Comparative Analysis of Water Quality Applying Statistic and Machine Learning Method: A Case Study in Coyuca Lagoon and Tecpan River, Mexico

Abstract

The water quality monitoring of lotic and lentic ecosystems allows for informing the possible use in human activities and the consumption of the vital liquid. This work measures the biochemical parameters in Coyuca Lagoon and Tecpan River, localized in Guerrero, Mexico. A comparative statistical analysis of six physicochemical factors in lentic and lotic ecosystems was carried out, finding individual pH values slightly higher for the lagoon ecosystem and lower for the river. For electrical conductivity, we find river sites with parameters lower than 500 μS/cm ideal for human use and consumption. On the contrary, in sites of the lagoon system, the conductivity was higher. As for the total hardness of the river, the values are within the Mexican standard; however, for the lagoon ecosystem, the water has a higher amount of calcium and magnesium salts and is not recommended for human consumption. For chlorides, the lagoon system exceeds the limits of regulations for human consumption; otherwise, it happens with the lotic system. The values of total alkalinity and total dissolved solids are higher for the lentic system than for the lotic one. Finally, the machine learning method shows the importance of measuring other parameters to determine the water quality, especially the salinity and calcium hardness.

Keywords

• artificial intelligence
• aquatic ecosystems
• applied mathematics
• punctual water condition