Correo Electrónico
DNINFOA - SIA
Bibliotecas
Convocatorias
Identidad U.N.
Published
Performance of Random Forest in predicting soil loss based on values calculated by USLE
Desempeño del algoritmo Random Forest en la predicción de la pérdida de suelo basada en valores calculados por la USLE
DOI:
https://doi.org/10.15446/esrj.v29n4.121271Keywords:
NDVI, Machine Learning, Topographic factor (LS), Erosion processes (en)NDVI, Aprendizaje automático, Factor topográfico (LS), Procesos de erosión (es)
Downloads
Soil erosion directly affects agricultural productivity and water resource quality, but estimating soil loss is complex and costly. This study proposes a machine learning (ML) approach to predict soil loss using selected factors from the Universal Soil Loss Equation (USLE) and the Normalized Difference Vegetation Index (NDVI). We applied the Random Forest (RF) algorithm to train and validate two models using different combinations of predictors: (1) NDVI, topographic factor (LS), and land cover/management factor (CP); and (2) NDVI, LS, and soil erodibility factor (K). These variables represent land use, conservation practices, and topographic conditions in the Sorocabuçu River Basin (SRB), part of Brazil’s Atlantic Forest biome with high environmental and socioeconomic value. Soil loss was classified into three classes (in ton/ha): low (0–10.0), moderate (10.1–50.0), and high (≥50.1). A total of 3348 samples were randomly selected and proportionally distributed to reflect class representation across the study area. We used a 70/30 train-test split and standardized parameters (50 trees and four variables per node) to enable reproducibility. The model using NDVI, LS, and CP achieved 93.43% accuracy with a kappa index of 0.90. The performance was especially strong for the low-loss class, the most prevalent in the area. The second model using NDVI, LS, and K achieved 97.14% accuracy with a kappa index of 0.90, showing excellent results, particularly for the high-loss class, which poses the greatest environmental risk. These models prove effective in identifying areas at risk of severe erosion using fewer, more accessible parameters. The approach offers a scalable and practical tool for decision-makers, environmental managers, and public agencies to monitor and mitigate soil degradation, particularly in sensitive and ecologically important regions.
La erosión del suelo afecta directamente la productividad agrícola y la calidad de los recursos hídricos; sin embargo, la estimación de la pérdida de suelo es un proceso complejo y costoso. Este estudio propone un enfoque de aprendizaje automático (Machine Learning (ML)) para predecir la pérdida de suelo utilizando factores seleccionados de la Ecuación Universal de Pérdida de Suelo (USLE) y el Índice de Vegetación de Diferencia Normalizada (NDVI). Se aplicó el algoritmo Random Forest (RF) para entrenar y validar dos modelos con diferentes combinaciones de variables predictoras: (1) NDVI, factor topográfico (LS) y factor de cobertura y manejo del suelo (CP); y (2) NDVI, LS y factor de erodabilidad del suelo (K). Estas variables representan el uso del suelo, las prácticas de conservación y las condiciones topográficas en la cuenca del río Sorocabuçu (SRB), ubicada en el bioma de la Mata Atlántica de Brasil, una región de alto valor ambiental y socioeconómico. La pérdida de suelo se clasificó en tres categorías (en t/ha): baja (0–10,0), moderada (10,1–50,0) y alta (≥50,1). Se seleccionaron aleatoriamente un total de 3348 muestras, distribuidas proporcionalmente para reflejar la representatividad de las clases en el área de estudio. Se utilizó una división de los datos del 70% para entrenamiento y 30% para validación, junto con parámetros estandarizados (50 árboles y cuatro variables por nodo) para garantizar la reproducibilidad del análisis. El modelo basado en NDVI, LS y CP alcanzó una precisión del 93,43% y un índice kappa de 0,90, con un desempeño destacado en la clase de baja pérdida de suelo, la más frecuente en el área. El segundo modelo, que utilizó NDVI, LS y K, obtuvo una precisión del 97,14% y un índice kappa de 0,90, mostrando resultados excelentes, especialmente en la clase de alta pérdida de suelo, que representa el mayor riesgo ambiental. Los resultados demuestran que ambos modelos son eficaces para identificar áreas con riesgo de erosión severa utilizando un conjunto reducido de parámetros más accesibles. Este enfoque constituye una herramienta práctica y escalable para la toma de decisiones por parte de gestores ambientales y organismos públicos, contribuyendo al monitoreo y la mitigación de la degradación del suelo, particularmente en regiones sensibles y de gran importancia ecológica.
References
Arantes, L. T., Santos, A. P., Silva, C. V., Nery, L. M., Toledo, M. V. L., Simonetti, V. C., Silva, D. C. C., & Lourenço, R. W. (2024a). Socioeconomic spatial analysis through fuzzy system as a tool for territorial planning applied to watersheds. International Journal of River Basin Management, 1–17. https://doi.org/10.1080/15715124.2024.2387579
Arantes, L. T., Santos, A. P., Silva, D. C. C., & Lourenço, R. W. (2024b). Indicador de vulnerabilidade ao carreamento de sedimentos integrado ao SIG e SR. Geo UERJ, (45). https://doi.org/10.12957/geouerj.2024.74164.
Amundson, R., Berhe, A. A., Hopmans, J. W., Olson, C., Sztein, A. E., & Sparks, D. L. (2015). Soil and human security in the 21st century. Science, 348(6235), 1261071. https://doi.org/10.1126/science.1261071
Andreoti, C. E. (2012). Avaliação da eficiência de um sistema agroflorestal na recuperação de um solo degradado por pastoreio. Dissertação (mestrado – Programa de Pós-graduação em Geografia Física) – Faculdade de Filosofia, Letras e Ciências Humanas – São Paulo, Brasil. https://doi.org/10.11606/D.8.2012.tde-09012013-121619
Bertoni, J., & Lombardi Neto, F. (1999). Conservação do solo. 4. ed. São Paulo, SP: Ícone.
Boardman, J., & Poesen, J. (2006). Soil erosion in Europe. Ed. John Wiley & Sons Ltd. West Sussex. 855 p. https://doi.org/10.1002/0470859202.ch36
Breiman, L. (2001). Random Forests. Journal Machine Learning, 45, 5-32. https://doi.org/10.1023/A:1010933404324
Cheng, Z., Lu, D., Li, G., Huang, J., Sinha, N., Zhi, J., & Li, S. (2018). A Random Forest-Based Approach to Map Soil Erosion Risk Distribution in Hickory Plantations in Western Zhejiang Province, China. Remote Sensing, 10, 1-20. https://doi.org/10.3390/rs10121899
Costa, R. V. F., Leite, M. G. P., Leao, L. P., Nalini Junior, H. A., Silva, D. C. C., & Valente, T. M. F. (2025). Hydrogeochemistry of surface waters in the Iron Quadrangle, Brazil: High-Resolution Mapping of Potentially Toxic Elements in the Velhas and Paraopeba River Basins. Water, 17, 2446. https://doi.org/10.3390/w17162446
Dubreuil, V., Fante, K. P., Planchon, O., & Sant'Anna Neto, J. L. (2017). Les types de climats annuels au Brésil: une application de la classification de Köppen de 1961 a 2015. EchoGéo, 41, 1-27. https://doi.org/10.4000/echogeo.15017
Empresa Brasileira de Pesquisa Agropecuária [EMBRAPA]. (2018). Solo e relevo: Influência no uso da terra. https://www.embrapa.br/agencia-de-informacao-tecnologica/cultivos/eucalipto/pre-producao/escolha-da-area/solo-e-relevo Access: 02 nov. 2025.
FAO – Food and Agriculture Organization of the United Nations. (2023). The State of Food and Agriculture 2023. Revealing the true cost of food to transform agrifood systems. From: https://www.fao.org/documents/card/en/c/cc7724en. Access: 09 ago. 2023.
FAO – Food and Agriculture Organization of the United Nations. (2021). The state of the world land and water resources for food and agriculture 2021. Disponível em: https://www.fao.org/3/cb7654en/online/cb7654en.html. Access: 09 ago. 2023.
Filho, J. P. (2014). Capacidade preditiva de Modelos Credit Scoring em inferência dos rejeitados. Dissertação (Mestrado em Estatística) – Centro de Ciências Exatas e de Tecnologia, Universidade federal de São Carlos, São Carlos, 95p.
Ganasri, B. P., & Ramesh, H. (2016). Assessment of soil erosion by RUSLE model using remote sensing and GIS-A case study of Nethravathi Basin. Geoscience Frontiers, 7(6), 953-961. https://doi.org/10.1016/j.gsf.2015.10.007
Ghimire, B., Rogan, J., & Miller, J. (2010). Contextual land-cover classification: incorporating spatial dependence in land-cover classification models using random forests and the Getis statistic. Remote Sensing Letters, 1(1), 45-54. https://doi.org/10.1080/01431160903252327
Ghosal, K., & Das Bhattacharya, S. (2020). A review of RUSLE model. Journal of the Indian Society of Remote Sensing, 48, 689-707. https://doi.org/10.1007/s12524-019-01097-0
Glaros, A. G., & Kline, R. B. (1998). Understanding the accuracy of tests with cutting scores: The sensitivity, specificity, and predictive value model. Journal of clinical psychology, 44(6), 1013-1023. https://doi.org/10.1002/1097-4679(198811)44:6%3C1013::aid jclp2270440627%3E3.0.co;2-z
He, Q., Zhao, H., Feng, Y., Wang, Z., Ning, Z., & Luo, T. (2024). Edge computing-oriented smart agricultural supply chain mechanism with auction and fuzzy neural networks. Journal of Cloud Computing: Advances, Systems and Applications, 13(1). https://doi.org/10.1186/s13677-024-00626-8
Helmi, A. M. (2023). Quantifying catchments sediment release in arid regions using GIS-based Universal soil loss equation (USLE). Ain Shams Engineering Journal, 14(8), 102038. https://doi.org/10.1016/j.asej.2022.102038
IBGE – Instituto Brasileiro de Geografia e Estatística. (2021). Banco de Dados de Informações Ambientais. From: https://bdiaweb.ibge.gov.br/. Acess 09 jun. 2023.
Kashiwar, S. R., Kundu, M. C., & Dongarwar, U. R. (2022). Soil erosion estimation of Bhandara region of Maharashtra, India, by integrated use of RUSLE, remote sensing, and GIS. Natural Hazards, 110(2), 937–959. https://doi.org/10.1007/s11069-021-04974-5
Köpen, W. (1948). Climatologia. Buenos Aires: Gráfica Panamericana. 478p.
Kulimushi, L. C., Choudhari, P., Mubalama, L. K., & Banswe, G. T. (2021). GIS and remote sensing-based assessment of soil erosion risk using RUSLE model in South-Kivu province, eastern, Democratic Republic of Congo. Geomatics Natural Hazards and Risk, 12(1), 961–987. https://doi.org/10.1080/19475705.2021.1906759
Kulimushi, L. C., Bashagaluke, J. B., Prasad, P., Heri-Kazi, A. B., Kushwaha, N. L., Masroor, M., Choudhari, P., Elbeltagi, A., Sajjad, H. & Mohammed, S. (2023). Soil erosion susceptibility mapping using ensemble machine learning models: A case study of upper Congo river sub-basin. Catena, 222, 106858. https://doi.org/10.1016/j.catena.2022.106858
Lepsch, I. F. (2010). Formação e conservação dos solos. 2. ed. São Paulo, SP: Oficina de Textos.
Maia Júnior, L. P. M. & Lourenço, R. W. (2020). Impactos das mudanças no uso e cobertura da terra sobre a variabilidade do albedo na Bacia Hidrográfica do Rio Sorocabuçu (Ibiúna-SP). Revista Brasileira de Climatologia, 27, 443-462. https://doi.org/10.5380/abclima.v27i0.72761
Nery, L. M., Sabonaro, D. Z. & Silva, D. C. C. (2023). A multicriteria analysis for decision making. Environment, Development and Sustainability, 25, 1-19. https://doi.org/10.1007/s10668-023-03261-6
Nery, L. M., Gomes, G., Nicomedes, N. P., Sabonaro, D. Z., & Silva, D. C. C. (2024). Análise socioambiental de sistemas de integração: quais seus benefícios, desafios e oportunidades? RISUS. Journal on Innovation and Sustainability, 15, 177–192. https://doi.org/10.23925/2179-3565.2024v15i2p177-192
Nery, L. M., Toniolo, B. P., Santos, A. P., Martins, A. C. G. & Silva, D. C. C (2025). Challenge of political integration in the territorial management of a protected area based on the analysis of land use and land cover change. Journal of Environmental Studies and Sciences, 15, 845–860. https://doi.org/10.1007/s13412-024-00990-6
Nguyen, K. A., Chen, W., Lin, B. S., Seeboonruang, U. & Thomas, K. (2019). Predicting Sheet and Rill Erosion of Shihmen Reservoir Watershed in Taiwan Using Machine Learning. Sustainability, 11, 1-18. https://doi.org/10.3390/su11133615
Noi, P. T. & Kappas, M. (2017). Comparison of random forest, k-nearest neighbor, and support vector machine classifiers for land cover classification using Sentinel-2 imagery. Sensors, 18(1), 18. https://doi.org/10.3390/s18010018
Pacheco, F. A. L., Fernandes, L. F. S., Valle Júnior, R. F., Valera, C. A. & Pissarra, T. C. T. (2018). Land degradation: Multiple environmental consequences and routes to neutrality. Current Opinion in Environmental Science & Health, 5, 78-86. https://doi.org/10.1016/j.coesh.2018.07.002
Panagos, P., Borrelli, P., Poesen, J., Ballabio, C., Lugato, E., Meusburger, K., Montanarella, L. & Alewell, C. (2015). The new assessment of soil loss by water erosion in Europe. Environmental Science & Policy, (54), 438-447. https://doi.org/10.1016/j.envsci.2015.08.012
Pandey, A., Chowdary, V. M. & Mal, B. C. (2007). Identification of critical erosion prone areas in the small agricultural watershed using USLE, GIS and remote sensing. Water Resources Management, 21(4), 729-746. https://doi.org/10.1007/s11269-006-9061-z
Paula, A. L., Pereira dos Santos, A., Belfort Poletti, F., & Lourenço, R. W. (2025). Adjustment of the conservation practices factor calculation in estimating soil loss. Ra’e Ga: O Espaço Geográfico em Análise, 63(1), 125–151. https://doi.org/10.5380/raega.v63i1.100335
QGIS. (2023). QGIS Geographic Information System. QGIS Association. From: http://www.qgis.org. Acess: 09 jun. 2023.
Rizzo, F. A., Santos, A. P., & Silva, D. C. C. (2024). Técnicas de geoprocessamento aplicadas para análise temporal do microclima na bacia hidrográfica do córrego do Pequiá, Maranhão. Boletim Goiano de Geografia, 44, e78032. https://doi.org/10.5216/bgg.v44i1.78032
Rossi, M. (2017). Mapa pedológico do Estado de São Paulo: Revisado e ampliado. Instituto Florestal. https://www.infraestruturameioambiente.sp.gov.br/institutoflorestal.
Rouse, J. J. R., Haas, R. H., Schell, J. A. & Deering, D. W. (1973). Monitoring the vernal advancement and retrogradation (green wave effect) of natural vegetation. Remote Sensing Center Texas A&M University College Station, Texas. 93p. From: https://core.ac.uk/download/pdf/42887948.pdf. Acess: 09 jun. 2023.
RStudio Team (2023). RStudio: Integrated Development Environment for R. RStudio, PBC, Boston. From: http://www.rstudio.com/. 09 jun. 2023.
Santos, A. P., Silva Junior, A. X., Nery, L. M., Gomes, G., Toniolo, B. P., da Cunha e Silva, D. C., & Lourenço, R. W. (2025). Random forest algorithm applied to model soil textural classification in a river basin. Environmental Monitoring and Assessment, 197, 330. https://doi.org/10.1007/s10661-025-13786-0.
Sheikh, A. H., Palria, S. & Alam, A. (2011). Integration of GIS and Universal Soil Loss Equation (USLE) for soil loss estimation in a Himalayan watershed. Recent Research in Science and Technology, 3(3), p. 51-57. https://www.researchgate.net/publication/286921198_INTEGRATION_OF_GIS_AND_UNIVERSAL_SOIL_LOSS_EQUATION_USLE_FOR_SOIL_LOSS_ESTIMATION_IN_A_HIMALAYAN_WATERSHED
Silva, D. C. C., Albuquerque Filho, J. L., Sales, J. C. A. & Lourenço, R. W. (2017). Identificação de áreas com perda de solo acima do tolerável usando NDVI para o cálculo do fator C da USLE. Ra'e Ga, 42, 72-85. http://dx.doi.org/10.5380/raega.v42i0.45524
Simonetti, V. C., Silva, D. C. C., & Rosa, A. H. (2022). Correlação espacial compartimentada dos padrões de drenagem com características morfométricas da bacia hidrográfica do rio Pirajibu-Mirim. Revista Brasileira de Geomorfologia, 23, 1134–1154. https://doi.org/10.20502/rbg.v23i1.2037
Tanyas, H., Kolat Ç. & Süzen, M. L. (2015). A new approach to estimate cover-management factor of RUSLE and validation of RUSLE model in the watershed of Kartalkaya Dam. Journal of Hydrology, 528, 583-598. https://doi.org/10.1016/j.jhydrol.2015.06.048
Tarek, Z., Elshewey, A. M., Shohieb, S. M., Elhady, A. M., El-Attar, N. E., Elseouf, S., Shams, M. Y. (2023). Soil Erosion Status Prediction Using a Novel Random Forest Model Optimized by Random Search Method. Sustainability, 15(9), 7114. https://doi.org/10.3390/su15097114
Toniolo, B. P., Nery, L. M., & Silva, D. C. C. (2024). Modelagem espacial para identificação de áreas potenciais à geração de poluição difusa na Bacia Hidrográfica do Rio Cotia - SP. URBE. Revista Brasileira de Gestão Urbana, 16, e20220207. https://doi.org/10.1590/2175-3369.016.e20220207.
UN – United Nations. The 17 goals. (2023). From: https://sdgs.un.org/goals. Acess: 30 nov. 2023.
Van Stralen, K. J., Stel, V. S., Reitsma, J. B., Dekker, F. W., Zoccali, C. & Jager, K. J. (2009). Diagnostic methods I: sensitivity, specificity, and other measures of accuracy. Kidney international, 75(12), 1257-1263. https://doi.org/10.1038/ki.2009.92
Weiss, S. M. & Zhang, T. (2003). Performance analysis and evaluation. In: The handbook of Data Mining. Lawrence Erlbaum Associates Publishers, Mahwah, NJ, 14, 425 – 440.
Wischmeier, W. H. & Smith, D. D. (1978). Predicting rainfall erosion losses – A guide to conservation planning. Washington, USDA, 1978. 58p. (USDA AH-537). file:///C:/Users/simio/Downloads/USLE.pdf
Xiao, Y. G. B., Lu, Y., Zhang, R., Zhang, D., Zhen, X., Chen, S., Wu, H., Wei, C., Yang, L. & Zhang, Y. (2021). Spatial–temporal evolution patterns of soil erosion in the Yellow River Basin from 1990 to 2015: impacts of natural factors and land use change. Geomat Nat Hazard Risk, 12(1), 103–122. https://doi.org/10.1080/19475705.2020.1861112
Yang, D., Kanae, S., Oki, T., Koike, T. & Musiake, K. (2003). Global potential soil erosion with reference to land use and climate changes. Hydrological Processes, 17, 2913-2918. https://doi.org/10.1002/hyp.1441
Yang, T., Siddique, K. H. M. & Liu, K. (2020). Cropping systems in agriculture and their impact on soil health - A review. Global Ecology and Conservation, 23, e01118. https://doi.org/10.1016/j.gecco.2020.e01118
Zhang, H. K. & Roy, D. P. (2017). Using the 500 m MODIS land cover product to derive a consistent continental scale 30 m Landsat land cover classification. Remote Sensing of Environment, 197, 15-34. https://doi.org/10.1016/j.rse.2017.05.024
Zhang, X., Song, J., Wang, Y., Deng, W. & Liu, Y. (2021). Effects of land use on slope runoff and soil loss in the Loess Plateau of China: A meta-analysis. Science of The Total Environment, 755(1), 142418. https://doi.org/10.1016/j.scitotenv.2020.142418
How to Cite
APA
ACM
ACS
ABNT
Chicago
Harvard
IEEE
MLA
Turabian
Vancouver
Download Citation
License
Copyright (c) 2026 Earth Sciences Research Journal
This work is licensed under a Creative Commons Attribution 4.0 International License.
Earth Sciences Research Journal holds a Creative Commons Attribution license.
You are free to:
Share — copy and redistribute the material in any medium or format
Adapt — remix, transform, and build upon the material for any purpose, even commercially.
The licensor cannot revoke these freedoms as long as you follow the license terms.
The Earth Sciences Research Journal is the copyright holder for these license attributes.
