Research of chlorophyll-a concentration inversion at different depths in Hong Kong offshore waters based on gaussian process regression

Abstract

By inverting the changes in chlorophyll-a (chl-a) concentration values, seawater’s pollution level and phytoplankton’s growth can be effectively assessed. In this paper, based on Sentinel-3 OLCI remote sensing image data and actual measured chl-a concentration data, the Gaussian Process Regression (GPR) model was used to invert the surface chl-a concentration in Hong Kong waters in 2018, and the average R2 of GPR model inversion was 0.68. The results show a clear vertical decreasing variation feature based on the GPR model to invert the chl-a concentrations in the surface, middle, and bottom layers of Hong Kong waters month by month in 2018.