Stratosphere-Troposphere Coupling: Stratospheric Polar Vortex Extremes

Last update: 2023-08-22

This POD assesses stratospheric polar vortex extremes, and the tropospheric circulation and surface patterns that precede and follow them. Extremes in the stratospheric polar vortex are closely linked to the tropospheric circulation and surface climate both before and after the event. The occurrence of polar stratospheric circulation extremes in the Northern Hemisphere (NH), such as sudden stratospheric warmings (SSWs) and polar vortex intensifications (VIs), are important aspects of stratospheric variability that rely on realistic representations of the stratosphere and the troposphere. Extremes in the strength of the Arctic polar stratospheric circulation are often preceded by known near-surface circulation patterns, and then subsequently followed by shifts in weather patterns (sometimes for weeks). SSWs in the Southern Hemisphere (SH) are rare (only one event in the satellite record), while VIs occur more often, but both events can have persistent impacts on SH mid-latitude weather.

The definition for SSW events used in this POD is the most commonly used one (Charlon and Polvani 2007): a reversal of the 10 hPa 60 deg latitude daily-mean climatological westerly zonal winds between November and March, which returns to westerly for at least 10 consecutive days prior to 30 April (so that final warmings are not included). SSWs are independent events if they are separated by at least 20 days of consecutive westerlies. The definition for VI events used in this POD is adapted from previous studies (Limpasuvan et al. 2005, Domeisen et al. 2020): an increase of the 10 hPa 60 deg latitude daily-mean zonal-mean zonal winds above the daily 80th percentile value calculated across the full input data time period, which persists for at least 10 consecutive days. VIs are independent events if they are separated by at least 20 consecutive days below the 80th percentile.

Models often show a different SSW seasonality compared to reanalysis (Ayarzaguena et al. 2020), and may vary in their simulation of tropospheric circulation/surface patterns both preceding and following extreme stratospheric events (Ayarzaguena et al. 2020). Models with low model lids (>1 hPa in pressure) may show less persistent downward coupling than models with higher model lids (Charlton-Perez et al. 2013). SSWs and their precursor patterns and impacts have been heavily studied (Baldwin et al. 2021), but VIs less so (Limpasuvan et al. 2005).

This POD makes three kinds of figures from provided model data:

  • Barplots showing the frequency of events by month over the input period

  • Pressure versus lag contour plots of polar cap geopotential height anomalies, composited around all detected SSWs and VI events. These types of plots are sometimes referred to “dripping paint” plots in the scientific literature.

  • Polar stereographic maps of surface air temperature and 500 hPa geopotential height anomalies averaged over the 30 days before and after all detected SSW and VI events.

Additionally, the POD outputs text files of the detected SSW and VI event dates in each hemisphere. These plots are made for both hemispheres, and require at least one event to be detected in order for the POD to create the figure.

Version & Contact info

  • Version/revision information: v1.0 (Aug 2023)

  • Project PIs: Amy H. Butler (NOAA CSL) and Zachary D. Lawrence (CIRES/NOAA PSL)

  • Developer/point of contact: Amy Butler (


This POD is driven by the file, with a helper script of The driver script reads in the model fields, performs a few preparatory actions such as averaging the geopotential heights over the polar cap and removing the daily climatology to obtain anomalies, and selecting the 10 hPa zonal-mean zonal winds. The script then creates three plots (in both hemispheres, so 6 plots in total) and outputs to text files the SSW and VI dates.

The observational data this POD uses is based on ERA5 reanalysis (Hersbach, et al., 2020), and includes pre-computed daily-mean zonal mean zonal winds and geopotential heights (with dimensions of (time,lev,lat)), and gridded daily-mean 500 hPa geopotential heights and surface air temperatures (with dimensions of (time,lat,lon)).

Required programming language and libraries

This POD requires Python 3, with the following packages:

  • numpy

  • pandas

  • datetime

  • xarray

  • matplotlib

  • statsmodels

  • cartopy

  • scipy

Required model output variables

The following daily-mean fields are required:

  • Zonal-mean zonal wind, ua as (time,lev,lat) (units: m/s)

  • Zonal-mean geopotential heights, zg as (time,lev,lat) (units: m)

  • Geopotential Heights at 500 hPa, zg as (time,lat,lon) (units: m)

  • Surface air temperature, tas as (time,lat,lon) (units: K)


Charlton, A. J., and L. M. Polvani, 2007: A new look at stratospheric sudden warmings. Part I: Climatology and modeling benchmarks. Journal of Climate, 20, 449–469.

Limpasuvan, V., D. L. Hartmann, D. W. J. Thompson, K. Jeev, and Y. L. Yung, 2005: Stratosphere-troposphere evolution during polar vortex intensification. Journal of Geophysical Research, 110, D24101,

Domeisen, D. I. V., and Coauthors, 2020: The role of the stratosphere in subseasonal to seasonal prediction Part I: Predictability of the stratosphere. Journal of Geophysical Research: Atmospheres, 125, e2019JD030920,

Ayarzagüena, B., and Coauthors, 2020: Uncertainty in the Response of Sudden Stratospheric Warmings and Stratosphere-Troposphere Coupling to Quadrupled CO2 Concentrations in CMIP6 Models. Journal of Geophysical Research: Atmospheres, 125, e2019JD032345,

Baldwin, M. P., and Coauthors, 2021: Sudden Stratospheric Warmings. Reviews of Geophysics, 59, e2020RG000708,

Hersbach, H. and coauthors, 2020: The ERA5 global reanalysis. Q J R Meteorol Soc., 146, 1999-2049,

More about this POD

Confidence intervals for frequency of events

This POD calculates the total frequency of SSW and VI events over the input period, and then determines what fraction of those events occur in each month of the winter season. Because the event either occurs or doesn’t in any given month, we calculate the binomial proportion confidence interval using the Wilson score interval, for the 95% level.

Significance for vertical composites

The lag-pressure composites (“dripping paint”) plots provided by this POD show the composite-mean values of standardized polar cap geopotential height anomalies. In these figures, significance is evaluated at the 95% level using a one-sample t-test, and assumes that the population mean has an anomaly value of 0 and that the sample mean comes from a normally distributed population. This may not be a robust assumption, but here this test is chosen for a computationally inexpensive estimate of significance. In these plots, values that are insignificant by this test are stippled.