Impacts of subsurface chlorophyll-a distribution on remote sensing reflectance spectra

Traditional ocean color remote sensing inversion algorithms often overlook the influence of the subsurface chlorophyll maximum layer (SCML) on remote sensing reflectance (Rrs), potentially leading to errors in estimating surface chlorophyll-a (Chl-a) concentration. To address this issue, in this study, we parameterize Chl-a vertical profiles using a Gaussian model and perform spectral simulations of Rrs under varying SCML scenarios with the Hydrolight radiative transfer model. The aim is to explore the response mechanism of Rrs to the SCML for more accurate retrieval of surface Chl-a concentration. By comparing the Rrs spectra from vertically inhomogeneous (including SCML) and homogeneous Chl-a profiles, three key findings were obtained about the SCML. First, the presence of an SCML introduces considerable biases in traditional retrieval models. Second, a global sensitivity analysis reveals that SCML parameters have a dominant influence on Rrs, with a total sensitivity index exceeding 50 % within the blue bands (395–475 nm), whereas Rrs(485)/Rrs(535) exhibits the lowest sensitivity for surface Chl-a estimation. Third, an empirical retrieval model based on Rrs(488)/Rrs(531) is proposed, which outperformed the MODIS standard algorithm. Independent validation datasets from three different ocean regions (the Mediterranean, Atlantic, and Pacific) confirm its robustness in Case-1 waters with SCML, achieving a Root Mean Square Error (RMSE) of less than 0.012 mg m −3 . This study advances current ocean color methods by systematically quantifying the impact of SCML on Rrs and providing a validated algorithm that mitigates this effect. These findings offer a pathway to improve global Chl-a products, which are critical for applications in global carbon cycle monitoring and primary productivity estimation. • Subsurface chlorophyll maximum layer (SCML) introduces considerable biases in traditional ocean color remote sensing inversion models for surface chlorophyll-a retrieval. • Global sensitivity analysis reveals that SCML parameters (depth, thickness, and magnitude) have dominant influence on remote sensing reflectance, with total sensitivity index exceeding 50 % in blue bands (395–475 nm). • A novel band-ratio algorithm based on Rrs(488)/Rrs(531) effectively mitigates SCML interference and outperforms the MODIS standard algorithm with RMSE less than 0.012 mg m −3 .