The capacity of CO₂ uptake in the Chukchi Sea is particularly sensitive to rapid physical and biological changes. However, scarce field observations pose a challenge in understanding the long-term trend of CO₂ uptake capacity on this continental shelf. We adopted a machine-learning-based approach to construct a 17-years (2003–2019) long-term time series of summer sea surface partial pressure of CO₂ (pCO₂) from remote sensing products. We show that the long-term increase in CO₂ uptake capacity can be attributed to strong and enhanced biological uptake. In addition, the intraseasonal variability of surface pCO₂ in early summer confirms the crucial role of sea ice melt and the subsequently enhanced photosynthesis as soon as the surface ocean converts into an open system. Our results thus highlight the use of remote sensing data in interpolating/extrapolating the highly dynamic carbonate system in the continental shelf sea and shed light into future studies involving machine learning or algorithms.

中文翻译：

---

基于卫星的估计揭示了夏季北极楚科奇海二氧化碳吸收能力的增加

楚科奇海的 CO ₂吸收能力对快速的物理和生物变化特别敏感。然而，稀有的实地观测对了解该大陆架上CO ₂吸收能力的长期趋势构成了挑战。我们采用基于机器学习的方法，从遥感产品中构建了 17 年（2003-2019）夏季海面 CO ₂分压( p CO ₂ )的长期时间序列。我们表明，CO ₂吸收能力的长期增加可归因于强大且增强的生物吸收。此外，地表p CO ₂的季节内变化初夏证实了海冰融化的关键作用，一旦表层海洋转变为开放系统，光合作用就会增强。因此，我们的结果突出了遥感数据在大陆架海中高动态碳酸盐系统的内插/外推中的使用，并为涉及机器学习或算法的未来研究提供了启示。