J. Geogr. Sci. 2016, 26(9): 1277-1288 DOI: 10.1007/s11442-016-1326-8 © 2016
Science Press
Springer-Verlag
Decrease in snowfall/rainfall ratio in the Tibetan Plateau from 1961 to 2013 WANG Jie, *ZHANG Mingjun, WANG Shengjie, REN Zhengguo, CHE Yanjun, QIANG Fang, QU Deye College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, China
Abstract: On the basis of two gridded datasets of daily precipitation and temperature with a spatial resolution of 0.5°×0.5°, and meteorological station data released by the National Meteorological Information Center (NMIC) during 1961–2013, the spatial and temporal variations of total amount of precipitation, amount of rainfall, amount of snowfall and snowfall/rainfall ratio (S/R) in the Tibetan Plateau (TP) are analyzed using Sen’s slope, the Mann–Kendall mutation test, Inverse Distance Weighting (IDW) and the Morlet wavelet. Total amount of precipitation and amount of rainfall generally show statistically significant increasing trends of 0.6 mm·a–1 and 1.3 mm·a–1, respectively, while amount of snowfall and S/R have significant decreasing trends of –0.6 mm·a–1 and –0.5% a–1, respectively. In most regions, due to significant increasing trends in total amount of precipitation and amount of rainfall, and significant decreasing trends in amount of snowfall, S/R shows a decreasing trend in the TP. Abrupt changes in total amount of precipitation, amount of rainfall, amount of snowfall and S/R are detected for 2005, 2004, 1996 and 1998, respectively. Total amount of precipitation, amount of rainfall, amount of snowfall and S/R are concentrated in cycles of approximately 5 years, 10 years, 16 years and 20 years, respectively. The trend magnitudes for total amount of precipitation and amount of rainfall all show decreasing-to-increasing trends with elevation, while amount of snowfall and S/R show decreasing trends. Keywords: Tibetan Plateau; gridded data; snowfall/rainfall ratio; precipitation
1
Introduction
The different phases (solid or liquid) of precipitation greatly influence surface runoff, energy flow and material cycles on the Earth (Mizukami et al., 2013; Dingman, 2015), which can usually be assessed using the snowfall/rainfall ratio (S/R). Snowfall is an important water source for irrigation and drinking water in high and cold regions. Changes in snowfall and snowpack volume can greatly influence regional water supply (Scipión et al., 2013). HowReceived: 2016-01-25 Accepted: 2016-03-10 Foundation: National Basic Research Program of China (973 Program), No.2013CBA01801; National Natural Science Foundation of China, No.41461003, No.41161012 Author: Wang Jie (1989–), MS Candidate, specialized in global change and sustainable development. E-mail:
[email protected] * Corresponding author: Zhang Mingjun, Professor, E-mail:
[email protected]
www.geogsci.com
www.springerlink.com/content/1009-637x
1278
Journal of Geographical Sciences
ever, changes in S/R may also affect snowpack in the spring, which is often critical for spring flood development (Berghuijs et al., 2014). In addition, S/R will affect changes in surface albedo (snow and rain have different impacts on the absorption and reflection of solar radiation), which will alter the energy balance of the Earth–atmosphere system (Screen and Simmonds, 2012). In terms of glacier mass balance, net radiation plays a key role in glacier melt, and variation in glacier albedo will affect the amount of glacier melt that takes place (Mernild et al., 2015). An increase in snowfall will contribute to an increase in a glacier’s surface albedo, which will slow down glacier melt; however, a decrease in snowfall will accelerate glacier melt. So S/R can be used as an indicator for climate change, and plays an extremely important role in measuring surface runoff, energy flow and material cycles. Changes in S/R have been reported across the globe during recent decades. In New England, S/R has shown a significant decreasing trend, which is related to the North Atlantic Oscillation and the Pacific–North America index (Huntington et al., 2004). A similar decreasing trend is found in Switzerland, especially at low altitudes (Serquet et al., 2011). Feng and Hu (2004) indicated that S/R has greatly decreased in the Pacific Northwest and the central United States, but that the magnitude of the decrease has been slight in the eastern United States; the winter temperature in winter is higher than freezing point, which affects precipitation patterns in these regions. Guo and Li (2015) concluded that S/R has experienced a downward trend in the Chinese Tianshan Mountains, and that temperature is the major factor. Due to its high elevation and low surface temperature, the Tibetan Plateau (TP) is extensively covered by glaciers and permafrost. In recent decades, the rapid shrinkage of the cryosphere in the TP has been reported (Yao and Zhu, 2006). Variation in the amount of snowfall may greatly alter the surface albedo and hydrological processes, which influence regional and global climate (Wu et al., 2012; Wang et al., 2013), so it is necessary to assess recent changes using S/R in this region. However, in situ observations on the plateau are fairly limited. High-resolution gridded meteorological datasets are applied to study the characteristics of the variation in S/R from 1961 to 2013.
2 2.1
Data and methods Study areas
The TP covers an area of about 2572.4×103 km2 with a mean elevation is 4500 m, lying between 26°00′12″N to 39°46′50″N in latitude and 73°18′52″E to 104°46′59″E in longitude (Zhang et al., 2002). Precipitation mainly occurs between May and October, accounting for about 90% of the annual total (Wang, 2007). In addition, spatial dependency of air temperature and precipitation is also widely reported for this region (Li and Kang, 2006). 2.2
Data sources
Input data for this study include data from meteorological stations and gridded data. They are acquired from the National Meteorological Information Center (NMIC). To acquire the critical temperature for identifying precipitation phases in the study region, the meteorological records for precipitation are from stations in the TP and its surrounding areas. In this
WANG Jie et al.: Decrease in snowfall/rainfall ratio in the Tibetan Plateau from 1961 to 2013
1279
study, daily precipitation and daily mean temperature at 146 meteorological stations from 1961 to 1979 are used; these stations are located in seven provinces or autonomous regions including Gansu (26), Inner Mongolia (1), Qinghai (35), Sichuan (36), Tibet (20), Xinjiang (16) and Yunnan (12) (Figure 1). Strict quality controls are conducted by the NMIC.
Figure 1
Spatial distribution of meteorological stations and grid boxes in the TP and its surrounding areas
Gridded datasets of precipitation and daily mean temperature with a spatial resolution of 0.5°×0.5° from 1961 to 2013 were also obtained from the NMIC. The datasets are based on an interpolation method using daily meteorological parameters from more than 2000 stations in China, and describe the spatial pattern of temperature and precipitation in China very accurately. The accuracy of the precipitation dataset is assessed in a previous study (Zhao et al., 2014). 2.3
Methods
Precipitation can generally be divided into rainfall and snowfall. Without manual or instrumental records of precipitation type, identification of snowfall and rainfall is usually based on measured meteorological parameters, including temperature profiles and other atmospheric conditions (Rauber et al., 2001), and single-threshold and dual-threshold surface air conditions (Kang and Ohmura, 1994; Yang et al., 1997; Gustafsson et al., 2001). On the basis of existing statistical methods (Han et al., 2010; Chen et al., 2014a, 2014b), a statistical method was applied in this study to diagnose precipitation types. In the NMIC’s records for 146 meteorological stations, weather codes are available from 1961 to 1979: snow (including sleet and snowstorms) is marked as 31XXX (XXX is the actual amount), and a mixture of rain and snow is marked as 30XXX. Daily mean temperatures for the two types are selected, and then temperature ranges for the two types are acquired for each station. The two ranges have a temperature intersection, which is then divided into a number of subintervals of 0.5℃ steps. Probabilities of rain and snow for the subintervals are calculated, and then the lower limit of the turning point of the probability values is identified
1280
Journal of Geographical Sciences
as the critical temperature between rain and snow. This critical temperature is applied to other regions where there are no meteorological stations using Inverse Distance Weighting (IDW) on the ArcGIS platform (Figure 2). To investigate inter-annual variation in S/R, the period from August 1 to July 31 the following year is defined as a hydrological year (i.e., the year of 1961 means from August 1961 to July 1962), and the change in S/R in the TP during 1961–2013 is studied.
Figure 2
Spatial distribution of snow and rain critical temperatures in the TP
In this paper, snow and rain critical temperatures are interpolated from points to the surface by IDW (Figure 2). The nonparametric Sen’s method (Sen, 1968) is employed to calculate the linear trends in total amount of precipitation, amount of rainfall, amount of snowfall and S/R, and the statistical significance is examined using the Mann–Kendall test. The spatial distribution of trend magnitudes for these elements is drawn using ArcGIS. In addition, the Morlet wavelet is used to analyze the cycles.
3 3.1
Results Temporal variation
Total amount of precipitation ranges between 374.1 mm and 497.4 mm (with an average of 438.2 mm) in the TP from 1961 to 2013 (Figure 3a); it shows an increasing trend at 0.6 mm·a–1 (p<0.05) during the study period, which corresponds to the previous research (Hu and Liang, 2013; Tong et al., 2014; You et al., 2015; Lin et al., 2015). According to the moving curves, the total amount of precipitation shows a slightly rising trend with fluctuations during 1961–1992; after 1992, it sharply increases. Figure 3b shows inter-annual variation in the amount of rainfall between 1961 and 2013, which shows an increasing trend at 1.3 mm·a–1 (p<0.001); after a short-term decrease in 1960–1970, it started to increase during 1970–2013. In Figure 3c, the amount of snowfall generally shows a decreasing trend at a rate of 0.6 mm·a–1 (p<0.01), from an increase in 1961–1974, to a sharp decrease with some fluctuation after 1974. With an increase in the amount of rainfall and a decrease in the amount of snowfall, S/R shows a decreasing trend at a rate of 0.5% a–1 (p<0.001) over the analysis period, although it alters from an increase to a decrease in 1975 (Figure 3d). 3.2
Spatial variation
Figure 4a shows the trend magnitudes for total amount of precipitation during 1961–2013.
WANG Jie et al.: Decrease in snowfall/rainfall ratio in the Tibetan Plateau from 1961 to 2013
1281
Figure 3 Inter-annual variation in total amount of precipitation (a), amount of rainfall (b), amount of snowfall (c) and S/R (d) in the TP during 1961–2013
The trend magnitude changes from –12.1 mm·a–1 to 14.1 mm·a–1; 76.9% and 23.1% of the grid boxes show positive and negative trends, respectively, but only 45.1% and 20.9% of the respective grid boxes are statistically significant at the 0.05 level (Table 1). Also, most of the grid boxes experience positive trends. The areas with significant values are in the middle, northwest, northeast and southeast of the TP, and there are more positive grid boxes than the negative ones. The areas with decreasing trend magnitudes are situated in the southeast, east and west of the TP (Figure 4b). Table 1 Intervals of annual trend magnitude rate and percentages of grid boxes for different variation trends, for total amount of precipitation, amount of rainfall, amount of snowfall and S/R in the TP during 1961–2013
Total precipitation
Percentage (%)
Trend magnitude range (mm·a–1 or a–1)
Positive
Significantly positive
Negative
Significantly negative
–12.1 to 14.1
76.9
45.1
23.1
20.9
Rainfall
–4.6 to 14.0
87.2
65.1
12.8
14.9
Snowfall
–12.1 to 3.3
25.2
13.6
74.8
41.8
S/R
–6.2 to 13.1
8.6
10.5
91.4
59.1
Trend magnitudes of the annual amount of rainfall vary from –4.6 mm·a–1 to 14.0 mm·a–1. Some 87.2% of the grid boxes show increasing trends, and 65.1% of them are statistically significant at the 0.05 level (Table 1). Significant positive trends are located in the middle and southeast of the TP. Decreasing trends are concentrated in the southeast, east and west of the TP. Grid boxes of significant increases and decreases are in most areas of the TP except for the east, and there are significantly more grid boxes with increasing trends (Figures 4c and 4d).
1282
Journal of Geographical Sciences
Figure 4 Spatial distribution of trend magnitudes for total amount of precipitation (a and b), amount of rainfall (c and d), amount of snowfall (e and f) and S/R (g and h) in the TP during 1961–2013
Trend magnitudes of the amount of snowfall vary from –12.1 mm·a–1 to 3.3 mm·a–1 (Table 1). Some 25.2% of the grid boxes show increasing trends, and 13.6% of them are statistically significant at the 0.05 level. Increasing trends are located in the middle and northwest of the TP (Figure 4e). Grid boxes of significant increases and decreases are scattered, and there are more positive grid boxes than the negative ones (Figure 4f). For S/R, trend magnitudes vary from –6.2 a–1 to 13.1 a–1 (Table 1). Some 8.6% of the grid boxes have positive trends, which are distributed in the northeast and southeast of the TP; 91.4% show decreasing trends in most regions, and significantly decreasing trends are scattered in the middle of the TP (Figures 4g and 4h). The most statistically significant grid boxes show decreasing trends in the TP.
WANG Jie et al.: Decrease in snowfall/rainfall ratio in the Tibetan Plateau from 1961 to 2013
3.3 3.3.1
1283
Characteristics of abrupt change and change cycles Abrupt change
The abrupt change diagram reveals the trend in total amount of precipitation in the TP (Figure 5a). The UF curve of total amount of precipitation shows a decreasing trend with some fluctuation during 1961–1994 and then an increase to 2013. An abrupt change in total amount of precipitation is detected in about 2005 (significant at the 0.05 level). The total amount of precipitation increases by 35.9 mm after an abrupt change (Table 2). The UF curve of the amount of rainfall shows a decrease before 1970, and then an increase with fluctuations during 1970–2013 (Figure 5b). The amount of rainfall has an abrupt change around 2004 and then an increase of 52.4 mm after 2004 (Table 2). The UF curve of the amount of snowfall shows an increase during 1961–1968, that it remains stable after 1968, and that it starts to decline after 1980 (Figure 5c). An abrupt change in the amount of snowfall appears in 1996 (significant at the 0.05 level), and then it decreases by 20.3 mm after 1996 (Table 2). The abrupt change in S/R is similar to that of the amount of snowfall. An abrupt change appears in 1998, and then S/R decreases by 20% after 1998 (Figure 5d and Table 2).
Figure 5 Mann–Kendall statistics for total amount of precipitation (a), amount of rainfall (b), amount of snowfall (c) and S/R (d) in the TP during 1961–2013
3.3.2
Change cycles
Cycles for total amount of precipitation are approximately 6 years, 16 years and 20 years in the TP over the past 53 years (Figure 6a). The 16-year cycle is predominant, and has been observed since 1980. The 6-year cycle occurs more frequently during 1961–2013, and
1284
Journal of Geographical Sciences
Table 2 Years of abrupt change for total amount of precipitation, amount of rainfall, amount of snowfall and S/R in the TP during 1961–2013 Abrupt year
Average before abrupt change
Average after abrupt change
Difference
Total precipitation (mm)
2005
433.4
469.3
35.9
Rainfall (mm)
2004
266.4
318.8
52.4
Snowfall (mm)
1996
172.1
151.8
–20.3
S/R (%)
1998
70
50
–20
Figure 6 Variance in change cycles for total amount of precipitation (a), amount of rainfall (b), amount of snowfall (c) and S/R (d) in the TP during 1961–2013
appears to be weaker in the early 1990s and stronger after 1990. The 20-year cycle appears before 1980 (Figure 7a). Cycles for the amount of rainfall are about 5 years, 8 years and 20 years. The predominant cycle is 20 years (Figure 6b), which occurs about the same time as the 8-year cycle in the late 1980s. The 5-year cycle appears in the early 1980s (Figure 7b). Cycles for both the amount of snowfall and S/R are about 5 years, 10 years and 16 years. But the major cycles for the amount of snowfall and S/R are 5 years and 10 years, respectively (Figures 6c and 6d). The 5-year cycle occurs more frequently; the 10-year cycle occurs before 1980; and the 16-year cycle is superimposed on the 10-year cycle before 1990 (Figures 7c and 7d). 3.4
Elevation analysis of S/R
Figure 8 shows trend magnitudes for the total amount of precipitation, amount of rainfall, amount of snowfall and S/R with elevations from 1961 to 2013. The 995 grids are divided into 9 elevation ranges according to 500 m intervals. It is clear that fluctuations in total
WANG Jie et al.: Decrease in snowfall/rainfall ratio in the Tibetan Plateau from 1961 to 2013
1285
Figure 7 Change cycles for total amount of precipitation (a), amount of rainfall (b), amount of snowfall (c) and S/R (d) in the TP during 1961–2013
amount of precipitation and amount of rainfall are more pronounced at lower elevations than at higher elevations; they show decreasing trends below 2500 m, and increase up to 5000–5500 m (Figures 8a and 8b). Figure 8c shows that changes in the amount of snowfall are relatively stable below 3000 m, greatly fluctuate above 3000 m, and show an increasing trend above 5000 m. The change in S/R is relatively stable below 4500 m, and has a decreasing trend above 4500 m (Figure 8d).
4
Discussion
Previous research indicates that rate of warming is amplified with elevation: that is, surface air temperatures increase more rapidly at higher elevation (Mountain Research Initiative EDW Working Group, 2015). The TP is a unique geographical region with an average height of approximately 4000 m above sea level and an area of approximately 2.5×106 km2. It is sensitive and vulnerable to climate change, and thus its tendencies imply an early signal of global change (Feng et al., 1998). Yan and Liu (2014) found that the rate of increase in annual mean temperature above 2000 m in the TP is greater than the global rate, suggesting that the TP is a region that is sensitive to global warming. In this study, the trend magnitude of the amount of snowfall is relatively stable below 3000 m, and has a decreasing trend above 3000 m in the TP. This reflects the fact that changes in the amount of snowfall are related to warming at higher elevations, which is consistent with the research result of the
1286
Journal of Geographical Sciences
Figure 8 Change in trend magnitude for total amount of precipitation (a), amount of rainfall (b), amount of snowfall (c) and S/R (d) in the TP with change in elevation during 1961–2013 (a line within the box marks the median)
Mountain Research Initiative Elevation Dependent Warming Working Group (2015). Due to a change in warming at higher elevations, some snow is being converted into rain, inducing a decreasing trend in the amount of snowfall and an increasing trend in the amount of rainfall at higher elevations of the TP. Furthermore, the TP is the largest modern glacier distributed area at low–mid latitude in the world (Liu et al., 2000). So glacier mass balance has been impacted in the TP. In the hydrological model, the critical temperature of snow and rain is an important parameter, and is used as a fixed threshold in some schemes such as the BATS model (Yang et al., 1997), Snow cover model (Loth et al., 1993), DHSVM model (Wigmosta et al., 1994) and HBV model (Lindström et al., 1997). However, precipitation types are affected by the complicated landforms and other natural factors. In this study, we obtain the critical temperature values for stations in the TP, and then these critical temperatures are applied to other regions where there are no meteorological stations using the IDW approach on the ArcGIS platform, especially for extremely complex terrain at high altitudes. However, this study has some shortcomings in obtaining a critical temperature between rain and snow. It should be noted that the length of the time series is limited, and the impact of global warming on the critical temperature is not fully considered after 1979. In addition, in the west of
WANG Jie et al.: Decrease in snowfall/rainfall ratio in the Tibetan Plateau from 1961 to 2013
1287
the TP, meteorological stations are scarce, which will cause inaccuracy in critical temperature. Therefore, the accuracy of the critical temperature between rain and snow needs to be further improved.
5
Conclusions
On the basis of the daily precipitation and daily mean temperature from 0.5°×0.5° gridded datasets and meteorological stations during 1961–2013 released by the National Meteorological Information Center, the spatial and temporal change, abrupt change and change cycles of annual total amount of precipitation, amount of rainfall, amount of snowfall and S/R in the TP for the past 53 years are analyzed. (1) Total amount of precipitation and amount of rainfall generally show increasing trends of 0.6 mm·a–1 (p<0.05) and 1.3 mm·a–1 (p<0.001), respectively, while the amount of snowfall and S/R show decreasing trends of –0.6 mm·a–1 (p<0.01) and –0.5% a–1 (p<0.001), respectively. (2) For annual total amount of precipitation, 76.9% of the grid boxes show positive trends, and 45.1% of the grid boxes are statistically significant at the 0.05 level. The amount of rainfall shows increasing trends in 87.2% (65.1% of the grid boxes are statistically significant at the 0.05 level) of the grid boxes. However, the amount of snowfall shows decreasing trends in 74.8% of the grid boxes (41.8% of the grid boxes are statistically significant at the 0.05 level). So S/R shows decreasing trends in 91.4% of the grid boxes in most regions and 59.1% of the grid boxes are statistically significant at the 0.05 level. (3) Abrupt changes in the total amount of precipitation, amount of rainfall, amount of snowfall and S/R are found in 2005, 2004, 1996 and 1998, respectively. Total amount of precipitation, amount of rainfall, amount of snowfall and S/R are concentrated in cycles of approximately 5 years, 10 years, 16 years and 20 years, in which the major cycles are 16 years, 20 years, 5 years and 10 years, respectively. (4) Fluctuation in trends of total amount of precipitation and amount of rainfall is more pronounced at lower elevations than at higher elevations, and the total amount of precipitation and amount of rainfall show decreasing trends below 2500 m and increasing trends at 5000–5500 m. Changes in amount of snowfall are relatively stable below 3000 m, greatly fluctuate above 3000 m, and show a decreasing trend above 5000 m. Changes in S/R are also relatively stable below 4500 m, while significant decreasing trends are seen above 4500 m.
References Berghuijs W R, Woods R A, Hrachowitz M, 2014. A precipitation shift from snow towards rain leads to a decrease in streamflow. Nature Climate Change, 4(7): 583–586. Chen R, Liu J, Song Y, 2014b. Precipitation type estimation and validation in China. Journal of Mountain Science, 11(4): 917–925. Chen R S, Kang E S, Ding Y J, 2014a. Some knowledge on and parameter of China’s alpine hydrology. Advances in Water Science, 25(3): 307–316. (in Chinese) Dingman S L, 2015. Physical Hydrology. Long Grove: Waveland Press. Feng S, Hu Q, 2004. Changes in the agro-meteorological indicators in the contiguous United States: 1951–2000. Theoretical and Applied Climatology, 78(4): 247–264. Feng S, Tang M C, Wang D M, 1998. New evidence for the Qinghai-Xizang (Tibet) Plateau as a pilot region of climatic fluctuation in China. Chinese Science Bulletin, 43(6): 633–636. (in Chinese) Guo L, Li L, 2015. Variation of the proportion of precipitation occurring as snow in the Tian Shan Mountains, China. International Journal of Climatology, 35(7): 1379–1393.
1288
Journal of Geographical Sciences
Gustafsson D, Stähli M, Jansson P E, 2001. The surface energy balance of a snow cover: Comparing measurements to two different simulation models. Theoretical and Applied Climatology, 70(1–4): 81–96. Han C T, Chen R S, Liu J F et al., 2010. A discuss of the separating solid and liquid precipitations. Journal of Glaciology and Geocryology, 32(2): 249–256. (in Chinese) Hu H R, Liang L, 2013. Temporal and spatial variations of rainfall at the east of Qinghai-Tibet Plateau in last 50 years. Plateau and Mountain Meteorology Research, 33(4): 1–7. (in Chinese) Huntington T G, Hodgkins G A, Keim B D et al., 2004. Changes in the proportion of precipitation occurring as snow in New England (1949–2000). Journal of Climate, 17(13): 2626–2636. Kang E S, Ohmura A, 1994. Energy, water and mass balance and runoff models in Tianshan glacier affected area. Science in China (Series B), 24(9): 983–991. (in Chinese) Li C L, Kang S C, 2006. Review of the studies on climate change since the last inter-glacial period on the Tibetan Plateau. Acta Geographica Sinica, 16(3): 327–335. (in Chinese) Lin H B, You Q L, Jiao Y et al., 2015. Spatial and temporal characteristics of the precipitation over the Tibetan Plateau from 1961 to 2010 based on high resolution grid-observation dataset. Journal of Natural Resources, 30(2): 271–281. (in Chinese) Lindström G, Johansson B, Persson M et al., 1997. Development and test of the distributed HBV-96 hydrological model. Journal of Hydrology, 201(1): 272–288. Liu Z X, Su Z, Yao T D et al., 2000. Resources and distribution of glaciers on the Tibetan Plateau. Resources Science, 22(5): 49–52. (in Chinese) Loth B, Graf H F, Oberhuber J M, 1993. Snow cover model for global climate simulations. Journal of Geophysical Research: Atmospheres, 98(D6): 10451–10464. Mernild S H, Malmros J K, Yde J C et al., 2015. Albedo decline on Greenland’s Mittivakkat Gletscher in a warming climate. International Journal of Climatology, 35(9): 2294–2307. Mizukami N, Koren V, Smith M et al., 2013. The impact of precipitation type discrimination on hydrologic simulation: Rain-snow partitioning derived from HMT-west radar-detected bright band height versus surface temperature data. Journal of Hydrometeorology, 14(4): 1139–1158. Mountain Research Initiative Elevation Dependent Warming (EDW) Working Group, 2015. Elevation-dependent warming in mountain regions of the world. Nature Climate Change, 5(5): 424–430. Rauber R M, Olthoff L S, Ramamurthy M K et al., 2001. Further investigation of a physically based, nondimensional parameter for discriminating between locations of freezing rain and ice pellets. Weather and Forecasting, 16(1): 185–191. Scipión D E, Mott R, Lehning M et al., 2013. Seasonal small-scale spatial variability in alpine snowfall and snow accumulation. Water Resources Research, 49(3): 1446–1457. Screen J A, Simmonds I, 2012. Declining summer snowfall in the Arctic: Causes, impacts and feedbacks. Climate Dynamics, 38(11/12): 2243–2256. Sen P K, 1968. Estimates of the regression coefficient based on Kendall’s tau. Journal of American Statistical Association, 39: 1379–1389. Serquet G, Marty C, Dulex J P et al., 2011. Seasonal trends and temperature dependence of the snowfall/precipitation-day ratio in Switzerland. Geophysical Research Letters, 38(7). doi: 10.1029/2011GL046976. Tong K, Su F, Yang D et al., 2014. Tibetan Plateau precipitation as depicted by gauge observations, reanalysis and satellite retrievals. International Journal of Climatology, 34(2): 265–285. Wang J A, 2007. China Geography Tutorial. Beijing: Higher Education Press. (in Chinese) Wang T, Peng S, Lin X et al., 2013. Declining snow cover may affect spring phenological trend on the Tibetan Plateau. Proceedings of the National Academy of Sciences, 110(31): E2854–E2855. Wigmosta M S, Vail L W, Lettenmaier D P, 1994. A distributed hydrology-vegetation model for complex terrain. Water Resources Research, 30(6): 1665–1679. Wu Z, Jiang Z, Li J et al., 2012. Possible association of the western Tibetan Plateau snow cover with the decadal to interdecadal variations of northern China heatwave frequency. Climate Dynamics, 39(9/10): 2393–2402. Yan L, Liu X, 2014. Has climatic warming over the Tibetan Plateau paused or continued in recent years? Journal of Earth, Ocean and Atmospheric Sciences, 1(1): 13–28. Yang Z L, Dickinson R E, Robock A et al., 1997. Validation of the snow submodel of the biosphere-atmosphere transfer scheme with Russian snow cover and meteorological observational data. Journal of Climate, 10(2): 353–373. Yao T D, Zhu L P, 2006. The response of environmental changes on Tibetan Plateau to global changes and adaptation strategy. Advances in Earth Science, 21(5): 459–464. (in Chinese) You Q, Min J, Zhang W et al., 2015. Comparison of multiple datasets with gridded precipitation observations over the Tibetan Plateau. Climate Dynamics, 45(3): 791–806. Zhang Y L, Li B Y, Zheng D, 2002. A discussion on the boundary and area of the Tibetan Plateau in China. Geographical Research, 21(1): 1–8. (in Chinese) Zhao Y F, Zhu J, Xu Y, 2014. Establishment and assessment of the grid precipitation datasets in China for recent 50 years. Scientia Meteorologica Sinica, 34(4): 414–420. (in Chinese)