Data from: Meta-analysis shows overall benefits of cattle silvopastoral systems for biodiversity in fragmented forest landscapes
Data files
Mar 18, 2026 version files 92.86 KB
-
DataS1_OverviewStudiesMeta-analysis.xlsx
32.74 KB
-
DataS2_ExtractedBiodiversityData_MetaAnalysis.csv
32.40 KB
-
DataS3_ObservedVsRarefiedRichness_MetaAnalysis.csv
5.49 KB
-
README.md
22.24 KB
Abstract
This submission contains three interrelated datasets supporting a systematic review and meta-analysis of biodiversity responses to cattle-based silvopastoral systems (SPS) in fragmented forest landscapes (Perez-Alvarez et al., Ecological Applications, 2026). The study synthesizes evidence from 45 studies across 15 countries, four biogeographic regions, and seven taxonomic groups to evaluate how SPS affect biodiversity (species diversity and abundance) relative to treeless pastures and natural forests.
The datasets are organized to reflect the sequential workflow of the meta-analysis: from literature screening and study selection, through effect-size calculation, to sensitivity analyses. They are linked by a common study_id identifier that allows records to be traced across files.
Associated publication:
Perez-Alvarez, R., Chará, J., Snyder, L.D., Bonatti, M., Sieber, S., & Martin, E.A. (2026). Meta-analysis shows overall benefits of cattle silvopastoral systems for biodiversity in fragmented forest landscapes. Ecological Applications.
This repository contains three interrelated datasets supporting the meta-analysis described above. They are organized to reflect the sequential workflow of the study: from literature screening and study selection (DataS1), through effect-size calculation (DataS2), to sensitivity analyses (DataS3). All datasets are linked by a common study_id identifier that allows records to be traced across files.
Note on missing values: Cells containing n/a across all three datasets represent values that were not recorded in the original source or are not applicable.
Table of Contents
- Authors
- Funding Information
- License
- DataS1 — Overview of Studies
- DataS2 — Extracted Biodiversity Data
- DataS3 — Observed vs Rarefied Richness
Authors
| # | Name | Affiliation | Contact |
|---|---|---|---|
| 1 | Ricardo Perez-Alvarez | Institute of Animal Ecology and Systematics, Justus-Liebig-University of Giessen, 35390 Giessen, Germany | Ricardo.Perez-Alvarez@allzool.bio.uni-giessen.de |
| 2 | Julián Chará | Centre for Research on Sustainable Agriculture, CIPAV, 760042 Cali, Colombia | |
| 3 | Lauren D. Snyder | TIB – Leibniz Information Centre for Science and Technology, 30167 Hannover, Germany | |
| 4 | Michelle Bonatti | Leibniz Centre for Agricultural Landscape Research (ZALF), 15374 Müncheberg, Germany | |
| 5 | Stefan Sieber | Leibniz Centre for Agricultural Landscape Research (ZALF), 15374 Müncheberg, Germany | |
| 6 | Emily A. Martin | Institute of Animal Ecology and Systematics, Justus-Liebig-University of Giessen, 35390 Giessen, Germany |
Funding Information
The work was supported by the Leibniz Centre for Agricultural Landscape Research (ZALF) (Auftrag ZALF 111450) to MB and SS. RPA was supported by a Fulbright-Minciencias fellowship. JC received support from IKI-BMU (Project 18_III_106_COL_A_Sustainable Production Strategies). LDS received funding through a Leibniz University Hannover Flexible Funds Grant (project number: 51171322).
License
All datasets in this repository are released under a Creative Commons Zero (CC0) waiver. All data deposited at Dryad are subject to the CC0 license, placing them in the public domain with no restrictions on use.
https://creativecommons.org/publicdomain/zero/1.0/
DataS1 — Overview of Studies
File: DataS1_OverviewStudiesMeta-analysis.xlsx
Date of dataset creation: 2026-03
Keywords: agroforestry, cattle ranching, forest conservation, land use changes, livestock grazing, meta-analysis, research synthesis
Dataset Description
This dataset is a curated inventory of primary sources screened and/or included in a systematic review and meta-analysis assessing biodiversity responses to cattle silvopastoral systems (SPS) in fragmented forest landscapes. The file provides bibliographic information and study-level metadata for each source, including study location, biogeographic region, taxonomic group, and biodiversity outcome type.
- Temporal coverage (publication years): 1998–2022
- Spatial coverage: 15 countries spanning Tropics, Subtropics, Temperate, and Mediterranean regions
Each row represents a single record for a given study_id and a specific combination of country/region, taxonomic_group, and biodiversity_metric (diversity or abundance); therefore, a study may appear in multiple rows.
File Contents
| Sheet | Name | Description |
|---|---|---|
| 1 | Selected_studies_meta-analysis | Studies included in the meta-analysis with bibliographic metadata and coding variables (location, region class, taxonomic group, outcome type, and notes). |
| 2 | Articles_Excluded | Screened studies excluded from effect-size calculation due to unavailable variability/uncertainty information (SD or SE absent from main text and supplementary materials, or not supplied by authors upon request). |
Methods
Study metadata were compiled during the systematic literature search, screening, and data-extraction workflow for the associated publication. Definitions of inclusion/exclusion criteria, screening steps, and coding rules (including how studies were grouped under study_id and categorized by region, taxonomic group, and outcome type) are provided in the associated publication and its supplementary materials.
Column Descriptions
Columns are present in Sheet 1 (Selected_studies_meta-analysis) unless noted. full_text_URL appears exclusively in Sheet 2 (Articles_Excluded). All other columns are shared across both sheets.
| Column | Sheet | Description |
|---|---|---|
additional_notes |
1 | Any relevant information about how a study was coded in the meta-analysis. For example, notes clarifying when two papers share the same dataset and were therefore assigned the same study_id. |
authors |
1, 2 | Full list of authors of the source document. |
biogeographic_region |
1, 2 | Broad climatic zone used for subgroup analyses. Assigned using study location (latitude and/or source-reported climate/biome region). Allowed values: Tropics | Subtropics | Temperate | Mediterranean |
biodiversity_metric |
1 | Type of biodiversity measure used in the comparison between silvopastoral systems and the reference system. Broadly classified as diversity (e.g., species richness, diversity indices) or abundance. Allowed values: diversity | abundance |
country |
1, 2 | Country where the study was conducted. |
DOI_ISBN_URI |
1, 2 | Persistent identifier for the source document: Digital Object Identifier (DOI), ISBN, or URI link, as available. If a DOI is available, it is listed; otherwise a persistent URI/handle is provided. ISSN/ISBN may also appear for non-DOI sources. |
full_text_URL |
2 | Direct URL to the full text of the article or document. Primarily populated for excluded studies; may be blank for included studies when access is via DOI or paywalled sources. |
journal_volume_issue |
1, 2 | Journal name, volume, issue, and page numbers where the study was published. |
region |
1, 2 | Specific location within the country where the study was conducted. This may refer to a political division (e.g., state, department, or province) or a recognized geographic region, as reported in the original paper. |
source_title |
1, 2 | Full title of the article, thesis, or report from which biodiversity data were extracted. |
study_id |
1, 2 | Unique identifier assigned to each independent study. Studies sharing the same study_id draw from the same underlying dataset, even if published as separate papers. This identifier is consistent across DataS1, DataS2, and DataS3, and can be used to link records across files. |
study_No |
1, 2 | Sequential number assigned to each study in the dataset. A single study may appear in multiple rows if it reported data for multiple metrics or taxonomic groups. |
study_year |
1, 2 | Year the study was published. This may differ from the year(s) in which biodiversity data were collected in the field. |
taxonomic_group |
1 | Broad taxonomic category used for subgroup analyses in the meta-analysis. Groups individual taxa into higher-level categories. Allowed values: insects | birds | mammals | plants | dung beetles | macroinvertebrates | bacteria and fungi |
DataS2 — Extracted Biodiversity Data
File: DataS2_ExtractedBiodiversityData_MetaAnalysis.csv
Date of dataset creation: 2026-03
Keywords: agroforestry, cattle ranching, forest conservation, land use changes, livestock grazing, meta-analysis, research synthesis, silvopasture
Dataset Description
This dataset contains the extracted biodiversity summary statistics used to calculate effect sizes in the meta-analysis of cattle silvopastoral systems (SPS). The file includes study-level means, measures of dispersion (SD), and sample sizes (n) for biodiversity outcomes measured in SPS and in the reference systems (treeless pastures and/or forests).
- Temporal coverage (publication years): 1998–2022
- Spatial coverage: 15 countries spanning Tropics, Subtropics, Temperate, and Mediterranean regions
The dataset is in wide format: each row contains summary statistics for SPS and, where available, treeless pasture and/or forest reference systems. Missing reference systems are recorded as NA.
File Contents
One CSV file; single table.
Methods
Biodiversity summary statistics (means, SD/SE/CI-derived SD where applicable, and sample sizes) were extracted from primary sources during the systematic review workflow for the associated publication. When original biodiversity measures differed across studies, values were standardized to the analysis categories (diversity vs abundance) following the procedures described in the associated publication and its supplementary materials.
Column Descriptions
| Column | Description |
|---|---|
biogeographic_region |
Broad climatic zone used for subgroup analyses. Assigned using study location (latitude and/or source-reported climate/biome region). Allowed values: Tropics | Subtropics | Temperate | Mediterranean |
country |
Country where the study was conducted. |
covariance_group |
Grouping variable used to link diversity and abundance records drawn from the same sampling unit within a study, enabling within-study covariance to be modeled. Records from the same study, location, and taxonomic assemblage share a covariance_group identifier. |
latitude_absolute |
Absolute value of latitude (decimal degrees), i.e., distance from the equator, with values ranging from 0 (equator) to 90 (poles). |
mean_forest |
Mean value of outcome_type per assemblage for the forest treatment. |
mean_sps |
Mean value of outcome_type per assemblage for the silvopastoral systems treatment. |
mean_treeless_pasture |
Mean value of outcome_type per assemblage for the treeless pasture treatment. |
n_forest |
Sample size for the forest treatment. |
n_sps |
Sample size for the silvopastoral system treatment. |
n_treeless_pasture |
Sample size for the treeless pasture treatment. |
original_metric |
Original biodiversity metric reported in the source (e.g., species richness, Shannon index, abundance). These were transformed into Hill numbers and aggregated into two general categories: diversity_hill_numbers or abundance. See outcome_type. |
outcome_type |
Type of biodiversity measure used in the comparison between silvopastoral systems and the reference system. Allowed values: diversity_hill_numbers | abundance |
region |
Specific location within the country where the study was conducted. This may refer to a political division (e.g., state, department, or province) or a recognized geographic region, as reported in the original paper. |
row_id |
Unique identifier for each row (one extracted outcome record). |
sd_forest |
Standard deviation of outcome_type for the forest treatment. |
sd_sps |
Standard deviation of outcome_type for the silvopastoral system treatment. |
sd_treeless_pasture |
Standard deviation of outcome_type for the treeless pasture treatment. |
specific_taxa |
Specific taxonomic group reported in the study, at the resolution reported in the original source. These taxa were grouped into broader categories for subgroup analyses. See taxonomic_group. |
study_id |
Unique identifier assigned to each independent study. Studies sharing the same study_id draw from the same underlying dataset, even if published as separate papers. This identifier is consistent across DataS1, DataS2, and DataS3, and can be used to link records across files. |
study_year |
Publication year of the source study. This may differ from the year(s) in which biodiversity data were collected in the field. |
taxonomic_group |
Broad taxonomic category used for subgroup analyses in the meta-analysis. Groups individual taxa into higher-level categories. Allowed values: insects | birds | mammals | plants | dung beetles | macroinvertebrates | bacteria and fungi |
DataS3 — Observed vs Rarefied Richness
File: DataS3_ObservedVsRarefiedRichness_MetaAnalysis.csv
Date of dataset creation: 2026-03
Keywords: agroforestry, cattle ranching, forest conservation, land use changes, livestock grazing, meta-analysis, research synthesis, silvopasture
Dataset Description
This dataset contains observed and coverage-standardized (rarefied/extrapolated) species richness estimates used in sensitivity analyses for the meta-analysis of cattle silvopastoral systems (SPS). The file includes study-level richness estimates (Hill number q = 0), measures of uncertainty, and sample sizes for SPS and reference systems (treeless pastures and/or forests).
- Temporal coverage (publication years): 2012–2022
- Spatial coverage: 5 countries spanning Tropics, Subtropics, and Temperate regions
The dataset is in wide format: each row contains values for SPS and, where available, treeless pasture and/or forest reference systems. Missing reference systems are recorded as NA. Rows are duplicated by richness_type (observed vs rarefied) for the same study/outcome record where both are available.
File Contents
One CSV file; single table.
Methods
We compiled studies that reported individual-based abundance data as species-by-assemblage matrices (14 studies for SPS vs treeless pastures; 12 studies for SPS vs forests). Studies reporting only pre-calculated rarefied richness were excluded because rarefaction methods are not fully interchangeable.
Coverage-based rarefaction/extrapolation was implemented using iNEXT.4steps in R. For each study and land-use category, we estimated sample coverage from observed abundance distributions, rarefied/extrapolated richness to a common target coverage, and obtained uncertainty estimates via bootstrap resampling (n = 100). Observed richness and coverage-standardized richness are provided in separate rows identified by richness_type (observed | rarefied). Effect sizes (lnRR) and sampling variances were then calculated from these richness estimates as described in the associated publication and Appendix S2.
Column Descriptions
| Column | Description |
|---|---|
biogeographic_region |
Broad climatic zone used for subgroup analyses. Assigned using study location (latitude and/or source-reported climate/biome region). Allowed values: Tropics | Subtropics | Temperate |
country |
Country where the study was conducted. |
mean_forest |
Mean species richness per assemblage for the forest treatment, either observed or coverage-standardized depending on richness_type. |
mean_sps |
Mean species richness per assemblage for the silvopastoral treatment, either observed or coverage-standardized depending on richness_type. |
mean_treeless_pasture |
Mean species richness per assemblage for the treeless pasture treatment, either observed or coverage-standardized depending on richness_type. |
n_forest |
Number of assemblages/replicates for the forest treatment. |
n_sps |
Number of assemblages/replicates for the silvopastoral system treatment. |
n_treeless_pasture |
Number of assemblages/replicates for the treeless pasture treatment. |
region |
Specific location within the country where the study was conducted. This may refer to a political division (e.g., state, department, or province) or a recognized geographic region, as reported in the original paper. |
richness_type |
Indicates whether richness values are observed (raw richness / Hill q = 0 from the source data) or rarefied (coverage-standardized richness estimated via iNEXT). Allowed values: observed | rarefied |
row_id |
Unique identifier for each row (one richness record for one study/taxon/location and one richness_type). |
sd_forest |
Standard deviation of species richness per assemblage for the forest treatment. For richness_type = rarefied, uncertainty was calculated via bootstrap resampling. |
sd_sps |
Standard deviation of species richness per assemblage for the silvopastoral system treatment. For richness_type = rarefied, uncertainty was calculated via bootstrap resampling. |
sd_treeless_pasture |
Standard deviation of species richness per assemblage for the treeless pasture treatment. For richness_type = rarefied, uncertainty was calculated via bootstrap resampling. |
specific_taxa |
Specific taxonomic group reported in the study, at the resolution reported in the original source. These taxa were grouped into broader categories for subgroup analyses. See taxonomic_group. |
study_id |
Unique identifier assigned to each independent study. Studies sharing the same study_id draw from the same underlying dataset, even if published as separate papers. This identifier is consistent across DataS1, DataS2, and DataS3, and can be used to link records across files. |
study_year |
Publication year of the source study. This may differ from the year(s) in which biodiversity data were collected in the field. |
taxonomic_group |
Broad taxonomic category used for subgroup analyses in the meta-analysis. Groups individual taxa into higher-level categories. Allowed values: insects | birds | plants | dung beetles |