The decline of Arctic sea ice in recent decades is closely associated with amplified warming and the heightened occurrence and intensity of storms in the region. Atmospheric Rivers which are usually associated with an extratropical cyclone, have the potential to impact Arctic sea ice through both thermodynamic and dynamic mechanisms. However, our current understanding of the evolving nature of ARs in a warming climate and their specific influence on sea ice remains incomplete.

To address these knowledge gaps, we employed the Atmospheric River Detection and Tracking algorithm (ARDT) that utilizes meridional integrated vapor transport to identify ARs in the Arctic within the Community Earth System Model, Version 2 (CESM2). We evaluated CESM2's capacity to simulate intense ARs by comparing it to MERRA2 reanalysis data. Our findings demonstrate that CESM2 generally captures the patterns and prominent hotspots of ARs, exhibiting agreement with MERRA2 data during the 1980-2015 period. Furthermore, we investigated changes in AR behavior under the SSP370 scenario, which assumes high greenhouse gas emissions.

We explore three different methods to detect ARs under a changing climate scenario by modifying the minimum threshold criteria of the AR detection algorithm. 1. Defining extremes based on present climate thresholds; 2. Scaling the threshold with projected changes in moisture for the future; and 3. Calculating a unique threshold for each decade. Our results reveal an increasing trend in AR occurrence for all the three methods, during winter, fall, and spring, accompanied by intensified ARs throughout all seasons.

Moreover, we assessed the impacts of ARs on sea ice for all the three different methods and consistently observed a net sea-ice loss impact through both dynamic movement of ice and thermodynamic heating. Overall, our findings suggest that the escalating frequency and intensity of ARs could exacerbate the decline of Arctic sea ice, leading to significant consequences for the Arctic ecosystem and global climate.