Mesoscale eddies in the Arctic Ocean modulate the transport of heat, freshwater, salt and nutrients, playing a key role in Arctic Ocean circulation and increasingly recognized as hotspots of biogeochemical activity and marine productivity. Nevertheless, their observation and modelling remain challenging due to sea-ice coverage, polar nights, strong seasonal variability and the small Rossby radius (7–30 km compared with 20–100 km at mid-latitudes). Recent advances in high-latitude altimetry (e.g., CryoSat-based products), SWOT observations and in situ measurements (e.g., MOSAiC, Ice-Tethered Profilers and Fram Strait moorings) provide an unprecedented opportunity to evaluate high-resolutionArctic Ocean models.
Although recent studies analysed Arctic vortices using SWOT observations, FESOM/FESOM2 simulations or NEMO CREG12, studies using the ultra-high-resolution CREG36 configuration remain absent or underrepresented. This study addresses this gap by evaluating NEMO’s capability to reproduce surface mesoscale eddies across the Arctic Ocean (60–90°N) and quantifying the impact of increasing horizontal resolution from 1/12° to 1/36°. 
The analysis focuses on January 2020–June 2022, when both NEMO simulations and satellite observations are available. Validation is based primarily on the new regional DUACS Arctic altimetry product and supplemented by the global AVISO product, providing a dual observational benchmark. SWOT 1/8° daily observations are further used to compare eddy population statistics. Although not overlapping temporally with simulations, SWOT data highlight the benefit of new high-resolution altimetry for studying Arctic mesoscale dynamics.
Datasets were remapped onto a common grid and temporally harmonized, using SSH for NEMO simulations and ADT for satellite products. Daily sea ice masks combining native NEMO and AVISO masks were applied to all datasets to limit analysis to open-ocean regions and ensure comparable domains. After detection, features identified as sea ice or land were quantified to estimate their impact on eddy detection. Eddy detection and tracking were performed using PyEddyTracker. Parameters were based on existing literature and adapted to Arctic conditions and dataset scales. The same configuration was applied to model and observational datasets, including a 500-km Bessel high-pass filter, 0.01-m contour interval, and minimum eddy lifetime of 6 days.
A first exploratory analysis indicates that CREG36 detects substantially more mesoscale eddies than CREG12 while preserving the main spatial patterns of Arctic eddy activity. The complete validation against satellite observations will quantify whether this increase reflects improved mesoscale dynamics representation and assess the added value of the higher-resolution configuration.