Evaluation of the Threshold for an Improved Surface Water Extraction Index Using Optical Remote Sensing Data

Area of Science:

• Environmental Science
• Remote Sensing
• Geospatial Analysis

Background:

• Accurate lake surface water detection is crucial for environmental monitoring.
• Traditional water extraction methods often struggle with precise boundary delineation.
• Optical remote sensing data presents challenges in defining water-nonwater thresholds.

Purpose of the Study:

• To enhance the accuracy of automatic lake surface water detection.
• To improve threshold value determination for the Automatic Water Extraction Index (AWEI).
• To validate the improved AWEI model on diverse Indonesian lakes.

Main Methods:

• Developed an improved AWEI threshold model using the split-based approach (SBA).
• Utilized Google Earth Engine (GEE) for Landsat 8 annual mosaic creation.
• Applied geostatistical analysis (smart quantiles) for threshold calculation.

Main Results:

• The improved AWEI threshold (≥ -0.23) achieved nearly 100% overall accuracy.
• This represents a 2% accuracy increase compared to the standard AWEI threshold (≥ 0.00).
• Applied to Indonesian lakes, accuracy ranged from 94% to 100%.

Conclusions:

• The SBA-based smart quantile approach significantly improves AWEI threshold accuracy.
• The enhanced AWEI model is effective for detecting lake surface water across varied conditions.
• This method offers a more precise tool for hydrological and environmental studies.