Analysis of gravity wave activity during stratospheric sudden warmings in the northern hemisphere

1.   Introduction

Gravity waves (GWs) are one of the common disturbances in the middle and upper atmosphere; they play a vital role in atmospheric motion (Fritts and Alexander, 2003; Chang SJ et al., 2020; TengChen KM et al., 2021; Yi W et al., 2023a). They are an important factor affecting changes in stratospheric wind and temperature fields. The breaking of upward propagating GWs is the main source of small-scale turbulence and mixing in the middle and upper atmosphere (He Y et al., 2020a). The momentum flux caused by the generation, propagation, and breaking of GWs affects atmospheric circulation (Bian JC et al., 2005; Li T et al., 2007; Gu SY et al., 2020; Ji QQ et al., 2022; Yi W et al., 2021, 2023b).

In recent years, GWs have been studied using different instruments and methods, such as ground-based radar observation, observation from meteorological rockets, radiosonde and satellite observation, and the application of deep learning (Xiao CY and Hu X, 2010; Chang SJ et al., 2019; He Y et al., 2020b, 2020c; Xue XH et al., 2020; He Y et al., 2021, 2022a, 2023; Wu Y et al., 2022; Zhang J et al., 2023). Satellite observation is superior to ground observation because the area measured by a satellite is often larger than can be observed by ground instruments, and some satellite observations also provide better temporal and vertical resolution than ground-based systems (Jin S and Feng GP, 2011; Jin S et al., 2013). A large number of satellite-mounted instruments have been used to observe temporal and spatial distribution characteristics of GWs (Wu DL and Waters, 2013; He Y et al, 2022b; Ge W et al., 2023). The distribution characteristics of global GW activity have been studied, and the impact of background wind field on GWs has been discussed (Tsuda et al., 2016; Jia MJ et al., 2018; Wu JF et al., 2018; Ge W et al., 2019; Yi W et al., 2019; He Y et al., 2020d; Ern et al., 2022). Although results obtained by data from these satellites are sufficiently detailed, there are still deficiencies, such as low spatial resolution and low data density, leading to limited knowledge of spatial distribution and temporal variation of GWs. The Constellation Observing System for Meteorology, Ionosphere and Climate (COSMIC) mission provides data with global coverage and high vertical and temporal resolution. Therefore, COSMIC data are particularly suitable for studying the activities of global GWs (Wang L and Alexander, 2009; Wright et al., 2011; Mcdonald, 2012).

Stratospheric sudden warming (SSW) is a phenomenon that occurs in the polar regions of the winter hemisphere. Normally, the temperature in the winter hemisphere's stratosphere gradually decreases as the latitude increases, and the zonal mean zonal winds in the polar stratosphere are eastward. However, in certain winters, the state of polar regions is severely disturbed. The temperature of the polar stratosphere can increase rapidly in a short period of time; at an altitude of 10 hPa near the pole, the temperature can rise by 40−60 K within a week. As the temperature rises, the eastward polar vortex is weakened, and winds even reverse from eastward to westward (Harvey et al., 2002).

The SSW phenomenon was discovered by German scientist Richard Scherhag in 1952 when analyzing radiosonde data (Scherhag, 1952). Radiosonde observations provided useful data on the troposphere and stratosphere during SSW events. Later, probe rocket observations added data that extended to the height of the mesosphere. Today, satellite observation data provide ideal spatiotemporal coverage, making it possible to study the global impact of SSW events (Zhou B et al., 2023).

Recent observations and theoretical investigations have shown that the effects of SSW events are global (Laskar et al., 2019; Kogure et al., 2021; Watanabe et al., 2022). SSWs can not only alter the temperature structure of the polar atmosphere and the polar night jet, but can also alter the meridional circulation, thereby affecting mid to low latitudes (Laskar et al., 2019). The climatology of stratospheric polar vortices and anticyclones has been investigated (Harvey et al., 2002). Therefore, studying changes in atmospheric parameters on a global scale during SSW events can help us understand the dynamic processes of the atmosphere.

The SSW is a dynamical phenomenon of the wintertime stratosphere, caused by interaction between the stratospheric zonal-mean flow and planetary rossby waves propagating from the troposphere. There are two different types of such warmings, identified as the vortex displacement type (dominant wave number is 1, i.e. wavenumber 1 [WN1]), and the vortex splitting type (dominant wave number is 2, i.e. wavenumber 2 [WN2]) (Charlton and Polvani, 2007; Seviour et al., 2013; Butler and Domeisen, 2021). The occurrence of an SSW is mainly due to interaction in winter between enhanced planetary waves (PWs) and the atmospheric background mean flow. SSWs are very common in the northern hemisphere, where PWs are enhanced by the interaction between sea and land; in the southern hemisphere, where middle and high latitudes are mostly oceanic, lack of sea−land interactions results in many fewer PWs and SSWs. The rapid changes of temperature and wind in the upper and middle atmosphere of the winter hemisphere that characterize an SSW have significant impact on the atmospheric circulation (Tang L et al., 2023). During SSWs, the temperature of the polar stratosphere increases rapidly — within a few days, accompanied by weakenings of the polar vortex and easterly winds. The breaking of polar vortex and the reverse of winds happen even during the main warming period. In high latitudes, the main dissipation region of GWs is above the stratosphere. Due to the significant changes they make in the background atmosphere of the high latitude stratosphere over a short period of time, SSWs can significantly affect the propagation and energy distribution of GWs. Reduced GW activity was observed during the 2009 major SSW event (Kalisch and Chun HY, 2021). During the September 2019 SSW event in the Southern Hemisphere, strongly negative vertical fluxes of zonal momentum in the stratosphere were observed around the edge of the polar vortex (Minamihara et al., 2022). Positive feedback between GW forcing and the state of the polar vortex has been discussed (Wicker et al., 2023); they report that the associated polar stratospheric cold bias following SSW events in sub-seasonal hindcasts can be halved by increasing vertical model resolution, suggesting a potential sensitivity to GW forcing. PWs are considered the main driving force of SSWs. However, theoretical studies have suggested that GWs can interact with background circulation and contribute to the occurrences of SSWs (Cullens and Thurairajah, 2021). More work is needed to study the characteristics of GW activities during SSWs to help elucidate the interaction mechanisms between SSWs and GWs.

The purpose of this study is to use COSMIC radio occultation (RO) temperature data to analyze the characteristics of GW activity in the Northern Hemisphere during SSWs. The results can serve as a reference for future work studying interaction mechanisms between SSWs and GWs.

2.   Data and Methods

The COSMIC mission was launched on April 15, 2006. It consists of six low-earth-orbit satellites with an inclination of about 72° and an interval of about 30°. Radio occultation (RO) occurs when tracking occludes global positioning system (GPS) satellites above or below the earth's horizon. Theoretically, COSMIC can obtain about 3000 RO profiles every day, but some profiles cannot obtain reasonable atmospheric physical quantities; roughly 2000 profiles a day can be used to calculate temperature data. COSMIC observation data are sufficient to achieve global, all-weather coverage with high vertical resolution.

We selected dry temperature profiles of COSMIC RO level 2 data collected from January to March in 2014−2016. The GW activity is evaluated by the potential energy density (${E}_{p}$) calculated from each profile. The altitude resolution of COSMIC temperature profile data is a few tenths of 1 km. Therefore, to achieve temperature profiles at altitude intervals of 0.1 km, we have interpolated temperatures from COSMIC data as needed. We extract GW disturbances as follows:

(1) within each latitude interval of 5°, daily temperature profiles are arranged by increasing longitude;

(2) least squares harmonic fitting is performed at each altitude (Zeng XY et al., 2017; Zeng XY et al., 2021) with a zonal wavenumber of 0−6, to obtain background temperature and large-scale fluctuations with zonal wavenumber of 1−6;

(3) background and large-scale fluctuations are subtracted from the original temperature to obtain the temperature residual;

(4) a sixth-order Butterworth bandpass filter with band width of 3−10 km is used to filter out large-scale vertical wavelength fluctuations, so as to obtain GW disturbance;

(5) the obtained GW disturbance is used to calculate the ${E}_{p}$.

The steps above allow interference of tides or PWs to be basically eliminated from the ${E}_{p}$ calculation.

${E}_{p}$ is calculated as follows:

where $g$ is the acceleration due to gravity, $\overline{T}$ is the background temperature, $z$ is the altitude, ${c}_{p}$ is the isobaric heat capacity, $N$ is the Brunt−Väisälä frequency, and $\widehat{T}$ is the temperature amplitude.

The zonal wind distribution during SSWs has also been considered, using data provided by The European Centre for Medium-range Weather Forecast (ECMWF), which are re-analyses of observations.

3.   Results and Discussion

We focus our research in the mid to high latitudes of the Northern Hemisphere in the first three months of 2014−2016, the season when SSWs usually occur (Manney et al., 2015; Gong Y et al., 2018; Ma Z et al., 2020a; Idolor et al., 2021), and analyze the temporal and spatial distribution of temperature and GW ${E}_{p}$. To create a grid to display temperature and ${E}_{p}$ values, we divided the global projection into a grid by longitude × latitude × altitude × day (with intervals of 10° × 5° × 0.1 km × 1 day). Temperature and ${E}_{p}$ profile data within each grid interval were identified and averaged to arrive at values for that interval. Here are the results.

3.1   The Temperature Variation during SSWs

Figure 1 shows the distribution of temperature at the altitude interval of 20−30 km between 30°N and 90°N latitude from January to March of the three years 2014 to 2016, as seen looking down from above the North Pole. Significant SSW events occurred each year in those months. From the perspective of the Northern Hemisphere, high-temperature and low-temperature regions alternated when SSWs began. Between days of the year (DOY) 040 to 052 in 2014, when an SSW event occurred (Idolor et al., 2021), temperature increased near 60°W−120°W. The other warming occurred near 60°E on DOYs 004−020 in 2015, and near 180°E DOY 036−040 in 2016, when SSW events also happened (Manney et al., 2015; Gong Y et al., 2018). We can see an interesting phenomenon that the area of high temperatures expanded to cover the polar region on DOY 40 in 2014 and DOY 004 in 2015. Besides, late in March of 2014 and 2016, the temperature in the mid to high altitude areas of the entire Northern Hemisphere increased.

Figure  1.  Images, overlooking the North Pole, show distribution of temperature, from 30°N to 90°N latitude, between January and March in the three years 2014−2016. Longitude is 180° at the top to 0° at the bottom of each image. Each temperature distribution image is labeled by year and day of the year.

DownLoad: Full-Size Img PowerPoint

The images in Figure 1 show temperature distribution characteristics during the first three months of 2014 to 2016, when SSWs frequently occurred. In the pre-warming period (before the onset or central date of an SSW), the temperature structure can indicate a dominant wave component (e.g., waves 1 or 2). In the post-warming period of an SSW, high temperatures can be transported to another hemisphere (e.g., from 120°E to 60°W) via the polar region, showing temperature enhancements in two longitudinal regions. This longitudinal dependence is caused by the anticyclones (Ma Z et al., 2020a). The average temperature in the mid to high altitude areas of the entire Northern Hemisphere increased late in March of 2014 and 2016. Those are stratospheric final warmings without a recovering of zonal mean zonal winds (Yamazaki et al., 2019). Final warmings in 2014−2016 are related to wave 1 (Butler and Domeisen, 2021). As a consequence, each year there was only one high temperature zone during final warmings.

3.2   Vertical Propagation of GWs during SSWs

To understand the variation of GW activities during SSWs, we first analyzed the vertical propagation of GWs. Figure 2 shows the longitude-altitude distribution of temperature (colored contour map) and GW ${E}_{p}$ (the red line marks the value ${E}_{p}$ = 3 J/kg) in the Northern Hemisphere from January to March in 2014−2016. We can see that when SSWs occurred, obvious enhancements of GW ${E}_{p}$ were observed. The enhancement of GW ${E}_{p}$ has an eastward and downward extension. We can see that during the period of SSWs in 2014−2016 mentioned in Section 3.1, the enhanced area of GW ${E}_{p}$ expanded and covered a larger range of longitude and altitude. The altitudes at which ${E}_{p}$ values exceeded 3 J/kg expanded down to below 20 km. The vertical propagation of GWs proved to be significant, which may have been caused by the GW sources and the wind field (Cullens and Thurairajah, 2021).

Figure  2.  Longitude−altitude distribution of temperature and GW ${E}_{p}$ at 50°N−70°N from January to March in 2014−2016 (the day of year is marked on each image; the GW ${E}_{p}$ value of 3 J/kg is marked as the red line).

DownLoad: Full-Size Img PowerPoint

3.3   The Longitude and Latitude Distribution of GWs

In addition to the vertical variation, the horizontal spatial variation of GWs is also significant. Figure 3 shows the longitude-latitude distribution of temperature and GW ${E}_{p}$ within 20−30 km at 30°N−90°N from January to March in 2014−2016. GW ${E}_{p}$ values are drawn as a gray contour map. The temperature values of 205, 210, 220, 225 K are marked as the purple, blue, orange, and red lines. During the period of SSWs in 2014−2016 mentioned in Section 3.1, enhanced areas of GW ${E}_{p}$ are shown to expand and cover a larger range of longitude and latitude. Usually, GWs are enhanced at the junction of high (marked by the red line) and low (marked by the blue line) temperature zones, sometimes slightly closer to the high temperature zone. The possible reason is that the area locates at the edge of the polar vortex where the background wind was strong, and the GWs were enhanced (Jia Y et al., 2015; Cullens and Thurairajah, 2021).

Figure  3.  Longitude−latitude distribution of temperature and GW ${E}_{p}$ at 30°N−80°N from January to March in 2014−2016 (Images are labeled as in Figure 1; the temperature values of 205, 210, 220, 225 K are marked as the purple, blue, orange, and red lines, respectively).

DownLoad: Full-Size Img PowerPoint

To illustrate the onset and end of each SSW event, we show (Figure 4) the daily variations of the zonal mean temperature and zonal wind associated with the three events of (1) 9 February 2014, (2) 26 January 2015, and (3) 8 February 2016 (mentioned by Ma Z et al., 2020b). At the onset of SSWs, the zonal mean temperature increased to over 230 K. The longitudinal area with zonal mean temperature higher than 220 K expanded from west of 90°W to east of 90°W. During SSWs, the ${E}_{p}$ usually increased, while partially filtered by the westerly zonal winds. When the zonal winds weakened or turned westward, ${E}_{p}$ began to strengthen. After SSWs, the ${E}_{p}$ usually decreased.

Figure  4.  (a, c, e) Longitude-DOY distribution of temperature (K) at 60°N−90°N from January to March in 2014−2016; (b, d, f) the same as (a, c, e) but for zonal wind (m/s) (the values of 2.2 and 3 J/kg are marked as the orange and red solid lines, respectively; the zonal wind of 0 m/s is marked as the white dashed line).

DownLoad: Full-Size Img PowerPoint

Of course, vortex dynamics/type evolutions during SSWs are not in evidence in the three years from 2014 to 2016. The time period we have chosen to investigate and display did not cover all SSW events in those years, nor the complete process of each SSW event included in our study. SSWs in these three years were all minor ones, and the vortices should be the displacement type (Ma Z et al., 2020b). More work can be done in the future to analyze GW behavior during different types of SSW events.

4.   Conclusions

The SSW is one of the most severe atmospheric disturbances in the winter hemisphere. The drastic changes in temperature and wind during SSWs are believed to be the main reasons for abnormal enhancement of atmospheric wave energy in the upper and middle hemispheres observed during winter. Using temperature data provided by COSMIC, we have analyzed global GW activity at 15−40 km in the Northern Hemisphere from January to March in 2014−2016, during periods when SSWs tend to occur. Our findings of the distribution and variation characteristics of temperature and GW ${E}_{p}$ in longitude, latitude, and altitude, have been presented here and discussed. The conclusions are as follows:

(1) During warming period, the stratospheric temperature was enhanced in one or two longitudinal regions, and the heating area extended to the east of 90°W.

(2) During SSWs, the enhanced area of GW ${E}_{p}$ expanded and covered a larger range of longitude and altitude. Significant enhancement of GW ${E}_{p}$ was observed with an eastward and downward extension during SSWs.

(3) During SSWs, the ${E}_{p}$ usually increased, while partially filtered by eastward zonal winds. When the zonal winds weakened or turned westward, ${E}_{p}$ began to strengthen. After SSWs, the ${E}_{p}$ usually decreased.

Our results can serve as a reference for future work analyzing the interaction mechanisms between SSWs and GWs. More work is needed to understand the detailed mechanisms of GW changes during SSWs and whether the changing GWs have a driving effect on SSWs. Widespread deployment of satellites and radar will provide richer observational data related to the propagation and evolution of GWs. These data can be expected to further reveal details of the excitation processes and activity patterns of GWs in the middle and upper atmosphere during SSW periods.