Evaluating Causal Arctic‐Midlatitude Teleconnections in CMIP6

Abstract

To analyse links among key processes that contribute to Arctic‐midlatitude teleconnections we apply causal discovery based on graphical models known as causal graphs. First, we calculate the causal dependencies from observations during 1980‐2021. Observations show several robust connections from early to late winter, such as atmospheric blocking within central Asia via the Ural blocking and Siberian High, the North Atlantic Oscillation (NAO) phase and the polar vortex. The polar vortex is affected by poleward eddy heat flux at 100 hPa, which is also directly connected with the Aleutian Low. We then evaluate climate models participating in the Coupled Model Intercomparison Project Phase 6 (CMIP6) by comparing their causal graphs with those derived from observations. Compared to observations, CMIP6 historical and future simulations do not robustly capture Arctic‐midlatitude teleconnections arising from Arctic sea ice variability. This highlights the role of atmospheric internal variability in modulating the Arctic‐midlatitude teleconnections. However, we find several distinct patterns that are simulated by most of the analyzed climate models. For example, both historical and future model simulations robustly capture observed atmospheric blocking in central Asia. But contrary to observations, model simulations show a robust link between the Arctic temperature and sea ice cover over Barents and Kara seas. The analysis of future changes also reveals that the connection between the Aleutian Low and the poleward eddy heat flux at 100 hPa is expected to become more robust towards the end of the 21st century than in the analysed past.

Warm Arctic–Cold Eurasia pattern driven by atmospheric blocking in models and observations

“…Our reanalysis results are consistent with prior causal network analyses that identified two-way causality between Barents-Kara sea-ice extent and Ural sea-level pressure at the sub-monthly time scale (Kretschmer et al 2016, McGraw andBarnes 2020) and monthly time scale (Galytska et al 2023). However, we additionally find that the causal effect of sea-ice loss is intermittent in observations (based on bootstrap resampling) and not captured in an ensemble of historical simulations.…”

“…Our analysis builds upon the growing body of causal inference studies that highlight intermittent, two-way interactions between Barents-Kara sea-ice extent and midlatitude circulation (Kretschmer et al 2016, Siew et al 2020, Galytska et al 2023. In spite of this intermittency, we identify an atmospheric driver of the Warm Arctic-Cold Eurasia pattern that is robust across climate states in both models and observations.…”

“…This study seeks to identify and quantify causal pathways in the Warm Arctic-Cold Eurasia Pattern. To achieve this goal, we use a statistical causal inference method (Runge et al 2019), which has enabled robust quantification of two-way Arctic-midlatitude interactions in observational studies (Kretschmer et al 2016, Siew et al 2020 and allowed for more direct model-observation comparison without targeted perturbation experiments (Galytska et al 2023). Building on these recent advances, our study uses causal inference to compare observed linkages with historical simulations in a fully coupled large ensemble, separating forced responses from internal variability.…”

Causally‐Informed Deep Learning to Improve Climate Models and Projections

“…As a further test of the credibility of our causal feature selection methodology and its stability with a changing input distribution, we also explore its sensitivity to climate change (Galytska et al., 2023; Karmouche et al., 2023). Thermodynamic features driving different atmospheric processes are “climate invariant”, that is, they govern the same processes regardless of the climate state of the system, as physics does not change with climate change.…”

“…Our selection algorithm is the PC 1 phase of the PCMCI method (Runge, Nowack, et al, 2019), which is based on the PC algorithm (Spirtes & Glymour, 1991), and the Momentary Conditional Independence (MCI) test. PCMCI, and its different flavors, have been widely used in recent years in climate sciences, such as for better understanding teleconnections in the Earth system (Kretschmer et al, 2016(Kretschmer et al, , 2018Runge et al, 2014;Siew et al, 2020) and their pathways (Galytska et al, 2023;Karmouche et al, 2023;Kretschmer et al, 2021;Runge et al, 2015) or to investigate land-atmosphere interactions (Krich et al, 2020).…”