Spatiotemporal distributions and influences on snow density in China from 1999 to 2008

LiYun Dai , Tao Che

1. Cold and Arid Regions Remote Sensing Observation System Experiment Station, Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou, Gansu 730000, China

2. Graduate University of the Chinese Academy of Sciences, Beijing 100039, China

Spatiotemporal distributions and influences on snow density in China from 1999 to 2008

LiYun Dai1,2, Tao Che1*

1.Cold and Arid Regions Remote Sensing Observation System Experiment Station, Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou, Gansu 730000, China

2.Graduate University of the Chinese Academy of Sciences, Beijing 100039, China

Ground snow observation data from 1999 to 2008 were used to analyze the temporal and spatial distribution of snow density in China. The monthly maximum density shifted from north to south during the period from October to the following January, and then moved back from south to north during the period from January to April. The maximum snow density occurred at the border between Hunan and Jiangxi provinces in January, where snow cover duration was short and varied remarkably. Snow density in Northeast China and the Xinjiang Uygur Autonomous Region were also high and showed less variation when the snow cover duration was long. Ground observation data from nine weather stations were selected to study changes of snow density in Northeast and Northwest China. A phase of stable snow density occurred from the middle ten days of November to the following February;non-stationary density phases were observed from October to the first ten days of November and from March to April. To further investigate the effects of climatic factors on snow density, correlations between snow density and precipitation, air temperature,snow depth and wind velocity for Northeast and Northwest China were analyzed. Correlation analysis showed that snow depth was the primary influence on snow density.

snow; density; spatiotemporal distribution

1. Introduction

Snow density is one of the key factors for retrieving snow depth or snow water equivalent (SWE) using passive microwave remote sensing data. Many researchers used passive microwave brightness temperature difference of 19 GHz and 37 GHz to retrieve snow depth (Cheet al., 2004,2008; Tedesco and Narvekar, 2010). These algorithms are based on the fact that the difference will increase with the increase of snow depth. However, when snow depth remains unchanged, the difference will decrease with the increase of snow density. So, it is very important to understand the temporal and spatial distribution of snow density when retrieving snow depth from passive microwave satellite data.

Snow density is the mass per unit volume of snow(g/cm3), and the product of snow density and snow depth(cm) is snow pressure, which is defined as mass per unit area of snow (g/cm2). Hence, snow density can be defined as the quotient of snow depth and snow pressure. Snow density is one of the fundamental and significant physical parameters used in water balance research, snowmelt runoff modeling and avalanche forecasting (Margreth, 2007; Lazar and Williams, 2008). Understanding the temporal-spatial distribution of snow density is important for studying the temporal-spatial distribution of snow water resources and understanding the practical use of snow water resources.

There have been many studies on the temporal-spatial distribution of snow cover (Haoet al., 2009; Wanget al.,2009). However, only a few studies of snow density have been performed. Yang studied seasonal snow distribution and density variation of a mountain at the source of the Urumqi River (Yanget al., 1992a, b); Li Zhen successfully retrieved dry snow density information using airborne polarized radar data (Liet al., 2001); Huang and his colleagues analyzed the temporal-spatial distribution of snow density and the factors that affected snow density in the Xinjiang Uygur Autonomous Region (Huanget al., 2007).

China is a vast country that spreads across tropical and temperate zones. Snowfall occurs in most regions during spring and winter, and snow density is related to precipitation, air temperature, snow depth and wind velocity (Yanget al., 1992b; Huanget al., 2007). This study utilized snow ground observation data for the years from 1999 to 2008 to analyze the temporal-spatial distribution of snow density in China and to understand snow density characteristics in different regions.

2. Data and methods

In this study, we used daily snow depth and snow pressure data from 756 weather stations from 1999 to 2008. According to Che (Cheet al., 2008), snow occurred from September to June of the following year in Tibet; however,weather stations in the Tibetan Plateau are rare and not representative of the entire region. Furthermore, snowfall occurs mainly in the southeastern, western and southern mountains of Tibet, but weather stations are distributed in flat areas (Figure 1). Therefore, it is meaningless to analyze snow characteristics in Tibet using data from these stations,and this region was not considered in this paper.

Snow depth was observed every day and snow pressure was observed once every five days. When snow depth was less than 5 cm, snow pressure was not recorded. Snow density data were obtained from depth and pressure parameters,and snow density data were available every five days from 1999 to 2008. Ten-year monthly average values of snow depth and snow density were calculated for each station.Finally, monthly average snow depth and snow density data for the entire country were calculated using a neighboring interpolation method.

Figure 1 Distribution of snow stations and averaged snow depth in January

3. Monthly average snow density characteristics in China for 1999-2008

China extends over 40 degrees in latitude from north to south, and elevation ranges from 0 m in the east to over 5,000 m in the west. Northeast China borders the ocean,whereas the northwestern region is far from the ocean. The snow season begins in different times across the regions,therefore the temporal and spatial characteristics of snow are also different in various regions.

In China, snowfall generally begins in October and ends in the following April. In September, there is a slight amount of snow in the Northeast, the Xinjiang Uygur Autonomous Region and Tibet, although snow depths are so small that no snow pressure is observed. Based on monthly average snow depth and snow density data from all weather stations, the neighboring interpolation method was used to obtain monthly average snow depth and snow density distribution figures for China from 1999 to 2008(Figure 2). Snow depth and snow pressure data originated from station observations, and snow density was the quotient of the two parameters. Therefore, density was the mass per unit volume of snow including dry snow, melted water and water vapor (Sun, 2005).

Figure 2 Distribution of ten-year monthly averaged snow depth and snow density from December to the following April during 1999 and 2008 (left: snow density (g/cm3); right: snow depth (cm)). --to be continued

Figure 2 Distribution of ten-year monthly averaged snow depth and snow density from December to the following April during 1999 and 2008 (left: snow density (g/cm3); right: snow depth (cm))

According to Figure 2, snow cover, represented by snow depth, was generally consistent with snow density. However,snow pressure was not measured when snow depths were less than 5 cm, therefore snow density distribution range was slightly smaller than the distribution of snow depth.

In October, snow was mainly distributed in Northwest and Northeast China and the Tibetan Plateau, where average monthly snow depths were less than 1 cm, except for a few parts of Da Hinggan Ling, Xiao Hinggan Ling and the southeastern region of Qinghai Province, where snow depth was over 2 cm and exceeded 5 cm in some areas. Snow density was approximately 0.03 g/cm3. Figure 2 also shows that there was very shallow snow in North China in October,although snow density was over 0.1 g/cm3.

In November, the deepest snow of about 13 cm was observed in the northern region of Northeast China. The next deepest snow was recorded in the northern region of Xinjiang Uygur Autonomous Region, followed by other locations within Northeast China, Hebei Province and the eastern region of Qinghai Province. However, the maximum snow density of over 0.1 g/cm3occurred in North China and the northern region of Northeast China, where snow depths were not the deepest. Liquid water content within snow was much higher in North China than in other locations, which led to high snow density values in this region. The northern area of Xinjiang Uygur Autonomous Region and the eastern area of Northeast China had snow densities of 0.06-0.1 g/cm3.

All of China entered into the snow season during December. Snow depths were up to 20 cm in Northeast China and the north of Xinjiang Uygur Autonomous Region, and were over 5 cm in many places; snow depths were approximately 1 cm in the eastern and central regions of China. The maximum snow density occurred at the boundary of Hunan and Hubei provinces, which had abundant precipitation and density values near 0.2 g/cm3. The northern region of Northeast China and the Xinjiang Uygur Autonomous Region had density values between 0.1 g/cm3and 0.15 g/cm3.

In January, the maximum snow depth of over 30 cm occurred in the northern region of the Altai Mountains. The next deepest snow occurred in Da Hinggan Ling and Xiao Hinggan Ling. The maximum snow density of over 0.2 g/cm3was observed in the northern region of the Altai Mountains and the boundary of Hunan and Hubei provinces;density values at Da Hinggan Ling and Xiao Hinggan Ling were between 0.1 g/cm3and 0.15 g/cm3.

Snow began to melt in February throughout the country,with the exception of Da Hinggan Ling and Xiao Hinggan Ling. There was no snow in the southeastern and central regions of China in February. The maximum snow depth of over 30 cm was observed in the northern region of the Altai Mountains; Northeast China and other places within the Xinjiang Uygur Autonomous Region had snow depths of over 20 cm. The maximum snow densities of 0.15-0.17 g/cm3occurred at the Altai Mountains, Da Hinggan Ling and Xiao Hinggan Ling; however, snow density was greater than 0.1 g/cm3at most locations. In regions at south of the Yangtze River, where had occasional light snow, snow was very wet with densities greater than 0.1 g/cm3.

In March, snow had melted in most locations. The maximum snow depth was greater than 20 cm in some areas of the Altai Mountains, Da Hinggan Ling and Xiao Hinggan Ling, and there was occasional snow in the Huainan area.The maximum snow density of over 0.1 g/cm3occurred in the middle of Henan Province. In April, snow existed only in the northern part of the Xinjiang Uygur Autonomous Region and Northeast China, where snow depths were approximately 2 cm and snow densities were between 0.13 g/cm3and 0.17 g/cm3.

In summary, the area of maximum snow density shifted from the northern regions in October to the southern regions in the following January, and moved back to the northern regions from January to April. Snow density values were very large in the eastern regions of China, where the air was wet and the precipitation was much greater than in other areas, which led to high liquid water content and large snow density. In Da Hinggan Ling, Xiao Hinggan Ling, and the northern region of the Xinjiang Uygur Autonomous Region,snow depths were deep and density values were also large.

Although there are many snow cover areas in the Tibetan Plateau, there are few snow observation stations in this region. Other two main snow cover regions are Northeast and Northwest China. In this study, nine stations were selected in Northeast and Northwest China to analyze the variance of snow density characteristics in detail (Table 1). Average and maximum snow depth and snow density data were calculated at ten-day intervals from October to the following April during 1999 and 2008. Snow pressure data were not available when snow depths were less than 5 cm; therefore snow depth and density data points where snow density was not observed were not used in calculations, and the values were considered as zero.

Table 1 Location of nine selected stations and their ID

Figure 3 Changes of averaged snow depth and snow density of every ten days at selected stations from October to the following April in northeastern China

Figures 3 and 4 show the ten-day average snow depth and density variation trends from October to the following April during 1999 and 2008 in Northeast and Northwest China, respectively. The average snow density increased steadily from November to the following March; the density values were between 0.1 g/cm3and 0.2 g/cm3in both regions, and the maximum value varied from 0.15 g/cm3to 0.35 g/cm3. During this period, snow depths also increased steadily, and the number of snow days increased markedly.The stations entered into the snow season at different dates:it began to snow at some stations in October, whereas at other stations, the first snow was observed in November.During the middle of March to April, snow densities decreased abruptly as snow depths and the number of snow days decreased, which is the melting period. Some stations entered into the melting period during the first ten days of April, and melting began at some stations during the last ten days of April. These two periods were the unstable snow phases.

Figure 4 Changes of averaged snow depth and snow density of every ten days of selected stations from October to the following April in northwest China

4. Analysis of climatic impacts on snow density in Northeast and Northwest China

Snow density is strongly impacted by climatic factors(Wanget al., 2009), such as precipitation, air temperature and wind velocity. To further investigate the influence of environmental parameters on snow density, a correlation analysis was performed using snow density, precipitation, air temperature, snow depth and wind velocity information.Data from 126 weather stations in Northeast China and 102 stations in Northwest China were selected for this analysis.

These data included daily snow depth, snow pressure,average air temperature, precipitation and average wind velocity from 1999 to 2008. Ten-day average values of snow density, snow depth, average air temperature, precipitation and average wind velocity data were calculated for the period from October through the following April during these ten years. Where the snow density was equal to zero, snow density and the data of other corresponding parameters were excluded; these data were then normalized to maintain consistent units for these parameters, then we did a stepwise regression analysis of normalized snow density, snow depth,average air temperature, precipitation and wind velocity values to obtain the following equations:

whereZdenis the normalized snow density,Zdepis the normalized snow depth,Zpreis the normalized precipitation andZtemis the normalized air temperature.

The regression equations show that the extent to which snow density was impacted by climate factors varied in different regions. In Northeast China, snow density was mainly affected by snow depth, and the regression coefficient was 0.886. Deeper snow corresponded with higher snow pressure and accelerated snow densification. Air temperature and precipitation had weaker influences on snow density than on snow depth. Snow melts due to a rise in air temperature, and melted water flowing down into the snow increases snow density. In this region, wind velocity had little impact on snow density.

In Northwest China, snow depth had the largest effect on snow density, with a regression coefficient of 0.814; snow depth was followed by precipitation and air temperature.This region has a dry and cold climate (Weiet al., 2001); air temperature is low during snow periods, the climate is dry,and melted water evaporates quickly. Therefore, a rise in air temperature had little impact on snow densification. Wind velocity was excluded from the stepwise regression analysis.

5. Conclusion

Based on monthly average snow density data from October in 1999 to April in 2008, snow density distribution was spatially inconsistent in China. There were three regions that had high density values: Da Hinggan Ling, Xiao Hinggan Ling, the Altai Mountains, and regions south of the Yangtze River. Temporally, the maximum snow density moved toward the south from October to the following January. From January to April, snow melted in South China, and the maximum snow density was observed in North China,where snow density was mainly influenced by snow depth.Spatially, snow density had varied characteristics in different regions. The overall variance trend was obvious in Northwest and Northeast China. The phase of stable snow density occurred from the second ten-day period of November to the first ten-day period of March; unstable phases occurred from October to the first ten-day period of November and from the second ten-day period of May to April. Snow density increased and decreased abruptly in other locations due to short snow durations.

The effect of climatic factors on snow density was dissimilar in different regions based on correlation analyses between snow density and snow depth, precipitation, air temperature and wind velocity. On a regional scale, snow depth had the greatest contribution to snow density in Northeast and Northwest China, and wind velocity had little impact on snow density. However, snow density may be influenced locally by other factors, such as wind velocity,the number of days with blowing snow, elevation (Lazar and Williams, 2008; Wanget al., 2009), sub-surface vegetation types or air moisture (Liet al., 2002).

The data in this study were provided by China Meteorological Administration. The work was funded by the China State Key Basic Research Project (No.2007CB411506,No.2007CB714403) and the Natural Science Foundation of China (No.40601065, No.40971188).

Che T, Li X, Gao F, 2004. Estimation of snow water equivalent in the Tibetan Plateau using passive microwave remote sensing data (SSM/I).Journal of Glaciology and Geocryology, 26(3): 363-368.

Che T, Li X, Jin R, Armstrong R, Zhang TJ, 2008. Snow depth derived from passive microwave remote-sensing data in China. Annals of Glaciology,49: 145-154.

Hao XH, Wang J, Che T, Zhang P, Liang J, Li HY, Li Z, Bai YJ, Bai Y, 2009.The spatial distribution and properties of snow cover in Binggou Watershed, Qilian Mountains: Measurement and Analysis. Journal of Glaciology and Geocryology, 31(2): 284-292.

Huang WJ, Huang Z, Cui CX, Feng ZM, Li Q, Wang L, 2007. Temporal and spatial distribution of Xinjiang snow density and its consequence. Journal of Glaciology and Geocryology, 29(1): 66-72.

Lazar B, Williams M, 2008. Climate change in western ski areas: potential changes in the timing of wet avalanches and snow quality for the Aspen ski area in the years 2030 and 2100. Cold Regions Science and Technology, 51(2-3): 219-228.

Li J, Wang XH, Cai TJ, Guan ZH, Wang XM, 2002. The effect of snow-melting in the watershed of Xigou reservoir. Journal of Northeast Forestry University, 30(2): 69-72.

Li Z, Guo HD, Shi JC, 2001. Retrieving dry snow density with SIR-C polarimetric SAR data. Chinese Science Bulletin, 46(14): 1211-1214.

Margreth S, 2007. Snow pressure on cableway masts: analysis of damages and design approach. Cold Regions Science and Technology, 47(1-2):4-15.

Sun SF, 2005. The Physical, Biochemistry Mechanism and Parametric Models of Land Process. Beijing: China Meteorological Press, 104.

Tedesco M, Narvekar PS, 2010. Assessment of the NASA AMSR-E SWE Product. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 3(1): 141-159.

Wang CH, Wang ZL, Cui Y, 2009. Snow cover of China during the last 40 years: spatial distribution and interannual variation. Journal of Glaciology and Geocryology, 31(2): 301-310.

Wei WS, Qin DH, Liu MZ, 2001. Properties and structure of the seasonal snow cover in the northwest regions of China. Arid Land Geography,24(4): 310-313.

Yang DQ, Wang CZ, Zhang YS, Zhang ZZ, 1992a. Distribution of seasonal snow cover and variation of snow density on the headwater of Urumqi River Basin. Geographical Research, 12(4): 87-96.

Yang DQ, Zhang YS, Zhang ZZ, 1992b. A study on the snow density in the head area of Urumqi River Basin. Acta Geographica Sinica, 47(3):260-266.

10.3724/SP.J.1226.2011.00325

*Correspondence to: Tao Che, Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences. No.320, West Donggang Road, Lanzhou, Gansu 730000, China. Tel: +86-931-4967966; Email: chetao@lzb.ac.cn

21 February 2011 Accepted: 28 April 2011