Modelado numérico y clasificación del riesgo ecológico por transporte de contaminantes en cuerpos de aguas superficiales

Abstract

This paper examines water pollution and its health risks. The primary goal is to identify ecological risk zones using a hybrid approach that combines numerical solutions and machine learning. The two-dimensional advection-diffusion equation is used to model pollutant dispersion, creating a dataset for training a supervised classification model. The results demonstrate high accuracy, surpassing 95%, in predicting risk zones (low, medium, or high). It is concluded that this method is an effective tool for water quality management thus having the potential to adapt to real conditions.