Dataset for depth-dependent effects of grazing enclosure on soil carbon, carbon stock, nitrogen, and bulk density in Sub-Saharan African grasslands
Description
This dataset supports a meta-analysis on the effects of grazing enclosure systems on soil properties in Sub-Saharan African grasslands. It compiles quantitative observations extracted from 31 peer-reviewed studies published between 2007 and 2024, covering 229 study–parameter–depth combinations. The dataset includes treatment and control values for enclosed and grazed systems, with means, standard deviations, and sample sizes extracted for four soil response variables: bulk density, soil nitrogen, soil carbon, and soil carbon stock. Observations are harmonized into four standardized soil-depth classes: ≤10 cm, ≤15 cm, ≤20 cm, and ≤30 cm. The dataset also includes information required to calculate standardized effect sizes, including Hedges’ g, and to conduct random-effects meta-analysis by soil property and depth class. It is intended to support reproducibility, transparency, and reuse of the evidence base underlying the manuscript on depth-dependent effects of grazing enclosure on soil carbon dynamics, nitrogen, and bulk density in Sub-Saharan African rangelands.
Files
Steps to reproduce
The dataset was produced through a systematic literature search conducted in Scopus and Web of Science using terms related to grasslands, grazing management, enclosure or fencing, soil carbon, soil nitrogen, carbon stock, bulk density, and Sub-Saharan African countries. Retrieved records were screened to retain peer-reviewed English-language studies conducted in grassland or rangeland systems, where grazing enclosure or exclosure was evaluated against a comparable grazed or conventionally managed control. Studies were included only if they reported at least one target soil variable and provided extractable quantitative data, including means, standard deviations, and sample sizes for both treatment and control groups. For each eligible study, the study identifier, authors, publication year, country, soil variable, soil depth, treatment mean, treatment standard deviation, treatment sample size, control mean, control standard deviation, and control sample size were extracted. Observations were restricted to the 0–30 cm soil layer and harmonized into four standardized depth classes: ≤10 cm, ≤15 cm, ≤20 cm, and ≤30 cm. Each unique study–soil variable–depth combination was treated as a separate observation. Standardized mean differences were then calculated as Hedges’ g using the extracted treatment and control statistics with small-sample correction. The compiled dataset was analyzed using random-effects meta-analysis separately for each soil variable and depth class, generating pooled effect sizes, 95% confidence intervals, p-values, heterogeneity statistics, treatment–control scatter plots, and forest plots.