An Approach for Monitoring Chlorophyll-A and Turbidity Levels in Saduperi and Kaveripakkam Lakes Located in and Around Vellore Region, Tamil Nadu, India, Using Spectral Unmixing Technique

The paper examines the levels of Chlorophyll-a and turbidity in lake water, which are essential indicators of freshwater ecosystem health. These parameters, indicative of algal presence and suspended matter, respectively, significantly impact aquatic life and human activities. Conventional assessment methods have limitations in coverage, spatial and temporal resolution, and cost, necessitating innovative approaches. This study adopts a comprehensive strategy by merging advanced remote sensing techniques with seasonal monitoring to address a significant gap in understanding Chlorophyll-a and turbidity dynamics in Saduperi and Kaveripakkam Lakes using ESA Sentinel-2 satellite imagery. By leveraging remote sensing technology and advanced algorithms, the research overcomes the shortcomings of traditional monitoring methods. A robust multispectral unmixing model is developed, incorporating carefully chosen endmembers representing Chlorophyll-a and turbidity spectral signatures. Validation with laboratory-derived ground truth data confirms the accuracy and reliability of the unmixing results, achieving high correlation coefficients (R² values) and low root mean square error (RMSE) values for all seasons. Specifically, the correlation analysis for chlorophyll-a showed an R² value above 0.8 and RMSE values below 0.5, indicating strong linear relationships and high accuracy in the estimation process. For turbidity, R² values were above 0.7, with RMSE values less than 0.4 across all seasons. Findings reveal distinct spectral signatures for Chlorophyll-a and turbidity across seasons. During winter, Chlorophyll-a absorption peaks in the red and blue spectral bands (bands 2 and 4), while turbidity variations are noticeable in shorter wavelengths. In summer, Chlorophyll-a levels rise significantly, accompanied by increased turbidity. The monsoon season shows localized responses, with fluctuating Chlorophyll-a and turbidity levels. Linear spectral unmixing effectively estimates Chlorophyll-a and turbidity concentrations, offering insights into water quality dynamics. The study explores the methodology of spectral signature matching, elucidating the spectrum’s concentration range. This innovative approach enables accurate estimation, mapping, and temporal analysis of Chlorophyll-a and turbidity in lake water, bridging gaps in remote sensing applications for water quality assessment. Integration with Sentinel-2 imagery enhances understanding of freshwater environments, aiding informed decision-making for sustainable resource management.