Empirical algorithm for detecting coccolithophore blooms through satellite observation in the Barents Sea.

Coccolithophores are a key phytoplankton group with a major influence on the carbon cycle in global oceans. Understanding its phenological responses to climate change is vital, particularly in Arctic regions that are experiencing warming faster than the global average. Ocean color remote sensing is a powerful tool that provides the global distribution and spatial extent of coccolithophore blooms. However, remotely observing coccolithophore blooms in Arctic regions is challenged by severe conditions, such as heavy cloud cover. In this work, an ocean color index (CI)-based approach is developed to detect coccolithophore blooms in the Arctic Sea by using Moderate Resolution Imaging Spectroradiometer (MODIS) images. The optical properties of clouds and coccolithophore blooms are characterized by employing the band difference of Rayleigh-corrected reflectance (ρR) at 412 and 469 nm (CI 412–469) and 555 and 645 nm (CI 555–645), respectively. A hybrid threshold is further determined to classify coccolithophore bloom and nonbloom waters. Results show that the algorithm can reduce the negative effects of heavy clouds and persistent fog on detecting coccolithophore blooms and increase the coverage of satellite observation. Compared with the traditional particulate inorganic carbon concentration-based approach, the algorithm yields more coverage and frequency of coccolithophore blooms. Global evaluation reveals that the CI-based algorithm has great potential applications in global oceans and different ocean color sensors. • A new algorithm is developed to detect coccolithophore blooms in the Barents Sea. • The algorithm can reduce the effect from heavy cloud cover in high-latitude regions. • The approach has potential applicability in global oceans and next-generation ocean color remote sensing sensors. [ABSTRACT FROM AUTHOR]