A Scheme of Atmospheric Correction Using Same-day Observations

Robust atmospheric correction to obtain accurate remote-sensing reflectance (R_rs, sr^-1) of a water body in the upper layer from satellite ocean color measurements is still a challenging task for coastal waters and/or scenes involving strongly absorbing aerosols. Here, with a revised POLYMER (POLYnomial based approach applied to MERIS data) atmospheric correction model, we describe a novel retrieval scheme that determines oceanic and atmospheric optical properties simultaneously combining multiple satellite observations with different sun-sensor viewing geometries through the spectral optimization approach. The new algorithm is first validated with a synthetic dataset where the results suggest that this new scheme can derive highly accurate retrieved R_rs, with R² ≥ 0.96, where the root mean square error (RMSE) is less than ~3.2 × 10^-4 sr^-1 and the unbiased root mean square (uRMS) is less than ~26.4%. The scheme is further applied to VIIRS data, and validated against in-situ measurements in Massachusetts Bay. It is found that the retrieved R_rs from VIIRS agree with in-situ measurements within ~20.4% at the visible wavelengths. In contrast, the NOAA’s operational R_rs products have a difference of ~164.9% at 410 nm for these measurements. The results show high potential of this new scheme for atmospheric correction in coastal waters involving strongly absorbing aerosols.