The effect of desertification on frozen soil on the Qinghai-Tibet plateau

Journal Pre-proofs The effect of desertification on frozen soil on the Qinghai-Tibet Plateau Luyang Wang, Qingbai Wu, Guanli Jiang PII: DOI: Reference:

S0048-9697(19)34631-5 https://doi.org/10.1016/j.scitotenv.2019.134640 STOTEN 134640

To appear in:

Science of the Total Environment

Received Date: Revised Date: Accepted Date:

8 July 2019 19 August 2019 23 September 2019

Please cite this article as: L. Wang, Q. Wu, G. Jiang, The effect of desertification on frozen soil on the QinghaiTibet Plateau, Science of the Total Environment (2019), doi: https://doi.org/10.1016/j.scitotenv.2019.134640

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Elsevier B.V. All rights reserved.

The effect of desertification on frozen soil on the Qinghai-Tibet Plateau Luyang Wang a, b, Qingbai Wu a, *, Guanli Jiang a a

State Key Laboratory of Frozen soil Engineering, Northwest Institute of Eco-Environment and

Resources, Chinese Academy of Sciences, Lanzhou 730000, China b University

*

of Chinese Academy of Sciences, Beijing 100049, China

Corresponding author at: 320 West Donggang Road, Northwest Institute of Eco-Environment and

Resources, Chinese Academy of Sciences, Lanzhou 730000, Gansu Province, China. E-mail addresses: [email protected] (L. Wang), [email protected] (Q. Wu), [email protected] (G. Jiang).

ABSTRACT

Under the influences of climate change and human activities, desertification has become widespread on the Qinghai-Tibet Plateau (QTP). However, the effect of desertification on frozen soil is still debated. Here, soil temperatures are observed through 14 boreholes at Honglianghe River Basin on the QTP to study the relationship between desertification and frozen soil. The results showed soil temperatures change with the thickness of sand cover. With increasing sand thickness, maximum soil temperatures at shallow depths (0.05-6.00 m) increase by 0.25-1.57 oC, but minimum temperatures decrease by 0.21-1.49 oC, on average. Temperatures at deep depth (≥ 6.00 m) exhibit a rising trend that temperatures increase by 0.01-0.05 oC on average with each increment of 10 cm in sand thickness. Furthermore, aeolian sand enhances seasonal thawing processes, resulting in an increase of 7.70-9.50 cm in active layer thickness with each increment of 10 cm in sand thickness. Meanwhile, aeolian sand weakens seasonal freezing processes, resulting in a decrease of 1.07-13.00 cm in seasonal freezing depth with each increment of 10 cm in sand thickness. Moisture contents of aeolian sand and vegetation coverages on the sand cover surface 1

influence energy state and thermal regime of frozen soil. Annual heat budgets of soil under aeolian sand increase from -57.97 MJ·m-2 to -26.28 MJ·m-2 as water content of sand layer decreases from 13.42 % to 3.61 %. Annual range of ground temperatures of soil at shallow depths (0.05-1.60 m) increase by 2.19-6.17 oC on average as vegetation coverage increases from 5 % to 20 %. Due to the effects of aeolian sand on frozen soil, desertification accelerates, and can even cause, the degradation of frozen soil on the QTP. Our study provides an important reference for future research about the interaction between desertification and frozen soil in other regions.

Keywords Aeolian sand, Permafrost, Seasonally frozen ground, Thermal regime of soil, Qinghai-Tibet Plateau 1. Introduction

Permafrost is a product of long-term interaction between the ground and atmosphere (Williams and Smith, 1989; Qiu and Cheng, 1995; Burn and Nelson, 2006), and its distribution is controlled by many factors (French, 2007). The macroscale distribution of permafrost is dominated by the climate; however, local-scale factors such as terrain or landform, vegetation, snow cover, swamps and water bodies can significantly alter the general characteristics of permafrost distribution, resulting in spatial abnormalities (Cheng, 2004; French, 2007; Pang et al., 2011). The impact of surface cover conditions on permafrost, as an important local-scale factor, has been studied by many researchers (Cheng, 2004; French, 2007; Pang et al., 2011; Luo et al., 2018a, 2018b). Only a few studies (Wang and Zhang, 1985; Wang and Xie, 1998; Wang et al., 2002), however, give adequate attention to the impact of aeolian sand cover. Desertification is widespread in the permafrost region on the Qinghai-Tibet Plateau (QTP) (Li et al., 2001; Li et al., 2

2004; Wang, 2004). The strongly increasing desertification not only destroys the fragile permafrost ecological environment but also seriously threatens routine operations and maintenances of the Qinghai-Tibet Highway (QTH) and the Qinghai-Tibet Railway (QTR) (Dong, 2001; Zhang et al., 2008; Zhang et al.,2010, 2011a, 2011b; Wu et al., 2016). Since the late Pleistocene, the QTP has repeatedly experienced climate fluctuations with different amplitudes and periods, resulting in a climate that alternates between dry-cold and moist-warm (Qiu and Cheng, 1995; Wang et al., 1979). Therefore, the evolution of permafrost and the development of desertification are extremely complex (Yang et al., 2004). According to previous studies (Wang and Zhang, 1985; Wang, 1989; An et al., 1991; Huang et al., 1993; Jin et al., 2006, 2018), the aeolian sand on the QTP can be classified into four categories based on the chronological sequence as follows. (1) During the Last Glacial Maximum (LGM, 21-10.8 ka BP), permafrost began to aggrade in most areas of the QTP due to the cold-dry climate caused by plateau rapid apophysis (Wang, 1989). At that time, the extent of LGM permafrost on the QTP was about three times than that of today (Jin et al., 2018). Meanwhile, wind-blown sands had widely covered on the ground surface in the plateau during this period because of strong wind erosion (Huang et al., 1993). One example is a 3.00-m-thick aeolian sand sediment formed at about 12.7±0.8 ka BP (tested by 14C dating of the humus interlayer at the 1.00-1.10 m depth) in the steep slope of a river gully in northern Wudaoliang (Wang and Zhang, 1985). Currently, this type of sand cover remains widespread on the QTP. However, it is covered by vegetation and new deposits and is thus difficult to distinguish at present. (2) During the early Holocene (10.8 ka BP to 8.5-7.0 ka BP), the climate on the QTP shifted 3

from dry-cold to moist-warm (Jin et al., 2006). In such climate conditions, the southern and northern limits of permafrost on the QTP retreated northwards and southwards, respectively, and the permafrost on the interior of the QTP area remained stable. Although the permafrost degraded in this period, its extent remained 1.4-1.5 times that of today (Jin et al., 2006, 2018). Due to the increases in air temperature and precipitation caused by the warmer and wetter climate, the vegetation effectively stabilized sand dunes that had formed or was forming. Consequently, during this period, sand cover presented mostly as fixed or semi-fixed sand dunes, longitudinal dunes, and deflation hollows. One example is a longitudinal sand dune (formed at about 9716±270a BP from

14C

dating result of the plant residuals) surrounding a lake 2 km east of the Wudaoliang

(Wang, 1989). (3) Since the late Holocene, particularly during the global Little Ice Age 500 years ago, climate on the QTP became dry and cool. As a result, the thawed soil refroze, and the extent of permafrost on the QTP increased to 1.15-1.20 times than that of today (Jin et al., 2006, 2018). During this period, some wind-blown sand was fixed into sand dunes by vegetation, but most remained to extensively cover the plateau ground surface. Meanwhile, the reactivated surfaces of some fixed sand dunes were recovered by drift sand, leading to continued desertification on the QTP (Jin et al., 2018). (4) Since the last century, the extent of permafrost on the QTP has decreased to approximately 1.06-1.59×106 km2 (Jin et al., 2018) due to global warming. During this period, the intensive wind erosion formed numerous mobile sand dunes and forced the process of grassland degradation to some extent (Huang et al., 1993). As a result, the desertification phenomenon was accelerated on the QTP, especially in the areas impacted by human activities and engineering 4

facilities (e.g., Xidatan, Honglianghe, and the east coast of Honghai) (Huang et al., 1993). The wind-blown sand accumulations that covered partial sections of the QTH and the QTR contributed to severe damage to infrastructure. Due to the impacts of the desertification on permafrost and engineering activities on the QTP, studies about this subject have been carried out since the 1990s. Unfortunately, the effects of desertification on permafrost are still rarely investigated, and no significant conclusions have been drawn. The traditional geocryological viewpoint is that aeolian sand can accelerate or even cause the degradation of permafrost because ground temperatures in regions of wind-blown sand in permafrost regions are significantly higher than that in regions without sand cover. In addition, the active layers in the regions covered with sand are also thicker than these in regions lacking in the sand cover (Huang et al., 1993; Pang et al., 2011; Wu et al., 2017; Wang et al., 2018). However, recent studies have concluded that the sand layer may play a role in protecting permafrost because the permafrost temperatures were lower underneath sand cover (Xie and Qu, 2013; Xie et al., 2012, 2013, 2015, 2017a, 2017b). Furthermore, the thicker the sand cover, the stronger the protective effect. Interestingly, other researches have shown that the dual-influence of sand cover on permafrost because of the different thicknesses of sand layer (Wang and Xie, 1998; Wang and Zhao, 1999; Wang et al., 2002; Yang et al., 2004; Lü et al., 2008). These studies have found that the thick sand cover can accelerate the degradation of permafrost, while thin covers can actually protect permafrost. Why these varied and even contradict conclusions? The main reasons are the lack of experimental sites, discontinuous time observations, and insufficient considerations of other factors. Compared with the recent studies (Xie et al., 2017a, 2017b), a ground temperature 5

observation within related large zone (fourteen observation sites) on desert region on the QTP is first carried out in our study. These consecutive temperature data are important to assessing the thermal state and the development trend of frozen soil, and also make our study more effective and reasonable. In previous researches (Huang et al., 1993; Wang and Xie, 1998), soil temperatures were utilized to estimate the thermal regime of frozen soil. However, this would ignore the effect of soil energy state which is the intrinsic factor causing the variation in temperature. Accordingly, in this paper, the balance theory of annual heat budget of soils is creatively used to explain the energy mechanism of frozen soils under aeolian sand, and the results of our study would be more reliable due to this explanation.

2. Materials and methods

2.1 Study area

The study area is located in the Honglianghe River basin on the western side of the QTH (35°02′43″N, 93°00′51″E, average elevation: 4640 m a.s.l.). This area typifies the continuous permafrost region (Wang et al., 1979). According to the data from meteorological station in study region, mean monthly air temperatures (MMAT) in the summer season (JJA) vary from 4.61 °C to 7.80 °C, with an average of 6.21 °C. MMATs in the winter season (DJF) vary from -14.76 °C to -11.84 °C, with an average of -13.30 °C. Maximum and minimum albedo occur in summer and winter, and the mean albedos in summer and winter are 0.42 and 0.24, respectively. The mean annual albedo is 0.30. Precipitation is mainly concentrated in summer, and the total precipitation for the year is approximately 253 mm (Wu et al., 2017). In the Honglianghe River basin, modern aeolian sand activities are quite extensive, and the

6

yearly total sand transport quantity from all wind directions is about 434.33 kg·m-1 (Xie et al., 2018). As a result, thick sand covers and longitudinal dunes are widespread. In addition to the modern sand sediments, fixed or semi-fixed sand dunes also cover large areas near the Honglianghe River. Nonetheless, most of these dunes are covered by sparse vegetation, making the desert landscape difficult to distinguish. In addition, some sand sediments at relatively deep depths exhibited obvious differences in states with the top sand layers. Because of repeated freeze-thaw action and cryogenic weathering during the glacial-interglacial cycles 500 years ago, these deep sand layers, which comprise primary silt, clay particles, and plant residuals, present as sandy clay or clayey sand. Moreover, the water contents of these sand layers are generally high with slight seasonal changes. Consequently, such sand sediments might be a part of the permafrost table (Wang, 1989). In addition to the aeolian sand, most areas near the Honglianghe River are sparsely covered by dwarf alpine vegetation. The vegetation coverages of those sites were obtained by typical vegetation sample investigation method. The quadrat of 100 m × 100 m was set up in each site. Based on the photos taken by an UAV, the proportion of the projected area of the vegetation canopy to the total quadrat was calculated in order to obtain the vegetation coverage (Zhao and Sheng, 2015). The sections covered by vegetation are a typical alpine desert vegetation landscape (Fig. 1a), and the sections without vegetation are desert landscapes (Fig. 1b). Due to the widespread frozen soil and the intensive aeolian sand activities, Honglianghe River basin is an ideal locality for studying the effects of desertification on frozen soil.

2.2 Site design and data collection

Fourteen boreholes were drilled to measure soil temperature in the Honglianghe River basin 7

(Fig. 2). Soil temperatures at HLH-01-S through HLH-12-P sites were recorded from 23, August 2017 to 23, August 2018. and the soil temperatures at QH4 and QH5 sites were from 28, November 2013 to 23, January 2016. Soil temperatures were measured at depths from 0.05 to 18.00 m using a thermistor probe (assembled by the State Key Laboratory of Frozen Soil Engineering at Lanzhou, China). The temperature accuracy of the sensors is ±0.05 °C. The data were collected every four hours by an automatic data logger (CR3000, Campbell Scientific Inc., USA) and transmitted to a terminal by a remote transfer component. 2.3 Study method In this paper, to remove the influence of vegetation, the variations in soil temperature (max and min temperatures, mean monthly temperatures and annual average temperatures) were studied among sites with same vegetation coverage and frozen soil condition but different sand thickness. However, the deep temperatures (such as MAGT) were affected less by vegetation. In order to obtain the trend of MAGT with increasing in thickness of sand cover, statistical analyses were performed with a unary linear regression analysis by RStudio software, and the statistical formulas are as follows: (1)

𝑀𝐴𝐺𝑇 = 𝑎ℎ + 𝑏 𝑎=

𝑏=

∑𝑀𝐴𝐺𝑇𝑖

―𝑏

∑ℎ𝑖

𝑛 𝑛 𝑛∑ℎ𝑖𝑀𝐴𝐺𝑇𝑖 ― ∑ℎ𝑖∑𝑀𝐴𝐺𝑇𝑖 𝑛

where MAGT is the mean annual ground temperature (°C), h is the thickness of sand cover (cm), a and b are the regression parameters, n is the number of data. Then, the coefficient of determination (R2) was used to assess the accuracy of the regression analysis. Similar to the analysis of soil temperatures, statistical comparisons of thawing and freezing 8

process among sites were carried out. Based on the measured daily soil temperatures, the geothermal isolines of sites were plotted by RStudio software. The time that 0 °C isoline began to move downward from the ground surface was used as the onset of seasonal thawing or freezing (Kudriavtsev, 1992). The maximum depth of 0°C isoline represented ALT or SFT, and the time when 0 °C isoline reached the maximum depth was used as the time of thawing or freezing maximum depth (Kudriavtsev, 1992; French, 2007). The interval between onset of thawing or freezing and time of thawing or freezing maximum depth was the duration of seasonal thawing or freezing. Furthermore, due to the importance of water content within sand cover to the heat budget of frozen soil, according to the Fourier Law and the formula (Feldman, 1982; Wu et al., 2017) as followed, we discussed the annual heat budgets of soils at 0.05-0.60 m depths (the water contents of soil are obtained from 0.05-0.60 m depths) among sites with same sand thickness and vegetation coverage but different moisture content. 𝑇2 ― 𝑇1

𝜏

𝑄 = ∫𝜏1 ―𝜆

Δ𝑍

2

𝑑𝜏

(2)

where λ is the thermal conductivity of soil (W·m-1·k-1), T2 and T1 are the temperatures of the upper and lower surfaces of the soil layer, respectively (°C), Δ Z is the thickness of the soil layer (55 cm). τ1 and τ2 are the times of direction changes of heat flux on the ground surface. The influence of other factors (such as precipitation process) on heat budget have all included in measured soil temperatures. Finally, the effect of vegetation on ground temperatures at study region was studied between sites with same sand thickness but different vegetation coverage.

3. Results 9

3.1 Frozen soil conditions

According to our results (Table 1), the permafrost has occurred only at four sites. Moreover, the permafrost presents a warm thermal state (Wang et al., 1979), with MAGTs at 15.00 m ranging from -0.42 °C to -0.96 °C. Correspondingly, the ALTs range from 1.90 m to 4.10 m. The other nine sites have seasonal frozen soil whose MAGTs range from 0.15 ° C to 1.04 ° C and whose the SFTs range from 2.94 m to 6.50 m.

3.2 Ground temperature

To remove the influence of vegetation, comparisons in soil temperatures and seasonal freeze-thaw processes among these boreholes must focus on the sites that have the same vegetation coverage and similar frozen soil conditions, such as permafrost sites without vegetation cover (HLH-12-P and QH5-P sites) and seasonally frozen ground sites with 5 % vegetation coverage (HLH-04-S and HLH-05-S sites). Fig. 3 shows the minimum and maximum ground temperatures at shallow depths (0.05-6.00 m) for every group of sites. The results reveal that the shallow-layer maximum ground temperatures under thicker sand cover are higher than those under the thinner one. In contrast, the minimum ground temperatures underlying the thicker sand cover are lower than those of ground covered by thinner sand at corresponding depths. For instance, at depths between 0.05 m and 4.00 m, maximum ground temperatures at QH5-P (a 115-cm- thick sand cover) are approximately 1.79 ° C, higher than those at HLH-12-P (a 10-cm-thick sand cover) during the warm seasons; however, the minimum ground temperatures at QH5-P are about 0.47 ° C, cooler than those at HLH-12-P during the cold season (Fig. 3a). Similar patterns of variations in ground temperature

10

have also been found at other site groups (Fig. 3b-d). The data demonstrate that ground temperatures vary seasonally at shallow depths with the changes in sand cover thickness. It is evident that the soil temperatures increase during the warm seasons and decrease during the cold seasons as the thicknesses of sand cover increase. The maximum soil temperatures at shallow depth increase with an increment of 0.25 oC to 1.57 oC on average, and the minimum soil temperatures decrease with a diminution of 0.21 oC to 1.49 oC on average. Consequently, the annual range of ground temperatures (ARGT) increase with increasing sand cover. Therefore, the ground temperatures significantly change due to sand cover on the ground surface. In comparison with the variations of soil temperature in a shallow layer, deep ground temperatures barely vary regardless of the season. Fig. 4 shows that deep ground temperatures vary over time. The ground temperatures under thicker sand cover are generally higher than those under thinner sand cover. For example, the temperatures at a depth of 6.00 m at HLH-07-P (with a 40-cm-thick sand cover) are approximately 0.10 ° C higher than those at the same depth at HLH-08-P (with a 20-cm-thick sand cover) (Fig. 4a). In addition, at a depth of 6.00 m, the temperatures at HLH-04-S (with a 180-cm-thick sand cover) are approximately 0.86 ° C higher than those at HLH-05-S (with a 60-cm-thick sand cover), on average (Fig. 4c). This pattern is similar to other locations, as Fig. 4b and Fig. 4d shows. These data show that deep ground temperatures rise with increasing sand layer thickness, and deep temperatures increase 0.01-0.05 oC

on average with each increment of 10 cm in sand thickness. In particular, the soil temperatures

at some sites in Fig. 4c-4d during October 2017 to Mar 2018 are higher than that during other months. Moreover, the trend of MAGTs at study sites (Fig. 5) also illustrates that the ground temperatures increase as the thicknesses of aeolian sand cover increase. The MAGTs increase 11

0.04 °C on average with each increment of 10 cm in sand thickness. Fig. 6 shows average ground temperatures at all depths. Generally, aeolian sand cover causes an obvious rise in soil temperatures with increasing sand layer thickness at each group of sites. For example, the ground temperatures at depths of 5 cm to 18 m at QH5-P (a 115-cm-thick sand cover) are all higher than those measured at HLH-12-P (a 10-cm-thick sand cover) (Fig. 6a), and the difference in temperatures between these two sites peaks at a maximum value of 0.65 °C at 1.00 m depth. At depths of 0.05 m to 0.30 m, soil temperatures at HLH-05-S are slightly lower than those at HLH-04-S (Fig. 6c); however, the temperature drop between these two sites begins to rise when the depth increases from 0.30 m, ultimately peaking at 0.76 °C below 2.00 m depth. 3.3 Freezing-thawing process

In addition to causing variations in ground temperature, aeolian sand also influence the soil freeze-thaw process because of the essential connection between thermal regime and seasonal freezing-thawing process. Taking HLH-07-P and HLH-08-P as examples (Fig. 7a and b, Table 2), we can see that the onset of thawing downward from the ground surface at HLH-07-P (May 10, 2018) is almost equal to that at HLH-08-P (May 11, 2018). However, the time of thawing maximum depth at HLH-07-P (October 7, 2018) is earlier than that at HLH-08-P (October 15, 2018). At the same time, the ALT at HLH-07-P (2.09 m) is larger than that at HLH-08-P (1.90 m). These data demonstrate that durations of seasonal thawing shorten and the ALTs increase as aeolian sand cover thickness increases. The ALT increases 7.70-9.50 cm on average with each increment of 10 cm in sand thickness. Therefore, the aeolian sand cover increases heat transfer efficiency downward, enhancing the seasonal thawing of the active layer. The effects of aeolian sand cover on the seasonal freezing process contrast with the effects on 12

the seasonal thawing process. Taking HLH-04-S and HLH-05-S as examples (Fig. 7c and d, Table 2), we can see that the onsets of freezing downward from the ground surface at HLH-04-S and HLH-05-S were the same on October 25, 2017. The time of freezing to the maximum depth at HLH-04-S was April 2, 2018, whereas at HLH-05-S, it was July 8, 2018. Accordingly, the duration of seasonal freezing at HLH-04-S (159 days) is 97 days shorter than at HLH-05-S (256 days). Meanwhile, the SFT at HLH-04-S was 3.01 m, compared with 4.57 m at HLH-05-S. The above analyses show that the thinner the sand cover, the shorter the duration of seasonal freezing, and the thinner the SFT. The SFT decreases 1.07-13.00 cm on average with each increment of 10 cm in sand thickness. In other words, sand cover weakens the freezing process of seasonally frozen ground. The seasonal freezing-thaw cycle is bound up with the states of permafrost and seasonally frozen ground. Consequently, sand cover causes a trend of permafrost degeneration by intensifying the seasonal thawing process of the active layer and degrading the seasonally frozen ground through weakening the seasonal freezing process. 3.4 Annual heat budget The annual heat budgets of soils under aeolian sand cover are obtained by formula (2) and are shown in Table 3. The results illustrate that the annual heat budgets of the soil below the sand cover progressively increase as the sand moisture contents decrease. The annual heat budgets of soils increase from -57.97 MJ·m-2 to -26.28 MJ·m-2 as water content of sand layer decreases from 13.42 % to 3.61 %.

4. Discussion

4.1 The effect of desertification on distribution of frozen soil 13

Aeolian sand cover pronouncedly changes the thermal regime of frozen soils and the freezing-thawing processes of soils. Consequently, the spatial distribution of permafrost largely depends on the spatial distribution of desertification. Generally, the distribution of desertification is controlled by aeolian sand source, wind, terrain, and slope. At Honglianghe River, sand-moving wind usually originates from the north. Therefore, aeolian sandy particles can easily accumulate in low depressions or steep north-facing slope (Ding, 2010; Dong et al., 2012; Yao et al., 2015). For example, on NE-SW trending geological profiles (Fig. 1a in Supplementary material), sandy particles transported by the north wind settled easily in depressions on the northern side of the HLH-01-S site, as well as on the northward steep slope of HLH-01-S. As a result, thick aeolian sand sediments were found in these places, including 300-350-cm-thick aeolian sand accumulations at HLH-02-S and HLH-03-S. In contrast, sandy particles are difficult to deposit on higher terrains such as those found at HLH-07-P and HLH-08-P, which were covered only by 20-40-cm thick sand layers. Interestingly, under such an aeolian sand environment, permafrost is common in areas covered by thinner sand layers, such as HLH-07-P and HLH-08-P; in areas covered by thicker sand layers, no permafrost has been found, and soils experience seasonal freezing (examples include HLH-02-S, HLH-03-S, HLH-04-S, and HLH-05-S sites). A similar relationship between sand layer thickness and frozen soil distribution has also been observed in NW-SE trending geological profiles (Fig. 1b in Supplementary material). Permafrost frequently occurs in areas where the aeolian sand layer is thin. In contrast, seasonally frozen ground is generally present in areas with thick sand cover. These observations are consistent with traditional geocryological knowledge. The relationship between aeolian sand and occurrence of permafrost is consistent with the relationship between sand thickness and 14

ground temperature. The soil temperatures show an increased trend with an increase in thickness of sand cover. The thicker the sand cover, the higher the soil temperature. Accordingly, on the one hand, the permafrost is more unstable and begin to degrade into seasonal frozen soil. On the other hand, the thickness of seasonal frozen soil also decreases. Therefore, aeolian sand cover accelerates the degradation of frozen soil.

4.2 Impact of the thickness of sand cover on thermal regime and freezing-thawing processes in frozen soil

The endothermic and exothermic properties of aeolian sand are directly controlled by the thickness of the sand layers (Wang and Zhao, 1999; Wang et al., 2002). When a sand layer is deposited on the ground surface, the original ground temperature field is disturbed. As a result, the maximum ground temperature at shallow depths increased but the minimum decreased. The reasons behind these variations are based on the special thermophysical properties and high infiltration rates of aeolian sand (Chen et al., 2014; Wu et al., 2017). On the one hand, compared with fine-grained soils such as clays, the specific heat capacity and thermal conductivity of sand cover are small; however, the thermal diffusivity is relatively large (Baver et al., 1983; Wang and Xie, 1998). Therefore, the temperature of the aeolian sand layer can easily change. On the other hand, because the sand layer cannot maintain high water contents, water can easily infiltrate into the ground below the aeolian sand cover, leading to strong convective heat transfer (Wu et al., 2017). During warm seasons with high air temperatures and large amount of precipitation, the aeolian sand layer can absorb radiation and then transfer heat into the shallow layer, rising soil temperatures. In the meanwhile, the convective heat, which is transferred by infiltrating water, can increase the soil temperatures further. During cold seasons, the air temperature is low, and there is 15

little precipitation on the QTP. Accordingly, convective heat transfer is quite weak, and the variations on soil temperature primarily rely on heat conduction. As a result, soil temperatures at shallow depths below sand cover decrease due to the relatively large value of sand thermal diffusivity. However, the extent of the decrease in minimum soil temperatures is less than the extent of increase in maximum temperatures. As a result, the aeolian sand cover mainly plays a role in warming the frozen soil. The result of variations of maximum and minimum soil temperatures at shallow depths is that the ARGTs increase, enhancing the freeze-thaw effect and extending the duration of freezing and thawing, resulting in an acceleration in the weathering rate of the shallow soil layer (Huang et al., 1993). Inevitably, intensified cryogenesis loosens soil structure and increases porosity (Huang et al., 1993; Wang et al., 2002; Yang et al., 2004), meaning that heat flux and water can easily pass through sand cover into the deep ground, where it increases the heat budget of the soil underneath the sand layer. From the above discussion, it is evident that permafrost and seasonally frozen ground will progressively degrade due to the rise of ground temperatures caused by an increasing soil heat budget. The response mechanism to variations of aeolian sand cover thickness is certainly different between seasonal freezing process and thawing process. For a seasonal freezing process, decreases in minimum soil temperature caused by thick sand layers can shift the minimum soil temperature curve to left, leading to an increase in the maximum seasonal freezing depth (MSFD) (Kudriavtsev, 1992; Yershov, 2016). However, the increases in MAGT caused by aeolian sand cover decrease MSFD. Therefore, the MSFD is a result of declining minimum soil temperature and increasing of MAGT (Fig. 8a). According to the above analysis, in desert regions, the effect of the aeolian sand 16

layer on soil temperature reduction is confined to shallow depths. Therefore, the influence of decreasing shallow soil temperature on MSFD have been limited. As a result, at the Honglianghe River basin, the decrease in MSFD caused by a rise in MAGT is larger than the increase of MSFD caused by declining of minimum soil temperature. These two trends cancel out each other, meaning that the MSFD decreases as the thickness of the sand cover increases. For the seasonal thawing process, the rise in the maximum soil temperature caused by the thick sand layer induces a rightward shift in the maximum soil temperature curve, resulting in an increase in the maximum seasonal thawing depth (MSTD) (Kudriavtsev, 1992; Yershov, 2016). Meanwhile, rising MAGT caused by the sand layer also increases MSTD. These two effects significantly increase the MSTD as the thickness of the sand cover increases (Fig. 8b).

4.3 Effect of water content of aeolian sand on heat budgets of soil

Previous studies have indicated that the thermal conductivity of aeolian sand decreases as the water content decreases (Chen et al., 2014; Wu et al., 2017). Because the aeolian sand layer is the heat exchange interface between the ground and the atmosphere, variations in the thermal conductivity of the sand layer can directly influence the annual heat budget equilibrium of the soil under the sand layer, changing the state and development trend of the frozen soil. Therefore, the water content of aeolian sand indirectly affects the frozen soil under sand cover. According to the results of annual heat budgets, the lower the water content of the sand layer, the stronger its heat absorption. Sand layers with low water contents are in a state of absorptive heat over the entire whole year. The strong endothermic effect of dry sand causes the increase in absorption of soils under sand cover, and then is adverse to the development of permafrost. Accordingly, dry sand is worse at protecting permafrost. Because the aeolian sand layer on the QTP usually has a high 17

infiltration rate, it is difficult to retain high water contents under drought conditions in a windy environment. As a result, aeolian sand layers in the QTP usually have low moisture contents, meaning they adversely affect the permafrost.

4.4 Effect of desert vegetation on shallow soil thermal regime

Vegetation on the surface of the aeolian sand layer can affect processes of ground evaporation and radiation, which are closely related to the ground temperature in the shallow layer (Brown, 1973; Smith, 1975; Jin et al., 2008). Therefore, the effect of vegetation on soil thermal regime cannot be ignored. Previous research (Jin et al., 2008; Lü et al., 2008) illustrates that the special climate of the QTP causes notable differences in heat loss and absorption on the ground surfaces covered by vegetation. These differences cause variations in ground temperatures. Generally, in the interior and eastern QTP region, if vegetation coverage is less than 20-30 %, shallow layer soil temperatures increase during warm seasons and decrease during cold seasons with increasing vegetation coverage (Jin et al., 2008). As a result, vegetation increases the values of the AGRTs. Vegetation coverages in the Honglianghe River basin are generally less than 20 %; thus, it can be inferred easily that sparse vegetation plays a role in increasing the AGRTs. For example, at HLH-07-P and HLH-06-S sites (Fig. 9a), at depths from 0.05 m to 1.60 m, the AGRTs that are differences between the maximum and minimum mean daily ground temperatures at HLH-07-P with a low vegetation coverage are obviously less than those at HLH-06-S, which has a relatively high vegetation coverage. Similar variations also have been found at HLH-08-P and HLH-01-S, both of which are covered by a 20-cm-thick sand layer (Fig. 9b). These data indicate that sparse vegetation in the deserts of the QTP can increase AGRT at shallow depths. The AGRT increases 18

with an average increment from 2.19 oC to 6.17 oC with an increase in vegetation coverage. resulting in an increase in deep soil temperatures. Therefore, in the desert of the QTP, sparse vegetation is not conducive to the preservation and development of frozen soil, and even causes a degradation trend of frozen soil.

5. Conclusions

Deposits of aeolian sand originating from desertification on the QTP have significant effects on the frozen soil beneath. Due to the thermal properties of aeolian sand, as the thickness of the sand cover increases, maximum soil temperatures at shallow depths (0.05-6.00 m) increase with an average increment from 0.25 oC to 1.57 oC, but minimum temperatures decrease with an average diminution from 0.21 oC to 1.49 oC. This result in an increase in the annual range of ground temperature variation, resulting in rising soil temperatures at deep depths (≥ 6.00 m). Deep soil temperatures increase 0.01-0.05 oC on average with each increment of 10 cm in sand thickness. Aeolian sand enhances the seasonal thawing process in permafrost areas but weakens freezing process in seasonally frozen ground by increasing the active layer thickness (an increase of 7.70-9.50 cm on average in ALT with each increment of 10 cm in sand thickness) and decreasing seasonally frozen ground thickness (a decrease of 1.07-13.00 cm on average in SFT with each increment of 10 cm in sand thickness). Consequently, desertification on the QTP will have adverse effects on the preservation and development of frozen soil, causing degradation in permafrost and seasonally frozen ground. The extent of the degradation is enhanced with increasing aeolian sand cover thickness. The effects of aeolian sand cover on frozen soil are not entirely determined by the thickness of the sand layer. The moisture content and vegetation coverage can also influence the impacts of 19

aeolian sand on frozen soil. The annual heat budgets of soil layers covered by sand layers increase from -57.97 MJ·m-2 to -26.28 MJ·m-2 as the water contents decrease from 13.42 % to 3.61 %. Therefore, aeolian sand layers with low water contents caused by drought and windy environment, combined with high infiltration rates, play an adverse role in preserving frozen soil. Vegetation coverages are generally low and range from 5 % to 20 % in the QTP desert region. The sparse vegetation increases the AGRTs with an average increment from 2.19 °C to 6.17 °C at shallow depths (0.05-1.60 m) and is not conducive to the preservation and development of frozen soil. Large thicknesses of low-moisture aeolian sand resulting from strong expansion of desertification on the QTP, in combination with sparse vegetation cover, accelerate the degradation of frozen soil on the QTP.

Acknowledgments

This research was financially supported by the Key Program for Frontier Sciences of the Chinese Academy of Sciences (Grant No. QYZDJ-SSW-DQC011) and the National Natural Science Foundation of China (Grant No. 41690144). The authors are also grateful to engineers and members of the Beiluhe Observation and Research Station on Frozen Soil Engineering and Environment in Qinghai-Tibet Plateau, who helped the authors to carry out the fieldwork.

References

An, Z.S., Wu, X.H., Lu, Y.C., Zhang, D.E., Sun, X.J., Dong, G.R., Wang, S.M., 1991. Paleoenvionmental changes in China in

the recent 18000 years. Advance in Nature Science: Report of State Key Laboratory. (2), 153-159.

Baver, L.D., Gardiner, W.H., Gardiner, W.R., 1983. Soil Physics. Agriculture Press, Beijing, China.

Brown, R.J.E., 1973. Influence of climatic and terrain factors on ground temperatures at three locations in the permafrost region

20

of Canada. Proceedings of the 2nd International Conference on Permafrost. National Academy Press, Washington,

D.C., pp. 27-34.

Burn, C.R., Nelson, F.E., 2006. Comment on “A projection of severe near-surface permafrost degradation during the 21st century”

by David M. Lawrence and Andrew G. Slater. Geophysical Research Letters. 33 (21).

Chen, L., Yu, W.B., Yang, C.S., Yi, X., Liu, W.B., 2014. Conductivity of aeolian sand on the Tibetan Plateau based on

microstructure. Journal of Glaciology and Geocryology. 36 (5), 1220-1226.

Cheng, G.D., 2004. Influences of local factors on permafrost occurrence and their implications for Qinghai-Xizang Railway

design. Science in China Series D: Earth Sciences. 47 (8), 704-709.

Ding, G.D., 2010. Blown Sand Physics, second ed. China Forestry Publishing House, Beijing.

Dong, Y.X., 2001. Study on the control of land desertification and its project construction in Tibet autonomous region. Journal of

Natural Resources. 16 (2), 145-151.

Dong, Z.B., Hu, G.Y., Yan, C.Z., Lu, J.F., Wei, Z.H., 2012. The Desertification in Yangtze and Yellow Rivers, first ed. Science

Press, Beijing.

Feldman, G.M., 1982. Computing method of temperature regime of frozen soil, first ed. Science Press, Beijing.

French, H.M., 2007. The Periglacial Environment, third ed. Wiley Press, Chichester,UK.

Huang, Y.Z., Guo, D.X., Zhao, X.F., 1993. The desertification in the permafrost region of Qinghai-Xizang Plateau and its

influences on envrionment. Journal of Glaciology and Geocryology. 15 (1), 52-57.

Jin, H.J., Jin, X.Y., He, R.X., Luo, D.L., Chang, X.L., Wang, S.L., Marchenko, S.S., Yang, S.Z., Yi, C.L., Li, S.J., Harris, S.A.,

2018. Evolution of permafrost in China during the last 20 ka. Science China Earth Science. 61.

Jin, H.J., Sun, L.P., Wang, S.L., He, R.X., Lü, L.Z., Yu, S.P., 2008. Dual influences of local environmental variables on ground

temperatures on the interior-eastern Qinghai-Tibet Plateau (Ⅰ): Vegetation and Snow Cover. Journal of Glaciology

and Geocryology. 30 (4), 535-545.

21

Jin, H.J., Zhao, L., Wang, S.L., Guo, D.X., 2006. Evolution of permafrost and environmental changes of cold regions in eastern

and interior Qinghai-Tibet Plateau since the Holocene. Quaternary Sciences. 26 (2), 198-210.

Kudriavtsev, V.A., 1992. Fundamentals of Frost Forecasting in Geological Engineering Investigations. Lanzhou University Press,

Lanzhou.

Li, S., Dong, Y.X., Dong, G.R, Yang, P., Zhang, C.L., 2001. Regionalization of land desertification on Qinghai-Tibet Plateau.

Journal of Desert Research. 21 (4), 418-427.

Li, S., Yang, P., Gao, S.Y., Chen, H.S., Yao, F.F., 2004. Dynamic changes and developmental trends of the land desertification

in Tibetan Plateau over the past 10 years. Advance in Earth Sciences. 19 (1), 63-70.

Luo, D.L., Jin, H.J., He, R.X., Wang, X.F., Muskett, R.R., Marchenko, S.S., Romanovsky, V.E., 2018a. Characteristics of

water-heat exchanges and inconsistent surface temperature changes at an elevational permafrost site on the

Qinghai-Tibet Plateau. Journal of Geophysical Research-Atmospheres. 123, 10057-10075.

Luo, D.L., Jin, H.J., Wu, Q.B., Bense Victor F., He, R.X., Ma, Q., Gao, S.H., Jin, X.Y., Lü, L.Z., 2018b. Thermal regime of

warm-dry permafrost in relation to ground surface temperature in the Source Areas of the Yangtze and Yellow rivers

on the Qinghai-Tibet Plateau, SW China. Science of the Total Environment. 618, 1033-1045.

Lü, L.Z., Jin, H.J., Wang, S.L., Xue, X., He, R.X., Yu, S.P., 2008. Dual influences of local environmental variables on ground

temperatures on the interior-eastern Qinghai-Tibet Plateau (Ⅱ): Sand-layer and surface water bodies. Journal of

Glaciology and Geocryology. 30 (4), 546-555.

Pang, Q.Q., Zhao, L., Li, S.X., 2011. Influences of local factors on ground temperatures in permafrost regions along the

Qinghai-Tibet Highway. Journal of Glaciology and Geocryology. 33 (2), 349-356.

Qiu, G.Q., Cheng, G.D., 1995. Permafrost in China: past and persent. Quaternary Sciences. 15 (1), 13-22.

Smith, M.W., 1975. Microclimatic influences on ground temperatures and permafrost conditions , Mackenzie Delta Northwest

Territories. Canadian Journal of Earth Sciences. 12 (8), 1421-1438.

22

Wang, J.C., Wang, S.L., Qiu, G.Q., 1979. Permafrost along the Qinghai-Tibet Highway. Acta Geographica Sinica. 34 (1), 18-32.

Wang, L.Y., Wu, Q.B., Jiang, G.L., 2018. Numerical simulation of the effect of aeolian sand accumulation on permafrost.

Journal of Glaciology and Geocryology. 40 (4), 738-747.

Wang, S.L., 1989. Formation and evolution of permafrost on the Qinghai-Xizang plateau since the late Pleistocene. Journal of

Glaciology and Geocryology. 11 (1), 69-75.

Wang, S.L., Xie, Y.Q., 1998. Study on the ground temperature of sandy area in the Qinghai-Tibet Plateau. Journal of Desert

Research. 18 (2), 137-142.

Wang, S.L., Zhang, W.X., 1985. On permafrost evolution in the Qingshui River Region of the Qinghai-Xizang Plateau since the

late Pleistocene. Journal of Glaciology and Geocryology. 7 (1), 15-26.

Wang, S.L., Zhao, L., Li, S.X., 2002. Interaction between permafrost and desertification on the Qinghai-Tibet Plateau. Journal of

Desert Research. 22 (1), 33-39.

Wang, S.L., Zhao, X.M., 1999. Analysis of the ground temperatures monitored in permafrost regions on the Tibetan Plateau.

Journal of Glaciology and Geocryology. 21 (4), 351-356.

Wang, T., 2004. Study on sandy desertification in china-3.Key regions for studying and combating sandy desertification. Journal

of Desert Research. 24 (1), 1-9.

Williams, P.J., Smith, M.W., 1989. The frozen earth. New York : Cambridge University Press,

Wu, Q.B., Yu, W.B., Jin, H.J., 2017. No protection of permafrost due to desertification on the Qinghai–Tibet Plateau. Scientific

Reports. 7.

Wu, Z.Q., Zhang, M.Y., You, Z.L., Wang, J.W., 2016. Study of the effect of sand cover on the albedo of railway ballast layer.

Journal of Glaciology and Geocryology. 38 (6), 1598-1606.

Xie, S.B., Qu, J.J., 2013. Effect of sand sediments accumulated in sand-control projects on the thermal regime of underlying

permafrost and its mechanism. Journal of the China Railway Society. 35 (12), 77-82.

23

Xie, S.B., Qu, J.J., Lai, Y.M., Xu, X.T., Pang, Y.J., 2015. Key evidence of the role of desertification in protecting the underlying

permafrost in the Qinghai-Tibet Plateau. Scientific Reports. 5.

Xie, S.B., Qu, J.J., Xu, X.T., Pang, Y.J., 2017a. Interactions between freeze–thaw actions, wind erosion desertification, and

permafrost in the Qinghai–Tibet Plateau. Natural Hazards. 85, 829-850.

Xie, S.B., Qu, J.J., Xu, X.T., Pang, Y.J., Wang, T., 2017b. Experimental analysis of effect of sandy sediments produced by the

sand-control projects of Qinghai-Tibet Railway on the surface heat exchange. Journal of the China Railway Society.

39 (7), 159-165.

Xie, S.B., Qu, J.J., Zu, R.P., Zhang, K.C., Han, Q.J., 2012. New discoveries on the effects of desertification on the ground

temperature of permafrost and its significance to the Qinghai-Tibet Plateau. Chinese Science Bulletin. 57 (8),

838-842.

Xie, S.B., Qu, J.J.., Zu, R.P., Zhang, K.C., Han, Q.J., Niu, Q., 2013. Effect of sandy sediments produced by the mechanical

control of sand deposition on the thermal regime of underlying permafrost along the Qinghai-Tibet Railway. Land

Degrad. Develop. 24 (5), 453-462.

Xie, S.B., Yu, W.B., Qu, J.J., Pang, Y.J., 2018. Dynamic environment of blown sand at Honglianghe River of Qinghai- Tibet

Plateau. Journal of Desert Research. 38 (2), 219-224.

Yang, M.X., Wang, S.L., Yao, T.D., Gou, X.H., Lu, A.X., Guo, X.J., 2004. Desertification and its relationship with permafrost

degradation in Qinghai-Xizang (Tibet) plateau. Cold Regions Science and Technology. 39 (1), 47-53.

Yao, Z.Y., Li, X.Y., Dong, Z.B., 2015. Causes and processes of desertification in Madoi County in the source regions of the

Yellow River. Journal of Glaciology and Geocryology. 37 (5), 1245-1256.

Yershov, E.D., 2016. Principles of Geocryology: Dynamic Geocryology, first ed. Lanzhou University Press, Lanzhou.

Zhang, K.C., Niu, Q.H., Han, Q.J., 2011a. Characteristics of wind-blown sand and dynamic environment in the section of

Wudaoliang-Tuotuo River along the Qinghai-Tibet Railway. Environmental Earth Sciences. 64 (8), 2039-2046.

24

Zhang, K.C., Qu, J.J., Liao, K.T., Niu, Q.H., Han, Q.J., 2010. Damage by wind-blown sand and its control along Qinghai-Tibet

Railway in China. Aeolian Research. 1 (3), 143-146.

Zhang, K.C., Qu, J.J., Niu, Q.H., Yao, Z.Y., Han, Q.J., 2011b. Protective mechanism and Efficiency of sand-blocking fences

along Qinghai-Tibet Railway. Journal of Desert Research. 31 (1), 16-20.

Zhang, T.J., Baker, T.H.W., Cheng, G.D., 2008. The Qinghai-Tibet Railroad: A milestone project and its environmental impact.

Cold Regions Science and Technology. 53 (3), 229-240.

Zhao, L., Sheng, Y., 2015. Permafrost Survey Manual. Science Press, Beijing, China, pp. 55-59.

Fig. 1. The landscape of Honglianghe River basin, (a) alpine desert vegetation landscape, (b) desert landscape.

25

Fig. 2. Map of study sites at the Honglianghe River. A and B are two geological profiles. The red dot represents the location of Honglianghe River basin in the Qinghai-Tibet plateau

26

Fig. 3. Maximum and minimum ground temperatures (Tmax and Tmin) at study sites during the observation period from August 23, 2017 to August 23, 2018, and the soil temperatures of QH5-P are measured from January 23, 2015 to January 23, 2016. The acronyms of “P” and “S” mean permafrost and seasonal frozen soil, respectively. (a) temperatures at HLH-12-P and QH5-P site, (b) temperatures at HLH-08-P and HLH-07-P site, (c) temperatures at HLH-05-S and HLH-04-S site, (d) temperatures at HLH-01-S, HLH-11-S and HLH-10-S site.

27

Fig. 4. Variations in soil temperatures at different depths at study sites, (a) temperatures at HLH-08-P and HLH-07-P site, (b) temperatures at HLH-12-P and QH5-P site, (c) temperatures at HLH-05-S and HLH-04-S site, (d) temperatures at QH4-S, HLH-03-S and HLH-02-S site; soil temperatures for QH4-S and QH5-P were measured from August 2014 to July 2015.

Fig. 5. MAGTs with different thicknesses of sand cover at study sites.

Fig. 6. Annual average ground temperatures at study sites, (a) temperatures at HLH-12-P and QH5-P site, (b) temperatures at HLH-08-P and HLH-07-P site, (c) temperatures at HLH-05-S and HLH-04-S site, (d) temperatures at HLH-02-S, HLH-03-S and QH4-S site; the average temperatures at QH4-S and QH5-P are measured from 2014 28

to 2015.

Fig. 7. Seasonal thawing and freezing processes: (a) seasonal thawing process at HLH-07-P, (b) seasonal thawing process at HLH-08-P, (c) seasonal freezing process at HLH-05-S, and (d) seasonal freezing process at HLH-04-S.

Fig. 8. Characteristic parameters of seasonal freezing and thawing processes with the impacts of sand cover; (a) 29

seasonal freezing process, and (b) seasonal thawing process. Note: A: annual range of ground temperature; Tm: mean annual ground temperature; Tmin and Tmax: minimum and maximum soil temperature; ξ: maximum seasonal freezing or thawing depth; Δ ξ: increment or decrement of ξ; h: depth of zero annual amplitude of ground temperature; subscripts 1 and 2: respectively represent thin and thick sand layers.

Fig. 9. The annual range of ground temperatures from August 23, 2017 to August 23, 2018 at shallow depths. (a) annual range of ground temperatures at HLH-07-P and HLH-06-S, (b) annual range of ground temperatures at HLH-08-P and HLH-01-S.

Table 1 Sand thickness, surface condition, mean annual ground temperature (MAGT), active layer thickness (ALT), and seasonally frozen soil thickness (SFT) at 14 study sites. Frozen soil

Site

condition

Sand thickness

Surface condition

(cm)

MAG

ALT/SFT

T

(m)

(°C) Permafrost

HLH-12-P

10

Sand cover

-0.54

3.29

HLH-08-P

20

Sparse vegetation (coverage:5%)

-0.96

1.90

HLH-07-P

40

Sparse vegetation (coverage:5%)

-0.92

2.09

QH5-P

115

Sand cover

-0.42

4.10

Seasonally

HLH-01-S

20

Sparse vegetation (coverage:10%)

0.73

3.21

frozen

QH4-S

30

Sand cover

0.25

4.95

ground

HLH-06-S

40

Sparse vegetation (coverage:20%)

0.15

6.50

HLH-05-S

60

Sparse vegetation (coverage:5%)

0.30

4.57

30

HLH-11-S

160

Sparse vegetation (coverage:10%)

0.97

3.06

HLH-04-S

180

Sparse vegetation (coverage:5%)

0.94

3.01

HLH-09-S

180

Sparse vegetation (coverage:15%)

1.02

3.02

HLH-10-S

220

Sparse vegetation (coverage:10%)

1.04

2.94

HLH-03-S

300

Sand cover

0.75

3.92

HLH-02-S

350

Sand cover

0.96

3.46

31

Table 2 Feature parameters of seasonal thawing and freezing process. Frozen

soil

Site

h

C

ALT/SFT

(cm)

(%)

(m)

HLH-08-P

20

5

1.90

2018/5/11

2018/10/15

157

HLH-07-P

40

5

2.09

2018/5/10

2018/10/7

150

HLH-12-P

10

0

3.29

2018/4/16

2018/10/23

190

QH5-P

115

0

4.10

2015/4/25

2015/11/16

205

Seasonally frozen

QH4-S

30

0

4.95

2014/11/5

2015/8/24

293

ground

HLH-03-S

300

0

3.92

2017/10/25

2018/7/6

254

HLH-02-S

350

0

3.46

2017/10/30

2018/4/19

171

HLH-05-S

60

5

4.57

2017/10/25

2018/7/8

256

HLH-04-S

180

5

3.01

2017/10/25

2018/4/2

159

HLH-01-S

20

10

3.21

2017/10/25

2018/6/19

237

HLH-11-S

160

10

3.06

2017/10/29

2018/5/20

203

HLH-10-S

220

10

2.94

2017/10/25

2018/4/29

186

condition Permafrost

t0

tm

Δtm (d)

Note: h: thickness of sand cover; C: vegetation coverage; t0: onset of seasonal thawing-freezing from the ground surface; tm: date of maximum seasonally thawing-freezing depth; Δtm: duration of the seasonal thawing-freezing process in days.

32

Table 3 Weighted mean sand water content, thermal conductivity of soil and the annual heat budget in soil. Surface condition

Site

Sand

w

λf

λu

Annual

thicknes

(%)

(W·m-2·°C -1)

(W·m-2·°C -1)

heat budget (MJ·m-2)

s (cm) Aeolian

sand

HLH-02-S

350

3.61

1.35

1.40

-36.66

HLH-03-S

300

6.27

1.74

1.58

-40.52

HLH-12-P

10

13.42

2.42

2.02

-45.51

Desert with sparse

HLH-10-S

220

4.48

1.47

1.46

-26.28

vegetation

HLH-11-S

160

5.00

1.54

1.50

-42.03

(coverages:10%)

HLH-01-S

20

7.33

1.88

1.64

-57.97

cover

Note: w: weighted mean water content (%); the soil water contents were obtained from our in-site works by drying soil samples in dry oven. λf and λu (W·m-2·°C -1): linear interpolated thermal conductivities.

HIGHLIGHTS

33

1. Aeolian sand cover warms the frozen soil. 2. The thicker the aeolian sand, the stronger the soil warming effect. 3. Desertification accelerates the degradation of frozen soils.

34