FUZZY INFERENCE SYSTEM TO EVALUATE THE QUALITY OF GROUNDWATER IN MEXICAN WATER BODIES 

Abstract

Introduction: This paper presents the development of a fuzzy inference system to assess groundwater quality using data from 2012-2021 from the national water information system (SINA). The objective was to create a simplified semaphore based on fuzzy logic, classifying groundwater into CONAGUA’s three traditional categories (green, yellow, red) while incorporating a degree of membership for each condition. Methodology: CONAGUA classifies water quality using 14 crisp variables, but we employed eight fuzzy variables as inputs to a Mamdani inference system. Results: Our fuzzy system achieved 84% similarity with CONAGUA’s classification while providing an intraclass distribution for each semaphore color. A robustness evaluation using 2021 data showed comparable classification distribution (67% green, 62% yellow, and 49% red). The system effectively classifies gradual quality using key indicators: conductivity, hardness, total dissolved solids (TDS), and metal levels, aligning with CONAGUA’s classical semaphore. Conclusion: Despite the existence of a superficial water semaphore, we propose using the groundwater semaphore instead. The superficial classification does not consider metals, yet preliminary multidisciplinary findings indicate metal presence in the Tampamachoco Lagoon. Therefore, the groundwater semaphore could be a suitable tool for assessing Tampamachoco Lagoon’s water quality in future studies.