Over the last four years (where does the time go?!), I’ve been working extensively on understanding weather regimes over North America. With our new paper recently published in Journal of Climate, I thought now might be a good time to link together all the publications in one short blog post.
Weather regimes represent recurrent, persistent and quasi-stationary states in the atmosphere, typically lasting for about a week (although sometimes much longer; instances of the same regime for >40 days have occurred). Regimes give a more complete picture than looking at individual modes of variability alone (like the PNA or NAO). Broadly speaking, the concept of regimes is more popular in Europe (something I hope these papers play a small part in changing). But in contrast to the traditional “zoo of indices” needed to describe flow patterns over North America, regimes are much easier to deal with, especially at subseasonal lead-times.
“Wintertime North American weather regimes and the Arctic stratospheric polar vortex” (Lee, Furtado, Charlton-Perez; GRL, 2019)
Written during my summer 2019 visit to the University of Oklahoma, this short paper focuses on understanding the relationship between the strength of the lower-stratospheric polar vortex (100 hPa 60°N zonal winds split into terciles) and the likelihood of four wintertime regimes over the North American continent. We built on earlier work justifying the use of four regimes and termed our regimes “Arctic High”, “Arctic Low”, “Alaskan Ridge” and “Pacific Trough”. We found that the Arctic High regime is about seven times more likely when the vortex is weak versus strong, while the Arctic Low and Pacific Trough regimes are about twice as likely when the vortex is strong versus weak. In contrast, while the Alaskan Ridge regime is the one most likely to bring extreme cold to the largest area of North America, it is equally likely for weak or strong polar vortex states — thus, for North America, “weak vortex” does not equal extreme cold.
“How do stratospheric perturbations influence North American weather regime predictions?” (Lee, Charlton-Perez, Woolnough, Furtado; J. Climate, 2022)
In this final paper from my PhD, we build a statistical model to explain how changes to the strength of the lower-stratospheric polar vortex can shift a point between different regimes within a 3D principal component space. We then run stratospheric nudging experiments using OpenIFS — a portable version of the ECMWF IFS model — to see how a ‘perfect stratosphere’ forecast shifts the ensemble distribution. The results of the experiment broadly supported the simple linear theory, and implied that while the response to stratospheric perturbations was generic, the overall result depends on the “unperturbed” state (analogy: you always take the same train for 30 minutes, but you board the train at different stops, so you end at different places). The paper also provided a simple explanation for why the Alaskan Ridge regime is insensitive to changes in the polar vortex strength: transition vectors between it and other regimes lie almost orthogonal to the linear influence of the stratosphere. A subset of the data for this study is archived on Zenodo (10.5281/zenodo.4818043).
“Stratospheric downward wave reflection events modulate North American weather regimes and cold spells” (Messori, Kretschmer, Lee, Wendt; WCD, 2022)
Motivated by the similarity between the Alaskan Ridge regime and previously-defined patterns associated with stratospheric wave reflection, this paper develops a simple diagnostic for defining wave reflection “events” that can be easily applied to weather and climate model output without computing complex wave activity fluxes. We then show that these events are associated with a fast (within one week) transition from a Pacific Trough precursor — associated with unusually warm conditions across North America — toward an increased risk of the Alaskan Ridge regime, and thus much colder conditions. Crucially, the precursor warmth yields an even stronger average cooling than merely the post-event anomalies themselves. We also explicitly show that the Alaskan Ridge regime dominates the coldest cold spells in central North America. When combined with the previous two papers on the influence of the vortex strength, we now have a comprehensive view of the downward influence of the polar vortex on regimes over North America. The data for this study are archived on Zenodo (10.5281/zenodo.7126679).
“A New Year-Round Weather Regime Classification for North America” (Lee, Tippett, Polvani; J. Climate, 2023)
Thus far, we had been working exclusively on wintertime regimes, which typically receive more focus due to the more coherent large-scale structures in winter and greater low-frequency variability. But this limits the use of regimes to an arbitrary choice of months, and re-defining regimes for a desired time period can lead to uncertainty with regard to the robustness of the classification or a lack of reproducibility. Motivated by the success of a year-round Atlantic-European regime classification by Grams et al. (2017), we set out to define a set of regimes that can be used year-round. We normalise the 500 hPa geopotential height anomalies by the seasonal cycle in the domain-average standard deviation in order to give equal weight to summer and winter circulation patterns. We then demonstrate using four objective methods that only four regimes emerge as the best choice for year-round use. We term them “Greenland High”, “Alaskan Ridge”, “Pacific Ridge” and “Pacific Trough” to distinguish them from previous classifications and use a consistent naming convention. We also add a new “No Regime” category for states in which the anomalies are not distinct from climatology. In addition, we use the regimes to assess climate variability since 1979, finding that the Pacific Trough has become less frequent overall, while the Greenland High has become much more common during summer.
The year-round classification supersedes the previous season-specific definitions. The regimes data are freely available on Zenodo (10.5281/zenodo.8165164) — please do feel free to use these data (citing both the dataset and the paper).
We have other studies in preparation, and plenty of other ideas for the use of regimes diagnostics. If you’re interested in weather regimes, please reach out!