889 / 2024-09-19 18:29:47
PACE Hyperspectral Data Enables Enhanced Identification of Algal Bloom Types in Coastal Waters
PACE, coastal waters, atmospheric correction, algal bloom types
Session 54 - Remote sensing of coastal zone and sustainable development
Abstract Accepted
Renhu Li / East China Normal University
Fang Shen / East China Normal University
Harmful algal blooms (HABs) pose a significant threat to marine ecosystems worldwide, requiring advanced approaches for effective monitoring and classification. The Plankton, Aerosol, Cloud, ocean Ecosystem (PACE) mission offers hyperspectral measurements of remote sensing reflectance (Rrs) , providing enhanced spectral resolution for improved algal bloom classification. Yet, acquiring accurate hyperspectral Rrs from PACE is still challenging for coastal waters. In this effort, we present a novel chief band atmospheric correction algorithm (CB-AC) specifically designed for PACE in coastal waters. The CB-AC algorithm estimates aerosol reflectance and transmittance at key wavelengths (380 nm, 412 nm, 443 nm, 490 nm, 555 nm, 620 nm, 660 nm, 745 nm, and 865 nm), and interpolates these values across all PACE bands for the atmospheric correction. Validation against measured Rrs from eight AERONET sites suggests that CB-AC outperforms the that NASA standard algorithm. Specifically, the CB-AC-derived Rrs achieved an R² of 0.84, an APD of 32.64%, and an RMSE of 0.0014 sr-1, compared to an R² of 0.58, an APD of 33.26%, and an RMSE of 0.0019 sr-1 for NASA Rrs product. Additionally, we developed a PACE-specific algal bloom classification algorithm, termed Hyperspectral-Slope-Algal-Class, using simulated hyperspectral Rrs data of various algal bloom types. The Hyperspectral-Slope-Algal-Class algorithm differentiates diatom blooms from dinoflagellate blooms through spectral slope differences at 610 (Slope(610)) nm, 640 nm (Slope(640)), and 680 nm (Slope(680)), The Slope(680) can reflect the biomass of algal blooms. As the biomass changes, the Slope(610) and Slope(640) of diatom and dinoflagellate blooms show significant differences. This algorithm uses the Slope(680) under different conditions as the baseline to distinguish between diatom and dinoflagellate blooms. When the difference between Slope(640) and Slope(610) is greater than the baseline, it is identified as a dinoflagellate bloom; otherwise, it is identified as a diatom bloom. The classification results by Hyperspectral-Slope-Algal-Class align well with existing bloom measurements of diatom and dinoflagellate blooms in the East China Sea, highlighting the great potential of PACE data in advancing algal bloom monitoring and differentiation.