Origin of soft-sediment deformation structures in Nihewan Basin

De-Chen Su , Ai-Ping Sun , Zhao-Li Li ,Song-Yong Chen , Zhen-Jie Wu

a Key Laboratory of Deep-Earth Dynamics of Ministry of Natural Resources, Institute of Geology, Chinese Academy of Geological Sciences, 26 Baiwanzhuang Rd, Beijing, 100037, China

b Jiangsu Donghai Continental Deep Hole Crustal Activity National Observation and Research Station, China

Abstract The Nihewan Basin is a rift basin at the junction of northern Shanxi Province and northwestern Hebei Province in north China. The basin is known for its rich paleontological fossils and ancient human remains.There are also abundant soft-sediment deformation structures(SSDS)in the thick lacustrine sediments.Previously, most SSDS have been interpreted as ice-edge features or ignored entirely. Recently, the authors have carried out several field surveys in the Nihewan Basin and found that many SSDS are sandwiched between normal lacustrine strata at multiple sections.In the excavation pit at the 10th Locality of Maliang Site(ML10),10 horizontal SSDS layers and two vertically developed geological features have been identified. Based on genesis analysis and related criteria, these features are divided into two categories: cryoturbation-triggered SSDS and earthquake-triggered SSDS.Among them,a special type of ancient ice-wedge pseudomorph(SSDS-8)of ML10 is recognized in the basin for the first time. The other 9 horizontal SSDS are mainly caused by earthquake-triggered liquefaction and slumping. They can be further divided into 14 seismic event layers.These findings indicate that the tectonic activity in the Nihewan Basin is very strong and frequent, and there were cold periods in the geological history of the basin.At the same time,the SSDS with distinct morphological characteristics and stable horizontal distribution in the basin can be used as an important indicator of stratigraphic correlation.

Keywords Nihewan Basin, Soft-sediment deformation structures (SSDS), Paleoearthquake, Freezing-andthawing, Ice-wedge, Palaeolithic site

1.Introduction

The Nihewan Basin is one of several rift basins of the Fenhe River-Weihe River Graben System in the middle of the North China Plate. Starting from the Weihe Basin in the southwest,following in a northeast direction, is the Yuncheng Basin, Linfen Basin, Taigu Basin (Taiyuan Basin), Hutohe River Basin, Nihewan Basin, and Yanqing-Huailai Basin (Wang, 1925; Xing et al., 2005; Yuan et al., 2009; Hu et al., 2011). Between 2 million and 3 million years ago,due to strong crustal movement, the Nihewan Basin, with a total area of about 10,000 square kilometers, was finally formed at the border of northern Shanxi Province and northwestern Hebei Province, North China (Fig. 1)(Yuan et al., 1996; Wei, 1997, 2009, 2012; Liu et al.,2018).

The earliest scientific research on the stratigraphy and paleontology of the basin began in the 1920s.From 1921 to 1924, the French priest Ernest Vincent discovered a large number of animal fossils in and near Nihewan village (Ni-Ho-Wan) and informed E. Licent,and G.B. Barbour successively. Between 1924 and 1929, G.B. Barbour, E. Licent and P. Tellhard de Chardin conducted a series of geological investigations in Nihewan village (40°15′30.36′′N, 114°40′54.86′′E,813 m) and surrounding areas on the banks of the Sanggan River(Sangkan Ho)in Yangyuan County,Hebei Province. In 1924, at the end of paper Preliminary Observations in the Kalgan Area, Barlbour added an extra note on his latest and short geological inspection on the Nihewan village and its vicinity in the Sanggan River Basin. He named the thick lacustrine deposits between Xiaodonggou(Hsiao Tung K'ou)(40°13′9.25′′N,114°37′35.82′′E, 815 m) and downstream Nihewan as“Nihewan beds” (Nihowan beds). He also named the basin after Sanggan River (Sangkan Basin) (Barbour,1924). In 1925, Barbour published another paper and identified the age of Nihewan beds as Early Pleistocene by comparing the fossils with that of Henan Province(Barbour, 1925). During the winter of 1925, E. Licent and G. Barbour sent some mammalian fossils of the Nihewan Basin to Prof. Teilhard de Chardin and Prof.Boule. After a more careful study, Prof. Teilhard de Chardin confirmed that these mammals lived in the very early Pleistocene or very late Pliocene(Teilhard,1926). Also in 1926, Barbour, Licent, and Teilhard coauthored the Geological Study of the Deposits of the Sangkanho Basin (Sanggan River Basin), which briefly summarized the vertebrate fossils and the lacustrine Nihewan layers of the basin and correlated the geological age to the Villafranchian period in Europe(Barbour et al., 1926). In 1948, the resolution of the 18th International Geological Congress suggested that the Villafranchian in Europe could be the lower boundary of the Pleistocene.The comparable Nihewan beds can be used as the lower boundary of the Pleistocene in China.This scheme was formally affirmed by the Chinese geoscience community in 1954 (Chen et al., 1986; Xia, 2001). Since then, the little-known Nihewan village has become a fruitful confluence of Palaeolithic archaeology, Quaternary geology, and paleo-mammalogy in northern China and people call this basin as the Nihewan Basin or Sanggan River Basin(Fig.1A)(Wei,1976;Wei et al.,1977;Wu,1980;Chen et al., 1986; Xia, 2001; Wei et al., 2002; Min et al.,2006; Zhu et al., 2007; Wei, 2009; Ge and Wang,2010; Xie, 2018). More recently, the term Nihowan Beds was replaced by Nihewan Formation in 1964,which is used to define the whole fluvio-lacustrine sequence in the Nihewan Basin (Wang and Pan, 1982;Yuan et al.,1996;Deng et al.,2007;Zhu et al.,2007).

The Nihewan Basin has been in a state of very frequent and intense tectonic activities from the very beginning to the present. The basin is surrounded by several large and deep faults (Fig. 1A), all of which have long geological histories (Wang, 1925). For example, the activity of Kouquan Mountain Fault on the west boundary(Fig.1A,B)can even be traced back to the Jurassic (Shanxi Bureau of Geology and Mineral Resources, 1989; Luo et al., 2019; Ding et al., 2021).Kouquan Mountain Fault is, in general, a right-lateral strike-slip fault. It can sometimes be divided into several segments, and sometimes it becomes two or more parallel branches (Fig. 1B) (Ding et al., 2021).Dozens of east-west-directed gullies are offset rightlaterally for 8-20 m when crossing the fault in one section of Kouquan Fault near Songjiazhuang village.Two big trenches revealed a line of evidence of the Holocene activities of Kouquan Fault and 4 paleoearthquake events since 13.73 ka BP(Xie et al.,2003).

The Liuleng Mountain Fault is another important and active fault that controls the basic structure of the basin and surrounding mountain ranges. It has a total length of 130 km (Deng et al., 1994; Cheng and Yang,1996; Yuan et al., 2009, 2011). The topographical features show that the Sanggan River Valley, alluvial fans,etc.,are interrupted or controlled by the Liuleng Mountain Fault or its branches (Fig. 1A, C, E). The study on the kinematic characteristics of the Liuleng Mountain Fault shows that it is a dip-slip normal fault that is still active,with an average vertical sliding rate of 0.43-0.75 mm/a from the late Pleistocene to the Holocene period(Deng et al.,1994).In addition to the evidence of new tectonic activities visible on the surface, two ancient seismic events were recognized in the trenches across the fault zone in 1997. According to infrared luminescence (IRSL) dating, the two earthquakes occurred most probably around 14.2ka and 6.7ka(Yin et al., 1997).

Field investigations and analysis of satellite images show that the active faults of smaller scales are relatively common in Quaternary unconsolidated sedimentary layers (e.g., Dong et al., 1996; Yang et al.,2019) (Fig. 1E, G). As a result, earthquakes of different scales triggered by neo-tectonic movement occur from time to time. For example, in October 1989,an earthquake with a maximum magnitude of 6.1 and multiple aftershocks occurred near the Liuleng Mountain Main Fault. In 1991 and 1999, another two earthquakes with a magnitude greater than 5(Ms >5)occurred nearly in the same area (Fig. 1A, Ma, 1992;Deng et al., 1994; Feng et al., 2016A).

During the existence of Nihewan Lake, it was also the time of intense crustal activity in North China(Yuan et al.,2009,2011).Between 900,000 and 90,000 years ago,the tectonic-magmatic activity of the basin reached a climax. A large amount of basalt magma spewed from the deep bottom of the basin, forming a volcano group composed of more than 30 volcanoes in the east of Datong City (Yin, 1976; Hu et al., 2017.Fig. 1A, F). The latest high-resolution tomographic model of the upper mantle indicates that a plume has played an important role in the forming process of the basalt magma (e.g., Lei, 2012; Cai et al., 2021). Four major volcanic eruption cycles have been recognized,each major cycle has many smaller eruption events(Hu et al., 2011, 2017). Based on the evidence of geomorphology, sedimentology, and palaeontology, Xia(1992) found 4 ancient lake shorelines and there have been several periods of lake level changes during the whole evolutionary history of the lake, and the main reason for the changes are neo-tectonic movement(Long et al., 1991; Xia,1992;Wei, 1997).

In August 2020, we learned from Prof. Wei that there are many abnormal deformation features in the archaeological excavation sites in the Nihewan Basin.He has been puzzled by these features for a long time because some of them could not be explained by freeze-and-thaw action. Therefore, we went to Donggutuo village from August 15th to 18th,2020,and inspected several archaeological excavation pits under the leadership of Prof. Wei. We found a lot of softsedimentdeformationstructures(hereinafter referred to‘SSDS’for both singular and plural)that are interbedded with normal lacustrine strata and have obvious characteristics of liquefaction or fluidization,as well as syn-sedimentary fractures and gravitational phenomena that cut across multiple layers.From then on, we checked several other excavation pits at Maliang and Hongya,and found many spectacular SSDS there. We also visited several tectonic sections and Datong Volcanic Park to investigate the activities of faults and volcanoes and their relationship with the earthquakes. The characteristics and causes of the deformation of these soft sediments will be discussed below according to the exposed locations.

2.Previous studies on SSDS in Nihewan Basin

In 1936,Chan Kuoda(Chen Guoda)reported for the first time the “subaquatic deformation” of unconsolidated sediments in the Nihewan layer in Taigu County,Taiyuan Basin, which belongs to Fenhe River-Weihe River Graben System. After excluding various possible causes (surface slippage, differential consolidation and gravity load, gypsum expansion due to hydration), Chan pointed out that the deformation in this layer is due to the micro-tectonic movement acting on the ongoing sedimentation. The deformational feature is “originated by lateral compressions,which are caused by tectonic movement, and intimately related to the large regional structure”(Chan,1936).There was no concept of seismite or SSDS at that time. However, Chan's paper is definitely the earliest research article regarding the origin of SSDS in China.

From 1963 to 1975, Wang Kejun and Pan Jianying conducted field geological surveys in the Nihewan Basin. They found the purple-red moraine and glacial deposits of the Pliocene at the bottom of the sedimentary sequences near Hongya village(Fig.1)on the left bank of the Huliu River(Zhou,1981;Wang and Pan,1982).

In 1964, geologists from Tianjin Institute of Geology and Mineral Resources discovered periglacial phenomena in the Nihewan Basin, but there was no detailed report (Yuan, 1989). In June 1974, Yuan discovered the freeze-and-thaw folds at the top of the Nihewan layer near Hutouliang village. This discovery immediately attracted the attention of Young Chung-Chien (Yang Zhongjian), Pei Wen-Chung (Pei Wenzhong), and Sun Tien-Ching (Sun Dianqing). They were all well-known geologists in Quaternary geology and paleoanthropology in China at that time and they all went to the field to conduct on-site inspections and confirmed the great significance of this discovery(Yuan, 1989).

Fig. 1 Simplified geomorphological map of Nihewan Basin and some field photos, showing the geomorphological features, the main faults,and the volcanoes in the basin.A)The relief map is drawn based on Google Earth and our field exploration.The main faults are based on Map of Active Tectonics in China (Deng et al., 2007) and the location of the epicenter of the Nihewan Basin is from the article by Feng et al.,2016). The magnitude information of the 1989 and 1991 earthquakes is from Ma Jin's paper (Ma, 1992); B) Photo taken by a drone near Xiaoyukou village on the southwest side of the Nihewan Basin(39°47′33.71′′N,112°56′38.33′′E).The mountain on the left is mainly composed of Upper Cambrian and Ordovician limestones (C-O), the Archean gneiss (Ar) exposed here is about 400 m wide, and the right side of the photo is the Quaternary sediments (Q) of Sanggan River valley; C) Photo shot by a drone near Xiaodukou village (40°13′3.42′′N,114°37′24.65′′E). The photo shows the positions of Nihewan village, Sanggan River, Cenjiawan village, and Maliang archeological site. The Liuleng Mountain main fault passes near Cenjiawan,leaving clear features of the faulted landform.The former Nihewan Ancient Lake and the current Sanggan River are both greatly affected by this fault and its branching fault.The straight distance between Maliang Site and this fault is about 900 m;D)Photo further showing the characteristics of the Liuleng Fault plane,which is a high-angle normal fault.Cenjiawan village is located on the hanging wall of the fault.The footwall of the fault is Proterozoic(Pt)sandstone and carbonate,which are severely broken by the main fault and its branches. Photo was taken at east of Cenjiawan (40°14′10.43′′N, 114°40′37.40′′E); E) Photo taken in Majiayao village,where the main fault and its branch have an obvious angle of intersection(40° 6′0.19′′N,114°28′19.36′′E).The Quaternary(Q)alluvial fan in the piedmont is staggered by the fault,forming a steep ridge with a straight boundary;F)Photo taken in Datong Volcano Geopark.The photo shows Langwoshan Volcano(in the centre)(40° 6′6.12′′N,113°38′10.42′′E)and Jinshan Volcano(40° 6′46.75′′N,113°37′0.76′′E)in the distance(to the left).Nearly all surrounding area of the volcanoes is covered by Malan loess,including the inside of the Langwo crater.Before it was opened as part of a volcano park, the loess inside the Langwo crater was reclaimed by local peasants as farmland. With the volcano as the midpoint,there are radial gullies to the periphery;G)A high-angle normal fault in the Quaternary lacustrine sediments beside the tributary of Nihewan,Chenzhuang village,Datong City(39°58′14.18′′N,113°31′43.07′′E),Shanxi Province.The dip angle of the fault is between 60 and 80°and the fault displacement is about 90 cm. Similar small faults are often encountered in the basin. The person in the photo is 165 cm tall.

During the excavation of the Xujiayao Palaeolithic Site from March to June in 1976,Chia Lang-P'o and Wei Chi'i (Jia Lanpo and Wei Qi) observed “freeze-andthaw folds” formed under periglacial climatic conditions (Chia et al., 1979). In the late 1970s, Wu Zirong also noticed the interlayered folded features in Hutouliang section and Hongya section(Fig.1),where the gravel layer was abnormally “twisting and rotating”, some mud between the gravel layers has“folding deformation” (Wu, 1979; Wu et al., 1980).Based on the strange appearance and deformation pattern, Wu and his collaborators infer that these deformations are“produced under the drive of freezingand-thawing”, and named them freeze-and-thaw deformations. The freezing-and-thawing process also envelops the overlying Malan loess and calcareous nodules.Therefore,it was concluded that the Nihewan Basin has experienced some periodic “cold and warm climate”(Wu et al., 1980).

At the end of 1980, Zhou et al. (1982) found some ancient ice-wedge pseudoforms under the freeze-andthaw folds near Yujiagou (Fig. 1A), a small gully of Hutouliang village. The height of these ice-wedge pseudoforms is normally between 1.5 and 1.7 m.Zhou reported that these ice-wedge pseudoforms and the folds are all “periglacial phenomena” and they were the products of two successive freezing events.Zhou et al. (1982) did a chronological study on the freeze-and-thaw folds by analyzing the C14content in a thin layer of calcareous membrane crust between the 1.5-m-thick silt layer and the lower gravel layer at Yujiagou. The result is 27,675 ± 745 years, which represents the deposition time before the beginning of freezing-and-thawing. Before that, Liu et al. (1978)had already achieved thermoluminescence age of the top of Malan loess at Luochuan County as B.P.8000 ± 400 years. Accordingly, Zhou et al. (1982)deduced that the age of this periglacial environment at Yujiagou was between 10,000 and 27,000 years.Yang et al.(1983)reported the discovery of more than 30 ancient ice-wedges in the south of Xubao village,Datong County(Fig.1). The height of most of the icewedges exceeds 1 m.In 1988,Chen et al.conducted a multidisciplinary study on the Nihewan Beds, and made a very systematic summary of 13 geological sections. The term “periglacial accumulation” and“thawing fold” appeared once respectively in the section descriptions. After entering the 1990s, few scholars have paid attention to the deformation phenomenon in the Nihewan Beds, and only briefly mentioned in a few articles. For example, in the description of the stratigraphic profile of Donggutuo Site,the“curly contact”of light orange silty clay was mentioned (Hou et al., 1999). Since then, nearly no researchers have reported or even mentioned similar deformation features in their research papers.

3.Field observations of the SSDS

From August 2020 to August 2021,we found a lot of SSDS both in the outcrops at Yujiagou and Hongya village and in the archaeological excavation pits at Maliang Site near Donggutuo village. From July to August in 2021,we also visited some geological sites in Datong Volcanic Park and checked some larger faults around the Nihewan Basin, in order to find the relationship between the activities of the faults and volcanoes with the origins of these SSDS.The strata in the Nihewan Basin are generally in a loose state and are highly prone to denudation.The surface of the basin is generally covered with Quaternary loess. Therefore,the SSDS in the strata are easily covered by loess,they can only be observed clearly on the walls of the recently excavated pits or newly constructed roadcuts. Often after a few showers of rain, they become completely unrecognizable.

3.1. Yujiagou of Hutouliang village

“Gou”means valley or gully in Chinese.Yujiagou is a gully of Hutouliang village.It is located on the north bank of the Sanggan River in the middle of the Nihewan Basin. In the 1970s, Gai and Wei discovered several late Palaeolithic sites in Yujiagou (Gai and Wei, 1977)(40°9′48.42′′N, 114°28′51.42′′E). Later, this place was named Yujiagou Microliths Site and a protection monument was built. A very thick gravel layer covered by Malan loess is exposed on the east side of the road leading to the monument. Within a road cut about 150 m long, a large number of continuous involuted and penetrating structures both in the gravel layer and the Malan loess can be observed (40°9′52.33′′N,114°28′44.47′′E) (Fig. 2). The thickness of the curled gravel layer is between 2.5 m and 3 m(Fig.2A).Under the action of external forces, the gravel was almost completely changed from its original normal state of deposition and re-arranged in a“twisting and rotating”state (Wu, 1979, 1980). Some gravels even appear upright or inverted (Fig. 2B, E). The scale and degree of curling deformation of the gravel layers are exceptionally large, and some of the overlying Malan loess is stirred in,which makes the loess layer appear as an obvious traction fold(Fig.2B,C).We will discuss the causes of these structures in the next section.

3.2. SSDS near Hongya village

The Hongya section is located on the Huliu River Fault, which could be regarded as a branch of the Liuleng Mountain Fault. In this work, we conducted preliminary observations on the upper and the middle part of the section and found two types of SSDS.

The first type of SSDS is located just between the top layer of Malan loess and the gravel layer. We observed some wedge-shaped features composed of loess and gravel in a road-cut section (40°8′13.86′′N,114°39′10.16′′E)(Fig.3A;Fig.1A).This kind of feature looks particularly like gravity-loaded structures at first glance. On closer inspection, however, we find intrusion marks from soft formation with more loess and sand to a hard formation dominated by gravel. The feature in Fig. 3B is the intrusion of loess and sanddominated body into the gravel, which illustrates a similar situation more clearly even though it is only 20 cm high.

The second kind of SSDS occurred 8 m below the first type. The deformed structure is in the form of interlayer folds and flames (40°8′16.01′′N,114°39′12.61′′E)(Fig.3C).Careful observation reveals that there is a characteristic of the upward flow of sand,the deformation phenomenon in this layer tends to weaken from bottom to top(Fig.3D). In the lower part of the layer, the bedding of the sediment was disturbed or almost disappeared,replaced by a spiral structure (the green arrow). In the upper part of the same layer,the original lacustrine beddings were also disturbed but still visible. We will discuss the deformation mechanism with other SSDS in the next section.

3.3. SSDS at Maliang Palaeolithic Sites near Donggutuo village

Donggutuo is a very small village located on the eastern edge of the Nihewan Basin (Fig. 1,40°13′8.86′′N;114°40′54.50′′E).Many palaeolithic sites have been discovered around this village. The most famous one is the Xiaochangliang Site(40°13′10.00′′N;114°39′44.00′′E), which is about 1.2 km on the west side of Donggutuo village. A lot of palaeoliths were found by You and his colleagues in 1978 and were first reported in 1979.The age of the palaeolithic layer was judged to be between 2.43 and 2.55 Myr by comparing the paleomagnetic data of a nearby section(You et al.,1979,1980).Now the age has been revised to 1.36 Myr(Zhu et al., 2001). In 1981, Wang Wenquan, a local villager,first discovered a Palaeolithic tool in Xujiapo,about 1 km on the northwest side of Donggutuo village(Fig.4A).After that,Prof. Wei presided over the field excavation and named the locality as Donggutuo Site(40°13′23.00′′N; 114°40′16.00′′E) (Wei, 1985, 1991,2014). In 1983, Wei discovered some Palaeolithic products and animal fossils in Maliang, a small place about 500 m northwest of Donggutuo village(Fig.4A).In the same year, Wei named it Maliang Site(40°13′23.02′′N; 114°40′40.47′′E) and did some simple excavation there. Further excavations of Maliang Site were made in 2006 and 2008 (Yuan et al., 2011; Liu et al., 2018). In 2005 and 2008, the Hougou Site(40°13′27′′N, 114°40′49′′E) and the Sankeshu Site(40°13′23′′N, 114°40′50′′E) were successively discovered closely on the east of Maliang Site (Yuan et al.,2011; Zuo et al., 2012). There are more newly discovered Palaeolithic sites around Maliang Site,which could not be mentioned one by one in this paper.Maliang Site has become “one of the sites of ancient human activities in the Early Pleistocene with the largest number of excavations and unearthed remains in the Nihewan Basin” (Li et al., 2010).

Fig.3 Field characteristics of the deformed layers near the Hongya village.Fig.3A,B shows the deformation in the upper gravel layer.Note the progressive multi-phase deformation signs shown by the red arrows in Fig.3A,B.Fig.3C,D shows the curling deformation and flame-like structure in the lacustrine sand layer,which are formed by the liquefaction of unconsolidated sand.Green arrow in Fig.3C shows scale and in Fig. 3D a spiral structure.

Fig.4 SSDS in the earlier excavation pits of Maliang Site Group,near Donggutuo village.Fig.4A is a satellite photo of Maliang Site(Google Earth). Note that the terrain of Donggutuo village is high in the east and low in the west, with high mountains on the east and south sides.Fig. 4B is the excavation pit of Maliang Site (photographed in 2006). The person in the photo is Professor Wei, who is 182 cm tall. The sequential numbers are based on the correlation with Fig. 5. There is an obvious deformation layer about 60-80 cm thick. Sand bodies of different sizes and colors are interspersed with each other,accompanied by crumpling or convolution(Fig.4C).Fig.4D is the excavation pit of ML5(photographed in August,2020).Layer ②and Layer ④are two obvious event layers on the north wall of the pit.The upper event layer is more than 1 m thick, and the upward deformation is aggravated; the lower event layer is severely weathered on the main section of the excavated pit. But it is still possible to see clear involutions and small flow structures on both sides (Fig. 4E,F). Fig. 4G, H shows the main features of ML9(40°13′20.48′′N,114°40′40.43′′E).A group of small normal faults on the east wall and irregular collapse feature on the south wall. Green arrows in Fig. 4C, D indicate positions of scales. The hammer in Fig. 4H is 38 cm for scale.

In 2016,Hebei Provincial Institute of Cultural Relics implemented a project called“Oriental Human Source Exploration Project in Nihewan”, and carried out a special survey of ancient human remains in Maliang area. A series of excavation pits were constructed at Maliang and named by numbers(such as ML3 and ML5).At the 10th locality of Maliang Site(ML10),researchers discovered a cultural layer that is 3 m below the original Maliang palaeolithic layer (Liu et al., 2018)(40°13′21.9′′N, 114°40′40.2′′E). In 2020, a further archaeological excavation was carried out on the basis of ML3 and ML10. The merged new excavation pit carries the previous name of ML10.

Fig. 5 A) A panoramic photo of the ML10 taken by a drone (August, 2021), showing the spatial relationship of SSDS, the normal lacustrine sediments,and the paleo human cultural layers(CS-1 to CS-4).The total thickness of the exposed succession is 15 m.Further description and discussion are detailed below. The person on the left is 165 cm high, the persons in the middle (SSDS-8 layer) are taking paleomagnetic samples.B)Photo of ML3 and ML10 in 2006(from Liu et al.,2018).The CS of ML in Liu's article is the SSDS-4 layer.The CS of ML10 is the CS-4.Note the similarities between the layers of Fig. 5A,B (especially the areas of green arrows).

3.3.1. Excavation pits constructed before 2020

Fig. 4 shows the main features observed in three old excavation pits at Maliang Site, ML5, and ML9.Fig. 4B was taken in 2006 by Prof. Wei. Fig. 4C is a partial enlargement of Fig. 4B. There is an obvious deformation layer on the east wall of the excavation pit, and sand layers of different grain sizes and colors interact with each other. Both the upper and lower layers of the deformed sand layer are basically normal lacustrine fine sandstone.

ML5excavationpit(40°13′20.08′′N,114°40′43.93′′E) is about 6 m in depth and its section can be divided into four layers(Fig.4D).On the north wall, two event layers containing obvious deformations can be observed (Layer ②and Layer ④in Fig.4 D).The lower event layer is severely weathered on the main section of the excavation pit, but strong convolutions and flame structures can be seen on two side walls (Fig. 4E,F). The thickness of the upper deformed layer(Layer ④in Fig.4 D)exceeds 2 m,with an upward increasing deformation density.

Fig.4G,H show the east wall and main section(south wall) of ML9 that opens to the north (40°13′20.48′′N,114°40′40.43′′E). On the east wall, there is a group of near-upright small normal faults with a fault distance of 5-10 cm (Fig. 4G); the original sediment layering can be seen on the south wall.The phenomenon of rupture,the original rock gravels of different sizes “float”intermittently in the sand layer, and there is still a looming relationship between the normal rock layers that have not cracked on both sides(Fig.4H).

3.3.2. Excavation pit which began construction in 2020

The current ML10 excavation pit is located about 30 m south of Maliang Site. ML10 opens to the south and west respectively and has a depth of 15 m. Four ancient human cultural layers have been unearthed there, which are marked as CS1-CS4 from top to bottom in accordance with archaeological work (Fig. 5).We have found 10 SSDS layers formed by obvious abnormal events. They are named SSDS-1～SSDS-10 from bottom to top. There is another vertical type of deformation features that penetrate downward through multiple layers of sediment on the north and east walls of the excavation pit and are collectively named as SSDS-11 (Fig. 5).

Fig.6 The lowest lacustrine layers at ML10.The rhythm of the sand layer and argillaceous silt layer in the lowest lacustrine deposit of ML10 is cut by a vertical microcrack to form a sand vein. The sand vein of the uppermost layer is connected with the horizontal sand layer at the middle (green arrow) to form a branch vein with a straight angle. The people are working on the excavation pit to search for artifacts and bones.The vein ended at the boundary between the SSDS-1 and the erosion surface.A thrust of slump deposit(SSDS-2)cuts both the vein and the cultural layer (CS-4).

SSDS-1 layer is located at the lowest part of ML10,on top of it is the fourth cultural strata (CS-4) (Figs. 5 and 6). SSDS-1 actually is a sand dike intruded into a set of interbedded lacustrine sandy and argillaceous sediments with pronounced horizontal bedding. A micro-fault with a width of 1-2 cm occurs nearly vertical.The hanging wall of the fault is shifted downwards by 1-2 cm, and the fault fissure is filled with sandy sediments to form a sand vein. The inner sand of the vein is not sorted, but the vein itself is obviously different from the sedimentary bedding around them.

SSDS-2 layer occurs above the fourth cultural layer(Figs. 5 and 7A) with a thickness of 60-80 cm of fine sand and silt. There is a group of small interlayered thrust faults arranged at intervals of 80-120 cm. The average dip angle of these faults is 22°. The thrust surface tends to face the center of the Nihewan ancient lake on the west side.The upper walls of these thrusts have clear traction folds (Figs. 5 and 7A). The boundary between SSDS-2 layer and the normal sediments above it is not clear. Judging from the stacking pattern of the thrust faults and their properties, we infer that SSDS-2 is an underwater slump. We will discuss its origin later.

Fig.7 Field characteristics of SSDS-2 to SSDS-6 at ML10,the colors of sediments have changed from time to time.Fig.7A shows the thrusts of SSDS-2. Fig. 7A, B shows that there is a “normal” lacustrine layer between SSDS-3 and SSDS-4. The normal lacustrine layer completely disappeared in the east part of Fig. 7C. Fig. 7D-G shows detailed parts of SSDS-3 and SSDS-4.

The SSDS-3 layer is located 80 cm above SSDS-2.The bottom of SSDS-3 is straight and clear (Figs. 5 and 7A). The contact between SSDS-3 and SSDS-4 is more complicated.On the west of the excavation pit,there is a 40-50 cm thick normal sedimentary layer sandwiched between them (Figs. 5 and 7b). From the western side of the pit to the inside,the normal layer rapidly wrinkled and thinned until disappeared. As a result,the contact between SSDS-3 and SSDS-4 quickly changed from clear to blurred and the thickness of SSDS-3 has also changed a lot.Its thickness on the west side of the excavation pit is generally about 80 cm thick, with various interspersed, folds and other deformed structures in the middle. As it gradually moved closer to the center of the excavation pit, its upper part became violently undulating. The bottom of SSDS-4 layer is first in contact with the normal lacustrine wedge and then transitions to directly cover SSDS-3. Therefore, the bottom boundary of SSDS-4 layer is an uneven wavy surface (Figs. 5,7A, B, EG).

Above the layer of SSDS-4 is a thick cultural layer with a large number of stone artifacts and animal bone fossils.From bottom to top,this cultural layer is further divided into three cultural sections and named as CS-3,CS-2, and CS-1 respectively (Figs. 5, 7A, C, F). It was also the target layer of the nearby Maliang Site (e.g.,Schick et al.,1991;Wei,1991;Wang et al.,2005).

From the contact surface between SSSD-4 and CS-3, it can be easily recognized that CS-3 is a wedgeshaped layer (Figs. 5, 7C,F), thicker in the southeast,and thinner in the northwest. CS-2 and CS-1 are basically horizontally distributed.The sediments in SSDS-4 layer below the boundary line have obvious wrinkles and abnormal flow. Their origins will be further analyzed later.

About 20 cm above CS-1, there are two deformed sedimentary layers next to each other (SSDS-5 and SSDS-6).The color and grain size of the two deformed layers are quite different, and the boundary between them is clear (Fig. 5). The thickness of SSDS-5 at the lower part is between 1 m and 1.2 m. It is mainly composed of yellow-green silt and gray-white silty clay layers,with more sandy content than the bottom.As the exposure time increases, the surface of the layer gradually turns white, which is caused by the precipitation of calcium in the sediments (Figs. 5,8A,B). Many sand bodies of different grain sizes and colors curl and overlap each other to form a series of oblique flame-like structures or dense folds in the upper part.The axis of most of the folds is inclined to the east with an inclination angle of about 30°(Fig. 8C). There is noticeable zoning in the vertical direction and can be further divided into three smaller SSDS (Figs. 8C and 10).

The SSDS-6 layer is composed of interbedded sandy sediments and argillaceous sediments. The thickness varies between 40 and 60 cm. The lowermost part is medium-fine sand with a darker color and extremely weak consolidation. We can see obvious flame structures interspersed into the nearby muddy sediments in Fig. 8D. But this flame layer is not stable and it disappears towards the right(Fig.8A,D).The upward fine sand layer containing silt and clay has a relatively higher degree of consolidation,with obvious folds and curling deformations(the arrows in Fig.8D).The upper part of the layer is cut by an erosion surface, upon which is a normal lacustrine layer (Figs. 5, 8AD). By field tracing and comparison,it is the same level as the deformation in Fig. 4B of Maliang Site.

Fig.8 The basic features of SSDS-5 to SSDS-11.Fig.8A shows the features on the north wall of ML10.The wedge in the upper left is almost the same as the ice-wedge pseudomorph found in the coal mine near Lissa in eastern Germany(see Fig.9).There is also a sacculus feature,which is high on the left and low on the right. The geological body in the vertical direction as shown in Fig. 8B is also an ice wedge pseudomorph left after the ice intruded into the stratum by freezing is replaced by later sediments.Fig.8C is the close-up view of SSDS-5.For the explanation of other parts, see the main text.

Fig.9 The paleo-ice wedge pseudomorphs of Nihewan Basin and their genetic models.The Ice-wedges at Yujiagou of Hutouliang(Zhou et al.,1982)are shown in Fig.9A.The ice-wedges found in a basalt quarry(40°9′31.76′′N,114°6′14.44′′E)near Huiquanzi village of Yangyuan County are shown in Fig. 9B. The ice-wedge pseudomorphs formed by freezing-and-thawing at ML10 in different geological periods (SSDS-11 and SSDS-8)are shown in Fig.9C.The left side of the picture shows the ice-wedge developed on the north wall of ML10,which is very similar in structure to the ancient ice-wedge in the Lissa coal mine in eastern Germany (French, 2018). The nearly horizontally distributed pits in the middle of the picture (SSDS-8) are actually ice-wedges formed earlier in the periglacial environment. Fig. 9D and E shows vertical sections through artificial ice-wedge pseudomorphs by Harris and Murton(2005).The morphological features of Fig.9D and E are almost identical to those in Figs. 9C and 4H.

Fig. 10 Relationship of SSDS,seismic events,cultural sections and the relative water level of Nihewan Basin.

Both SSDS-5 and SSDS-6 are cut by the vertical feature of SSDS-11 on the east wall.

SSDS-7 layer is the most significant and typical soft-sediment deformation layer in ML10, with a thickness of about 1.4 m. The stratum with a high degree of consolidation in the lower part is mainly crumpled, and the main body is the interpenetrating,curling, and large flame-like structures between the sand and argillaceous layers(Figs.5,8A,B,E).There is also a minor deformed layer of flame-like sand,which is only 1-2 cm thick (the red arrow in Fig. 8E).

SSDS-8 is mainly composed of silt and fine sand,with obvious color and morphological characteristics.The color changed from light grayish-green to the current white within a few months. The shape was irregularly wavy on both walls of the pit, and both were cut off by the vertically distributed SSDS-11(Figs.5, 8A,B). Truncated part of this layer can be seen on the east wall,as it“drops”into the vertical feature of SSDS-11 (Fig. 8B).

SSDS-9 and SSDS-10 layers are two deformed layers with similar thickness (20 cm) and deformed style. Both are represented by the interspersion between the yellow fine sand layer and light gray argillaceous silt layer. Standard flame-like structures can be seen locally. The two layers are separated by a 10-15 cm normally deposited muddy silt layer(Fig. 8A,B).

There are two independent vertical geological features on the east and north walls of the excavation pit (Figs. 5, 8A,B). The one on the east wall is embedded from SSDS-9 down to SSDS-5 or even lower.But there is a vein at the top that intrudes upward into the normally deposited formation (Fig. 8B). The vertical feature on the north wall penetrates vertically downwards into the normal sedimentary layer below SSDS-7(Figs.5 and 8A).The characteristics of the two vertical geological features are collectively named SSDS-11,and will be further discussed in the following genetic analysis.

4.Interpretation of the SSDS origins

There are hundreds of excavation pits in the Nihewan Basin, most of which have been covered by loess or damaged by erosion. We focused on a limited number of excavated pits near Maliang Site. From previous description, it is evident that there are abundant SSDS in Nihewan Formation, but they have not attracted enough attention from geologists and archaeologists, because most of these features are simply considered to be the periglacial phenomena of freeze-and-thaw processes. Geologists who study neotectonic movement and earthquake prediction have focused on fault activities and surface features near the fault zones. Their exploration trenches are relatively shallow and short,revealing only the ancient earthquake relics at 10,000-year level.Archaeologists are dedicated to the discovery of ancient human fossils and stone tools in the strata.So far,there has been no systematic study on the SSDS in the huge Nihewan layers.

Correctly identifying the triggers of the SSDS is important for interpreting the tectonic formation process and the paleoclimatic evolution history of the sedimentary basin(Mills,1983;van Vliet-Lano¨e,1988;Hibsch et al., 1997; Wheeler, 2002; van Vliet-Lano¨e et al., 2004; Feng et al., 2016b; Qiao et al., 2017).For example, the recurrence interval of the major earthquakes is usually several thousand years (Xie et al., 2003), while the history of modern seismic instruments is only more than one hundred years, and the longest history of human beings with written records is only a few thousand years.The geologic study of seismic-induced SSDS can make up for the limitations of these records (Obermeier, 1996, 1998;Wheeler, 2002; Sakai et al., 2015; Qiao et al., 2017;Lima et al., 2021). Sometimes, earthquake-induced SSDS can provide an accurate isochronous comparison basis for regional stratigraphic correlation (Qiao and Li,2009;Qiao et al.,2017; Morsilli et al., 2020).

4.1. Triggers and identification? Criteria of SSDS

Many triggers or factors can lead to the formation of SSDS in the strata of a sedimentary basin (e.g.,Owen and Moretti, 2011; Shanmugam, 2016; Feng et al., 2016b). Some factors belong to the depositional processes of the basin and were termed autokinetic by Leeder (1987). Owen and Moretti (2011)named them autogenic or internal triggers, including cyclic liquefaction by waves, the impact of breaking waves,turbulent pressure fluctuations in strong water flows, tsunamis, tidal shear, rapid sediment loading,periglacial thawing in poorly drained sediments, or groundwater movements.Some factors are external to the depositional processes and are named allokinetic(Leeder,1987)or allogenic triggers(Owen and Moretti,2011). Typical examples are earthquakes, volcano eruptions, and extra-terrestrial impacts.

The Nihewan Basin has been evolving for more than two million years. It has both the conditions for the formation of internal dynamic deformation structures and the conditions for external dynamic deformation.But it is impossible to have all the known factors that lead to the formation of SSDS in the world. Some triggers are easily eliminated by the method of exclusion. For example, the extra-terrestrial impact could form unique geomorphological formations on the ground and could trigger a large tsunami in the lake.For a basin with a history of nearly 100 years of geological research, such geomorphological formations have never been found, so the extra-terrestrial impact can be completely ruled out. Similarly, based on previous researchers’ work, comprehensive comparison research of similar environments in the world,and our field investigations,we can roughly determine the autogenic triggers that led to the formation of SSDS in the Nihewan strata were mainly freeze-thaw processes and rapid accumulation. The allogenic triggers are mainly earthquakes caused by tectonic activity and volcanic eruptions.

We first discuss the climate of the Nihewan Basin during the geohistorical period.

As early as 1926, Barbour et al. inferred that the mammalian types collected in the Nihewan beds belong to an environment of moderately warm and moist climate from Late Pliocene to Early Pleistocene age. On the other hand, the carbonate lime and the presence of gypsum in the upper part of the Nihewan beds indicate a growing aridity, and perhaps the complete drying up of the basin(Barbour et al.,1926).After that, many researchers have conducted extensive studies on the environmental evolution of the Nihewan Basin based on special geological and geomorphological features, fossils of animals and plants in the strata, pollen analysis, and paleomagnetic data (Zhou, 1981; Zhou et al., 1982; Wei, 1991;Schick and Dong, 1993; Yuan et al., 1996; Chi et al.,2002; Wang et al., 2005; Deng et al., 2007; Zhu et al., 2007; Hun et al., 2011; Zuo et al., 2012; Yang et al., 2019). A large number of pieces of evidence prove that most of the geological history of Nihewan Basin is generally temperate or even close to a subtropical environment. During the Middle Pleistocene,there were also semi-arid to arid steppe temperate environments (Yuan et al., 2011), and even multilayered stromatolites growing in the warm environments (Xia et al., 1993; Chi et al., 2001).

There is also a lot of evidence to prove that there are some short cold periods of “periglacial environment” in the development process of Nihewan Basin(Wang et al.,1982;Yuan 1989; Yuan et al., 1996; Xia,2001).

The formation and evolution of the Nihewan Basin are in the Quaternary glacial period. The basin has been in a relatively high terrain and a relatively high latitude area. The comprehensive research results of Yuan et al. (2009) showed that the peak period of Nihewan ancient lake development was mainly in a warmer temperate environment or even subtropical environment (Yuan et al., 2009), but the pollen analysis data prove that firs and spruce,etc.,which prefer low-temperature environments,are rich in the bottom layer (early Pleistocene) as well as in the upper layer(upper Pleistocene)of the Nihewan Basin(Zhou et al.,1982).The lower cold period stratum is exposed in the Nangou of Hongya village, which is called the Nangou cold period, during which, the annual average temperature was between −1.5°C and 3°C (Zhou et al.,1982). The upper cold strata were discovered in Houjiayao Site, which is located in Liyigou Gully between Xujiayao village of Shanxi Province and Houjiayao village of Hebei Province (Chia and Wei, 1976; Zhou et al., 1982; Hun et al., 2011). Zhou et al. (1982)and Yang et al. (1983) found dozens of ice wedges in the Late Pleistocene strata in Hutouliang village(Fig.9A).and Xubao village(Fig.1A)respectively.The authors of this paper also found several ice wedges in the basalt just under the Malan loess in a quarry (40°9′31.76′′N, 114°6′14.44′′E) near Huiquanzi village,Yangyuan County(Fig.9B).Ice wedges can basically be considered as an indicator of a past permafrost environment (Washburn, 1980; Cui, 1980; Yang et al.,1983; Van Vliet-Lano¨e, 1988; Cui et al., 2002; Harris et al., 2018). Based on the occurrences of icewedges, sand-wedges, frost cracks, and periglacial involutions, Cui et al. (2002) established the permafrost boundary of North China since the late Pleistocene and four frost stages. The western part of Nihewan Basin is just located in the south boundary of the permafrost area. Based on analysis results of thermoluminescence and carbon samples from the pollen contents in the wedges,Cui et al.(2002)found four major periglacial stages, from 33 ka to 8 ka. All these facts show that the Nihewan Basin had the conditions for the formation of freezing-and-thawing or cryogenic deformations in some specific periods of its geological history.

As an active rift basin,the Nihewan Basin has been in a stage of continuous and strong activity. In the early 1930s,Teilhard and Young have found evidence that the Shanxi Plateau (Shansi Plateau) has been moving throughout the Pliocene period (Teilhard and Young,1933). Researchers have found evidence that the Fenhe River-Weihe River Graben System(including the Nihewan Basin)has experienced continuous and strong extension,at least since the late Miocene(e.g.,Kuo and Hsia,1956;Cen et al.,2015).As mentioned before,the active faults of smaller scale are very common in Quaternary unconsolidated sedimentary layers(e.g.,Dong et al.,1996;Yang et al.,2019)(Fig.1E,G).From 900 ka to 90 ka, the eruption of Datong Volcanoes was quite frequent (Figs. 1, 9D,E) (Hu et al., 2011). Frequent eruptions of Datong volcanoes are bound to be accompanied by frequent earthquakes.The tectonic setting is the prerequisite before we further analyze the triggers and formation mechanisms of all SSDS in Nihewan Basin.Therefore, the frequency of earthquake events in the ancient Nihewan lake is much higher than that of freezing-and-thawing events.

In the study of the causes of SSDS, it has to be recognized that there is basically no complete one-toone correspondence between the trigger and the formation mechanism of SSDS.The same trigger can lead to the formation of different types of SSDS. While,similar SSDS can have completely different triggers.For example, seismic liquefaction can form various forms of SSDS such as flame structures,liquefied veins,liquefied folds, and so on. The curling deformation in the stratum can be caused by either seismic liquefaction or freezing-and-thawing caused by climatic factors. To confuse the trigger and the formation mechanism is a fundamental problem in the current SSDS research (Shanmmgan, 2016; Feng et al., 2016;Feng, 2018; Su and Qiao,2018).

Many scholars have made useful contributions and proposed multiple criteria and methods to determine the triggers of SSDS(e.g.,Seilacher,1984;Owen,1987;Obermeier, 1996; Montenat et al., 2007; Van Loon,2009; Qiao and Li, 2009; Ettensohn et al., 2011;Owen and Moretti, 2011; Moretti and van Loon, 2014;Su and Qiao, 2018). In summary, there are mainly the following criteria and consensus:

1) Tectonic settings: Most SSDS are related to paleofaulting activities, distributed in a belt along a large area, that is, they are distributed on a considerable scale in space, or are located in a basin with tectonic activity at that time(e.g.,Qiao et al., 2017; Jones and Omoto, 2000). The sedimentary basin should have experienced tectonic activity at the time when the seismic-triggered SSDS were formed (Moretti and van Loon, 2014).

2) The instantaneity and isochronism:An earthquakeinduced SSDS formed more suddenly and violently than a non-seismic alternative feature (Wheeler,2002). For example, the 2008 Wenchuan Earthquake lasted less than 100 s and the longest sand boil process lasted about 20-30 min (Chen et al.,2013). While the deformations caused by repeated freezing-and-thawing are mostly formed over a long period of time,often over hundreds to thousands of years (Van Vleit-Lan¨oe, 2004; Liljedahl et al.,2016). Researchers have discovered the oldest relic ice in the Canadian Yukon that has been over 740,000 years old(Froese et al., 2008).

3) Vertical repetition: Normally, a seismic event mainly affects the most recently deposited strata(e.g., Selley, 1969; Owen, 1995). The strata in the basin must have multiple deformed layers in vertical direction, reflecting the periodicity or repetition of ancient earthquakes. The seismic triggered SSDS should occur in laterally continuous recurring horizons, separated by undeformed beds (Sims 1973, 1975; Owen and Moretti, 2011; Moretti and van Loon, 2014).

4) Zoned distribution: Large-scale earthquakes can form a large number of deformed structures over a large area almost simultaneously. The distribution of seismic-triggered SSDS is closely related to the magnitude of the earthquake and the distance from the epicenter. The greater the magnitude, the larger the deformation area (Kuribayashi and Tatsuoka, 1975; Wu, 1980; Allen, 1986; Obermeier et al., 1991; Pope et al., 1997; Obermeier et al.,1998; Wheeler, 2002). For example, the size of the seismic-triggered sand dikes usually decreases with increasing distance from the epicenter of the earthquakes Conversely, the zoned distribution features can also be used to estimate the location and intensity of paleo-earthquakes (Obermeier et al.,1991,1998).

5) The actualism:The SSDS should be comparable with structures triggered by modern earthquakes (Sims,1973,1975;Obermeier,1996,1998;Moretti and van Loon, 2014).

6) Synchroneity:Much of the geologic record of a large earthquake forms synchronously(Wheeler,2002).If two or more different types of SSDS which are mostly interpreted as seismic origin occur together in the same layer, they are mostly triggered by an earthquake (Obermeier, 1996; Qiao et al., 2017;Lima et al., 2021).

7) Scientific exclusion method: There are many triggers for SSDS, but specific to a certain region or a certain basin in a certain historical period,the main or exclusive trigger can be obtained by the elimination of other impossible triggers(e.g.,Jones and Omoto, 2000; Qiao et al., 2017). For example, the large meteorite impacts can be easily eliminated because there were no corresponding impact features in Nihewan Basin.

It must be stressed here that each of the above criteria has its own limitations and exceptions. There are even contradictions or objections between different criteria. They must be applied carefully. Different scholars have their own opinions.It is often necessary to apply two or more criteria at the same time to determine the trigger and formation process of SSDS.

4.2. Soft deformation structures triggered by cryoturbation

The melting of the ice in the strata could cause the supersaturation of the water in underground sediments,resulting in the formation of various SSDS,such as involutions, gravity collapses and load structures,etc. (e.g., Van Vleit-Lan¨oe, 1988; French, 2018).

In Nihewan Basin,deformation structures triggered by cryoturbation include folding and curling features mainly found in the upper gravel layer (Figs. 2, 3A,B)and the wedge-shaped features penetrated into the lacustrine layers (Figs. 4H, 5, 8A,B).

The most typical cryoturbation folds occur at the upper part of a road-cut near the monument of Yujiagou Microliths Site(Fig.2).We can see from Fig.2 that the loess actively intrudes into the gravel layer(red arrows in Fig.2B,D,E).The closer the distance to the gravel layer is, the more gravel is mixed into the loess. Based on the topography, it can be roughly judged that that loess has invaded downward at least four times (Fig. 2B,E).

Under normal circumstances, it is impossible for soft loess to intrude into the gravel layer so deeply and actively. These phenomena cannot be explained by normal wind or hydrodynamic forces. Can these folds be formed by seismic liquefaction? The multi-phase intrusion phenomenon shows that the formation time is rather long (at least several years), while the earthquake vibration can only last for dozens of seconds. The liquefaction effect after the earthquake,including the adjustment effect after liquefaction,generally last for several or tens of minutes. An extreme example is the sand boil process of the 1976 Tangshan Earthquake in China,which lasted about 3-5 days (Wu, 1980). Therefore, such structural phenomena cannot be interpreted by loading or by earthquake vibrations, or earthquake-triggered liquefaction.

The morphological characteristics of the “loess intrusion”are similar to the“ice wedge”formed in the periglacial environment. Considering that Zhou et al.(1982) and Yang et al. (1983) found dozens of typical ice wedges locally or nearby (Fig. 8A), the causes of the deformation structure shown in Fig. 2B, D, E can only be explained by freezing-and-thawing.During the downward development of the freezing-and-thawing process, as the volume of the ice body changes, the surrounding consolidated gravel produces a flow-like deformation phenomenon, the occurrence of gravel changes accordingly (shown by the green arrows in Fig. 2C,E), and the ice body containing loess that intrudes into the gravel layer does not form a typical wedge but forms a non-standard column or “boots”shape. During the melting of ice, the ice space was gradually replaced by muddy water containing fine sand and “loess”, and finally, the current structural phenomenon was formed.

Similarly, the curling and folding features in the gravel layer near the upper part of the Hongya section(Fig. 3A,B) have occurred in the same position as that of Yujiagou.The upper loess and fine sands also showed downward intruding into the gravel layers. So, the features like Fig. 3A,B in Hongya are freezing-andthawing products.

In the excavation pit at M10 site, there are two vertically developed geological features along the north and east wall(labeled as SSDS-11 in Figs.5,8,9).The feature on the north wall looks very complicated but still has a certain regularity: there are small fractures in the weakly consolidated stratum on both sides,the strata on the right part are nearly horizontal with well-developed en echelon downthrows. The horizontal rock strata that have been broken off are arranged in steps. The strata on the left are disordered. The lower the rock strata, the greater the tilt angle to the right,and the more severe the damage to the original structure of the strata.Only huge pressure from above can form such a complicated feature.The main part of the columnar body in the middle is wedgelike, wider at the top and narrower at the bottom. It roughly starts at the boundary between SSSD-10 and SSDS-9 and cuts down through SSDS-9, SSDS-8,and intersects with the SSDS-7 layer, where a rather huge capsule of sand locates directly on the bottom of SSDS-11, without any bedding (Figs. 8A and 9C). The total height of SSDS-11 and the capsule exceeds 5 m. The part above the capsule is almost identical to the icewedge pseudomorph in the Lissa coal mine in eastern Germany (Eissmann, 2002; French, 2018). Therefore,SSDS-11 on the north wall is an ice-wedge pseudomorph. It is untenable to explain such a complicated feature by the erosion of water, or other geological effects.

The other vertical feature on the east wall of the pit is also an ice-wedge pseudomorph with a huge dimension (at least 10 m high) and a more irregular shape (Fig. 8B). No formal excavation reports or archaeological papers of this area have recorded such features. The person in charge of the excavation believes that it was formed by the erosion of flowing water.However,in its internal fillings,we can see rock fragments that have collapsed just from the original site or nearby surrounding rocks, without any sign of water flow and erosion. In its upper part, there is a cone-shaped 170 cm-long sand vein, which could only be explained by the injection of liquefied sand under great pressure.Usually,this kind of sand vein is a good indicator of seismic liquefication (e.g., Obermeier,1996; Obermeier et al., 2005). However, the melting of a huge ice-wedge can also lead to the supersaturation of the sediments filling in the underground space and forming a variety of SSDS. Cryoturbation-induced intrusion features have been found in many places in Europe and have the name of “amorphous deformations, injection tongues and intrusion features”(Eissmann,2002;French,2018).According to our field observations, this geological body is basically an irregular huge ice-wedge-like feature, similar to the one on the north wall. The current excavation pit has not bottomed out,and the total depth is estimated to exceed 10 m. This 170 cm-high sand vein must have formed by the injection of liquefied sand,but it is not certain whether it was caused by seismic liquefaction,or caused by freezing-and-thawing triggered liquefaction.There is also another possibility:the freezingand-thawing action of a huge ice-wedge provided the material basis (water-saturated sediment), and the later earthquake provided a strong driving force, so that the sediment was liquefied and injected into the unconsolidated strata above.

Harris and Murton (2005) conducted a series of laboratory simulations on forming processes of icewedges. SSDS-11 is remarkably similar to their results of experiment 2,and the phenomena in Fig.4G,H in ML 9 of this paper are almost the same as the results of their experiment 5 (Fig. 9H). Therefore, we believe that the stepped faults on the east wall of the ML9 excavation pit(Fig.4G)and the collapsed structure on the south wall (Fig. 4H) were also formed by a large ice-wedge, but this ice-wedge was not fully exposed.The last time we visited there in August 2020,ML9 has been backfilled.

The sizes of SSDS-11 at Maliang are much larger than the ice-wedges discovered by Zhou et al. (1982)and Yang et al. (1983) in Hutouliang and Xubao village,but their stratigraphic position is basically the same.All these features have occurred just under the Malan loess. Since we have not done age-dating work yet, we temporarily consider the formation time of SSDS-11 at ML10 (Figs. 5, 8 and 9) and ice wedge structures at ML 9(Fig.4H)have the same age as that of Yujiagou valley of Hutouliang (Fig. 2).

SSDS-8 is a structure of multiple connected ditches with wider tops and narrower bottoms, exposed on both the north and east walls of ML10. The depths of ditches vary between 20 cm and 80 cm, but the top surface is basically at the same elevation on both the north and east walls of excavation pit. The on-site construction personnel interpreted it as an ancient surface flowing water gully,but could not explain why there were two flow directions intersecting at right angles on the same surface. In fact, their original spatial distribution was a “lattice” of connected troughs,which were formed by the melted ice-wedges in ancient permafrost(Fig.9C)(Friedman et al.,1971;Liljedahl et al., 2016).

4.3. Earthquake-triggered SSDS

We have examined the cryoturbation folds at Yujiagou, the ice-wedges and ice-wedge pseudomorphs at both Yujiagou and Maliang.Is it true that all the SSDS in Nihewan Basin are cryoturbation products?The answer is“No”.Then,how to identify the SSDS of seismic origins from that of cryoturbation?

The SSDS caused by freezing-and-thawing can only develop in a cold surface environment. For the Cenozoic lakes like Nihewan, ancient ice wedges or freezing-and-thawing folds can only develop at or near the erosion surface after the lake level has receded and the temperature is cold enough. Continuous underwater SSDS occurring in the lacustrine stratum cannot be the product of freezing-and-thawing because subaqueous sediments are usually not affected by periglacial activity (Van Vleit-Lan¨oe,2004). Furthermore, cryoturbation-induced features are usually exceptionally large in scale, from a few meters to even tens of meters thick (Vandenberghe et al., 2016; French, 2018; Zhong et al., 2020), while most of the seismic-triggered SSDS are less than meterscale,and their upper and lower surrounding rocks are mostly normal lacustrine deposits.

4.3.1. Earthquake-triggered SSDS at ML10

A total of 10 SSDS layers, which are interbedded with undeformed normal lacustrine layers,were found at ML10 excavation pit.Among them,the wavy feature of SSDS-8 is interpreted as the well-drained trough network from the degradation of ancient patterned ice-wedges.

Now we will analyze the origins of the other 9 deformation layers from bottom to top.A prerequisite must be kept in mind,that is,the Nihewan Bain is a rift basin with strong tectonic activity and frequent earthquakes. The distance from Maliang Site to the Liuleng Mountain Fault is less than 1 km.

SSDS-1 layer is located at the lowest part of ML10.Above the SSDS-1 layer is an eroded surface,on which,a large number of artifacts and some fragmentary bones have been unearthed and labeled as CS4 by archeologists (Figs. 5 and 6). SSDS-1 is a micro-fault with a width of less than 2 cm and occurs nearly vertical. The hanging (right) wall of the fault is shifted downwards by 1-2 cm, and the crack is filled with sandy sediments to form a sand dike,which cuts across several layers of lacustrine sands.

SSDS-2 is composed of a set of piggyback sequences with back-steepening of imbricate thrust faults.Based on the slump patterns, it can be judged that the direction of the slump is from west to east(Fig.7A),and the slump formed very quickly.

Due to the very limited exposure condition, we cannot yet determine the true scale of the slump.However, Maliang is located in a clearly easterly direction from the proper center of the basin, and the topography of the bedrock at the lake bottom should generally be high in the east and low in the west.If the slump is caused by the difference in the topography of the lake bottom or the rapid sedimentation of excessive sediments around the basin, the direction of the slump is more reasonable from east to west. But it is obviously inconsistent with the field observations,indicating that these factors are not the case.Considering that Maliang is located on the footwall(east side)of the Liuleng Mountain Fault,and the main activity of the main fault is a normal fault dipping westward, it is more reasonable to explain this slump as triggered by a sudden activity of the Liuleng Mountain Main Fault (relatively rapid rise of the footwall).

Alsop et al.has conducted a systematic study on a large number of slump structures in the Dead Sea(e.g.,Alsop and Marco 2012a, 2012b, 2013; Alsop et al.,2016). Their research methods and results can be used as models for similar comparative studies in the Nihewan Basin in the future.

Both the upper and lower layers of SSDS-3 are generally normal lacustrine sand layers. Although SSDS-3 and SSDS-4 are severely deformed, they still retain some of the original lacustrine features,such as laminae and varved mud. The lacustrine strata sandwiched between the SSDS-3 and SSDS-4 were originally horizontal. After experiencing an abnormal liquefaction process, most of the normal sediments were liquefied and mixed together with sediments from both SSDS-3 and SSDS-4 layers. Hence the sandwiched wedge-shaped normal sediments are the residues of large-scale liquefaction. Observing SSDS-3 layer closely, we can clearly see the folds and micro synsedimentary faults of the muddy sediments in the layer. The sandy sediments are wrapped between the folds of muddy sediments.We can also see the flow or interspersed phenomenon of fine sands(Fig.7BD).The sandy sediment of SSDS-4 layer has an obvious abnormal flow in the vertical direction and formed relatively large crumples,diapirs,and flame-like sand veins.The most typical diapir structure and flame-like structure in this excavation pit are all found in this layer (Fig. 7A,B,7D).

The strong liquefaction of SSDS-3 and SSDS-4 layers as well as the “disappearing” of the east part of the normal wedge-shaped sedimentary layer between them are the results of either freezing-and-thawing in cold conditions or earthquakes.The paleo-cultural layers on the paleo-eroded surfaces indicate that the Nihewan Basin at that time was a warm environment, very suitable for animals,plants,and humans,rather than a very cold periglacial zone. As a result, the cryoturbation mechanism can be eliminated. Hence it is quite feasible that from SSDS-3 layer to SSDS-4 layer,at least three phases of earthquake-triggered liquefaction processes have occurred. The fact that SSDS-4 layer is cut by the erosion surface indicates that the seismic events that led to the formation of SSDS-4 and SSDS-3 occurred before the paleo-eroded surface was formed.

The erosion feature between SSDS-4 layer and CS-3 (Fig. 5) indicates that Maliang's topography at that time was high in the northwest and low in the southeast. This feature is basically consistent with the geomorphic features indicated by the SSDS-2 slump.

The lake retreat process after SSDS-4 lasted for a long time. As the water level of the lake slowly increased,the cultural layer transitioned from CS-3 to CS-1.Afterwards,a rapid process of lake transgression took place. Human civilization moved elsewhere, and the stone tools and animal fossils used by the ancients were left at the bottom of the thick layer of lake sediments(CS-1 in Fig.5).Due to the rapid deposition,the early mud and sand are mixed deposited in the lower part.They have a higher degree of consolidation and appear smooth and delicate on the surface of the excavation pit, with a lighter color. The upper sandy sediments are yellowish in color and the surface is relatively rough. When an earthquake occurs, the sediments with high mud content and a high degree of consolidation in the middle and lower parts were dominated by larger plastic folds, and the sediments with high sand content in the upper part were dominated by small but very dense flame structures. The original bedding features often disappear due to liquefaction.SSDS-5 exhibits longitudinal deformation zoning and can be further divided into three smaller event layers (Figs. 8A, C, 10).

The SSDS-6 layer can also be divided into two smaller event layers (Figs. 8D and 10). The sediments of lower part are dominated by liquefied flow structure, mainly flame structures. The mud content becomes richer to the upper part.As a result,the degree of consolidation is relatively strong, and the deformation shows obvious folding and curling. The top of SSDS-6 layer is cut by a small erosion surface(Fig.8D).

The strata in the Nihewan Basin are often damaged by faults or earthquakes of different scales. In addition,the original lacustrine strata usually lack obvious marks. After the ancient lake vanished, the semiconsolidated lacustrine strata are highly susceptible to erosion, contamination, or coverage of sediments also occur from time to time. Nearly all strata still contain a small amount of water. Once they are excavated and exposed, the surface features such as the colors of the sediments will change time after time. These factors have created problems for stratigraphic correlation.Sometimes the strata of very close Palaeolithic sites cannot be correlated clearly, and even misjudgments are made.

Compared with the deformation feature of Maliang Site (Fig.4B,C), it is not difficult to find that they are the product of the same deformation event as SSDS-6.This shows that the same deformation layer can extend far, and the deformation feature can be used as a marker for stratigraphic correlation.

The most intensely deformed layer in ML10 is SSDS-7, which is exemplified by the interpenetrating,enveloping, and large flame-like structures between sand and mud layers(Fig.8A,B,E).The SSDS-7 layer is also the thickest event layer in ML10(SSDS-5 is thicker but it consisted of 3 smaller event layers). All these features show that the event is the strongest.

The flame-like structures in SSDS-9 and SSDS-10 are formed by small-scale liquefied sand. The original rocks of SSDS-5～SSDS-7 and SSDS-9 and SSDS-10 are all normal lacustrine sand, silt sand, and muddy deposits with different contents.

The 15-m-deep pit revealed 10 nearly horizontally distributed SSDS layers. Except for SSDS-8, which was proved to be the cryoturbation SSDS,all other SSDS can be explained by earthquake-related products of liquefaction and slump.

4.3.2. Earthquake-triggered SSDS at Hongya village and Maliang Site

Now we will discuss SSDS in other lacustrine sections in the Nihewan Basin.

Fig. 3C shows the curling feature seen in the silts and fine sands of the Hongya section. The layer of Fig.3C is only several meters below the gravel layers,but the component is mainly the lacustrine sand. The curling feature shown in Fig. 3C is sandwiched between its overlying and the underlying normal lacustrine layers.Several liquefaction flame structures can be recognized in this layer. Combined with the background of active faults developed nearby and especially other similar liquefaction SSDS in Maliang, we interpret the feature in Fig. 3C as the liquefied convolution triggered by earthquake.

The photos of Fig.4B,C were provided by Prof.Wei.They were taken in 1996 from the original Maliang Site.The thickness of the deformed layer shown in Fig.4B is between 60 and 80 cm.Both its upper and lower layers are normal sediments with very thin and clear lacustrine laminae. The original lacustrine laminae of Fig. 4C were thoroughly dispersed, and sand grains of different sizes form a complex pattern but still with flowing traces(Fig.4C).Thus,the feature of this layer is interpreted as seismic liquefaction convolution.It is the same as SSDS-6 of ML10(Figs. 5, 8A,E).

There are two event layers with strong liquefaction SSDS in ML5 (Layer ②and Layer ④in Fig. 4D). Event layer 2 shows a clear upwardly weaker liquefaction bed.The fine sand at the bottom experienced the most serious liquefaction,and the original bedding can only be faintly visible or almost completely disappeared.The liquefied sand surged upwards, forming some obvious diapir-like structures (the red arrows in Fig.4DF).The sediments closely adjacent to the diapir structures in the middle were severely folded or crumpled. As they continued to approach the top of the layer, the crumpled phenomenon of the original lacustrine bedding gradually weakened until it returned to its original flat state.

Layer ④is mainly liquefied convolution, which tends to become stronger upwards,possibly due to the superposition of the superimposed weight of loess and some cold weather effect. Layer ② shows clear liquefaction characteristics and obvious flow direction from bottom to top. Some diapir-like structures are formed by the liquefied fine sand flowing upward in the lower part of the second layer (see the red arrows in Fig. 4E,F). Near the top of this layer, the original lacustrine bedding has not been destroyed, showing a series of crumples.

4.4. The recurrence of seismic events and the application of SSDS

One of the purposes of studying earthquaketriggered SSDS is to evaluate the crustal activity and the seismic recurrence period in Nihewan Basin.Although such an evaluation may not directly help earthquake prediction, it can provide long-period seismicity of this area.

Numerous examples prove that lake sediments are more sensitive to earthquake shaking (e.g., Sims,1973; Sims 1975; Hibsch et al., 1997; Jones and Omoto, 2000; Su et al., 2013). One of the most typical examples is the Van Reservoir in California,USA.After the San Fernando,California earthquake in 1971,Sims found three deformed structural zones in a 1-m-thick layer at the bottom of the Van Normal Reservoir. These deformed structures are related to the moderate earthquakes that shook the San Fernando area in 1930,1952,and 1971(Sims,1973,1975).The three earthquakes correspond perfectly to the three deformed layers, and Sims' observations are consistent with the experimental results of Owen(1996), that is, earthquakes only affect the last deposited layers, and only once. This has become a precondition for judging earthquake frequency when conducting paleoseismic research. It should be mentioned that Owen has pointed out that his experiments were all based on the behaviour of liquefaction in surface sediments, without considering the liquefaction developed in the buried layer.The earthquakeinduced liquefaction, fluidization initially developed in a buried layer behaved quite different from his experiment (Owen, 1996).

After basically ascertaining that most of the SSDS exposed at ML10 were earthquake-origin, we can discuss the earthquake recurrence in the Nihewan Basin.This requires two types of data,one is the age of the deformed layers, and the other is the earthquake times.

For the age of SSDS-11, we use the C14age value obtained by Professor Zhou Tingru for the lower part of the Jiagou gravel layer in 1982,that is,27.675 ka.

The magnetostratigraphic data in the early 1990s show that the 50 cm thick Maliang palaeolithic layer is in magnetically normal sediments just above the Matuyama-Brunhes boundary and the age of the palaeolithic layer is a little less than 0.78 myr (Schick and Zhuan, 1993). More recently detailed magnetostratigraphic studies have moved the Brunhes/Matuyama boundary above the palaeolithic layer of Maliang Site, and found the palaeolithic layer at Donggutuo Site is just below the Jaramillo onset.Therefore, the ages of the Maliang and Donggutuo cultural layers are considered to be 0.78 myr and 1.1 myr,respectively(Wang et al.,2005;Zhu et al.,2007;Deng et al., 2007; Yang et al.,2019).

SSDS can be used as a good marker for strata correlation. Field correlation shows that the CS1～CS3 unearthed in ML10 are equivalent to the Maliang palaeolithic layer (Fig. 5) and should have the same age of 0.78 myr. The CS-4 is equivalent to the palaeolithic layer which was first discovered in 2016 (Liu et al. in 2018). The latest age data of CS4 has not yet been published yet. However, Liu calculated the age of CS4 as 0.80-0.90 myr by deposition rate as well as referring to the optical luminescence age of nearby archaeological sites and the position of the B/M boundary.

The linear distance between Maliang Site and the Donggutuo Site is only 580 m (Fig. 4A). Searching and excavation of palaeolithics near the two sites have been carried out for many years. The possibility of rediscovering the palaeolithic layer between CS-4 and the Donggutuo cultural layer is very small. It is also reasonable to infer that the CS-4 in ML10 is the Palaeolithic layer of Donggutuo Site and has the same age of 1.1 myr.

Since the SSDS-1 layer is closely below CS-4,we can roughly use CS-4's age as the age of SSDS-1. Thus, we currently have two age data for SSDS-1, namely 0.8-0.9 million and 1.1 million. For the convenience of calculation,we take the average age number of CS-4 as 0.85 million years. So, the time interval from SSDS-1 to SSDS-11 is roughly 0.82 ma or 1.07 ma.

The number of seismic events is basically determined by the number of deformed layers triggered by the earthquake.However,it should be noted that due to the inconsistent standards for layering in the field,multiple deformation layers are sometimes classified as one layer.For example,SSDS-5 actually consists of 3 secondary deformation layers.SSDS-6 and SSDS-7 have two event layers, and SSDS-3 to SSDS-4 are formed by at least three seismic events. Thus, from SSDS-1 to SSDS-10,there are at least 14 seismic events(Fig.10).

Assuming that these 14 earthquakes are all basinscale events,the earthquake recurrence period of the Nihewan Basin is approximately 59 ka～76 ka.Either of these two ages is far greater than the general empirical value and also larger than the results revealed by the trenches(Yin et al.,1997;Xie et al.,2003).

There are three most likely reasons for this result.First, under the control of tectonic activities and climate change, the water level of the Nihewan ancient lake often fluctuates up and down. Xia's research (Xia, 1992) shows that there were four obvious lake transgression and lake regression processes in the Nihewan Basin. During the regression period, no new earthquake shaking can be recorded,and the previous earthquake records in loose sediments are exposed on the surface and are easily eroded away.The rapidly rising lake water during lake transgression can also damage or transform the previous paleo-earthquake records.Maliang is located in the northeast marginal area of Nihewan Lake, and the unsolid SSDS are easily affected or damaged by the fluctuation of lake water level. Three large erosion surfaces were identified in ML10, two of which were occupied by ancient humans, and the other was occupied by permafrost ice wedges due to the cold climate.

The second possible reason is that our criteria for discriminating between deformed layers and the number of events need to be carefully applied or modified.Only those SSDS with definite characteristics can be regarded to have formed by one earthquake.For example, SSDS-2 is formed by an underwater slump, with a set of distinctive back-steepening of imbricate thrust faults.

Searching the evidence of multi-phase seismic disturbance in a thick deformed sedimentary layer needs careful observation. Sometimes it is necessary to observe the interspersed relationship between sediments under a microscope to determine which are formed by the deposition process and which are caused by seismic liquefaction.

The third possible reason is that dating technique is not precise enough, and the time of the sedimentary discontinuity could not be accurately constrained.It is hoped that a more accurate and reliable dating method can be found. Future research on the composition and chronology of the possible syndepositional volcanic eruptive material may identify which earthquake-triggered SSDS are caused by the Datong volcanic eruptions.

Finally, all the event layers and larger erosion surfaces we identified and the cultural layers identified by archaeologists at ML10 are shown in Fig.10.

5.Conclusions

The Nihewan Basin has a pivotal position in the fields of Quaternary Geology, Paleontology, Paleomagnetism,Palaeolithic Archaeology,and Early Human Evolution(Zhu et al.,2007).Many stone tools and bone fossils have been unearthed in Maliang Site and surrounding sites. There are still much more treasures buried in the thick strata below and beyond Maliang.The densely populated SSDS is like a treasure guide,leading us to explore the evolutionary history of Nihewan Basin from another direction.

The main conclusions of this article are as follows:1) There are a large number of SSDS found in the

Nihewan Basin.According to their genesis,the SSDS can be divided into two categories: cryoturbationrelated SSDS in the periglacial environments and the earthquake-triggered SSDS.

2) Based on the comparison with typical periglacial features in Europe and North America, we identified that the special structures vertically distributed on the east and north walls of ML10 are the remains of large ancient ice-wedges formed by freezing-and-thawing in the periglacial environment. The pointed vein that developed upward from the top of the ancient ice wedge on the east wall is probably the result of the dual effects of freezing-and-thawing and ancient earthquakes.SSDS-8 and its related erosion surface are identified for the first time in this work. SSDS-8 is a relic of ancient ice wedges developed in the lacustrine strata during the cold period. It can be used as a good indicator when conducting stratigraphic comparison studies for the distinctive shape and stable distribution.

3) A total of 9 earthquake-triggered SSDS layers and 14 seismiceventshavebeenidentifiedinML10.TheSSDS have a variety of forms, including flame-like structures, convolutions, diapirs, inject filling and piercing veins along the cracks, and underwater slump,etc.They have important reference values for the long-period seismicity and tectonic evolution history of Nihewan Basin. The cause of the earthquake may be attributed to the activities of the surrounding fault zones,or the magma eruption of the Datong Volcanic Group that began 900,000 years ago.

4) The original lacustrine facies strata in the Nihewan Basin lack obvious signs for correlations.Once they are excavated and exposed, the surface features such as the color of the sediments will change quickly. The semi-consolidated or even completely loose lacustrine facies stratum is highly susceptible to erosion or contamination or coverage by sediments in other layers. The strata are often damaged by faults or earthquakes of different scales. All these reasons have caused many problems for archaeological and stratigraphic correlations.Sometimes the strata of very close fossil sites cannot be correlated normally, and even misjudgments are made.

The SSDS sandwiched between normal strata usually have distinctive shapes,and they can be regarded as the benchmarks for stratigraphic correlation.In the study of paleomagnetism and palynology, special attention should be paid to the disturbances of paleoerosion, the paleo-cryoturbation, and paleoearthquakes on the strata. The sampling and statistical methods at equal intervals need to be improved based on actual geological conditions.

This article is mainly the author's preliminary summary of the field geological phenomena in the Nihewan Basin in the past two years. More extensive field investigations and systematic sampling and testing need to carried out.It is hoped that in our future study of the Nihewan Basin,we can integrate the SSDS investigation with sedimentology, stratigraphy, paleomagnetism,paleontology, neotectonic and volcanic activities for comprehensive research and judgment.

Funding

This work was financially supported by the National Natural Science Foundation of China (41772116).

Authors' contributions

De-Chen Su:Conceived and designed the analysis,collected the data,contributed data or analysis tools,performed the analysis, wrote the pater;

Ai-Ping Sun: Collected the data, contributed data or analysis tools, performed the analysis;

Zhao-Li Li:Collected the data,contributed data or analysis tools, performed the analysis;

Song-Yong Che:Collected the data,performed the analysis;

Zhen-Jie Wu: Contributed data or analysis tools,performed the analysis.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

During the field investigation and thesis writing,we have always received strong support and help from Professor Qi Wei.During the field investigation,we also received help from Mr.Shi-Jun Bai,Mr.Chi-Chun Huang and Mr. Shang Gao. Together we want to express our gratitude to them! We would especially like to thank Professor Ian D. Somerville, who helped us carefully reviewing the manuscript, providing many useful suggestions and comments,and polished the English of the manuscript. This research was supported by the National Natural Science Foundation of China(41772116).

We would like to dedicate this document to Mr.Xiu-Fu Qiao, a Chinese geologist.