Interdependent Extratropical Atmospheric Responses to Arctic Sea Ice Loss, QBO, and ENSO

The quasi-biennial oscillation (QBO) and El Niño–Southern Oscillation (ENSO) both influence the strength of the stratospheric polar vortex, with weakening of the polar vortex during the easterly QBO and El Niño. The polar vortex is also thought to weaken in response to Arctic sea ice loss, related to an equatorward shift of the midlatitude jet stream. However, it is unclear how these factors impact the polar vortex in combination and whether their influences are linearly additive. Here, using atmospheric general circulation model experiments, we show that in our model, in the presence of Arctic sea ice loss, the polar vortex weakens significantly during easterly QBO, but not during westerly QBO and weakens significantly during neutral ENSO, but not during El Niño. This suggests that the polar vortex responses to sea ice loss, QBO, and ENSO are interdependent and nonadditive. This interdependence cannot be explained by the polar vortex background state, since the easterly QBO and El Niño both separately weaken the polar vortex but have an opposite-signed influence on the response to sea ice loss. Instead, we surmise the modulation is governed by how QBO and ENSO affect anomalous wave propagation into the stratosphere in response to sea ice loss. At the surface, sea ice loss enhances the Siberian high and causes cooling over Eurasia in QBO easterly phase (QBO-E) but not QBO westerly phase (QBO-W). Both QBO-W and El Niño damp the equatorward tropospheric jet shift in response to sea ice loss, relative to QBO-E or neutral ENSO, respectively. The existence of these nonlinearities has implications for subseasonal-to-interannual climate prediction.
Significance Statement

This work aims to understand how the atmospheric changes triggered by sea ice loss in the Arctic will differ depending on the background state of the atmosphere. We find that the response to sea ice loss is stronger under certain conditions, meaning sea ice loss is more impactful to humans in years where these conditions are present. The results of this paper have implications for accurate climate prediction, particularly on the expected impacts of climate change.