Data from: Global patterns and drivers of extinction risk in island endemic terrestrial vertebrates
Data files
Jul 07, 2026 version files 13.31 MB
-
Data_S1._Taxonomic_filtering_and_dataset_construction_for_32_214_non-Data_Deficient_terrestrial_vertebrate_species..xlsx
3.29 MB
-
Data_S10_Univariate_phylogenetic_logistic_regression_model_results_implemented_in_phyloglm.xlsx
49.67 KB
-
Data_S11_Model-averaged_phylogenetic_logistic_regression_results_implemented_in_phyloglm.xlsx
39.62 KB
-
Data_S2._Summary_of_non-DD_Terrestrial_vertebrate_species.xlsx
1.94 MB
-
Data_S3._Raw_data_for_the_30_711_analyzed_tetrapod_species_before_any_imputation.xlsx
6.17 MB
-
Data_S4._Species_richness_of_island_endemic_terrestrial_vertebrates_by_archipelago_and_taxonomic_class.xlsx
31.93 KB
-
Data_S5._Taxonomic_clustering_of_extinction_risk_among_island_endemic_terrestrial_vertebrate_families.xlsx
1.51 MB
-
Data_S6._Contingency_table_analyses_testing_the_association_between_major_IUCN_threat_types_and_Red_List_categories_in_threatened_non-Data_Deficient_terrestrial_vertebrates.xlsx
69.05 KB
-
Data_S7._Species-level_correlation_matrices.xlsx
117.59 KB
-
Data_S8._Univariate_phylogenetic_linear_model_results_implemented_in_phylolm.xlsx
50.91 KB
-
Data_S9._Model-averaged_phylogenetic_linear_model_results_implemented_in_phylolm.xlsx
28.12 KB
-
README.md
19.60 KB
Abstract
This Dryad repository contains the supplementary datasets supporting The Global Patterns and Drivers of Extinction Risk in Island Endemic Terrestrial Vertebrates. This repository integrates taxonomic information, distribution pattern data, compiled ecological and life-history traits, environmental and anthropogenic variables, as well as phylogenetic comparative analyses and statistical model outputs generated during the study. The datasets include more than 31,000 non-Data Deficient terrestrial vertebrate species, including birds, mammals, amphibians, and reptiles, based on data compiled from the IUCN Red List (version 2024-1) and subsequently standardized and refined.
The repository is organized into 11 supplementary datasets. These datasets include: (1) taxonomic filtering procedures and dataset construction workflows; (2) species-level taxonomic information, endemicity categories, conservation status, and population trends; (3) raw ecological and life-history variables prior to imputation and standardization; (4) archipelago-level summaries of species richness and high-risk endemic richness; (5) family-level analyses of phylogenetic clustering in extinction risk; (6) contingency table analyses linking IUCN threat categories with extinction risk across taxonomic and endemicity groups; (7) correlation matrices for predictor variables; and (8–11) outputs from phylogenetic linear models and phylogenetic logistic regression analyses, including both univariate and model-averaged results.
The repository contains categorical, binary, ordinal, and continuous variables compiled and harmonized from multiple global biodiversity databases. Missing values in raw datasets are consistently coded as NA. Statistical outputs include parameter estimates, confidence intervals, odds ratios, significance tests, model importance values and other related analytical outputs. All datasets are provided in Excel file format, with detailed descriptions of variables and worksheet structures documented in the Supplementary Materials.
These datasets have broad reuse potential for comparative macroecology, conservation biogeography, extinction risk modeling, island biodiversity research, phylogenetic comparative analyses, and studies examining taxon-specific responses to environmental and anthropogenic pressures. The repository also provides standardized data structures and reusable analytical outputs that may facilitate future cross-taxon conservation studies and meta-analyses.
All data were compiled from publicly available databases, and the original data sources are indicated within the corresponding Excel files. The repository contains no human subject data or sensitive personal information. Species nomenclature was taxonomically reconciled during dataset construction. All derived datasets are intended for scientific research and conservation purposes, with appropriate citation of the original data sources and associated publication.
Dataset DOI: 10.5061/dryad.kprr4xhkf
Description of the data and file structure
This README describes the structure, content, and purpose of all Supplementary Data files associated with the study on global patterns and drivers of extinction risk in island-endemic terrestrial vertebrates. All datasets are provided as separate files and are referenced as Data S1–Data S11 in the main text and Supplementary Information. NA indicates data that were unavailable or not collected from the sources; Not_applicable indicates variables or analyses that did not apply to a particular taxonomic group or model structure; and Not_included indicates variables that were excluded from model averaging or multivariable analyses following variable screening, model selection, or significance filtering procedures.
Files and variables
File: Data_S1._Taxonomic_filtering_and_dataset_construction_for_32_214_non-Data_Deficient_terrestrial_vertebrate_species..xlsx
Description: Data S1 documents the full taxonomic reconciliation, habitat filtering, and data-availability screening pipeline applied to terrestrial vertebrates from the IUCN Red List (version 2024-1). It provides transparency and reproducibility for all inclusion and exclusion steps.
Key variables：
- iucn_assessment_id, iucn_taxon_id: Unique IUCN Red List identifiers
- scientific_name: Accepted binomial name after taxonomic reconciliation
- iucn_category: IUCN Red List category (EX and EW retained and treated as extinct)
- excluded_dd: Indicator for exclusion of Data Deficient species
- synonym_or_subspecies, synonym_type: Resolved synonyms or excluded subspecies
- marine_system, freshwater_system, terrestrial_system: Habitat system flags
- excluded_taxonomic_issue, excluded_non_terrestrial: Filtering indicators
- retained_in_final_dataset: Species retained after taxonomic and habitat filtering (n = 31,892)
- missing_core_traits: Species lacking required ecological or life-history traits
- included_in_analysis: Species entering phylogenetic imputation and modelling (n = 30,711)
File: Data_S2._Summary_of_non-DD_Terrestrial_vertebrate_species.xlsx
Description: Data S2 provides core taxonomic, biogeographic, and conservation assessment information for all non-Data Deficient terrestrial vertebrates included in the study.
Key variables：
Scientific name and IUCN assessment identifier
Taxonomic classification (class, order, family)
Endemicity category:
1 = island endemic
2 = continental endemic
3 = dual-range (islands and continents)
IUCN Red List category and assessment criteria
Criteria version (IUCN v3.1)
Population trend (IUCN 2024-1)
File: Data_S3._Raw_data_for_the_30_711_analyzed_tetrapod_species_before_any_imputation.xlsx
Description: Data S3 contains the original, unimputed species-level data compiled from multiple sources. All variables are provided exactly as assembled prior to taxonomic imputation, log-transformation, or standardization. Missing values are indicated as NA (not available).
Key variables：
| Variable | Description |
|---|---|
| ScientificName | Scientific name of the species. |
| GenusName | Genus of the species. |
| BodyMass | Adult body mass (g). |
| BodyLength | Maximum adult body length (mm; amphibians only). |
| Verticality | Vertical stratum preference (0 = fossorial, 1 = semifossorial, 2 = terrestrial/aquatic, 3 = semi-arboreal, 4 = arboreal/aerial). |
| TrophicLevel | Trophic level (0 = omnivore, 1 = herbivore, 2 = carnivore). |
| Nocturnality | Activity pattern (0 = nocturnal, 1 = diurnal, 2 = cathemeral/crepuscular). |
| TypeGrow | Larval development mode for amphibians (1 = indirect, 2 = direct, 3 = viviparous). |
| MajorHabitatSum | Number of IUCN major habitat types occupied (1–13). |
| Latitude | Range centroid latitude (decimal degrees). |
| AnnuMeanTemp | Annual mean temperature (°C). |
| AnnuPrecip | Annual mean precipitation (mm). |
| TempSeasonality | Temperature seasonality (SD × 100). |
| PrecipSeasonality | Precipitation seasonality (coefficient of variation). |
| Elevation | Mean elevation (m). |
| RangeSize | Geographic range size (number of occupied 1 km × 1 km grid cells). |
| HumanFootPrint | Human Footprint Index (0–50). |
| InvasiveSpeciesResidual | Residual invasive species richness after controlling for range size. |
| DescriptionTime | Years since species description (2024 − year of description). |
| DirectSourceBodyMass/BodyLength | Original data source for body mass or body length. |
| DirectSourceTrophicLevel | Original data source for trophic level. |
| DirectSourceNocturnality | Original data source for activity pattern. |
| DirectSourceVerticality | Original data source for verticality. |
| DirectSourceHabitat | Original data source for habitat information. |
| DirectSourceGeo | Original data source for geographic and environmental variables. |
| SourceAnthropogenicPressures | Original data source for anthropogenic pressure variables. |
File: Data_S4._Species_richness_of_island_endemic_terrestrial_vertebrates_by_archipelago_and_taxonomic_class.xlsx
Description: Data S4 summarizes patterns of island endemic species richness and high-risk richness across global archipelagos and taxonomic classes. This dataset contains 11 sheets: (1) Archipelago label and name correspondence table; (2) Overall endemic richness (all tetrapods); (3) high-risk endemic richness (all tetrapods); (4–7) endemic richness for birds/mammals/reptiles/amphibians; (8–11) high-risk endemic richness for each class.
Key variables：
| Variable | Description |
|---|---|
| Label | Numeric identifier assigned to each archipelago and used in the figures. |
| Archipelago | Name of the archipelago. |
| Endemic_Richness | Number of island endemic species recorded for their primary area of occurring archipelago . |
| High_Risk_Endemic_Richness | Number of island endemic species classified as Vulnerable (VU), Endangered (EN), Critically Endangered (CR), or Extinct (EX) for each archipelago. |
File: Data_S5._Taxonomic_clustering_of_extinction_risk_among_island_endemic_terrestrial_vertebrate_families.xlsx
Description: Data S5 reports family-level analyses of extinction-risk clustering among island endemic terrestrial vertebrates, conducted separately for each vertebrate class. This dataset contains four worksheets, one for each vertebrate class.
Key variables：
| Variable | Description |
|---|---|
| Family | Taxonomic family name. |
| N_Species | Number of island endemic species in the family. |
| N_High_Risk | Number of high-risk island endemic species (VU, EN, CR, and EX) in the family. |
| High_Risk_Prop | Proportion of high-risk species in the family. |
| P_Value | Two-sided exact binomial test P value. |
| Direction | Direction of deviation from the expected proportion ("Higher" or "Lower"). |
| Significant | Indicates whether the result remained statistically significant after Dunn–Šidák correction. |
| Observed_Prop | Observed proportion of high-risk species in the family. |
| Expected_Prop | Expected proportion of high-risk species based on the corresponding vertebrate class. |
| Ci_Lower | Lower bound of the Clopper–Pearson confidence interval. |
| Ci_Upper | Upper bound of the Clopper–Pearson confidence interval. |
| Alpha_Original | Original significance level (α = 0.05). |
| Alpha_Adjusted | Dunn–Šidák-adjusted significance level. |
| Confidence_Level | Adjusted confidence level corresponding to the Dunn–Šidák correction. |
| N_Comparisons | Number of family-level comparisons used for multiple-testing correction. |
| Method | Statistical method applied to the family. |
| Risk_Category | Classification of the family's extinction risk pattern. |
| Test_Performed | Indicates whether a formal statistical test was performed. |
| Test_Note | Additional notes describing the statistical analysis or explaining why no test was performed. |
File: Data_S6._Contingency_table_analyses_testing_the_association_between_major_IUCN_threat_types_and_Red_List_categories_in_threatened_non-Data_Deficient_terrestrial_vertebrates.xlsx
Description: Data S6 contains contingency table analyses testing associations between major IUCN threat types and Red List categories in threatened non-DD terrestrial vertebrates. This Excel workbook contains 14 sheets.
Sheet 1: Summary of Pearson’s χ² tests of independence for each of the eight taxonomic class × endemism combinations (island endemic, continental endemic, dual-range). Columns include: taxonomic class, endemism status, total χ² statistic, degrees of freedom, raw p-value, and significance level (***p < 0.001, **p < 0.01, *p < 0.05, ns = not significant).
Sheets 2–13: Detailed cell-by-cell results for each class × endemism combination (e.g., "Aves_Island_Endemic", "Aves_Continental_Endemic", "Aves_Dual-Range"). Columns report, for every combination of the 11 major IUCN threat types (threat1 to threat11) and the four Red List categories (VU, EN, CR, EX): observed count, expected count under the null hypothesis of independence, standardized Pearson residual ((observed - expected)/√expected), χ² contribution ((observed - expected)²/expected), raw cell p-value, FDR-adjusted p-value (Benjamini–Hochberg correction within each group), and significance of the cell-level χ² contribution based on adjusted p-value (***p_adj < 0.001, **p_adj < 0.01, *p_adj < 0.05, blank otherwise).
Sheet 14: IUCN threats classification scheme (version 3.1; Threat 12 (Other threat) and Unknown were excluded.
Key variables：
| Variable | Description |
|---|---|
| Threat_Type | IUCN major threat category (Threat 1–Threat 11, following the IUCN Threats Classification Scheme v3.1). |
| RedlistGroup | IUCN Red List category (VU, EN, CR, or EX). |
| Count | Observed number of species in the corresponding threat type and Red List category. |
| Expected | Expected count under the null hypothesis of independence. |
| Residual | Standardized Pearson residual, calculated as (Observed − Expected)/√Expected. |
| Chi_Squared | Cell contribution to the Pearson χ² statistic, calculated as (Observed − Expected)²/Expected. |
| P_Value_Cell | Raw P value for the individual cell. |
| P_Adj | False discovery rate (FDR)-adjusted P value using the Benjamini–Hochberg procedure. |
| Significance_Cell | Significance level based on the adjusted P value (*** < 0.001, ** < 0.01, * < 0.05; blank = not significant). |
NOTE: Variables included in DATA S7-11 are defined in the Common variable definitions for Data S7–S11 section below.
File: Data_S7._Species-level_correlation_matrices.xlsx
Description: Data S7 contains 32 sheets of correlation matrices showing pairwise Spearman correlation coefficients among all univariate predictor variables used in the extinction risk analyses. Correlations were calculated for four vertebrate classes (birds, mammals, amphibians, reptiles) and multiple species subgroups defined by endemicity and IUCN criteria, respectively. Notably, “criteria2” sheets correspond to subsets of species within each class and endemicity type for which species classified as threatened under IUCN Criteria B or D2 were excluded. For each dataset, Spearman correlation coefficients were first Fisher z-transformed, averaged across 100 paired datasets, and then inverse-transformed to obtain mean correlation values. Each sheet corresponds to one class-subgroup combination, with variable names mapped to standardized analysis identifiers. Values range from –1 to 1, with positive values indicating positive associations between variables.
File: Data_S8._Univariate_phylogenetic_linear_model_results_implemented_in_phylolm.xlsx
Description: Data S8 reports the results of univariate phylogenetic linear models (PLMs) implemented using the phylolm framework, evaluating the association between individual predictor variables and extinction risk (ordinal IUCN categories, LC = 0 to EX = 5) across global terrestrial vertebrates. For each predictor, effect estimates, confidence intervals, and statistical significance are reported separately for each vertebrate group and endemicity category. These univariate analyses were used to characterize the marginal effects of species traits and external variables prior to multivariable model averaging.
File: Data_S9._Model-averaged_phylogenetic_linear_model_results_implemented_in_phylolm.xlsx
Description: Data S9 summarizes model-averaged parameter estimates derived from phylogenetic linear models (PLMs) implemented using the phylolm framework fitted to 100 paired datasets. Estimates incorporate both within-dataset and between-dataset variance, and are reported for each vertebrate group, endemicity category, and extinction risk criteria subset.
File: Data_S10_Univariate_phylogenetic_logistic_regression_model_results_implemented_in_phyloglm.xlsx
Description: Data S10 contains the results of univariate phylogenetic logistic regression models implemented using the phyloglm framework, evaluating the association between individual predictor variables and extinction risk across global terrestrial vertebrates. Extinction risk is treated as a binary response based on IUCN Red List category. For each predictor variable, model outputs include effect estimates, standard errors, z-values, p-values, and 95% confidence intervals. In addition, phylogenetically corrected model parameters, including the phylogenetic signal parameter where applicable, are reported. Results are provided separately for each vertebrate class, endemicity category, and extinction risk criterion subset.
File: Data_S11_Model-averaged_phylogenetic_logistic_regression_results_implemented_in_phyloglm.xlsx
Description: Data S11 summarizes model-averaged parameter estimates derived from phylogenetic logistic regression models implemented using the phyloglm framework fitted across multiple candidate model sets. Estimates incorporate both within-dataset and between-dataset variance, and are reported for each vertebrate class, endemicity category, and extinction risk criteria subset. For each predictor variable, the dataset provides model-averaged effect estimates, standard errors, z-values, p-values, and model importance values, together with odds ratios and their corresponding 95% confidence intervals.
Common variable definitions for Data S7–S11
Dataset identifiers
| Variable | Description |
|---|---|
| ClassName | Vertebrate class. |
| EndemicType | Endemism category analyzed. |
| CriteriaB/D2 | Indicates whether species listed solely under IUCN Criteria B or D2 were included in the analysis. |
Predictor variables
The predictor variables (BodyMass, BodyLength, GrowthType, Verticality, Nocturnality, TrophicLevel, MajorHabitatSum, RangeSize, Latitude, AnnuMeanTemp, AnnuPrecip, TempSeasonality, PrecipSeasonality, Elevation, HumanFootPrint, InvasiveSpeciesResidual, and DescriptionTime) correspond to the species traits and environmental variables described for Data S3.
Model output variables
| Variable | Description |
|---|---|
| Estimate | Rubin’s rule pooled regression coefficient obtained from repeated phylogenetic model fits across 100 paired datasets. |
| SE / StdErr | Rubin’s rule pooled standard error, combining within-dataset variance and between-dataset variance across repeated analyses. |
| t_stat | t statistic from univariate phylogenetic linear models (PLMs) fitted using phylolm. |
| z.value / z_value | Wald z statistic. |
| p.value / p_value | Two-sided P value derived from the Wald statistic under the standard normal distribution assumption. |
| Lower_CI | Lower bound of the 95% confidence interval derived from Rubin-pooled uncertainty. |
| Upper_CI | Upper bound of the 95% confidence interval derived from Rubin-pooled uncertainty. |
| Mean_Lambda | Mean Pagel’s λ estimated from phylogenetic models (phylolm) across 100 paired datasets. |
| Mean_Alpha | Mean α parameter from univariate phylogenetic logistic regression models (phyloglm) across 100 paired datasets. |
| Degrees_Freedom | Residual degrees of freedom from individual phylogenetic model fits (varies across iterations; not Rubin pooled). |
| Mean_R2 | Mean coefficient of determination (R²) across repeated phylogenetic linear model fits. |
| Mean_adjR2 | Mean adjusted R² across repeated phylogenetic linear model fits. |
| Mean_Pseudo_R2 | Mean pseudo-R² across phylogenetic logistic regression models fitted in repeated analyses. |
| Mean_AIC | Mean Akaike Information Criterion (AIC) across all candidate models prior to model averaging (phyloglm). |
| Importance | Variable importance derived from model averaging weights. |
| Odds_Ratio | Exponentiated pooled coefficient (exp(Estimate)), interpreted as multiplicative effect size in phylogenetic logistic models. |
| OR_Lower | Lower bound of the 95% confidence interval for the odds ratio. |
| OR_Upper | Upper bound of the 95% confidence interval for the odds ratio. |
Code/software
Microsoft Excel is recommended to view our data.