The Influence of Vegetation Cover on the Settlement Behavior of Permafrost Subgrade in the Greater Khingan Mountains Forest Region

Abstract

The subgrade construction in the permafrost forest region will aggravate the degradation of frozen soil, which will lead to the settlement of the subgrade. Based on the road project of National Highway 332 in the Great Khingan Mountains, by means of field observation, experimental testing, and numerical simulation, a thermo-hydro-mechanical coupling numerical model of the permafrost subgrade considering vegetation cover was established to analyze the influence of vegetation cover on the settlement behavior of the permafrost subgrade. The study indicates that vegetation cover mainly influences the seasonal active layer temperature of permafrost, and its cooling effect on permafrost in the warm season is more significant compared with the warming effect in the cold season. The volumetric water content of the subgrade with vegetation cover is greater than that without vegetation cover in the cold season. The situation is just the opposite in the warm season. The damage to the subgrade is mainly reflected in the settlement caused by the thawing of frozen soil. The maximum settlement of the subgrade with and without vegetation cover is 8.3 mm and 9.5 mm at the foot of the subgrade slope. After construction, the settlement behavior of the permafrost subgrade will undergo a degradation period of 3 years, a restoration period of 2 years, and finally, tend to be stable.

1. Introduction

Permafrost is defined as a rock and soil mass containing various substances that exist continuously for two years or more at or below 0 °C [1]. China is the third largest frozen soil country in the world. The permafrost area is about 2.15 × 10⁶ square kilometers, accounting for 22.4% of China’s total land area [2,3]. Permafrost is mainly distributed in the Qinghai Tibet Plateau at high altitude and the Greater and Lesser Khingan Mountains at high latitudes [4,5]. As one of the most important forestry bases in China, the Great Khingan Mountains are covered by a large area of virgin forests with rich vegetation types [6]. Therefore, the Great Khingan Mountains permafrost belongs to the high latitude forest region permafrost, with a total area of about 0.39 × 10⁶ square kilometers [7,8].

Permafrost is characterized by poor thermal stability, sensitivity to climate warming, and strong hydrothermal activity [9,10]. In the context of global warming, permafrost is gradually deteriorating [11], which will lead to the rise of underground temperature, the increase of seasonal active layer thickness, and the increase of greenhouse gas emissions [12,13]. Moreover, the degradation of permafrost will also lead to ground settlement and deformation, which will seriously affect the stability of engineering structures [14].

In the forest region of the Great Khingan Mountains, the cold region ecosystem formed by the interaction of permafrost and vegetation is more vulnerable to damage due to environmental changes [15,16]. Vegetation cover will change the energy exchange process in the soil layer and directly affect the hydrothermal state of permafrost. Guo et al. [17] proposed that vegetation can effectively slow down the degradation of permafrost caused by climate warming by regulating the surface freeze–thaw cycles of permafrost. In order to reflect the influence of vegetation on surface heat transfer more accurately, Liu et al. [18] proposed an improved productivity simulator for the northern ecosystem and further studied the role of vegetation in the process of soil surface freezing and thawing. Jia et al. [19] found that revegetation can effectively improve the permafrost stability of degraded grassland in the Qinghai–Tibetan Plateau and enhance the service functions of alpine grassland ecosystems. Szymański et al. [20] studied the impact of tundra vegetation type on topsoil temperature in central Spitsbergen. Zhao et al. [21] studied the influence of freeze–thaw erosion on the stability of steep slopes with high ice content in the Great Khingan Mountains. Yan et al. [22] studied the temperature load mode of a railway bridge in the permafrost region of the Qinghai–Tibet railway. Permyakov et al. [23] developed an algorithm for numerical prediction of frost heaving and thaw settlement of a railway embankment on permafrost in Central Yakutia. However, these engineering activities did not take into account the factors of vegetation cover.

It can be seen that in the Greater Khingan Mountains forest region, scholars have studied the interaction between vegetation and permafrost, mainly focusing on the influence of vegetation on the active layer of permafrost. However, under the condition of vegetation cover, there are relatively few studies on the influence of engineering construction, especially road engineering construction, on permafrost.

Based on the road project of National Highway 332 in the Great Khingan Mountains permafrost forest region, this study analyzed the influence of vegetation cover on the settlement behavior of the permafrost subgrade by means of field observation, experimental testing, and numerical simulation. The purpose of this study is (1) to capture the temporal and spatial dynamic changes of temperature, moisture, and settlement of permafrost subgrade in the forest region; (2) to establish a thermo-hydro-mechanical coupling numerical model of permafrost subgrade considering vegetation cover; and (3) to calculate multiple freeze–thaw cycles to study the influence of vegetation cover on the settlement behavior of permafrost subgrade.

2. Materials and Methods

2.1. Study Region

As shown in Figure 1a, the road project (red block) is situated in the permafrost region of the Great Khingan Mountains, in northeast China. The longitude of this region is between 118°12′01″ E and 127°10′24″ E, and the latitude is between 46°58′02″ N and 53°13′37″ N [3]. This project is the middle connecting section of National Highway 332 Gagadaki Ali River Section Highway and Gen River Labudalin Section Highway. The starting point is situated in the west of Alihe Town at K0+000, and the ending point is situated in the west of Kubuchun Forest Farm at K116+700. The total length of this highway is 116.7 km. The main line is constructed according to the standard of Class I highway, with a design speed of 80 kilometers per hour and a subgrade width of 12.75 m.

Permafrost sections are mainly between K41+910 and K116+700. This section is covered with dense vegetation. The natural permafrost table is shallow and the volume ice content is high, so it is extremely sensitive to temperature changes. The subgrade construction here very easy to damage the underlying permafrost, which will affect the stability of the permafrost subgrade. In this section, as shown in Figure 1b, the permafrost subgrade section (red dot) at K67+300 is the object of our on-site observation.

The K67+300 subgrade section has a latitude of 50°30′18″ N, a longitude of 122°55′30″ E, and an altitude of 574.45 m, as shown in Figure 2. It is characterized by dramatic changes in temperature, short summer, rapid cooling in autumn, cold winter, groundwater development, and abundant precipitation. According to the field observation data, the annual average temperature is between −2.7 °C and −1 °C. The highest temperature in a year is 30 °C, the lowest temperature is −40 °C, and the annual temperature difference reaches 70 °C. The average annual precipitation is 460 mm, which is mainly concentrated in summer [24]. The freezing period lasts from October to May of the next year. The annual sunshine hours are 2500 to 3100 h, and the sunshine rate is 65 to 71% [7].

According to the engineering geological survey report, the surface layer of the proposed subgrade is covered with 0.5 m thick clay, and the underlayer is gravelly soil. The depth of the permafrost is shallow, with the permafrost table being 0.9 m, and the lower limit is 4 m. Permafrost is mainly attached to gravelly soil. The type of frozen soil is mainly ice-rich frozen soil.

2.2. Vegetation

The area is densely forested, and the surface vegetation is developed, as shown in Figure 3. The tree layer is composed of coniferous forests dominated by Larix gmelinii and broad-leaved forests represented by a small number of Betula platyphylla. There are many species in the shrub layer, among which Betula fruticosa, Rhododendron dauricum, Ledum palustre, and Vaccinium uliginosum are the characteristic species of the permafrost vegetation community [25]. There are most species in the herb layer, among which Carex subpediformis is the dominant one. The ground cover is mainly covered by various peat mosses.

2.3. Field Observation Methods

A permafrost subgrade observation section was established at K67+300, as shown in Figure 4. The K67+300 subgrade section was located in the gentle slope section in front of the mountain and the valley plain landform area. It was designed to observe the temperature, moisture content, and settlement of subgrade [26,27]. The temperature sensors, moisture sensors, and settlement observation points were set in this section. The observation period was from October 2019 to December 2020. The subgrade construction period was from October 2019 to May 2020, and the observation period after construction was from June 2020 to December 2020. Considering that the subgrade was still under construction in the early stage of observation, it was not convenient to directly set observation sensors on its surface. Therefore, the observation points were set outside the subgrade slope angle. The subgrade settlement observation started after the construction. To reflect the influence of vegetation cover more intuitively, the subgrade observation section was divided into left and right halves. The left half was covered with vegetation, while the right half was not.

In order to obtain the temperature data of the subgrade section, two temperature observation boreholes denoted as T1 and T2 were set up 1 m outside the slope toe on both sides of the subgrade. Each borehole was 15 m deep, and 15 PT100 thermocouple temperature sensors were set in it. The measuring range of this instrument was between −50 °C and 180 °C, which was suitable for use in permafrost regions. After being bundled, the temperature sensors were fixed on the steel pipes and put into the boreholes. The length of the temperature measurement steel pipe was 15 m. Cement was used to seal the bottom. Periodically, temperature reading instruments were used to collect temperature data for the subgrade section.

In order to obtain the moisture data of the subgrade section, two 2 m deep moisture observation profiles denoted as W1 and W2 were set up 2 m outside the slope toe on the left and right sides of the subgrade. A solid moisture sensor called RS458 was set every 0.5 m from the surface to 2 m underground. This sensor can transmit data to computer software through a USB connection to realize real-time monitoring. Then, the soil was backfilled and compacted. Periodically, moisture data from the subgrade section were collected.

In order to obtain the settlement data of the subgrade section, two settlement observation marks were fixed in the boreholes T1 and T2 to act as base marks. They were protected by two iron buckets to collect data stably for a long time. Settlement observation points were selected at the center of the subgrade, slope toe on the left and right sides of the subgrade, and 4 m outside each slope toe. A total of 5 points were set from left to right, numbered 1#–5#. A digital level called DiNi03 was used to observe the height difference between the base marks and the settlement observation points and calculate the settlement of the subgrade by observing the change in the height difference.

2.4. Establishment of Permafrost Subgrade Model in ABAQUS

As the subgrade is a strip structure and can be considered as an infinite extension along the longitudinal direction, the influence of the spacing effect can be ignored, and a two-dimensional model can be used for simulation. This model consists of a natural foundation and embankment. Considering the influence of subgrade construction on underground frozen soil, the width of the natural foundation is 50 m, and the depth is 15 m. The width of the embankment is 12.75 m, the height is 3 m, and the slope is 1:1.5. The subgrade filling materials are mainly sand gravel, and rubble. The sand gravel layer is 2.5 m thick, and the rubble layer is 0.5 m thick. From 0 m to 0.5 m below the ground is silty clay, and from 0.5 m to 15 m is gravelly soil.

ABAQUS was used to establish the finite element model of the permafrost subgrade in the forest region. The grid was divided into 3342 units in total, and the size of each unit was 0.5 m. The boundary conditions of the model include temperature boundary, moisture boundary, and displacement boundary. For temperature boundary conditions, the upper boundary of the model adopted the average land surface temperature measured in the field during the observation period, including two parts with and without vegetation cover, as shown in Figure 5. The left and right boundaries were set as thermal insulation. According to the engineering geological survey report, the temperature below 10 m underground is constant at −1 °C in this area, so the temperature of the lower boundary of the model was set to −1 °C. For the water boundary condition, it was assumed that the two sides and the bottom of the model were impermeable boundaries, and the upper boundary was an open boundary. The influence of groundwater recharge on the water separation field within the calculation scope was not considered. For displacement boundary conditions, the displacement in the x direction was fixed on both sides of the model, the displacement in the x and y directions was fixed on the lower boundary, and the upper boundary was free.

The material parameters in the model were obtained through field and laboratory tests [28,29], as shown in

Table 1. Physical parameters in the permafrost subgrade model.

Material	${\mathit{\rho }}_{\mathit{d}}$ (kg/m3)	${\mathit{\theta }}_{\mathit{u}}$ (%)	$\mathit{k}$ (m/s)	$\mathit{\mu }$	${\mathit{E}}_{\mathit{u}}$ (Mpa)	${\mathit{E}}_{\mathit{f}}$ (Mpa)
Gravel	2010	17	7.2 × 10−4	0.29	2.35	90
Rubble	2400	10	1.1 × 10−2	0.30	2.65	130
Clay	1700	23	1.5 × 10−8	0.23	2.05	58
Gravel soil	1950	15	6.9 × 10−3	0.25	2.20	70

Note: ${\rho }_d$ = density; ${\theta }_u$ = volumetric water content; $k$ = hydraulic conductivity; $\mu$ = Poisson’s ratio; $E_u$ = elastic modulus of unfrozen soil; $E_f$ = elastic modulus of frozen soil.

and

Table 2. Thermal parameters in the permafrost subgrade model.

Material	${\mathit{\lambda }}_{\mathit{u}}$ (W/(m·°C))	${\mathit{\lambda }}_{\mathit{f}}$ (W/(m·°C))	${\mathit{C}}_{\mathit{u}}$ (J/(kg·°C))	${\mathit{C}}_{\mathit{f}}$ (J/(kg·°C))
Gravel	2010	17	7.2 × 10−4	2.35
Rubble	2400	10	1.1 × 10−2	2.65
Clay	1700	23	1.5 × 10−8	2.05
Gravel soil	1950	15	6.9 × 10−3	2.20

Note: ${\lambda }_u$ = thermal conductivity of unfrozen soil; ${\lambda }_f$ = thermal conductivity of frozen soil; $C_u$ = specific heat of unfrozen soil; $C_f$ = specific heat of frozen soil.

. The materials consisted of gravel, rubble, clay, and gravel soil. A linear elastic model was adopted for subgrade filling materials and underlying rock and soil mass. To simplify the calculation, the thermal conductivity, specific heat capacity, and elastic modulus only considered the freeze–thaw state in this simulation process.

3. Results

3.1. Temperature Variation Regularity of K67+300 Subgrade Section

After installation and debugging, the temperature sensor could accurately measure the temperature of underground rock and soil mass. Temperature data were obtained from October 2019 to December 2020.

Figure 6a,b show the variation trend of the underground temperature of the left half of the subgrade (with vegetation cover) and the right half of the subgrade (without vegetation cover) with the depth during the observation period. The following information can be obtained.

At the same depth, no matter whether there was vegetation cover, the underground temperature decreased with the decrease of external temperature in the cold season and increased with the increase of external temperature in the warm season. When the underground depth was below 0.9 m, the annual underground temperature became negative, which was consistent with the permafrost table of 0.9 m in the geological survey report.

When the underground depth was less than 5 m, the underground temperature changed greatly. The buried depth of this layer was shallow, and there was an active layer of frozen soil in the upper part. As a consequence, it was easily affected by temperature, precipitation, vegetation, and other external conditions. The underground temperature was maintained at −0.8 °C below the depth of 5 m. It can be seen that in this depth range, the influence of the external environment was small. The underground temperature was constant.

In the same month, for example, the underground temperature in December 2020 was slightly higher than that in December 2019. In other words, the underground temperature of the subgrade after construction was slightly higher than that before construction. Especially for the subgrade without vegetation cover, the permafrost table after construction was 1.2 m, which was 0.3 m lower than that before construction of 0.9 m. Because the subgrade construction had disturbed the underlying frozen soil, destroyed its original water and heat balance state, and replaced more heat into the ground, resulting in the reduction of the permafrost table.

According to Figure 7, the changing trend of ground temperature at different depths of borehole T1 (with vegetation cover) and borehole T2 (no vegetation cover) in 2020 can be obtained.

When the underground depth was 0.5 m, the underground temperature reached the lowest value in February and the highest value in August. However, with the increase of burial depth, the influence of the external environment on underground rock and soil mass gradually weakened. In addition, the temperature transfer in the rock and soil layers had a hysteresis effect. These factors led to the lowest underground temperature in March and the highest underground temperature in September.

When the underground depth was 0.5 m, the amplitude of the temperature rise and fall in borehole T2 (no vegetation cover) was significantly greater than that in borehole T1 (with vegetation cover). The maximum temperature difference was 3.4 °C. However, with the increase in depth, the temperature difference became smaller and smaller. When the underground depth was 1 m, the temperature of the two boreholes was approximately the same. According to the geological survey report, the seasonal active layer depth of frozen soil was 0.9 m. It can be seen that the surface vegetation mainly influences the ground temperature of the seasonal active layer.

From the perspective of time, the period with a large temperature difference between borehole T1 (with vegetation cover) and borehole T2 (no vegetation cover) was mainly in July, August, September, and October. However, as time went on, the difference gradually decreased. Since November, the temperature of the two boreholes gradually approached one another. It is shown that the influence of surface vegetation on ground temperature was mainly concentrated in summer and autumn, that is, the warm seasons.

The vegetation could reflect a large amount of solar radiation in the warm seasons, reducing its reach to the ground, which could prevent the ground from warming too fast. In winter, that is, the cold season, vegetation could increase the underground temperature. However, due to the more significant impact of snow cover, the regularity of the impact of vegetation on temperature was not obvious in the Greater Khingan Mountains forest region. Snow cover had the characteristics of high reflectivity, high latent heat of phase change, and low thermal conductivity. In the permafrost region of the Great Khingan Mountains, snowfall activities were frequent. The snow cover period was long, from November to May of the next year. This would undoubtedly have a great impact on the heat exchange between the frozen soil and the atmosphere. On the one hand, the high albedo of snow was conducive to the cooling of the surface and the underlying soil layer. On the other hand, the snow had a good thermal insulation effect due to its low thermal conductivity, which could reduce the heat loss in rock and soil in winter. However, in spring, the heat dissipation of rock and soil would be correspondingly reduced and delayed, leading to the rise of soil temperature, and aggravating the melting of permafrost.

In general, for permafrost, vegetation cover has a cooling effect in the warm season and a warming effect in the cold season. Relatively, the cooling effect of vegetation cover on permafrost in the warm season is more significant.

3.2. Moisture Variation Regularity of K67+300 Subgrade Section

After installation and commissioning, the moisture sensors could accurately measure the volumetric water content of unfrozen water in underground rock and soil mass, excluding solid ice. Moisture data were also obtained from October 2019 to December 2020. Figure 8 shows the changing trend of volumetric water content during the observation period, respectively, at different depths within 2 m underground in section W1 (with vegetation cover) and section W2 (no vegetation cover).

With the increase in depth, no matter whether there was vegetation cover, the volumetric water content of underground rock and soil mass increased. It reached the maximum when the underground depth was 1.5 m, and decreased when the underground depth was 2 m.

The volumetric water content of underground rock and soil mass changes with the seasonal change in external temperature. From October 2019, the surface temperature gradually decreased and transferred to the underground. The capillary water in the unfrozen soil migrated to the surface at a lower temperature under the action of matrix suction, which led to the continuous decrease of the water content of underground rock and soil mass. By the middle of February 2020, the volumetric water content of rock and soil mass was the lowest under the continuous influence of negative temperature. After entering the spring thawing period, the ground temperature gradually rose, and the ice and snow gradually melted. Furthermore, the precipitation in this region was mostly concentrated in summer. Permafrost in the forest region was very sensitive to temperature, so the volumetric water content of rock and soil mass gradually increased under the continuous influence of positive temperature. By the middle of October 2020, the volumetric water content of rock and soil mass was the largest.

At the same depth, in the cold season, the value of moisture sensors in profile W1 (with vegetation cover) was larger than that in profile W2 (no vegetation cover). In the warm season, the opposite was true. This phenomenon was more obvious within 1 m underground. It is because the vegetation cover has the functions of heat insulation, wind, snow resistance, and transpiration of moisture, which has a great influence on the water and heat cycling process of rock and soil mass. When the cold season comes, the canopy, shrub, and moss cover can prevent the cold air from entering the rock soil mass, and the root water in the upper part of the soil layer freezes to release a lot of heat energy, delaying the occurrence of freezing. When the temperature rises in the warm season, the vegetation insulation layer also prevents the warm air from entering the rock and soil mass. Snow in the forest and ice melting in the upper part of the soil layer consumes a lot of heat energy, which hinders the melting process. The transpiration of vegetation becomes more intense because of the temperature rise, which makes more water in the soil be absorbed and transported away. As a result, the volume water content of rock and soil mass will decrease. It can be seen that when the active layer of permafrost melts in the warm season, the volume moisture content of the underlying rock and soil mass of the subgrade with vegetation cover is lower than that of the subgrade without vegetation cover, which is more beneficial to the stability of the subgrade.

3.3. Settlement Variation Regularity of K67+300 Subgrade Section

There were 5 settlement observation points, numbered 1#–5#. Observation points 1# and 5# were located 4 m outside the slope toe on the left and right sides of the subgrade, respectively. Observation points 2# and 3# were located at the left and right slope toe of the subgrade, respectively. Observation point 3# was in the center of the subgrade. The subgrade settlement observation could only be started after the construction had been completed. Therefore, the subgrade settlement data were obtained from June to December 2020. The following information can be obtained from Figure 9.

From the center of the subgrade to the toe of the slope, and then to the outside of the subgrade, the subgrade settlement was increasing. It is shown that the closer to the center of the subgrade, the smaller the magnitude of subgrade freeze–thaw deformation.

The subgrade settlement observation started just after the completion of subgrade construction in May 2020. Therefore, the subgrade at this time was at the initial position, and the settlement was regarded as 0 mm. The settlement values in June and the next months were increments relative to the initial position. As the temperature rose, part of the ice in the underground soil melted into water, causing the subgrade to change from previous frost heave to gradual subsidence. So, all charts intersected at the point “0” on August 15. In November, the settlement reached the maximum. The subgrade settlements calculated from the observation data of observation points 1#–5# were 19.3 mm, 8.3 mm, 4.8 mm, 9.5 mm, and 21.1 mm, respectively. It can be seen that the settlement at the center of the subgrade and the slope toe was relatively small, which could ensure the safe operation of the subgrade. The subgrade settlement of the observation points 1# and 2# (with vegetation cover) is less than that of points 4# and 5# (no vegetation cover).

3.4. Numerical Model Validation

Based on the model of permafrost subgrade in the forest region, combined with the multi-field coupling theory, the thermo-hydro-mechanical coupling numerical model of permafrost subgrade with vegetation coverage was established through the secondary development of ABAQUS. By calculating the temperature field, moisture field, stress field, and displacement field of the K67+300 subgrade section, the settlement behavior of the permafrost subgrade with vegetation cover was obtained.

The subgrade settlement numerical results and field observation results of the subgrade center and the left slope toe of the subgrade were selected for comparison to verify the accuracy of the numerical model. The settlement comparison period was from June to December 2020. As shown in Figure 10, the results of numerical simulation and field observation were in good agreement. Therefore, this numerical model was suitable for simulating the settlement behavior of the permafrost subgrade.

4. Discussion

Based on the road project of National Highway 332 in the Great Khingan Mountains, this study captured the temporal and spatial dynamic changes of temperature, moisture, and settlement of permafrost subgrade in the forest region, and established a thermo-hydro-mechanical coupling numerical model of permafrost subgrade with vegetation cover. After the completion of subgrade construction, the subgrade will aggravate the thawing settlement deformation every time it goes through a freeze–thaw cycle. After a long period of accumulation, the safety of road operations will be seriously threatened. Therefore, this numerical model can be used to simulate the change process of the displacement field of the permafrost subgrade in the next 10 years and reveal its settlement behavior.

In winter, the external temperature gradually decreases, and part of the water in the underground rock and soil mass freezes into ice, resulting in gradual frost heave and uplift of the subgrade. In summer, the external temperature gradually rises, and part of the ice melts into water, leading to the gradual transformation of the subgrade into subsidence. It can be seen that the settlement amplitude is greater than the uplift. Therefore, the subgrade deformation is mainly caused by thawing settlement [30].

At the moment of maximum subgrade settlement, that is, in November of each year, the subgrade settlement cloud maps of 3, 5, and 10 years after construction were drawn, as shown in Figure 11. To study the influence of vegetation cover on subgrade settlement, the settlement data of slope toes on both sides were intercepted to draw the settlement curve, as shown in Figure 12. It can be seen from the figures that the thawing settlement at the slope toes will continue to increase within 3 years after the subgrade construction. The trend of subsequent increase gradually slows down. After the fifth year, the subgrade settlement is stable with little change.

In addition, as shown in Figure 12, the settlement of the subgrade with vegetation cover is less than that without vegetation cover, because the temperature of the subgrade with vegetation cover is lower than that without vegetation cover. Similarly, the volumetric water content of subgrade with vegetation cover is also lower than that without vegetation cover. Therefore, the settlement deformation amplitude of the subgrade with vegetation cover is smaller than that of the subgrade without vegetation cover. It can be seen that vegetation cover is conducive to reducing the settlement of frozen soil subgrade in the warm season.

In general, the settlement behavior of permafrost subgrade from construction to operation can be summarized into three periods: the degradation period, recovery period, and stability period. The degradation period occurs because the water and heat balance of permafrost is destroyed by the subgrade construction, which causes the ground temperature to rise, and the upper limit of permafrost to move down, resulting in a relatively rapid increase of subgrade settlement within 3 years after the subgrade construction. As time goes on, the magnitude of the settlement of permafrost will decrease with the weakening of the influence of subgrade construction on permafrost. At this time, the subgrade settlement is in the recovery period. Finally, the hydrothermal transformation of permafrost will reach equilibrium again. Then, the subgrade settlement is in the stability period.

In the process of establishing the numerical model of the permafrost subgrade in the forest region, ABAQUS can calculate some simple coupling problems of the temperature field, moisture field, and stress–strain field. However, it does not support complex and extensive multi-physical field coupling numerical calculation.

5. Conclusions

Based on the road project of National Highway 332 in the Great Khingan Mountains permafrost forest region, a study on the influence of vegetation cover on the settlement behavior of the permafrost subgrade was carried out through field observation, experimental testing, and numerical simulation. The conclusions are as follows:

(1)

In the forest region of the Great Khingan Mountains, the underground temperature reaches its lowest value in March and the highest value in September. Vegetation cover mainly influences the seasonal active layer temperature of permafrost and its cooling effect on permafrost in the warm season is more significant compared with the warming effect in the cold season.

(2)

The volumetric water content of underground rock and soil mass reaches the maximum when the underground depth is 1.5 m. At the same depth, the volumetric water content of the subgrade with vegetation cover is greater than that without vegetation cover in the cold season. The situation is just the opposite in the warm season.

(3)

The closer to the center of the subgrade, the smaller the magnitude of subgrade freeze–thaw deformation. The damage to the subgrade is mainly reflected in the settlement caused by the thawing of frozen soil. The maximum settlement of the subgrade with and without vegetation cover is 8.3 mm and 9.5 mm at the foot of the subgrade slope, respectively. After construction, the settlement behavior of the permafrost subgrade will undergo a degradation period of 3 years, a restoration period of 2 years, and finally, tend to be stable.