Developing a semi-automated technique of surface water quality analysis using GEE and machine learning: A case study...

Area of Science:

• Environmental Science
• Remote Sensing
• Machine Learning

Background:

• The Sundarbans ecosystem faces significant threats from climate change and anthropogenic activities.
• Traditional water quality monitoring methods are often resource-intensive and impractical for large, remote areas.

Purpose of the Study:

• To develop and validate a semi-automated approach for assessing water quality in the Sundarbans.
• To integrate field and remotely-sensed data using machine learning for enhanced prediction accuracy.
• To provide a scalable and efficient tool for continuous water quality monitoring in ecologically sensitive regions.

Main Methods:

• Utilized machine learning (ML) algorithms and Empirical Bayesian Kriging (EBK) for predicting water quality parameters (SST, TSS, Turbidity, Salinity, pH).
• Leveraged Google Earth Engine (GEE) and AutoML with deep learning libraries for dynamic model creation.
• Integrated field measurements with remotely-sensed data for model calibration and validation.

Main Results:

• ML-based models demonstrated high accuracy in capturing spatial and temporal water quality variations.
• The semi-automated approach showed strong agreement with field measurements.
• The EBK model effectively interpolated predicted water quality parameters.

Conclusions:

• The developed semi-automated technique is a valuable tool for continuous and efficient water quality monitoring in the Sundarbans.
• This approach supports climate resilience and policy-making by enabling early detection of water quality degradation.
• The study advances water quality assessment methods for vulnerable ecosystems, promoting sustainable management practices.