The proportion of low abundance species is a key predictor of plant β-diversity across the latitudinal gradient
Data files
Jan 09, 2025 version files 126.41 KB
-
Metadata_Xiao_et_al_simulation_vary_k.docx
13.21 KB
-
Metadata_Xiao_et_al_simumation_prediction_SAD.docx
13.54 KB
-
README.md
2.05 KB
-
Xiao_et_al_a_simulation_PL_beta_relationship_different_k_ranges.xlsx
9.20 KB
-
Xiao_et_al_code_simulation_and_null_model.zip
27.32 KB
-
Xiao_et_al_simulation_k_from_-0.5_to_5.xlsx
22.48 KB
-
Xiao_et_al_simumation_prediction_SAD.xlsx
38.62 KB
Abstract
The diversity of life displays very strong patterns of disparity across the Earth. Beta (β)-diversity (species compositional differences among sites) of woody plants, for instance, has usually been documented to decline with increasing latitude. Understanding these patterns, however, remains a grand challenge in ecology and evolution.
We develop a mathematical model to explain patterns of β-diversity across multiple landscapes. The model effectively predicts β-diversity in simulated and natural communities, regardless of the types of species abundance distributions.
Our model provides the novel insight that the proportion of species in the lowest abundance category (PL), which represents the share of relatively rare species in the regional species pool, is the key predictor of plant β-diversity. By applying the model to global forest inventories sampled from 40.7° S to 60.7° N, we find that PL explains nearly 85% of the variation in plant β-diversity along the global latitudinal gradient. Through a series of numerical simulations, we further show that the predictive power of PL on plant β-diversity on a global scale is largely determined by the variation of intraspecific aggregation among different communities.
Synthesis: We develop a new sampling model to predict patterns of β-diversity and find that the proportion of species in the lowest abundance category explains the majority of the variation in plant β-diversity along the latitudinal gradient. Our work provides a new tool in analyzing β-diversity and advances the theoretical understanding of large-scale β-diversity patterns across environmental gradients.
README: The proportion of low abundance species is a key predictor of plant β-diversity across the latitudinal gradient
https://doi.org/10.5061/dryad.n8pk0p35r
Description of the data and file structure
The data were simulation data that are generated using the R software (version 3.6.1). R codes for generating the data are also attached in the zip file "Xiao_et_al_code_simulation_and_null_model.zip"
Files and variables
Xiao et al. Simulation of communities of different SADs to test the model
This dataframe contains simulation data for sampling from communities with four different species abundance distributions (SADs, including lognormal, gamma, exponential and truncated-hyperbolic distributions).
Xiao_et_al_simumation_prediction_SAD.xlsx
-This dataframe contains α-diversity and β-diversity predicted from the model and observed in simulated communities of each SAD.
Metadata_Xiao_et_al_simumation_prediction_SAD.docx
Xiao et al. Simulation of different aggregation parameters
These dataframes contain simulation data for sampling from simulated communities with different values of proportion of species in the lowest abundance category (P_L) and different values of aggregation parameters (k).
Xiao_et_al_a_simulation_PL_beta_relationship_different_k_ranges.xlsx
-This dataframe contains the values of r-square for the relationship between P_L and β-diversity under different ranges for randomly generating k.
Xiao_et_al_simulation_k_from_-0.5_to_5.xlsx
-This dataframe contains β-diversity in one simulation where aggregation parameters are randomly generating between a range of -0.5~5.
Metadata_Xiao_et_al_simulation_vary_k.docx
Xiao et al. Codes for simulations and null model analyses
R codes for simulations and the null model analyses conducted in the current study. A detailed README.docx file is within the ZIP file that explains the codes in detail.
Xiao_et_al_code_simulation_and_null_model.zip