Desempeño de algoritmos de aprendizaje automático para mapear variabilidad de materia orgánica de suelo a escala de lote

Autores/as

DOI:

https://doi.org/10.24215/15146774e105

Palabras clave:

interpolación espacial, bosques de regresión, redes neuronales, agricultura de precisión, incertidumbre de la predicción

Resumen

Los mapas de variabilidad de materia orgánica del suelo (MO), a escala de lote, usualmente son generados mediante modelos geoestadísticos que permiten la predicción sitio-específica y la obtención de medidas de incertidumbre. Alternativamente, la predicción espacial puede realizarse mediante aprendizaje automático. Los bosques de regresión cuantílica (QRF) pueden proveer predicciones espaciales con su incertidumbre. Las redes neuronales artificiales (ANN), de potencial uso para predicción desde múltiples covariables, presentan desafíos para medir incertidumbre en bases de datos pequeñas, como las de muestreos edáficos. En este trabajo se comparan las capacidades predictivas de estos algoritmos aplicados al mapeo de MO intralote. Se evaluaron QRF y ANN utilizando como covariables de sitio datos obtenidos por sensores proximales. Como método de referencia se empleó el modelo geoestadístico regresión Kriging (RK). Las predicciones espaciales se realizaron en siete lotes agrícolas ubicados en las provincias de Córdoba y Santiago del Estero, Argentina. Los ajustes se realizaron teniendo en cuenta la información de todos los lotes (modelo global) y de manera individual para cada lote (modelo local). En todos los casos, ANN presentó el mejor desempeño, mostrando los menores valores de error de predicción sitio-específica, aunque con incertidumbre de la predicción mayor a la obtenida mediante QRF. La cantidad de observaciones por lote, así como la variabilidad de la MO y las covariables, influyeron en el desempeño de modelos globales y locales. La ANN resultó la mejor herramienta como soporte para la toma de decisiones de manejo a escala de lote en agricultura de precisión. 

Referencias

Adamchuk, V. I., Hummel, J. W., Morgan, M. T. y Upadhyaya, S. K. (2004). On-the-go sensors for precision agriculture. Computers and Electronics in Agriculture, 44(1), 71–91. https://doi.org/10.1016/j.compag.2004.03.002

Bashie, A. L., Ayankukwa, A. U., Owojoku, O. D., Nicholas, O. G., Ishoro, A. P. y Christiana, U. A. (2024). Predicting environmental covariates of soil organic matter at sub-regional scale for sustainable agricultural development in Southeast Nigeria. Polish Journal of Environmental Studies, 34(3), 2011–2021. https://doi.org/10.15244/pjoes/186888

Bergstra, J. y Bengio, Y. (2012). Random search for hyper-parameter optimization. Journal of Machine Learning Research, 13(1), 281–305.

Bhat, S.A. y Huang, N. (2021). Big Data and AI revolution in precision agriculture: Survey and challenges. IEEE Access, 9, 110209–110222. https://doi.org/10.1109/ACCESS.2021.3102227

Candel, A. y LeDell, E. (2024). Deep learning with H2O. (6ª ed.). H2O ai.

Cannon, A. (2010). Quantile regression neural networks: Implementation in R and application to precipitation downscaling. Computers & Geoscience, 37(9), 1277–1284. https://doi.org/10.1016/j.cageo.2010.07.005

Córdoba, M., Paccioretti P. A., Giannini Kurina, F., Bruno, C. I. y Balzarini, M. G. (2020). Guía para el análisis de datos espaciales: aplicaciones en agricultura. (1ª ed.). Editorial Brujas. https://ri.conicet.gov.ar/handle/11336/195087

Córdoba, M. y Balzarini, M. (2020). Mapeo de materia orgánica del suelo a escala de campo [ponencia]. Congreso Argentino de Agroinformática-Jornadas Argentinas de Informática e Investigación Operativa. Ciudad Autónoma de Buenos Aires, Argentina. https://sedici.unlp.edu.ar/handle/10915/115515

Córdoba, M. y Balzarini, M. (2021). A random forest-based algorithm for data-intensive spatial interpolation in crop yield mapping. Computers and Electronics in Agriculture, 184, 106094. https://doi.org/10.1016/j.compag.2021.106094

Córdoba, M., Hang, S., Bozzer, C., Álvarez, C., Faule, L., Kowaljow, E., Vaieretti, M., Bongiovanni, M. y Balzarini, M. (2025). Spatial Variability and Temporal Changes of Soil Properties Assessed by Machine Learning in Córdoba, Argentina. Soil Systems, 9(4), 109. https://doi.org/10.3390/soilsystems9040109

Dash, P., Ferhatoglu, C. y Miller, B. (2025). Influence of sample size and machine learning algorithms on digital soil nutrient mapping accuracy. Environment Monitoring and Assessment, 197, 996. https://doi.org/10.1007/s10661-025-14322-w

Efron, B. (1979). Bootstrap Methods: Another Look at the Jackknife. The Annals of Statistics, 7(1), 1–26. https://doi.org/10.1214/aos/1176344552

Fryda, T., LeDell, E., Gill, N., Aiello, S., Fu, A., Candel, A., Click, C., Kraljevic, T., Nykodym, T., Ahoyoum, P., Kurka, M., Malohlava, M., Poirier, S., Wong, W., Rehak, L., Eckstrand, E., Hill, B., Vidrio, S., Jadhawani, S., … Novotny, M. (2024). H2o: R Interface for the ‘H2O’ Scalable Machine Learning Platform. CRAN.R. https://doi.org/10.32614/CRAN.package.h2o

García Tomillo, A., Mirás Avalos, J. M., Dafonte Dafonte, J. y Paz González, A. (2017). Estimating soil organic matter using interpolation methods with a electromagnetic induction sensor and topographic parameters: a case study in a humid region. Precision Agriculture, 18, 882–897. https://doi.org/10.1007/s11119-016-9481-6

Goodfellow, I., Bengio, Y. y Courville, A. (2016). Deep learning. MIT Press. https://mitpress.mit.edu/9780262035613/deep-learning/

Gräler, B., Pebesma, E. y Heuvelink, G. (2016). Spatio-Temporal Interpolation using gstat. The R Journal, 8, 204-218. https://doi.org/10.32614/RJ-2016-014

Guo, P., Wu, W., Sheng, Q., Li, M., Liu, H. y Wang, Z. (2013). Prediction of soil organic matter using artificial neural network and topographic indicators in hilly areas. Nutrient Cycling in Agroecosystems, 95, 333–344. https://doi.org/10.1007/s10705-013-9566-9

H2O.ai. (2024). Deep learning (Neural Networks). https://docs.h2o.ai/h2o/latest-stable/h2o-docs/data-science/deep-learning.html

Haynes, K., Lagerquist, R., McGraw, M., Musgrave, K. y Ebert-Uphoff, I. (2023). Creating and evaluating uncertainty estimates with neural networks for environmental-science applications. Artificial Intelligence for the Earth Systems, 2(2), 220061. https://doi.org/10.1175/AIES-D-22-0061.1

Heil, J., Jörges, C. y Stumpe, B. (2022). Fine-scale mapping of soil organic matter in agricultural soils using UAVs and Machine Learning. Remote Sensing, 14(14), 3349. https://doi.org/10.3390/rs14143349

Hengl, T., Nussbaum, M., Wright, M., Heuvelink, G.y Gräler, B. (2018). Random forest as a generic framework for predictive model for spatial and spatio-temporal variables. PeerJ, 6, e5518. https://doi.org/10.7717/peerj.5518

Hiemstra, P., Pebesma, E., Twenhofel, C. y Heuvelink, G. (2008). Real-time automatic interpolation of ambient gamma dose rates from the Dutch Radioactivity Monitoring Network. Computers & Geosciences, 35(8), 1711–1721. https://doi.org/10.1016/j.cageo.2008.10.011

Honorato Fernandes, M. M., Prates Coelho, A., Fernandes, C., da Silva, M. F. y Campos Dela Marta, C. (2019). Estimation of soil organic matter content by modelling with artificial neural networks. Geoderma, 350, 46–51. https://doi.org/10.1016/j.geoderma.2019.04.044

Jurafsky, D. y Martin, J. H. (2024). Speech and language processing: An introduction to natural language processing, computational linguistics, and speech recognition with language models [PDF]. (3ª ed.). Inédito.

https://web.stanford.edu/~jurafsky/slp3/ed3book_Jan25.pdf

Kakhani, N., Alamdar, S., Kebonye, N. M., Amani, M. y Scholten, T. (2024). Uncertainty quantification of soil organic carbon estimation from remote sensing data with conformal prediction. Remote Sensing, 16(3), 438. https://doi.org/10.3390/rs16030438

Kamilaris, A. y Prenafeta-Baldú, F. X. (2018). Deep learning in agriculture: A survey. Computers and Electronics in Agriculture, 147, 70–90. https://doi.org/10.1016/j.compag.2018.02.016

Kuang, B. y Mouazen, A. (2012). Influence of the number of samples on prediction error of visible and near infrared spectroscopy of selected soil properties at the farm scale. European Journal of Soil Sciences, 63, 421–429. https://doi.org/10.1111/j.1365-2389.2012.01456.x

Kuhn, M. (2008). Building predictive models in R using the caret package. Journal of Statistical Software, 28(5), 1–26. https://doi.org/10.18637/jss.v028.i05

Kvamme, K. L. (1993). Spatial statistics and GIS: An integrated approach. En J. Andresen, T. Madsen e I. Scollar (Eds.), Computing the past: Computer applications and quantitative methods in archaeology (pp. 91-103). Aarhus University Press.

Kweon, G., Lund, E. y Maxton, C. (2013). Soil organic matter and cation-exchange capacity sensing with on-the-go electrical conductivity and optical sensors. Geoderma, 199, 80–89. https://doi.org/10.1016/j.geoderma.2012.11.001

Lantz, B. (2015). Machine learning with R. (2ª ed.). Packt Publishing Ltd. https://faculty.cengage.com/works/9780357387528

LeCun, Y., Bengio, Y. y Hinton, G. (2015). Deep learning. Nature, 521, 436–444. https://doi.org/10.1038/nature14539

Liu, Q., He, L., Guo, L., Wang, M., Deng, D., Lv, P., Wang, R., Jia, Z., Hu, Z., Wu, G. y Shi, T. (2022). Digital mapping of soil organic carbon density using newly developed bare soil spectral indices and deep neural network. CATENA, 219, 106603. https://doi.org/10.1016/j.catena.2022.106603

Mohammadpour, M., Roshan, H., Arashpour, M. y Masoumi, H. (2024). Machine learning assisted Kriging to capture spatial variability in petrophysical property modelling. Marine and Petroleum Geology, 167, 106967. https://doi.org/10.1016/j.marpetgeo.2024.106967

Meinshausen, N. y Ridgeway, G. (2006). Quantile Regression Forests. Journal of Machine Learning Research, 7(6), 983–999.

Odeh, I. O. A., McBratney, A. B. y Chittleborough, D. J. (1995). Further results on prediction of soil properties from terrain attributes: Heterotopic cokriging and regression-kriging. Geoderma, 67(3-4), 215–226. https://doi.org/10.1016/0016-7061(95)00007-B

Oldfield, E. E., Bradford, M. A., Augarten, A. J., Cooley, E. T., Radatz, A. M., Radatz, T. y Ruark, M. D. (2022). Positive associations of soil organic matter and crop yields across a regional network of working farms. Soil Science Society of America Journal, 86(2), 384–397. https://doi.org/10.1002/saj2.20349

Oliver, M. A. (2010). Geostatistical applications for precision agriculture. (1ª ed.). Springer. https://link.springer.com/book/10.1007/978-90-481-9133-8

Oliver, M. A. y Webster, R. (2015). Basic steps in Geostatistics: The variogram and kriging. (1ª ed.). Springer. https://link.springer.com/book/10.1007/978-3-319-15865-5

Omonode, R. A. y Vyn, T. J. (2006). Spatial dependence and relationships of electrical conductivity to soil organic matter, phosphorus, and potassium. Soil Science, 171(3), 223–238. https://doi.org/10.1097/01.ss.0000199698.94203.a4

Paccioretti, P., Córdoba, M., Giannini Kurina, F. y Balzarini, M. (2024). Paar: Precision agriculture data analysis. CRAN.r. https://doi.org/10.32614/CRAN.package.paar

Paranavithana, T. M., Karunaratne, S. B., Wimalathunge, N., Malone, B. P., Macdonald, B., Bishop, T. F. A. y Ratnayake, R. R. (2024). Unravelling spatial drivers of topsoil total carbon variability in tropical paddy soils of Sri Lanka. Geoderma Regional, 26, e00745. https://doi.org/10.1016/j.geodrs.2023.e00745

Pebesma, E. y Bivand, R. (2023). Spatial Data Science: With applications in R. Chapman and Hall/CRC. https://doi.org/10.1201/9780429459016

Piccini, C., Marchetti, A. y Francaviglia, R. (2014). Estimation of soil organic matter by geostatistical methods: Use of auxiliary information in agricultural and environmental assessment. Ecological Indicators, 36, 301–314. https://doi.org/10.1016/j.ecolind.2013.08.009

R Core Team. (2025). R: A Language and Environment for Statistical Computing (Versión 4.5.1) [Software]. R-Foundation. https://www.R-project.org/

Sekulić, A., Kilibarda, M., Heuvelink, G. B. M., Nikolić, M. y Bajat, B. (2020). Random forest spatial interpolation. Remote Sensing, 12, 1687. https://doi.org/10.3390/rs12101687

Schmidlinger, J., Schröter, I., Bönecke, E., Gebbers, R., Ruehlmann, J., Kramer, E., Mulder, V., Heuvelink, G. y Vogel, S. (2024). Effect of training sample size, sampling design and prediction model on soil mapping with proximal sensing data for precision liming. Precision Agriculture, 25, 1529–1555. https://doi.org/10.1007/s11119-024-10122-3

Tagasovska, N. y López-Paz, D. (2018). Frequentist uncertainty estimates for deep learning. ArXiv, 1811.00908.

Vallejos, R., Osorio, F. y Bevilacqua, M. (2020). Spatial Relationships Between Two Georeferenced Variables: with Applications in R. Springer. https://link.springer.com/book/10.1007/978-3-030-56681-4

Vázquez, J. M. (2016). Manejo sitio específico. En J. M. Gueçaimburu, V. Rojo, G.Reposo, J. M. Vázquez y M. Martínez (Eds.), Manejo del sistema agropecuario (1ª ed., pp. 91–103). Editorial Universidad Nacional de Luján (EdUNLu). https://ri.unlu.edu.ar/xmlui/handle/rediunlu/1360

Wang, Y., Shi, L., Hu, Y., Hu, X., Song, W. y Wang, L. (2024). A comprehensive study of Deep learning for soil moisture prediction. Hydrology and Earth System Sciences, 28(4), 917–943. https://doi.org/10.5194/hess-28-917-2024

Xie, J., Huang, J., Zheng, C., Huang, S. y Burton, G. (2022). A generic framework for geotechnical subsurface modeling with machine learning. Journal of Rock Mechanics and Geotechnical Engineering, 14(5), 1366–1379. https://doi.org/10.1016/j.jrmge.2022.08.001

Ying, X. (2019). An overview of overfitting and its solutions. Journal of Physics: Conference Series, 1168, 022022. https://doi.org/10.1088/1742-6596/1168/2/022022

Zhao, Y. y Shi, X. (2010). Spatial prediction and uncertainty assessment of soil organic carbon in Hebei Province, China. En J. Boettinger, D. Howell, A. Moore, A. Hartemink y S. Kienast-Brown (Eds.), Digital soil mapping. Springer. https://doi.org/10.1007/978-90-481-8863-5_19

Zhang, W., Quan, H. y Srinivasan, D. (2019). An improved quantile regression neural network for probabilistic load forecasting. IEEE Transactions on Smart Grid, 10(4), 4425–4434. https://doi.org/10.1109/TSG.2018.2859749

Zhang, Y., Luo, C., Zhang, W., Wu, Z. y Zang, D. (2025). Mapping soil organic matter in black soil cropland areas using remote sensing and environmental covariates. Agriculture, 15(3), 339. https://doi.org/10.3390/agriculture15030339

Descargas

Publicado

2026-05-27

Cómo citar

García Seleme, F. A., Paccioretti, P., Balzarini, M., & Córdoba, M. (2026). Desempeño de algoritmos de aprendizaje automático para mapear variabilidad de materia orgánica de suelo a escala de lote. SADIO Electronic Journal of Informatics and Operations Research, 25(2), e105. https://doi.org/10.24215/15146774e105