Data from: Functional diversity shapes the stability of reef fish biomass under global change

Mahaut, Lucie 1 ; Loiseau, Nicolas2; Villéger, Sébastien2; Auber, Arnaud3; Hautecoeurh, Cyril2; Maire, Anthony4; Mellin, Camille5; Mouquet, Nicolas6; Stuart-Smith, Rick7; Violle, Cyrille1; Mouillot, David2

Published Apr 15, 2025 on Dryad. https://doi.org/10.5061/dryad.1g1jwsv5z

Data files

Apr 15, 2025 version files 94.63 KB

final_data.rds
88.93 KB
README.md
5.70 KB

Abstract

Understanding how environmental and human pressures impact the temporal stability of fish community biomass on shallow reefs is essential for effective conservation and management. These pressures influence community stability directly, by affecting species stability and asynchrony in species’ fluctuations. However, their effects may also indirectly depend on the functional traits of the species composing the community, which remains poorly understood. Here, we examine both direct and indirect, trait-mediated effects of environmental variability and human impacts on species biomass stability and asynchrony in 215 Australian shallow reefs. These communities span a 10-degree sea surface temperature (SST) gradient and have been monitored over 14 years. Our results indicate higher asynchrony in tropical reefs due to higher trait diversity and trait redundancy and higher species stability in temperate communities due to higher mean trophic level. Human impacts, through its negative effects on species stability and trait diversity, was the main destabilising factor of fish community biomass. Temporal change in SST destabilised species biomass while increasing mean trophic level in fish communities. Overall, our findings show that a comprehensive analysis of the multiple facets of functional diversity is crucial to better understand the long-term stability of marine ecosystems under global change.

https://doi.org/10.5061/dryad.1g1jwsv5z

Description of the data and file structure

We used raw data, which are freely available. URLs are given in Supplementary Table 3, which accompanies our submitted manuscript. The R code used for data processing and analysis is given in this Dryad folder.

Files and variables

File: final_data.rds

Description: This is the processed data used in our analysis. It contains site attributes, fish biodiversity facets and abiotic variables. Missing values are NA.

More specifically, the 39columns correspond to:

site_code : name of the site (i.e., fish community) as in Edgar et al. (2023)
Longitude of the site
Latitude of the site
mean _ annual_biomass: averaged of annual total fish biomass across the 14 years of survey (unit: g) for each site. For each community, we summed the annual biomass of each species to get annual total fish biomass. Then we averaged across years. Because not all sites underwent monitoring throughout the 14-year period, we extrapolated data for missing year. The process is detailled in the Materials and Methods section of the related paper.
sd annual biomass: standard deviation of annual total fish biomass for each site (unit: g)
mean raw annual biomass: averaged of annual total fish biomass across the 14 years of survey (unit: g) for each site before filling in missing annual data - see details in Materials and Methods
sd raw data : standard deviation of annual total raw fish biomass (unit: g)
cv biomass: coefficient of variation of annual total fish biomass. It is computed as the ratio between the sd and the mean of annual fish biomass (unitless)
cv biomass raw: coefficient of variation of raw annual fish biomass (unitless)
cv sp biomass: average coefficient of variation of annual species biomass. For each species in the community, we compute the coefficient of variation of its annual biomass across the 14 years. Then, we averaged the coefficient of variations of all species presented in the community (unitless).
cv sp raw biomass: same as cv sp biomass but with raw biomass data (unitless)- see Materials and Methods for details.
synchrony: quantifies the covariation between the fluctuation of annual species biomass. See De Mazancourt & Loreau (2008) for details (unitless)
synchrony raw : same as synchrony but based on raw annual biomass data (unitless)
n: number of year with biomass information. varies between 7 and 14
richness: number of fish species in each community (site)
td1: taxonomic diversity of order 1 that is, exponential of shanon index (unitless) (not used in this study)
fd1: functional diversity of order 1 (unitless) - see Materials and Methods for details
fr1: functional richness (unitless) (not used in this study)
cwm_trophic level: Community Weighted Mean value of fish trophic level (unitless) - see Materials and Methods for details
cwm_length: - Community Weighted Mean value of fish length (unit: cm) - see Materials and Methods for details
cwm_depth: - Community Weighted Mean value of fish maximum depth (unit:m) - see Materials and Methods for details
cwm_grate: - Community Weighted Mean value of fish growth rate (unit:cm per year) - see Materials and Methods for details
n_distinct; number of functionally distinct species - see Materials and Methods for details
cool; warm and tropical: We classified each fish in three biogeographic groups (i.e. cool, tropical or warm; extracted from Edgar et al., 2023). Then for each community, we computed the summed of species relative biomass in each biogeographic group (respectively the column cool, warm and tropical) (unit: %)
high, low and medium: We classified each fish species in three classes of trophic level (low, medium and high) by splitting the distribution of the trophic level of fish species into three classes of equal effective. Then, we computed for each communities the sum of the relative biomass of species in the low, medium and high trophic level groups (i.e. dominance of low/medium/high trophic level) (unit: %)
mean sst: mean sea surface temperature (unit: °C) across the 14 years of survey
change sst: we quantify temporal changes in SST using linear models with ‘annual SST’ and ‘year’ as dependent and independent variables, respectively. We use the coefficient of regression of ‘year’ to quantify the annual rate of change in SST (unit: °C per year)
sd sst: standard deviation of the annual sea surface temperature across the 14 years of survey (unit: °C)
cv sst: coefficient of variation (i.e., sd/mean) of annual sea surface temperature across the 14 years of survey (unitless)
mean_chloro : mean chlorophyll content across the 14 years of survey (unit %), see Materials and methods for details
sd chloro: standard deviation of annual chlorophyll content across the 14 years of survey (unit %), see Materials and methods for details
cv chloro: coefficient of variation (sd/mean) of annual chlorophyll content
mean depth: mean site depth (unit: m) - see Materials and methods for details
max depth: maximal site depth (unit: m) - see Materials and methods for details
gravity: an index of human impact on shallow reef ecosystems - see Materials and methods for details

Code/software

We use R version 4.0.0 to process and analyse our data. URLs of the R packages that we used are given in the supplementary table 3 of our submitted manuscript. The script to process and analyse the data is organised in order to make the workflow easily interpretable.