1. Introduction
China is a developing country, and agriculture is important to its economy. Hailstorms are one of the main disasters/risks to farmers. The hail damage in China was reported to be valued as high as billions of U.S. dollars in 2004 alone (Dong et al. 2006). Therefore, the study of hail climatology and improvements in the prediction of hailstorms are important to disaster prevention/reduction.
Hailstorms were routinely reported at many places around China from the early nineteenth century, but were not documented until Liu and Tang (1966), who showed the hail spatial and temporal distributions in China. Using 811 surface meteorological observations between 1951 and 1960, Liu and Tang (1966) documented the annual and seasonal variations of hailstorms in China. Their work is limited by a short dataset (10 yr), with 40% stations containing only 7 yr. Nevertheless, Liu and Tang’s work has been the only one that produced the hail climatology for all of China.
The early research on hail climatology in China was reviewed briefly by Xu (1983). There have been several recent publications with their focus on the hail climatology in parts of China. For example, Yang and Ma (2003) described the hail climatology of northern China based on data between 1991 and 2000. Yang (2002) discussed the hail characteristics in the Xinjiang Uighur Autonomous Region. However, a study of the country-wide hail climatology in China with long-term hail observations has not yet been carried out.
Hail climatology has been well studied in Europe and North America. Changnon (1977) reviewed hail information at various time and space scales, and pointed out that the principal hail area in North America is to the lee side of the Rocky Mountains where hail is both frequent and intense. Dessens (1986) analyzed hail damage in southwestern France during a 29-yr period between 1952 and 1980. Mean spatial hail patterns between 1931 and 1975 in Greece were revealed by Kotinis-Zambakas (1989). With the help of a mesoscale weather network and hail observational network with hail pads in the last two decades, more and more hail data have become available and have been analyzed to better represent the hail climatology in some countries, such as France (Vinet 2001) and Canada (Etkin and Brun 1999). In those countries, where meteorological observations began in the nineteenth century or even earlier, a long-term climatology could be constructed (Changnon 2000; Schuster et al. 2005; Webb et al. 2001).
China occupies the eastern part of the vast Eurasian continent, with a variety of topographies and landscapes. In particular, China has the world’s highest plateau (the Tibetan Plateau) to its west. Its hail climatology is thus expected to be quite different from many other countries and also is varied from one region to another within the country. Therefore, there is a need to extend the recent regional efforts mentioned earlier to include the whole country with the use of all available observations. This is made possible by a new dataset for hail observations from 753 meteorological stations since 1951 that has been recently released by the National Meteorological Information Center (NMIC), China Meteorological Administration (CMA). We note, however, that although most surface meteorological stations in China were established in the 1950s, the data quality and continuity became much more stable after 1961 due to historical reasons. As a result, our analysis will be based on hail observations during the 45-yr period between 1961 and 2005.
The objective of this paper is to provide an update to the hail climatology in China, including the geographical distribution of mean annual hail frequency and seasonal and diurnal variations of hail occurrence. The rest of the paper is organized as follows: section 2 describes briefly the dataset used in the study, section 3 discusses in detail the geographical distribution of mean annual hail frequency and seasonal and diurnal variations of hail occurrence in China, a discussion is given in section 4, and a summary is in section 5.
2. Data
The observational hail data used in this study were obtained from NMIC, CMA. The NMIC provides a complete historical hail dataset, which includes hail data for all weather stations in the surface meteorological observational network over all of China from 1951 to 2005. The hail data are coded routinely with the World Meteorological Surface Observation Criteria (information online at http://www.wmo.ch/web/www/WMOCodes.html). After decoding and quality control, the first comprehensive collection of hail data was released in May 2006 by NMIC.
Hail in China is defined as precipitation that is in the form of balls or irregular lumps of ice with diameters that are larger than 2 mm, has an opaque inner core, and is covered by a clear ice layer or several clear–opaque–clear layers (China Meteorological Administration 2003). The differences between small hail (with diameter between 2 and 5 mm) and sleet are also described in the CMA handbook (China Meteorological Administration 2003). Hail is hard, while sleet is loose in shape and fragile. The hail data include records of time of occurrence and duration, but not of hailstone size or other information. Hail data are recorded daily and manually.
In our analysis, if there are more than 6 days of observations missing in a month, the month is regarded as an invalid month; if there is an invalid month in a year, that year is regarded as an invalid year. Invalid years are not considered in our analysis. There are 753 stations in this dataset, but not all stations cover the entire period from 1951 to 2005. To ensure a relatively large and continuous data record, we chose stations with complete observations from 1961 to 2005. The resulting 604 stations contain 4 stations that are missing records of 3 yr, 11 stations that are missing records of 2 yr, and 43 stations that are missing records of 1 yr (Fig. 1). Note that because of the high elevation and sparse population, there are relatively few surface meteorological stations in the western part of the Tibetan Plateau (Fig. 1).
In this study, a hail day is defined as a day during which hail is observed and recorded at each station. The (mean) annual hail frequency at a station is defined as the mean number of hail days per year. For example, for a station, its mean annual hail frequency is equal to all hail days in 45 yr divided by 45. The mean monthly hail frequency is defined as the average number of hail days in each month during 1961 and 2005; namely, for a station, its monthly hail frequency is equal to all hail days of one particular month in 45 yr divided by 45. The hail peak month is defined as a month in which the hail days reach to 10% of the total hail days in a year.
3. Results
a. Geographical distribution of annual hail frequency
Mean annual hail frequency is displayed in Fig. 2. China has a terrain-related geographical distribution, similar to that of the United States (Changnon 1977) and Canada (Etkin and Brun 1999). The mean annual hail frequency pattern is similar to that in Liu and Tang (1966), but the following two areas indicate considerable differences. Liu and Tang (1966) showed more than five annual hail days over the Yinshan Mountains and in the southwestern part of the Great Xing’an Mountains based on a 10-yr period of observations between 1951 and 1960. Based on the complete 45 yr of observational data, our results show that the mean annual hail days are either about 2–3 or a little more than 3 days in these areas, and that there is a very small area over the Yinshan Mountains where there are more than four hail days. There are more than five hail days per year in the Changbai Mountain area shown in Liu and Tang (1966), while there are only about two hail days per year in our analysis. These differences could be attributed to different time periods of the datasets. As indicated previously, Liu and Tang (1966) studied the hail events during 1951 and 1960, while we have examined the hail events during a much longer time period from 1961 to 2005.
The general geographical distribution of annual hail frequency from our analysis can be summarized as follows:
The highest mean annual hail frequencies in China are located in the Tibetan Plateau and the Qilian Mountain areas, where hail frequencies are larger than 5 days yr−1. Two centers with the highest mean annual frequencies are located in the central region of the Tibetan Plateau and over the Qilian Mountains, respectively. Naqu [World Meteorological Organization (WMO) identification: 55299; 31.48°N, 92.07°E] has the peak value of 33.4 days yr−1.
High mean annual hail frequency exists at the Tian Shan area and the southwest of Altai. There are some small, scattered high-frequency centers in these areas, with more than two hail days per year. In the west and southwest parts of Tian Shan, the mean annual frequencies are even more than 5 days. There is a belt with more than two hail days per year to the south of Altai. This area was not included in the analysis of Liu and Tang (1966).
Hail frequency over the Yunnan–Guizhou Plateau is higher than 1 day yr−1 and higher than 2 days yr−1 in some areas. The detailed distribution is closely related to topography. Generally, areas with high hail frequency are located at the southeast slope of the high mountains.
The Yinshan Mountains, north of the Taihang Mountains, Great Xing’an Mountains, Changbai Mountain, and Dongbei (wide areas east of the Great Xing’an Mountains, almost half of them are plains) are hail-productive areas with mean annual frequencies exceeding 1 day. There is only an exception on the east slope of the Great Xing’an Mountains where the annual hail frequency is below 1 day.
There are some relatively isolated high hail frequency areas associated with isolated mountains over the eastern and southeastern plains of China. For example, the mean annual hail frequency for the Taishan Mountain and Wuyi Mountain is more than 1 day yr−1.
Few hail events occur in the Sichuan basin, southern Hainan Island, and on the coastal areas of southeast China. Hail frequency in the Tarim basin and a small area south of the Tibetan Plateau is unknown due to sparse meteorological stations in those areas.
b. Seasonal variation of hail occurrence
The mean monthly hail frequency is shown in Fig. 3 for all 12 months. There is a general trend of a northeastward migration of the high monthly hail occurrence from the Northern Hemisphere spring to summer in China. The high hail season appears in the southwest in January and February and then all of southern China in March, and extends to northern and northwestern China from April to September. This is followed by a considerable reduction of hail occurrence in October. Hail seldom occurs in the cold season in November, December, and January. Hail seasons in China are distributed with diversity. Most places experience hailstorms more frequently in summer, while the southern China has a high hail occurrence in spring, as compared with few occurrence in other seasons.
Based on the spatial seasonal distribution in Fig. 3, China can be divided into five regional types of seasonal variation of hail occurrence. Figure 4 shows the peak month of hail occurrence, and Fig. 5 gives the detailed description of mean monthly hail days of example stations.
Type I: There are two peaks in this type. The first peak in hail occurrence is in May and June (or in April–June, in some areas), and the second peak is in September (or September–October, in some areas). The former appears to be the main peak, while the latter appears to be a subpeak. This type is found in the eastern part of northeastern China, in some areas in central China, and in the western and northwestern parts of the Sichuan basin (Fig. 5a).
Type II: Most hail occurrences are from May to September. This type is found in both the western part of northeastern China and northwestern China (Fig. 5b).
Type III: In this type, the hail season begins later than that in type II. Most hail occurrences are from June to September. This type can be found in the Tibetan Plateau, Qilian Mountains, and northwestern China (Fig. 5c).
Type IV: This type can be called the spring type (Fig. 5d). Most of this hail occurs during February or from March to April or May. A hail peak appears in January in the western part of southwest China. Then, in February all of southwest and South China begins their hail season. The hail season does not start until March in the lower reaches of the Yangtze River.
Type V: Most hail occurrences are from April to September. This type is found in the western and northwestern parts of the Xinjiang Uighur Autonomous Region (Fig. 4, not shown in Fig. 5).
In addition, there are some areas with no significant peaks. In those areas hailstorms are rare events, such as in the coastal areas of southeastern China.
c. Diurnal variation of hail occurrence
Reliable information about the starting time of each hailstorm event is available for each hail record in the NMIC dataset. To examine the diurnal variation of hail occurrence, we divided a day into four periods, namely, 0000–0600, 0600–1200, 1200–1800, and 1800–0000 local time (LT). A map for hail diurnal variation was generated and is given in Fig. 6. The majority of hails occur in the afternoon (1200–1800 LT), which is similar to some areas in Australia (Schuster et al. 2005), while there are some regions where hail occurrence peaks at nighttime. One such region is around Tian Shan in the Xinjiang Uighur Autonomous Region, in which the hail mostly occurs in the early nighttime (1800–0000 LT). The other regions are Guizhou and Hubei Provinces. Hailstorms mostly occur at early night in Guizhou Province (1800–0000 LT) and late at night in Hubei Province (0000–0600 LT). Figure 7 shows the typical diurnal variations with nighttime hail occurrences at three stations. The hail ratio in Fig. 7 is defined as the hail occurrence in each hour divided by the total hail occurrence for a station in the 45-yr period. It is clear that hailstorms mostly occur between 1600 and 2200 LT in the Tian Shan area (Fig. 7a), 1400 and 2200 LT in Guizhou Province (Fig. 7b), and 2300 and 0600 LT in Hubei Province (Fig. 7c). Overall, 66% of the hail occurs between 1400 and 2000 LT, and 38% of the hail occurs between 1500 and 1800 LT in China.
4. Discussion
Hail is one of the phenomena associated with thunderstorms that is most likely to occur in an unstable atmosphere with an abundant low-level moisture supply, in the presence of strong vertical wind shear (Das 1962), and accompanied by a triggering mechanism that can release instability (Longley and Thompson 1965). Hail is also favored by both strong updrafts (to allow particle growth to occur) and low freezing levels in the atmosphere (Dessens 1986). It is a consensus that hail is favored by the front of two different air masses (Dessens 1986). The high-level jet stream also favors hail formation by offering strong vertical wind shear (Dessens 1960; Longley and Thompson 1965). As a result, topography and atmospheric circulation play important roles in hail formation (Etkin and Brun 1999). Ludlum (1980) and Kessler (1986) reviewed the global occurrence of severe thunderstorms and hail. The hail formation mechanism may be complex in China. As discussed by Vinet (2001), the formation of hail could result from a combination of multiple factors, but their relative importance varies from case to case and from one region to another.
As shown in Fig. 2, the occurrence of hail in China is closely terrain related. Almost all of the mountainous areas are hail concentrated. This is consistent with the finding of Lott (1999), who indicated that high mountains offer a favorable uplift background for hailstorms to form. Some high hail frequency areas, such as the Yunnan–Guizhou Plateau and northeastern China, are situated in hail-favored atmospheric circulation systems, such as the frontal systems and cut-off lows (Nieto et al. 2005; Kentarchos and Davies 1998). In a frontal system, the atmosphere, in general, is conditionally unstable with an upper-level jet and thus strong vertical wind shear (Hsieh 1949; Tao 1980). The rainbands organized by a cutoff low always bring hail, lightning, heavy rain, and strong winds (Tao 1980). Therefore, some plains of northeastern China are abundant with hail occurrence. The south branch of the westerly jet stream is always located to the south of the Tibetan Plateau in winter and spring, and brings warm and wet air to southwestern China (Duan et al. 1998; Yan et al. 2005). When this warm and wet air meets the cold air from the north, convection often forms and favors the formation of hailstorms.
The high hail occurrence in China exhibits a general trend of northeastward migration on a seasonal time scale from spring to summer, with a considerable reduction of hail occurrence in October. This seasonal variation can be explained by the large-scale circulation in East Asia, where the summer monsoon exhibits a similar northeastward progression over China. In particular, the East Asian summer monsoon flow provides plenty of low-level moisture for hailstorm formation (Tao 1980). The quasi-stationary frontal systems associated with the monsoon flow together with the low-level southwesterly jets provide conditionally unstable atmospheric conditions, favorable for deep convection, and thus hailstorm formation (Dessens 1986).
On a daily time scale, hail events are most likely to occur in the afternoon and early night. The peak is mainly a result of surface heating resulting from daytime solar radiation. Solar radiation processes destabilize the atmospheric column, favoring the development of convection and enhancing the existing convective systems. However, hailstorms mostly occur at early night in Guizhou Province (1800–0000 LT) and late at night in Hubei Province (0000–0600 LT). These two provinces are influenced by the same convective systems according to satellite observations (Liu et al. 2006). Because Hubei Province is located to the east of Guizhou Province, hailstorms may develop in Guizhou Province in the early night, and then propagate to the east or northeast following the synoptic weather systems, and finally arrive at Hubei Province after midnight.
There are some issues yet to be addressed. One issue is how sensitive the formation of hailstorms is to the environmental conditions. From the spatial and seasonal distributions of hail in China, one may speculate that hail could be sensitive to air humidity. This seems to be the case because hail is rare in the coastal areas of southeast China and other very wet areas. A typical case is the eastern part of northeast China, where hail is rich in May, June, and September, while hail is relatively rare in July and August during the wet and rainy season. This issue could be addressed with other adequate meteorological observational data in the future. Another issue that we have not touched upon is the long-term trend and decadal and interannual variability of hail activity in China and the involved physical mechanisms causing the trend and variability. This will be an interesting topic for future study.
5. Summary
A previous hail climatology of China was studied mainly based on the observations during a 10-yr period from 1951 to 1960 (Liu and Tang 1966). We have provided an update to the hail climatology in China with the use of 45 yr of observation from 1961 to 2005, which were collected and released in May 2006 by NMIC. The mean annual geographical distribution of hail frequency and seasonal and diurnal variations of hail occurrence are shown and can be summarized as follows:
Hailstorms are common in China. High hail frequency is found in the high mountainous areas and northern plains. Northern China has a higher hail frequency than southern China. The highest hail frequency occurs over the central Tibetan Plateau.
The season of hail occurrence in China has a diverse geographical distribution. Hail season starts from late spring to early autumn in northern and western China; in southwestern China the season mainly occurs in spring. Five types of geographical distribution of seasonal hail occurrence are identified.
Most hail events occur in the afternoon and early night. About 66% of the hail events occur between 1400 and 2000 LT and 38% occur between 1500 and 1800 LT. Hailstorms occur predominantly early at night in Guizhou Province and late at night in Hubei Province.
Acknowledgments
The authors express their sincere thanks to the editor and three anonymous reviewers for their constructive comments and suggestions, which helped to improve the representation of the paper. This study is supported by Chinese State 973 Key Program (2004CB418301), the Beijing Institute of Urban Meteorology Foundation under Grant UMRF200503, and the Chinese National Science Foundation under Grant 40675022; YW is supported in part by the JAMSTEC through its sponsorship to the International Pacific Research Center at the University of Hawaii. We thank NMIC for allowing us to access their hail data.
REFERENCES
Changnon, S. A., 1977: The scales of hail. J. Appl. Meteor., 16 , 626–648.
Changnon, S. A., 2000: Long-term fluctuations in hail incidences in the United States. J. Climate, 13 , 658–664.
China Meteorological Administration, 2003: Handbook of Surface Meteorological Observation. (in Chinese). China Meteorological Press, 151 pp.
Das, P., 1962: Influence of the wind shear on the growth of hail. J. Atmos. Sci., 19 , 407–414.
Dessens, H., 1960: Severe hailstorms are associated with very strong winds between 6,000 and 12,000 meters. Physics of Precipitation, Geophys. Monogr., No. 5, Amer. Geophys. Union, 333–338.
Dessens, J., 1986: Hail in southwestern France. I: Hailfall characteristics and hailstorm environment. J. Climate Appl. Meteor., 25 , 35–47.
Dong, W. J., Q. Zhang, J. X. Guo, and Y. Chen, 2006: Annals for Chinese Nature Disasters in 2004. (in Chinese). China Meteorological Press, 193 pp.
Duan, X., Y. Li, and Y. Zhou, 1998: Ambient field characteristics of high wind and hail over south Yunnan in spring (in Chinese). Meteor. Mon., 24 , 39–43.
Etkin, D., and S. E. Brun, 1999: A note on Canada’s hail climatology: 1977–1993. Int. J. Climatol., 19 , 1357–1373.
Hsieh, Y-P., 1949: An investigation of a selected cold vortex over North America. J. Atmos. Sci., 6 , 401–410.
Kentarchos, A. S., and T. D. Davies, 1998: A climatology of cut-off lows at 200 hPa in the Northern Hemisphere, 1990–1994. Int. J. Climatol., 18 , 379–390.
Kessler, E., 1986: Thunderstorm Morphology and Dynamics. 2nd ed. University of Oklahoma Press, 411 pp.
Kotinis-Zambakas, S. R., 1989: Average spatial patterns of hail days in Greece. J. Climate, 2 , 508–511.
Liu, K. Y., L. Li, and S. S. Wang, 2006: Analyses of Convective Systems Using Satellite and Lightning Observation: Collection of Weather Forecast Papers. (in Chinese). China Meteorological Press, 164 pp.
Liu, Q. G., and M. C. Tang, 1966: Climatological characteristics of hail in China (in Chinese). Acta Geogr. Sin., 32 , 48–65.
Longley, R. W., and C. E. Thompson, 1965: A study of causes of hail. J. Appl. Meteor., 4 , 69–82.
Lott, F., 1999: Alleviation of stationary biases in a GCM through a mountain drag parameterization scheme and a simple representation of mountain lift forces. Mon. Wea. Rev., 127 , 788–801.
Ludlum, F. H., 1980: Clouds and Storms: The Behavior and Effect of Water in the Atmosphere. Pennsylvania State University Press, 405 pp.
Nieto, R., and Coauthors, 2005: Climatological features of cutoff low systems in the Northern Hemisphere. J. Climate, 18 , 3085–3103.
Schuster, S. S., R. J. Blong, and S. S. Milton, 2005: A hail climatology of the greater Sydney area and New South Wales, Australia. Int. J. Climatol., 25 , 1633–1650.
Tao, S. Y., 1980: Heavy Rain in China. (in Chinese). Science Press, 224 pp.
Vinet, F., 2001: Climatology of hail in France. Atmos. Res., 56 , 309–323.
Webb, J. D. C., D. M. Elsom, and D. J. Reynolds, 2001: Climatology of severe hailstorms in Great Britain. Atmos. Res., 56 , 291–308.
Xu, J-L., 1983: Some hail research in China. Bull. Amer. Meteor. Soc., 64 , 124–132.
Yan, H. M., Z. N. Xiao, X. L. Zhang, and J. T. Li, 2005: Numerical simulation of mesoscale convective system of southern branch trough heavy rainfall in Yunnan region (in Chinese). Plateau Meteor., 24 , 672–684.
Yang, G. M., and Z. P. Ma, 2003: Characteristics of hailfall in North China (in Chinese). Meteor. Monogr., 29 , 31–34.
Yang, L. M., 2002: Climatic characteristics of hail in Xinjiang and the prevention (in Chinese). J. Catastrophol., 17 , 26–32.
Locations of the 604 stations in China with quality hail records for the 45-yr period from 1961 to 2005.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1603.1
The geographical distribution of mean annual hail frequency in China during 1961–2005. The contour interval is 0 (dashed line), 0.5, 1 (thick line), 2, 3. . . . Terrain is shaded, with scale bar on the right (m).
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1603.1
The geographical distribution of mean monthly hail frequency (hail days per month) during 1961–2005 (January–December).
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1603.1
(Continued)
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1603.1
The geographical distribution of the hail peak month in China during 1961–2005. Roman numbers (I, II, III, IV, V) in circles are the number of hail peak types defined in the text. Regular numbers with 5 digits are the WMO identification of stations whose monthly hail frequencies from 1961 to 2005 will be shown in Fig. 5.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1603.1
The mean annual monthly hail days during 1961–2005 at example stations in different regions, as shown in Fig. 4: types (a) I, (b) II, (c) III, and (d) IV.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1603.1
The geographical distribution of the maximum hail occurrence period in a day. Four periods are defined at 0000–0600, 0600–1200, 1200–1800, and 1800–0000 LT. Numbers are the WMO identification of three stations whose mean diurnal variation of hail frequency during 1961–2005 will be shown in Fig. 7.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1603.1
The diurnal variation of the hail ratio for typical stations that have a high frequency of nighttime hail occurrences. The hail ratio is defined as hail occurrences in each hour divided by total hail occurrences for a station during 1961–2005. The loci of stations are plotted in Fig. 6.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1603.1
