Can a history of crop rotations improve the prediction of soil organic carbon in the Andes? Integrating machine learning multi-annual crop classification as a proxy of soil management

Soil organic carbon (SOC) is a crucial component related to various processes that ensure soil health and function. Its modeling is vital for assessing and monitoring soil degradation caused by the potential impact of agricultural activities. This study aimed to model SOC in the Northern highlands o...

Descripción completa

Detalles Bibliográficos
Autores principales: Bueno, M., Loayza, H., Ninanya, J., Rinza, J., Briceño, P., Silva, L., Mestanza, C., Otiniano, R., Kreuze, Jan F., Ramirez, D.
Formato: Preprint
Lenguaje:Inglés
Publicado: 2025
Materias:
Acceso en línea:https://hdl.handle.net/10568/178773
_version_ 1855540600252661760
author Bueno, M.
Loayza, H.
Ninanya, J.
Rinza, J.
Briceño, P.
Silva, L.
Mestanza, C.
Otiniano, R.
Kreuze, Jan F.
Ramirez, D.
author_browse Briceño, P.
Bueno, M.
Kreuze, Jan F.
Loayza, H.
Mestanza, C.
Ninanya, J.
Otiniano, R.
Ramirez, D.
Rinza, J.
Silva, L.
author_facet Bueno, M.
Loayza, H.
Ninanya, J.
Rinza, J.
Briceño, P.
Silva, L.
Mestanza, C.
Otiniano, R.
Kreuze, Jan F.
Ramirez, D.
author_sort Bueno, M.
collection Repository of Agricultural Research Outputs (CGSpace)
description Soil organic carbon (SOC) is a crucial component related to various processes that ensure soil health and function. Its modeling is vital for assessing and monitoring soil degradation caused by the potential impact of agricultural activities. This study aimed to model SOC in the Northern highlands of Peru, characterized by a high amount of SOC, which is being affected by crop expansion. Crop rotation (CR) was incorporated into a modeling exercise using remote sensing data, fieldwork, and farmer surveys. A multi-year classification model with seven cropland classes was developed using data collected from 534 fields across 2022-2024, including 189 soil samples. Each cropland field was represented as a polygon delineating its boundaries and indicating its dominant crop cover. Time series of multispectral Sentinel-2 Level-2A Top of Canopy imagery were used to derive phenological features—such as the timing of maximum canopy cover and the length of the growing period—based on Normalized Difference Vegetation Index (NDVI) time series. A Random Forest classifier was used as the baseline model. The cropland classification model demonstrated strong overall performance, with F1 scores ranging from 0.81 to 0.98 across the different classes. The model performed well for lupin and pasture but scored lower for beans and potatoes. Predictions of cropland classes from 2019 to 2022 were created, resulting in frequency layers that represent crop rotations. Four feature configurations were evaluated: (i) including all features as a benchmark, (ii) excluding climatology, (iii) excluding crop rotation history, and (iv) excluding soil properties. Configurations including all features and excluding crop rotation history showed the highest performance (R2 = 0.63), while those excluding climatology or soil properties performed worse (R2 ≈ 0.52–0.53). Although soil features were the most important, fallow frequency emerged as the most critical predictor of SOC in crop rotations. When soil data were excluded, fallow frequency, combined with climatic features, explained over half of the SOC variability. The findings emphasize the importance of incorporating CR into SOC mapping efforts.
format Preprint
id CGSpace178773
institution CGIAR Consortium
language Inglés
publishDate 2025
publishDateRange 2025
publishDateSort 2025
record_format dspace
spelling CGSpace1787732025-12-11T22:46:40Z Can a history of crop rotations improve the prediction of soil organic carbon in the Andes? Integrating machine learning multi-annual crop classification as a proxy of soil management Bueno, M. Loayza, H. Ninanya, J. Rinza, J. Briceño, P. Silva, L. Mestanza, C. Otiniano, R. Kreuze, Jan F. Ramirez, D. potatoes carbon sequestration soil organic matter climate-smart agriculture Soil organic carbon (SOC) is a crucial component related to various processes that ensure soil health and function. Its modeling is vital for assessing and monitoring soil degradation caused by the potential impact of agricultural activities. This study aimed to model SOC in the Northern highlands of Peru, characterized by a high amount of SOC, which is being affected by crop expansion. Crop rotation (CR) was incorporated into a modeling exercise using remote sensing data, fieldwork, and farmer surveys. A multi-year classification model with seven cropland classes was developed using data collected from 534 fields across 2022-2024, including 189 soil samples. Each cropland field was represented as a polygon delineating its boundaries and indicating its dominant crop cover. Time series of multispectral Sentinel-2 Level-2A Top of Canopy imagery were used to derive phenological features—such as the timing of maximum canopy cover and the length of the growing period—based on Normalized Difference Vegetation Index (NDVI) time series. A Random Forest classifier was used as the baseline model. The cropland classification model demonstrated strong overall performance, with F1 scores ranging from 0.81 to 0.98 across the different classes. The model performed well for lupin and pasture but scored lower for beans and potatoes. Predictions of cropland classes from 2019 to 2022 were created, resulting in frequency layers that represent crop rotations. Four feature configurations were evaluated: (i) including all features as a benchmark, (ii) excluding climatology, (iii) excluding crop rotation history, and (iv) excluding soil properties. Configurations including all features and excluding crop rotation history showed the highest performance (R2 = 0.63), while those excluding climatology or soil properties performed worse (R2 ≈ 0.52–0.53). Although soil features were the most important, fallow frequency emerged as the most critical predictor of SOC in crop rotations. When soil data were excluded, fallow frequency, combined with climatic features, explained over half of the SOC variability. The findings emphasize the importance of incorporating CR into SOC mapping efforts. 2025-12 2025-12-11T22:37:52Z 2025-12-11T22:37:52Z Preprint https://hdl.handle.net/10568/178773 en Open Access Bueno, M., Loayza, H., Ninanya, J., Rinza, J., Briceño, P., Silva, L., Mestanza, C., Otiniano, R., Kreuze, J., & Ramírez, D. A. (2025). Can a history of crop rotations improve the prediction of soil organic carbon in the Andes? Integrating machine learning multi-annual crop classification as a proxy of soil management (Version posted December 10, 2025). bioRxiv. https://doi.org/10.64898/2025.12.06.692759
spellingShingle potatoes
carbon sequestration
soil organic matter
climate-smart agriculture
Bueno, M.
Loayza, H.
Ninanya, J.
Rinza, J.
Briceño, P.
Silva, L.
Mestanza, C.
Otiniano, R.
Kreuze, Jan F.
Ramirez, D.
Can a history of crop rotations improve the prediction of soil organic carbon in the Andes? Integrating machine learning multi-annual crop classification as a proxy of soil management
title Can a history of crop rotations improve the prediction of soil organic carbon in the Andes? Integrating machine learning multi-annual crop classification as a proxy of soil management
title_full Can a history of crop rotations improve the prediction of soil organic carbon in the Andes? Integrating machine learning multi-annual crop classification as a proxy of soil management
title_fullStr Can a history of crop rotations improve the prediction of soil organic carbon in the Andes? Integrating machine learning multi-annual crop classification as a proxy of soil management
title_full_unstemmed Can a history of crop rotations improve the prediction of soil organic carbon in the Andes? Integrating machine learning multi-annual crop classification as a proxy of soil management
title_short Can a history of crop rotations improve the prediction of soil organic carbon in the Andes? Integrating machine learning multi-annual crop classification as a proxy of soil management
title_sort can a history of crop rotations improve the prediction of soil organic carbon in the andes integrating machine learning multi annual crop classification as a proxy of soil management
topic potatoes
carbon sequestration
soil organic matter
climate-smart agriculture
url https://hdl.handle.net/10568/178773
work_keys_str_mv AT buenom canahistoryofcroprotationsimprovethepredictionofsoilorganiccarbonintheandesintegratingmachinelearningmultiannualcropclassificationasaproxyofsoilmanagement
AT loayzah canahistoryofcroprotationsimprovethepredictionofsoilorganiccarbonintheandesintegratingmachinelearningmultiannualcropclassificationasaproxyofsoilmanagement
AT ninanyaj canahistoryofcroprotationsimprovethepredictionofsoilorganiccarbonintheandesintegratingmachinelearningmultiannualcropclassificationasaproxyofsoilmanagement
AT rinzaj canahistoryofcroprotationsimprovethepredictionofsoilorganiccarbonintheandesintegratingmachinelearningmultiannualcropclassificationasaproxyofsoilmanagement
AT bricenop canahistoryofcroprotationsimprovethepredictionofsoilorganiccarbonintheandesintegratingmachinelearningmultiannualcropclassificationasaproxyofsoilmanagement
AT silval canahistoryofcroprotationsimprovethepredictionofsoilorganiccarbonintheandesintegratingmachinelearningmultiannualcropclassificationasaproxyofsoilmanagement
AT mestanzac canahistoryofcroprotationsimprovethepredictionofsoilorganiccarbonintheandesintegratingmachinelearningmultiannualcropclassificationasaproxyofsoilmanagement
AT otinianor canahistoryofcroprotationsimprovethepredictionofsoilorganiccarbonintheandesintegratingmachinelearningmultiannualcropclassificationasaproxyofsoilmanagement
AT kreuzejanf canahistoryofcroprotationsimprovethepredictionofsoilorganiccarbonintheandesintegratingmachinelearningmultiannualcropclassificationasaproxyofsoilmanagement
AT ramirezd canahistoryofcroprotationsimprovethepredictionofsoilorganiccarbonintheandesintegratingmachinelearningmultiannualcropclassificationasaproxyofsoilmanagement