Optimization of water quality monitoring programs by data mining

Area of Science:

• Environmental Science
• Data Science
• Water Resource Management

Background:

• Water quality monitoring is crucial but faces resource limitations in developing nations.
• Existing optimization methods for monitoring programs are insufficient.
• Data mining applications for optimizing water quality monitoring are underexplored.

Purpose of the Study:

• To develop and test data-based models using data mining to reduce water quality parameters in monitoring programs.
• To estimate laboratory water quality variables using association rules and field parameters.
• To assess the effectiveness of this data mining approach across diverse Brazilian river basins.

Main Methods:

• Utilized data mining techniques, specifically association rules, to identify patterns in historical water quality data.
• Integrated field parameters measured by automatic probes to estimate laboratory-analyzed variables.
• Applied the methodology to a large dataset from 35 monitoring stations across 27 Brazilian river basins (1971-2021).

Main Results:

• Successfully estimated 56% of laboratory water quality parameters using the developed data-based models.
• Found that monitoring programs with 20 or more parameters showed the highest optimization potential (≥44%).
• Environmental characteristics influenced optimization capacity, but methodology was robust across different water quality levels and anthropogenic impacts.

Conclusions:

• Data mining offers a promising alternative for optimizing water quality monitoring programs by reducing laboratory analyses.
• This approach can significantly increase the spatial and temporal coverage of water quality monitoring networks.
• The number of parameters in a monitoring program is a key factor influencing optimization potential.