A Data-Based Framework for Identifying a Source Location of a Contaminant Spill in a River System with Random...

Area of Science:

• Environmental Science
• Water Resource Management
• Data Science

Background:

• Contaminant spills in river systems pose significant environmental risks.
• Accurate source identification is crucial for effective remediation.
• Sensor networks often provide data with inherent measurement errors.

Purpose of the Study:

• To develop and validate a novel framework for identifying contaminant spill source locations in river systems.
• To address the challenge of random measurement errors in sensor observations.
• To provide a robust method for pinpointing spill origins with high accuracy.

Main Methods:

• A three-step framework: spill detection, data preprocessing, and source identification.
• Application of statistical process control charts for spill detection with a controlled false alarm rate.
• Utilization of nonlinear regression for breakthrough curve estimation and random forest models for source localization.

Main Results:

• The framework successfully detects spills and identifies potential source locations.
• The random forest model outputs possibility values (0-1) for candidate locations.
• The approach was validated using data from the Altamaha River system.

Conclusions:

• The proposed framework offers an effective solution for contaminant spill source identification in rivers.
• The integration of statistical process control and machine learning enhances accuracy despite measurement errors.
• This method aids in rapid and precise response to river pollution incidents.