lotvs.csic.es/). Due to ownership by various authors and organisations, the raw data cannot be freely shared. However, extensive metadata describing each plot and the TPL calculation results are shared.
The vegetation plots span 92 distinct datasets, each representing a unique location and experimental design. These datasets vary in abundance quantification methods (e.g., aboveground biomass, species cover estimates, and species frequencies), plot sizes (median = 1 m², range = 0.04–3000 m²), vegetation types (grassland, shrubland, savanna, forest understory, tundra, and salt marsh), and management regimes (e.g., grazing, fertilization, burning). The dataset encompasses a total of 11,676 individual plots across 92 localities, with 1,216,339 abundance observations.
This data was used to test the hypothesis of a globally consistent TPL relationship within plant communities, with implications for understanding community stability on a worldwide scale.
Files and variables
File: original_data.txt
Description: this file contains the main dataset, encompassing the results of Taylor’s Power Law (TPL) calculations for 11,676 plant communities. Due to varying species presence across communities, some lacked a sufficient number of species to reliably estimate the TPL relationship, resulting in NA values for certain TPL estimates. To qualify for TPL calculation, each community had to contain at least four species with more than 15% presence across the sampling years.
This dataset is subsequently used to create different data subsets for analysis. Each subset applies additional filters to account for methodological factors and to ensure robust analyses in the manuscript.
Variables
- LOTVS_ID: single identification name for each community (vegetation plot).
- Dataset: single number identification the original dataset. Each dataset constitutes a group of plots sharing same location and methodology.
- Name_Dataset: name of the dataset
- Plot: name of the plot inside the dataset
- Data_type
- TreatmentS: dicotomic variable indicating if the plot has any treatment (Perturbation) or not (Control)
- Treatment: categorical variable with 25 categories describing the treatments applied.
- Treatment_details: brief description of the treatment.
- Trajectory: variable describing if the plot has gone through successional changes through the sampling years. 0= no succession; 1 = some degree of directional change; 2 = high degree of directional change
- Country
- Continent
- Habitats: categorical variable of 5 broad vegetation habitats refering to the vegetation height. Grassland; Grassland-Shrubland; Scrubland; Forest; Other
- Biomes: categorical variable of the reference Biome based on Whittaker (1975)
- Plot_size: in meters
- Initial_year: initial year of sampling
- End_year: last year of sampling
- Duration: whole duration of the sampling in years, including years without measure
- n_years: number of years with measure
- nsp: maximum number of species present through the whole sampling period
- nsp_mean: mean number of species per year
- simpson: mean Simpson’s dominance index
- ENS_simpson: mean effective number of species based on Simpson’s index
- bio1: Annual Mean Temperature (ºC*10)
- bio10: Mean Temperature of Warmest Quarter (ºC*10)
- bio11: Mean Temperature of Coldest Quarter (ºC*10)
- bio12: Annual Precipitation (mm)
- bio13: Precipitation of Wettest Month (mm)
- bio14: Precipitation of Driest Month (mm)
- bio15: Precipitation Seasonality (Coefficient of Variation - no units)
- bio16: Precipitation of Wettest Quarter (mm)
- bio17: Precipitation of Driest Quarter (mm)
- bio18: Precipitation of Warmest Quarter (mm)
- bio19: Precipitation of Coldest Quarter (mm)
- bio2: Mean Diurnal Range (Mean of monthly (max temp - min temp); ºC *10)
- bio3: Isothermality (BIO2/BIO7) (×100)
- bio4: Temperature Seasonality (standard deviation ×100)
- bio5: Max Temperature of Warmest Month (ºC * 10)
- bio6: Min Temperature of Coldest Month (ºC * 10)
- bio7: Temperature Annual Range (BIO5-BIO6) (ºC * 10)
- bio8: Mean Temperature of Wettest Quarter (ºC * 10)
- bio9: Mean Temperature of Driest Quarter (ºC * 10)
- Dim.1: first axis of a PCA performed for the climatic variables
- Dim.2: second axis of a PCA performed for the climatic variables
- LDMC: Community Weighted Mean of Leaf Dry Matter Content (g g-1)
- Plant.height: Community Weighted Mean of plant height (m)
- Plant.height.generative: Community Weighted Mean of height of the generative part of the plant (m)
- Plant.height.vegetative: Community Weighted Mean of height of the vegetative part of the plant (m)
- Seed.dry.mass: Community Weighted Mean of the seed dry mass (g)
- LA_leaf_avg: Community Weighted Mean of the leaf area (mm2)
- LA_leavlet_avg: Community Weighted Mean of the leaflets area (for compound leaves) (mm2)
- SLA_avg: Community Weighted Mean of the Specific Leaf Area (mm2 mg-1)
- Phanerophyt: Community Weighted Mean of the Phanerophyt category in Raunkiers classification (%)
- Therophyte: Community Weighted Mean of the Therophyte category in Raunkiers classification (%)
- Hemicryptophyte: Community Weighted Mean of the Hemicryptophyte category in Raunkiers classification (%)
- Chamaephyte: Community Weighted Mean of the Chamaephyte category in Raunkiers classification (%)
- Cryptophyte: Community Weighted Mean of the Cryptophyte category in Raunkiers classification (%)
- Fern: Community Weighted Mean of the dicotomic variable Fern/No Fern (%)
- Graminoid: Community Weighted Mean of the dicotomic variable Graminoid/No Graminoid (%)
- Herb: Community Weighted Mean of the dicotomic variable Herb/No Herb (%)
- Herb-Shrub: Community Weighted Mean of the dicotomic variable Herb-Shrub/No Herb-Shrub (%)
- Shrub: Community Weighted Mean of the dicotomic variable Shrub/No Shrub (%)
- Shrub-Tree: Community Weighted Mean of the dicotomic variable Shrub-Tree/No Shrub-Tree (%)
- Tree: Community Weighted Mean of the dicotomic variable Tree/No Tree (%)
- Woody: Community Weighted Mean of the dicotomic variable Woody/No Woody (%)
- Non_Woody: Community Weighted Mean of the dicotomic variable Non-Woody/No Non-Woody (%)
- slope.OLS: slope estimate of the lm(log(variance)~log(mean)) with the Ordinary Least Squares method Equivalent to the* b* value (exponent of the power relationship) in TPL
- slope.OLS.lim.inf: inferior limit of the 95% confidence interval of the slope estimate with the OLS method
- slope.OLS.lim.sup: superior limit of the 95% confidence interval of the slope estimate with the OLS method
- intercept.OLS: intercept estimate of the lm(log(variance)~log(mean)) with the Ordinary Least Squares method. Equivalent to the* a* value in TPL
- OLS.p:* p-value* of the difference between the slope estimate with OLS method and the value 2. Obtained with the slope.test() function in the package* lmodel2* that uses the Warton et al., (2006) method
- slope.SMA: slope estimate of the lm(log(variance)~log(mean)) with the Standardised Major Axis method. Equivalent to the* b* value (exponent of the power relationship) in TPL
- slope.SMA.lim.inf: inferior limit of the 95% confidence interval of the slope estimate with the SMA method
- slope.SMA.lim.sup: superior limit of the 95% confidence interval of the slope estimate with the SMA method
- intercept.SMA: intercept estimate of the lm(log(variance)~log(mean)) with the SMA method. Equivalent to the* a* value in TPL
- SMA.p: p-value of the difference between the slope estimate with SMA method and the value 2. Obtained with the slope.test() function in the package* lmodel2* that uses the Warton et al., (2006) method
- r2: variance explained measured as R-squared of the lm(log(variance)~log(mean)) model. It is valid for OLS and SMA methods
File: comm_1.txt
Description: this file contains the abundance data of one example community necessary to draw the first figure. Name of the community and the species have been erase for data protection reasons.
Variables
- LOTVS_ID: simmulated unique LOTVS_ID number
- Sp_ID: summulated species ID
- Year: year of sampling
- Abun: estimated abundance sampled.
File: coms20.txt
Description: this file contains the abundance data of 20 randomly selected communities necessary to draw the first figure. Name of the community and the species have been erase for data protection reasons.
Variables
- LOTVS_ID: simmulated unique LOTVS_ID number
- b: estimated community TPL scaling factor b
- delta: estimated dominance effect calculated as complementary of delta
- CVratio: estimated dominance effect calculated as the ratio of CVs
File: R_code.Rmd
Description: RMarkdown file to run all the analysis in the paper. The RMarkdown file that accompanies this data is ready to be used in an R environment. The only precaution to be made is to place the RMarkdown file and the data files in the same folder so it can be accessed.
Code/software
RStudio (Software capable of running R) is used for statistical analysis and data visualization.
Access information
Data was derived from the following sources:
- LOTVS Database of permanent vegetation plots (https://lotvs.csic.es/)