Thermal regime of conventional embankments along the Qinghai–Tibet Railway in permafrost regions

Abstract

Thermal regime of conventional embankments (CEs) was investigated by compiling ground temperature data from 13 monitoring sites along the Qinghai–Tibet Railway in permafrost regions. The investigation included shallow ground temperatures within embankment, variations of permafrost table and dynamic changes of deep permafrost temperature beneath embankment. The results indicated that the shallow ground temperatures within two embankment shoulders differed from each other; the difference at depth of 50 cm was less than 1 °C in warm seasons but more than 2.7 °C in cold seasons. Statistic analyses of permafrost table variations beneath CEs confirmed existence of a reasonable height range of CE in regions with mean annual ground temperature (MAGT) < − 0.6 to − 0.7 °C. When the height of CE lay in the range, underlying permafrost tables all had obvious upward movements and could be maintained well with time. By contrast, in regions with MAGT > − 0.6 °C, permafrost tables either declined (beneath sunny shoulders) or maintained near original level (beneath shady shoulders) at majority sites no matter what the height of CE was. As these permafrost tables were not stable; ground temperatures near them increased slowly with time. Analyses of ground temperature profiles indicated that the deep permafrost beneath CEs all warmed in the contexts of climate warming and construction activities, and this warming trend was more pronounced in warm permafrost regions. Supra-permafrost talik developed and/or permafrost degraded from both permafrost table and permafrost base beneath embankment at some sites with MAGT > − 0.5 °C.

Introduction

From an engineering viewpoint, the bearing strength of permafrost is a function of its temperature (Esch and Osterkamp, 1990, Lunardini, 1996). Therefore, in order to reduce the risk of structure failure in a permafrost environment, thermal stability of the ground has to be the main goal (Harris et al., 2009). As a major linearity engineering on the Qinghai–Tibet Plateau (QTP), the Qinghai–Tibet Railway (QTR) was a landmark project in permafrost regions (Zhang et al., 2008). The successful construction of the railway benefited from detailed geotechnical investigation, advanced design idea, high-quality construction process and precise monitoring system during all project phases (Ma et al., 2011). In unstable permafrost regions, the railway adopted the advanced design idea, “cooled roadbed”, to proactively lower underlying ground temperature in the contexts of construction activities and climate warming (Cheng, 2003, Cheng, 2005, Ma et al., 2002). A number of active cooling measures including crushed rocks, thermosyphons, ventilation ducts, awnings and dry bridges were applied to cool the permafrost beneath embankment, and all of them performed satisfactorily at present (Cheng et al., 2008, Feng et al., 2006, Ma et al., 2006, Ma et al., 2009, Niu et al., 2006, Wu et al., 2008, Wu et al., 2010, Zhang et al., 2005). But, except these embankments with cooling measures, there were still some sections of the embankment constructed with conventional technique, which only placed gravel pad directly upon the undisturbed ground surface without any cooling measures, called as conventional embankment (Wu et al., 2007). The thermal regime of conventional embankment (CE) determines to a large extent the embankment performance relative to settlement and stability, and therefore need detailed investigation.

In order to prevent decline of underlying permafrost table after construction phase and formation of talik during the operation phase, CE should be built within a reasonable height range (Cheng et al., 1983, Wu et al., 1988, Wu et al., 1998). In cold permafrost regions (mean ground temperature close to − 11 °C), a gravel embankment thickness of about 1.5 m is usually adequate to prevent thawing of the underlying permafrost (Berg and Quinn, 1977). But on the QTP, permafrost is characterized by high mean annual ground temperature (MAGT), ranging between 0 and − 4.0 °C, and consequently by a weak thermal stability (Wu et al., 2010a). So only in regions with MAGT or mean annual air temperature (MAAT) lower than certain values, the reasonable height range of CE does exist. According to field observations of freezing and thawing degree days and n factors along the Qinghai–Tibet Highway, Ding and He (2000) pointed out that the certain value of MAAT was − 3.8 °C for embankment constructed with fine-grained soil. Combining with observation of experimental embankment at Beiluhe Basion, Zhang et al., 2005, Zhang et al., 2006 conducted some numerical simulations and proposed that the certain values of MAGT and MAAT were − 0.6 °C and − 3.5 °C, respectively. However, up to now, all these values had not been validated by engineering practice. Additionally, in locations where it is impractical to build a thicker embankment, placement of insulation in thinner embankment provides an alternative. Insulation layer was initially installed in roadways and airfield runways in North American (Esch, 1973). At Beiluhe Basin, some experimental embankments with insulation layer were constructed, but the results showed it was not suitable for warm permafrost regions with high embankment surface temperature (Sheng et al., 2003, Wen et al., 2008).

Another problem of CE that researchers and engineers need to take into consideration is thermal effect related to embankment side-slopes (Cheng et al., 2003). After embankment construction, the two side-slopes with different aspects receive different amounts of solar radiation and thermal difference between each other is the result. This thermal difference could be considerable since the combination of low latitude, high elevation, thinner atmosphere and higher atmosphere transparency makes the QTP one of the areas with strongest solar radiation on the earth (Chen et al., 2006). Based on filed experiments at Beiluhe Basin, Hu et al. (2002) studied the relationship between solar radiation and temperatures of the two side-slopes with different embankment orientations and proposed a method to calculate both of them. Sheng et al. (2005) observed temperatures of two side-slopes of an experimental embankment at Beiluhe Basin for one year, and concluded that high temperature of south-facing slope in winter contributed to a large extent to the thermal difference between south- and north-facing slopes. Chou et al., 2008a, Chou et al., 2008b researched the thermal difference between the two side-slopes and their influence on underlying permafrost table distribution by numerical simulations. Wu et al. (2010b) examined the freezing and thawing periods of side-slopes, average temperatures of active layer and permafrost tables beneath crushed rock embankments and CEs with different orientations along the QTR.

However, researches above always focused on active layer thickness and underlying permafrost table distribution but seldom cared about deep permafrost temperature beneath embankment. Investigation of crushed rock embankments along the QTR founded that the deep permafrost had some warming trends even when permafrost tables rise obviously in some regions (Ma et al., 2008, Mu et al., 2010, Wu et al., 2006b). These warming trends may be more considerable beneath CEs and consequently undermine the embankment stability. Secondly, researches about CE above generally conducted at some specific regions, such as Beiluhe Basin. Thus results from these researches could not be easily applied to other regions along the QTR since the complicated environment conditions and permafrost distribution on the QTP. In this paper, the thermal regime of CEs along the QTR in 550 km of permafrost regions was comparatively investigated based on ground temperature data compiled from 13 monitoring sites. Shallow ground temperatures within embankment, permafrost table variations and dynamic changes of deep permafrost temperatures beneath embankment were investigated to evaluate the thermal performances of CEs and their influences on long-term embankment stability. The investigation is essential for precise diagnosis of embankment failures and adoption of appropriate remedial measures, and meanwhile provides high-quality in-situ data series for numerical simulation in future.