Agroecological practices and soil ecosystem services in cacao agroforestry farms of the Colombian Amazon: farm-level dataset
Description
Farm-level and sampling-point data from twenty-three smallholder cacao agroforestry farms in Belén de los Andaquíes, Caquetá, Colombia, sampled in 2024. Agroecological practices were characterized with the 20 categorical indicators of the five biophysical elements of the FAO Tool for Agroecology Performance Evaluation (TAPE, Step 1), and farms were classified into three typologies by multiple correspondence analysis and hierarchical clustering. Soil was sampled at three points per farm (0–30 cm composite for physicochemical determinations; undisturbed cores for bulk density and visual evaluation of soil structure). Four soil ecosystem services (erosion control, plant growth support, carbon storage, nutrient availability) were quantified as soil quality indices (0–1) following Silva-Olaya et al. (2022), and compared among typologies with linear mixed models. The dataset contains the TAPE indicator scores and typology of each farm, the measured soil properties and the four service indices at each of the 66 sampling points, site covariates from public cartography, the scoring thresholds used to build the indices, and the Python scripts that reproduce the analyses, supplementary tables and figures of the article. Farms are identified by anonymous codes (P01–P23); producer names and coordinates are not included. Data supporting: Duran-Bautista, E.H., Celis-Daza, F., Angel-Sanchez, Y.K. Agroecological practices shape soil ecosystem service provision in cacao agroforestry of the Colombian Amazon. Geoderma Regional (under review).
Files
Steps to reproduce
1. Download all files into one folder (five CSV files and eight Python scripts). 2. Install Python 3.10 or later with numpy 2.2, pandas 2.3, scipy 1.15, statsmodels 0.14, scikit-learn 1.7, prince 0.19.0 and matplotlib (e.g. pip install numpy pandas scipy statsmodels scikit-learn prince==0.19.0 matplotlib). 3. Open a terminal in that folder and run: python typology.py (farm typology, Table S4 and the values of Fig. S1) python analysis.py (service indices, Tables S3 and S5-S9, Spearman correlations and co-inertia; about 3 min; set PERM=50 for a quick run) python fig2_mca_biplot.py python fig3_profiles.py python fig4_services_panel.py python fig5_heatmap.py python figS1_cluster_selection.py
Institutions
- Universidad de la AmazoniaCaquetá, Florencia