Thermal state of permafrost and active layer in Central Asia during the international polar year


  • Lin Zhao,

    Corresponding author
    1. Cryosphere Research Station on the Qinghai–Tibet Plateau and State Key Laboratory of Cryospheric Sciences, Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences
    • Cryosphere Research Station on the Qinghai-Tibet Plateau and State Key Laboratory of Cryospheric Sciences, Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, 320 Donggang West Rd., Lanzhou, Gansu 730000, China.
    Search for more papers by this author
  • Qingbai Wu,

    1. State Key Laboratory of Frozen Soil Engineering, Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences
    Search for more papers by this author
  • S.S. Marchenko,

    1. Geophysical Institute, University of Alaska Fairbanks, AK 99775-7320, USA
    Search for more papers by this author
  • N. Sharkhuu

    1. Institute of Geography, Mongolian Academy of Sciences, Ulaanbaatar 210620, Mongolia
    Search for more papers by this author


Permafrost in Central Asian is present in the Qinghai–Tibet Plateau in China, the Tien Shan Mountain regions in China, Kazakhstan and Kyrgyzstan, the Pamirs in Tajikistan, and in Mongolia. Monitoring of the ground thermal regime in these regions over the past several decades has shown that the permafrost has been undergoing significant changes caused by climate warming and increasing human activities. During the International Polar Year, measured mean annual ground temperature (MAGT) at a depth of 6 m ranged from −3.2°C to 0.2°C on the Qinghai–Tibet Plateau and the active-layer thickness (ALT) varied between 105 and 322 cm at different sites. Ground temperatures at the bottom of the active layer (TTOP) warmed on average by 0.06°C yr−1 over the past decade. In Mongolia, MAGT at 10–15 m depth increased by up to 0.02–0.03°C yr−1 in the Hovsgol Mountain region, but by 0.01–0.02°C yr−1 in the Hangai and Hentei Mountain regions. The increase in permafrost temperatures in the northern Tien Shan from 1974 to 2009 ranged from 0.3°C to 0.6°C. At present measured permafrost temperatures vary from −0.5°C to −0.1°C. The ALT increased from 3.2 to 4 m in the 1970s to a maximum of 5.2 m between 1995 and 2009. Copyright © 2010 John Wiley & Sons, Ltd.


The Central Asian region encompasses the largest area underlain by alpine (mountain) permafrost in the world, extending across parts of southern Siberia, Mongolia, China, Kazakhstan, and adjacent countries (Figure 1) (Brown et al., 1997). It covers 3.5 × 106 km2, about 15% of the total areal extent of permafrost in the Northern Hemisphere (Romanovsky et al., 2007). Climate warming during the 20th century and increasing human activities (Jin et al., 2006; Zhang et al., 2008) have had major effects on the contemporary thermal state of this permafrost as documented in a number of papers, including a series within the Proceedings of the Ninth International Conference on Permafrost (Wu and Liu, 2004; Cheng and Wu, 2007; Marchenko et al., 2007; Sharkhuu et al., 2007; Kane and Hinkel, 2008; Wu and Zhang, 2008; Wu et al., 2008; Zhao et al., 2008a, b).

Figure 1.

Digital elevation map (ETOPO30/USGS) of Central Asia showing permafrost distribution (Brown et al., 1997) and locations of boreholes.

Changes in the permafrost thermal regime can have significant impacts on local hydrology, land surface energy and moisture balances, carbon exchange between the land and the atmosphere, and ecosystems, as well as on engineering infrastructure (Cheng and Zhao, 2000; Nelson et al., 2001; Zhang et al., 2005). In addition, increases in active-layer thickness in certain areas may cause thaw settlement as ice-rich soils in the upper permafrost layers thaw (e.g. Nelson et al., 2001), resulting in an increase in slope instability (Harris et al., 2001). These changes in permafrost temperatures and in the depth of seasonal freezing/thawing are indicators of changes in climate, and also influence surface and subsurface hydrology. Under the auspices of the International Permafrost Association, a world-wide permafrost monitoring network was set up and thermal state monitoring was undertaken during the International Polar Year (IPY) to obtain a snapshot of global permafrost temperatures and active-layer measurements (Brown and Romanovsky, 2008; Smith and Brown, 2009). These in situ measurements are essential to calibrate and to verify regional and global climate change models (IPCC, 2007; Romanovsky et al., 2007). This paper reports on recent monitoring results from Central Asia.


Qinghai-Tibet Plateau

Owing to its high elevation and extremely cold climate, the Qinghai–Tibet Plateau is the largest and highest permafrost region in the mid- to low-latitude regions of the world. The lower limit of permafrost descends northward and eastward gradually, and roughly corresponds to the −2.0 to −2.5°C isotherm of mean annual air temperature (Zhou and Guo, 2000). It rises by about 150 to 200 m per 1° change in latitude southward. Permafrost temperatures decrease by 0.6 to 1°C and permafrost thickness increases by 15–20 m as the elevation increases by 100 m (Zhou and Guo, 2000). Permafrost on the Qinghai–Tibet Plateau is classified as predominantly continuous (occupying 70–90% of the area), as widespread island permafrost (30–70%), or as sparse island permafrost (<30%) (Zhou and Guo, 2000). A permafrost monitoring network along the Qinghai–Tibet Highway includes more than 20 boreholes with depths from 20 to 127 m (Table 1), 10 sets of active-layer measurement systems (Table 2), four automatic weather stations, and two eddy covariance systems (Cheng and Wu, 2007; Zhao et al., 2004; Xie et al., 2008). Thermistors with an accuracy of better than 0.1°C were installed in all the boreholes, which were cased using 5 cm diameter steel pipe, and were also inserted in soils at the active-layer monitoring sites to measure the ground temperatures. All the sites were equipped with Campbell Scientific, Inc. data-loggers (Model: CR1000, CR10X or CR3000) to record the measurements.

Table 1. Location and surface conditions of the monitoring boreholes in the study region
SiteLocationLongitude (°E)Latitude (°N)Elevation (m a.s.l.)RegionsMonitoring periodSurface vegetationPerson responsible
QTB 01Xidatan94.8335.734530QTP, China2001–2008Sparse grassy dry steppeLin Zhao
XDTXidatan94.8335.734530QTP, China2005–2008meadowLin Zhao
QTB 02Kunlun Pass94.0735.634753QTP, China2001–2008Sparse grassy steppeLin Zhao
KM1Kunlun Pass94.0735.634770QTP, China1996–2002Sparse grassy steppeQingbai Wu
KM2Kunlun Pass94.0735.624759QTP, China1996–2002Sparse grassy steppeQingbai Wu
QTB 0366 Squad93.7935.534560QTP, China2001–2008Sparse grassy dry steppeLin Zhao
XSHXieshuihe93.5335.404482QTP, China1996–2002Sparse grassy dry steppeQingbai Wu
CM1Cumaer93.7335.514482QTP, China1996–2008Sparse grassy dry steppeQingbai Wu
QTB 04Qingshuihe93.6035.444488QTP, China2005–2008Grassy meadowLin Zhao
QTB 05Chumaer River93.4635.374520QTP, China2005–2008Desert steppeLin Zhao
QTB 06Wudaoliang93.2735.294563QTP, China2005–2008Alpine steppeLin Zhao
QTB 07Wudaoliang93.0835.204656QTP, China2004–2008Alpine meadowLin Zhao
QTB 09Kekexili93.0335.134740QTP, China2005–2008Alpine meadowLin Zhao
WDLWudaoliang93.7535.134707QTP, China1996–2002Alpine steppeQingbai Wu
KKXLKekexili93.0435.134707QTP, China1996–2008Alpine meadowQingbai Wu
FHSFenghuoshan92.8934.694938QTP, China1996–2008Alpine meadowQingbai Wu
QTB 11Wuli92.6634.394623QTP, China2005–2008Alpine steppeLin Zhao
QTB 15Wenquan91.9133.114960QTP, China2004–2008Alpine meadowLin Zhao
QTB 16Tanggula91.9433.085100QTP, China2005–2008Grassy meadowLin Zhao
QTB 18Liangdaohe91.7431.824808QTP, China2004–2008Swamp meadowLin Zhao
China09Source of Urumqi River86.8543.123500Tien Shan, China1992–2008MeadowLin Zhao
China10Source of Urumqi River86.8143.113900Tien Shan, China1992–2002Bed rock, vegetation freeLin Zhao
M1aBaganuur108.2647.691350Hentei, Mongolia1996–2009Sparse grassy steppeN. Sharkhuu
M3Argalant106.5547.921385Hentei, Mongolia1999–2009Sparse grassy steppeN. Sharkhuu
M6aTerkh99.3748.082050Hangai, Mongolia2002–2009Sparse grassy steppeN. Sharkhuu
M7aChuluut100.3948.041870Hangai, Mongolia2002–2009Sparse grassy dry steppeN. Sharkhuu
M8Sharga98.6649.491855Hovsgol, Mongolia2002–2009Sparse grassy dry steppeN. Sharkhuu
M4aBurenkhan100.0349.791705Hovsgol, Mongolia1996–2009mountain grasslandN. Sharkhuu
K0 76.9443.043330Tien Shan, Kazakhstan1974–2009Bare groundS.S. Marchenko
K1 76.9543.043328Tien Shan, Kazakhstan1974–2009Grass patchesS.S. Marchenko
K2 76.9543.043325Tien Shan, Kazakhstan1998–2009GrasslandS.S. Marchenko
C70 76.9443.043330Tien Shan, Kazakhstan1998–2009Bare groundS.S. Marchenko
Table 2. Locations and surface conditions of the active-layer monitoring sites in the study region
SiteLocationLongitude (°E)Latitude (°N)Elevation (m a.s.l.)RegionsMonitoring periodSurface vegetationPerson responsible
China06Kunlun Pass94.0735.624746QTP, China2005–2008sparse grassy steppeLin Zhao
China02Suoladajie93.6035.434488QTP, China1999–2008steppeLin Zhao
QT01Kekexili93.0535.154734QTP, China2004–2009alpine meadowLin Zhao
QT02Beiluhe192.9234.824656QTP, China2004–2009alpine meadowLin Zhao
QT03Beiluhe292.9234.824656QTP, China2004–2009alpine meadowLin Zhao
China01Fenghuoshan92.9034.734896QTP, China1998–2009alpine meadowLin Zhao
QT05Kaixinling92.4033.954652QTP, China2004–2009desert steppeLin Zhao
QT06Tongtianhe92.8733.584650QTP, China2004–2007steppeLin Zhao
QT04Tanggula91.0232.975100QTP, China2006–2009meadowLin Zhao
China04Liangdaohe91.7331.824808QTP, China1999–2009swamp meadowLin Zhao
China09Source of Urumqi River86.8543.123500Tien Shan, China1992–2008MeadowLin Zhao
M1aBaganuur108.2647.691350Hentei, Mongolia1996–2009Sparse grassy steppeN. Sharkhuu
M3Argalant106.5547.921385Hentei, Mongolia1999–2009Sparse grassy steppeN. Sharkhuu
M6aTerkh99.3748.082050Hangai, Mongolia2002–2009Sparse grassy steppeN. Sharkhuu
M7aChuluut100.3948.041870Hangai, Mongolia2002–2009Sparse grassy dry steppeN. Sharkhuu
M8Sharga98.6649.491855Hovsgol, Mongolia2002–2009Sparse grassy dry steppeN. Sharkhuu
M4aBurenkhan100.0349.791705Hovsgol, Mongolia1996–2009mountain grasslandN. Sharkhuu


Permafrost zones occupy almost two-thirds of Mongolia, predominantly in the Hentei, Hovsgol, Hangai and Altai Mountains and surrounding areas. The region is characterized by mountain and arid-land permafrost, sporadic to continuous in extent, and occupies the southern fringe of the Siberian permafrost zones. Most of the permafrost is at temperatures close to 0°C and so is thermally unstable in relation to climate change and human activities (Sharkhuu et al., 2008). Permafrost has been monitored consistently in Mongolia since 1996 and during the IPY, five additional boreholes (10–30 m deep) were drilled at sites with differing surface conditions (vegetation, snow thickness and soil moisture content). There are now 16 monitoring sites, including 46 of both GTN-P and CALM active boreholes in Mongolia (Figure 1). In the past 5–10 years, 13 boreholes with depths of 10–15 m were re-drilled near older sites where temperature measurements were made 15–40 years ago. All the re-drilled boreholes are cased by parallel steel and plastic pipes, 3–5 cm in diameter. The main variables being monitored are active-layer thickness (ALT) and mean annual ground temperature (MAGT) at the depth of the zero annual amplitude (about 10–15 m depth) and the geothermal gradient. Temperatures are measured by thermistors with an accuracy better than 0.1°C at the corresponding depths of the old boreholes, and temperature data-loggers (Onset HOBO U12) are installed in most of the boreholes. Six typical boreholes with longer data series were selected for this study (Tables 1 and 2).

Tien Shan Mountains

Continuous permafrost occurs above 3600 m above sea-level (a.s.l.) in the central northern Tien Shan Mountains. The discontinuous zone extends from 3200 to 3600 m a.s.l., and the sporadic zone from 2700 to 3200 m a.s.l. However, small isolated patches of permafrost can be found below 2700 m a.s.l. at the base of north-facing or shaded slopes, inside coarse blocky debris or beneath mossy spruce forest even at 1800 m a.s.l. where the MAAT is 3.0–4.0°C (Gorbunov et al., 2004). Permafrost and seasonally frozen ground investigations have been carried out over the past 35 years in selected areas using a variety of methods including measurements of ground and spring water temperatures, DC resistivity soundings, and modelling (Gorbunov and Nemov, 1978; Gorbunov and Titkov, 1989; Seversky and Seversky, 1990; Jin et al., 1993, 1998; Gorbunov, 1996; Marchenko, 2003; Marchenko et al., 2007). Before 2003, ground temperature measurements were carried out with MMT-4 thermistors with a resolution of 0.02°C and an accuracy not better than 0.05°C. Since 2004 in the framework of the Thermal State of Permafrost (TSP) project and especially during the IPY, many boreholes were equipped with Onset HOBO two and four channel 12-bit resolution data-loggers, which provide high accuracy. Two boreholes in middle Tien Shan of China were drilled in 1991 with occasional monitoring of the ground temperatures (Table 1). One of the sites (China10) was destroyed in 2002, but the other was equipped with a Campbell 10X data-logger that started recording in 2000 using 107-type probes with an accuracy of 0.1°C.


The Qinghai-Tibet Plateau

Permafrost temperatures on the Qinghai–Tibet Plateau decrease as the elevation increases. The MAGT at the 50-cm depth at sites with an elevation from 4650 to 4700 m a.s.l. varied between −1.0 and −2.0°C, while at sites above 4700 m a.s.l., the MAGT generally varied between −2.0 and −3.0°C. The MAGT at 50-cm depth declines northwards. Similar spatial distribution characteristics for MAGT exist at the base of the active layer (TTOP) (see Figure 5). Ground surface condition, such as vegetation and snow cover, etc., also influence the MAGT. For example, better developed meadow vegetation cover at Site XDT caused the MAGT at 6 m depth to be somewhat lower than that at Site QTB1 (Table 1 and Figure 2) where sparse steppe vegetation is present, even though these two sites are located at nearly the same location and have the same elevation.

Figure 2.

Qinghai-Tibet Plateau sites: (a) borehole elevation; (b) ground temperature at 6 m depth; and (c) scattergram of borehole latitude and elevation labelled with the ground temperatures measured during the International Polar Year.

The ALT varied between 105 and 322 cm at the monitoring sites during 2007–2008. Generally, the ALT decreased with increasing elevation. Active-layer thickness at sites with elevations greater than 4700 m a.s.l. varied between 100 and 200 cm, while it was generally more than 200 cm at sites with elevations below 4700 m a.s.l. However, the ALT at Tanggula was more than 300 cm due to the influence of latitude and ground surface conditions.


Active-layer thickness in Mongolia varied spatially in different regions during the IPY period from 1–1.5 to 5–8 m, but was predominantly from 2.5 to 4.0 m (Table 3). The maximum seasonally thawed layer is observed in January–February in boreholes such as M1a and M3, located at permafrost sites with relatively high MAGTs, low to medium ice content, or bedrock. Full refreezing of these seasonally thawed zones continues until February–April. Relatively shallow active layers are characteristic of boreholes such as M6a, M7a and M8, located at sites where permafrost has a relatively low MAGT or ice-rich fine-grained sediments. The shallow seasonally thawed layer completely refreezes in November–December.

Table 3. Active-layer thickness (ALT), mean annual ground temperature (MAGT) and geothermal gradient at selected boreholes, in Mongolia
RegionBorehole name and numberMeasured yearsALT (cm)MAGT at 10–15 m depth (°C)Geothermal gradient at 15–50 m depth (°C m−1)
HenteiBaganuur M1a1976355−0.450.020
 Argalant M31989600−0.48
HangaiTerkh M6a1969205−2.040.022
 Chuluut M7a1969125−0.720.038
HovsgolSharga M81968265−2.350.026
 Burenkhan M4a1987285−1.000.026

Tien Shan

Initial geothermal observations (1974–1977) in boreholes in the Kazakh part of the northern Tien Shan showed that the permafrost temperatures within unconsolidated deposits and bedrock at an elevation of 3300 m a.s.l. varied from −0.4°C to −0.8°C (Gorbunov and Nemov, 1978) and the maximum ALT reached 3.5–4.0 m. Coarse blocky debris of various origins occupies a large part of this high mountainous territory and promotes convective mass and heat transfer, especially during cold seasons, because of its high porosity. Measurements during 1974–1987 show that the temperatures inside the coarse debris are typically 3–5°C colder than the mean annual air temperature (MAAT) (Gorbunov et al., 2004). Figure 3 demonstrates that cooling propagated more deeply during winter 2007 in borehole C70. Normally, in Zailiysky Alatau at an elevation 3300 m a.s.l. seasonal, freezing penetrates from 3.5 to 5.2 m within permafrost-free areas or at sites with degrading permafrost. At two observation sites in the Zailiysky Alatau Range the long-term thawing of permafrost started at the end of the 1990s when a supra-permafrost talik began to develop. The MAGT at 160 cm during this period at China09 in the Tien Shan, China was −1.15°C. This site is located nearly at the valley bottom of the headwaters of Urumqi River. The MAGT at the depth of zero annual amplitude (10 m) was −1.07°C (Figure 4).

Figure 3.

Soil temperatures at 4 m and 6 m depths in the C70 borehole (Kazakhstan, Zailiysky Alatau Range, 3340 m a.s.l.) during the International Polar Year.

Figure 4.

Mean monthly ground temperatures at China09 during the International Polar Year.


The Qinghai–Tibet Plateau

Up to a decade of mean annual TTOP, MAGT50 (mean annual ground temperature at the depth of 50 cm) and the ALT have been compiled for 10 sites (Figure 5). The TTOP values increased between 0.02 and 0.19°C yr−1 (mean 0.1°C yr−1) up to 2006, then declined between −0.4°C yr−1 and −0.1°C yr−1 over the following two years. Rates of change were generally higher in regions where the permafrost is colder and lower where it is near 0°C. Over the past 10 years as a whole, TTOP for China01, China02 and China04 increased at an average rate of 0.06°C yr−1. Trends in MAGT50 are quite similar to those of TTOP at the different sites, with increases until 2006 and decreases since then (Figure 5b).

Figure 5.

The thermal state of the active layer at sites along the Qinghai–Tibet Highway: (a) mean annual ground temperature (MAGT) at the bottom of the active layer (TTOP); (b) MAGT at 50 cm below surface (MAGT50); (c) active-layer thickness (ALT).

Active-layer thickness is generally increasing. Maximum values were observed in 2006 or 2007 at most of the sites. The ALT for the sites with longer time series (China01, China02 and China04) increased between 35 cm and 61 cm from 1999 to 2006/2007, representing an average increase of about 4 cm yr−1. Observation sites with shorter series, such as QT01, QT05 and China06, showed ALT increases of 4–15 cm from 2004 to 2006/2007 except for QT06.

Mean annual ground temperatures increased at 15 of 17 boreholes along the Qinghai–Tibet Highway during the period of record, the exceptions being XDT and QTB16 where monitoring started only in 2005 (Figure 6). The rate of increase in MAGT from 1996 to 2008 was inversely proportional to the MAGT at KM1, KKXL and FHS (0.5°C, 0.45°C and 0.4°C per decade, respectively). The average rate of increase in MAGT at 6 m depth from seven sites was from 0.02 to 0.07 °C yr−1 from 1996 to 2002 but has declined recently, in part because of cooler air temperatures.

Figure 6.

(a) Annual ground temperatures at 6 m depth and (b) rates of change in temperature during the monitoring period at sites along the Qinghai–Tibet Highway. Note: values in (b) are influenced by variable monitoring period lengths (see text).


The ALT monitored at CALM boreholes increased at varying rates (see Figure 7a). The highest rates of increase in ALT of 20–40 cm yr−1 were observed over the past decade at boreholes M1a and M3 where the MAGT was close to 0°C. A low rate of increase usually occurred in shallow active layers over ice-rich and colder permafrost. Small increases in ALT (0.5–2.0 cm yr−1) in some boreholes such as M6a and M7a over the past few years may also be due to soil drying and overgrazing under climate warming. Generally, the rate of increase in ALT during the past 10–15 years was much higher than during the previous 15–20 years. However, the ALT in most boreholes in 2009 was less than that in the previous two years.

Figure 7.

Changes of (a) active-layer thickness (ALT) and (b) mean annual ground temperature (MAGT) at monitoring sites in Mongolia.

The MAGT at permafrost sites in Mongolia varies from 0 to −0.5°C in sporadic permafrost zones and from −1.0 to −3.0°C in continuous permafrost. Mean annual ground temperature increased in all the boreholes over the past 10–40 years, and the geothermal gradient in the upper 15–50 m depth decreased by 0.01–0.02°C m−1 over the same period (Figure 7b). The rate of increase in MAGT and decrease in geothermal gradient varied from site to site, but were greater in bedrock and sandy sediments than in unconsolidated and ice-rich fine sediments. The average rate of increase in MAGT at 10–15 m depth was 0.02–0.03°C yr−1 in the Hovsgol Mountain region, and 0.01–0.02°C yr−1 in the Hangai and Hentei Mountain regions (Sharkhuu et al., 2007). Permafrost warming was greater during the past 15–20 years than during the previous 15–20 years (1970s–1980s) and the average rate of increase in MAGT in Mongolia was about 0.15°C per decade (Sharkhuu et al., 2008). Average trends of recent permafrost warming in Mongolia are similar to those in Central Asia and in the European mountains (Harris and Haeberli, 2003, Marchenko et al., 2007; Harris et al., 2009) but lower than those in eastern Siberia and Alaska (Osterkamp and Romanovsky, 1999, Gavrilova, 2003, Pavlov and Perlshtein, 2006).

Tien Shan

Permafrost temperature observations during 1974–1977 and 1990–2009 indicate that the ground has warmed in the Kazakh part of Tien Shan Mountains over the past 35 years. The increase from 1974 to 2009 varies from 0.3°C to 0.6°C at depths of 14–25 m. Based on interpolation of borehole temperature data, the active layer increased in thickness from 3.2 to 3.4 m in the 1970s to a maximum of 5.2 m in 1992 and to 5.0 m in 2001 and 2004. The average ALT for all measured sites increased by 23% in comparison to the early 1970s (Marchenko et al., 2007).

Monitoring in the headwaters of the Urumqi River in the eastern Tien Shan Mountains, China, indicates that permafrost changed significantly as the air temperature increased (Zhao et al., 2010). Mean annual ground temperatures measured in the China09 borehole increased from −1.6°C in 1992 to −1.0°C in 2008 (Figure 8a). Temperatures at various depths increased by 0.4–0.9°C, a rate of about 0.3–0.5°C per decade, while the depth of zero annual ground temperature amplitude increased from 10 m in 1992 to 12 m in 2008. Active-layer thickness has been increasing since 1992 and reached a maximum (1.6 m) in 2007, which is 0.35 m deeper than that in 1992 (Table 3). The MAGT in China10 (Figure 8b) is much lower than that in China09, probably due to the different elevations, landforms and vegetation covers of the two boreholes (Table 1). Ground temperature in China10 borehole increased by 0.2–0.5°C at the depths of 5–18 m from 1992 to 2002 (Figure 8b), which is greater than in China09 during the same period (0.1–0.4°C).

Figure 8.

Ground temperature variations in (a) China09 and (b) China10 in headwaters of the Urumqi River, eastern Tien Shan Mountains, China.


Measurements in boreholes in Central Asia show that much of the permafrost is currently at temperatures close to 0°C. Longer-term records, where available, demonstrate that permafrost has warmed, active layers have become thicker, and geothermal gradients have decreased over the past few decades. For example, ground temperatures increased by 0.3–0.5°C per decade at the China09 site in the eastern Tien Shan from 1992 onwards. Although there is significant spatial variation in temperatures due to elevation, substrate and vegetation cover, the temporal trends are consistent.

Rates of change in permafrost temperatures vary regionally and in relation to site conditions. Warming is greater at colder sites and at those in bedrock or ice-poor sediments than in warmer, ice-rich sites. Most sites with longer records show warming up to 2006, but several show slight cooling in the IPY period due to colder air temperatures in 2007 and 2008. Mongolian sites seem to be warming more slowly than those on the Qinghai–Tibet Plateau.

This compilation of results for the largest expanse of mountain permafrost in the Northern Hemisphere demonstrates that much of it is sensitive to climate change and consequently, that changes have been significant over the past several decades. Ongoing monitoring at numerous sites is expected to reveal continued warming of ground temperatures and thicker active layers in the future.


The authors thank Dr Jerry Brown for his encouragement, suggestions, editing and other advice and assistance. Thanks to Dr Ren Li and Qiangqiang Pang for their contributions, including data processing and preparing the figures. The authors would like to thank the editor, Professor Antoni Lewkowicz for his assistance, including detailed editing of the English. The Chinese part of the study was funded by the National Basic Research Program (2005CB422003), the Outstanding Youth Foundation Project to Dr Qingbai Wu, and the Chinese National Science Foundation (Grant Numbers 40830533 and 40625004). Both the Mongolian and Kazakhstan projects were support by US National Science Foundation grants under the CALM (NSF OPP 03529580) and TSP (NSF ARC 0632400) projects at the University of Delaware and the University of Alaska, respectively.