Topographic effects of Svalbard on warm and moist air intrusions into the Central Arctic

Abstract. Warm air intrusions (WAIs) along the North Atlantic pathway are key drivers of warm extremes in the central Arctic. The Svalbard archipelago acts as a major topographic barrier in the middle of this gateway, but its role in modulating WAIs and their impacts has not been studied in detail. We combine (i) high-resolution regional ICON simulations with and without Svalbard’s topography, (ii) Lagrangian back-trajectories, and (iii) observations from the MOSAiC expedition to analyze a strong WAI event in mid-April 2020, and extend the analysis with (iv) climatological composites from an ICON simulation for 2000–2022. Based on the April 2020 case study, we show that Svalbard's influence can extend ∼ 500 km downstream over sea ice and was observed near 84° N during MOSAiC. The response depends on the static stability of the impinging flow: stable conditions favor flow-around response, characterized by accelerated barrier winds along the eastern and western flanks of Svalbard and gap winds through the Hinlopen Strait, together with a broad lee wake north of the archipelago. In this wake, wind speed, near-surface temperature, and column-integrated water vapor are reduced by > 5 m s−1, > 3 K, and > 1 kg m−2, respectively. Under less stable flow-over conditions, föhn signatures yield lower-tropospheric warming (> 1 K) and drying, reduced low-level cloud cover (> 20 %), and decreased (increased) downwelling longwave (shortwave) radiation (> 20 W m−2). Springtime composites reveal that these signals recur during southerly advection events, can extend several hundred kilometers into the central Arctic, and vary in character with poleward wind speed, moisture transport, and static stability linked to the synoptic situation. Together, the results demonstrate that Svalbard's topography systematically modulates the dynamical and thermodynamic imprint of WAIs, with effects detectable far downstream in both model experiments and MOSAiC observations.