Predicting Water Availability in the Chalcas River Basin: Application of Artificial Neural Networks for Sustainable Resource Management

Hemerson Lizarbe Alarcón; Jhac Taboada-Valenzuela; Edward León-Palacios; José Estrada-Cárdenas; Rocky Giban Ayala Bizarro; Main G. Tenorio-Palomino; Rualth G. Bravo-Anaya; Alex Ircañaupa · International Journal of Environmental Impacts (2026) · DOI: 10.56578/ijei090210

Resumen

This study aimed to develop a predictive model of water availability using artificial neural networks (ANN) in the Chalcas River basin, located in the district of San Pedro de Palco, Ayacucho, Peru.A quantitative, predictive, and non-experimental longitudinal design was applied.Hydrological data were used, including monthly average precipitation (ranging from 3.44 mm in June to 123.49 mm in February), weighted crop coefficients (Kc), monthly evapotranspiration (ETo), and a drainage density of 5.81 km/km 2 .A multilayer ANN was structured and trained over 2000 epochs, achieving an average accuracy of 90.62% and a normalized mean absolute error (MAE) of 0.0528.The model determined the flow rate for the period 2003-2030 period, identifying critical seasonal patterns: a peak of 317.45 l/s in January 2028 and a minimum of 28.55 l/s in July 2026.These findings highlight the need to implement water storage strategies during wet seasons and optimize water use during dry periods.Ultimately, the ANN-based model enhances water resource management, reduces scarcity-related risks, and promotes the sustainability of the irrigation system.This methodology demonstrates broad applicability and can be replicated in other basins facing similar hydrological challenges, using the ANN model.

Instituciones: Universidad Nacional de San Cristóbal de Huamanga