In the winter of 2008, China experienced once-in-50-yr (or once in 100 yr for some regions) snow and ice storms. These storms brought huge socio economical impacts upon the Chinese people and government. Although the storms had been predicted, their severity and persistence were largely underestimated. In this study, these cases were revisited and comprehensive analyses of the storms’ dynamic and thermodynamic structures were conducted. These snowstorms were also compared with U.S. east coast snowstorms. The results from this study will provide insights on how to improve forecasts for these kinds of snowstorms. The analyses demonstrated that the storms exhibited classic patterns of large-scale circulation common to these types of snowstorms. However, several physical processes were found to be unique and thought to have played crucial roles in intensifying and prolonging China’s great snowstorms of 2008. These include a subtropical high over the western Pacific, an upper-level jet stream, and temperature and moisture inversions. The combined effects of these dynamic and thermodynamic structures are responsible for the development of the storms into one of the most disastrous events in Chinese history.
China’s great snowstorms of 2008 have been identified as the third-most memorable event in China during 2008, after Beijing’s summer Olympic Games and the devastating Sichuan earthquakes. The snowstorms occurred on such a grand scale that 20 out of China’s 34 provinces and special districts were affected by this disastrous event. According to Chinese Department of Civil and Internal Affairs, more than 100 million people suffered from these storms, and the economic loss totaled about 100 billion Chinese yuan (about U.S. $15 billion; information online at http://ncc.cma.gov.cn/upload/upload2/yxpj/ehpj_m080200.doc).
From early January to early February 2008, China experienced once-in-50-yr (or once in 100 yr for some regions) snow and ice storms. These snow and ice storms were organized into four episodes, occurring during 10–16 January, 18–22 January, 25–29 January, and 31 January–2 February. Most of southeast China was impacted by these large-scale storms. Figure 1a shows the distribution of accumulated total precipitation during this period (10 January–2 February) over China. Seven provinces exhibited total precipitation of over 100 mm. If one uses a conservative snow–liquid conversion factor of 0.2 (NRCS 2009), these provinces received more than 50 cm (∼20 in.) total snowfall during this period. Figure 1a also plots the distribution of the snowfall and freezing rain areas. Most snow occurred in the north, which gradually transitioned into freezing rain in the south. The affected areas covered most of southeast and east-central China.
Figure 1b plots China’s topography. From the southeast to the northwest, the altitude rises up in three steps. Such a topographic distribution leads to cold-air intrusion from Mongolia into northeastern China, then to southeastern China (shown as black arrows in Fig. 1b). Based on these observations, we plotted regionally averaged precipitation and temperature variations during 10 January–3 February, averaged over stations east of 105°E within a latitudinal band of 25°–35°N (Fig. 2a). Figure 2a shows that four significant precipitation episodes occurred during this period, accompanied by sustained low temperature, which exhibited a dramatic drop during the first precipitation episode.
The snow depth on 0000 UTC 28 January 2008 indicated that heavy snow was being accumulated in the region of the middle and lower reaches of the Yangtze River, among which parts of the Anhui and Jiangsu Provinces had reported snow depths of over 30 cm (Fig. 2b). In Table 1, we list the averaged temperature and accumulated total precipitation and corresponding anomalies (calculated based on a climatology averaged from 1971 to 2000) for 13 provinces strongly affected by these snowstorms. One can see that most of these provinces showed an anomalous increase in accumulated total precipitation during this period. The situation was made worse by the anomalously low temperature, which meant that a lot of the precipitation took the form of snow or ice accumulating on the ground. The snowstorms paralyzed both surface and air transportation, cut off electricity and other energy supplies, and damaged agricultural crops. Even worse, these storms occurred right before the Chinese traditional New Year (5 February), when more than two hundred million people were on the move to their homes or on holiday trips. The storms trapped millions of people in railway stations, airports, and other transportation shelters.
In addition to the unprecedented socioeconomic impacts brought about by these storms, more than 50 000 Chinese meteorological professionals and scientists were under the media’s spotlight—being questioned as to what they knew about how these extreme events happened on such a grand scale, and how much predictability they could provide for future occurrences of storms like these. In recent years, many Chinese meteorologists and climatologists (Xu et al. 1999; Wang 2001; Li and Ding 2005; Shi et al. 2007) have indicated a weakening trend in the intensity of the East Asian winter monsoon since 1980. In response to the reduced monsoon intensity, China had experienced persistently warm winters for several years, and winter precipitation had also been decreasing (Wang and Ding 2006; Qian and Zhang 2007; Q. Li et al. 2007). There has been a large body of work that has concentrated on various mechanisms responsible for the recent warm winters in China. These studies pointed to the North Pacific Oscillation (NPO; Wu and Wang 2002; Gong and Wang 2003; Gong et al. 2004; Chen and Kang 2006), El Niño and the Southern Oscillation (ENSO; Tao and Zhang 1998; Gong and Wang 1999; He et al. 2007), the Siberian high (Qian et al. 2001; Gong et al. 2002), the western Pacific teleconnection (Y. Li et al. 2007), and the East Asian jet stream (Yang et al. 2002; Mao et al. 2007); all of which may have influenced China’s winter precipitation and temperature to some extent. On the other hand, Huang et al. (2007) suggested that the winter stationary planetary wave played an important role in the variation of the East Asian winter monsoon.
Some synoptic studies for China winter snowstorms began to occur in the 1970s. Wang and Xu (1979) first investigated China’s Inner Mongolia “77.10” extraordinary snowstorm. They proposed a “north ridge–south trough” antiphased conceptual model for snowstorms in northern China. Gong and Li (2001) analyzed the relationship between Inner Mongolian snowstorms and low-level jets and pointed out the importance of southerly low-level jets in the development of snowstorms in this region. Many other researchers also carried out diagnostic and modeling studies on snowstorms over the Tibetan Plateau (Deng et al. 2000; Zhang and Cheng 2000a,b; Wang and Chen 2002). In recent years, there have been some case studies on abnormal snowstorms in the lower latitudes of southern China. For example, Zhu and Shou (1994) investigated secondary circulation and frontogenesis associated with a snowstorm that occurred in the Yangtze and Huaihe River basins. Yang et al. (2006) also conducted a modeling study of this storm using the fifth-generation Pennsylvania State University–National Center for Atmospheric Research Mesoscale Model (MM5). Nevertheless, knowledge of snowstorms in southern China is still quite limited, and understanding the physical principles that drive abnormal winter precipitation in lower-latitude southern China is still lacking.
During this grand snowstorm episode, the meteorological services from the central government to various local provinces all provided relatively accurate forecasts for each episode. The 2–5-day forecasts did indicate that the storms and precipitation events would cover a large area of southeast China. However, being subject to the forecast lead times, the predictions from numerical weather forecast models were not able to give sufficient warnings for persistent low temperature and continued snowfall for more than 20 days. Therefore, the severity and persistence of these storms were underestimated. Furthermore, forecasters also realized that such extreme weather events were very improbable; even 50- and 100-yr climate records did not include such persistent cold temperatures, consecutive heavy snowfalls, and ice storms for these areas.
The present study tries to provide a comprehensive understanding of how China’s great snowstorms of 2008 occurred and developed, and presents a picture of the synoptic- and storm-scale dynamic and thermodynamic structures. We need to clarify that the term “storms” used in this study does not denote a series of surface cyclones moving along a quasi-stationary frontal boundary through southeast China. The storms here are a prolonged period of warm advection in the midtroposphere far to the north along a quasi-stationary frontal boundary. However, because of their large amount of snowfall, we refer to this series of events as snowstorms.
To investigate these snowstorms, we focused our data analyses on the period from 0000 UTC 10 January 2008 to 2000 UTC 2 February 2008. In accord with the four precipitation periods (10 January–16 January, 18 January–22 January, 25 January–29 January, and 31 January–2 February), we divided the data into four episodes and investigated each separately. In addition, the composite of these four episodes provided a comprehensive understanding of the physical mechanisms for the snow and ice storm’s genesis, development, and reinforcement. These episodes also constituted four individual samples for comparison studies. We concentrated on the analyses of surface precipitation observations from 2519 weather stations located throughout China and provided by the China Meteorological Administration (CMA). For other meteorological variables, we used the National Centers for Environmental Prediction–National Center for Atmospheric Research reanalysis dataset (Kalnay et al. 1996), which provided geopotential height, meridional and zonal winds, humidity, and air temperature. The spatial resolution for this dataset is 2.5° × 2.5° latitude–longitude. We used 6-hourly data for both surface precipitation observations and the reanalysis data.
3. Synoptic conditions and large-scale atmospheric circulations
Figure 3a shows averaged sea level pressure (SLP) and surface wind fields over the Asian region during 10 January–2 February 2008. Two main features that can be identified from this chart are the existence of the Siberian high (encompassed by the blue rectangle in Fig. 3) and the prevailing northerly wind in eastern China and the adjacent ocean (encompassed by the orange rectangle). To understand the variation characteristics of the Siberian high during the snowstorm period, we plotted the latitude–time evolution diagram of the SLP (contours) and its anomaly (color shaded) averaged over the 70°–130°E meridional belt (Fig. 3b). One can see that this Siberian high pressure system was sustained over the entire snowstorm period and intensified during the four significant precipitation episodes. Compared with the climatology (averaged over 1971–2000), the SLP (between 40° and 70°N) exhibited positive anomalies, which indicates that the Siberian high had been anomalously strong during the period of the snowstorms. This situation was also accompanied by an intensification of corresponding northerly winds between 10° and 40°N in the 110°–130°E longitude band (Fig. 3c). The sustained negative anomalies in the meridional wind indicated that there had been stronger than normal northerly winds in eastern China and the adjacent ocean. The anomalous northerly winds brought significant cold air into eastern China.
Figures 4a–d are composites of 500-hPa geopotential height and its anomaly fields over the Asian region, corresponding to the four significant precipitation periods. There existed a blocking ridge north of China during all of these cases. The corresponding blocking centers are respectively at 62°N, 90°E; 55°N, 70°E; 55°N, 85°E; and 48°N, 70°E. In the meantime, the western Pacific subtropical high (WPSH) between 10° and 25°N also showed above-normal strength. During the four precipitation episodes, the positive–negative–positive patterns of the geopotential anomalies were sustained. While the variations of the northern blocking ridge affected cold air moving southward, the variations of the WPSH controlled warm–moist air transport northward. In normal winter years, a large-scale cold-air intrusion from the north can often cause a temperature drop over a large area of China. When combined with a supply of moisture from the south, a winter weather system often brings snow in northern China and rain in southern China. Such a winter system typically lasts between 5 and 7 days. With the breakdown of the northern blocking ridge and the weakening of the WPSH, the cold–dry northerly wind and warm–moist southerly flow weakened. The typical large-scale circulation pattern for winter precipitation in eastern China would either move out to sea or dissipate entirely. However, China’s great snowstorms of 2008 occurred during a 20-day period with heavy precipitation during four distinct episodes of enhanced precipitation, each lasting a little over 5 days. From the large-scale synoptic weather patterns indicated in Figs. 3 and 4, we observed that in the present case the typical large-scale circulation pattern formed but never moved on or dissipated. There was a strong negative height anomaly over Japan and north of the WPSH, which would result in the stronger than normal jet over northeast China that will be discussed later in this paper. This would act to keep the entrance region of the jet persistently over eastern China.
In Fig. 5a, we plot the 700-hPa heights and temperatures, averaged over the period of 10 January–2 February 2008 (24 days). The geopotential height (solid, black) developed a ridge in northwestern China and a trough over central China. This pressure pattern caused confluent flows downstream (indicated by two arrows). There was also a thermal trough (marked by the thick dashed line) in the vicinity of the flow confluence region in southeast China (around 20°–35°N and 105°–120°E). The northern branch of the convergent flow brought in cold–dry air, while the southern branch brought in warm–moist air. The development of the confluence zone led to an uplifting of warm–moist air over the cold–dry air; thus, heavy snows occurred along the confluence region. All of these features were very similar to those found in a study (Fig. 5b) of the 1977 snowstorm over Inner Mongolia by Wang and Xu (1979). Their findings indicate that the “north ridge–south trough” antiphase pressure field coupled with confluence may characterize the typical synoptic conditions for this type of snowstorm. From the latitude–time evolution diagram (Fig. 6) for the 700-hPa geopotential height (averaged over 70°–110°E), one can also see that the signature of the north ridge–south trough corresponds to each precipitation process. In each episode the trough (dashed ellipse) was pushed by the strong blocking ridge (solid ellipse) from the north deep into the south, below 30°N. As noticed previously, the presence of a strong subtropical high off the southeast coast (at 20°N) of China might contribute to the persistence of snowstorms. The subtropical high sustained the precipitation not only by supplying a large amount of warm–moist water vapor to southeastern China but also by also blocking the troughs from moving out to sea.
To examine the wind, temperature, and precipitation fields in the low-level atmosphere, in Fig. 7 we further plotted these fields at 925 (left column) and 850 hPa (right column) for each episode from top to bottom, respectively. During each significant precipitation process, a cold tongue (or thermal trough) was evident in the lower levels of the atmosphere, just west of the heavy precipitation (color shaded). The cold tongue was stronger at lower levels than above because at 925 hPa the dominant wind direction (arrow) in this region was northeasterly (Figs. 7a, 7c, 7e, and 7g), while at 850 hPa flows were mostly southwesterly (Figs. 7b, 7d, 7f, and 7h). The 925- and 850-hPa fields also seem to show some evidence of cold-air damming, especially with the strong ageostrophic northeast flow at 925 hPa.
To better understand the variations of synoptic forcing and precipitation processes, we conducted detailed analyses for a particular storm episode. The third precipitation episode (25–29 January 2008) was selected for this study. Figure 8 shows the SLP over the north of China every 24 h. One can see that the Siberian high intensified from 25 to 26 January (Figs. 8a and 8b). From 27 to 29 January (Figs. 8c–e), the Siberian high pressure decreased slightly, but it persisted and continuously expanded southward. Corresponding to this, the northerly winds (at the 925-hPa level) followed a strengthening–weakening–strengthening pattern, but were persistent. At 500 hPa (Fig. 9), the Ural-blocking high and a weak antiphased trough established on 26 January persisted throughout the storm episode. During the same period, the WPSH also showed a strong signal (see the 588-dm contour), which expanded westward on 26 and 27 January. It began to weaken and retreat to the east over the course of 28 and 29 January. Associated with the setup of the WPSH and upper-level trough, strong water vapor transport occurred from the Bay of Bengal and the South China Sea, seen at 700 hPa by the wind (vectors) and specific humidity (color shaded) analysis in Fig. 10. On 29 January, both water vapor transport and specific humidity displayed significant decreases. Figure 11 shows that heavy precipitation occurred on 25 January (color shaded). It weakened on 27 January, regained strength on 28 January, and began to dissipate on 29 January. The contours plotted in Fig. 11 are 200-hPa zonal winds, which show the development of the upper-level jet streams. Figures 11a and 11d show that the greatest precipitation falls in the right-entrance region of the jet maximum to the east of China and is associated with the large-scale confluence. The precipitation weakens in Figs. 11b and 11c when the jet streak propagates through southern China and the precipitation region is located in the right-exit region. Figures 11b and 11c also suggest that the primary dynamic mechanism for the precipitation is not the jet-streak circulations. Rather, the precipitation is enhanced when it is in a favorable location with respect to the jet. This indicates that the upper-level divergence in the right-entrance region of a jet stream poses a “pumping effect” to maintain an upward motion, thus promoting precipitation in the region. As the jet core moved out of China (Fig. 11e), the precipitation began to dissipate.
Comparing these results with U.S. northeast snowstorms (Kocin and Uccellini 2004, hereafter KU), we find that many of the main features between the snowstorms on the two continents are strikingly similar. For example, the large-scale pattern of a dominant upper-level trough coupled with a weak ridge to its north resulting in confluent flows east of the pressure centers was also found in Fig. 7-4 of KU. In addition, the high pressure system in the north leading to cold-air damming in Figs. 7-4, 7-6, and 8-1 of KU is conceptually the same as the cold tongue intrusion from the north in the Chinese snowstorms analyzed in this study. However, a distinctive feature shown in the Chinese snowstorms was the presence of a subtropical high over the coast. As we discussed above, this subtropical high might play an important role in blocking the snowstorms exiting to the sea, which explains why the snowstorms lasted on and off for almost a month inland. Of course, the subtropical high also helped to advect warm and moist air northward. In the United States, this subtropical high blocking effect was typically not seen. Furthermore, differences in geographic features from the two countries may also result in differences in snowstorms (e.g., storm locations). China’s great snowstorms of 2008 occurred along the southeast coast of China, between 20° and 35°N, while U.S. snowstorms typically occurred along the northeast coast, around 35°–45°N. In China, the world’s largest and highest plateau and mountains, the Tibetan Plateau and Himalayas, are located upstream of the snowstorms, directly to the west. The plateau blocks water vapor transported from low-latitude oceans to northern China. Instead, it deflects the moisture transport to the east, which promotes a storm development pattern along the southeast coast of China.
4. Vertical wind shear, upper-level jet, and temperature–moisture inversion
Note that all four episodes of snowstorms essentially occurred in the same large areas shown in Fig. 1a. All features discussed above are common to all four snowstorms that occurred during this period. The only difference is the varying strength of these features from one storm to another. Thus, we can examine the flow fields by taking an average of the four episodes. We plotted the averaged meridional wind (contours) and its anomaly (color shaded) in a south–north vertical cross section (Fig. 12a) and averaged the zonal wind (contours) and its anomaly (color shaded) in a west–east vertical cross section (Fig. 12b). It can be seen that in the snowy region (blue shaded), vertical wind shear was evident in both meridional and zonal wind components at lower levels. For example, for the meridional wind in Fig. 12a, the southerly wind prevailed above 850 hPa, while below 850 hPa the northerly wind prevailed. The anomaly showed negative–positive coupling from the bottom up, indicating a stronger southerly wind at upper levels and a stronger northerly wind at lower levels in comparison to climatology. Similarly, for the zonal wind in Fig. 12b, the westerly wind prevailed above 850 hPa and a very strong westerly jet was located at the upper troposphere (around 200 hPa) with a positive anomaly, while below 850 hPa the easterly wind prevailed with a negative anomaly. The low-level northeasterly wind brought in cold–dry air, which undercut the warm–moist air brought by the southwesterly wind at middle levels.
The vertical wind shear also suggests a thermodynamic structure of cold-air intrusion at low levels and warm-air crossover at middle levels. During the period of persistent freezing rain and snow, the third episode (25–29 January 2008) was the most severe snowstorm. Figures 13 and 14 show the vertical cross sections of temperature and specific humidity coupled with vertical atmospheric circulations in this episode. In the south–north vertical cross section (Fig. 13a), the isotherms (color shaded) between the 925- and 700-hPa levels tilted toward the north with increasing altitude, indicating a temperature inversion and warm advection in the region of heavy precipitation (Fig. 13c). In the west–east vertical cross section (Fig. 13b), a cold-air center existed in the low levels around 110°E. This cold-air center is a slab view of the cold tongue that we saw in Fig. 7, representing a cold-air intrusion from north to south in the lower atmosphere. On the east side of this cold-air center, precipitation occurred over a large area (Fig. 13d). Over this region between the 925- and 700-hPa levels, the isotherms (color shaded in Fig. 13b) tilted westward with the increase of altitude, which also indicates a temperature inversion.
Very similar vertical structure is also found in the humidity field. In the south–north vertical cross section (Fig. 14a), the specific humidity (color shaded) tilted toward the north with increasing altitude between the 925- and 700-hPa levels, right above the snowfall area (Fig. 14c). This indicated moisture inversion in the region of heavy precipitation as well. In the west–east vertical cross section (Figs. 14b and 14d), a dry-air center existed in the low levels around 110°E, in association with the cold-air intrusion from north to south at lower levels. On the east side of this cold-air center and between the 925- and 700-hPa levels, the specific humidity contours tilted toward the north with increasing altitude, also depicting a moisture inversion. The moisture inversion manifested itself as the moisture was lifted over the cold dome of air, presumably condensing in the upper levels to produce the snow–frozen precipitation. The other three precipitation episodes showed similar temperature and moisture inversions in the vertical.
According to the distribution of wind fields at the 200-hPa level for the particular episode (Fig. 11), the jet stream was located in a latitudinal belt between 22.5° and 35°N. To understand the relationship between the variational characteristics of the jet stream and the surface snowfall for the entire storm period, we plotted a meridional–time evolution of zonal wind fields at the 200-hPa pressure level, averaged over the latitudinal belt (Fig. 15). The shaded regions denote the locations (both in time and longitude) of the four precipitation episodes. It can be seen that the jet stream intensified with time as the snowstorms developed from one episode to the next. Significant precipitation occurred at the locations below the entrance region of the jet streak on the equatorial side; that is, the heavy precipitation could be correlated to an upper-level divergence in the right-entrance region of the jet steak. As the jet intensified with time, the pumping effect due to the divergence field at the right-entrance region of the jet streak became dominant. In response, the storms began to organize into the classical configuration with a complete coupling of surface- and upper-level systems, and more precipitation followed. These features are consistent with the conceptual models of ageostrophic flows associated with the entrance and exit regions of coupled jets, as summarized in Fig. 7-4 in KU.
Combining these analyses, an important thermodynamic feature stands out during these snowstorm episodes. The cold–dry-air intrusion from the north seemed to play a central role in enhancing the baroclinic zone. This provided favorable conditions for a prolonged period of warm advection in the midtroposphere far to the north along a quasi-stationary frontal boundary, thus initiating and maintaining the snowstorm. This cold–dry-air intrusion formed a “cold wedge” (see Fig. 7). Because this cold air was wedged in at lower atmospheric levels, warm–moist air from the south was forced upward, providing a mechanism for clouds and precipitation to form. The formation of the “cold wedge” is consistent with the pattern of the north ridge–south trough in the large-scale dynamic system (Fig. 5a), as discussed in the previous section.
In this study, we assembled composite analyses as well as a case study for China’s great snowstorms of 2008. Through these analyses, we obtained the following physical pictures of the large-scale dynamic and thermodynamic structures of the snowstorms (Fig. 16).
From analyses of the SLP and 500-hPa geopotential height fields, we found that a prominent blocking ridge existed over the northern Asian continent during these snowstorms. Meanwhile, all four snowstorm episodes presented a “north ridge–south trough” pressure distribution at 700 hPa, which led to a southward intrusion of cold air. The maintenance of a WPSH promoted the transport of warm–moist air into southeast China. The WPSH also played a crucial role in blocking the storm system from moving eastward and out to sea. The confluence of the two airflows—the northern cold air and the warm, moist air from the south—provided favorable conditions for a prolonged period of warm advection in the midtroposphere far to the north, along a quasi-stationary frontal boundary. Further analyses of atmospheric vertical circulations indicated that there was strong vertical wind shear at low atmospheric levels. In the snowstorm region, northeasterly wind dominated below 850 hPa, while southwesterly wind prevailed above 850 hPa. Where the two airflows met, the warm–moist air was lifted up over the cold air. This process caused both a temperature and a moisture inversion and thus formed a “cold-air wedge” with a “warm–moist cover” sliding over a “cold–dry base.” The upper-level jet stream was intensified, and was consistent with the intensifications of surface lows, in the sense that for the heavy precipitation events the right-entrance region of the jet streak aligned itself with the surface precipitation. All of these in-phase dynamic and thermodynamic structures were responsible for sustaining and intensifying the storms, which evolved into one of the most extraordinary and disastrous weather events in recent Chinese history.
The authors would like to express their sincere gratitude to Jeff Waldstreicher for his excellent editorship and insightful suggestions that improved this paper significantly. This research was jointly funded by the International Sci.-Tech. Cooperative Project (2007DFB20210), the Ministry of Science and Technology of the People’s Republic of China, the Key Project of Basic Scientific Research and Operation fund of the Chinese Academy of Meteorological Sciences (2008Z006), the Independent Research Project of LaSW (2008LASWZI04, 2009LASWZF02), and the China–Japan Meteorological Disaster Cooperative Research Center, which is China–Japan intergovernmental cooperation program that is supported by the Japan International Cooperation Agency (JICA). We are also grateful to Annie Reiser for providing typographical, grammar, and style edits for this paper.
Corresponding author address: Dr. Chungu Lu, GSD David Skaggs Research Center, 325 S. Broadway, Boulder, CO 80305. Email: firstname.lastname@example.org