Validation of remote sensing chlorophyll-a concentration with considering effect of non-uniform vertical profiles

This is the authors' abstract. We don't add key points for this paper.

The validation of satellite-derived chlorophyll-a concentration (Chlasat) products is crucial for ocean colour remote sensing applications. The rapid development and deployment of Biogeochemical Argo (BGC-Argo) floats have provided a large amount of time-series Chla vertical profile data for validation. When only using the surface Chla (average within top 10 metres, Chla10 m) as the ‘ground truth’ value, the validation in the open ocean of Eastern Pacific indicates Chlasat to be significantly overestimated, particularly in the region with low Chla10 m ( < 0.1 mg m−3), and the mean absolute percent difference (MAPD) peaks at 170%. The nonuniform vertical distribution of Chla was found to be mainly responsible for this high uncertainty. The more prominent the nonuniform vertical distribution (such as a high ratio of deep chlorophyll-a maximum (DCM) to Chla10 m and shallow mixing layer depth (MLD)), the higher the difference between Chlasat and Chla10 m. Radiative transfer simulations of remote sensing reflectance (Rrs) confirmed these findings and demonstrated that Chla distribution below the DCM depth had little impact on Rrs. Unlike previous studies relying on surface Chla or fixed-depth averages, we proposed the average Chla within the DCM depth (ChlaDCM) as the more suitable ‘ground truth’ value, such that the MAPD between ChlaDCM and Chlasat was only 31.77%, satisfying the IOCCG requirement of < 35%. This indicates that Chlasat cannot be simply regarded as surface Chla but reflects the vertical distribution of Chla within the DCM depth, which should be advantageous for a more accurate evaluation of marine primary productivity in the open ocean when using Chlasat products.