Remote Sensing of Particulate Organic Carbon on the Western Antarctic Peninsula Shelf Using a Color Index-Based Algorithm

A color index (CI)–based algorithm (CI $_{\mathbf {POC}}$ ) for estimating particulate organic carbon (POC) concentrations from satellite observations is evaluated for the Western Antarctic Peninsula (WAP) shelf. The CI utilizes three spectral bands centered approximately at 490, 550, and 670 nm of the ocean color satellite sensors. The CI $_{\mathbf {POC}}$ algorithm is calibrated and validated using field-measured POC concentration data and match-up satellite-derived remote-sensing reflectance ( $\text{R}_{\mathbf {rs}}$ ) data. For comparison, the performance of the blue–green (BG) band ratio algorithm is also evaluated using the same match-up dataset. Results show that the CI $_{\mathbf {POC}}$ algorithm outperforms the BG band ratio algorithm in determining the POC concentration for the WAP shelf with good statistical parameters of algorithm performance for the CI $_{\mathbf {POC}}$ algorithm. Given that the CI $_{\mathbf {POC}}$ algorithm is built upon a three-band-difference concept and is less sensitive to artificial errors in $R_{\mathbf {rs}}$ than the BG algorithm, CI $_{\mathbf {POC}}$ can correct POC concentration by retrieving uncertainties caused by sea ice and cloud contaminations. Further evaluation showed that the CI $_{\mathbf {POC}}$ algorithm can derive POC from the next-generation ocean color satellite sensors with robust performance and noise tolerance.