Synoptic Climate Settings and Moisture Supply for the Extreme Heavy Snowfall in Northern China during 6–8 November 2021

Under current global warming, extreme weather and climate events, such as extreme heatwaves and severe precipitation, are occurring frequently worldwide (Chen D. L. et al., 2020; Herring et al., 2020, 2022; Zhou et al., 2020). Extreme heavy snowfall (EHS), as another example of extreme weather, has drawn attention because of its potentially serious socioeconomic impacts (Zhou et al., 2011; Wang et al., 2016). Moreover, under global warming, EHS is also becoming more and more frequent—a trend that is expected to continue in the future (Ding et al., 2008; Sun et al., 2010; Kawase et al., 2016; Chen H. P. et al., 2020; Quante et al., 2021; IPCC, 2022). In recent decades, EHS has shown an increasing trend in northern China, on both interannual and interdecadal timescales (Zhao et al., 2016; Zhou et al., 2018, 2021), and it easily causes severe damage. For example, extraordinarily frequent and long-lasting snowstorms in January 2008 caused above-normal precipitation, below-normal temperature, and severe icing conditions over central–southern China (Ding et al., 2008; Wang et al., 2008; Wen et al., 2009). Naturally, therefore, more and more attention has been paid to the attribution of such extreme events in China (Zhao, 2020; Sun et al., 2021), and elucidating the mechanism driving EHS in China will be beneficial to forecast of this type of extreme weather.

Abundant atmospheric moisture is one of the necessary conditions for EHS (Guan et al., 2010). For instance, the increased evaporation enabled by Arctic sea-ice loss can provide moisture for EHS over Eurasia (Liu J. P. et al., 2012; Bailey et al., 2021). Thus, exploring the moisture sources will help to comprehensively reveal the mechanism of extreme precipitation. To identify the moisture sources, Lagrangian diagnosis has been successfully applied over several regions during different seasons (Drumond et al., 2011; Jiang et al., 2017; Huang et al., 2018). Sun and Wang (2015) indicated that the moisture source for wintertime precipitation over South China is in the western Pacific, while that for precipitation over the middle and lower reaches of Yangtze River valley and North China is primarily the land areas around East China. Further research on extreme precipitation events over East China using Lagrangian diagnosis has shown that the moisture sources for snowfall and rainfall originate from land areas and sea areas, respectively (Yang et al., 2019). However, the moisture sources for EHS over northern China have been less well studied.

Abnormal atmospheric circulation is also an important condition causing EHS events. For example, Pei et al. (2022) indicated an anomalous anticyclone over the northwestern Pacific and a weakened East Asian winter monsoon (EAWM) contributed to a record-breaking wintertime precipitation event over Beijing. More specifically, Zhou et al. (2021) found that anomalous atmospheric circulation similar to the northern mode of the EAWM may favor the interaction of cold air with moist airflows over northern China, which is conducive to interdecadal increases in localized heavy snowfall. Besides, the North Atlantic Oscillation (NAO) and Arctic Oscillation (AO) also play key roles in climate and weather changes over Eurasia (Cohen et al., 2010; Li and Wu, 2012). In terms of snowfall, some studies have found that the NAO and AO tend to be in a negative phase when more intense snowfall events occur in East Asia (Park et al., 2010; Wang and Zhou, 2018; Sun et al., 2021). Also, the negative phase of the NAO/AO is related to sea-ice loss and a weak polar vortex (Honda et al., 2009; Nakamura et al., 2015). Kim et al. (2014) proposed that Arctic sea-ice loss in early winter causes the upward propagation of planetary-scale waves to the stratosphere, especially over the Barents and Kara Seas, which will lead to a weakening of the polar vortex in late winter and ultimately to a negative phase of the AO. The weakening of the polar vortex can explain 60% of extreme cold events in winter in the Eurasian midlatitudes in recent decades (Kretschmer et al., 2018). In addition, some studies have confirmed that EHS in China is closely related to Arctic sea-ice loss (Liu N. et al., 2012; Sun et al., 2019, 2021).

A record-breaking EHS event occurred in northern China during 6–8 November 2021. As shown in Fig. 1, the snowfall process not only resulted in abnormally high snowfall in northern China in early November 2021, but also abnormally high total snowfall in the whole of November, compared with the climatology in November 1981–2010. There were two maximum snowfall centers, in Tongliao and Tianjin, where the snowfall was abnormally exceeding eight times more than the climatology. Besides, the regions around the two centers had more than twice as much snowfall as normal. Comparing Figs. 1a and 1b, it can be seen that the EHS event resulted in a snowfall anomaly for the whole November 2021. As the center of the EHS, Tongliao City in Inner Mongolia, reported a total snowfall amount (hereafter, snowfall refers to snow water equivalent in this study) that reached 87.4 mm. Disasters were experienced in 8 districts, inflicting direct economic losses of 34.5 million Yuan.

Figure 1.  Distributions of snowfall anomalies (%) in northern China in (a) early November 2021 and (b) November 2021 compared with the climatology in November 1981–2010.

Elucidating the mechanism of typical EHS events is useful in a broader context, to further improve the predictability of extreme snowfall in this region (northern China). Accordingly, in this study, the aim was to reveal the mechanism of this EHS event from the perspective of the two key conditions (cold air and water vapor). Following this introduction, the data and methods applied in our study are described in Section 2. In Section 3, the developments of the snowfall, weather systems, horizontal temperature advection, and moisture supply related to the EHS event are described. Then, in Section 4, the water vapor sources and the dominant factors involved in the development of the weather systems are investigated. In Section 5, the precursory signals of the snowfall are discussed, and then a summary of the study’s key findings is provided in Section 6.

Hourly and monthly data from the fifth major global reanalysis produced by ECMWF (ERA5), with a spatial resolution of 0.25° × 0.25° (Hersbach et al., 2020), were employed in this study. The variables used included specific humidity, horizontal wind vector, air temperature, geopotential height, snowfall, surface pressure, and mean sea level pressure. Meanwhile, the daily snowfall recorded at 205 stations in Northeast China, collated by the China Meteorological Administration (CMA), was also used to verify the reliability of the reanalysis data.

The daily AO index in October 2021 was obtained from the Climate Prediction Center of the NOAA (https://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/ao.shtml).

The change in local relative vorticity is often measured by the local relative vorticity tendency, $ \dfrac{\partial \zeta }{\partial t} $ (Kirk, 2003; Fu et al., 2017), which is an indicator of the development in geopotential height. On the synoptic scale, scale analysis shows that the local relative vorticity tendency equation can be written as

where $ f=2\mathrm{\Omega }\mathrm{sin}\varphi $ is the Coriolis parameter, $ \mathrm{\Omega } $ and $ \varphi $ are earth’s rate of rotation and latitude, respectively, and $ {{\boldsymbol{V}}}_{\rm {h}} $ denotes the horizontal wind vector. Equation (1) shows that the intensification of relative vorticity ($ \zeta $) in a region is mainly controlled by two dynamic processes: absolute vorticity divergence (term A) and absolute vorticity advection (term B).

NOAA’s Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model (Stein et al., 2015) is widely applied in studies of moisture sources (Li et al., 2016; Jiang et al., 2017; Shi et al., 2020; Xu et al., 2020). In this study, the HYSPLIT model is used to identify the moisture sources of the water vapor for this EHS event in northern China (38°–55°N, 113°–135°E). The target region is divided into two areas: North China [NC; purple box (38°–45°N, 113°–119°E)] and Northeast China [NEC; red box (39°–50°N, 119°–130°E)], according to the center of the EHS (Fig. 2e). The two areas are then further divided into 1° × 1° fixed grid points as starting parcels (NC: 48 parcels and NEC: 144 parcels). Using the user-specified matrix function in the model, parcels are released at 5 altitude levels: 500, 1000, 2500, 5000, and 8000 m. These levels approximately correspond to the pressure levels of 900, 850, 700, 500, and 300 hPa, respectively. The parcels are released every 6 hours during the period of the EHS (4 times per day; NC: from 0000 UTC 6 to 1800 UTC 7 November 2021; NEC: from 0000 UTC 7 to 1800 UTC 8 November 2021). Therefore, there are 384 and 1152 trajectories in total for NC and NEC, respectively, at every starting altitude level. Air parcels are integrated backwards at hourly time steps until 240 h (10 days). Furthermore, the variables related to the air parcels, including their position (latitude, longitude, and altitude) and meteorological variables (specific humidity and pressure) are stored for further analysis and diagnosis.

Figure 2.  Spatial distributions of daily snowfall in northern China in the ERA5 data during 6–9 November.

To calculate the contribution and height of moisture sources, the specific humidity and pressure at trajectory points are interpolated to 1° × 1° global fixed grid points. Then, the method proposed by Jiang et al. (2017) is employed, in which the local specific humidity is applied as a weighting factor to reflect the capability of the water vapor supply for the target region. The calculation can be expressed as follows:

where $ {Q}_{\rm {sum}} $ represents the average number of air parcels weighted by the local specific humidity at every level, L (= 5) indicates the number of starting altitude levels , T (= 8 in both NC and NEC) represents the time of air-parcel release, and $ {q}_{i,j} $ is the specific humidity of air parcels in the global fixed grid.

For the average pressure height ($ {p}_{\rm {avg}} $) of moisture sources,

where $ {p}_{i,j} $ is the pressure of air parcels at grid point (i,j) and $ {l}_{i,j} $ indicates the number of times the air parcels passed through that point.

To diagnose changes in water vapor in northern China, the vertically integrated moisture flux Q and the divergence of the moisture flux $ {Q}_{\rm {div}} $ were calculated as follows:

where $ {p}_{1} $= 300 hPa, $ {p}_{\rm s} $ is the surface pressure, g is the acceleration due to gravity, and q is specific humidity.

Atmospheric blocking, known as quasi-stationary high-pressure, is one of the major features of midlatitude weather systems. The blocking highs in key regions of Eurasia, such as the Scandinavian, Ural, and Baikal regions, have an important influence on the weather and climate over East Asia (Li et al., 2017; Hwang et al., 2022; Pang et al., 2022; Song et al., 2022). In order to explore the role of atmospheric circulation anomalies in those regions for this EHS event, the 500-hPa geopotential height was used to calculate the average in the Scandinavian region (red box: 45°–75°N, 5°–40°E), Urals region (orange box: 45°–70°N, 50°–70°E), and Baikal region (purple box: 45°–60°N, 90°–120°E) (Fig. S1), as three atmospheric circulation indexes. Besides, the northeast cold vortex (NECV) index was calculated by the regionally averaged 500-hPa geopotential height in the key regions (black box: 40°–50°N, 120°–130°E) of Northeast China (Liu H. B. et al., 2012) (Fig. S1).

The record-breaking EHS event studied in this paper hit northern China (38°–55°N, 113°–135°E) during 6–8 November 2021. As shown in Fig. 2, the center of the EHS was over NC on 6 November (Fig. 2a), with a maximum snowfall intensity exceeding 24 mm day⁻¹. During 7–8 November, the snowfall center moved towards the northeast, mainly lying in NEC, and its intensity strengthened further to over 50 mm day⁻¹ (Figs. 2b, c). Meanwhile, the station observations from the CMA show that the maximum snowfall occurred in Liaozhong, Liaoning Province (61.7 mm; station ID: 54332) and Qianan, Jilin Province (51 mm; station ID: 50948) on 7 and 8 November, respectively. On 9 November, the snowfall event came to an end (Fig. 2d). The total cumulative snowfall during this process exceeded 85 mm in central Liaoning Province, western Jilin Province, and Tongliao City, Inner Mongolia (Fig. 2e). The station-observed spatial distribution and magnitude of the snowfall were generally consistent with the ERA5 reanalysis data over NEC (Fig. S2).

During the EHS event, the polar vortex was deflected to the Eastern Hemisphere at 500 hPa, and its center was consistently located over Severnaya Zemlya, but its intensity decreased with time (Figs. 3a–d). On 6 November, a deep, positive-tilt 500-hPa trough was situated over northwestern China, and the pathway of the trough-front jet stream followed a strong geopotential height gradient from Inner Mongolia to Northeast China (Fig. 3a). At the lower level, there was convergence in the Beijing–Tianjin–Hebei region, caused by the upward movement along the front below the 500-hPa trough (Fig. S3a). At sea level, the central strength of the Siberian high reached 1051 hPa, and the cold front (with a strong pressure gradient from 1020 to 1040 hPa) was also located over NC (Fig. 3e), which indicates that the cold air had arrived in NC on 6 November. As the trough moved further east, a cut-off low pressure was forming in northern NC, and the front associated with the trough was located over the Yellow–Bohai Sea and NEC (Fig. 3b). At lower levels, convergence in front of the trough (Fig. S3b) contributed to the formation of a Yellow–Bohai Sea cyclone on 7 November (Fig. 3f). The upper cut-off low pressure, combined with the lower Yellow–Bohai cyclone, ultimately formed an NEC cold vortex. The geopotential height of the cold vortex center was 5278 and 5192 gpm on 8 and 9 November, respectively (Figs. 3c, d). With the development of the NEC cold vortex and its equivalent barotropic structure, the cyclone not only narrowed the range of the Siberian high at sea level (Figs. 3g, h) but also further contributed to southeastward cold air advection.

Figure 3.  (a–d) 500-hPa geopotential height (purple contours; gpm) and horizontal wind speed ($ \left|{\mathit{V}}_{\rm{h}}\right| $; color-shaded; m s⁻¹) during 6–9 November 2021 (blue dashed curves: trough lines), and (e–h) mean sea level (MSL; hPa) pressure during 6–9 November 2021.

As the atmospheric circulation developed, the trajectory of the cold air can be reflected in the horizontal temperature advection (Fig. 4). On 6 November, the junction of cold and warm temperature advection was located in front of the trough at 500 hPa (Figs. 3a, 4a), and cold and warm air masses at 850 hPa met directly in central NC (Fig. 4e). Besides, the front of the cold air stretched from the west to 115°E below 600 hPa, which promoted the warm air east of this position to lift (Fig. 4i), coincident with the location of the EHS (Fig. 2a). On 7 November, the frontogenesis region was where the junction of cold and warm advection was located, in western Liaoning, with the trough moving eastwards (Figs. 4b, f). Also, the bottom of the frontogenesis region extended eastwards to 118°E, and the lifting effect of cold air together with the updraft triggered by the Yellow–Bohai Sea cyclone (Fig. S3b) contributed to the uplifting of air in the warm advection regions where the center of the EHS was located (Fig. 2b). With the formation of the NEC cold vortex, its center, with a strong updraft, was located between 120° and 125°E (Fig. 4k). The cold advection behind the cold front was mainly located from Liaoning to eastern Jilin, while the prefrontal warm advection was located from Heilongjiang to western Jilin on 8 November (Fig. 4g). The frontogenesis line had reached 123°E by 8 November (Fig. 4k), and the EHS center was mainly located in the area east of 123°E (Fig. 2c). When the cold vortex matured further, the boundary between the cold and warm advection became less distinct, and the remaining warm advection in the cold vortex was mainly located over northern NEC (Figs. 4d, h). The updraft also became weak as the frontogenesis region weakened or disappeared (Fig. 4l). In conclusion, the center of heavy snowfall was located adjacent to the region of warm air advection, indicating that the frontogenesis region where the lifting of warm air forced by the undercutting of cold air determined the location of the extreme snow center in the event.

Figure 4.  (a–d) 500-hPa horizontal temperature advection (color-shaded; 10⁻⁴°C s⁻¹), temperature (yellow contours; °C), and horizontal wind vector (black vectors; m s⁻¹) during 6–9 November 2021 (green boxes: location of frontogenesis region); (e–h) as in (a–d), but for at 850 hPa; and (i–l) vertical profiles of the average zonal–vertical wind field (vectors) and temperature field (color-shaded; °C) over 39°–45°N during 6–9 November 2021, respectively (y-axis: pressure, hPa; red boxes: extreme snowfall center; green dashed curves: frontogenesis lines; and black shading: terrain).

To explore the water vapor transport during this EHS event, the vertically integrated moisture flux ($ {Q}_{\rm {flux}} $) and moisture flux divergence ($ {Q}_{\rm {div} }$) were used to analyze the moisture transport pathways and sources/sinks. For the EHS in NC on 6 November, it is clear that the Beijing–Tianjin–Hebei region was the moisture transport sink, with an intensity of about −9 × 10⁻⁵ kg m⁻² s⁻¹. Also, the water vapor entered the area from the southeast and west (Fig. 5a). On 7 November, the water vapor sink moved northeast to Liaoning–Jilin–Heilongjiang and arrived in NEC via the Bohai–Yellow Sea. The water vapor flux exceeded 100 kg m⁻¹ s⁻¹, which created water vapor conditions conducive to the EHS in NEC (Fig. 5b). With the formation of the NEC cold vortex, the water vapor sinks were mainly located in Heilongjiang–West Jilin–East Inner Mongolia (Fig. 5c), indicating that the center of the EHS was located in this region (Fig. 2c). On 9 November, the EHS event came to an end as the magnitudes of the water vapor sinks, which were mainly located in northeastern Inner Mongolia, decreased (Fig. 5d).

Figure 5.  Changes in vertically integrated water vapor flux ($ {Q}_{\rm {flux}} $, vectors; kg m⁻¹ s⁻¹) and water vapor flux divergence ($ {Q}_{\rm {div}} $, color-shaded; 10⁻⁵ kg m⁻² s⁻¹) from the surface to 300 hPa during 6–9 November 2021.

The local vorticity budget, which can reflect changes in geopotential height, is an effective indicator for diagnosing the development of a trough or ridge. Using Eq. (1), the changes in local vorticity were roughly divided into the contributions from absolute vorticity divergence (DIV) and advection (ADV) on the synoptic scale. Firstly, the location of the 500-hPa trough front was almost entirely covered by a positive relative vorticity tendency during 6–8 November (Figs. 6a–c), which determined where the center of the trough/vortex would be the next day (Figs. 6b–d). Secondly, the magnitude of ADV (Figs. 6i–l) was larger than that of DIV (Figs. 6e–h) during 6–9 November. Therefore, ADV was more important in determining the movement of the trough/vortex or ridge than DIV, which indicates that horizontal airflow played an important role in guiding the movement of the trough/vortex by conveying vorticity.

Figure 6.  Variations in (a–d) local vorticity tendency (VRT) at 500 hPa and the contribution of (e–h) divergence terms (DIV) and (i–l) advection terms (ADV) during 6–9 November (color-shaded; 10⁻⁹ s⁻²), and geopotential height (hgt, purple contours; gpm).

Having analyzed the changes in water vapor transport during the snowfall event by using water vapor flux and divergence, next, the HYSPLIT model outputs were used to further quantitatively estimate the contributions of moisture sources with Eqs. (3) and (4). According to the spatial distribution of the sum of specific humidity at the grid points that the event passed through, the external water vapor sources for the EHS in NC and NEC were divided into 5 regions: East Asia (20°–45°–60°N, 100°–120°–160°E), Asia–Africa (20°–45°N, 0°–100°E), Eurasia (45°–60°N, 0°–120°E), the Barents and Kara Seas (BKS; 60°–90°N, 0°–120°E), and North Atlantic (30°–60°N, 0°–60°W). During the EHS in NC (Figs. 2a, b), local water vapor accounted for 5.81% of the total water vapor, and the rest (94.19%) was from outside the region (Fig. 7a). That is, the proportion was the highest from Asia–Africa, accounting for 32.43%. The regions in the mid-to-high latitudes accounted for 31.29% of the total, which provided 19.52% and 11.77% of the water vapor in Eurasia and BKS, respectively. The contributions from East Asia and North Atlantic accounted for 18.65% and 9.08% of the total, respectively. As for the height of the moisture sources (Fig. 7b), most of the dominant regions were mainly located at the mid-to-low levels (500–800 hPa), except for East Asia, where sources were concentrated at the near-surface layer (900–1000 hPa).

Figure 7.  (a) Contributions of different moisture sources (kg kg⁻¹) and (b) average pressure height (hPa) for the EHS in NC during 6–7 November. (c, d) As in (a, b), but for the EHS in NEC during 7–8 November.

The contributions of water vapor sources during the EHS in NEC (Fig. 7c) were different from those in NC. The total contribution of mid-to-high latitude water vapor increased to 38.13%, becoming the highest proportion. However, the proportion from Asia–Africa, which was the dominant water vapor source for the EHS in NC, decreased to 11.71% during the EHS over NEC. Meanwhile, the contributions from East Asia and local regions increased to 27.49% and 11.83% of the total, respectively. This may be related to the increase in water vapor associated with the formation of the Bohai–Yellow Sea cyclone at low levels (Fig. 3f). The heights of the water vapor sources during the EHS event were relatively consistent in NC and NEC (Figs. 7b, d); that is, most of the water vapor from East Asia was concentrated in the near-surface layer, and contributions from other areas were also mainly located in the mid-to-low layers.

Analysis of the HYSPLIT model outputs identified differences in the moisture sources for the EHS over NC and NEC. The water vapor in NC mainly came from the mid-to-low levels over the Asian–African region and Eurasian mid–high latitudes, while the source of water vapor in NEC was mainly the near-surface layer over East Asia and local regions. Under the current phenomenon of Arctic amplification, the sea ice in the Arctic region is continually declining (Swart, 2017; Gao et al., 2019). Some studies proposed that Arctic sea-ice loss has contributed to the increase in Eurasian snowfall through the associated abundant moisture supply (Cohen et al., 2012; Liu J. P. et al., 2012; Wegmann et al., 2015; Yeo et al., 2017), and this may also be an important reason why the Eurasian mid-to-high latitudes were one of the main moisture sources for the EHS event in northern China in this study—an idea that needs to be further explored in future research. Besides, abnormal atmospheric circulation systems, such as the NEC cold vortex, constitute another reason for the differences in moisture sources for the EHS event in NC and NEC.

Kolstad et al. (2010) indicated that cold anomalies develop over the Eurasian continent along with a weakening of the stratospheric polar vortex. In October 2021, an abnormal high-pressure ridge (blocking) with northward extension persisted between Canada and Greenland, causing the 200-hPa polar vortex to weaken and shift southwards in the Eastern Hemisphere (Figs. 8a–c). This feature was beneficial to cold-air outbreaks over Eurasia. Song et al. (2018) revealed that negative height anomalies associated with a weakening polar vortex over North Atlantic can propagate towards East Asia along an intraseasonal wave train, ultimately leading to the occurrence of cold events over eastern China by enhancing the strength of the Siberian high and East Asian trough. As shown in Figs. 8d–f, the 500-hPa geopotential height anomaly center also moved from Atlantic to Eurasia in October 2021. In early- and mid-October (Figs. 8d, e), geopotential height anomaly centers were mainly situated over North Atlantic, Scandinavia, and western Siberia, which also constituted a wave train. As the intraseasonal wave propagated downstream in late October (Fig. 8f), positive 500-hPa height anomalies were located in southern Scandinavia and the Lake Baikal region, while negative anomalies were located over North Atlantic and Urals region. Therefore, a southward-shifted polar vortex, along with the establishment and collapse of the Lake Baikal ridge/blocking high, were the basis for cold-air outbreaks during the EHS event.

Figure 8.  (a–c) 200- and (e–f) 500-hPa geopotential height fields (contours; gpm) and anomaly fields (color-shaded; gpm) relative to the climate mean (October 1981–2010) in (a, d) early, (b, e) mid, and (c, f) late October 2021.

Meanwhile, results from three atmospheric circulation indexes were also used to display the development of cold air in the EHS event. As shown in Fig. 9, the atmospheric circulation indices in Scandinavia and the Urals region changed from showing a positive to negative anomaly on 12 and 18 October, respectively. However, the indices showed positive anomalies in the Baikal region during 12–30 October, demonstrating that the Baikal ridge/blocking high had intensified and stabilized. With the movement of cold air from upstream to downstream, cold air continued to accumulate and strengthen until the Baikal ridge/blocking high collapsed on 1 November. Pang et al. (2022) also indicated that a negative Scandinavian pattern can bring more frequent occurrences of blocking over central Siberia (i.e., the Baikal region) via the eastward propagation of wave activity fluxes. But what triggered the collapse of the Baikal ridge/blocking high before the EHS event in this study?

Figure 9.  Daily atmospheric circulation index anomalies in the Scandinavian (red curve), Urals (orange curve), and Baikal (purple curve) regions, defined as the deviation from the climatology in October and November 1981–2010. The black curve represents the daily NECV index anomaly.

A weak stratospheric polar vortex can propagate downwards to influence tropospheric circulation by modulating the phase of the AO (Baldwin et al., 2003; Wang and Chen, 2010). Woo et al. (2015) found that regional coupling between the stratosphere and troposphere can influence the strength of the East Asian trough and thus the activity of cold events over East Asia during a weak stratospheric vortex. Besides, previous studies have concluded that the AO can affect cold events over eastern China by modulating the Siberian high, East Asian trough, and polar front jet (Gong et al., 2001; Wu and Wang, 2002; Jeong and Ho, 2005). Thus, change in the AO phase could act as a bridge connecting the polar vortex and cold-air weather systems in East Asia. In the present study, further analysis also showed that the AO was closely related to the change in the Baikal ridge/blocking (Fig. 10). Generally, the AO index was in a negative phase (−0.13) in October, which was beneficial to the exchange of meridional cold and warm air advection (Fig. 10a). When the AO index was in a positive phase, there were positive height anomalies over the Scandinavian and Baikal regions but negative height anomalies over the Urals region (Fig. 10b). The distribution of geopotential anomalies was similar to that in late October at 500 hPa (Fig. 8f). The daily AO index and the Baikal ridge/blocking high were both positive during 20–28 October 2021 (Figs. 9 and 10a), which indicates the accumulation of cold air associated with the establishment of the Baikal ridge/blocking high. The AO then shifted to a negative phase on 29 October, and the Baikal ridge/blocking high presented as a negative anomaly on 1 November, indicating that cold air had begun to break out southeastwards. Therefore, the phase shift of the AO is a useful indicator for the outbreak of cold air in this EHS event.

Figure 10.  (a) Daily AO index anomaly (relative to the climatology during 1981–2010) in October 2021 and (b) daily AO index anomaly regressed onto the 500-hPa height field in October 2021, in which the dotted area indicates statistical significance at the 95% confidence level.

A record-breaking EHS event occurred in northern China during 6–8 November 2021. Data analysis shows the existence of two extreme snowfall centers, located in NC and NEC, where the maximum cumulative snowfall exceeded 20 and 40 mm, respectively. In this study, the contributions to this event of two key factors at the synoptic scale—cold air and water vapor—were analyzed. The analysis focuses on quantitative diagnoses of the physical mechanisms behind the development of synoptic systems and moisture sources associated with the event. Although the results presented are only applicable to this specific case, the approach used and the conclusions obtained add new evidence and insights towards improving our understanding of such extreme events and their predictability.

For the EHS during 6–7 November over NC, a positive-tilt 500-hPa trough over northwestern China played an important role. On the one hand, trough-front cold advection continuously transported cold air from high latitudes into NC; on the other hand, uplifting along the front below the trough contributed to the convergence at lower levels, which promoted NC as the moisture sink. With the northeastward movement of the 500-hPa trough, cold air advection reached NEC on 7 November. Meanwhile, the uplift below the trough triggered a Yellow–Bohai Sea low-level cyclone, which promoted low-level water vapor transport to NEC. Therefore, a heavy snowfall center appeared in Liaoning Province and to its north on 7 November. As the 500-hPa trough moved further northeast, an NEC cold vortex developed when the 500-hPa trough coincided with the center of the low-level Yellow–Bohai Sea cyclone; and this NEC cold vortex was the key synoptic feature driving the EHS in Heilongjiang and Jilin, coincident with warm, moist air advection on 8 November. On 9 November, the EHS ended as the cold vortex matured. Therefore, the location of the frontogenesis region in the front of the 500-hPa trough, where ascending warm air was forced by the undercutting of cold air, played an important role in this EHS event.

To detect the development of the trough/ridge, the local relative vorticity tendency was calculated, indicating that absolute vorticity advection dominated the movement of the trough/ridge, rather than the absolute vorticity divergence. Furthermore, the contributions of moisture sources for this EHS event were quantitatively diagnosed by using the HYSPLIT model. The results indicated that in NC, the main moisture sources were the Asian–African region (32%) and the Eurasian mid-to-high latitudes (31%), and the height was mainly in the mid-to-low levels (500–800 hPa). However, in NEC, the two dominant moisture sources were the Eurasian mid-to-high latitudes (38%) at mid-to-low levels and East Asia (28%) in the near-surface layer. A schematic diagram of the mechanism that modulated this EHS event is presented in Fig. 11.

Figure 11.  Schematic diagram showing the mechanism that modulated the EHS event during 6–8 November 2021 in northern China.

Analysis of the atmospheric circulation anomalies before the EHS event showed that the polar vortex had shifted eastwards owing to its weakening, thereby providing favorable conditions for cold polar air to move southward. Also, the blocking high collapsed in the Scandinavian and Urals region but prolonged in the Baikal region, which contributed to the accumulation of cold air over western Siberia. Finally, a cold-air outbreak occurred in northern China after the collapse of the Baikal ridge/blocking high. Further analysis showed that the phase shift of the AO from positive to negative contributed to the collapse of the Baikal ridge/blocking high. Therefore, a phase shift in the AO related to weakening of the polar vortex is a useful indicator of cold-air outbreaks in such an EHS event.

Although previous studies have stressed the key role of a negative-phase AO in cold surges over East Asia (Park et al., 2011; Wu et al., 2011; Woo et al., 2012; Yuan and Li, 2019), it has also been found that the relationship between midlatitude cold-air outbreaks and the AO is less robust in the Northern Hemisphere winter (Yu et al., 2015). Liu J. P. et al. (2012) also indicated that changes in atmospheric circulation, linked to Arctic sea-ice loss that can cause winter snowfall in the midlatitudes, show different interannual variability than the classical AO. Therefore, the impact of an AO-like pattern on midlatitude weather and climate changes may differ across timescales and regions. Due to the case-based approach in this study, the phase-shift of the AO as a precursor signal may not necessarily hold true for all such EHS events, and so its prevalence throughout regions such as northern China needs to be investigated in further research on historical and future EHS events at synoptic scales.

The results show that the NECV promoted this EHS event over NEC in the early formation stage. Previous studies have also stressed that the NECV is a major synoptic system causing cold-related damage, extreme precipitation, and severe convective storms over NEC (Sun et al., 2002; Liu G. et al., 2012; Liu H. B. et al., 2012; Xie and Bueh, 2015; Wu et al., 2021). However, the mechanism of NECV formation and maintenance are still under debate (Lian et al., 2016), and thus further NECV-related research is a key scientific issue moving forwards.

The authors wish to thank Dr. Zhi’ang Xie for valuable suggestions related to this study.