Predictive Modeling of Effective Electrical Conductivity for Irrigation Water in Mendoza, Argentina
Year: 2026
Type: article
Abstract: This study develops five predictive models based on regression trees to estimate the Effective Electrical Conductivity (CEE) of irrigation water in the Green Belt of Mendoza, Argentina. Using 468 water samples and physicochemical measurements from the Chachingo and Pescara canals, the models incorporate between one and six predictor variables. The most prominent model includes values of Actual Electrical Conductivity (CEA), calcium, magnesium, and bicarbonate ions, with a correlation coefficient of 0.99 and a mean absolute error of 26.5 µS cm-1. CEE is a key indicator for assessing water quality in areas with moderately soluble salts. These models provide a cost-effective tool for identifying salinity risks in irrigation water of varying quality. Their simplicity and robustness make them suitable for Mendoza or other regions with similar conditions, supporting efficient water resource management in agriculture. Highlights: Los modelos de árboles de regresión permitieron predecir con precisión la conductividad eléctrica efectiva (CEE) en aguas de riego bajo condiciones locales. Los modelos se desarrollaron utilizando 468 muestras de agua del sistema de riego del Cinturón Verde de Mendoza. El Modelo 3 alcanzó un alto desempeño predictivo (R = 0,99) utilizando solo cuatro variables de fácil medición. El enfoque propuesto reduce la necesidad de análisis fisicoquímicos completos, disminuyendo costos operativos. Esta metodología constituye una herramienta práctica para el monitoreo eficiente de la calidad del agua de riego en regiones semiáridas.
DOI: 10.48162/rev.39.214
Authors: Daniela Virginia Cónsoli, María Flavia Filippini
Institution: Universidad Nacional de Cuyo
Language: English
Subfield: Aquatic Science
Domain: Life Sciences
Open Access status: gold
Estimated APC paid: $300
OpenAlex ID: W7159770258
FWCI:
Cites:
Cited by:
Related to: