Permafrost thermal dynamics at a local scale in northern Da Xing’anling Mountains

Permafrost in Northeastern China is not only controlled by latitude and elevation, but also locally environmental factors, such as vegetation cover and human activities. During 2009–2022, thinning active layer, increasing annual maximum frost depth in talik zones and lowering ground temperature above the depth of dividing point (DDP) between permafrost cooling and warming have been observed in many places, possibly due to the global warming hiatus (GWH). However, the responses of permafrost below DDP did not show a clear trend to the GWH, despite an evident ground warming. The warming and degradation of permafrost below DDP in the Da Xing’anling Mountains are more strongly influenced by the overall climate warming than by regional GWH. This study improves our understanding of changing permafrost temperature and its drivers. It also helps to provide data support and references for the management of the ecological and hydrological environment of the northern Da Xing’anling Mountains and the Heilongjiang-Amur River Basin.


Introduction
Permafrost is a major component of the cryosphere.Its distribution and hydrological, thermal and physical conditions are controlled or influenced by various factors and exhibit significant spatiotemporal variability.At a global scale, permafrost is primarily controlled by the latitudinal zonality of climate (Yershov et al 1988, Zhou et al 2000a).At a regional scale, the control and influence of hydroclimate factors, such as elevation and climate aridity become significant to permafrost formation.At a local scale, ground hydrothermal regimes are highly dependent on various factors, such as snow cover, vegetation cover, slope angles and aspect, lithology, and soil hydrology and moisture contents, resulting in great variability in permafrost distribution and other characteristics (Tutubalina andRees 2001, Cao et al 2020).
Northeastern China hosts the only region of latitudinal permafrost and the second largest permafrost region in China (Zhou et al 2000a).As a result of climate warming, permafrost in this region has been experiencing active layer deepening, temperature rising, taliks and thermokarst landforms occurrences, areal continuity declining and extent shrinking since the late 20th century (Jin and Ma 2021).Numerous studies show that permafrost degradation casts significant impacts not only on hydrological and ecological processes (Zhang et al 2017(Zhang et al , 2018) ) and the release of perennially frozen organic carbon and nitrogen (Wu et al 2022), but also on infrastructure (Liew et al 2022), food and water security (Maslakov et al 2020), public health (Revich et al 2022), industrial production (Wang et al 2019), and socio-economic development (Sjöberg et al 2020, Streletskiy et al 2023).
So far there are significant progresses in studies on permafrost degradation in many regions (e.g.Ishikawa et al 2018, Angelopoulos et al 2020, Cao et al 2020, Yin et al 2021, Jin et al 2021a, Malkova et al 2022, Smith et al 2022), but research on permafrost in the Da Xinganling Mountains has been limited (He et al 2021, 2022, Chang et al 2022, Li et al 2022).This study aims to investigate the thermal characteristics of permafrost and active layer in the Da Xing'anling Mountains, and examine the temperature changes over the past two decades (2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018)(2019)(2020)(2021)(2022) by analyzing long-term data of air and ground temperatures from national meteorological stations and our observation sites in Mangui, Genhe, Yituli'he, and Nanwenghe in the northern part of Northeast China.These study results can provide timely, important guidance for ecological and watershed management in the Da Xing'anling Mountains and Heilongjiang-Amur River Basin and can contribute to regional sustainable development in an environment of degrading permafrost under a changing climate.

Study area and data
The northern Da Xing'an Mountains are predominantly covered by cold temperate coniferous forests, with extensive by-occurrences of shrub, swamp, and moss/peat layers.Due to the combination of relatively higher latitudes, low mean annual air temperatures (MAAT), widespread and stable temperature inversion layers in winter, dense forest vegetation, and large thermal insulation semiconductivity of moss/peat layers, the unique Xing'an-Baikal permafrost (Jin et al 2007), with complex southern limit and features, has been developed and preserved in this region.
Currently, long-term monitoring stations for permafrost temperature have been established in several locations in the Da Xing'anling Mountains, including Yituli'he, Genhe, Alongshan, Mangui, Mo'he, Xinlin, and Nanwenghe.These yielded data have been published in the National Tibetan Plateau Data Center (Chang et al 2022).However, due to the damage to temperature sensors and the impacts of the COVID pandemic and ensued traffic control, some data were missing for the periods of January to June 2014 and 2019-2021.There are some more data gaps resulting from frequently failed batteries of borehole GH-8 caused by the low temperatures in Northeast China, and the MG-3 site was destroyed in 2016.Other detailed information about boreholes are included in table 1.

Changes in active layer thickness (ALT)
Long-term, continuous measurements of ground temperature in boreholes indicate that, in the Mangui, Genhe, Yituli'he, and Nanwenghe in the northern permafrost zone of the Da Xing'anling Mountains since the start of observation in 2009 (figures 1(a)-(e)).The only exception is borehole MG-3 was destroyed in 2016.Among them, the most rapid thinning of active layer is found in boreholes GH-1 and NE-2, which were observed for 5 and 6 years, respectively, with decreasing rates of ALT at 26.7 and 15.6 cm a −1 (figures 1(c) and (d)).The data from boreholes MG-2, YTLH-1, and YTLH-2 showed annual declining rates of ALT at approximately 7.0, 5.6, and 4.0 cm a −1 , respectively (figures 1(a) and (b)), while a relatively small declining rate, about 3.0 cm a −1 for boreholes MG-1, GH-8, GH-2, and NE-1.However, ALT remained relatively stable in the perennially waterlogged Carex tato wetlands (GH-3) and the grass/sedges wetlands (GH-9).
Figure 1(F) shows the trend of annual maximum depth of frost penetration (AMDFP) observed at the GH-Q and GH-721 boreholes in the talik zone of Genhe from 2009 to 2019.The Borehole GH-Q is located immediately outside the meteorological station of Genhe city, with low (maintained lawn) grass cover, while the Borehole GH-721 is situated in the outskirts of Genhe city, with vegetation cover consisting of grasses and white birch shrubs.Both locations are considered as the through talik zone.Since 2009, the AMDFP at GH-Q increased from 0.8 to 1.9 m, at an increasing rate of approximately 7.0 cm a −1 , and; that at the borehole GH-721 fluctuated between 1.5 and 3.0 m, increasing at 2.8 cm a −1 .
In summary, in 2009-2019, ALT in the northern Da Xing'anling Mountains was on a decline, but AMDFP in talik zones was increasing.

Change rates of permafrost temperature
Boreholes with depths ⩾20 m in Mangui, Genhe, and Yituli'he were chosen to analyze the variations in ground temperature at depths greater than 2.5 m during 2009-2019 (figure 2).Based on the position of dividing point between permafrost cooling and warming, where change rate of permafrost temperature is 0 (dashed line in this figure), three kinds of change rate curve can be observed.These three cases indicate that the permafrost in the Da Xing'anling Mountains undergoes complex changes under a warming climate.In the first and second cases, the deep permafrost layers all show a warming trend.In the third case, a warming rate of 0.001 • C/10a at the depth of 20 m was yielded for the observations from the Borehole YTLH-2.Thus, a gradual permafrost warming below 20 m was inferred.According to the statistics in table 2, the temperature change rate in the permafrost of Nanwenghe (P1 site) is negative at depths of less than 30 m and positive at depths greater than 30 m.However, the negative-positive position is roughly found at 15 m in depth for the Boreholes XL-2 and XL-3 in Xinlin town, i.e. both ground cooling (smaller than 15 m in depth) and warming (greater than 15 m in depth) occurred in permafrost layers of this region during 2011-2022.The change rate of permafrost temperatures in borehole XL-1 exhibited a negative variation in ground temperature at depths shallower than 20 m, with much smaller change rate with increasing depth, in accord with the third case.This suggested that during the observation, permafrost here persisted in cooling above 20 m, but at a certain depth greater than 20 m, its temperature variation might have approximated zero and subsequently become positive.In other words, the third case of temperature change rate curve in the permafrost is considered to be an extension of the second case, with the negativepositive turning position ground at greater depths than 20 m, which was not observed in our study.According to statistical results, the positions where the temperature change rates become zero were located at depths of 5.0, 4.5, and 7.0 m for MG-1, MG-2, and GH-9, respectively, 15 m for XL-2 and XL-3, 20 m for YTLH-2, and above 20 m for XL-1, while it is at 30 m for P1.

Causes of active layer thinning
The Sixth Assessment Report (AR6) of the Intergovernmental Panel on Climate Change (IPCC) highlights that high-altitude permafrost in the Eurasian continent has experienced an increase in ALT since the mid-1990s (Biskaborn et al 2015).Similarly, ALT in Arctic Europe and Russia has generally increased since the 21st century (Zhong et al 2022).Analysis of over 230 observation sites of ALT reveals that during 1990-2013, approximately 47% of locations in Alaska exhibited a significant increase in ALT, with a maximum annual change rate of 0.69 cm a −1 (Wang et al 2015).In Russia and on the Qinghai-Tibet Plateau, around 90% of sites showed an increase trend in ALT, with maximum annual change rates of 5.9 and 23.77 cm a −1 , respectively (Yang et

Depth of dividing point (DDP) between permafrost warming and cooling
We refer to the depth at which the permafrost temperature change rate (obtained from linear regression) equals to zero during the observation period, i.e. the DDP between permafrost cooling and warming as DDP (denoted by arrows in figure 3).Based on the statistical analysis of long-term changes in permafrost temperature in section 3.2, DDP varies between 4.5-30.0m.This means that permafrost in different regions exhibit distinct responses to the GWH, possibly due to some local factors, such as vegetation cover and soil moisture content.The DDP of permafrost in the northern Da Xing'anling Mountains descends in the following order: MG-2 > MG-1, XL-1 > XL-3 > XL-2, and P1 > P2.The vegetation at MG-2, XL-1, and NW-P1 consists of marsh wetlands, with higher soil moisture content.In contrast, MG-1, XL-3, XL-2, and P2 are covered by deciduous forests or shrubs, with relatively lower soil moisture content.It can be inferred that sedges (Carex tato) swamp with higher soil moisture content exhibit a more significant response to climate change.The presence of deciduous forests or shrubs, due to their capability in redistributing solar radiation and producing a thicker layer of litterfall, leads to a less pronounced response.In the case of the boreholes in Genhe, GH-9 is characterized by grass vegetation, and the DDP is observed at 7.0 m.GH-8, located in the interior of the pristine forest, is covered by birch and larch forests.However, due to a considerable amount of missing data, its response to the GWH seemed insignificant, with only a slight decrease in ALT during the observation period.

Warming and degradation of permafrost below the DDP
Although the active layer in permafrost regions in the Da Xing'anling Mountains has mostly thinned and temperatures of near-surface permafrost have lowered during 2009-2019, the deep permafrost below the DDP has experienced a consistent warming.This indicates that in these areas, permafrost below the DDP has not yet responded to the GWH and is undergoing a state of warming and degradation.
Currently, there is a widespread thermal state of increasing temperature in permafrost regions in the Northern Hemisphere, yet with strong regional variations in warming rates (Biskaborn et al 2019).For example, in the western Arctic Russia, the warming rate of permafrost ranged from 0.35 to 0.56 • C/10a during 1970-2020 (Vasiliev et al 2020), while in the high-latitude permafrost regions of northern Canada, the warming rates at depths of 15 and 24 m are 0.6 and 0.4 • C/10a, respectively, and ground temperature at the depth of 50 m has increased by 0.5 • C over the past 40 years, or 0.125 • C/10a (Smith et al 2019).In Mongolia, permafrost warmed at rates of 0.1-0.3• C/10a at depths of 10-15 m during 1980s-2017 (Dashtseren 2021), and; in Alaska, the rate fluctuated between 0.1 and 1.0 • C/10a (Wang et al 2015).According to table 2, the warming rates of permafrost at depths below the DDP in the Xing'anling Mountains ranged from 0 to 0.52 • C/10a during 2009-2019, similar to those of 0.1-0.6 • C/10a on the Qinghai-Tibet Plateau (Zhao et al 2010, 2020, Wu et al 2012), but lower than those in the Arctic (0.4-0.6 • C/10a) during the period from 1974 to 2019 (Zhong et al 2022).In addition, the warming rates of permafrost in most areas were lower than those rising rates of MAAT increase in the Da Xing'anling Mountains region (0.3-0.5 • C/10a) from 1960 to 2015 (Zhou et al 2020b).
Among them, the borehole MG-3 exhibits the highest warming rates at all depths (figure 2), which can be attributed to two factors.First, the observation period (2012-2016) was relatively short compared to other boreholes, which may strongly affect the outcomes (Wu et al 2022).Second, the borehole is located in a backyard close to a local residence, where everyday life activities, such as heating with firewood, and vegetation clearance may result in abnormally high warming rates compared to other adjacent boreholes.The warming rate of permafrost at depths greater than 15 m in this location is similar to that of other boreholes, approximately 0.2 • C/10a, suggesting that human activities have not significantly impacted the deeper permafrost.In contrast, borehole YTLH-2, under a dramatically reduced human disturbances in the Yituli'he region since 1998 (Chang et al 2022), and possibly also the influences of the GWH, shows a cooling trend in the near-surface permafrost at depths of greater than 20 m.The change rates of ground temperatures of deep permafrost (⩾10 m) in boreholes MG-1, GH-8, P2, and MG-2 have not exceeded 0.3 • C/10a (table 2).It is notable that in borehole MG-1, the warming rates increase with depth, indicating an accelerated warming or degradation of the permafrost in this area.Permafrost temperatures in borehole GH-9 remained almost unchanged during 2009-2019, maintaining around 0.04-0.05• C/10a, similar to the changing rates of ground temperatures at 15 m of patchy permafrost on the upper Mahanshan Mountains 40 km east of Lanzhou, China (0.02 • C/10a) (Wu et al 2022).However, due to the thick active layer at this location (6.5 m) and warm permafrost (-0.3 • C), it is inferred that the permafrost here is on the verge of degradation and disappearance, and it is highly probable that a supra-permafrost subaerial talik will soon occur.
In summary, the warming and degradation of deep permafrost in the Da Xing'anling Mountains are more strongly influenced by the overall climate warming than by the GWH.However, changes in human activities can correspondingly modify the response of permafrost temperature at depths.This can result in the warming rate at the depth of 10 m in Borehole MG-3 being more than twice that of the same depth in other boreholes, or the cooling trend in the permafrost at depths of above 20 m in Borehole YTLH-2 in the northern Da Xing'anling Mountains, Northeast China.

Conclusions
The thermal dynamics of the Xing'an permafrost in the northern Da Xing'anling Mountains are quite complicated and closely related to local factors, such as vegetation cover, snow cover, soil moisture content, and human activities.
During the observation period (2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018)(2019)(2020)(2021)(2022), ALTs in the study areas show a thinning trend, decreasing by 3.0-26.7 cm annually, while the AMDFP in taliks increased by 2.8-7.0 cm per year.Permafrost at depths above DDP (between permafrost cooling and warming) exhibited a cooling trend, possibly as a response to the GWH, but permafrost below this depth showed a consistent warming trend and was insensitive to the GWH.There is significant local variability in the position of DDP due to the strong influences of local factors and human activities.Particularly, the influences of human activities are noteworthy, which can exacerbate or mitigate the changes in permafrost temperature along with other factors.
Overall, the warming and degradation of deep Xing'an permafrost are more strongly influenced by the global warming than by regionally temporary GWH.All the findings highlight the complex interplay between regional and global factors, as well as the joint influences of local factors and human activities, in shaping the distributive and other characteristics, and the responses of Xing'an permafrost to climate changes of different spatiotemporal scales.
Borehole positions and permafrost features in permafrost zones of the northern Da Xing'anling Mountains in Northeast China.annual ground temperature (MAGT) at the depth of zero annual amplitude (D ZAA ).b Denotes MAGT at the maximum observation depth.* MALT, mean active layer thickness during the observation.
al 2010, Zhao et al 2010).Research conducted by Jin et al (2007) indicated a fluctuating increase in ALT in the Da Xing'anling Mountains, particularly in the Yituli'he region, during the 1990s.However, in central Yakutia, approximately one-third of boreholes showed a decrease in ALT, while nearly half of the boreholes exhibited negligible changes, not exceeding 5 cm/10a (Varlamov et al 2019).In this study, a fluctuatingly decreasing trend in ALT was observed in Mangui, Genhe, Yituli'he, Nanwenghe, and other areas in northern Da Xing'anling Mountains since 2006.Additionally, MG-1, MG-2, GH-5, and YTLH-2 experienced significant ground cooling at shallow depths of latitudinal permafrost.These observations could be attributed to factors, such as local climate, snow cover, and vegetation dynamics (Chang et al 2022, Zhang et al 2023).Numerous studies indicate a global warming hiatus (GWH) in 1998-2013 (Yan et al 2016, Medhaug et al 2017), when the increase rate of

Table 2 .
Thermal change rates ( • С/10a) below 10 m for different boreholes in northern Da Xing'anling Mountains, Northeast China during 2009-2019.The asterisk * means ground temperatures of boreholes in taliks; the observation period of boreholes is the same as that in table 1.
(Zhou et al 2000a)hou et al 2020b), and similar results were also found in Northeast China(Sun et al 2018, Zhou et al 2020b).For example, analysis Climate, snow cover, and vegetation interact with each other and collectively influence ALT and ground temperatures of near-surface permafrost(Wang et al 2024).It is well-known that permafrost temperatures can lag behind climate change(Zhou et al 2000a).Therefore, the observed active layer thinning and ground cooling in the near-surface permafrost after 2008 are likely responses to the above mentioned GWH.