Artificial intelligence approach applied to the time of concentration of hydrographic units

López-Silva M.; Alca-Vásquez B.; Ricaldi-Marquez V. · Tecnologia y Ciencias del Agua (2026) · DOI: 10.24850/j-tyca-2026-03-07

Resumen

The present study proposed an artificial intelligence approach for obtaining the time of concentration in hydrographic units. First, 30 empirical equations were examined on a database of 199 records and second, gene expression programming (GEP) was guided on a data matrix of 199 records from different continents. Finally, the prediction models were analyzed under three error metrics, the coefficient of determination (R2), the mean absolute error (MAE) and the mean square error (MSE). Five accurate models were generated under the GEP technique, the most optimal being the fourth with R2 = 0.991, MAE = 0.152 and MSE = 0.086. However, when validated with real field data, the equation of model I presented the lowest relative error, demonstrating superior reliability compared to empirical equations such as Kirpich. It is concluded that, although more complex models can achieve high accuracy, the simplicity of the equations is crucial for their applicability in various hydrographic units. This artificial intelligence-based approach offers an effective alternative for hydraulic design and management in ungauged regions, with potential for future extensions of its applicability in hydrological studies.

Instituciones: Universidad Ricardo Palma