Characteristics of abrupt changes of snow cover and seasonal freeze-thaw layer in the Tibetan Plateau and their impacts on summer precipitation in China

Rong Gao , HaiLing Zhong , WenJie Dong , ZhiGang Wei

1. Laboratory of Climate Studies, National Climate Center, China Meteorological Administration, Beijing 100081, China

2. State Key Laboratory of Earth Surface Processes and Ecology Resource, Beijing Normal University, Beijing 100875, China

3. Cold and Arid Regions Environmental & Engineering Research Institute, Chinese Academy of Sciences, Lanzhou, Gansu 730000, China

*Correspondence to: Dr. RongGao, the senior engineer of National Climate Center, China Meteorological Administration, Beijing 100081, China. Tel: 86-10-68400090. Email: gaor@cma.gov.cn

Characteristics of abrupt changes of snow cover and seasonal freeze-thaw layer in the Tibetan Plateau and their impacts on summer precipitation in China

Rong Gao1*, HaiLing Zhong1, WenJie Dong2, ZhiGang Wei3

1. Laboratory of Climate Studies, National Climate Center, China Meteorological Administration, Beijing 100081, China

2. State Key Laboratory of Earth Surface Processes and Ecology Resource, Beijing Normal University, Beijing 100875, China

3. Cold and Arid Regions Environmental & Engineering Research Institute, Chinese Academy of Sciences, Lanzhou, Gansu 730000, China

*Correspondence to: Dr. RongGao, the senior engineer of National Climate Center, China Meteorological Administration, Beijing 100081, China. Tel: 86-10-68400090. Email: gaor@cma.gov.cn

In this paper, a variation series of snow cover and seasonal freeze-thaw layer from 1965 to 2004 on the Tibetan Plateau has been established by using the observation data from meteorological stations. The sliding T-test, M-K test and B-G algorithm are used to verify abrupt changes of snow cover and seasonal freeze-thaw layer in the Tibetan plateau. The results show that the snow cover has not undergone an abrupt change, but the seasonal freeze-thaw layer obviously witnessed a rapid degradation in 1987, with the frozen soil depth being reduced by about 15 cm. It is also found that when there is less snow in the plateau region, precipitation in South China and Southwest China increases. But when the frozen soil is deep, precipitation in most of China apparently decreases.Both snow cover and seasonal freeze-thaw layer on the plateau can be used to predict the summer precipitation in China. However,if the impacts of snow cover and seasonal freeze-thaw layer are used at the same time, the predictability of summer precipitation can be significantly improved. The significant correlation zone of snow is located in middle reaches of the Yangtze River covering the Hexi Corridor and northeastern Inner Mongolia, and the seasonal freeze-thaw layer exists in Mt. Nanling, northern Shannxi and northwestern part of North China. The significant correlation zone of simultaneous impacts of snow cover and seasonal freeze-thaw layer is larger than that of either snow cover or seasonal freeze-thaw layer. There are three significant correlation zones extending from north to south: the north zone spreads from Mt. Daxinganling to the Hexi Corridor, crossing northern Mt.Taihang and northern Shannxi; the central zone covers middle and lower reaches of the Yangtze River; and the south zone extends from Mt. Wuyi to Yunnan and Guizhou Plateau through Mt. Nanling.

Tibetan Plateau; snow cover; seasonal freeze-thaw layer; precipitation

1. Introduction

Abrupt climate change is an important phenomenon that generally exists in the climate system, and it can be defined as a "jump" from a stable climate status (or a stable continuing variation trend) to another stable mode (or a different continuing variation trend). Such abrupt climate changes fall into four categories: (1) change in terms of mean value departures; (2) change in variance; (3) sea-saw change; and (4)change at a catastrophe point (Fu and Wang, 1992). In reality,an abrupt climate change often takes place in combination of two or more aforementioned changes.

The Tibetan Plateau is the world’s largest and highest landmass, located in the upper stream of China’s climate system, has an important impact on the climate of China,and is regarded as an initial zone (Fenget al., 1998) –– a climate driver or multiplier. The available findings show that an abrupt change in annual mean temperature (Ma and Li,2003) once occurred in the plateau region in the mid 1980s,without leading to any significant abrupt changes in precipitation (Ma and Hu, 2005). The Tibetan Plateau, known for its large snow cover and frozen soil in China, is an important factor that influences land-air energy and water exchanges.Previous studies focused on the distribution characteristics,inter-annual and decadal variations of snow cover and frozen soil in the Tibetan Plateau (Wanget al., 2001; Weiet al.,2002; Gaoet al., 2003, 2004). However, snow and frozen soil changes are mainly subject to the impacts of precipitation and temperature (Gaoet al., 2003), whether or not the snow cover and frozen soil have undergone an abrupt change because of the rapid temperature rise.

Many previous studies addressed the impacts of plateau snow cover and frozen soil on precipitation in China during the flood-prone season. Chenet al. (2000) investigated the mechanism of snow cover impacts on summer monsoon precipitation by using a statistical analysis method; Wei(2003) gave a comprehensive analysis on impacts of plateau snow through analysis and numerical simulations; Wanget al. (2003) analyzed the impacts of the plateau maximum frozen soil depth on summer precipitation; and Gaoet al.(2005) analyzed the impacts of earlier and later freezing in the plateau region on summer precipitation. However, these studies all addressed the impacts of individual factors. As snow and seasonal frozen soil on the plateau are the major features of the land surface, their variations and abrupt changes all have an important impact on the heat source in the plateau region, thus influencing summer precipitation in China as a whole. By using the variation series of both snow cover and seasonal freeze-thaw layer, we analyze the characteristics of their abrupt changes on the plateau, at the same time we also make a preliminary discussion on precipitation predictions by incorporating the two influential factors—plateau snow cover and seasonal freeze-thaw layer,which prove to be superior relative to any other single element-based approach.

2. Data and methods

In this paper, Tibetan Plateau (Figure 1) winter and spring (October–May) snow series (Figure 2a) is established by using the daily snow-depth data from 72 meteorological stations from 1965 to 2004 in the Tibetan Plateau and the means of Weiet al. (2002). The snow cover in the Tibetan Plateau, prior to the late 1970s, was reduced compared to the snow cover from the late 1970s to the late 1990s, then decreased again from 1999 onwards. At the same time, the method of Gaoet al. (2008) is used to establish the plateau seasonal freeze-thaw layer series (Figure 2b) with winter daily frozen soil depth data from 20 selected stations (Figure 1). It is clear that the seasonal frozen soil on the plateau underwent an evident thinning in frozen depth around 1987.To study the snow cover and frozen soil in relation to summer precipitation in China, summer (June–August) precipitation data of 550 stations in China for the same period are also used. Snow depth data, seasonal frozen soil and precipitation are made available from the National Meteorological Information Centre under the China Meteorological Administration.

The analysis methods used in this paper includes sliding T-test, M-K test and B-G segmentation. The sliding T-test(Afili and Azen, 1972) is a method that determines a catastrophic point by checking the significance level of the difference of mean values from the two samples. But this method is artificial in choosing the length of a sub-sequence,which is likely to shift a catastrophic point. The M-K test(Goossens and Berger, 1986) is a non-parametric testing approach, which is less subject to a data sequence, thus appropriate for detecting a catastrophe point. For a non-smooth time series, Bernaolaet al. (2001) from Boston University proposed a segmentation approach (B-G algorithm) to detect an abrupt change in a time series as a segmentation process.Specifically, an original sequence can be divided into a certain number of different sub-sequences according to their mean values, and the segmentation point can be regarded as a catastrophe point.

Figure 1 The station distribution of snow and frozen soil on the Tibetan Plateau(solid dots: frozen soil stations; hollow dots and solid dots: snow cover stations)

Figure 2 Variations of snow cover (a) and seasonal freeze-thaw layer (b) on the Tibetan Plateau

3. Characteristics of abrupt changes in snow cover and seasonal frozen soil

In this section, sliding T-test, M-K test and B-G algorithm are used to detect catastrophe points of both snow cover and seasonal frozen soil in the Tibetan Plateau. Figure 3 gives sliding T-test curves representing variation amplitudes of the snow cover and seasonal freeze-thaw layer with the sub-sequence length being set to 5. It is quite clear that the curve shows a process of abrupt change in 1998 in figure 3, exceeding the significance level (α＝0.05). The T statistical value is positive, showing an evident snow decrease in 1998. Moreover, figure 2 also presents a snow- decreasing trend after 1999. In the sliding T-test curve representing the seasonal freeze-thaw layer on the Tibetan Plateau, there is one point which is above the significance level (α＝0.05).The catastrophe point appeared in 1986. This point gave a positive value for the seasonally frozen soil on the plateau,indicating the frozen soil became obviously thin in 1987.Figure 4 gives M-K curves for both snow cover and seasonal freeze-thaw layer on the Tibetan Plateau. From the snow UF curve, it is clear that the snow on the plateau continues increasing before 1998, but decrease after 1999, especially the trend of snow increase exceeded the significance level (α＝0.05) in 1987–1998. However, since there is no evident overlapping between UF and UH curves, the snow cover on the plateau witnessed no clear abrupt change. The UF curve for the seasonal freeze-thaw layer on the plateau declined after 1986, and exceeded the significance level (α＝0.05) after 1990, showing that the seasonal frozen soil was constantly de-freezing. Based on the overlapped proportions of UF and UB curves, it was found that the seasonal frozen soil on the plateau witnessed an abrupt change in 1987, in other words, the seasonal frozen soil became thin starting from 1987. The B-G algorithm allows detection of abrupt changes on different scales and amplitudes, and it is less subjective to noises. If the window width is set to 20 for segmentations, then catastrophe points of the snow cover and frozen soil on the Tibetan Plateau are derived, as shown in figure 1. Two catastrophe points of the plateau snow cover were detected in 1972 and 1999. The 40-year variation of the plateau snow can be broadly divided into three mean value segments with relatively smooth changes, out of which the central mean-value segment dominates the whole period. However, the two detected catastrophe points could only pass 87% in the confidence test, suggesting that these two catastrophe points were insignificant. Only one catastrophe point of the seasonal frozen soil on the plateau is detected using the B-G algorithm (Table 1). This point is found in 1987, which passes the significance test (α＝0.01). Before and after 1987, the seasonal frozen soil in the plateau region underwent relatively smooth changes. In contrast, the seasonal frozen soil witnessed a clear degradation after 1987 relative to the period before 1987.

Figure 3 Sliding T-test curves of snow cover (a) and seasonal freeze-thaw layer (b) on the Tibetan Plateau(dashed line: threshold value α=0.05 significance level)

Figure 4 M-K curves of snow cover (a) and seasonal freeze-thaw layer (b) on the Tibetan Plateau(dashed line: threshold value α=0.05 significance level)

In general, snow cover on the Tibetan Plateau may have witnessed an abrupt change in 1999, but the detected catastrophe point was apparently insignificant. The seasonally frozen soil on the plateau experienced a significant abrupt change in 1987, which well corresponded to the findings in the studies of plateau temperature and precipitation changes by Ma and Li (2003) and Ma and Hu (2005).According to the literature (Gaoet al., 2003), snow cover is mainly subject to precipitation changes. No significant precipitation change was detected on the plateau, thus the snow cover on the plateau did not witness an abrupt change. The frozen soil is mainly affected by temperature.Around 1987, the average annual temperature on the plateau increased significantly; therefore the average depth of the seasonal freeze-thaw layer on the plateau became remarkably thin after 1987.

Table 1 B-G algorithm-based catastrophe points of snow cover & seasonal freeze-thaw layer on the Tibetan Plateau

4. Correlation of the snow cover and seasonal freeze-thaw layer variations on the Tibetan Plateau with summer precipitation in China

Although there is no abrupt change in snow cover on the Tibetan plateau, its decadal variations still can be characterized with three phases,i.e. there was less snow both in 1965–1971 (SN1) and 1999–2004 (SN3), but there was more snow in 1972–1998 (SN2). However, the seasonal freeze-thaw layer can be divided into two periods:1965–1986 (FR1) and 1987–2004 (FR2), when abrupt changes took place. As shown in Figure 5, the percentage differences of summer precipitation as a result of the abrupt changes in snow cover and seasonal freeze-thaw layer on the plateau can be derived from the precipitation differences between the more snow period (SN2) and less snow periods(SN1 and SN3), and between deep (FR1) and thin (FR2)frozen-soil periods. Figure 5 also shows that when snow is less on the plateau, precipitation in South China and Southwest China will witness more than that of more snow in the flood-prone season; and increased rainfall in most areas north of the Yangtze River when the plateau is covered with

more snow (Figure 5a, b). On the other hand, when the frozen soil layer is deep in the plateau region, precipitation in most of China will be less than normal, with exceptions in central Northwest China and North China (Figure 5c). These changes of summer precipitation due to abrupt changes of snow cover and seasonal frozen soil in the Tibetan Plateau are mainly attributed to the following processes: snow cover change affects surface albedo, and the change of frozen depth may modify both thermal capacity and heat conductivity, thus leading to the changes of the heat source on the plateau, and eventually changing the general circulation in East Asia. These processes were proved in many previous studies.

5. Distributive characteristics of impacts of Plateau snow cover and seasonal freeze-thaw layer on summer precipitation in China

Although the seasonal freeze-thaw layer on the plateau underwent an evident abrupt change on average, it was not evident that the snow cover witnessed a similar change. The patterns reflect impacts of different plateau snow extents are quite clear, but it shows the impacts of decadal variations of the freeze-thaw layer are rather chaotic. In this paper, the decadal variation of plateau seasonal freeze-thaw layer has been filtered (i.e. seasonal freeze-thaw layer either before or after 1987 subtract the mean value separately) and snow cover variation is used to calculate correlation coefficient of summer precipitation data from 550 stations in China from 1965 to 2004. The findings suggest that the negative correlations between snow cover and summer precipitation are mainly distributed in South China and the southern part of Southwestern China, however positive correlations are found in the rest of China. The significant positive correlations that pass the confidence level (α=0.05) are found in the Yangtze River, Hexi Corridor in Gansu and northeastern Inner Mongolia (Figure 6a). The correlation patterns of the freeze-thaw layer excluding the decadal variation is opposite to that of snow cover,i.e. a positive correlation in South China and the southern part of Southwest China versus a negative correlation in almost the rest of China. The significant positive correlation passing the confidence level (α=0.05) is located in the Mt. Nanling region, while a significant negative correlation is noted in the northern part of North Shaanxi and northwestern part of North China (Figure 6b). From the correlation of both snow cover and freeze-thaw layer on the plateau with summer precipitation in the rest of China, they are basically the same in nature,which reflects the impacts of the increased snow cover and the thin frozen soil on the plateau heat source, leading to precipitation increase in the middle reaches of the Yangtze River.

Figure 5 Percentage differences of summer precipitation in China related to the abrupt change of snow cover and seasonal freeze-thaw layer on the Tibetan Plateau (unit: %)

If the plateau snow cover and freeze-thaw layer have certain predictability for summer precipitation in China, then they can be used to better predict summer precipitation in China by giving a comprehensive consideration of their variations. Therefore,complex correlation coefficients of the plateau snow cover and freeze-thaw layer relative to summer precipitation in China are calculated (Figure 6c). It is found that the significant correlation zone that passes confidence level (α=0.05) is larger than any individual correlation area, showing a pattern of three zonal belts extending from north to south. The north zone spreads from Mt.Daxinganling to the Hexi Corridor, crossing Mt. Taihang and Shanxi. The central zone is located in the lower and middle reaches of the Yangtze River, and the south zone spreads from Mt. Wuyi to the central Yunnan-Guizhou Plateau through Mt.Nanling. The correlation patterns of both snow cover and freeze-thaw layer on the plateau with summer precipitation is most likely related to the plateau’s heat source change that impacts China’s summer monsoon and eventually leading to the northward jump of the summer precipitation belt.

6. Conclusions and discussions

Due to its unique geographical features, the Tibetan Plateau has a large snow cover and extensive frozen soil. Snow cover is affected by precipitation to a greater extent, while variation of frozen soil is closely related to temperature change. Therefore, no evident abrupt change in snow cover has been detected in the plateau region so far. But around 1987, the seasonal freeze-thaw layer witnessed an abrupt change, with the depth of frozen soil being remarkably decreased. It is found that when snow is less on the plateau, precipitation will be more in South China and Southwest China.Conversely, the thicker the frozen soil depth is, the less the precipitation will be in most parts of China.

Figure 6 Distribution of partial and complex correlation coefficients (c) fitting the summer precipitation in China by the snow cover (a)and seasonal freeze-thaw layer (b) in Tibetan Plateau.

In the plateau region, the unique dynamic and thermodynamic effects are one of the key factors that shape the East Asian monsoon climate. Surface heat sources are affected by snow cover through changing albedo of land surface,snowmelt water through changing soil moisture, variation of frozen soil through changing thermal nature of soil (i.e.thermal capacity and heat conductivity), and other factors that may reduce water infiltration. Any change of snow cover or seasonal freeze-thaw layer on the plateau may lead to changes of heat sources, eventually influencing summer precipitation pattern in China. But correlation of snow cover with seasonal freeze-thaw layer is rather complex. Generally,with less snow on the plateau, the temperature may be lowered due to higher atmospheric radiation reflection of the snow cover, which is favorable for soil freezing. In contrast,with deep snow cover, it may suppress the evolution of frozen soil due to heat preservation effect of the snow cover.Therefore, taking into account all impacts of snow cover and frozen soil on summer precipitation in China, it may help understand the impacts of heat source over the plateau on summer precipitation to improve forecast and predictions. In this regard, there are three significant correlation zones,i.e.the north zone extending from Mt. Daxinganling to the Hexi Corridor through northern Mt. Taihang and northern Shannxi; the central zone is located in the middle and lower reaches of the Yangtze River; and the south zone spreads from Mt. Wuyi to the central part of the Yunnan-Guizhou Plateau crossing Mt. Nanling.

The authors are very thankful to reviewers for proposing good suggestions for this manuscript. This research is supported by the National Key Basic Research Program(2007CB411505) and S&T Support Project(2007BAC29B06), National Natural Science Foundation(40705031).

Afili AA, Azen AP, 1972. Statistical Analysis a Computer Oriented Approach.Academic Press, Harcourt Brace Jovanonich Publishers, New York, 366.

Bernaola GP, Ivanov Ch, Nunes Amaeal LA, Stanley HE, 2001. Scale invariance in the nonstationarity of human heart rate. Physical Review Letters, 87(16): 168105-1–4.

Chen QJ, Gao B, Zhang Q, 2000. Studies on relation of snow cover over the Tibetan Plateau in winter to the winter-summer monsoon change. Chinese Journal of Atmospheric Sciences, 24(2): 477–491.

Feng S, Tang MC, Wang DM, 1998. New evidence for the Qinghai-Tibetan Plateau as a pilot region of climatic fluctuation in China. Chinese Science Bulletin, 43(6): 633–636.

Fu CB, Wang Q, 1992. The definition and detection of the abrupt climatic change. Sciencia Atmospherica Sinica, 16(4): 482–493.

Gao R, Wei ZG, Dong WJ, Dong WJ, Wang CH, Zhong HL, 2003. Variation of the snow and frozen soil and their response to climatic change on the Qinghai-Tibetan Plateau on twenty century evening. Plateau Meteorology, 22(2): 191–196.

Gao R, Wei ZG, Dong WJ, 2004. Analysis of the cause of the differentia in interannual variation between snow cover and seasonal frozen soil in the Tibetan Plateau. Journal of Glaciology and Geocryology, 26(2):153–159.

Gao R, Wei ZG, Dong WJ, Zhong HL, 2005. Impact of the anomalous thawing in the Tibetan Plateau on summer precipitation in China and its mechanism. Advances in Atmospheric Sciences, 22(2): 238–245.

Gao R, Wei ZG, Dong WJ, 2008. The character of temporal and spatial distribution of seasonal frozen soil in Tibetan Plateau. Journal of Glaciology and Geocryology, 30(5): 740–744.

Goossens CH, Berger A, 1986. Annual and seasonal climatic variations over the Northern Hemisphere and Europe during the last century. Ann. Geophys., 4: 385–400.

Ma XB, Hu ZY, 2005. Precipitation variation characteristics and abrupt change over Qinhai-Tibetan Plateau in recent 40 Years. Journal of Desert Research, 25(1): 137–139.

Ma XB, Li DL, 2003. Analyses on air temperature and its abrupt change over Qinghai-Tibetan Plateau in modern age. Plateau Meteorology, 22(5):507–512.

Wang CH, Dong WJ, Wei ZG, 2001. The feature of seasonal frozen soil in Qinghai-Tibet Plateau. Acta Geographica Sinica, 56(5): 523–531.

Wang CH, Dong WJ, Wei ZG, 2003. Study on relationship between the frozen-thaw process in Qinghhai-Tibetan Plateau and circulation in East-Asia. Chinese Journal of Geophysics, 46(3): 309–315.

Wei ZG, 2003. The large scale variation and impact of land surface in western China on summer precipitation, PH. D. Dissertation, Graduate School of the Chinese Academy of Sciences, 61–118.

Wei ZG, Huang RH, Chen W, Dong WJ, 2002. Spatial distributions and interdecadal variations of the snow at the Tibetan Plateau Weather Stations. Scientia Atmospherica Sinica, 26(4): 496–508.

10.3724/SP.J.1226.2011.00024

16 June 2010 Accepted: 18 August 2010