LiJuan Ma,DaHe Qin

1.National Climate Center,Beijing 100081,China

2.China Meteorological Administration,Beijing 100081,China

3.State Key Laboratory of Cryospheric Sciences,Cold and Arid Regions Environmental and Engineering Research Institute,Lanzhou,Gansu 730000,China

Temporal-spatial characteristics of observed key parameters of snow cover in China during 1957–2009

LiJuan Ma1*,DaHe Qin2,3

1.National Climate Center,Beijing 100081,China

2.China Meteorological Administration,Beijing 100081,China

3.State Key Laboratory of Cryospheric Sciences,Cold and Arid Regions Environmental and Engineering Research Institute,Lanzhou,Gansu 730000,China

Using observed snow cover data from Chinese meteorological stations,this study indicated that annual mean snow depth,Snow Water Equivalent (SWE),and snow density during 1957–2009 were 0.49 cm,0.7 mm,and 0.14 g/cm3over China as a whole,respectively.On average,they were all the smallest in the Qinghai-Tibetan Plateau (QTP),and were greater in northwestern China(NW).Spatially,the regions with greater annual mean snow depth and SWE were located in northeastern China including eastern Inner Mongolia (NE),northern Xinjiang municipality,and a small fraction of southwestern QTP.Annual mean snow density was below 0.14 g/cm3in most of China,and was higher in the QTP,NE,and NW.The trend analyses revealed that both annual mean snow depth and SWE presented increasing trends in NE,NW,the QTP,and China as a whole during 1957–2009.Although the trend in China as a whole was not significant,the amplitude of variation became increasingly greater in the second half of the 20th century.Spatially,the statistically significant (95%-level) positive trends for annual mean snow depth were located in western and northern NE,northwestern Xinjiang municipality,and northeastern QTP.The distribution of positive and negative trends for annual mean SWE were similar to that of snow depth in position,but not in range.The range with positive trends of SWE was not as large as that of snow depth,but the range with negative trends was larger.

snow cover;snow density;snow depth;snow water equivalent;climate change

1.Introduction

Snow cover is one of the main components of the cryosphere.Its impacts on climate,water resources,and the environment should not be ignored.Snow cover is widely,but unevenly distributed in China.The characteristics of snow cover in China,especially over the Qinghai-Tibetan Plateau (QTP),were researched broadly since ground-based observations were recorded.Limited by the parameters and length of observations,most researchers focused on snow depth and the number of snow cover days before retrievable data from satellites were available (Li and Mi,1983).The research of snow cover extent and snow depth for the regions without or with less meteorological stations,such as the central and western QTP,were enriched by the improved accuracy of the National Oceanic and Atmospheric Administration (NOAA) weekly snow cover extent data since 1973 and the distribution of retrieved snow depth data from Scanning Multichannel Microwave Radiometer (SMMR) and Special Sensor Microwave Imager (SSM/I) (Qinet al.,2006).However,remote sensing data cannot accurately describe the features of snow cover in China,especially in western China due to the complicated underlying surface.Although an advanced retrieval algorithm for snow depth in China was proposed by correcting the algorithm for global SWE using part of observing snow depth data,problems stillexist,e.g.,overestimation over the QTP (Cheet al.,2008).Atmospheric layer thickness between surface and satellite over the QTP is far less than that in other regions due to high altitude of the QTP.As a result,performing atmospheric corrections can effectively decrease the overestimation of satellite-retrieved snow depth data (Savoieet al.,2009).However,since there is no atmospheric-corrected snow cover production in China at present,observation data recorded by surface instruments become the main data analyzing changes of snow cover in China (Wanget al.,2009).

At present,the analyses of temporal-spatial characteristics of snow cover focus on snow depth and the number of snow cover days,but seldom on snow water equivalent(SWE).There are two ways in which snow cover feeds back to the climate.One is to impact surface energy balance through high albedo of snow cover.The other is to delay the rising of surface temperature by affecting soil moisture.The latter relates to SWE and is crucial.There is no direct record of SWE in Chinese meteorological stations,although snow pressure is observed every five days,which can be converted to SWE together with snow depth.Therefore,characteristics of snow depth,SWE,and snow density in China are analyzed by using observation data from surface meteorological stations.

2.Data sets and pre-processing

The data sets used in this study include daily snow depth(cm) and pentad snow pressure (g/cm2),which were obtained from the National Meteorological Information Center,China Meteorological Administration.These data sets are available for 754 Chinese meteorological stations covering the period from the beginning of records (approximately January 1,1951) until December 31,2009.Pentad snow pressure was recorded in the 5th,10th,15th,20th,25th,and the last day of the month when snow depth is equal or greater than 5 cm (CMA,2003).

The data were pre-processed before analyses.According to the specifications for surface meteorological observation,the value was recorded as trace amount when snow depth was less than 0.5 cm,and hence,"trace" records of daily snow depth were all reassigned to a value of 0.5 cm.For daily snow density data,pentad snow pressure and snow depth in the corresponding date was first converted to pentad snow density (g/cm3).Then,for other days with snow cover,the value of snow density closest to the selected date was supplied in consideration of little change of snow density.Daily SWE was hence computed by daily snow density and daily snow depth.According to the 30-year climatology convention,this study took 1971–2000 as climatology period,and rejected stations with effective data less than 20 years in this period to ensure the continuity of the selected data.After rejection,the number of stations was 669 for annual data,and they were 669,668,669,and 670 for autumn,winter,spring,and summer.According to the feature of snow season,we took the period from September 1 of the previous year to August 31 in the current year as a whole snow year.The previous September,October,and November composed the current autumn,the previous December,current January and February composed the current winter,and so on.Each snow year began with autumn and ended with summer.In addition to snow cover over China as a whole,we also analyzed some regions with stable snow cover,such as northeastern China including eastern Inner Mongolia (NE hereafter;40°–55°N,115°–135°E),northwestern China (NW hereafter;40°–50°N,73°–105°E),and the QTP (25°–40°N,73°–105°E;elevation ≥2,000 m).

The following principles were abided by when computing monthly mean snow depth/SWE for one station.If the days with effective values were no less than 20 days in a month,we obtained the monthly mean snow depth/SWE by averaging all effective values in this month;but if they were less than 20 days,we recorded as a missing value.For seasonal mean value,if the monthly mean value was missing in any month of the season,it was recorded as missing.Otherwise,we averaged all the three monthly mean values and obtained the seasonal mean value.Similarly,if the seasonal mean value was missing in any season of the year,the annual mean value was recorded as missing.Otherwise,we averaged all the four seasonal mean values and obtained the annual mean value.When computing mean snow depth/SWE over a region,it was recorded as missing when the number of stations with effective data was less than half of the total.Since there is no perfect objective interpolation method for snow cover in China,this study applied the arithmetic average method to avoid huge biases being introduced when performing interpolation.After the aforementioned pre-processing,the time series of snow cover data began with 1957 approximately.

Additionally,there exists a certain deficiency although station observation data are considered as accurate first-hand data all along.For example,the data will be affected by the relocation of stations and the individual difference in operation of meteorological equipment.Moreover,the spatial representativeness of ground-based observation data is defective because of the uneven distribution of stations.Over the QTP,the meteorological stations are distributed mainly in the middle and eastern sections,and consequently,the mean value over the QTP mentioned in this study implies the situation in these sections.

3.Results

3.1.Temporal variations

We firstly analyzed the inter-monthly variations of snow cover in NE,NW,the QTP,and China as a whole during 1957–2009.Figure 1 shows the inter-monthly variations of multi-year mean snow depth,SWE,and snow density.Generally,the characteristics of inter-monthly variations for snow depth and SWE were similar,but differed from that of snow density.Snow cover duration in China mainly began in October and ended with the next April,and it was slightly different in the three aforementioned regions.Snow coverduration was the longest in the QTP,from October to the next May,and snow cover might also occur during June and September.Over China as a whole,annual mean snow depth and SWE were 0.49 cm and 0.7 mm,respectively.Monthly mean snow depth was greatest in January,with mean value of 1.71 cm,but SWE in February,with mean value of 2.5 mm,was very close to that in January due to higher snow density.Regionally,mean snow depth,SWE,and snow density were all the smallest in the QTP,and were greater in NW.In October and November when snow began accumulating,snow depth and SWE in NE were the greatest in these three regions,but after that period,they were greatest in NW until the fully melting of snow cover appeared.The time occurring peak values of mean snow depth and SWE were consistent in all regions.For NE,they reached the maxima,5.47 cm and 8.0 mm,in January,and the peak values for NW,6.74 cm and 11.0 mm,and for the QTP,0.72 cm and 1.1 mm,all occurred in February.However,they were different for snow density.For NE and NW,it reached maximum in March and April,respectively,with values of 0.15 g/cm3and 0.17 g/cm3.Monthly mean snow density for the QTP was the first and the second largest in June and September,respectively,with values of 0.17 g/cm3and 0.14 g/cm3,and they were considerable in other months.For China as a whole,the inter-monthly variation was not apparent,and the annual mean value was 0.14 g/cm3.

Figure 1 Inter-monthly variations of multi-year mean (a) snow depth,(b) snow water equivalent,and (c) snow density over China as a whole,northeastern China,northwestern China,and the Qinghai-Tibetan Plateau during 1957–2009

The inter-annual variations of mean snow cover were further analyzed.The multi-year mean snow depth and SWE were 0.49 cm and 0.7 mm,and the minimum and maximum occurred in 1963 and 2001,with values of 0.33 cm and 0.82 cm,respectively.However,SWE in the corresponding years was not the greatest due to the inter-annual variation of snow density.Instead,the minimum and maximum occurred in 1976 and 1957,respectively,with values of 0.4 mm and 1.3mm.The multi-year mean values of annual mean snow depth and SWE for NE,NW,and the QTP were 1.57 cm and 2.2 mm,1.85 cm and 3.0 mm,and 0.29 cm and 0.4 mm,respectively.The amplitudes of variation in NE,NW,and the QTP in the past 53 years were 0.77–2.90 cm,0.86–3.42 cm,and 0.16–0.62 cm for snow depth,respectively,and were 1.0–4.4 mm,1.3–6.3 mm,and 0.2–1.2 mm for SWE.The linear trends are presented in Table 1.In annual scale,snow depth and SWE presented insignificantly increasing trends in all regions.In seasonal scale,the trends of annual mean snow depth and SWE over China as a whole were all negative except for insignificantly positive trends in winter,but they only reached 95% significant level in spring,with values of -0.02 cm/10yrs and -0.04 mm/10yrs.Regionally,only snow depth in NW in winter and in the QTP in summer presented significantly positive and negative trends,with values of 0.34 cm/10yrs and -0.002 cm/10yrs,respectively.For SWE,only negative trends over the QTP in spring and summer were significant,with values of -0.02 cm/10yrs and-0.004 cm/10yrs,respectively.

Table 1 Linear trend coefficients for annual and seasonal mean snow depth and snow water equivalent over China as a whole,northeastern China,northwestern China,and the Qinghai-Tibetan Plateau during 1957–2009

We standardized the time series of annual mean snow depth and SWE to present their amplitude of change.As presented in Figure 2,the variation of snow depth was consistent with that of SWE in all regions during 1957−2009.For China as a whole,the amplitude of variation increased gradually in the second half of the 20th century,and began decreasing in the first 10 years of the 21st century.Generally,the amplitude was greater in the period when snow cover was more than normal.However,the number of years with positive anomalies was less than that with negative anomalies.The proportions of positive to negative snow depth departures for China as a whole,NE,NW,and the QTP were 23:31,24:32,22:30,and 16:36,respectively.For SWE,the corresponding proportions were 25:29,26:30,22:30,and 15:37,respectively.

To quantify the impact of regional snow cover on snow cover in China as a whole,a linear regression analysis was performed on the standardizing time series of snow depth and SWE.Table 2 shows the regression coefficients and constants.It is clear that all constants were small,indicating that changes of snow cover for China as a whole can be determined by snow cover changes in NE,NW,and the QTP.As seen from the annual regression coefficients,annual mean snow depth changes in China as a whole can be explained by 76%,46%,and 20% of respective changes in NE,NW,and the QTP together.

Figure 2 Standardized departures of annual mean snow depth and snow water equivalent over (a) China as a whole,(b) northeastern China,(c) northwestern China,and (d) the Qinghai-Tibetan Plateau in 1957–2009 with respect to 1971–2000.

For SWE,the corresponding coefficients were 71%,50%,and 21%,respectively.This indicated that snow cover changes in China were closest to changes in NE,and their correlation coefficients were also the greatest in all regions,with values of 0.85 and 0.81 for snow depth and SWE,respectively (Table 3).The characteristics in autumn,winter,and spring,were similar with that in annual scale: snow cover changes in China were mainly determined by changes in NE and were hardly affected by changes in the QTP.In summer,however,snow cover changes in China were affected mainly by snow cover over the QTP because the main snow-covered field is located there.This was confirmed by the correlation analyses between snow cover in China and in these three regions,as presented in Table 3.No matter for snow depth or SWE,the correlations between snow cover in China,NE and NW in annual,autumn,winter,and spring were all significant,and the correlation coefficient was the greatest between snow cover in China and NE.In summer,snow cover in China was significantly correlated with those in the QTP and NW,and was insignificantly correlated with that in NE.It is noted that although the correlation between snow depth in China and in the QTP was not significant in autumn,it was significant for SWE.This is possibly related to higher snow density over the QTP in autumn,especially in September and October (Figure 1).Correspondingly,26%of SWE changes in the QTP were related to changes in China as a whole.The ratio was slightly higher than that of snow depth,which was 22% (Table 2).

Table 2 Fitting coefficients for annual and seasonal mean snow depth and snow water equivalent over China as a whole by using corresponding standardized values in northeastern China,northwestern China,and the Qinghai-Tibetan Plateau during 1957–2009

Table 3 Correlation coefficients between China and different regions for annual and seasonal snow depth and snow water equivalent during 1957–2009

3.2.Spatial variations

As presented in Figure 1,there exist obvious inter-monthly variations of snow cover in China.The spatial distributions of annual and seasonal mean snow density in 1971−2000 were first exhibited in Figure 3.Generally,annual mean snow density was below 0.14 g/cm3in most of China,and was higher in NE,NW,and the QTP.It was 0.14−0.18 g/cm3in most of Xinjiang municipality,northern NE,and middle and northwestern QTP.Larger values,basically between 0.18 g/cm3and 0.26 g/cm3,distributed mainly in South China and south-central QTP,and the maximum,0.43 g/cm3,occurred in central Yunnan Province.Seasonally,the spatial pattern of snow density was similar with that in annual scale,but with obvious seasonal evolution.In autumn,snow density was below 0.14 g/cm3in most of China,including northern parts of Xinjiang municipality and NE.It was 0.14−0.18 g/cm3in southern Xinjiang municipality and most of the QTP,with sections reaching 0.30 g/cm3.In winter,snow density in most of Xinjiang municipality and northern NE increased,and it was the greatest in western Yunnan Province,with maximum of 0.43 g/cm3.In spring,the region with snow density of 0.14−0.18 g/cm3enlarged in Xinjiang municipality and NE,and the maximum of 0.56 g/cm3occurred in southern Guizhou Province.In summer,the area covered by snow shrank tremendously,distributed mainly in the QTP.The maximum was located in central Sichuan Province,with a value of 0.36 g/cm3,and the second largest region was located in south-central QTP and Mount Tianshan.Generally,seasonal mean snow density was most stable in the QTP,and was variable in both NE and NW.

Figure 3 Spatial distribution of mean snow density in China for annual,autumn,winter,spring,and summer in 1971–2000

For spatial distribution of snow depth and SWE,their seasonal changes were almost synchronous (Figure 4).In annual scale,mean snow depth/SWE was greater in NE,northern Xinjiang municipality,and a small part of southwestern QTP,and was below 1 cm/1 mm in most of the other regions.For snow depth,the maxima in NE,NW,and the QTP were 6.92 cm,9.06 cm,and 4.71 cm,respectively,and were 9.9 mm,15.9 mm,and 10.5 mm for SWE.It is thus clear that maximum snow depth in the QTP was smaller than that in NE,but maximum SWE was greater due to higher snow density (Figure 3).In autumn,the zonal distribution of snow cover was obvious,and areas with larger values are located in northern NE and the northern edge of NW.In winter,snow cover increased substantially.For snow depth,the maxima were 18.88 cm,25.83 cm,and 13.41 cm in NE,NW,and the QTP,respectively,and were 26.5 mm,44.7 mm,and 30.3 mm for SWE.Moreover,the distribution pattern of snow cover was highly related to the local topography in NE and NW.The area with larger values in NE presented an inverted U-shape surrounding the Northeast Plain,and extended to north of Mount Tianshan in NW.In spring,snow cover in NE,NW,and the QTP all decreased substantially,the maxima were 6.95 cm,8.91 cm,and 4.85 cm for snow depth,respectively,and were 11.1 mm,18.2 mm,and 10.5 mm for SWE.In summer,snow cover only existed in parts of the QTP and Mount Tianshan due to their higher relief,but the magnitude was no less than 1 cm or 1 mm.

To signify the variability of snow cover with respect to 1971−2000,we further computed standardized departures of annual and seasonal mean snow depth and SWE in 1980–1989,1990–1999,2000–2009,and 1957–2009.Figure 5 shows the situation in annual scale.Generally,the negative amplitude was greater than the positive amplitude in the past 30 years.For both snow depth and SWE,the extent with positive anomalies was greater in an early stage,but the extent with negative anomalies became greater in the later stage.This indicated that,with climate warming,the range with decreasing snow cover gradually enlarged.In the 1980s(Figures 5b1and 5b2),the area with positive departures of snow depth was more extensive than that with negative departures,but on the contrary,the amplitude of positive departures was smaller than that of negative departures.The regions where snow depth was less than normal were located mainly in most of NE,part of northern North China,northern Xinjiang,central QTP,and southeastern Southwest China.The regions with more snow were located mainly in eastern Xinjiang,part of eastern QTP,most of North China,and most regions south of the Yangtze River.The pattern of SWE variability was consistent with that of snow depth expect for more distinct negative departures in central NE.In the 1990s,the variation of both snow depth and SWE was basically opposite to that in the 1980s.The region with more snow in former decades was covered by negative departures in later decades,andvise versa.Especially,decreases of snow cover in southern Xinjiang were considerable.Compared with that in the 1990s,the region with negative departures was more extensive in 2000−2009.Snow cover was less than normal in most of China,especially in most of Yunnan Province,except for part of northern Xinjiang,central and part of northern NE,part of northern North China,and parts of eastern and northern QTP.

Figure 4 Spatial distribution of snow depth (left) and snow water equivalent (right) in China for annual,autumn,winter,spring,and summer in 1971–2000

For the whole examined period,1957−2009,the multi-year mean of snow depth and SWE anomalies was positive in most of northern China except for parts of eastern Inner Mongolia and Xinjiang,but it was negative in most of southern China including the QTP.However,the variability of both snow depth and SWE was smaller with ±0.4 standard deviation at most.The maximum positive and negative variability of annual mean snow depth was 0.40 and −0.38,located in NE and the QTP,respectively.The corresponding values were 0.36 and −0.30 for SWE,both located in the QTP.As presented in Figures 5a1and 5a2,the multi-year mean of snow cover anomalies in the QTP was negative.This was consistent with what was presented in Figure 2d that snow cover on the QTP exhibited negative departures in two thirds of all years.However,the corresponding departures were basically positive in NW.This is,on the one hand,related to the comparative number of years with opposite signs,on the other hand,related to the greater positive variability (Figure 2c).Seasonally,the variability of snow depth and SWE remained consistent(figure not presented).The spatial distribution of snow cover variability in winter was similar to that in annual scale.In autumn,the range and amplitude of negative departures in NE were generally greater than those in annual scale.On the contrary,they were smaller in autumn than those in annual scale for the QTP.The negative departures in Central China in annual scale even turned into positive departures in autumn.In spring,the range and amplitude of negative departures were generally greater than those in annual scale for NW and the QTP,but the negative departures in NE in annual scale almost turned into positive departures in autumn.

Figure 5 Standardized departures of annual mean snow depth and snow water equivalent in 1980–1989,1990–1999,2000–2009,and 1957–2009 with respect to 1971–2000

Figure 6 presents the spatial distribution of linear trends of annual and seasonal snow depth and SWE during 1957−2009.In annual scale,the significant positive trends for snow depth are located mainly in eastern Inner-Mongolia,northern Northeast China,northwestern Xinjiang,and northeastern QTP,and the significantly negative trends are distributed mainly in southeastern Northeast China,most of North China,and southern QTP.The maximum positive and negative trends for snow depth were 0.11 cm/yr in northwestern Xinjiang and−0.05 cm/yr in southeastern Northeast China,respectively.

Figure 6 Trends for annual and seasonal snow depth (left) and snow water equivalent (right) in 1957–2009

For SWE,the position of significant trends was similar to that of snow depth,but the range was not very consistent.The range with positive trends was not as wide as that of snow depth,but the range with negative trends was wider.The maximum and minimum trends of SWE are located in the same regions with those of snow depth,with values of0.20 mm/yr and −0.30 mm/yr,respectively.In autumn,snow depth and SWE presented consistent decreasing trends distributed sporadically in parts of Xinjiang,NE,North China,and the QTP.The maximum negative trends were −0.04 cm/yr and −0.10 mm/yr.In winter,the distribution pattern of trends was similar to their respective annual situation,but the trends were more significant.The extreme values for snow depth were 0.36 cm/yr and −0.17 cm/yr,and were 0.60 mm/yr and −1.10 mm/yr for SWE,respectively.In spring,the negative trends dominated in China,distributing mainly in most of NE and NW,southern QTP,and parts of North China.The extreme values of trends were 0.07 cm/yr and−0.13 cm/yr for snow depth,and were 0.20 mm/yr and−0.30 mm/yr for SWE,respectively.In summer,the region covered by snow was located mainly over the QTP,and the significantly negative trends were dominant.It is thus clear that snow cover was decreasing significantly in warmer seasons in the past 53 years,and it presented significantly negative trends in most regions except for most of NE and northern Xinjiang.

4.Conclusions

This study analyzed not only the temporal-spatial characteristics of snow depth,but also changes of snow density and SWE,which were seldom analyzed before,by using ground-based observations from meteorological stations in China during 1957–2009.

Results indicate that the inter-annual changes of snow density were small,but its inter-monthly variations were apparent.The annual mean snow density was 0.14 g/cm3.For NE,NW,and the QTP,monthly mean snow density reached their maximum in March,April,and June,respectively.From November to the next April when snow cover was abundant,snow density was slightly higher in NW than that in NE,and was the smallest in the QTP.Spatially,annual mean snow density was below 0.14 g/cm3in most of China,and was relative higher in NE,NW,and the QTP.For snow depth and SWE,their inter-monthly variations were generally consistent.Their annual mean values were 0.49 cm and 0.7 mm,and the regions with greater values were located in northern NE and Xinjiang and parts of southwestern QTP.In October and November when snow began accumulating,snow depth and SWE were the greatest in NE,but then,they were the greatest in NW until snow cover disappeared completely.

The annual mean snow depth and SWE increased in all examined regions and in China as a whole with numerous fluctuations during 1957−2009.Although the trend was not significant,the changing amplitude increased gradually in the second half of the 20th century and decreased in the first decade of the 21st century.The changing amplitude of snow cover in China as a whole was basically determined by those of the three examined regions,but it was closely related to changes in NE.Snow cover decrease was affected by climate warming,and was more significant in warmer seasons.On the one hand,the sensitivity of snow cover to air temperature increased,which accelerated the melting of snow (Maet al.,2010);on the other hand,in the early period of snow accumulation,higher air temperature was not favorable for snow to accumulate.This included the decrease of the proportion of solid precipitation and the accumulation of solid precipitation (Maet al.,2011).In any case,persistent climate warming will increasingly threaten snow cover in warmer regions and seasons.

Acknowledgments:

This study was in part supported by the National Natural Science Foundation of China (40901045) and the China Meteorological Administration’s special funds for scientific research on public causes (GYHY200906017).

REFERENCES

Che T,Li X,Rui J,Richard A,Zhang TJ,2008.Snow depth derived from passive microwave remote sensing data in China.Annals of Glaciology,49(1): 145–154.

CMA,2003.Specification of Surface Air Observation.China Meteorological Press,Beijing,pp.1–62.

Li PJ,Mi DS,1983.Distribution of snow cover in China.Journal of Glaciology and Geocryology,5(4): 9–18.

Ma LJ,Qin DH,Bian LG,Xiao CD,Luo Y,2010.Sensitivity of the number of snow cover days to surface air temperature over the Qinghai-Tibetan Plateau.Advances in Climate Change Research,1(2): 76–83.DOI:10.3724/SP.J.0000.2010.00076.

Ma LJ,Qin DH,Bian LG,Xiao CD,Luo Y,2011.Assessment of snow cover vulnerability over the Qinghai-Tibetan Plateau.Advances in Climate Change Research,2(2): 93–100.DOI: 10.3724/SP.J.1248.2011.00093.

Qin DH,Liu SY,Li PJ,2006.Snow cover distribution,variability and response to climate change in western China.Journal of Climate,19(9):1820–1833.DOI: 10.1175/JCLI3694.1.

Savoie MH,Armstrong RL,Brodzik MJ,Wang JR,2009.Atmospheric corrections for improved satellite passive microwave snow cover retrievals over the Tibet Plateau.Remote Sensing of Environment,113(12):2661–2669.DOI: 10.1016/j.rse.2009.08.006.

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.

March 11,2012 Accepted: June 29,2012

*Correspondence to: Dr.LiJuan Ma,Associate Research Fellow of the National Climate Center,China Meteorological Administration.No.46,Zhongguancun Nandajie,Haidian District,Beijing 100081,China.Tel: +86-10-68406544;Email:malj@cma.gov.cn