Release of high-resolution dynamical downscaling dataset for the Arctic Ocean

The Arctic, Earth's northernmost region, is one of the most sensitive regions to global climate change and is undergoing unprecedented transformations. In the context of global warming, surface air temperatures in the Arctic are rising at more than twice the global average rate—a phenomenon known as Arctic Amplification. Climate change in the Arctic not only exerts significant impacts on local ecosystems but also plays a crucial role in regulating the global climate system.

Data form the critical foundation for understanding and addressing climate change. However, due to the unique geographical and climatic conditions in the Arctic, long-term continuous observations of the Arctic Ocean for climate research remain relatively sparse, which greatly limits scientific understanding of both the region and its changing climate. In response, scientists often rely on climate model simulations and projections to conduct relevant research. However, state-of-the-art climate models still face essential challenges in the Arctic Ocean, including low resolutions, substantial biases, and considerable inter-model spreads.

To address the scarcity of long-term observational data and the lack of high-resolution simulation and projection datasets for climate research in the Arctic Ocean, a Sino-German research team carried out close collaboration and successfully developed an Arctic Ocean dynamical downscaling dataset spanning the period from 1900 to 2100, with a horizontal resolution of up to 4.5 kilometers. This dataset is generated using the high-resolution sea-ice coupled model FESOM2 (version 2 of the Finite Volume Sea Ice-Ocean Model), driven by the bias-corrected surface field derived from the outputs of FIO-ESM v2.1 (the First Institute of Oceanography-Earth System Model, version 2.1). The dataset includes 115 years of historical simulations (1900-2014) and two 86-year future projections (2015-2100) following the SSP2-4.5 and SSP5-8.5 scenarios, which represent intermediate and high greenhouse gas emission pathways, respectively. This dataset contains 16 oceanic and sea ice parameters, primarily at monthly resolution, with select variables available at daily resolution..

Evaluation of the dataset shows that, compared to CMIP6 climate model output, it not only offers an increase in spatial resolution but also significantly reduces simulation errors for key variables such as seawater temperature, salinity, and sea ice thickness. Notably, the dataset also markedly improves the common biases in representing the Atlantic Water layer typically seen in climate model simulations.

The detailed descriptionof the dataset is published in Advances in Atmospheric Sciences and the dataset is archived in the Science Data Bank and is accessible via the following links.

Historical data: https://doi.org/10.57760/sciencedb.16206

SSP2-4.5 data: https://doi.org/10.57760/sciencedb.16286

SSP5-8.5 data: https://doi.org/10.57760/sciencedb.16319