Introduction

The response of permafrost to climate change and engineering construction is vitally important for studying the interactions between climate and permafrost, and between engineering infrastructure and permafrost, and to understand terrestrial and hydrological process in cold regions (Koster, 1993; Nelson et al., 1993; Zhang and Stamnes, 1998). Under global climate warming scenarios, permafrost is projected to undergo great change (Wu and Tong, 1995; Zhao and Wang, 1996; Li and Cheng, 1997). Permafrost change under a disturbed surface will be greater than under natural surfaces. The effect of global climate change on permafrost is less than that of engineering activity on it. According to drilling data from the 1980s and 1990s, the northern and southern limits of the Qinghai-Tibet Plateau permafrost have not changed substantially under natural conditions, with only minor changes in the thickness and temperatures of the permafrost. However, under asphalt pavement, permafrost has undergone great changes. Not only has the depth of the permafrost table changed significantly, permafrost also has degraded, resulting in a 5- to 8-km southward retreat of northern boundary and a 9- to 12-km northward retreat of the southern boundary. The permafrost table deepens greatly under an asphalt pavement (Wang and Mi, 1993; Wu and Tong, 1995).

Climate change and engineering construction are both expected to thicken the active layer and raise permafrost temperature. However, the responses of permafrost to climate change and to engineering construction differ substantially on the Qinghai-Tibet Plateau due to local variations in topography, geology, hydrology, vegetation, and soils, which is of interest to geotechnical engineers and scientists. Formerly, it was believed that cold (<−1.5°C) permafrost is more thermally stable and subsequently less sensitive to climate change and engineering activity than warm (≥−1.5°C) permafrost (Cheng, 1984; Tong and Wu 1996; Wang et al., 1996; Ding, 1998). However, monitoring of permafrost temperatures under the influences of a changing climate and the disturbances from the construction and operation of the Qinghai-Tibet Highway necessitate the reconsideration of this conclusion. The observations show that the rates of increase in permafrost surface temperatures (temperature at the permafrost table) and in permafrost temperature are substantially greater for areas with cold permafrost than areas with warm permafrost under a changing climate. Climate changes, however, have greater impacts on the thermal regimes of cold permafrost than engineering activities.

Much research on permafrost along the Qinghai-Tibet Highway are focused on either climate change or engineering activity (Yu et al., 2002; Wu et al., 2003; Wu and Liu, 2004). There are very few comparisons of the effects of climate change and engineering activities on the thermal regimes of permafrost. Seven field sites were installed along the Qinghai-Tibet Highway from Xidatan to Tanggula Pass in 1995 in order to monitor the responses of permafrost and active layer processes to climate changes and engineering construction and operations, and to find the important factors for consideration in engineering designs at different locations and at different timescales.

Study Area and Methods

Permafrost is present in most areas above 4100 m elevation along the Qinghai-Tibet Highway. On the Qinghai-Tibet Plateau, the effect of elevation on permafrost distribution is stronger than that of latitude (Cheng and Wang, 1982; Cheng, 1984). Generally, mean annual ground temperatures (MAGT) decrease with increasing elevations at rates of 8 to 9°C km⁻¹. The MAGTs rise 1°C and the lower limit of permafrost rises 100 to 110 m per degree of latitude (Tong and Li, 1982). Elevation is the controlling factor for the distribution and characteristics of permafrost on the Qinghai-Tibet Plateau. The MAGTs vary from −0.1 to −4.5°C, decreasing with increasing elevation and latitude. Permafrost thickness is from several meters to more than 100 m (Li, 1982). The permafrost table ranges from 0.8 to 3.8 m depth. Permafrost along the Qinghai-Tibet Highway has lower thermal gradients of about 46°C km⁻¹, compared with that in unfrozen soil under permafrost of about 61°C km⁻¹ (Wang and Li, 1983; Guo, 1985).

In order to investigate the response of permafrost to climate change along the Qinghai-Tibet Highway, seven monitoring sites were installed along the Qinghai-Tibet Highway according to the local MAGTs and sensitivity of the permafrost to climate change and engineering activity. The basic data for each site is shown in Tables 1 and 2.

Table 1

Information on permafrost monitoring sites along the Qinghai-Tibet Highway.

Table 2

Permafrost table depth under natural surfaces, 1996–2001.

At each monitoring site, thermistor cables were installed under the natural surface, the road shoulder, and centerline. Permafrost temperature and freeze-thaw processes under natural and engineering effects were monitored on 5th and 20th day of each month. The monitoring began in 1995 and lasted for 6 yr for sites 1, 2, 3, and 5 and it continues to the present for sites 4, 7, and 8.

Permafrost Change Under the Effect of Climate Change

The permafrost table deepened in all seven monitored natural surface sites from 1996 to 2001 (Table 2), at mean rates that ranged from 3.6 to 9.2 cm yr⁻¹. The permafrost table shows some minor interannual variations under natural surfaces (Table 2). Table 3 gives the variations of mean annual permafrost surface temperature at seven sites from 1996 to 2001. The mean annual rates of changes in the permafrost surface temperature are about 0.06 to 0.10°C yr⁻¹ for cold permafrost and about 0.01 to 0.04°C yr⁻¹ for warm permafrost. The mean annual temperature at 6 m depth of rose about 0.04 to 0.06°C yr⁻¹ for cold permafrost, and about 0.02 to 0.03°C yr⁻¹ for warm permafrost (Table 4). The rate of change of cold permafrost is obviously larger than that of warm permafrost, whether considering permafrost surface temperature or permafrost temperature at 6 m depth.

Table 3

Permafrost surface temperature under natural surfaces, 1996–2001.

Table 4

Permafrost temperature at 6 m depth under natural surfaces, 1996–2001.

Permafrost Change Under Engineered Surfaces

The mean annual permafrost surface temperature of warm permafrost under asphalt rose 0.08 to 0.14°C yr⁻¹, compared to a rise of 0.01 to 0.03°C yr⁻¹ for cold permafrost (Table 5). In accordance with the change of mean permafrost surface temperature, the depth to the permafrost table increased at a rate of 11 to 31 cm yr⁻¹ for warm permafrost and 1 to 4.6 cm yr⁻¹ for cold permafrost (Table 6). Permafrost temperature changes at 6 m depth accords with the change in permafrost surface temperature and permafrost table depth. The mean annual temperature at 6 m depth under asphalt rose faster for warm permafrost (0.11°C yr⁻¹) than for cold permafrost (0.02–0.05°C yr⁻¹) (Table 7). The effect of engineering activity on permafrost change is greater than the influence of climate change.

Table 5

Permafrost surface temperature under asphalt pavement, 1996–2001.

Table 6

Permafrost table depth under asphalt pavement, 1996–2001.

Table 7

Permafrost temperature at 6 m depth under asphalt pavement, 1996–2001.

Discussion

The effect of climate change is larger than the effect of engineering activity for cold permafrost, but the effect of climate change is smaller than the effect of engineering activity for warm permafrost. Actually, permafrost change beneath engineered embankments, for either cold permafrost or warm permafrost, is affected by both climate change and engineering activity so that permafrost changes beneath the embankment should be larger than that under the effect of climate change all the time. However, the monitoring results show that cold permafrost change beneath asphalt pavement is lower than cold permafrost change beneath natural surfaces. This is because of the difference in the permafrost response to the thermal disturbance of engineering construction and climate change.

Response of permafrost to climate change is a slow process; long-term effects of climate change produce a slower change of permafrost. However, response of permafrost to engineering activity is quick; the short-term effect of engineering activity produces a large change in permafrost. Over the short term, the much larger influence of engineering on permafrost dwarfs possibly the impacts of climate change. Figure 1 shows permafrost surface temperature change for cold and warm permafrost under the effect of climate change and engineering activity. The mean permafrost surface temperature in warm permafrost is about 0.5°C greater under asphalt than natural surfaces, and in cold permafrost is about 2.9°C greater under asphalt than natural surfaces (Fig. 1). Given the rates shown in Table 3, it will take at least 50 yr for climate change to raise permafrost surface temperatures under natural surfaces to the temperatures under asphalt for cold permafrost and 20 yr for warm permafrost. Figure 2 shows permafrost temperature change at 6 m for cold and warm permafrost under the effect of climate change and engineering activity. Figure 2 indicates that engineering activity can produce mean permafrost temperature at increase at 6 m depth of 0.50 to 1.5°C. Present climate change rates will produce this great a change at 6 m depth in 20 yr for cold permafrost, but only 5 to 8 yr for warm permafrost.

Figure 1.

Permafrost surface temperature under the natural and engineered (asphalt) surfaces. Solid points are under natural surface and open points are under the central roadbed.

Figure 2.

Permafrost temperature change at 6 m depth under the natural and engineered (asphalt) surfaces. Solid points are under natural surface and open points are under the central roadbed.

The effect of engineering thermal disturbance is only to raise the permafrost temperature for cold permafrost. For warm permafrost, energy is also used to thaw ground ice near the permafrost table. Cold permafrost beneath embankments with asphalt pavement produced a larger change in the first years after engineering construction compared to warm permafrost so that the change of cold permafrost beneath embankment with asphalt pavement is comparatively insensitive to climate change. Because of close relationships of permafrost with climate change and engineering activity, engineering designers must consider the effect of climate change and engineering activities on permafrost from the early stages of engineering construction for the areas with warm permafrost. However, engineering designers should not consider the effect of climate change until permafrost change is more than the change under the engineering state for the areas with cold permafrost.

Recent Climate Change

Temperature change of cold permafrost is greater than those of warm permafrost under the effect of climate change. There are only two meteorological stations, Wudaoliang and Tuotuohe Riverside, in permafrost regions on the Qinghai-Tibet Plateau. Wudaoliang Meteorological Station is located in a cold permafrost region and Tuotuohe Riverside Meteorological Station is located in a warm permafrost region. The change of air temperature at the Wudaoliang Station seems to be greater than at Tuotuohe Riverside Station (Figs. 3, 4). The mean annual air temperature (MAAT) at the Wudaoliang Station rose 0.7°C from about −5.9°C in the 1960s to about −5.2°C in the 1990s. The MAAT at the Tuotuohe Riverside Station rose 0.5°C from about −4.4°C in the 1960s, to about −3.9°C in the 1990s. Mean annual surface temperature (MAST) in the 1960s was about −2.0°C, and about −0.7°C in the 1990s, indicating an about 1.3°C rise of surface temperature during the past 40 yr at the Wudaoliang Meteorological Station. The MAST in the 1960s was about −0.78°C, and about −0.26°C in the 1990s, indicating a rise of about 0.52°C during the past 40 yr at the Tuotuohe Riverside. The change of the MASTs is about 0.8°C higher than that of the MAATs at the Wudaoliang Station. The amplitude of the changes at the Tuotuohe Riverside is smaller than that at the Wudaoliang Station. It seems that the climatic warming in the mountainous areas is more significant than that in the river valleys on the Qinghai-Tibet Plateau.

Figure 3.

MAAT (mean annual air temperature) and MAST (mean annual surface temperature) at the Wudaoliang Meteorological Station on the Qinghai-Tibet Plateau over the past 40 yr.

Figure 4.

MAAT and MAST at the Tuotuohe Riverside Meteorological Station on the Qinghai-Tibet Plateau over the past 40 yr.

Conclusions

The changes in cold permafrost are larger than that in warm permafrost on the Qinghai-Tibet Plateau under a warming climate. The mean annual increase in permafrost table depth is about 3.6 to 8.2 cm yr⁻¹ for cold permafrost and about 4.4 to 9.2 cm yr⁻¹ for warm permafrost under natural surfaces. The rate of mean annual permafrost surface temperature rise is about 0.06 to 0.10°C yr⁻¹ for cold permafrost and about 0.01 to 0.04°C yr⁻¹ for warm permafrost. The rate of change in the mean annual permafrost temperature at 6 m depth is about 0.04 to 0.06°C yr⁻¹ for cold permafrost and about 0.02 to 0.03°C yr⁻¹ for warm permafrost.

The change in warm permafrost is greater than cold permafrost under asphalt surfaces. The rate of the rise in the mean annual permafrost surface temperature is 0.08 to 0.14°C yr⁻¹ for warm permafrost and 0.01 to 0.03°C yr⁻¹ for cold permafrost. The mean annual increment of the permafrost table is 11 to 31 cm yr⁻¹ for warm permafrost and 1 to 4.6 cm yr⁻¹ for cold permafrost. The rate of the rise in the mean annual permafrost temperature at 6 m depth is 0.11°C yr⁻¹ for warm permafrost and 0.02 to 0.05°C yr⁻¹ for cold permafrost.

The response of permafrost to climate change and engineering activity differs significantly. The change of cold permafrost is larger than that of warm permafrost under the effect of climate change and is smaller than that of warm permafrost under the effect of engineering activity. Over time, the impacts of engineering activity on cold permafrost dwarfs that resulted from the climate change.

This research makes an important revelation for considering climate change for engineering construction in cold regions. Warm permafrost under engineered surfaces experiences the combined impacts of climate change and engineering several years after construction. On the other hand, cold permafrost under engineered surfaces would need to wait about 30 to 50 yr to undergo the combined influence of the climate change and engineering activity.