Aplicación de la inteligencia artificial en el monitoreo ambiental: modelos predictivos de elementos potencialmente tóxicos (PTEs) mediante machine learning

Contenido principal del artículo

Teresa Salazar-Rojas
Guillermo Calvo-Brenes

Resumen

La inteligencia artificial (IA), y en particular el aprendizaje automático (machine learning), ha emergido como una herramienta clave para mejorar el monitoreo ambiental, permitiendo desarrollar modelos predictivos capaces de estimar contaminantes a partir de variables fácilmente medibles. Este estudio aplicó técnicas de aprendizaje automático, específicamente Máquina de Vectores de Soporte (SVM), para predecir concentraciones de elementos potencialmente tóxicos (PTEs) en biomonitores vegetales y polvo de carretera de la Gran Área Metropolitana de Costa Rica. Se recolectaron 160 muestras de hojas de Casuarina equisetifolia y Cupressus lusitanica, junto con 80 muestras de polvo de carretera, durante campañas en 2020 y 2021. Las muestras se analizaron para determinar propiedades magnéticas y concentraciones de metales (Fe, Cu, Cr, Ni, Pb, V, Zn y Cd). Los modelos SVM, comparados con regresiones lineales simples y múltiples, mostraron mejor desempeño predictivo en la mayoría de los metales, con valores de R² superiores a 0.7 para Fe, Cu, Cr, V y Zn, y menores para Pb y Ni. C. equisetifolia presentó los mejores resultados en términos de ajuste (R²) y menor error (RMSE, MAE), superando a C. lusitanica; el polvo de carretera, por su parte, mostró una mayor variabilidad debido a sus múltiples fuentes de contaminación. Los resultados demostraron que la integración de la IA en el monitoreo biomagnético facilita la identificación de relaciones complejas entre las propiedades magnéticas y la contaminación por PTEs, confirmando su eficacia como herramienta rápida y de bajo costo para la evaluación ambiental.

Detalles del artículo

Cómo citar
Salazar-Rojas, T., & Calvo-Brenes, G. (2026). Aplicación de la inteligencia artificial en el monitoreo ambiental: modelos predictivos de elementos potencialmente tóxicos (PTEs) mediante machine learning. Revista Tecnología En Marcha, 39(5), Pág. 250–263. https://doi.org/10.18845/tm.v39i5.8507
Sección
Aplicaciones científicas y ambientales de la IA

Citas

[1] X. Zhang, K. Shu, S. Rajkumar, and V. Sivakumar, “Research on deep integration of application of artificial intelligence in environmental monitoring system and real economy,” Environ. Impact Assess. Rev., vol. 86, Jan. 2021, doi: 10.1016/j.eiar.2020.106499.

[2] E. Alotaibi and N. Nassif, “Artificial intelligence in environmental monitoring: in-depth analysis,” Discover Artificial Intelligence, vol. 4, no. 1, Dec. 2024, doi: 10.1007/s44163-024-00198-1.

[3] D. B. Olawade, O. Z. Wada, A. O. Ige, B. I. Egbewole, A. Olojo, and B. I. Oladapo, “Artificial intelligence in environmental monitoring: Advancements, challenges, and future directions,” Hygiene and Environmental Health Advances, vol. 12, Dec. 2024, doi: 10.1016/j.heha.2024.100114.

[4] S. M. Popescu et al., “Artificial intelligence and IoT driven technologies for environmental pollution monitoring and management,” Front. Environ. Sci., vol. 12, 2024, doi: 10.3389/fenvs.2024.1336088.

[5] E. C. McClure et al., “Artificial Intelligence Meets Citizen Science to Supercharge Ecological Monitoring,” Patterns, vol. 1, no. 7, Oct. 2020, doi: 10.1016/j.patter.2020.100109.

[6] T. Salazar-Rojas, F. R. Cejudo-Ruiz, and G. Calvo-Brenes, “Comparison Between Machine Linear Regression (MLR) and Support Vector Machine (SVM) as Model Generators for Heavy Metal Assessment Captured in Biomonitors and Road Dust,” Environmental Pollution, vol. 314, p. 120227, 2022, doi: https://doi.org/10.1016/j.envpol.2022.120227.

[7] S. Ferebee, “Predicting Tomorrow: A Review of Machine Learning’s Role in Shaping Environmental Forecasts,” Premier Journal of Science, vol. 2, Nov. 2024, doi: 10.70389/PJS.100008.

[8] Q. Dai, M. Zhou, H. Li, X. Qian, M. Yang, and F. Li, “Biomagnetic monitoring combined with support vector machine: a new opportunity for predicting particle-bound-heavy metals,” Sci. Rep., vol. 10, no. 1, pp. 1–11, 2020, doi: 10.1038/s41598-020-65677-8.

[9] Y. Xu et al., “Integrating machine learning for enhanced spatial prediction and risk assessment of soil heavy metal(loid)s,” Environmental Pollution, vol. 383, Oct. 2025, doi: 10.1016/j.envpol.2025.126919.

[10] P. I. Angulo-Bejarano, J. Puente-Rivera, R. Cruz-Ortega, and E. Hanus-Fajerska, “plants Metal and Metalloid Toxicity in Plants: An Overview on Molecular Aspects Metal and Metalloid Toxicity in,” Plants: An Overview on Molecular Aspects. Plants, vol. 10, p. 635, 2021, doi: 10.3390/plants.

[11] T. Salazar-Rojas, F. R. Cejudo-Ruiz, and G. Calvo-Brenes, “Magnetic and Chemical Testing in Plants, Road Dust and Soil, as Indicators of Atmospheric Pollution,” Water Air Soil Pollut., vol. 235, no. 9, Sep. 2024, doi: 10.1007/s11270-024-07333-9.

[12] S. Stankovic, P. Kalaba, and A. R. Stankovic, “Biota as toxic metal indicators,” Environ. Chem. Lett., vol. 12, no. 1, pp. 63–84, 2014, doi: 10.1007/s10311-013-0430-6.

[13] G. Chen-ming, W. Bo, and S. Xiao-han, “Biomagnetic monitoring of urban atmospheric pollution: A review of magnetic signatures from different types of plants,” Science of the Total Environment, vol. 963, Feb. 2025, doi: 10.1016/j.scitotenv.2025.178518.

[14] T. Salazar-Rojas, S. Murillo-Murillo, R. Ulate-Molina, F. R. Cejudo-Ruiz, and G. Calvo-Brenes, “Spatial assessment of potentially toxic elements (PTEs) pollution in the vicinity of a cement plant through magnetic and chemical testing in plants and soil,” Environmental Challenges, vol. 21, Dec. 2025, doi: 10.1016/j.envc.2025.101313.

[15] M. Arsyad, V. A. Tiwow, M. J. Rampe, and H. L. Rampe, “Relationships of Magnetic Properties and Heavy Metals Content of Guano in Bat Cave, South Sulawesi, Indonesia,” Karbala International Journal of Modern Science, vol. 8, no. 3, pp. 406–414, 2022, doi: 10.33640/2405-609X.3254.

[16] T. Salazar-Rojas, F. R. Cejudo-Ruiz, and G. Calvo-Brenes, “Assessing magnetic properties of biomonitors and road dust as a screening method for air pollution monitoring,” Chemosphere, vol. 310, no. September 2022, 2023, doi: 10.1016/j.chemosphere.2022.136795.

[17] Nation State Program, “Seguimiento al desarrollo humano sostenible,” 2019. [Online]. Available: https://estadonacion.or.cr/informes/

[18] Ministry of Public Construction and Transportation, “ANUARIO DE INFORMACIÓN DE TRÁNSITO 2019,” Costa Rica, 2020.

[19] M. A. E. Chaparro, M. A. E. Chaparro, A. G. Castañeda Miranda, H. N. Böhnel, and A. M. Sinito, “An interval fuzzy model for magnetic biomonitoring using the specie Tillandsia recurvata L,” Ecol. Indic., vol. 54, pp. 238–245, 2015, doi: 10.1016/j.ecolind.2015.02.018.

[20] B. Aguilar et al., “Ficus benjamina leaves as indicator of atmospheric pollution: A reconaissance study,” Studia Geophysica et Geodaetica, vol. 56, no. 3, pp. 879–887, Jun. 2012, doi: 10.1007/s11200-011-0265-1.

[21] S. Zamudio and E. Carranza, “Cupressaceae,” Flora del Bajío y de Regiones Adyacentes, vol. 29, Pátzcuaro, Michoacán, 1994. Accessed: Mar. 27, 2025. [Online]. Available: http://inecolbajio.inecol.mx/floradelbajio/documentos/fasciculos/ordinarios/Cupressaceae%2029.pdf

[22] F. Rojas-Rodríguez and G. Torres-Córdoba, “Casuarina,” Revista Forestal Mesoamericana Kurú, vol. 10, no. 25, pp. 32–33, 2013, doi: https://doi.org/10.18845/rfmk.v10i25.1375.

[23] U.S. EPA, “Method 3051A Microwave Assisted Acid Digestion of Sediments, Sludges, Soils, and Oils.,” no. February, USA, pp. 1–30, 2007. Available: https://www.epa.gov/sites/default/files/2015-12/documents/3051a.pdf

[24] J. A. Dearing, “Environmental Magnetic Susceptibility: Using the Bartington MS2 System,” Environmental Magnetic Susceptibility, p. 52, 1999, Accessed: Mar. 27, 2025. [Online]. Available: https://gmw.com/wp-content/uploads/2019/03/JDearing-Handbook-OM0409.pdf?srsltid=AfmBOooQ9qCZFrk_srScVdcwAG0c4Yt1k6uz-PM5ODYwv2XiGXpM3BiR

[25] RStudio Team, “RStudio: Integrated Development for R. RStudio, PBC.,” 2020, R Foundation for Statistical Computing, Boston, USA: 1.4.1717. [Online]. Available: http://www.rstudio.com/.

[26] E. Acuña, “Regresión Aplicada usando R,” 2015, [Online]. Available: https://www.researchgate.net/publication/335796371_REGRESION_APLICADA_USANDO_R

[27] A. Raj, “Unlocking the True Power of Support Vector Regression,” Towards Data Science. Accessed: Dec. 17, 2021. [Online]. Available: https://towardsdatascience.com/unlocking-the-true-power-of-support-vector-regression-847fd123a4a0

[28] H. Li, J. Wang, Q. Wang, C. Tian, X. Qian, and X. Leng, “Magnetic Properties as a Proxy for Predicting Fine-Particle-Bound Heavy Metals in a Support Vector Machine Approach,” Environ. Sci. Technol., vol. 51, no. 12, pp. 6927–6935, 2017, doi: 10.1021/acs.est.7b00729.

[29] H. Li, J. Wang, Q. Wang, C. Tian, X. Qian, and X. Leng, “Magnetic Properties as a Proxy for Predicting Fine-Particle-Bound Heavy Metals in a Support Vector Machine Approach,” Environ. Sci. Technol., vol. 51, no. 12, pp. 6927–6935, 2017, doi: 10.1021/acs.est.7b00729.

[30] G. Wang et al., “Magnetic properties of street dust in Shanghai, China and its relationship to anthropogenic activities,” Environmental Pollution, vol. 255, p. 113214, 2019, doi: 10.1016/j.envpol.2019.113214.

[31] A. Schmidt, R. Yarnold, M. Hill, and M. Ashmore, “Magnetic susceptibility as proxy for heavy metal pollution: A site study,” J. Geochem. Explor., vol. 85, no. 3, pp. 109–117, 2005, doi: 10.1016/j.gexplo.2004.12.001.

[32] S. G. Lu, S. Q. Bai, and Q. F. Xue, “Magnetic properties as indicators of heavy metals pollution in urban topsoils: A case study from the city of Luoyang, China,” Geophys. J. Int., vol. 171, no. 2, pp. 568–580, 2007, doi: 10.1111/j.1365-246X.2007.03545.x.

[33] O. O. Okedeyi, S. Dube, O. R. Awofolu, and M. M. Nindi, “Assessing the enrichment of heavy metals in surface soil and plant (Digitaria eriantha) around coal-fired power plants in South Africa,” Environmental Science and Pollution Research, vol. 21, no. 6, pp. 4686–4696, 2014, doi: 10.1007/s11356-013-2432-0.