Related Experiment Video
Updated: Jun 27, 2026
10:28
Investigating the Relationship between Sea Surface Chlorophyll and Major Features of the South China Sea with Satellite Information
Published on: June 13, 2020
A Synergistic Remote Sensing Inversion Study of Water Depth in Inland Lakes Integrating Chlorophyll-a Concentration
Junzhen Meng1, Yunfei Wang1, Jiajun Ren1
1School of Geomatics and Geographical Information Engineering, North China University of Water Resources and Electric Power, Zhengzhou 450045, China.
Sensors (Basel, Switzerland)
|June 26, 2026
Summary
Accurate inland lake bathymetry is crucial for water management. Machine learning models integrating chlorophyll-a and a Water Optical Index significantly improve remote sensing retrieval accuracy, outperforming traditional methods.
Area of Science:
- Earth and Environmental Sciences
- Remote Sensing
- Water Resource Management
Background:
- Accurate bathymetry is vital for inland lake management, but remote sensing retrieval faces challenges due to complex water optics and limited features.
- Conventional methods often lack the accuracy and robustness required for effective water resource monitoring and ecological research.
Purpose of the Study:
- To systematically compare machine learning models (Random Forest, XGBoost, AdaBoost) with a multiband logarithmic ratio model for inland lake bathymetric retrieval.
- To develop and evaluate a synergistic retrieval framework integrating chlorophyll-a concentration (Chla) and a Water Optical Index (WOI) for enhanced bathymetric accuracy.
Main Methods:
- Comparison of a multiband logarithmic ratio model with Random Forest (RF), XGBoost, and AdaBoost machine learning models.
- Development of a synergistic framework combining Chla and WOI with machine learning for bathymetric inversion.
- Validation of model performance using metrics such as R², Mean Absolute Error (MAE), and Root Mean Square Error (RMSE).
Main Results:
- Machine learning models, particularly Random Forest, significantly outperformed the traditional numerical model in bathymetric retrieval accuracy.
- The synergistic framework integrating Chla and WOI with machine learning models showed superior performance compared to models using only multispectral bands.
- The optimal Random Forest model achieved high-precision bathymetric inversion (R²=0.93, MAE=0.06 m, RMSE=0.14 m) and accurately estimated lake storage capacity with a -2.03% relative error.
Conclusions:
- The proposed synergistic retrieval framework offers higher applicability and accuracy for inland lake bathymetry in complex environments.
- Machine learning models, especially Random Forest, integrated with Chla and WOI provide reliable and highly accurate data for water resource management.
- This study demonstrates the potential of advanced remote sensing techniques for effective monitoring and management of inland water bodies.
