Construction and Evaluation of In-situ Observation Dataset and Its Climatology in Arctic Ocean

The Arctic Ocean has increasingly drawn widespread attention in global climate change system. However, due to the high-latitude air-sea characteristics and the seasonal distribution of sea ice, the on-site marine environment surveys are more challenging than other oceans.

To understand the ice‒sea thermal dynamic processes, we built the in-situ observation dataset based on a series of international in-situ observation plans carried out in the Arctic Ocean and Chinese Arctic Research Expedition. With the support of polar icebreakers Xuelong and Xuelong-2, China has carried out a series of scientific investigations in Arctic Ocean for special phenomena, and accumulated many first-hand in-situ observations.

We used quality control and data processing methods to analyze and re-arrange the data mentioned above and obtained nearly a million thermohaline profiles from1983 to 2023. Meanwhile, a monthly climatology dataset is established with a horizontal resolution of 0.25×0.25° and 57 vertical layers. The datasets can serve as a standard reference for future observation data quality control, and can also be used to correct the thermohaline results of existing ice-ocean coupled models.

In order to evaluate the quality of the in-situ observations dataset, we selected typical water exchange areas for water mass analysis and partial thermohaline profile analysis，the result shows a significant seasonal variation and has a high quality and effectively reflects the overall hydrological characteristics of the Arctic Ocean. Meanwhile we compared the climatology datasets with WOA18, and find out there is clearly positive feedback by using Chinese Arctic Research Expedition data in the climatology datasets we built. And the thermohaline has stronger continuity and more stable structure. In the key of Chinese Arctic Research Expedition area, the analysis can reflect the high temperature Pacific water flowing into the Arctic Ocean, with a clear meridional temperature stratification, and temperature gradually decreasing from south to north.

Evaluating Ocean Heat Content (OHC) with in-situ observations climatology datasets show that the climatology dataset reflects the accurate state of the OHC, and can be used to verify and evaluate the OHC calculated from different model.

Next step, for studying the thermohaline structure of the Arctic ocean, we will use AI models for training with reanalysis data to get the prediction field by using the observation datasets we built.