Performance analysis of inverting optical properties based on quasi-analytical algorithms

The inherent optical parameters play a very important role in determining the concentration of seawater components. One such inherent parameter, the absorption coefficient (a), is greatly significant when calculating each component’s content in water and simulating the water’s biological, physical and chemical properties. Quasi-analytical algorithm (QAA) is one of current inherent optical parameters inversion algorithms. In this study, we compared three versions of QAA, including QAA_V4, QAA_V5 and more recently QAA_V6, using IOCCG data set. The results showed that QAA_V4 model performs best for the inversion of the absorption coefficient at 443 nm, with MRE, RMSE and R2 values of 11.82%, 0.1995 and 0.9099, respectively, while QAA_V6 has the highest accuracy when inverting a at 555 nm and 670 nm wavelengths. Next, we tested the three models using MODIS product data, including data from case I and case II waters. The data is extracted from image, after position matching, then was imported into the three QAA models to calculate their RMSE, MRE and R2 values. The QAA model showed a high accuracy and robust applicability for absorption coefficient inversion at shorter wavelengths, such as 443 nm. Future research should include verification with more complicated water bodies, solidifying the foundation for implementing QAA into ocean color remote sensing research.