Evolutionary history and environmental variability structure contemporary tropical vertebrate communities
Data files
Apr 03, 2024 version files 194.30 KB
Abstract
Tropical regions harbor over half of the world’s mammals and birds, but how their communities have assembled over evolutionary timescales remains unclear. To compare eco-evolutionary assembly processes between tropical mammals and birds, we tested how hypotheses concerning niche conservatism, environmental stability, environmental heterogeneity, and time-for-speciation relate to tropical vertebrate community phylogenetic and functional structure. We used in-situ observations of species identified from systematic camera trap sampling as realized communities from 15 protected tropical rainforests in four tropical regions worldwide. We quantified standardized phylogenetic and functional structure for each community and estimated the multi-trait phylogenetic signal (PS) in ecological strategies for the four regional species pools of mammals and birds. Using linear regression models, we test three non-mutually exclusive hypotheses by comparing the relative importance of colonization time, paleo-environmental changes in temperature and land cover since 3.3 Mya, contemporary seasonality in temperature and productivity, and environmental heterogeneity for predicting community phylogenetic and functional structure. The phylogenetic and functional structure showed non-significant yet varying tendencies toward clustering or dispersion in all communities. Mammals had stronger multi-trait PS in ecological strategies than birds (mean PS: mammal = 0.62, bird = 0.43). Distinct dominant processes were identified for mammal and bird communities. For mammals, colonization time and elevation range significantly predicted phylogenetic clustering and functional dispersion tendencies, respectively. For birds, elevation range and contemporary temperature seasonality significantly predicted phylogenetic and functional clustering tendencies, respectively, while habitat diversity significantly predicted functional dispersion tendencies. Our results reveal different eco-evolutionary assembly processes structuring contemporary tropical mammal and bird communities over evolutionary timescales that have shaped tropical diversity. Our study identified marked differences among taxonomic groups in the relative importance of historical colonization and sensitivity to environmental change.
README: Evolutionary history and environmental variability structure contemporary tropical vertebrate communities
https://doi.org/10.5061/dryad.ht76hdrpm
This repository contains the data and R codes of standardized phylogenetic and functional community structure, multi-trait phylogenetic signals in ecological strategies, and all statistical analyses. The species-level phylogeny and ecological traits of mammals and birds used to measure community structures and multi-trait phylogenetic signals, as well as the environmental variables in this paper are all publicly available and can be accessed via the sources in the Sharing/Access information section. Species checklists of this study were obtained from the species occurrence data of the Tropical Ecology and Assessment Monitoring Network, which can be accessed through Wildlife Insights.
Description of the data and file structure
We include the taxonomic names of mammals and birds used to match species-level and ecological traits and outputs of the analyses in this study.
TEAM_SpList_unify_name.csv for the species list of 15 communities from the Tropical Ecology Assessment and Monitoring (TEAM) Network with corresponding scientific names among camera-trap data, and species-level datasets of traits and phylogeny.
A. Sci_name – scientific name in camera-trap data
B. site_name – full name of study site.
C. site_code – abbreviation of study site.
D. Genus – genus of species.
E. Class – class of species.
F. Sci_name_trait – scientific name in species-level trait datasets.
G. Sci_name_phylo – scientific name in species-level phylogeny.
TEAM_Aves_Trait_unify_mammal.csv for the bird species-level trait values that were reclassified into the same data structure and definition of mammal species-level traits. There were six ecological traits used in quantifying functional community structure and multi-trait phylogenetic signal, including adult body mass (g), dietary breadth, trophic level, foraging stratum, activity cycle, and habitat breadth. For mammals, the six ecological traits can be obtained from Soria et al. 2021 with missing values in habitat breadth filled based on the habitat information from the IUCN Redlist in the R code of 1_TEAM_SpList_PD_FD.R. For birds, we directly obtained the body mass (g) and activity pattern from Wilman et al. (2014) by the columns of BodyMass-Value and Nocturnal. We reclassified the dietary breadth, forage stratum, and trophic level from Wilman et al. (2014) based on the definition of Soria et al. (2021) by the columns of Diet, ForStrat, and Diet-5Cat. This CSV file includes the reclassified traits and habitat breadth acquired from the IUCN Redlist without replicating values from Wilman et al. (2014) and Soria et al. (2021).
A. Sci_name – scientific name in camera-trap data.
B. Sci_name_trait – scientific name in species-level trait dataset.
C. Genus – genus of species.
D. Class – class of species.
E. Dietary_breadth_n – numeric values of the number of prevalent (≥ 20%) EltonTraits dietary categories consumed.
F. Trophic_level – ordinal values with three levels, classified by the categorical diet, including one for Herbivore (i.e., PlantSeed and FruiNect), two for Omnivore, and three for Carnivore (i.e., Invertebrate and VertFishScav).
G. ForStrat – ordinal values with three levels, classified by the prevalent (≥ 20%) forage stratum, including G for ground-foraging, S for scansorial foraging around the ground to the canopy, and Ar for arboreal foraging around mid-height to the canopy.
H. Habitat_breadth_total – numeric values of the number of IUCN habitat types.
Phylogenetic_signal_niche_by_realm_raw.csv for the raw values of multi-trait phylogenetic statistics of regional species pool across 100 sampled trees from the 10-k credible set of species-level phylogeny obtained from the VertLife for mammals and birds.
A. Realm – four regions of study sites.
B. Phylo_signal – multi-trait phylogenetic signal values.
C. Phylo_tree – phylogeny tree identity of the 100 sampled trees.
D. Class – class of species.
Phylogenetic_signal_niche_by_global_species_pool_raw.csv for the raw values of multi-trait phylogenetic statistics of the global mammal species pool across 10 sampled trees.
A. Realm – Complete tree.
B. Phylo_signal – multi-trait phylogenetic signal values.
C. Phylo_tree – phylogeny tree identity of the 10 sampled trees.
TEAM_PD_FD_Model_variables.csv for the site-level information of location (center coordinates), sampling area size, monitoring period, tropical regions, and variables to perform linear regression on standardized phylogenetic and functional structure (i.e., colonization time, environmental covariates, standardized phylogenetic and functional structure).
A. Site_code – abbreviation of study site.
B. Full_name – full name of study site.
C. Country – country of study site.
D. Sample_area_2kmBuff – sampling area size of the 2-km-radius buffer of the study site center coordinates.
E. Sampling_start – sampling start year of the study site.
F. Sampling_end – sampling end year of the study site.
G. region_realm – the region of the study site, including four regions.
H. Mamma_SR – mammal species richness of the study site.
I. Bird_SR – bird species richness of the study site.
J. Temp_deep_CV – coefficient of variation in annual mean temperature over twelve time periods since 3.3 Mya using the bioclimatic map at 2.5 arcmins from PaleoClim.org (Brown et al., 2018). This is the mean value within a 2-km-radius buffer of each study site.
K. T0_10000BC – ordinal values of land-cover rank in prehistorical time 10,000 BC (i.e., Holocene) using the reconstructed land-use maps at 5 arcmins from HYDE 3.2 (Klein Goldewijk et al., 2017). The land cover rank defined by Rowan et al. (2020) with six levels, one for natural forests, two for semi-natural forests, three for range lands, four for croplands, five for villages, and six for urban and dense settlements. This is the mean value within a 2-km-radius buffer of each study site.
L. LC_change – difference values between the land-cover rank in 2015 and prehistorical time 10,000 BC. This is the mean value within a 2-km-radius buffer of each study site.
M. Modern_BIO1 – annual mean temperature (Celsius degree) using the bioclimatic map at 2.5 arcmins for 1979-2013 (Karger et al., 2017). This is the mean value within a 2-km-radius buffer of each study site.
N. Modern_BIO4 – temperature seasonality (standard deviation) using the bioclimatic map at 2.5 arcmins for 1979-2013 (Karger et al., 2017). This is the mean value within a 2-km-radius buffer of each study site.
O. EVI_mean_long – annual mean enhanced vegetation index values using the 16-day products of the Terra Moderate Resolution Imaging Spectroradiometer at 250 meters from 2000 to 2015 (Didan, 2015). This is the mean value within a 2-km-radius buffer of each study site.
P. EVI_sea_long – coefficient of variation in monthly average enhanced vegetation index values using the 16-day products of the Terra Moderate Resolution Imaging Spectroradiometer at 250 meters from 2000 to 2015 (Didan, 2015). This is the mean value within a 2-km-radius buffer of each study site.
Q. elev_mean – mean elevation (m) within a 2-km-radius buffer of each study site using the digital elevation map from the Shuttle Radar Topography Mission dataset at 3 arcseconds (Jarvis et al., 2008).
R. elev_range – range of elevation (m) within a 2-km-radius buffer of each study site using the digital elevation map from the Shuttle Radar Topography Mission dataset at 3 arcseconds (Jarvis et al., 2008).
S. Shannon – the Shannon index values of habitat types from the IUCN level 2 habitat map (Jung et al., 2020), generated by the global land-cover map in 2015 at ~100 m resolution.
T. M.mean.colonize.time – the mean stem age (Mya) of mammal community phylogeny over 100 sampled trees.
U. A.mean.colonize.time – the mean stem age (Mya) of bird community phylogeny over 100 sampled trees.
V. M.mean.DR – the mean DR metric values of mammal species in each study site over 100 sampled trees. DR metric is defined by Jetz et al. (2012).
W. A.mean.DR – the mean DR metric values of bird species in each study site over 100 sampled trees. DR metric is defined by Jetz et al. (2012).
X. mpd.obs.z – mean standard effect size of community structures among all species pairs in each study site by shuffling taxa labels on regional phylogeny or functional space 10,00 times. For phylogenetic community structure, the mean standard effect size was averaged over 100 sampled trees.
Y. mntd.obs.z – mean standard effect size of community structures among all the nearest species pairs in each study site by shuffling taxa labels on regional phylogeny or functional space 10,00 times. For phylogenetic community structure, the mean standard effect size was averaged over 100 sampled trees.
Z. Class – class of species
AA. index – index for phylogenetic community structure (PD) or functional community structure (FD).
TEAM_PD_global_species_pool.csv for the site-level mean values of standardized phylogenetic community structures (i.e., SES.MPD and SES.MNPD) relative to the global mammal species pool across 10 sampled trees.
A. Site_code – abbreviation of study site.
B. mpd.obs.z.global – mean standard effect size of phylogenetic community structure among all the species pairs in each study site by shuffling 10,00 times over 10 sampled complete trees.
C. mntd.obs.z.global – mean standard effect size of phylogenetic community structure among all the nearest species pairs in each study site by shuffling 10,00 times over 10 sampled complete trees.
Sharing/Access information
The original data of species-level phylogeny and ecological traits of mammals and birds used in this paper can be accessed via the following sources.
- The species-level phylogeny of mammals and birds is obtained from VertLife.
- The functional trait data are obtained from the published datasets of mammals from Soria et al. 2021 and birds from Wilman et al. 2014.
Code/Software
This repository includes four R scripts to quantify the phylogenetic and functional community structure and phylogenetic signal in ecological strategies, and one R script to conduct the statistical analyses in the study.
- 1_TEAM_SpList_PD_FD.R for standardized phylogenetic and functional community structure, with R packages of phytools v1.9-16, ape v5.7-1, picante v1.8.2, and mFD v1.0.6. R function tipRateFunctions.R published by Title and Rabosky (2019) was used to measure the DR metric values.
- 2_Niche_phylogenetic_signal.R for measuring phylogenetic statistics of ecological strategies, with R package adiv v2.2 and R script of self-defined functions to order nodes of phylogeny for phylogenetic statistics in PhylogenyFunctions.R
- 3_Statistics_Visualization.R for all statistical analyses on phylogenetic and functional community structure and phylogenetic statistics of ecological strategies in the study, with R package MuMIn v1.47.5
- 4_Compare_tree_effect.R for comparing the effects of defining regional species pools and the global species pool, which include species evolved from multiple biogeographic regions, on phylogenetic statistics and community phylogenetic structures. This script provides the same processes in 2_Niche_phylogenetic_signal.R and 3_Statistics_Visualization.R to measure the phylogenetic statistics and phylogenetic structure relative to the global mammal species pool.
- PhylogenyFunctions.R for self-defined functions to order node weights and quantify the phylogenetic statistic value in 2_Niche_phylogenetic_signal.R and 4_Compare_tree_effect.R.