Sciences in Cold and Arid Regions  2016, 8 (4): 269-296   PDF    

Article Information

JunJin Hui, LiChang Xiao, LiangLuo Dong, XiaHe Rui, Zhi Lü Lan, ZhongYang Si, Xin Guo Dong, Mei Chen Xue, A. Harris Stuart . 2016.
Evolution of permafrost in Northeast China since the Late Pleistocene
Sciences in Cold and Arid Regions, 8(4): 269-296
http://dx.doi.org/10.3724/SP.J.1226.2016.00269

Article History

Received: May 08, 2016
Accepted: June 22, 2016
Evolution of permafrost in Northeast China since the Late Pleistocene
JunJin Hui1, LiChang Xiao1,2, LiangLuo Dong1, XiaHe Rui1, Zhi Lü Lan1, ZhongYang Si1,3, Xin Guo Dong1, Mei Chen Xue1,4, A. Harris Stuart5     
1. State Key Laboratory of Frozen Soils Engineering, Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou, Gansu 730000, China;
2. National-Local Joint Engineering Laboratory of Geo-Spatial Information Technology, Hunan University of Science and Technology, Xiangtan, Hunan 411201, China;
3. GFZ German Research Centre for Geosciences, Telegrafenberg, D-14473 Potsdam, Germany;
4. College of Resources and Environment, Lanzhou University, Lanzhou, Gansu 730000, China;
5. Department of Geography, University of Calgary, Calgary, Alberta, Canada T2 N1 N4
Abstract: In Northeast China, permafrost advanced and retreated several times under the influences of fluctuating paleo-climates and paleo-environments since the Late Pleistocene. During the last 60 years, many new data were obtained and studies were conducted on the evolution of permafrost in Northeast China, but so far no systematic summary and review have been made. Based on sedimentary sequences, remains of past permafrost, paleo-flora and -fauna records, and dating data, permafrost evolution since the Late Pleistocene has been analyzed and reconstructed in this paper. Paleo-temperatures reconstructed from the remains of past permafrost and those from paleo-flora and -fauna are compared, and thus the southern limit of permafrost (SLP) in each climate period is inferred by the relationship of the permafrost distribution and the mean annual air/ground temperatures (MAAT/MAGT). Thus, the evolutionary history of permafrost is here divided into five stages:(1) the Late Pleistocene (Last Glaciation, or LG) (65 to 10-8.5 ka), the Last Glaciation Maximum (LGM, 21-13 ka) in particular, the coldest period in the latest history with a cooling of about 6~10℃, characterized by extensive occurrences of glaciation, flourishing Mammathas-Coelodonta Faunal Complex (MCFC), widespread aeolian deposits, and significant sea level lowering, and permafrost greatly expanded southwards almost to the coastal plains (37°N-41°N); (2) the Holocene Megathermal Period (HMP, 8.5-7.0 to 4.0-3.0 ka), 3~5℃ warmer than today, permafrost retreated to about 52°N; (3) the Late Holocene Cold Period (Neoglaciation) (4.0-3.0 to 1.0-0.5 ka), a cooling of 1~3℃, some earlier thawed permafrost was refrozen or attached, and the SLP invaded southwards to 46°N; (4) the Little Ice Age (LIA, 500 to 100-150 a), the latest cold period with significant permafrost expansion; and (5) climate warming since the last century, during which Northeast China has undergone extensive permafrost degradation. The frequent and substantial expansions and retreats of permafrost have greatly impacted cold-region environments in Northeast China. North of the SLP during the HMP, or in the present continuous permafrost zone, the existing permafrost was largely formed during the LG and was later overlapped by the permafrost formed in the Neoglaciation. To the south, it was formed in the Neoglaciation. However, many aspects of permafrost evolution still await further investigations, such as data integration, numerical reconstruction, and merging of Chinese permafrost history with those of bordering regions as well as collaboration with related disciplines. Of these, studies on the evolution and degradation of permafrost during the past 150 years and its hydrological, ecological, and environmental impacts should be prioritized.
Key words: Last Glaciation (LG)     Holocene Megathermal Period (HMP)     Neoglaciation     Little Ice Age (LIA)     Northeast China     permafrost evolution    
1 Introduction

Due to climate change and increasing human activities, natural environments have been greatly altered and human societies are facing many challenges regarding the natural resources and sustainable development in cold regions. Studies on changes in past climates and environments are important for improving the predictability of long-term trends. Amongst them, the reconstruction of evolutionary processes of natural environments since the Quaternary are key to understanding the controlling mechanisms and for future development tendencies of the nature-human coupled systems, as well as for the logical exploitation of natural resources and sustainable management of natural environments.

Among the many types of periglacial remains, extensive occurrence of cryogenic wedge structures, if well identified and clarified, and elaborated with other concurring cryogenic structures such as cryoturbations (involutions) and pingo scars, can serve as the most reliable and direct evidence of past permafrost (e.g., Romanovskii, 1977; Murton and Kolstrup, 2003; French, 2007). This is because they indicate the occurrence of past frost cracking at certain vertical and horizontal scales into the active layer and shallow (transient) permafrost. However, the extensive occurrence of cryogenic wedges can only largely be related to past air or ground (surface) temperatures (Péwé, 1966a, b; Harry and Gozdzik, 1988).

According to their different filling materials, cryogenic wedge structures can be further divided into ice wedge pseudomorphs (casts) and primary sand or soil wedges (veins). They differ in geometry and internal structures depending on ambient environments of initial formation and secondary modifications. Therefore, in order to better reconstruct the paleo-environment, one must first determine the cryogenetic types of these wedge structures. Regarding the scales and geometry of wedges, ice wedges generally have top widths of 2 to 3 m and penetrate 1 to 10 m in depth, and they should be under the permafrost table. In general, primary sand and soil wedges have top widths of about 1 m and their penetration depths are shallower than the depths of the permafrost table. In very cold environments, sand and gravel wedges can penetrate deeper than the permafrost table.

Wedge structures also differ in spatial distribution. Ice wedges generally are formed in the continuous permafrost zone. As a rule of thumb, ice wedge pseudomorphs (or casts) in fine-grained soils generally indicate the past occurrence of a discontinuous permafrost zone, while those in coarse-grained soils or weathered bedrocks could be indicators for continuous permafrost because polygonal cracking in coarse deposits requires a harsher periglacial environment. Sand and soil wedges are generally formed in discontinuous permafrost zones, but their occurrence in polar regions is not uncommon. In Siberia, presently existing ice wedges are generally found to the north of 70°N (continuous permafrost zone), while there are sand and soil wedges in discontinuous permafrost zones in the Siberia Lowlands (55°N-70°N) (e.g., Mackay, 1974; Black, 1976; Romanovskii, 1980, 1985; Washburn, 1980; Péwé, 1983; Harry and Gozdzik, 1988).

Larger (≥0.5 m in thickness) cryoturbations (involutions) generally occur in the degradation process of ice-rich permafrost and are therefore reliable indicators of past occurrence of permafrost (Van Vliet-Lanoë, 1988; Murton and French, 1995). Regarding their filling materials, an ice wedge and its casts are similar to sand and soil wedges, but they differ in wedge structures and the origins of the internal wedge materials. Ice wedge casts are formed by infilling after the vein ice melts and ambient soils generally slide unevenly into the wedges; thus, their filling materials are generally from ambient soils or rocks. Therefore, the wall of the wedge is irregular and ambient soils or rocks often have upturned structure towards the wedge wall. In primary sand and soil wedges, sand deposits or soil formation are generally concurrent with frost wedge and polygonal cracking, and gradually fill up with wedge expansion. Therefore, inside sand and soil wedges, layered structures parallel the wedge wall and vertical or slanted lamina are often observed (Xu and Pan, 1990).

In the early 1900s, the paleoclimatic significance of periglacial remains was first recognized, and periglacial evidence was applied for inferring regional environmental changes in Europe. Since then, periglacial remains and relevant environmental reconstruction have made great progress in Europe and North America (e.g., Black, 1976; Washburn, 1980; Kolstrup, 1986; Vandenberghe, 1992; Kasse et al., 1998). In the 1960s-1980s, scholars in the former USSR, Europe, and North America conducted numerous studies on polygonal wedge structures, and regarded them as reliable indicators for past permafrost occurrence (e.g., Péwé, 1966a, b, 1983; Mackay, 1974; Black, 1976; Romanovskii, 1980, 1985; Washburn, 1980; Harry and Gozdzik, 1988). In high-latitude permafrost zones, wedge structures are still active at present. In Alaska, the mean annual air temperature (MAAT) is at about −8 ℃ to −6 ℃ and −12 ℃, respectively, at the southern and northern limits of presently active ice wedges (Washburn, 1980); it is at −10 ℃ along the Atlantic coasts in Northern Siberia, and at about −8 ℃ to −3 ℃ for primary sand and soil wedges in the Siberia Lowlands (Romanovskii, 1980, 1985).

Air and ground temperatures for forming these wedge structures vary greatly (Romanovskii, 1977). In fine-grained soils, sand and soil wedges can form at a mean annual ground temperature (MAGT) of about −2 ℃ to −1 ℃, but ice wedges require a lower MAGT (about −5 ℃ to −4 ℃). In coarse-grained soils, sand and soil wedges generally occur at an MAGT of about −5.0 ℃ to −2.4 ℃, but ice wedges form at about −6 ℃ to −5 ℃ (Romanovskii, 1977). These criteria have been well established and extensively cited in Quaternary geocryology and geology for inferring present and past air and ground temperatures (e.g., Péwé, 1966b; Guo, 1979; Liang and Cheng, 1984; Harry and Gozdzik, 1988; Peng and Cheng, 1990; Xu and Pan, 1990; Tong, 1993; Murton and French, 1995; Huijzer and Vandenberghe, 1998).

However, the significance, reliability, and suitability of wedge structures for paleoclimatic and paleo-environmental reconstruction have been occasionally challenged by some scholars, who suggest that soil moisture, lithology, snow and vegetative covers, ground cooling rates, and many other local environmental variables are also important in evaluating the wedge-indicated paleo-environments and -climates (e.g., Murton and French, 1995; Murton and Kolstrup, 2003). In general, the MAAT for the formation of ice wedges is no warmer than −8 ℃ (coarse-grained soils) to −4 ℃ (fine-grained soils), and the average air temperature in the coldest month is no warmer than about −2 ℃ to 0 ℃ (Romanovskii, 1985; Murton and French, 1995; Van Huissteden et al., 2003).

Studies on the periglacial environment in China started late in China. The earliest work was reports on involutions (cryoturbations) in Huangshan, Harbin, and Jalainor, Inner Mongolia Autonomous Region (IMAR) by Pei (1957). Later, periglacial remains were identified extensively in Datong, Shanxi Province, and on the southern Ordos Plateau in North China (Yang et al., 1983; Cui and Xie, 1984; Dong et al., 1985; Vandenberghe et al., 2004). Reports also came from West China, such as the Qinghai-Tibet Plateau (e.g., Guo, 1979; Jin et al., 2007a), Hexi Corridor, and Tengger Desert (Wang et al., 2003). Recently, some new discoveries were also reported in Northeast China and on the Qinghai-Tibet Plateau (e.g., Cheng et al., 2006; Chang et al., 2011; Jin et al., 2011; Jiao et al., 2015), as well as in the Qilian Mountains (e.g., Vandenberghe et al., 2015). Some progress has also been made in the interdisciplinary fields of periglacial and aeolian geomorphology (e.g., Li et al., 2006).

Since the 1980s, natural environments in Northeast China and the Far East of Russia have drawn extensive and increasing attention due to their dramatic changes, and much progress has been made (e.g., Qiu et al., 1981a, b, 1984, 2005; Xie, 1984; Sun, 1985; Wang, 1987; Chen, 1988; Song and Xia, 1988; Lin et al., 1999; Alexeeva and Erbajeva, 2000; Li and Zhou, 2001; Sun and Liu, 2001; Liu et al., 2004; Yang et al., 2004). Major discoveries on Quaternary stratigraphy, sedimentary environments, flora and fauna, and the evolution of river channels and formation of plains and sand lands and paleoclimate in Northeast China Plain were achieved in the early 1990s (e.g., Jiang, 1991; Li, 1991a, b; Qiu, 1991a, b; Qiu et al., 1991; Song, 1991; Zeng, 1991; Zhang and Xie, 1991).

Since the 1960s, due to the needs for economic development and engineering construction, the Chinese Academy of Sciences (CAS) and governmental railway, forestry, and hydraulic engineering departments conducted many thematic and comprehensive investigations on permafrost in Northeast China. Although more in-depth understanding has been obtained on the distributive features, regional zones, and their relationships with climatic and environmental variables, investigations on the evolution of permafrost have been limited (e.g., Guo and Li, 1981; Guo et al., 1989; Peng and Cheng, 1990; Lu et al., 1993; Tong, 1993; Zhou et al., 2000; Jin et al., 2006, 2007b, 2011; Yang et al., 2009, 2015; Yang and Jin, 2010). On the basis of previous studies and recent progress, this paper reviews the evolutionary history of permafrost in Northeast China since the Late Pleistocene in order to provide a baseline for further investigations.

2 Study region 2.1 Natural environment

Northeast China extends from Mo'he (52°58'N), Heilongjiang Province in the north to Dalian (39°30'N), Liaoning Province on the south coast, a north-south distance of about 1, 700 km; and from Manzhouli (126°E), IMAR in the west to near Fuyuan-Boli (>135°E), Heilongjiang Province, a west-east span of about 1, 300~1, 400 km (Figure 1). Its general topography is characterized by higher mountains and hills in the eastern and western parts and lower river plains in the central part. From north to south a series of plains, includingthe Songhuajiang-Nenjiang (Songnen) Plain, the West Liao'he Plain, and the Coastal Plain, form the Songliao Plain, with elevations at 200~250 m a.s.l. In the western flank, the Da (Great) Xing'anling (Hinggan) Mountains stretch NE-NNE for about 1, 400 km at general elevations of 500~600 m a.s.l. in the north, 1, 300~1, 400 m a.s.l. in the middle, and as high as 2, 029 m a.s.l. at the Huanggangliang Mountains in the south. The Da Xing'anling Mountains are asymmetrical, with gentler western slopes and steeper eastern slopes and deeply-incised valleys, where many major rivers, such as the Nenjiang and West Liao'he rivers, originate and a great deal of weathered and eroded debris is carried onto the plains. In eastern Northeast China, the Xiao (Small) Xing'anling Mountains extend in the northwestern direction at elevations of 400~600 m a.s.l., with gentle slopes and truncated rivers. A series of mountains, including the Lao'yeling, Zhangguangcailing, Changbaishan, and Qianshan mountains, cascade to the south. The middle hilly part of the Xiao Xing'anling Mountains (the Changbai and Lao'yeling mountains) consists of volcanic rocks and gives birth to the headwaters of the Songhuajiang and Wusulijiang rivers, with elevations from 500~600 to 2, 200 m a.s.l. In the northeastern corner, i.e., the transition zone between the Xiao Xing'anling and Lao'yeling mountains, the Sanjiang (Heilongjiang, Wusulijiang and Songhuajiang) Plain is found at lower elevations of 50~150 m a.s.l.(Figure 1).

Figure 1 Topography and rivers in Northeast China (Data of administrative boundaries are from the National Geomatics Center of China. Same as below)

The climate of Northeast China is influenced by temperate continental and Pacific warm-wet monsoonal climates. In the northern part, the rigid continental monsoonal climate dominates with long, cold and dry winters and short, moist and warm summers; in the southern part, however, the temperate climate is characterized by warmer and wetter monsoonal influences. From southeast to northwest, MAAT lowers from 8~11 ℃ (Dalian area) to 3~5 ℃ (Harbin to Qiqi'har region), and further down to about −4.5 ℃ to −4.0 ℃ (northernmost Mo'he-Mangui region). The annual precipitation declines from 700~1, 000 mm in the southeast to 200~300 mm in the northwest. The summer precipitation accounts for about 70%~90% of the annual total, and the rainy and wet season to a certain degree suppresses the high summer temperatures. Therefore, the MAAT in northern Northeast China is about 3~4 ℃ lower than its counterparts at similar latitudes in Mongolia and the Russian Far East. This low MAAT provides crucial conditions for formation and preservation of permafrost. As a result, the southern limit of the Eurasian permafrost zones protrudes southwards into the Xing'anling Mountains.

Due to the control of the Siberia High in winter, northern Northeast China is under the influences of the winter atmospheric temperature inversion (WATI). The WATI is centered in the southern Yakutian Mountains in Eastern Siberia, and weakens and thins southwards. The WATI is about 500~1, 000 m in thickness in Northeast China, with an inversion gradient of about 10 ℃/km at Mo'he and 8 ℃/km at Nenjiang, Heilongjiang Province (China Institute of Meteorological Sciences, 1991). The presence of extensive and stable WATI exerts great influences on permafrost distribution in Northeast China and adjacent regions.

In the Xing'anling Mountains, ground surfaces are generally covered by dense forests, shrubs, wetlands, and peat layers. Dense forest coverage and moss-peat layers substantially reduce the insolation onto the ground surface and into the soils, retarding and attenuating ground warming and cooling. Due to the large difference of thermal conductivity of frozen and thawed ice-rich peat and long (7 to 9 months), cold winters, and resultant large thermal and surface offsets (about 2.2~5.5 ℃) (Chang et al., 2015), permafrost is well preserved in natural environments in the Xing'anling Mountains even under a warming climate, but it is more sensitive to human disturbances (Jin et al., 2007b).

2.2 Features of permafrost and periglacial environment

Permafrost in Northeast China can be divided into two types: latitudinal permafrost in the north and mountain permafrost on the Huanggangliang and Changbaishan mountains, with a lower limit of permafrost at about 1, 500~1, 800 m a.s.l.(Zhou et al., 2000). However, permafrost degradation may have already substantially reduced the areal extent of permafrost in Northeast China, by about 10, 000 km2, during the last 40 years, and further northward retreat of permafrost is anticipated (e.g., Jin et al., 2007b; Wei et al., 2011).

As the southeastern margin of the Eurasian permafrost zone, the southern limit of permafrost (SLP) in the Xing'anling Mountains sways between about 46°58'N and 48°00'N, which correlates with the MAAT isotherms of about −1 ℃ (west), 0 ℃ (middle), and +1 ℃ (east) (Figure 2). From the SLP northwards, the areal extent of permafrost increases from about 5%~20% to 60%~75%, and the MAGT, generally measured at the depth of zero annual amplitude (about 10~25 m), lowers from +0.5 ~ −0.5 ℃ to −2.5 ~ −1.5 ℃, occasionally to −4.2 ℃. The permafrost thickens from a few meters to several tens of meters, to about 50~70 m, occasionally to more than 100~120 m, and it changes from isolated patches to sporadic, discontinuous, and continuous permafrost.

Figure 2 Latest map for distribution of permafrost in Northeast China in the 2000(refer to Wang, 2006)

Distribution and other features of permafrost in the Xing'anling Mountains are controlled by the regional climate and are influenced by local variables, such as local atmospheric circulations (e.g., WATI), geology, and geography. Latitudinal climate zonation controls permafrost features, displaying evident regional zonation of permafrost. In the northern part, permafrost is well developed in areas densely covered by sedges (Carex tato) and mosses and with thicker (8~10 m) Quaternary sediments, generally in intermontane basins and wetlands and on lower river terraces, with thicknesses of permafrost at 60~80 m, or even greater than 100~120 m. On the other hand, on sunny or half-sunny slopes with thin coverage of forest and shrubs and thin Quaternary deposits, permafrost is poorly developed, with thicknesses of permafrost at about 20~30 m or less, or absent. On sun-shaded (shadowy) and half-shadowy slopes, although the thickness of Quaternary deposits is similar to that on the half-sunny slopes, vegetation is denser and with less insolation, so permafrost development is between that in intermontane basins, wetlands, and lower terraces and that on half-sunny slopes. In the south, permafrost generally occurs in the form of isolated patches or sporadic permafrost; the latter generally occurs in wetlands and on flood plains and lower terraces. In the 1970s, sporadic permafrost in Northeast China had an areal extent of 5%~30%. However, due to climate warming and increasing human activities, permafrost has degraded extensively (Jin et al., 2006, 2007b; Luo et al., 2014).

2.3 Geological control and influences on permafrost and periglacial environment

Permafrost in Northeast China has been controlled not only by climate but also by the neotectonics and volcanic eruptions, geomorphology, paleo-geography, and changes in river channels. Along the northeast direction there are three secondary geological structures: the Da Xing'anling fold system in the west, the Songliao platform syneclise in the middle, and the Laoyeling platform anticlines in the east. They are divided by deep faults with depression zones in the middle and uplifting zones in both sides. Since the late Paleozoic era, particularly since the activation of the China Platform from the Yanshan Movement, the NE-NNE and NW direction faults and folds were very active. They were accompanied by igneous activities, such as granite intrusions and volcanic eruptions, which further fragmented and disintegrated the abovementioned tectonic structures. These activities formed several lower-order tectonic and geological structure units. After the Tertiary period, because of the Himalaya Movement, the inherited palaeo-highs and -lows were accompanied by many basalt eruptions and igneous activities. Since the Quaternary period, modern landscapes have been shaped. A series of geological events affected the later formation and development of permafrost in Northeast China.

First of all, since the Tertiary, differential upwarping and subsidence of fault-blocks were accompanied by frequent igneous activities and basalt eruptions, particularly remarkable in the Xiao Xing'anling Mountains and eastern mountains. Nine basalt eruptions occurred in the Zhangguangcailing, Laoyeling and Changbaishan mountains since the Pliocene (Xu and Fan, 2015). Early eruptions were more extensive, mainly along major faults and for several hundreds of kilometers, forming the main ridge of the Zhangguangcailing and Laoyeling mountains. Although volcanic activities have been declining in intensity and scale, they have never ceased. The Lao'heishan and Huoshaoshan mountains (48°20'N, 126°30'E) were active and erupted as late as in 1720, and the Changbaishan Mountains erupted during the period of 18 to 4 ka before present, and in AD 1018, 1124, 1199-1200, 1265, 1373, 1401, 1573, 1668, 1702, and 1903(Liu et al., 2004). They have resulted in numerous hot springs, such as the Tianchi and Wudalianchi hot springs in the middle Da Xing'anling Mountains and those in the Changbaishan Mountains, and high anomalies in regional geothermal gradients. This interrupts the areal continuity of permafrost distribution since the Late Pleistocene, and still has a profound impact on the present distribution and thermal states of permafrost. For example, in Figure 2, in comparison with the western and eastern sides of the Xing'anling Mountains, the SLP shrinks northward by about 1°N (about 100 km) on the Songnen Plain in the middle. This can be attributed to:(1) the Songnen Plain is lower in elevation and thus higher in MAAT, and (2) affected by the Wudalianchi hot springs, the northern Songnen Plain has a higher geothermal gradient and thus more rapid permafrost degradation since the Last Glaciation Maximum (LGM).

In addition, since the end of the Tertiary to the early Quaternary, except for the Songliao and Sanjiang plains, the Xing'anling Mountains and eastern mountainous regions have been slowly and differentially uplifting and are subject to long-term denudation and leveling. Therefore, loose deposits are either very thin or absent. Because of the good conductivity of bedrocks, they cool and warm rapidly. For example, the Hola'he Basin (52°27'N-52°43'N), with an areal extent of about 60 km2 and consisting of mainly Mesozoic coal-bearing strata, is one of the subsidence troughs in northern Da Xing'anling Mountains. In the 1980s, the basin was investigated in detail and systematically assessed for coal measures and permafrost engineering geology. The results indicate that the permafrost thins and warms from the center of the basin to the peripheries, and vanishes (becomes talik zones) at the upper mountains and hills, hilltops and margins, such as in the northeastern and northwestern parts. Permafrost thickness varies from about 100~120 m in the center, to about 0~40 m on the peripheries, and then 15~20 m at the upper hills or piedmonts. In the same radial directions, MAGT warms from −2.9 ℃ in the center to about −0.5 ℃ on the margins (Guo et al., 1989). The formation, MAGT, and thickness of the permafrost are controlled or influenced by geology, lithology, surface paludification, and groundwater activities. Because of differential upwarping since the formation of the Hola'he Basin, the periphery mountains have been subject to denudation and weathering, resulting in well-developed crevices and fissures and thin (2~3 m) loose deposits on the hills. This is favorable for rainfall and melt-water infiltration and thus the resultant intense heat convection is adverse for the preservation of permafrost. The basin is in the subsidence zone, with more Quaternary deposits 6~10 m in thickness, gentle or flat topography, and poor drainage and resultant paludification, peat deposits, and denser vegetation. Therefore, permafrost is better formed and preserved in the basin.

Furthermore, the Liao'he, Songnen, and Sanjiang rivers basins were silted up in the Tertiary and Quaternary periods. Due to the upwarping of periphery mountains and subsidence of plains, these river channels have shifted numerous times; the present occurrences of numerous lakes and wetlands on the Sanjiang Plain and the northwestern Nenjiang Plain also resulted from river channel shifts (Qiu, 1991a). These processes further complicated the horizontal and vertical distribution of permafrost in Northeast China.

3 Evidence for past permafrost and periglacial environment 3.1 Direct evidence

Extensive occurrences of inactive ice wedges and ice wedge pseudomorphs, sand and soil wedges, pingo scars, cryoturbations (involutions), and many other periglacial remains are regarded as reliable direct evidence for past permafrost (e.g., Romanovskii, 1977; Washburn, 1980; French, 1999, 2007; Murton and Kolstrup, 2003; Jin et al., 2007a; Zhao et al., 2013; French et al., 2014; Vandenberghe et al., 2014). Extensive occurrences of these indicators have been identified for past permafrost in North and Northeast China during the past six decades, but with only limited research publications, mainly in Chinese (e.g., Cui and Xie, 1985; Xie, 1984; Sun et al., 1985; Jia et al., 1987; Song and Xia, 1988; Xu et al., 1989; Peng and Cheng, 1990; Li, 1991a; Tong, 1993; Yang et al., 1995, 2004, 2009, 2015; Owen et al., 1997; Vandenberghe et al., 2004, 2015; Jin et al., 2006, 2011; Yang and Jin, 2010). In this section, the direct evidences for past permafrost are reviewed in categories for reconstruction of permafrost evolution in Northeast China since the Late Pleistocene.

3.1.1 Cryogenic wedge structures

1) Inactive ice wedges at Wuma in northern Da Xing'anling Mountains

In June 1990, Tong et al.(1993)discovered seven ice wedges on the first terrace (52°45'N, 120°45'E; 300~400 m a.s.l.) of the Yilijiqi River, a second-order tributary of the Erguna River emptying to the Heilongjiang-Amur River in the northern Da Xing'anling Mountains. They were about 2 m in height, 1.0~3.3 m in top width, and the wedge tops were buried 1.6~2.0 m under the ground surface (Figure 3). The ground ice was pure and transparent, with vertical lamina and oblong air bubbles 1~2 mm in size. The hosting sediments were inter-bedded, ice-rich gravelly sand and silt, upturned at the contacts with the ice wedges (Figures 4 and 5). Layered ground ice 4~5 cm in thickness and 40%~50% in volumetric ice content were identified in the ambient strata.

Figure 3 Inactive ice wedges exposed by river-cut at the knee of the first terrace (52°58'N, 120°45'E) of the Yilijiqi-Wuma River in the northern Da Xing'anling Mountains (courtesy of Prof. Emeritus BoLiang Tong with the Cold and Arid Regions Environmental and Engineering Research Institute (CAREERI), Chinese Academy of Sciences (CAS), Lanzhou, China)
Figure 4 Closer look at cryostructures of one of the Wuma inactive ice wedges
Figure 5 Schematic diagram of Wuma ice wedge structure (modified from Tong, 1993).

Thaw concaves and refreezing traces were found on the wedge tops. The concaves were 15~55 cm in depth, 30~55 cm in width, and were filled with poorly graded gravelly sands. The presence of overlying humic soils and the absence of polygonal cracks or troughs indicate inactive ice wedges. The 14C-dating suggests an age of 14, 475±390 ~ 10, 668±75 a for the overlying and ambient sediments (Figure 5). Romanovskii (1977)suggested an MAGT no warmer than −6 ℃ for ice wedge development in sandy soils. Wuma now has an MAGT of about −1.5 ℃ to −1.0 ℃ (4.5~5.0 ℃ warmer than for the period of ice wedge formation). Local experiences indicate a difference of MAGT-MAAT at 3.8~4.3 ℃ (Guo et al., 1981). It should thus be possible that the MAAT was no warmer than about −10 ℃ to −9 ℃ in Wuma during the ice wedge formation period. On the basis of studies on ice wedges in Alaska, the MAAT ranges from −8 ℃ to −6 ℃ in the marginal areas of ice wedge distribution, and is about −12 ℃ in the central area (Péwé, 1964, 1966b). Although the Wuma region was not in the center of Eurasian permafrost zone during the Late Pleistocene, it should have been in the margin of the central permafrost zone. Thus, it should be reasonable to estimate an MAAT of no warmer than about −10 ℃ to −9 ℃ for the formation or activity of Wuma ice wedges.

2) Inactive ice wedges at Yituli'he in northern Da Xing'anling Mountains

In 1983, ice wedges were discovered on the first terrace and high floodplain (50°32'N, 121°29'E; 730 m a.s.l.) of the Yituli'he River in the northern Da Xing'anling Mountains (Jia et al., 1987; Peng and Cheng, 1990). The wedge tops were buried about 0.9~1.0 m in depth or about 20 cm under the permafrost table. The top widths of the wedges were about 10~40 cm and their exposed height was 0.9~1.5 m. The ice was transparent, pure, and milky-colored with vertical lamina. The topsoil was a peat layer of about 30~40 cm in thickness, underlain by humic soils about 50~60 cm in thickness. The hosting materials were grayish-brown sand clay and peat with an AMS-14C dating of 4.5 to 2.4 ka.

At a depth of 40 cm to the ice wedge top (about 0.9 m in depth), some secondary cryostructures of 45 cm in length, such as layered segregation ice, were visible, indicating the degradation history of the ice wedges. Therefore, these ice wedges were formed earlier than this dating, i.e., the Neoglaciation in the Late Holocene. Romanovskii (1977)suggested an MAGT of about −4 ℃ to −2 ℃ for the formation of ice wedges in peat or sandy clays. A difference of MAGT-MAAT at 3.5~4.0 ℃ was observed in the region today (Chang et al., 2015). Thus, when these ice wedges were formed, the MAAT should have been about −7.5 ℃ to −6.0 ℃, or about 2.0~3.5 ℃ colder than today (−4.0 ℃).

Recently, one of the ice wedges in Yituli'he was excavated (Figure 6). The ice wedge was selected after a trench survey for polygonal nets in three dimensions. Pollen, and oxygen and hydrogen isotopes of wedge ice, together with the validation for climate proxy, were further analyzed (Yang et al., 2009, 2015; Yang and Jin, 2010). Isotopic compositions were validated for evaluating the paleo-temperature changes, and three cooling periods (in comparison with MAAT of the end of the 1990s) of about 2.1 ℃ at 2.8 ka, 1.1 ℃ at 2.3 ka, and 1.3 ℃ at 1.9 ka were identified. The pollen assemblages suggested a colder and wetter period after 6 ka when the Yituli'he ice wedges formed and were active (Yang et al., 2009). Pollen records, AMS-14C dating, and depository strata all indicated a formation period of the ice wedge at about 3.3 to 1.6 ka (Yang et al., 2009, 2015), when the SLP advanced southwards by about 2°N (200 km).

Figure 6 Vertical cryostructure and photo of the Yituli'he pit profile.

3) Soil wedge at Wu'erxun, Xin Ba'erhu (Xin Barag) Right (West) Banner on the Hulun Buir High Plain

This 2-m-high exposed profile of soil wedges (48°25'N, 117°34'E; 556 m a.s.l.) by a roadcut was at Post Km 181 along Provincial Highway S203 at the knee of the first terrace in the Wu'erxun Valley east of Xin Barag Right (West) Banner on the Hulun Buir High Plain. Soil wedges, with top widths at 0.40~0.45 m, exposed heights of about 1.0 m, and irregular wedge surfaces, were identified in the lower soil strata of grayish-yellowish fine sands (Figure 7). Fallen blocks about 15 cm in size were visible in the overlying sand layer. The upper wedge consisted of grayish-black paleosols and aeolian sands with cryoturbations. The lithology of the wedge soil was similar to that of the overlying black soil layer.

Figure 7 Schematic diagram of soil wedge at Wu'erxun, Xin Barag Right (West) Banner, eastern IMAR.

4) Sand wedges (veins) in Eastern (Left) Xinba'erhu (Xin Barag) Banner on the Hulun Buir High Plain, Eastern IMAR

Sand wedges were found on a sand dune (48°16'N, 118°12'E; 592 m a.s.l.) on a gentle piedmont slope along Provincial Highway S203 40 km east of Eastern (Left) Xinba'erhu (Xin Barag) Banner on the Hulun Buir High Plain, Eastern IMAR. Sand veins were evident at the cross sections of sand pits. The soil profile consists of three layers of aeolian sands (Figure 8a) (Jin et al., 2011).

Figure 8 Sand veins at the eastern side of the Eastern (Left) Xin Barag (Ba'erhu) Banner.

Bulldozers cut eastern and western cross sections, exposing sand veins (Figures 8a and 8b), with upper widths at 0.6~0.7 m, vein widths generally at 3~4 cm, and vein lengths at 2.0~2.1 m extending into the bottom grayish sands. Vertical lamina in the eastern section were not as clear as those in the western side. In the western cross section, sand wedges (Figures 9a and 9b) were 15~20 cm wide on the top and about 1.1 m high, extending into the top layer of brownish-yellow sands, with many polygonal cracks 6~10 cm in width (Figure 9c). Vertical lamina in the sand wedges were evident, and the sand wedges consisted of mainly quartz grains.

Figure 9 Sand wedge and veins at the western side of the Eastern Xin Barag Banner

5) Soil wedges at Hui'he on Southern Hulun Buir High Plain

Soil wedges had a vertical profile naturally exposed by a river cut in the Hui'he Forest Farm area (48°04'N, 119°38'E; 779 m a.s.l.) on the southern Hulun Buir High Plain (Figures 10a and 10b).

Figure 10 Soil wedges in the Hui'he Forest Farm area. ① pedogenetic black fine sands; ②grayish-white fine aeolian sands, with some black soil blocks; ③ brownish-yellow sands mixed with soils; ④ pedogenetic black fine sands, with some sand blocks; and ⑤ grayish-white fine aeolian sands

There were many soil wedges, including two that were 2.8~3.0 m apart. On the left side, the wedge top width was 0.2 m and with a height of 1.0 m. The wedge extended downwards into brownish-yellow sand and soil mixtures in which cryoturbations were extensively observed on the bottom and at both sides. The left and right wedge surfaces were irregular, with evident falling structures. The left wedge was 0.25 m in top width and 0.9 m high.

6) Sand wedges south of the Ewenki Banner on central Hulun Buir High Plain

Sand wedges were encountered in a soil pit on a hill (48°55'N, 119°45'E; 820 m a.s.l.) at the western side of Provincial Highway S202, near Ewenki Banner 18~19 km south of Hailar (Hulun Buir) City on the central Hulun Buir High Plain. The exposed height of the profile was 2.4~2.5 m (Figure 11a). Two sand wedges were 2.5~3.0 m apart. The top width of the one at the left side was 0.25~0.3 m, and its height was about 1.3 m; the right-side wedge was 0.20~0.25 m in top width and 1.0~1.1 m in height. Sand wedges extended down to the blackish-grey silty loams, with irregular wedge surfaces. The filled materials were similar to the second overlying layer of yellowish silty sands (Figure 11b).

Figure 11 Sand wedges west of Provincial Highway S202 14 km south of Ewenki Banner. ① blackish-gray silts and sands; ② yellowish silts and fine sands; and ③ blackish-gray silts and sands

7) Sand wedges at Tianchi, A'ershan in southern Da Xing'anling Mountains

In the vicinity of the Tianchi Forest Farm (47°25'N, 119°40'E; 1, 200~1, 300 m a.s.l.), A'ershan Forest Bureau in the southern Da Xing'anling Mountains, sand wedges were found in sand and gravel layers from weathered basalt with good bedding overlying the Wuchagou basalt (N2 to Q1) (Figure 12). Above the sand wedges were Da'heigou basalt (Q31)4~5 m in thickness. The wedges were 0.7~0.8 m on the top and 2.5~2.8 m in height, and were filled with poorly graded basalt sands and gravels as large as boulders (10~20 cm) and as small as coarse sands. Inferred from the overlying basalt layer (Q31), these sand wedges should have been formed in the Middle Pleistocene cold periods (Guo and Li, 1981).

Figure 12 Sand wedge in Tianchi Forest Farm, A'ershan Forest Bureau, southern Da Xing'anling Mountains (modified from Guo and Li, 1981). ① black soil layer 1.0 m in thickness; ② Da'heigou basalt (Q13); ③ sands and gravels (basalt debris and sands)2.0~3.0 m in thickness; ④ sand wedges; ⑤ basalt sand and gravels, 15 m in thickness; and ⑥ Wuchagou basalt (N2 to Q1)

8) Soil wedges in Da'jiagou, De'hui County, Jilin and Sand wedges at Kuntouling, Ao'han Banner, IMAR

On a gentle water divide near Da'jiagou, De'hui, Jilin (44°38'N, 125°40'E; 250~300 m a.s.l.), soil wedges were found in the Harbin group (Q31) (hosting sediments) of the Late Pleistocene consisting of brownish-yellow and grayish-brown loess-like sandy clay and sandy loams (Sun, 1981). The wedges were overlain by Guxiangtun strata (Q32) consisting of yellow medium and fine sands. The upper wedge deposits were similar to the overlying soils, and the middle and lower wedge parts were filled with a grayish-brown sand layer and Guxiangtun-period (Q32) yellow sands. The top widths were 0.5-0.8 m and the heights were 3.0-5.5 m (Figure 13).

Figure 13 Soil wedge diagram at Dajiagou, De'hui County, Jilin Province (modified from Sun, 1981). ① black soil (Q4); ② yellow sands (Q32); ③ blackish-brown and grayish-brown, brownish-yellow sandy clay; ④brownish-yellow sandy loams; and ⑤ wedge-filled blackish-brown sandy clay and yellow sands

In the Kuntouling (42°20'N, 119°50'E; 500~600 m a.s.l.), Ao'han Banner, IMAR in the southern Da Xing'anling Mountains, sand wedges were found in the Harbin group (Q13) yellow sandy loams (Guo and Li, 1981; Sun, 1981). The wedges were infilled with medium and fine sands, and overlain by Guxiangtun strata (Q32) medium and fine sands. Wedges were 0.2~0.3 m wide on the top and 2~3 m in height.

The abovementioned two groups of soil and sand wedges, given their smooth wedge surfaces and the absence of infilled blocks in the wedges and visible cryoturbations in the hosting sediments, should be primary sand and soil wedges. To form wedges in sandy loams and clays, the MAAT should be about −3.5 ℃ to −5.0 ℃ (Romanovskii, 1977). Therefore, when these sand and soil wedges were formed, the MAAT was lower than today (8~10 ℃ at De'hui and 9.5~10 ℃ at Ao'han Banner). According to the overlying strata age (Q32), they should have been formed in the Late Pleistocene cold periods, or in the Guxiangtun cold period (Sun, 1981).

9) Soil and sand wedges in other places in Northeast China

In addition to those wedge structures mentioned above, many soil and sand wedges were identified to the north of 42°N in Northeast China, such as at Huangshan of Harbin City, Heilongjiang Province; Dun'hua (47°32'N, 122°52'E) and Panshi (42°50'N, 126°02'E), Nianzishan (47°32'N, 122°52'E) of Jilin in the eastern mountainous regions of Northeast China, and at Chahayang (47°56'N, 123°30'E), Gannan County, Heilongjiang in the downstream of the Nuomin River in the central Da Xing'anling Mountains (Guo and Li, 1981; Sun, 1981; Cui and Xie, 1984). In Pingtai and Pinganzhen of Baicheng (43°35'N, 122°41'E; 173 m), Jilin Province, frost cracks several tens of centimeters in width and about 1 m in depth were filled with yellow sands, and polygons were extensively observed on the ground surfaces.

3.1.2 Pingo scars and pingo-lake peatlands

Pingo scars and pingo-lake peatlands were mainly found on the first and second terraces of the Heilongjiang and Wusulijiang rivers on the Sanjiang Plain (47°20'N-48°30'N, 133°E-135°E; 34~60 m a.s l.), presently with an MAAT of about 1.5 ℃ and annual precipitation of 600~650 mm (Song and Xia, 1988). They generally had round or oblong shapes with sizes ranging from 100 to 300 m and depths of about 5 m. The pingo-thawed lakes generally had water with depths of 0.2 to 0.5 m; the lake bottom peat was generally 1 to 3 m thick and was underlain by grey sandy clay of the Guxiangtun group. At the periphery of the lake there was a fortress-like dike about 1.5~2.0 m in height and grown with birch (Betula spp.) stands. The bottom peat at a depth of 1.5~1.6 m had a 14C-dating of 9, 300±100 a; the transit layer between the peat layer and the Guxiangtun gray sandy clay was dated at 10, 295±305 a (Song and Xia, 1988). Therefore, the pingo lake was formed by thawing pingo at the end of the Late Pleistocene and Early Holocene, and the peat gradually silted up.

3.1.3 Cryoturbations (involutions)

1) Cryoturbations near Ping'an town, Baichengzi, Jilin Province

Cryoturbations (involutions) near Ping'an town (43°35'N, 122°41'E; 173 m a.s.l.), Baichengzi, Jilin Province were found in the contact zone between Early Pleistocene (Q1) grayish-white sandy gravels and Middle Pleistocene (Q2) brownish-red sandy gravels. Here, Q2 deposits were at a depth of about 2 m and the overlying black soils were not disturbed (Figure 14). It was thus estimated that after the formation of the Q2 strata, they were exposed to denudation under the Late Pleistocene cold climate and subjected to frost-thaw processes as deep as 2.0 m. They churned and upturned the underlying Q1 strata into cryoturbations. However, the overlying Q4 black soils were not disturbed. Therefore, these cryoturbation structures were formed during the Guxiangdun cold period in the Late Pleistocene (Guo and Li, 1981; Lin et al., 1999).

Figure 14 Profile of cryoturbations near Ping'an town, Baichengzi, Jilin Province (modified from Guo and Li, 1981. ① grayish-black soils (Q4); ② brownish-red sandy clays with gravels (Q2); ③ brownish-yellow sandy gravels (Q2); ④ grayish-white sandy clays (Q1); and ⑤grayish-white medium and coarse sands with gravels (Q1)

Cryoturbations were also identified in Guxiangtun group strata at Xiliao'he (43°24'N, 122°13.5'E; 186 m a.s.l.), Shuangliao (43°45'N, 123°30'E; 132 m a.s.l.), and Kailu (43°55'N, 120°20'E; 240 m a.s.l.), and in Holocene black soils at Baiquan (47°55'N, 126°20'E; 250 m a.s.l.) and Mingshui (47°20'N, 125°57'E, 240 m a.s.l.) (Guo and Li, 1981; Qiu et al., 1981a, b; Sun, 1981).

2) Cryoturbations east of Zhalinuor (Jalainur) near Hailar (Hulun Buir)

Cryoturbations were identified in a strip coal mine profile (49°27'N, 117°43E; 554 m a.s.l.) east of Zhalinuor (Jalainur) near Hailar (Hulun Buir) in eastern IMAR, to the north of the present SLP. However, the cryoturbations were evidently in a permafrost-free area, with the maximum depth of frost penetration at 1.5~1.7 m. The cryoturbations churned blackish-grey humic sandy clay and silty sandy loams, and fine sands as deep as 2.5 to 2.8 m (Figure 15). Because the burial depth of the cryoturbations was greater than the possible penetration depth of frost action at present, they were not formed in the modern climate. On the basis of relative strata and carbon dating, these cryoturbations were formed during the Neoglaciation of the Late Holocene.

Figure 15 Cryoturbations at a strip coal mine east of Zhalinuor (Jalainur) near Hailar (Hulun Buir) (modified from Pei, 1957; Guo and Li, 1981; Li et al., 1982). ① grayish-black sandy loam (14C-dating at 1, 450±140 a); ② blackish-grey humic sandy clay; ③grayish-white and grayish-black interbedded silty clayey sands and fine sands (3, 010±80 a); and ④ blackish-grey humic soils with plant remains (3, 770±70 a)

In addition, cryoturbations were extensively identified during the period from the 1950s to 1980s in the Huangshan group (Q21) near Harbin, in Xiangyangchuan (Q2) and Bie'la'hong (Q3) groups on the Sanjiang Plain, in alluvial sands on the second terrace of the Xilamulun River near Kailu and Shuangliao, and in Holocene black soils near Mingshui and Baiquan near Harbin (e.g., Guo and Li, 1981; Sun, 1981; Cui and Xie, 1984). Some direct evidences for past permafrost since the Late Pleistocene in Northeast China are summarized in Table 1.

3.2 Indirect evidence for evolution of permafrost and periglacial environment - Stratigraphic and pollen records

Since the Late Pleistocene, natural environments and geomorphology have displayed obvious differentiations in various parts of Northeast China: the periglacial environment in the north and in the Changbaishan Mountains differs from areas in the east that are generally under the influences of alluvial and fluvial processes; from areas in the west that have wind action with aeolian and fluvial sedimentary phases; and from alluvial processes and ocean wave actions in the south. Late Pleistocene deposits are usually more than 10 m thick, occasionally up to 30~40 m. Holocene deposits are generally 5~8 m in thickness, occasionally 10~30 m. According to sedimentary sequences, fossil flora and fauna, and 14C-dating, regional strata since the Late Pleistocene can be divided as follows:

1) Xiechen Group (Q34): Peat, fine sands, silt and gravels. Pollen spectra are characteristic of needle- and broad-leaved mixed forests dominated by pine (Pinus), in which birch (Betula) contents are significantly reduced. The 14C-dating indicates a range of 2, 000±340 to 700±90 a.

2) Dagushan Group (Q42): Humus-rich organic soils and lacustrine or marshy phase, peat, fine sands, silt and gravels. The lower part of the pollen spectra is characteristic of broad-leaved forests dominated by Quercus, while in the upper part, Betula is the major pollen type. The age from the bottom up is within 8.0 to 2.5 ka, corresponding to the Mid-Holocene (Q42).

3) Qindeli or Datushan Group (Q14): This is a transition between the Late Pleistocene and the Early Holocene. Lithology consists of silt, fine sands, sands and gravels, humus, and lacustrine or marshy peat. Pollen assemblages are characterized by the highest percentages of Betula(47%~72%) and Alnus(up to 28%).

4) Guxiangtun Group (Q33): Mainly identified on the Sanjiang, Songnen and Liao'he Plains. The upper part is brownish-yellow sandy clay or periglacial loess, and the lower part is silt and fine sand, and gravels, with grey clay or silt, rich in periglacial remains. This group is also rich in mammoth (Mammuthus primigenius) and woolly rhinoceros (Coelodonta antiquitatis) fossils, with 14C-dating back to 40-20 ka, or the Late Pleistocene. The regional vegetation was dark needle-leaved forests, mainly consisting of pine, spruce, and fir. However, the pollen percentages of the trees decline downwards in the profiles, and that of herbs increases, indicating a cooling periglacial environment. On the basis of studies on large herbivorous fauna fossils before and after the LGM, Zhang et al.(2015) posited these alternating successions of vegetation in Northeast China since the last 40 ka: from forest-steppe at 40, 30, and 18 ka, to sparse trees at 34.3, 22, 17, and 11 ka, and to treeless dry steppe at 14 ka, and then basically a tundra environment.

In summary, since the Late Pleistocene, Northeast China has experienced four paleo-vegetation periods:(1) dark needle-leaved forests (Q33);(2) birch forests (Q41);(3) broad-leaved forests (Q42); and (4) needle- and broad-leaved mixed forests (Q43).

The Guxiangtun Group (Q33) was deposited in the Late Pleistocene when dark needle-leaved forests dominated by spruce and fir, indicators of the periglacial environment (Markov and Popov, 1960), prevailed in Northeast China. Also, during that period large-scale ice wedges were formed and were active in the Wuma area along the Heilongjiang River (Tong, 1993), and numerous pingos were formed on the Sanjiang Plain (Song and Xia, 1988). Baulin and Chekhovskii (1973)believed that an MAAT no warmer than about −5 ℃ to −4 ℃ would be required for the formation of pingos. The present-day MAAT is about 2~3 ℃ on the Sanjiang Plain. Therefore, the Sanjiang Plain was then 6~7 ℃ colder.

During the Holocene, Northeast China experienced remarkable climatic and environmental changes. As the predominant species, birch occurred extensively from the Sanjiang Plain to the southernmost Liaodong (East Liao'ning Province) Peninsula in the Early Holocene. Oak and elm, as well as other broad-leaved trees, were the major components in eastern forests during the Early to Middle Holocene. Since ca. 6 ka, pine increased remarkably; at the same time, spruce and fir also expanded (Kong and Du, 1984; Xia, 1988; Liu, 1989; Sun and Weng, 1992; Ren and Zhang, 1998). Accordingly, during the Early to Middle Holocene, winters were slightly colder and summers were slightly warmer than today in Northeast China (Ren and Zhang, 1998).

During the Early Holocene, dark needle-leaved forests gradually moved northwards and eventually out of the southern part of Northeast China; as they climbed upwards into mountains, birch forests prevailed below. However, in comparison with today, spruce and fir were still more extensively distributed in the northern part, as suggested by the presence of spruce and fir pollens in the bottom of the Qindeli Group (Q41). Therefore, a relatively large amount of dark needle-leaved forests were still present in the periglacial environment in the north. During 10-6 ka, precipitation in Northeast China was relatively low, with the driest period occurring during 9-8 ka of the Early to Middle Holocene.

During the Middle Holocene (Q42, 8-5 ka), Northeast China was covered by broad-leaved forests, characterized by extensive occurrence of Quercus mongolica and Ulmus pumila. Some thermal and hygrophilous flora and fauna were unearthed from the Neolithic sites, indicating a warm and humid natural environment, corresponding tothe Holocene Megathermal Period (HMP). In east-central Northeast China, annual precipitation increased by 50~65 mm since 6 ka, as indicated by pollen records, paleosols in sand lands, marsh and peat deposits, sand dunes, loess, and black soils (Ren and Zhang, 1998).

During the Late Holocene (Q43), vegetation was needle- and broad-leaved mixed forests, with spruce and fir dominating and fewer broad-leaved taxa, indicating a humid and cool climate. Between the end of the HMP and the Neoglaciation (about 4, 000 to 1, 600 a) temperatures cooled by about 1~3 ℃, and during the Little Ice Age (LIA, 600 to 200 a) there was additional cooling of about 1~2 ℃; thus, the climate was cooling in general, but with some warming in between. The forest-steppe boundary shifted by certain degrees at a millennial scale in Northeast China. In comparison with the Early Holocene, the forest-steppe boundary at 4 ka shifted westwards by about 130 km (Ren and Zhang, 1998), and this trend continued during the Middle and Late Holocene. Therefore, there has been no forest grown from the central to the western part of the Songliao Plain, and only sparse forests in the eastern part. However, vegetation flourished during 1, 000-660 a, and during the last 3 ka there was a general declining trend for trees while herbs increased.

Table 1 Summary of key evidence for past permafrost Northeast China
Study sites Lat.(N) Long.(E) Elev.(m a.s.l.) Evidence Age (ka) Dating References
Wuma, E IMAR 52°45'120°45'300~400Ice wedges14.48±0.3914CTong et al., 1993
Yituli'he, E IMAR50°32'121°29'730Ice wedges2.6 to 2.314C, AMS-14CJia et al., 1987;Yang et al., 2015
Jalainur, NE IMAR49°27'117°43'554CryoturbationsQ42(Neoglaciation)StrataPu, 1984
Jalainur, NE IMAR49°25'117°47'554Mammoth fossils33.45±2.0014CPu, 1984
Jalainur, NE IMAR49°25'117°47'554Mammoth fossils and cryoturbations11.46±0.2314CCui and Xie, 1985
Hulun Lake, NE IMAR48°58'117°09'554 Mammoth and Rhinoceros fossils 3314CYang et al., 1995
Ewenki Banner, E IMAR48°55'119°45'820Sand wedges13.63±0.05 to8.47±0.04AMS 14CJin et al., 2011
Wu'erxun, Xin Barag R Banner48°25'117°34'556Soil wedges4.65±0.04 to2.29±0.02AMS 14CJin et al., 2011
Xin Barag E Banner, E IMAR48°16'118°12'592Sand wedges14.49±0.08 to8.96±0.06AMS 14CJin et al., 2011
Hui'he on S Hulun Buir Plateau48°04'119°38'779Soil wedges, Cryoturbations7.69±0.04 to1.44±0.02AMS 14CJin et al., 2011
Baiquan, Heilongjiang47°55'126°20'250Cryoturbations in black soils Q4StrataGuo and Li, 1981; Sun, 1981; Qiu et al., 1981b
Tianchi, A'ershan, S Xing'anling Mts.47°25'119°40'1, 200~1, 300Sand wedgesQ32StrataGuo and Li, 1981
Sanjiang Plain47°20'-48°30' 133°-135°34~60 Cryoturbations, pingo scars 10.30±0.30 to9.30±0.10 14CSong and Xia, 1988
Mingshui, Heilongjiang47°20'127°57'260Cryoturbations in black soilsQ4StrataSun, 1981; Qiu et al., 1981b; Song and Xia, 1988
Huangshan, Harbin, Heilongjiang45°47'126°44'240Cryoturbations, ice wedge casts30.00±0.7014C andstrataGuo and Li, 1981; Sun, 1981; Pu, 1984; Cui and Xie, 1984
Zhoujiayoufang, Yushu, Heilongjiang45°06'126°16'240Mammoth fossils29.75±1.2214CPu, 1984
Zhoujiayoufang, Yushu, Heilongjiang45°06'126°16'250Rhinoceros fossils31.80±0.9114CPu, 1984
Dajiagou, Dehui, Jilin44°38'125°40'300Soil wedgesQ33StrataSun, 1981
Kailu and Shuangliao 43°55', 43°45' 120°20', 123°30' 240, 132 CryoturbationsQ33StrataGuo and Li, 1981; Sun, 1981; Cui and Xie, 1984
Ping'an, Baichengzi, Jilin43°35'122°41'173CryoturbationsQ33StrataGuo and Li, 1981; Lin et al., 1999
Xiliao'he43°24'122°13.5'186CryoturbationsQ33StrataGuo and Li, 1981; Sun, 1981; Qiu et al., 1981b
Kuntouling, Ao'han Banner, IMAR42°20'119°50'550Sand wedgesQ33StrataGuo and Li, 1981;Sun, 1981
4 Evolution of permafrost in Northeast China since the Late Pleistocene

Since the Early Quaternary, the Da and Xiao Xing'anling Mountains as well as the eastern hilly regions have been lifting. Between them, there was a giant Songliao lake basin fed by the Liao'he, Tao'er'he, Nenjiang, and Songhuajiang rivers. Prior to the Middle Pleistocene, this basin emptied only northeastwards to the Sanjiang Plain. In the early part of the Late Pleistocene, the Songliao Water Divide between the Songhuajiang and Liao'he rivers was formed in the south-central Songliao Plain, and the giant lake shrank and disintegrated.

On the basis of sedimentary sequences, strata at depths above 80 m are between the Brunhes Normal Polarity Epoch and the Matuyama Reversed Polarity Epoch, and the divide for the Middle and Early Pleistocene lies slightly lower than the boundary. Strata at 43~45 m is on the boundary, with the age of ca. 730-750 ka. The Olduvai Event was identified at the basal clay layer at 67~75 m. Therefore, on the Songliao Plain the clay sediments with a thickness of 20~75 m were formed during the Early to Middle Pleistocene, indicating an extensive scale of the Songliao Lake in the Early Pleistocene. During the Middle Pleistocene, the Da Xing'anling and eastern mountains uplifted and the Songliao lake basin subsided, resulting in an expanding lake basin and lacustrine clay sediment as thick as 30~70 m (Qiu et al., 1984; Xia and Wang, 1991; Lin et al., 1999; Han et al., 2009). The pollen assemblages of these clay layers are dominated by herb pollens, up to 90%, mainly consisting of Artermisia, Chenopodiaceae, and Poaceae. Betula is also well represented through these layers. Overall, this area was covered by steppes with sparse birch, indicative of a cool and dry climate (Xia and Wang, 1991). In the late period of the Early Pleistocene, more trees, mainly Picea and Salix, were distributed in the meadow-steppe and the climate was cool but a little wetter than in the Middle Pleistocene.

In the early part of the Middle Pleistocene, pollen assemblages were dominated by Ephedra, Tamarix, and Chenopodium, with Chenopodiaceae (27%~ 33.6%), suggesting a dry steppe distributed in the region and thus there was a cold and dry climate (Qiu et al., 1984). In the late part of the Middle Pleistocene, the tree pollen increased significantly, especially these from broad-leaved trees such as Betula(up to 36%), Ulmus, Quercus, Juglans, and Tilia. In addition, a certain amount of hygrophilous plants and Pediastrum also occurred. The vegetation type in the area indicates a relatively warm and wet climate. Gravis (1974)discovered pingo scars and other periglacial remains in lacustrine sediments in the northern Hangai Mountains, Mongolia (48°N-50°N) during the start of Early Pleistocene, while permafrost formed above 2, 200 m a.s.l. in the Outer Baikal Mountains. The two regions are adjacent to the Xing'anling Mountains. Thus, it might be possible that the permafrost could have expanded into Northeast China. However, there is no reliable evidence to support this hypothesis that permafrost occurred in Northeast China during the Early Pleistocene.

In the early part of Middle Pleistocene, the Samrovskii Glaciation occurred north of the Heilongjiang-Amur River. Some ice wedges formed during this cold period are still preserved north of 64°N, and ice wedge casts have been found as far south as 56°N. Fotiev (1974)estimated that the SLP of this time reached as far south as Mongolia, and on the basis of ice wedge casts, the southern Outer Baikalia was no warmer than about −3 ℃ to −2 ℃. As mentioned above, sand and gravel wedges in the Tianchi Forest Farm (47°25'N, 119°40'E; 1, 200~1, 300 m a.s.l.) should have been formed during this cold period. This suggests an extensive occurrence of permafrost in the northern part of Northeast China. Where was the SLP then? With respect to the area at elevations of 1, 200~1, 300 m a.s.l., gravel and sand wedges should have been formed under an MAAT of about −7 ℃ to −6 ℃ (Romanovskii, 1977). In the Da Xing'anling Mountains lowered (southwards) down to about 500~600 m a.s.l., the MAAT elapse rate with elevation was about 6.5 ℃/km. Therefore, in comparison with the present, the SLP should have extended southwards by about 2°N to 3°N, or to about 45°N. The Songliao Plain, with elevations at 100~600 m a.s.l. was about 5°N-6°N south of the Russian location with ice wedge casts. Therefore, during the Middle Pleistocene cold period, permafrost was formed north of the Xiao Xing'anling Mountains.

In the later part of the Late Pleistocene, lakes and rivers extensively developed on the Songliao Plain, resulting in alluvial strata in the peripheries and lacustrine clay sediments in the central plain. The pollen records indicate that the vegetation was steppe or meadow-steppe with sparse birch trees. Additionally, fossils of Medvedev rhinoceros (Dicerorhi nusmerckii) were discovered in the deposits of the Middle Pleistocene in Zhoujiayoufang, Jilin Province and Harbin, Heilongjiang Province. In this period, reddish-brown weathering crust extensively developed, with brownish-red paleosols reaching as far north as 43°N-44°N (Liu, 1994). The yellowish-brown paleosols were identified on the third terrace in the Huma area in the Da Xing'anling Mountains (Guo and Li, 1981). They all indicate a warm and moist interglacial climate in the late part of the Middle Pleistocene. Then, ice wedge casts in Central Siberia reached 56°N. On the basis of these occurrences, Fotiev (1974)estimated the SLP of eastern Eurasia (110°E-140°E) was at about 57°N-58°N, i.e., permafrost retreated northwards and out of Northeast China. This conclusion is also backed up by the evidence of warm and moist climate and occurrences of brownish-yellow weathering crust and paleosols in the northern part of Northeast China.

Permafrost advanced and retreated several times under the influences of fluctuating paleo-climates and paleo-environments since the Late Pleistocene. On the basis of these direct and indirect evidences for past permafrost and periglacial environments, the evolutionary history of permafrost can be divided into five stages:(1) Late Pleistocene (Last Glaciation, LG) (65 to 10-8.5 ka); 2) Holocene Megathermal Period (HMP, 8.5-7.0 to 4.0-3.0 ka);(3) Late Holocene cold period (Neoglaciation) (4.0-3.0 to 1.0-0.5 ka);(4) Little Ice Age (LIA, 500 to 100-150 a); and (5) climate warming since the last century.

4.1 Late Pleistocene (Last Glaciation, LG) (65 to 10-8.5 ka)

The Late Pleistocene (ca. 65-60 to 11-8.5 ka) is marked by the Last Glaciation (LG) in East and Northeast China. Paleo-vegetation changes revealed from pollen records at 15 sites indicate three sub-stages of climate fluctuations: cold and dry climate during the early part (ca. 65-40 ka, or LG II, MIS 3b), the moist period of the middle part (MIS3c), and the late part (ca. 20-11 ka, LG III and IV (Younger Dryas, or YD)) (Cui et al., 2011). The period of 20-18 ka corresponds to the Last Glacial Maximum (LGM).

The vegetation compositions varied during the abovementioned climate periods in Northeast China. The northeast (Sanjiang Plain) and the southwest (Lower Liao'he Plain) were influenced by the monsoonal climate and the vegetation was characterized by forests, even during the LGM, consisting of cold-tolerant needle-leaved trees such as spruce, fir, and pine. However, in the interior far from oceans, especially under a very cold and dry continental climate during the LGM, steppes expanded to the Songnen Plain and trees vanished. Xu (1983)and Wang (1987)studied the pollen assemblages at different sites during different time periods in Northeast China and compared them with the modern vegetation; they then deduced the variations of the temperature in different periods of the growing season (as represented by the mean July air temperature). Their results indicate that the LGM was the coldest, and it got warmer southwards from the Sanjiang Plain (7.6 ℃) and Songnen Plain (6.9 ℃) to the Lower Liao'he Plain (5.4 ℃).

Matsuo (1983)discerned a similar cooling in August temperatures in Japan: northern Hokkaido (8~9 ℃), northernmost Honshu (7.7~8.7 ℃), central Honshu (6.8~7.2 ℃), southwest Honshu (6.5 ℃), and Kyushu (5.0~6.0 ℃). During the Late Glacial the region was highly influenced by arctic fronts and the climate was very cold and dry, with many consequences:

1) Glaciations: Inside the Changbaishan volcanoes, and above the water level of Tianchi Lake, six dry cirques of the LGM were identified (Xiao and Hu, 1988; Liu et al., 1990; Zhang et al., 2008). Dating by OSL, K-Ar, TIMS, ESR, and other methods indicates two ice advances at ca. 20 ka (LGM, larger-scale glaciation) and 11 ka (Late Glacial, smaller-scale glaciation).

2) Mammathas-Coelodonta Faunal Complex (MCFC): The Mammoth Faunas emergence cannot be attributed only to the climate cooling during the Neogene-Quaternary period. The formation of the Eurasian MCFC was a result of interacting tectonic, geographical, climatic, ecological, and phylogenetic processes. Kahlke (2014)believed that key environmental factors for controlling the origin and evolution of the MCFC were successive aridification and increasing continentality and rhythmic, prolonged and intensified climate cooling. More than 350 sites have unearthed the MCFC in the Guxiangtun Group of the Late Pleistocene in Northeast China (Zhang, 2008). These fossil sites are on the first and second terraces of rivers on the Songliao Plain; only a few are in the mountains. They are concentrated in a region of 122°E-128°E and 42°N-48°N, and the number with fossil discoveries increases northwards (Qiu et al., 1981b; Jiang, 1991; Dong et al., 1999). This further indicates a larger amplitude of climate cooling in the northern areas of Northeast China, where the climate was better for the survival and flourishing of MCFC. The composition of MCFC in Northeast China was similar to those in Siberia in the Late Pleistocene, but the time of the occurrence was different: In Siberia it was 14C age at 44-33.5 ka and in Northeast China it was 40-15 ka. Due to the extensive snow and ice coverage in chilly Siberia during the LG, the MCFC moved southwards, and these mega-faunal extinctions may have been "staggered" in Northeast China across the Late Quaternary (Turvey et al., 2013).

3) Aeolian deposits: Located on the west-central Songliao Plain, the Horqin Sand Land (118°E-124°31'E, 42°N-46°N) (Western Liao'he Plain and southern and eastern marginal hills) is on the eastern edge of the aeolian and loess deposition zone in North and Northeast China. Fixed, semi-fixed, and drifting sand dunes account respectively for about 10%, 50%~60%, and 25%~30% of the aeolian terrains in the western Horqin Sand Land; fixed and semi-fixed sand dunes prevail in the eastern Horqin Sand Land, and drifting sand dunes are minor. MCFC was also discovered in alluvial-fluvial and aeolian deposits, indicating an age of Late Pleistocene (Zhang et al., 2015). However, the Horqin Sand Land was first formed in the Middle Pleistocene (730-260 ka), and then it was subjected to shrinkage and sand-dune fixing (Qiu, 1989). During the LGM (21-13 ka), under the enhanced winter monsoonal climate, the Mu Us Desert in North China expanded southwards to 36°N-38°N, and deserts and loess deposits covered most of the eastern IMAR down to the lower reaches of the Yellow River basin, and permafrost and sand dunes were extensively developed in Mongolia (Owen et al., 1997; Sun et al., 1998; Yang et al., 2004; Zhao et al., 2013). During the Würm Glaciation (12-6 ka), the Horqin Sand Land expanded significantly to the west-central Songliao Plain and to the Song-Liao Water Divide, reaching its maximum. Yi et al.(2013)believed that four evident stages could be identified for the climate change and resultant fixation-activation of the Horqin Sand Land:(1) Extensive activation under a dry and cold climate during 12-6 ka, and transitory fixation and pedogenesis under a wetting and warming climate (but still quite cold and dry) during 12.0-9.5 ka;(2) basically fixation under a warm and moist climate during 9.5-2.5 ka;(3) several fixation and activation cycles under an alternating cold-dry and warm-moist climate since 2.5 ka, and;(4) evident activation since the last 1.5 ka.

The Songnen Sand Land on the western Songnen Plain was desert and semi-desert prior to 9 ka and experienced four expansion-fixation cycles (Li, 1991b). During 9-7 ka, it was again a semi-arid Artemisia or sparsely-treed Artemisia steppe; since 7 ka, it experienced three arid, two semi-arid, and one semi-humid periods, and in the last millennium it was dry again. Lu et al.(2013)and Zeng et al.(2013)concluded that the Hulun Buir Sand Land underwent dry-wet changes at the millennial scale during the last 165 ka, and those researchers showed the boundaries of sand lands and moving sands during the LGM and HMP. Their results indicate that during the LGM, in comparison with today, the sand lands expanded northward by about 60 km, eastwards by about 50 km, and the LGM desertification areal extent was 22, 337 km2(about 2.7 times of that today), while during the HMP, sand dunes fixation and pedogenesis occurred extensively and the aeolian action was localized.

In summary, during the Late Pleistocene it was cold and dry in Northeast China, as suggested by the extensive advances of mountain glaciers on the Changbaishan Mountains, the presence of aeolian sands and loess, and the MCFC and periglacial pollen records on sand lands and plains. As a result, Eurasian permafrost expanded southwards and invaded into the region, as manifested by ice wedge groups in Wuma in the northern Da Xing'anling Mountains, pingo groups on the Sanjiang Plain, and numerous other direct evidences of permafrost and periglacial remains as indicated in Figure 16. However, the position of SLP during the LGM has been argued among many Chinese geocryolgists (e.g., Yang and Xu, 1980; Guo and Li, 1981; Qiu et al., 1981b; Cui and Xie, 1984; Pu, 1985; Xu et al., 1989; Zhou et al., 2000; Cui et al., 2004; Jin et al., 2006, 2007b).

Figure 16 Periglacial and permafrost remains and inferred changes in the southern limit (SLP) of permafrost in Northeast China since the Late Pleistocene

The MAAT at the Late Pleistocene reconstructed from the inactive ice wedges at Wuma (52°50'N) was about −10 ℃ to −9 ℃ and that from pingo scars on the Sanjiang Plain (47°N-48°N) was about −6 ℃ to −5 ℃. According to a northward elapse rate of MAAT versus northern latitude of about 1 ℃/°N (varying at 0.8~1.1 ℃/°N, equivalent northern latitude) (Zhou et al., 2000; Wei et al., 2011), the 0 ℃ isotherm of MAAT during the LGM should have been at about 42°N (north of Shenyang with a latitude of 41°45'N) on the Songliao Plain. It is generally accepted that permafrost can form under subzero ground surface temperatures, and the present surface offset (difference between the mean annual ground surface temperature (MAGST) and MAAT, i.e., the value of MAGST-MAAT) is largely at 1.5 ℃ (Zhou et al., 2000, 2000; Wei et al., 2011; Chang et al., 2015). Thus, it is evident that the 0 ℃ isotherm of MAAT on the Songliao Plain during the LGM should have been about 41°N-42°N (Horqin Left-Back Banner to Changtu); there the present MAAT isotherm is 7~8 ℃. However, the LGM SLP was irregular due to topographic influences. Most of eastern mountains and hills were already above 500~600 m a.s.l. in elevation, and thus the SLP protruded southwards to the Liaodong Peninsula and connected with the Qianshan Mountains, i.e., surpassing southwards to 41°00'N -41°30'N in eastern Northeast China and 37°N-40°N in northern North China and western Northeast China (Zhou et al., 2000; Zhou et al., 2013). The SLP of the LGM can thus be delineated (Figure 16). However, Cui et al.(2004)believed that the SLP should have been more southerly and reached the Jinzhou Bay of Dalian on the Liaodong Peninsula. Yang and Yan (1958)reported many periglacial remains and deposits in the lower reaches of the Yangtze River, and Yang and Xu (1980)believed that the SLP of LGM in Eastern China should have reached as far as the lower reaches of the Yangtze River basin.

As inferred from the Wuma ice wedges, the paleo-temperatures in the continuous permafrost zone in the northern Da Xing'anling Mountains (to the north of 50°N) at the end of Late Pleistocene should have been about −10 ℃ to −7 ℃. This is similar to the present MAAT at Yakutsk in Eastern Siberia, where the present thickness of permafrost is about 150~250 m. It is possible that during the Late Pleistocene, permafrost was no thinner than 150 m in the northern Da Xing'anling Mountains and was continuous.

In the region to the north of SLP, as delineated in the 1970s, and on the Sanjiang Plain (e.g., Guo et al., 1981), the MAAT was about −10 to −5 ℃ during the Late Pleistocene, where the permafrost was continuous. To the south and until the SLP of the LGM, the MAAT varied from −5 ℃ to −1 ℃ and the permafrost was gradually changed from discontinuous to isolated patches due to local geology and geography. Since the Late Pleistocene, with uplifting eastern mountains and consequent westward shifts of the Songliao Plain subsidence zone, the Nenjiang and Songhuajiang river channels shifted many times, leaving behind clusters of lakes and wetlands. These also further complicated the horizontal and vertical continuity of permafrost distribution. Until the Late Pleistocene, the remaining heat from earlier basalt eruptions along river faults in eastern mountains and later basalt eruptions undoubtedly resulted in relatively higher geothermal anomalies, even in the Late Pleistocene. This may have adversely impacted regional formation of permafrost. Therefore, it is possible that permafrost was thinner with extensive taliks in eastern mountains in comparison with adjacent regions.

4.2 Holocene Megathermal Period (HMP, 8.5-7.0 to 4.0-3.0 ka)

Since the Holocene, the climate in Northeast China shifted from cold-dry to warm-moist, resulting in a series of successions in natural environments. Due to the significant climate warming, birch (Betula) forests developed rapidly and replaced fir forests to become the dominant species; the MCFC moved northwards and then disappeared. During the HMP (8-5 ka), birch forests were replaced by broad-leaved forests dominated by oak (Quercus monolica) and alder (Alnus cremastogyne), which extended southwards to the Sanjiang Plain. The vegetation indicates that the climate during this period was the warmest of the Late Glacial, corresponding to the HMP. Nowadays, these warm, temperate broad-leaved forests are mainly found in Hebei Province and the Shandong Peninsula (36°N-38°N). Thus, the MAAT on the Liaodong Peninsula was 13 ℃ during the HMP, 3~5 ℃ warmer than today (Guiyang Institute of Geochemistry, CAS, 1977; Zheng et al., 1998).

During the HMP, the Horqin Sand Land, the Loess Plateau, and the Mu Us Desert succeeded to steppe or forest-steppe landscape. Sand lands and dunes were fixed or semi-fixed, and inter-dune depressions contained lakes or wetlands, forming peat layers or sandy paleosols (14C-age at 9.5-2.8 ka). Among them, peat that formed during the period of 5.5-4.5 ka contained the least amount of sands, the organic carbon content was as high as >60%, and the deposition rate was 0.44 mm/a. Song and Xia (1988)estimated an aridity (evaporation/precipitation) of less than 1 and annual precipitation as high as 500~700 mm. Since the Late Glacial, climate warming and increased precipitation resulted in snow and ice melting and sea level rise, followed by sea transgressions of the southern coasts in Northeast China. During about 8-3 ka, sea transgression reached maximum, to coasts along the Liaodong Bay and the Yalüjiang Delta plain and coasts (Gao, 1986).

During the HMP, permafrost was greatly affected by the warming and wetting climate. Evidently, permafrost thawed extensively and the SLP retreated substantially northwards. During this period, pingos on the Sanjiang Plain collapsed into pingo lakes, and in the present-day zone of isolated patches of permafrost, permafrost degradation on the lower terraces and floodplains left behind numerous thermokarst depressions. However, permafrost had not completely backed out of Northeast China. It had a significant presence in the northern part with an SLP during the HMP at about 52°N, because:

1) At Amu'er in the northern Da Xing'anling Mountains (52°20'N, 123°11'E; 510 m a.s.l.), analyses on the light and heavy mineral contents indicated that, including the deposits of the HMP, the contents and variety of unstable and relatively stable minerals are low, and those of stable minerals are even less (~5%); in contrast, the rock debris contents are high (70%~95%) (Guo and Li, 1981). This indicates a slight physical weathering process in an arid and cold environment, i.e., it was not so warm and humid in the northern Da Xing'anling Mountains (north of 51°N-52°N) during the HMP.

2) The age of the hosting soils of ice wedges at Yituli'he in the northern Da Xing'anling Mountains is at 4.5-2.4 ka measured by the 14C-dating method, and the ice formation time was close to 2.4 ka (Peng and Cheng, 1990) or 3.3-1.6 ka (Yang and Jin, 2010; Yang et al., 2015). Evidently, there were no ice wedges present in the area during the HMP, but they should have been formed afterwards and the SLP should have retreated northward to about 51°N-52°N.

3) On the basis of pollen records, the MAAT on the Liaodong Peninsula was about 13 ℃ during the HMP. Assuming a northward cooling rate of 1 ℃/°N (MAAT/northern latitude), and from southern Northeast China at 39°N-40°N, the 0 ℃ isotherm of MAAT should have been at about 51°N-52°N. Under this circumstance, permafrost should have degraded but not disappeared completely.

4) In the profile of ice wedges at Wuma (120°45'E, 52°58'N) in the northern Da Xing'anling Mountains, the upturned upper parts of the hosting soils, instead of down-fallen materials, indicate that the ice wedges have been preserved continuously until the present rather than being refrozen, meaning that the SLP was still further south during the HMP. The thawed concaves 15~55 cm in depth on the top of the Wuma ice wedges (Tong, 1993) indicate significant degradation of permafrost in the region, such as certain warming and thinning, similar to the situation of the present-day island permafrost environment.

Given the above evidences and combined with the present southern boundary of the continuous permafrost zone in Northeast China, the SLP during the HMP was delineated (Figure 16), with the southernmost parts at about 51°N.

4.3 Late Holocene cold period (Neoglaciation) (4.0-3.0 to 1.0-0.5 ka)

From 3 ka, broad-leaved forests dominated by oaks, elm, and alder were replaced by needle- and broad-leaved mixed forests, suggesting modern-time climate conditions. The deposits for this period consist of peat or silt in the lower part and brownish and grayish-yellow sandy clay in the upper part. They indicate a cold-wet and warm-dry fluctuating climate; the early part was wetter and the later part was drying. In the meantime, the high sea level during the HMP left behind three low shell dikes. According to Zhu (1972), since the HMP and before 1 ka, China had two major cooling periods; the coldest were at ca. 1, 000 BC and AD 400. Under an 1~3 ℃ cooling climate in the Neoglaciation, the ground was refrozen. In the present northern part of the continuous permafrost zone (i.e., north of the HMP SLP), the old permafrost formed in the Late Pleistocene was cooled and thickened. To the south of the present continuous permafrost zone, new permafrost occurred and ice wedges were formed in Yituli'he (Peng and Cheng, 1990; Zhou et al., 2000; Yang and Jin, 2010; Yang et al., 2015).

In addition, in the vicinity of and to the north of the SLP delineated in the 1970s (Guo and Li, 1981), cryoturbations were identified in the black soil layer (14C-age at 7, 500~2, 500 a) and their equivalents (blackish-grey humic silt and fine sand with a 14C-age at 3, 010±80 a) in many places. This indicates that the distribution of permafrost during the Late Holocene cold period (Neoglaciation) expanded southwards and surpassed the SLP of modern permafrost (ca. 1970s), with better developed permafrost in the zones of island permafrost and isolated patches of permafrost (Figure 16). On the basis of the cooling (1~3 ℃) and cryoturbations in the black soil layer, the island permafrost zone was more extensively distributed than today but it should still be limited to north of 46°N. The SLP of Neoglaciation is thus delineated on the basis of the distribution of localities with cryoturbations in the black soil layer.

4.4 Local Little Ice Age (LIA, 500 to 100-150 a)

According to Zhu (1972), since the Little Ice Age (LIA) to the 20th century there were several climate cooling periods, and the coldest temperatures were at AD 400, 1200, and 1700. The latter two cooling periods were more intensive and persisted until the late 18th century. In particular, the cooling during AD 1650-1700 was more notable and the coldest during the last few hundred years, when the Heilongjiang River had river ice 1 m thicker than today (Zhu, 1972; Gong et al., 1979). The LIA in Northeast China should include the periods of the late 15th century to the early 16th century, the 17th century, and the 19th century, when the decadal (30~50 year average) MAAT departures were greater than −1.0(−0.5)℃ in comparison with the average of the 20th century (Wang and Gong, 2000). Until the 17th to 18th century, permafrost development may have reached a new peak in Northeast China, with well-developed alpine permafrost at elevations above 1, 550~1, 650 m a.s.l. in the Huanggangliang Mountains and above 1, 800~2, 000 m a.s.l. in the Changbaishan Mountains (Figure 16).

4.5 Last century

Since the 1750s, the climate has tended to get warmer; the late Holocene SLP gradually retreated to the SLP delineated in the 1970s (Guo and Li, 1981). During the last 45 years, the SLP has retreated further northwards by 50~200 km (Jin et al., 2006, 2007b) (Figure 16).

In summary, the present-day permafrost layers in the continuous permafrost zone in Northeast China were basically formed by the overlapping of the Late Pleistocene (LGM) and Late Holocene (Neoglaciation, 3-1 ka) permafrost; south of the present continuous permafrost zone, permafrost was developed during the cold periods since the Neoglaciation and has been preserved in favorable localities (Table 2). The above results indicate that the LGM SLP position in Northeast China is more northerly than that discussed in the study by Cui et al.(2004), who estimated an SLP of LGM at 37°N-40°N in North China and western Northeast China, i.e., further south to the Jinzhou Bay of Dalian on the Liaodong Peninsula, instead of north of Shenyang. Our result differs from an earlier but much rougher and bolder estimation by Yang and Yan (1958)and Yang and Xu (1980), that the LGM SLP of eastern China could have reached the lower reaches of the Yangtze River basin.

Table 2 Summary of permafrost evolution in Northeast China since the LG
Stage Start (ka) End (ka) Climate change (ΔMAAT in ℃) SLP Latitude (N) Estimated permafrost extent (in 1, 000 km2)
Min Max West Middle East
LG65.00-60.0011.00-8.50-5.0-8.037°-40°42°-43°41°-41.5°1, 100~1, 000
HMP8.50-7.004.00-3.00+2.0+5.052°51°52°30~25
Neoglaciation 4.00-3.00 1.00-0.50 -1.0 -3.0 ? 46° ? 750~700
LIA 1.00-0.50 0.15-0.10 −0.5 +0.5 41°-40° 45° 46° 650~550
Last century 1970s 0.15-0.10 0.05-0.015 0.00.047°-48°47°-48°47°-48°390~380
1990s0.025-0.0150.015-0.0000.92.247°-50°48°-51°48°-49°290~280
Note: Climate changes in ℃ as indicated by DMAAT in ℃ is in comparison with the 30-year average of MAAT during 1961-1990 for the last 100~150 years. For earlier periods, it is generally in comparison with the average of the 20th century.
5 Summary and conclusions 5.1 Summary

On the basis of sedimentary sequences and permafrost and periglacial remains, combined with pollen records, dating results, and basic features and boundaries of permafrost zones, the periglacial environment in Northeast China since the LGM was reconstructed and divided into five major stages:(1) the Late Pleistocene or Last Glacitation (LP or LG, 65-60 to 10.5-8.5 ka), LGM (21-13 ka) in particular;(2) the Holocene Megathermal Period (HMP, 8.5-7.0 to 4.0-3.0 ka);(3) the Late Holocene cold period (Neoglaciation) (4.0-3.0 to 1.0-0.5 ka); 4) the Little Ice Age (LIA, 500 to 100-150 a); and (5) the last 20th century. By cross-examination of paleo-temperatures inferred from permafrost and periglacial remains and from other climatic and environmental proxies, this paper focuses on the discussion and delineation of SLP positions since the LGM. We offer only very limited discussions on the evolution of permafrost prior to the Late Pleistocene due to limited reliable direct evidence and available understandings of earlier permafrost, largely based on comparisons with those of neighboring regions. These aspects should be further studied and discussed as more data become available.

On the basis of Wuma ice wedges in the northern Da Xing'anling Mountains, which have been preserved since the LGM and further verified by analyses on mineral compositions and pollen assemblages, we estimate that the SLP of the HMP retreated northwards to about 51°N-52°N, from the LGM SLP at roughly 41°00'N-41°30'N (East part) to 43°N (West part) in Northeast China. The permafrost north of 51°N-52°N should be the result of overlapping of permafrost formed during the LGM and Neoglaciation, while that to the south could have only been produced during the Neoglaciation. Our delineation of the LGM SLP position is more northerly than that of earlier conclusions, with rich evidences, which was 37°N-40°N in North China and western Northeast China, i.e., further south to the Jinzhou Bay of Dalian on the Liaodong Peninsula, instead of north of Shenyang. Our result also differs from an even earlier but bolder estimation, that the LGM SLP of eastern China could have reached the lower reaches of the Yangtze River basin, which needs more and further concrete evidence in North and East China.

5.2 Inadequacies

Systematic and interdisciplinary integration of studies on climate and environmental changes and Quaternary geocryology in Northeast China and bordering regions is very important for the reconstruction of the permafrost environment since the LGM. Although studies on past permafrost were active and productive in the 1980s, recently there have not been adequate studies even though new technology and supplementary studies have provided new evidence and tools for geocryology. For example, there has been less attention paid to the climate change during the Early Holocene and the Medieval Warming period in China due to the lack of direct evidence and measurements, making it very difficult to delineate boundaries of permafrost in these periods. Northeast China has even less available historical archives for similar studies.

As mentioned above, there are many inconsistencies in both results and conclusions for past permafrost features, such as the SLP of the LGM. It is thus necessary to calibrate the data, integrating regional and global data at different time periods. In particular, the reliable dating of occurrences of paleo-permafrost has always been a challenge; thus, systematic calibration of earlier dating data would enable cross-examination and inter-comparison of data from Northeast China and bordering regions.

Indirect evidences are important because they are more available, accurate, and systematic for indicating climate, environmental, and permafrost changes, such as pollen and tree-ring records, loess, and lacustrine deposits. In particular, during the warm or interglacial periods, when direct evidence is rare and can more easily be altered later, indirect evidences or proxies play more important roles in reconstructing the past permafrost.

Due to the large difference in thermal inertia for permafrost and glaciers, permafrost takes a longer time to build up and decay in response to climate cooling and warming. As a result, globally and locally, there should be issues in the largest areal and volumetric occurrence of glacial and permafrost/periglacial maxima and minima, such as the local and global Last Glacial Maximum and Minimum and global and local Last Permafrost or Periglacial Maximum and Minimum. Although the LGM, such as during 21±2 ka, could be global, to a certain degree it may not be the Last Permafrost Maximum (LPM) in either North America or on the Qinghai-Tibet Plateau. In the meantime, the least glaciation also may not be the time for the least or maximum occurrence of permafrost.

5.3 Prospects

Permafrost degradation in Northeast China since the 1850s, when the "Open Policy" of the Chinese Qing Dynasty was enacted, is very important for climate, environmental, and geocryological studies. This is not only because much more abundant historical records and civilian observations and records have been preserved and sorted out for scholarly research, but also because climate and environmental changes during the last 150 years have greatly shaped the modern landscapes in Northeast China, where human activities have played increasingly important roles. In particular, large-scale deforestation and cultivation of natural lands and recent rapid urbanization and resources development have reshaped Northeast China, because before the 1850s Northeast China was basically a reserve for royal hunting and recreation. Therefore, Northeast China is a pivotal place for studying changes in nature-human integration during the last 150 years.

Due to the marginal nature of Northeast China in Eurasia permafrost zones and its frequent major climatic and environmental shifts between glacial and interglacial periods, permafrost expanded and retreated numerous times during the Quaternary period, during the Middle and Late Pleistocene in particular. However, because many places rich in periglacial remains are in subsiding basins, most of strata with abundant evidence of past permafrost and periglacial environment are deeply buried, and therefore reliable direct evidences for past permafrost are hard to collect and integrate. However, new technology and data from many disciplines may help reveal the geocryological history prior to the LGM.

Although most glaciologists in China believe only very limited glaciation in high mountains (such as the Changbaishan Mountains) can be found during cold periods (such as the LGM), some scholars have identified more evidences for extensive occurrences of glaciers in the Xing'anling Mountains during the Quaternary periods. Due to the competing nature of permafrost, periglacial, and glacial environments, if these evidences were to be verified they would further complicate the reconstruction of Quaternary permafrost in Northeast China.

Studies on deserts and sand lands have made tremendous progress in North and Northeast China due to extensive concerns and resultant investment in relevant research. In general, direct evidences for past permafrost occurrences are scarce due to the lack of soil moisture in desert or semi-desert environments, but some direct and indirect evidences for geocryological studies have been discovered. They also point to the paradoxical and complex relationships of permafrost and deserts (or sand lands). For example, would permafrost degradation result in land desertification, such as on the former Horqin, Songen, and Hulun Buir grasslands?

At a regional scale (such as Northeast China), vegetation plays an important role in the formation and preservation of permafrost. Due to many major shifts of vegetation since the LGM, such as forestation and deforestation, and desertification and recent urbanization in Northeast China, thermal states and distributive features of permafrost may have been substantially altered due to the destabilizing hydrothermal balances. However, these mechanisms and processes have not yet been adequately understood, even for present situations.

Due to generally low elevations in Northeast China and many major changes in sea level and sea transgressions in the Quaternary, sea level changes may also have greatly affected permafrost in Northeast China. For example, during the LGM the sea level may have lowered as much as 120~150 m, extensively exposing the continental shelves of the Bo'hai and Yellow Seas. In addition, the colder and more-continental climate may have claimed the recently exposed shelves for permafrost. How have these sea-land shifts changed the permafrost distribution?

As discussed earlier, geological structures such as active fault zones and volcanic eruptions have greatly impacted the regional distribution of geothermal gradients and permafrost. Northeast China has been tectonically active and rich in volcanic activities since the Late Pleistocene. How they have affected permafrost evolution? What are the roles of geology in the formation and preservation of the Xing'an-Baikal permafrost? In addition, a strong, extensive, and persistent WATI plays important roles in the formation and preservation of permafrost in Eastern Siberia and the Far East. How has this WATI evolved and how has it affected permafrost?

Acknowledgments:

The studies in the paper were supported by the Sub-project No. XDA05120302 (Permafrost Extent in China during the Last Glaciation Maximum and Meg-athermal), Strategic Pilot Science and Technology Program of the Chinese Academy of Sciences (Identi-fication of Carbon Budgets for Adaptation to Changing Climate and the Associated Issues) (Grant No. XDA05000000), and under the auspices of the Inter-national Permafrost Association (IPA) Action Group on "Last Permafrost Maximum and Minimum (LPMM) on the Eurasian Continent."

Reference
Alexeeva NV, Erbajeva MA, 2000. Pleistocene permafrost in Western Transbaikalia. Quaternary International, 68-71: 5–12. doi: 10.1016/S1040-6182(00)00028-8
Baulin VV, Chekhovskii AL. 1973. Thickness of permafrost in Siberian lowlands. In:Proceedings of Second International Conference on Permafrost (Vol. 2). Yakutsk Press, pp. 280-312. (in Russian)
Black RF, 1976. Periglacial features indicative of permafrost:ice and sand wedges. Quaternary Research, 6: 3–26. doi: 10.1016/0033-5894(76)90037-5
Chang XL, Jin HJ, He RX, 2011. Formation and environmental evolution of sand wedges on the Tianshuihai North Lakeshore in the Western Kunlun Mountains. Quaternary Sciences, 31(1): 112–119.
Chang XL, Jin HJ, Zhang YL, et al, 2015. Thermal impacts of boreal forest vegetation on active layer and permafrost soils in northern Da Xing'anling (Hinggan) Mountains, Northeast Chi-na. Arctic, Antarctic and Alpine Research, 47(2): 267–279. doi: 10.1657/AAAR00C-14-016
Chen SH, 1988. Periglacial phenomena and their age in the valleys of eastern mountainous area of Heilongjiang Province. Journal of Glaciology and Geocryology, 10(4): 418–428.
Cheng J, Zhang XJ, Tian MZ, et al, 2006. Ice wedge casts discovered in the source area of the Yellow River, northeast Tibetan Plateau and their paleoclimatic implications. Quaternary Sciences, 26(1): 92–98.
China Institute of Meteorological Sciences. 1991. Historical Mete-orological Maps in Eurasia. China Meteorology Administration. (in Chinese)
Cui ZJ, Chen YX, Zhang W, et al, 2011. Research history, glacial chronology and origins of Quaternary glaciations in China. Quaternary Sciences, 31(5): 749–764.
Cui ZJ, Xie YY, 1984. On the southern boundary of permafrost and periglacial environment during the late part of late Pleistocene in North and Northeast China. Acta Geologica Sinica, 58(2): 165–176.
Cui ZJ, Xie YY, 1985. About the Late Pleistocene periglacial en-vironment in North China. Quaternary Sciences, 6(2): 115–119.
Cui ZJ, Yang JQ, Zhao L, et al, 2004. Discovery of a large area of ice-wedge networks in Ordos:Implications for the southern boundary of permafrost in the north of China as well as for the environment in the latest 20 ka BP. Chinese Science Bulletin, 49(11): 1177–1184. doi: 10.1360/03wd0211
Dong GR, Gao SY, Li BS, et al, 1985. Relict periglacial phenomena on the Ordos Plateau since the Late Pleistocene and their implications for climatic strata. Geographical Research, 4(1): 1–11.
Dong W, Xu QQ, Jin CZ, et al, 1999. Quaternary herbivore fauna in northeastern China:Evolution under climatic change. Chinese Science Bulletin, 44(Supplement): 129–132.
Fotiev SM, 1974. Geocryological Conditions in Central Siberia. Moscow: Science Press: 98-105.
French HM, 1999. Past and present permafrost as an indicator of climate change. Polar Research, 18(2): 269–274. doi: 10.1111/j.1751-8369.1999.tb00303.x
French HM, 2007. Evidence for past permafrost. West Sussex, England: John Wiley & Sons Ltd.
French HM, Millar SWS, 2014. Permafrost at the time of the Last Glacial Maximum (LGM) in North America. Boreas, 43: 667–677. doi: 10.1111/bor.12036
Gao FQ, 1986. The relation between formation of peat in seaside area of Bohai Sea and North Yellow Sea and sea level fluctuation since Late Würm-Glaciation. Geographical Science, 6(1): 56–63.
Gong GF, Chen EJ, Wen HR, 1979. The climatic fluctuations in the Heilongjiang Province, China. Acta Geographica Sinica, 34(2): 28–35.
Gravis GF, 1974. Major stages of permafrost. Moscow: Science Press: 184-186.
Guiyang Institute of Geochemistry, Chinese Academy of Sciences, 1977. Evolution of natural environments in Southern Liaoning Province during the last 10,000 years. Science in China, 7(6): 603–614.
Guo DX, 1979. Sand wedges on the Qinghai-Tibet Plateau. Journal of Glaciology and Geocryology, 1(1): 8–12.
Guo DX, Huang YZ, Wang JC, et al, 1989. Function of geological structure in the formation of permafrost conditions in Huola River basin, northern Da Hinggan Ling. Journal of Glaciology and Geocryology, 1(3): 216–222.
Guo DX, Li ZF, 1981. Preliminary approach to the history and age of permafrost in Northeast China. Journal of Glaciology and Geocryology, 3(4): 1–13.
Guo DX, Wang SL, Lu GW, et al, 1981. Division of permafrost regions in Daxiao Hinggan Ling, Northeast China. Journal of Glaciology and Geocryology, 3(3): 1–9.
Han JH, Wang YM, Li SQ, et al, 2009. Sequence structure and depositional filling of northern Songliao Basin during shrinking stage. Acta Sedimentologica Sinica, 27(3): 470–480.
Harry DG, Gozdzik JS, 1988. Ice wedges:growth, thaw transfor-mation, and palaeoenvironmental significance. Journal of Quaternary Science, 3: 39–53. doi: 10.1002/jqs.3390030107
Huijzer AS, Vandenberghe J, 1998. Climatic reconstruction of the Weichselian Pleniglacial in northwestern and central Europe. Journal of Quaternary Science, 13: 391–417. doi: 10.1002/(SICI)1099-1417(1998090)13:5<391::AID-JQS397>3.0.CO;2-6
Jia MC, Yuan F, Cheng GD, 1987. First discovery of ice wedges in Northeast China. Journal of Glaciology and Geocryology, 9(3): 257–260.
Jiang P. 1991. Studies on Mammuthus-Colelodonta Fauna and environment on the Songen Plain in Late Pleistocene. In:For-mation and Evolution of Natural Environments in Northeast China Plain in Quaternary. Harbin:Harbin Tushu Press, 24-29. (in Chinese)
Jiao SH, Wang LY, Sun CQ, et al, 2015. Discussion about the variation of permafrost boundary in Last Glacial Maximum and Holocene Megathermal, Tibetan Plateau. Quaternary Sciences, 35(1): 1–11.
Jin HJ, Chang XL, Guo DX, 2011. Holocene sand and soil wedges on the south-central Hulun Buir high plain in Northeast China. Quaternary Sciences, 31(5): 765–779.
Jin HJ, Chang XL, Wang SL. 2007a. Evolution of permafrost on the Qinghai-Xizang (Tibet) Plateau since the end of the late Pleistocene. Journal of Geophysical Research, 112:F02S09. DOI:10.1029/2006JF000521.
Jin HJ, Yu QH, Lü LZ, et al, 2007b. Degradation of permafrost in the Xing'anling Mountains, Northeastern China. Permafrost and Periglacial Processes, 18(2): 245–258. doi: 10.1002/ppp.589
Jin HJ, Yu SP, Lü LZ, et al, 2006. Preliminary assessment on the permafrost degradation and its trends in the Da and Xiao Xing'anling Mountains. Journal of Glaciology and Geocryology, 28(4): 467–476.
Kahlke RD, 2014. The origin of Eurasian Mammoth Faunas (Mammuthus-Coelodonta Faunal Complex). Quaternary Sci-ence Reviews, 96: 32–49. doi: 10.1016/j.quascirev.2013.01.012
Kasse C, Huijzer AS, Krzyszkowski D, et al, 1998. Weichselian Late Pleniglacial and Late-glacial depositional environments, Coleoptera and periglacial climatic records from central Poland (Bełchatów). Journal of Quaternary Science, 13(5): 455–469. doi: 10.1002/(SICI)1099-1417(1998090)13:5<455:AID-JQS398>3.0.CO;2-T
Kolstrup E, 1986. Reappraisal of the Upper Weichselian periglacial environment from Danish Frost Wedge Casts. Paleogeography, Paleoclimatology, Paleoecology, 56: 237–249. doi: 10.1016/0031-0182(86)90096-9
Kong ZC, Du NQ, 1984. The macrofossil plants and pollen assemblage on the Last Glacial in Sanjiang Plain. Scientia Geographica Sinica, 4(1): 76–80.
Li BL, Zhou CH, 2001. Sandy desertification in western Northeast China Plain in the past 10 years. Acta Geographica Sinica, 56(3): 307–315.
Li FH. 1991a. Paleo-periglacial structures on the Sanjiang Plain, Northeast China and their environmental implications. In:Formation and Evolution of Natural Environments in Northeast China Plain in Quaternary. Harbin:Harbin Tushu Press, pp. 202-208. (in Chinese)
Li QS, 1991b. The formation of Songnen Sandy Land and the changes of the environment. Journal of Desert Research, 11(3): 36–43.
Li X, Li XZ, Yi CL, et al, 2006. The periglacial blown-sand (aeo-lian) landforms and sediments:a review. Journal of Glaciology and Geocryology, 28(2): 199–210.
Li XH, Liu GL, Xu GY, et al, 1982. Periglacial phenomena at the east open-cut coal mine of Zhalinuoer in Inner Mongolia and their geochronology. Journal of Glaciology and Geocryology, 4(3): 65–72.
Liang FX, Cheng GD, 1984. Polygon-vein structures along the Qinghai-Tibet Highway and their paleo-climatic implications. Journal of Glaciology and Geocryology, 6(4): 45–59.
Lin NF, Tang J, Bian J, et al, 1999. The Quaternary environmental evolution and the problem of desertification in Northeast Plain. Quaternary Sciences, 19(5): 448–454.
Liu DS, 1994. Discussions on regional distributive features of Quaternary deposits in China. Beijing: Science Press: 80-87.
Liu J, 1989. Vegetational and climatic changes at Gushantun Bog in Jilin, NE China since 13000 Y. BP. Acta Palaeotologica Sinica, 28(4): 495–509.
Liu X, Sui WG, Xiang TY, et al, 2004. Holocene volcanic activities in Tianchi area, Changbai Mountains. Quaternary Sciences, 24(6): 638–644.
Liu ZJ, Tian J, Cui ZX. 1990. Features of geomorphology and geological processes in the Baitoushan region. In:Ning HY (ed.). Studies on Volcanoes in the Changbaishan Mountains. Yanbian, Jilin, China:Yanbian Univeersity Press, 104-119. (in Chinese)
Lu GW, Weng BL, Guo DX, 1993. Geographical southern limit of permafrost in Northeast China. Journal of Glaciology and Ge-ocryology, 15(2): 214–218.
Lu HY, Yi SW, Xu ZW, et al, 2013. Chinese deserts and sand fields in Last Glacial Maximum and Holocene Optimum. Chinese Science Bulletin, 58(23): 2775–2783. doi: 10.1007/s11434-013-5919-7
Luo DL, Jin HJ, Jin R, et al, 2014. Spatiotemporal variations of climate warming in northern Northeast China as indicated by freezing and thawing indices. Quaternary International, 349: 187–195. doi: 10.1016/j.quaint.2014.06.064
Mackay JR, 1974. Ice-wedge casts, Garry Island, Northwest Territories. Canadian Journal of Earth Sciences, 11: 1336–1383. doi: 10.1139/e74-133
Markov KK, Popov AI, 1960. Periglacial Environment in the USSR,: 44–100.
Matsuo T, 1983. Vegetation and climate during the last Glacial Maximum in Japan. Quaternary Research, 19(2): 212–235. doi: 10.1016/0033-5894(83)90006-6
Murton JB, French HM, 1995. Thermokarst involutions, Summer Island, Pleistocene Mackenzie delta, Western Canadian Arctic. Permafrost and Periglacial Processes, 4: 217–229. doi: 10.1002/ppp.3430040304
Murton JB, Kolstrup E. 2003. Ice-wedge casts as indicators of palaeotemperatures:precise proxy or wishful thinking? Pro-gress in Physical Geography, 27:155-170. DOI:10.1191/0309133303pp365ra.
Owen LA, Windley BF, Cunningham WD, et al, 1997. Quaternary alluvial fans in the Gobi Desert, southern Mongolia:evidence for neotectonics and climate change. Journal of Quaternary Science, 12: 239–252. doi: 10.1002/(SICI)1099-1417(199705/06)12:33.0.CO;2-P
Pei WZ, 1957. Preliminary study on involutions near Huangshan, Harbin and Jalainor, Inner Mongolia Autonomous Region. Chinese Science Bulletin, 1(1): 51–54.
Peng HY, Cheng GD. 1990. Ice wedges in the Da Xing'anling (Hinggan) Mountains in Northeastern China and their paleo-climatic implications. In:Proceedings of Fourth Chinese Conference on Glaciology and Geocryology (Geocryology Volume). Beijing, China:Science Press, pp. 9-16.
Péwé TL. 1964. Ice wedges in Alaska-classification, distribution, and climatic significance. In:Proceedings of the First Interna-tional Conference on Permafrost. Lafayette, Indiana, 11-15 November 1963. Special Paper Geological Society of America, 129. NAS/BRC (USA) Publication, No. 1287:76-81.
Péwé TL, 1966a. Paleoclimatic significance of fossil ice wedges. Biuletyn Peryglacjalny, 15: 65–73.
Péwé TL. 1966b. Ice wedges in Alaska-Classification, distribu-tion, and climatic significance. In:Proceedings of the 1st In-ternational Conference on Permafrost (Publication 1287). Na-tional Academy of Science, National Research Council of Canada, Ottawa, pp. 76-81.
Péwé TL. 1983. The periglacial environment in North America during Wisconsin time. In:Church M, Slaymaker O (eds.). Late-Quaternary of the United States. Vancouver, BC, Canada:University of British Columbia Press, pp. 154-165
Pu QY, 1984. Southern limit of permafrost in East China during the Dali Ice Age of the Late Pleistocene and its evolution. Hy-drogeology and Engineering Geology, 4: 49–51.
Pu QY. 1985. Where is the boundary of permafrost in China during the Late Pleistocene? In:Collection of Papers for the Chinese Symposium on Quaternary Glaciology and Periglacial Geo-morphology. Beijing:Science Press, pp. 216-220. (in Chinese)
Qiu SW, 1989. Study on the formation and evolution of Horqin Sandy Land. Scientia Geographica Sinica, 9(4): 317–328.
Qiu SW. 1991a. Genesis and environmental changes of lakes on the Songnen Plain. In:Formation and Evolution of Natural En-vironments in Northeast China Plain in Quaternary. Harbin:Harbin Tushu Press, pp. 146-154. (in Chinese)
Qiu SW. 1991b. Study on the formation and evolution of the Horqin Sandy Land. In:Formation and Evolution of Natural Environments in Northeast China Plain in Quaternary. Harbin:Harbin Tushu Press, pp. 185-201. (in Chinese)
Qiu SW, Jiang P, Li FH, et al, 1981b. Preliminary discussion on evolution of natural environment of Northeast China since the Late Glacial period. Acta Geographica Sinica, 36(3): 315–337.
Qiu SW, Li FH, Sui X, 1981a. Periglacial geomorphology in the Changbaishan Mountains. Journal of Glaciology and Geocry-ology, 3(1): 26–30.
Qiu SW, Xia YM, Li FH, et al, 1984. Study on paleogeography of Songliao Plain during the Middle Quaternary. Chinese Science Bulletin, 29(3): 172–174.
Qiu SW, Xia YM, Wang PF, et al. 1991. Study on Pleistocene strata and depositary environment on the Songliao Plain. In:Formation and Evolution of Natural Environments in Northeast China Plain in Quaternary. Harbin:Harbin Tushu Press, pp. 1-11. (in Chinese)
Qiu SW, Zhang B, Wang ZC, 2005. Analyses on current situation, causes of formation, and way of management of desertification in western Northeast Plain of China. Quaternary Sciences, 25(1): 63–72.
Ren GY, Zhang LS, 1998. A preliminary mapped summary of Holocene pollen data for Northeast China. Quaternary Science Review, 17: 669–688. doi: 10.1016/S0277-3791(98)00017-1
Romanovskii NN, 1977. Formation of Polygonal Structures. Russia: Science Press: 213-235.
Romanovskii NN. 1980. Principles of classification of polygonal-vein structures. In:Proceedings, Third International Conference on Permafrost; English Translations of 26 of the Soviet Papers. Ottawa, Canada:National Research Council of Canada, pp. 297-310.
Romanovskii NN. 1985. Distribution of recently active ice and soil wedges in the USSR. In:Church M, Slaymaker O (eds.). Field and Theory:Lectures in Geocryology. Vancouver, BC, Canada:University of British Columbia Press, pp. 154-165.
Song HY. 1991. Features of peat and paleo-environment in West-Central Songliao Plain. In:Formation and Evolution of Natural Environment on Northeast China Plain in Quaternary. Harbin:Harbin Tushu (Bibliography) Press, pp. 76-83. (in Chinese)
Song HY, Xia YM, 1988. Fossil-pingo lake peatlands in the San-jiang Plain. Journal of Glaciology and Geocryology, 10(1): 76–83.
Sun JM, Ding ZL, Liu DS, 1998. Desert distributions during the glacial maximum and climatic optimum:example of China. Episode, 21: 28–31. doi: 10.1016/S0277-3791(98)00017-1
Sun JM, Liu DS, 2001. Desertification in the Northeastern China. Quaternary Sciences, 21(1): 72–78.
Sun JZ, 1981. Quaternary periglacial stages of the Songliao Plain. Scientia Geographica Sinica, 1(2): 163–170.
Sun JZ, 1985. Preliminary reconstruction of the temperature curve of the last major climatic cycle in North China. Journal of Glaciology and Geocryology, 7(4): 317–321.
Sun JZ, Wang SY, Wang YZ, et al, 1985. Paleoenvironment in Northeast China during the Last Glaciation. Quaternary Sci-ences, 6(1): 82–89.
Sun XJ, Weng CY, 1992. Pollen records on the history of mixed conifer and hardwood forest in Northeast China. Acta Botanica Sinica, 34(5): 394–401.
Tong BL, 1993. Ice wedges in Northeastern China. Journal of Glaciology and Geocryology, 15(1): 41–46.
Turvey ST, Tong H, Stuart AJ, et al, 2013. Holocene survival of Late Pleistocene megafauna in China:a critical review of the evidence. Quaternary Science Reviews, 76: 156e166. doi: 10.1016/j.quascirev.2013.06.030
Van Huissteden K, Vandenberghe J, Pollard D, 2003. Palaeotem-perature reconstructions of the European permafrost zone during marine oxygen isotope Stage 3 compared with climate model results. Journal of Quaternary Science, 18: 453–464. doi: 10.1002/jqs.766
Van Vliet-Lanoë B, 1988. The genesis of cryoturbations and their significance in environmental reconstruction. Journal of Qua-ternary Science, 3: 85–96. doi: 10.1002/(ISSN)1099-1417
Vandenberghe J, 1992. Cryoturbations:a sediment structural analysis. Permafrost and Periglacial Processes, 3: 343–352. doi: 10.1002/ppp.3430030408
Vandenberghe J, Cui ZJ, Zhao L, et al, 2004. Ther-mal-contraction-crack networks as evidence for late-Pleistocene permafrost in Inner Mongolia, China. Permafrost and Periglacial Processes, 15(1): 21–29. doi: 10.1002/ppp.476
Vandenberghe J, French HM, Gorbunov A, et al, 2014. The Last Permafrost Maximum (LPM) map of the Northern Hemisphere:permafrost extent and mean annual air temperatures, 25-17 ka BP. Boreas, 43: 652–666. doi: 10.1111/bor.12070
Vandenberghe JF, Wang XY, Vandenberghe D, 2015. Very large cryoturbation structures of Last Permafrost Maximum age at the foot of the Qilian Mountains (NE Tibet Plateau, Chi-na). Permafrost and Periglacial Processes, 27(1): 138–143. doi: 10.1002/ppp.1847
Wang MH, 1987. Preliminary of paleovegetation and paleoclimate in Northeast Plain of China in the later part of the Late Pleis-tocene. Journal of Glaciology and Geocryology, 9(3): 229–237.
Wang NA, Zhao Q, Li JJ, et al, 2003. The sand wedges of the last ice age in the Hexi Corridor, China:paleoclimatic interpretation. Geomorphology, 51: 313–320. doi: 10.1016/S0169-555X (02)00243-X
Wang SW, Gong DY, 2000. Temperatures in China during the several characteristic periods in the Holocene. Advances in Natural Sciences, 10(4): 325–332.
Wang T, 2006. 1:4000000 Map of the Glaciers, Frozen Ground and Deserts in China. Beijing: Science Press.
Washburn AL, 1980. Permafrost features as evidence of climatic change. Earth Science Reviews, 15(4): 327–402. doi: 10.1016/0012-8252(80)90114-2
Wei Z, Jin HJ, Zhang JM, et al, 2011. Prediction of permafrost changes in Northeastern China under a changing climate. Sci-ence in China, 54(6): 924–935. doi: 10.1007/s11430-010-4109-6
Xia YM, 1988. Preliminary study on vegetational development and climatic changes in the Sanjiang Plain in the last 12000 years. Scientia Geographica Sinica, 8(3): 240–248.
Xia YM, Wang PF. 1991. Studies on pollen compositions, paleo-vegetation and paleo-climate on the Songnen Plain during the Late Neogene-Pleistocene. In:Formation and Evolution of Natural Environment on Northeast China Plain in Quaternary. Harbin:Harbin Tushu (Bibliography) Press, pp. 12-23. (in Chinese)
Xiao RH, Hu JB, 1988. Some characteristics of the climate landform in Northeast China since the Last Glacial Epoch. Journal of Glaciology and Geocryology, 10(2): 125–133.
Xie YY, 1984. Periglacial sedimentary environment of Huanggan-gliang region, the main peak of the Da Xing'anling Mountains in the Late Pleistocene. Journal of Glaciology and Geocryology, 6(4): 69–76.
Xu SY, Pan BT. 1990. Periglacial wedge structures and their for-mation environments on Eastern Qinghai-Tibet Plateau. In:Proceedings of Fourth Chinese Conference on Glaciology and Geocryology (Geocryology Volume). Beijing, China:Science Press, pp. 17-24. (in Chinese)
Xu SY, Xu DF, Pan BT. 1989. Problems on the boundary of per-mafrost zones in East China during the Last Glaciation. In:Collection of Papers for the Third Chinese Conference on Ge-ocryology. Beijing:Science Press, pp. 105-113. (in Chinese)
Xu WD, 1983. The relation between the distribution of edificatory and companion in azonal vegetation and water-temperature condition in Northeast China. Acta Botanica Sinica, 25(3): 261–273.
Xu YG, Fan QC, 2015. Cenozoic valcanism in Eastern China:Review and perspectives. Bulletin of Mineralogy, Petrology and Geochemistry, 34(4): 682–689.
Yang HR, Yan SY, 1958. Quaternary periglacial phenomena in the lower reaches of the Yangtze River. Quaternary Sciences, 1(2): 141–154.
Yang JC, Li SD, Sun JZ, et al, 1983. Ice wedge casts and sand wedges and natural environments in the Datong Basin, Shanxi Province, China during the Late Pleistocene. Chinese Geo-graphical Science, 3(4): 399–344.
Yang SZ, Cao XY, Jin HJ, 2015. Validation of ice-wedge isotopes at Yituli'he, northeastern China as climate proxy. Boreas, 44(3): 502–510. doi: 10.1007/s11430-010-4109-6
Yang SZ, Jin HJ, 2010. Isotopic records of oxygen and deuterium in ice wedges from Yituli'he, Northeast China and their paleo-climatic indications. Science in China (Series D:Earth Scienc-es), 54(1): 119–126.
Yang SZ, Jin HJ, Wen X, et al, 2009. Relationships between tem-perature change and microbial amount in inactive ice wedges in Yituli'he, Northeast China. Applied Ecology, 20(11): 2785–2789.
Yang XD, Wang SM, Xue B, et al, 1995. Vegetational development and environmental changes in Hulun Lake since Late Pleistocene. Acta Palaeontologica Sinica, 34(5): 647–656.
Yang XP, Rost KT, Lehmkuhl F, et al, 2004. The evolution of dry lands in northern China and in the Republic of Mongolia since the Last Glacial Maximum. Quaternary International, 118-119: 68–85. doi: 10.1016/S1040-6182(03)00131-9
Yi SW, Lu HY, Zeng L, et al, 2013. Paleoclimate changes and reconstruction of the border of Horqin (Northeastern China) since the Last Glacial Maximum. Quaternary Sciences, 33(2): 206–217.
Zeng JP. 1991. Buried peat and paleoenvironment on the Sanjiang Plain in the late part of Late Pleistocene. In:Formation and Evolution of Natural Environment on Northeast China Plain in Quaternary. Harbin:Harbin Tushu (Bibliography) Press, pp. 217-221. (in Chinese)
Zeng L, Lu HY, Yi SW, et al, 2013. Environmental changes of Hulun Buir Dune-field in northeastern China during the Last Glacial Maximum and Holocene Optimum. Quaternary Sci-ences, 33(2): 243–251.
Zhang HC, 2008. A review of the study of environmental changes and extinction of the Mammuthus-Colelodonta Fauna during the middle-late Late Pleistocene in NE China. Advances in Earth Science, 24(1): 49–60.
Zhang QY, Xie L. 1991. Discussions on genesis of loess in Changchun, Northeast China. In:Formation and Evolution of Natural Environment on Northeast China Plain in Quaternary. Harbin:Harbin Tushu (Bibliography) Press, pp. 109-116. (in Chinese)
Zhang W, Niu YB, Yan L, et al, 2008. Late Pleistocene glaciation of the Changbaishan Mountains in northeastern China. Chinese Science Bulletin, 53(17): 2672–2684. doi: 10.1007/s11434-008-0347-9
Zhang YX, Li YX, Xie K, et al, 2015. Succession of large herbivorous faunas and environmental evolution in Northeast China during the Last Glaciation Maximum. Quaternary Sciences, 35(3): 622–630.
Zhao L, Jin HJ, Li CC, et al, 2013. The extent of permafrost in China during the local Last Glacial Maximum (LLGM). Boreas, 43(3): 688–698. doi: 10.1111/bor.12049
Zheng Z, Yuan BY, Petit-Maire N, 1998. Paleoenvironments in China during the Last Glacial Maximum and Holocene Opti-mum. Episodes, 21(3): 152–158.
Zhou YW, Qiu GQ, Cheng GD, et al, 2000. Frozen Ground in China. Beijing: Science Press.
Zhu KZ, 1972. Preliminary study on climate change in China during the past 5000 years. Archeologists,: 28–40.