Warming shifts North Atlantic pressure centers north across 100 runs of one climate model

The design is a clean one. Every one of the 100 runs receives identical greenhouse gases, aerosols and volcanic eruptions, and each starts from slightly different initial conditions ^[5]. Whatever all the runs do together is the forced response. Whatever differs between them is the atmosphere's own noise ^[5]. A single simulation or a short observational record cannot make that split, and the authors say such approaches blur the signal for something as noisy as a weather regime ^[6].

The thing this doesn't tell you is whether CESM2 gets the forced response right. All 100 members are the same model ^[4]. Averaging across members removes weather noise. A bias in the model's physics is present in every member, so averaging leaves it in place.

The timing comes from a statistical changepoint test that found a clear signal emerging around 1995 in mid-tropospheric circulation patterns ^[7]. The team then compared regimes before 1995, when human influence was weak, with regimes after it, when that influence was dominant ^[7]. A winter climatology built on 1980 to 2020 would average across both periods ^[7].

With the forced signal left in, the region keeps the same number of regimes after 1995 ^[8]. Their pressure centers move north, some circulations intensify, and the balance between patterns changes ^[9]. In the most frequent regime, the low-pressure region sits clearly further north after 1995 ^[10]. "Climate change does not destroy the regimes; it reshapes them," the authors wrote ^[11].

Remove the forced signal and the leftover natural variability has fewer regime states after 1995, with one dominant regime much more persistent ^[12]. The authors take this to suggest that warming is suppressing part of the atmosphere's natural variability ^[12].

The North Atlantic Oscillation is the pressure seesaw between the subtropical Atlantic and Iceland that governs wind strength and storm tracks across the region ^[14]. In the forced runs it leans toward its positive phase for most of this century, a shift the authors call "subtle but detectable" ^[1]. It also swings less ^[13]. In the authors' summary, positive phases "are associated with stronger westerly winds, milder winters in northern Europe and wetter conditions in some regions" ^[2]. Associated is the right word here: the summary links the phase to those conditions without measuring how much of each winter it explains.

For a storm-risk model, I would treat the positive tilt as a direction to test against observed winters before using it. The authors' summary does not give the size of the shift, or say what the reorganized regimes mean for the eastern North American cold spells they list among the regimes' impacts ^[1][15].