Remote Sensing Estimations of the Seawater Partial Pressure of CO₂ Using Sea Surface Roughness Derived From Synthetic Aperture Radar

• Published in: IEEE transactions on geoscience and remote sensing 2024, Vol.62, p.1-13
• Main Authors: Wang, Yiren, Wu, Zelun, Lu, Wenfang, Yu, Shujie, Li, Shihui, Meng, Lingsheng, Geng, Xupu, Yan, Xiao-Hai
• Format: Article
• Language: English
• Subjects: Algorithms; Carbon; Carbon cycle; Carbon dioxide; Carbonates; Chemical analysis; Chlorophyll; Chlorophylls; Coastal zone; Cubist method; Error reduction; Independent variables; Learning algorithms; Machine learning; Marine systems; Mathematical models; Ocean dynamics; Ocean temperature; Oceans; Parameters; Partial pressure; Radar; Remote sensing; Root-mean-square errors; SAR (radar); Sea measurements; Sea surface; sea surface partial pressure of carbon dioxide (pCO2w); sea surface roughness (SSR); Sea surface temperature; Seawater; Spatial discrimination; Spatial resolution; Statistical analysis; Surface roughness; Surface salinity; Surface temperature; Synthetic aperture radar; synthetic aperture radar (SAR); Water analysis; Wave dynamics; Wind; Wind speed
• ISSN: 0196-2892; 1558-0644
• DOI: 10.1109/TGRS.2024.3379984

Remote sensing study of the carbon cycle in coastal marine systems using machine learning methods has received significant attention recently. The partial pressure of carbon dioxide (CO2) in seawater ( {p} CO2w) is a crucial parameter for quantifying the air-sea carbon dioxide exchange. However, previous studies did not consider the effect of sea surface roughness (SSR) on {p} CO2w caused by wind, waves, and other ocean dynamics. In this study, for the first time, we used SSR data derived from synthetic aperture radar (SAR), with sea surface temperature (SST), chlorophyll- {a} (Chl- {a} ) concentration, sea surface salinity (SSS) conventional remote sensing data to predict the {p} CO2w data along the North American East Coast from 2015 to 2021 using the Cubist algorithm. Results show that the semi-analytic algorithm, Cubist, performs best among 20 statistical and machine learning models. Moreover, compared with the control experiment without the SSR data, after adding SSR as an independent variable, the final Cubist model's coefficient of determination ( {R} ^{2} ) increased from 0.88 to 0.95, and the root mean square error (RMSE) reduced from 21.75 to 14.79~\mu atm. Our results showed significant improvement over the previous study ( {R} ^{2} = 0.8), proving the applicability of applying SSR data in retrieving high spatial resolution carbonate system parameters in the future, especially for coastal regions where wind and wave dynamics are more variable.