Assessment of Chlorophyll-a Algorithms Considering Different Trophic Statuses and Optimal Bands

Area of Science:

• Ocean optics and remote sensing
• Algorithmic development for water quality assessment

Background:

• Existing algorithms for chlorophyll-a retrieval in Case 2 waters show unsatisfactory accuracy.
• Multispectral and hyperspectral bands, combined with various regression approaches, are crucial for accurate remote sensing of water constituents.

Purpose of the Study:

• To evaluate the performance of seven distinct algorithms using multiple band combinations and regression techniques (linear, quadratic, power).
• To assess the accuracy of 43 algorithmic combinations using both simulated and measured datasets.
• To identify the strengths and limitations of different approaches for chlorophyll-a estimation in Case 2 waters.

Main Methods:

• Generation of two large simulated datasets (500,000 reflectance spectra each) with wide ranges of inherent optical properties (IOPs).
• Evaluation of 43 algorithmic combinations using linear (LN), quadratic polynomial (QP), and power (PW) regression on simulated and measured data.
• Comparative analysis of algorithm performance based on root mean square error (RMSE) and spatial distribution analysis.

Main Results:

• Linear regression approaches showed higher retrieval accuracy on simulated data compared to measured data.
• The 3-band (3b) algorithm incorporating 665-nm and 680-nm bands, along with band tuning, demonstrated superior performance on simulated data (RMSE: 15.87-19.05 mg·m⁻³).
• No single algorithm achieved outstanding accuracy across all conditions, particularly for low chlorophyll-a (Chla) and non-algal particle (NAP) concentrations, indicating the need for multi-algorithm strategies.

Conclusions:

• The choice of regression approach significantly impacts retrieval accuracy, with linear regression being most effective for simulated Case 2 water data.
• While specific band combinations (e.g., 665-nm) and tuning methods excel under certain conditions, a combination of algorithms is necessary for robust chlorophyll-a estimation.
• Future research should focus on developing integrated multi-algorithm systems to improve the reliability of chlorophyll-a concentration measurements in complex aquatic environments.