Machine Learning Models for Predicting Soil Organic Matter Dynamics in Regenerative Agroecosystems

Advances in Agriculture and Agricultural Sciences | Vol. 12, No. 3, March 2026 | pp. 55–60

DOI: 10.46882/2026.AAAS.120301

Research Article

Title: Machine Learning Models for Predicting Soil Organic Matter Dynamics in Regenerative Agroecosystems

Names of Authors: O'Connor, L. M.¹, Gallagher, P. S.¹, and Kennedy, T. R.²

Authors’ Affiliations: ¹School of Agriculture and Food Science, University College Dublin, Dublin, Ireland. ²Teagasc Environment Research Centre, Johnstown Castle, Wexford, Ireland.

Abstract: Accurately mapping soil organic matter (SOM) fractions is necessary to validate carbon tracking within regenerative agricultural frameworks. This study evaluated the performance of three machine learning algorithms—Random Forest (RF), Support Vector Machines (SVM), and Gradient Boosting Machines (GBM)—to predict topsoil SOM dynamics using multi-temporal satellite imagery and topographic indices. Field samples were collected from 120 temperate pasture plots showing diverse management histories, including multi-species swards and rotational adaptive grazing. The RF model out-performed alternative configurations, yielding a coefficient of determination (R²) of 0.84 and a Root Mean Square Error (RMSE) of 0.32%. Important predictive parameters included the Normalized Difference Vegetation Index (NDVI) extracted during early spring growth alongside the topographic wetness index. Model validations verified that converting traditional monoculture ryegrass pastures into multi-species swards increased SOM accumulation by an average of 1.24 ± 0.15% over a five-year monitoring cycle. This automated digital pipeline lowers verification costs for carbon credit markets while maintaining regional accuracy metrics.

Keywords: Predictive agrotechnology; Random forest; Soil carbon verification; Multi-species swards; Digital soil mapping; Remote sensing.

Manuscript Timeline: Received November 15, 2025; Revised January 08, 2026; Accepted February 10, 2026; Published March 12, 2026.

Citation: O'Connor, L. M., Gallagher, P. S., and Kennedy, T. R. (2026). Machine Learning Models for Predicting Soil Organic Matter Dynamics in Regenerative Agroecosystems. Advances in Agriculture and Agricultural Sciences, 12(3), pp. 55–60.

Subscribe to read the full article: https://internationalscholarsjournals.org/subscribe-to-read