Making soil health assessment accessible: a set of simple, low-input procedures for soil health indicators that discriminate between land-use intensity classes in the Andes of Ecuador

Résumé fourni par la source

This repository contains the dataset and R analysis code underlying the study by Forbes et al. (2026), which evaluates and compares structural–biological (S-B) and physicochemical (P-C) soil health indicators across a gradient of land-use intensity in the Ecuadorian Andes. Soil samples were collected from 52 sites spanning four Andean provinces (Imbabura, Pichincha, Cotopaxi, and Chimborazo) between September 2023 and January 2024. Sites represent a range of management types (Agroecological, Traditional, Conventional, Fallow, Natural, and Degraded) and corresponding land-use intensities (Low, Moderate, High, and Degraded). For each site the dataset includes field-based structural and biological measurements (soil aggregate stability via the SLAKES image-analysis method, soil colour, soil macrofauna abundance and family richness, earthworm abundance, microbial biomass, and soil respiration measured by both the Solvita and KOH-trapping methods) alongside laboratory physicochemical indicators (soil texture, total nitrogen, available phosphorus, exchangeable bases, micronutrients, soil organic matter, and pH), together with site geolocation, elevation, and climate data (mean annual temperature and precipitation). The repository comprises two files: The data file (Forbes_et_al_2026_EA_data.xlsx) contains a "Data" sheet with all site-level measurements (52 samples × 46 variables) and a "Metadata" sheet providing a full data dictionary with variable names, descriptions, units, and measurement methods. The R script reproduces the multivariate analyses presented in the paper, including PERMANOVA and PERMDISP tests for the effects of land-use intensity and province, pairwise PERMANOVA comparisons, constrained ordination (CAP / distance-based redundancy analysis) with environmental vector fitting, comparison of variable loadings, province × land-use-intensity interaction tests, and regression between the S-B and P-C indicator sets along the primary ordination axis. Analyses were conducted in R using the vegan, ggplot2, dplyr, tidyr, patchwork, ggrepel, pairwiseAdonis, and openxlsx packages. These materials are provided to support transparency and reproducibility of the published analyses, and may be reused for comparative soil health assessment and methodological work on field-based versus laboratory soil health indicators. Note on reuse: Please cite the associated publication (Forbes et al., 2026) and this dataset when using these materials.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.