1 Introduction
Warm permafrost (permafrost with a mean annual ground temperature higher than −1.5 °C) is widely distributed on the Qinghai−Tibet Plateau (QTP) in China [
1–
3]. Against the background of intensifying human engineering activities and global warming, the warm permafrost will be an important area of transportation infrastructure construction on the QTP in the future. Compared with general permafrost, warm permafrost has the characteristics of high temperature sensitivity, low mechanical strength, and strong thawing settlement [
4–
6]. During the operation of the Qinghai−Tibet Highway (QTH) and the Qinghai−Tibet Railway (QTR), it has been found that road engineering distresses, including longitudinal pavement cracks, thawing and subsidence, frost heaving, and slope instability, have been mainly distributed in sections with warm and ice-rich permafrost [
7,
8]. These distresses were largely related to the temperature state and water migration in embankment soils, while the main factors affecting thermal balance of the embankment were climate warming, engineering activities, and the effect of the shady–sunny slope [
9,
10]. Moreover, the embankment is subject to freeze–thaw patterns that are unbalanced on the two sides of the railway or roadway, due to the degradation of the underlying permafrost and the transverse temperature difference, which will eventually lead to various distresses caused by asymmetrical settlement of the permafrost embankment, affecting its normal operation and service life.
The Gonghe−Yushu Expressway (GYE) is the first expressway built in the warm permafrost regions of the QTP, has faced some new technical challenges. The first one is the thermal effect of the expressway embankment. Compared with other previously-built highways in permafrost regions, the GYE identifies itself by wider embankment, thicker pavement structure and black asphalt pavement. The black asphalt pavement absorbs more solar radiation than unpaved or gravel-paved embankment. The heat and water-vapor channels inside the embankment are blocked by the thicker pavement structure and it is difficult to transfer heat to the outside environment, resulting in heat collection and storage in the embankment. Studies have shown that wide asphalt-paved embankment has greater heat absorption intensity and more remarkable thermal effect on the thermal budget of underlying permafrost [
7,
11,
12]. The second challenge is the warm permafrost. The GYE is located on the south-eastern margin of the QTP, where the mean annual ground temperatures along the route are generally above −1 °C. The third challenge is climate warming. Maduo county had an average annual climate warming rate more than twice that of the average of the QTP, which undoubtedly accelerates the degradation of permafrost [
11,
13].
Regarding the technical challenges mentioned above, a variety of embankment structures and engineering measures were adopted in the construction of the GYE, including crushed-rock embankment (CRE), ventilation ducts embankment, L-shaped thermosyphon embankment, thermal insulation board embankment and combined structures of these. Among all the cooling measures, ventilation embankments were adopted most widely. Some studies regarding to the cooling mechanism and effect have been carried out. Pei et al. [
14] established the first coupling thermo-mechanical model for a CRE, which supplied a new method to effectively evaluate the embankment stability. Laboratory and numerical results showed that ventilation embankment structures, including CRE, ventilation embankment and combined ventilation and CRE, had significant effects on cooling permafrost [
15–
17]. To investigate the cooling performance of the cooling technologies and supply a reference for subsequent expressway construction in permafrost regions, an 8.06 km long scientific and technological test and demonstration section was designed and built by CCCC First Highway Consultants Co., Ltd. In-site monitor started in September of 2016 [
11].
In this study, the monitoring data for the ventilation embankments at the test and demonstration section were collected to study the thermal characteristics and stability of GYE ventilated embankments under the influence of wide asphalt pavement, climate warming, engineering activities, heat collecting effect and shady–sunny slope effect during operation process. The purpose was to allow analysis of the cooling performance of each type of embankment and to provide technical support for subsequent construction and maintenance of expressways in warm permafrost regions.
2 Description of the Gonghe−Yushu Expressway test and demonstration section and the monitoring system
2.1 The Gonghe−Yushu Expressway test and demonstration section
The construction of the GYE was completed in 2017. It has a total length of 635.9 km, of which 227.7 km passes through the permafrost regions. It is located at the south-western edge of the QTP. The permafrost temperature along the GYE was between −1.5 and 0 °C, which was classified as warm permafrost (as shown in Fig.1). The test and demonstration section was located between K566 + 400 and K574 + 460 on the GYE, with a total length of 8.06 km and had an average altitude of about 4300 m and a mean annual air temperature (MAAT) of 2.6 °C. Since the 1980s, the MAAT has rapidly increased at a rate of 0.069 °C/a, which was ten times higher than that before the 1980s [
13].
The area of the test and demonstration section is dominated by north-west winds throughout the year. The embankment of the test section consisted of separated embankment with a single embankment width of 12.25 m, and the thickness of the asphalt pavement was 65 cm. The pre-existing line (also phase I line) of the GYE was located on the left side of the test and demonstration section, and was completed in 2014. The pre-existing line was constructed earlier than the demonstration section, and the distance between the toes of two embankment slopes is about 15 to 25 m. According to the analysis results of geological exploration data, the mean annual ground temperature of the permafrost at the test site ranged from −0.7 to −0.27 °C, the natural permafrost table ranged approximately from 0.5 to 1.5 m, and the ice-rich permafrost accounted for over 65% of the total length of the test and demonstration section [
11,
18].
There are two main types of ventilation embankment used in the test and demonstration section, one is natural ventilation embankment, such as CRE, and the other one is forced ventilation embankment, such as ventilated ducts embankment (VDE). The ventilation embankments selected for the present study at the test and demonstration section are CRE, VDE, combined ventilated slab and CRE (VS
+ CRE), combined crushed-rock and heat-induced pavement embankment (CR
+ HIP), respectively. Among them, heat-induced pavement embankment has a gradient layered structure for heat conduction, and it improves the reflection of solar radiation by the pavement surface, decreases the thermal conductivity of the top layer. This type of structure also inhibits the heat conduction in the middle layer by adding additional materials with different coefficient of heat conduction or heat capacity, so as to reduce the heat flow into the embankment [
19,
20]. The structures of the CRE, VDE, CR + HIP, and VS + CRE are schematically shown in Fig.2. The geological conditions at the four ventilation embankment centers are shown in Fig.3.
2.2 Monitoring system
Embankment monitoring systems mainly consist of temperature monitoring. Temperature data is measured by the thermistors; the measuring accuracy is ±0.05 °C, and the data collection frequency is once every 4 h. The temperature values in this paper are mainly composed of embankments temperature and ground temperature.
The embankment temperature was measured by five transverse thermistor strings. In each thermistor string, the thermistors were arranged symmetrically and had a spacing of 2 m. Meanwhile, the ground temperature was measured by six vertical holes with depths of 15 m, which are respectively the geothermal holes corresponding to the left slope toe, left shoulder, embankment center, right shoulder, right slope toe, and the natural hole that was 15 m away from the right slope toe. The natural surface depth is set at 0.0 m. The thermistors were arranged along the cable at intervals of 0.5 m above −4 and 1.0 m between −4.0 and −15.0 m. The monitoring systems layout of the above embankment structures are shown in Fig.4.
3 Analysis of embankment temperature distribution
3.1 Embankment temperature distribution
To comparatively analyze the ground temperature distribution and variation rules of the four types of embankments, the contour maps of integral temperature for the moment of maximum thawing depth in 2018, 2019, and 2022 are plotted in Fig.5. As can be seen from Fig.5(a)–Fig.5(d), a warm temperature core was formed inside all the four embankments by October 15th of every year, which was caused by the intensive thermal effect of the wide expressway embankment with thick black asphalt pavement. The temperature of warm-temperature core varied with the intensity of solar radiation and the type of embankments. The warm-temperature cores in the CRE and CR + HIP were warmer than those of the VDE and VS + CRE, which was because the effective ventilation performance of the ventilation duct released the heat inside the embankment to the external environment.
Similarly, it can be seen from the isotherms that the temperature fields of the four embankments show asymmetric characteristics. The temperature of the four embankments on the left side were higher than that on the right side at corresponding positions. This was because of the shady–sunny slope effect. The right side was located on the windward side (north-west wind) and had a short lighting time, thus it was the shady slope, while the left side had a longer lighting time, and was the sunny slope. In particular, the shady–sunny slope effect for the CRE and CR + HIP embankments was relatively significant; the left-hand-side isotherm pattern was obvious concave, but the maximum thawing depth was located in the center of the embankment. The artificial permafrost table (APT) marked by solid red line in Fig.5 declined in depth from the surface continuously from 2018 to 2022. In particular, the isotherms of −0.25 and −0.5 °C moved downward and right rapidly year by year, indicating that the permafrost under the four types of embankments was constantly warming and degrading. There were two main reasons for the above phenomenon: 1) under the intensive thermal effect of black and thick asphalt pavement and shady–sunny slope effect, the accumulation and transfer of heat in the embankment were different, which lead to asymmetrical distribution of the warming amplitude of the embankment; 2) the left side of the pre-existing route of the GYE phase I had an obvious transverse thermal disturbance to underlying permafrost of the four ventilation embankments. Therefore, although the ventilation embankment structures changed the distribution characteristics of the temperature field inside the embankments and slowed down the rate of APT descent, they did not change the fact that permafrost gradually degraded.
The deformation and stability of embankment are closely related to the thermal state of the underlying permafrost. Warming of the underlying permafrost and decline of the permafrost table are the decisive factors leading to thaw settlement for the embankment [
21,
22]. Fig.6 shows the ground temperature variations at different embankments’ centers from September 2015 to September 2021. As can be seen from the temperature contour of the CRE in Fig.6(a), the temperature of underlying permafrost was generally high and getting warmer year by year, and the permafrost table was constantly declining. The APT declined from −1.20 m in 2018 to −1.94 m in 2022, and the range of the −0.3 °C isotherm gradually expanded, indicating that the permafrost was gradually degenerating. According to the temperature contours in Fig.6(b), the APT of the VDE decreased from −1.15 m in 2016 to −1.45 m in 2019. However, there was a notable exception in 2020. The permafrost table rose to −1.30 m, with an increase of about 15 cm. Meanwhile, the ground temperature decreased, the range of −0.3 °C isotherm expanded, and the area of −1.0 °C cold core enlarged. By 2021, the permafrost table had been descending to −1.42 m.
From the isotherms in Fig.6(c), it can be found that the permafrost table of the VS + CRE declined from −0.93 m in 2016 to −1.26 m in 2019, and then rose to −1.11 m in 2020 with a total increase of about 16 cm, accompanied by a decrease in ground temperature. The reason for this phenomenon was similar to that for the VDE. Therefore, increasing ventilation air-flow can effectively dissipate heat, reduce ground temperature, and inhibit the degradation of the permafrost. Moreover, the positive and negative temperatures of the ground for the VDE and VS + CRE appeared alternately, indicating that the embankments were in favorable working condition. Compared with the other two embankments, the temperature differences of the VDE varied more widely and had a deeper influence range. This was because the water content of the embankment was greater than that of the VS + CRE, and the high ice content could effectively maintain the stability of the ground temperature. In addition, it was found that the natural ground beneath the EC began to thaw in early June and reached a maximum thawing depth in late October. The duration of the thawing period of foundation soil was much longer than that of the natural ground, especially for the CRE. The reason for this was that the construction of embankment changed the energy balance between the permafrost and the atmospheric environment. The heat absorption and heat gathering effect of the black asphalt pavement was transferred to the ground through the embankment, resulting in increase in the total heat absorption and reduction of the dissipation [
23,
24].
The dynamic change process of the permafrost table under an embankment is an important index affecting the thermal stability of the embankment. As we already know, the settlement of the embankment is mainly caused by the inseparable, cyclical warming and thawing of the underlying permafrost. For the convenience of analysis, 0 °C is approximately taken as the thawing point of the permafrost. The mean annual thawing rates of permafrost under the four ventilation embankments from 2016 to 2022 are shown in Fig.7. It can be seen that the thawing rates of the CRE, CR + HIP, VS + CRE, VDE were 17.52, 15.90, 7.26, and 5.46 cm/a, respectively. That is, the thawing rates of the four embankments was CRE > CR + HIP > VS + CRE > VDE. The thawing rate was closely related to the ground temperature of the underlying permafrost. The higher the ground temperature was, the higher the thawing rates of the underlying permafrost. Fig.7 shows that the ground temperature of the CRE was the highest, while that of the VDE was the lowest. At the same time, monitoring data demonstrated that the cooling effect of the VDE and VS + CRE was remarkable, and the thawing rate was 50% less than that of the CRE. They played an important role in reducing the ground temperature and protecting the stability of embankment.
3.2 Shady–sunny slope effect of the ventilation embankments
The sunny slope of an embankment receives more solar radiation, while the shady slope receives greater wind speed, which makes the temperature fields of the shady–sunny slopes for embankments different over time. As a result, the two slopes have different thawing depths at the same time, resulting in uneven settlement of the embankment and even longitudinal cracks of the pavement in severe cases. To reveal the transverse geothermal state of the four ventilation embankments, Fig.8 shows the ground temperature at corresponding measuring points on the sunny and shady slopes for different embankments in the cold and warm seasons. As can be seen from Fig.8 that, on the one hand, the ground temperature on the sunny slope of the four ventilation embankments was always higher than that on the shady slope, and the temperature difference in the warm season was greater than that in the cold season. That is, the shady–sunny slope effect was more obvious in warm season. On the other hand, the temperature difference between the shady and sunny slope of the VS + CRE and VDE was small, and the shady–sunny slope effect was weaker. This shows that desirable forced convection heat transfer can effectively weaken the shady–sunny slope effect. In addition, it was found that the transverse temperature differences between the shady and sunny slopes were affected by solar radiation, slope direction, surface turbulence and slope boundary conditions [
25].
In combination with the analysis of thawing depth and ground temperature of the shady–sunny slopes of the embankment shown in Fig.5, the data shown in Fig.8 indicates that after the asphalt pavement was constructed, the warming rate of the permafrost on the sunny slope was faster than that on the shady slope. In Fig.8, The range of the −0.25 °C isotherm and −0.5 °C isotherm at the sunny slope side was larger, and the thawing depth of the foundation under the sunny slope was much greater than that under the shady slope. The differences of ground temperature and thawing depth between the shady and the sunny slopes increased over time. However, the shady slope, also the windward slope, was favorable for the heat in embankment to dissipate through forced convection heat exchange [
7]. Especially for the VS + CRE and VDE, the favorable ventilation and heat exchange effect made the permafrost table on the shady side change only slightly. Meanwhile, the monitoring results show that the thawing depth under slope shoulders was greater than that under slope toes, which was caused by the heat gathering effect of an embankment mainly occurring in the middle of the road. The thawing, deformation and even failure of the underlying permafrost in the embankment were closely related to each other [
26,
27]. For embankments with the significant shady–sunny slope effect, such as the CRE and CR + HIP, differential temperature distribution can result in differential settlement, which can aggravate the occurrence of embankment distresses and should be paid great attention to in design.
3.3 Influence of pre-existing embankment on the geothermal regime
At the test and demonstration section, the pre-existing embankment that was built three years earlier and was about 15 m away from the test embankment could exert thermal impact on the test embankment. This transverse thermal influence can also be illustrated by the 5 years’ observational data.
The deformation mechanism of the embankment in warm permafrost is related to the change characteristics of the APT and the transverse thermal regime. Fig.9 presents the maximum thaw depths and ground temperature under left shoulders at different time. It can be seen from Fig.9 that the APT under left shoulder (sunny side) of the CRE descended significantly, while the APT of the VS + CRE and VDE was lower at the left shoulder than that in the center (as shown in Fig.10), indicating that the CRE was greatly affected by solar radiation on the sunny slope. At the depth of 2 m, the temperature of the VS + CRE and the VDE was close to that at embankment center. In addition, the VS + CRE and the VDE can still decrease the ground temperature and raise the permafrost table (as shown by the green curve). In the range of 2 to 6 m, the temperature of the two embankments increases greatly, which shows that the ground temperature curve moves to the right, while the temperature of the CRE always increases at the 6 m depth range. The reason for the above phenomenon is that the favorable cooling effect in the VS + CRE and the VDE regulates the foundation within 2 m depth and weakens the thermal effect of the sunny slope. The geothermal gradients of the CRE, the VDE and the VS + CRE at a depth of 2 to 6 m are respectively −0.08, −0.10, and −0.05 °C/m, indicating that the VS + CRE is least affected by the pre-existing lines.
However, the ground temperature under the left shoulder of the embankment at a depth of 2 to 6 m was higher than the corresponding ground temperature in the center of the embankment. which was caused by the thermal influence of the pre-existing embankment on the left side. The heat absorption and heat storage characteristics of wide embankment not only made heat transfer vertically, but also caused it to spread transversely over time, which accelerated the increase of adjacent ground temperature. The higher the ground temperature is, the more significant the influence will be. The near-surface frozen soil was regulated by climate environment and embankment and had little effect on ground temperature. It was further indicated that the transverse thermal effect between embankments cannot be ignored and the reasonable transverse distance between embankments should be selected according to the geological condition and embankment type in practical engineering.
Fig.11 shows annual warming rate of permafrost beneath the sunny and shady shoulders at different depth for four ventilation embankments. It was found that warming rates of the permafrost under sunny slope for four ventilation embankments were always higher than those under the shady side, and the warming rates of underlying permafrost of the CRE and the CR + HIP gradually decreased with the increase of depth. Nevertheless, the warming rate of VS + CRE and VDE (as shown in Fig.11(c) and Fig.11(d)) increased first within the depth range of 2 to 6 m, and decreased below 6 m. The reason was that the pre-existing embankment on the left caused thermal disturbance to the underlying permafrost at a depth of 2 to 6 m, which accelerated the temperature rise of permafrost. In addition, the warming rate of permafrost below 6 m began to decrease, which was less and less influenced by the gathering effect, the shady–sunny slope effect, and the proximity of the pre-existing roadway embankment. Annual warming rate of the four ventilation embankments was CRE > CR + HIP > VDE > VS + CRE. In brief, the transverse thermal effects between embankments should be considered in practical engineering.
4 Heat flux and heat budget for four embankments
The characteristics and variations of ground temperature under different embankments in permafrost regions are closely related to vertical heat flux. According to heat transfer theory [
28,
29] and thawing depth of embankment, the depth range of 1.0–1.5 m below natural ground was selected as the computational domain of heat flux density (the depth near permafrost table). The heat flux density within this depth can be expressed as follows:
where
q refers to the heat flux density of soil layers (W/m
2);
Ti and
Tj are the ground temperatures at depths of 1.5 and 1.0 m below the natural ground;
λi,j refer to the heat conductivity coefficient of permafrost at the corresponding depth. The heat conductivity coefficients of the foundation soil for the CRE, the VDE, and the VS + CRE are 2.71, 1.68, and 1.88 W/(m∙K), respectively, in the frozen state, and are 1.76, 1.06, and 1.04 W/(m∙K), respectively, in thawed state [
30]. Δ
z refers to the soil thickness of 0.5 m.
Fig.12 shows the heat flux of soil layers under different embankment centers for one freeze–thaw cycle in 2018. Positive values of heat flux in soil indicate heat absorption, while negative values indicate heat release. The ground under the four ventilation embankments had the characteristics of heat absorption; for the CRE and the CR + HIP these were dominated by heat absorption, and the underlying permafrost inevitably rose. However, the VS + CRE and the VDE alternately absorbed and released heat, demonstrating that the embankments had smooth convection breathing.
The total heat at the depth of 1.0–1.5 m below the natural surface during 2018 can be calculated as follows:
The heat budget of the four ventilation embankments was calculated by Eq. (2), where t0 and t1 are the beginning and end time of heat flux density calculation, respectively. The algebraic sum of envelope area for heat flux curve and x-axis in Fig.12 was the heat flux value of the embankment, and their heat flux values were: 710.94 MJ/m2 for CRE, 517.07 MJ/m2 for CR + HIP, 272.84 MJ/m2 for VS + CRE, and 294.43 MJ/m2 for VDE (as shown in Fig.13). The relative heat budgets of four embankments followed the pattern CRE > CR + HIP > VDE > VS + CRE. The results indicated that VS + CRE had the lowest heat absorption capacity, while CRE absorbed the most heat. Both VS + CRE and VDE had smaller heat absorption capacity compared with CRE. Because of their favorable ventilation convection capacity, underground heat was taken away, ensuring that minimal heat was stored within the underlying permafrost.
5 Discussion
5.1 Variation characteristics of underlying permafrost temperature under the four ventilation embankments
After the embankment construction, the water-heat balance of the natural ground was broken, and the ground temperature changed year by year. It can be seen from the ground temperature variations at the embankment center in Fig.6(a) and Fig.6(c) that the APT rose and the temperature decreased compared to the original state from 2015 to 2016. The reason is that during the initial stage of embankment construction, the favorable thermal insulation effect of embankment fill postponed the transfer of heat absorbed by asphalt pavement to the underlying permafrost. Although the heat carried by embankment fill is greatly dominant at the early construction stage, its thaw effect is taken as limited. After that, the accumulated heat resulting from the long-time heat gathering effect of embankment gradually transferred downward to the underlying permafrost over time, causing the permafrost to continuously warm and degrade. The APT of CRE continuously decreased year by year, while the APT increased and ground temperature decreased again in 2020 for the VS + CRE and VDE.
Solar radiation and air convection are the two important factors affecting the temperature field of a permafrost embankment. Combining Fig.5, Fig.6, and Fig.8, it can be seen that solar radiation plays the main role in heating the embankment during the warm season. The analysis of the abnormal ground temperature in 2020 in Fig.6(b) shows that the ground temperature decline began in mid-November 2019 and ended in early June 2020. During this period, the plateau was in the cold season, the air temperature was below 0 °C, and the solar radiation was very weak. Therefore, the abnormal ground temperature of the VS + CRE and VDE in 2020, discovered in this study, was caused by the enhancement of convective heat transfer, rather than the weakening of solar radiation. In addition, an analysis of the −0.75 °C isotherm in Fig.5(a) and Fig.5(b) shows that warm permafrost had high thermal sensitivity and was vulnerable to impact by the external climate and environment.
5.2 Transverse thermal analysis between pre-existing embankments in permafrost regions
Through the analysis of ground temperature under the four ventilation embankments in Fig.9 and Fig.10, it can be seen that the transverse thermal influence of adjacent pre-existing embankments was mainly observed at 2 to 6 m below the ground depth, and the temperature was controlled by the embankment itself at the depth of 0 to 2 m. On the whole, the long-time heat absorption and heat gathering effect of the asphalt pavement led to significant permafrost degradation and continuous decline of APT, which decreased the cooling effect of the underlying permafrost layers. It was noteworthy that increasing the ventilation convection intensity of the VS + CRE and the VDE in winter could effectively take away the heat accumulated in the embankment, thus cooling the ground temperature obviously at depths up to 7 m depth and thereby raising the APT (refer to Fig.6, Fig.10, and Fig.12).
Therefore, When the ground temperature of adjacent embankments is warm and the distance is close, the VDE and the VS + CRE should be the preferred embankment types. That can ensure smooth breathing of embankments, prevent heat transfer downward, reduce the transverse thermal difference of embankments and maintain stability.
6 Conclusions
Through long-time monitoring of four ventilation embankments at the GYE of the QTP and analysis of their temperature fields, the main conclusions are drawn as follows.
1) Under the heat absorption and accumulation effect of the 4 embankments with black pavement and wide embankment, the temperature of underlying permafrost is warming up and degrading year by year, and the permafrost table is moving downward. Of the different embankment types, the CRE showed the fastest temperature rise and the largest thawing rate, while the thawing rate of VS + CRE and the VDE was slow. Their thawing rates were 17.52 cm/a for CRE, 15.9 cm/a for CR + HIP, 7.26 cm/a for VS + CRE, and 5.46 cm/a for VDE, indicating that VS + CRE and VDE had the most significant convective cooling effect.
2) The solar radiation and wind direction of the four embankments were different on the left and right shoulders, resulting in the shady and sunny effect. The temperature of sunny side for embankment was always higher than that of shady side, and this effect was most obvious for the CRE embankment. The transverse asymmetric distribution of temperature can aggravate the differential settlement of the embankment. However, the transverse temperature difference between VS + CRE and VDE was small, and favorable ventilation heat transfer weakened the shady–sunny slope effect.
3) The pre-existing GYE roadway caused thermal disturbance to the adjacent test and demonstration embankment, and this cannot be ignored. The monitoring data for the scientific and technological demonstration road showed that ground temperature at 2 m depth under natural ground of embankment was mainly regulated by the embankment body, and the temperature at the depth of 2 to 6 m was significantly affected by the thermal interaction with the adjacent embankment; this accelerated the degradation of permafrost and requires attention in practical engineering. Under the influence of climate environment, heat gathering effect, shady and sunny slope effect and pre-existing embankments, the warming rate of underlying permafrost for four types of embankment follow the pattern the VS + CRE > VDE > CR + HIP > CRE.
4) It was found from heat budget analysis that, for the CR + HIP and the CRE, the heat absorption was far larger than the heat release, resulting in a large net heat absorption and the underlying permafrost warmed up inevitably. Comparatively, the heat release for the VS + CRE and VED were larger and the net heat absorption were consequently smaller, ensuring minimal heat storage in the embankment and inhibiting warming up within the underlying permafrost.
PDF
(14143KB)
