Regression-transgression cycles of paleolakes in the Fen River Graben Basin during the mid to late Quaternary and their tectonic implication

Abstract

An investigation into lake terraces and their sedimentary features in the Fen River Graben Basin shows that several paleolake regression-transgression cycles took place during the mid to late Quaternary. The horizontal distribution of the lowest loess/paleosol unit overlying each lake terrace indicates the occurrence of four rapid lake regressions when paleosols S₈, S₅, S₂, and S₁ began to develop. The horizontal distribution of the topmost loess/paleosol unit underlying the lacustrine sediment in each transition zone between two adjacent terraces indicates that following a lake regression, a very slow lake transgression occurred. The durations of three lake transgressions correspond to those of the deposition or development of loess/paleosols L₈ to L₆, L₅ to L₃, and L₂. It is thereby inferred that regional tectonic movement is likely the primary factor resulting in the cyclical process of paleolake regressions and transgressions. Taking these findings along with published geophysical research results regarding the upper mantle movements underneath the graben basin into account, this paper deduces that a cause and effect relationship may exist between the paleolake regression-transgression cycles and the tectonic activity in the upper mantle. The occurrence of a rapid lake regression implies that the upwelling of the upper mantle underneath the graben basin may be dominant and resulting in a rapid uplifting of the basin floor. The subsequent slow lake transgression implies that the thinning of the crust and cooling of the warm mantle material underneath the graben basin may become dominant causing the basin floor to subside slowly. Four rapid paleolake regressions indicate that four episodic tectonic movements took place in the graben basin during the mid to late Quaternary.

Keywords

Fen River Graben Basin / lake terrace / paleolake regression/transgression / tectonic movement

Introduction

Climate change and tectonic movement are two important factors controlling the changes in lake water levels. Generally, understanding how dry-moist paleoclimate changes resulted in paleolake level fluctuations is straightforward (e.g.,Urabe et al., 2004; Briggs et al., 2005; Shanahan et al., 2006; Eriş, 2013; Massaferro et al., 2013), but comprehending how regional tectonic movements affected paleolake regression-transgression cycles is complicated. For tectonic movement, tectonically controlled differences in basin geometries, catchment sizes can result in paleolake regression-transgression cycles (e.g.,Bergner et al., 2009). The gradual subsidence/uplift of lake basins or volcanic activities in or near the basins can cause variations in lake levels (e.g.,Cukur et al., 2014; Ross et al., 2014) with the tectonic tilting movement causing dissimilar changes in lake levels on either side of the basin (e.g.,Delvaux et al., 1998).

To be extensional, the Fen River Graben Basin (FRGB), one of a series of graben basins around the stable Ordos Massif, has been tectonically active since the formation in the Pliocene epoch (e.g.,Xu and Ma, 1992; Li et al., 1998; Zhang et al., 1998). During the Quaternary, the basin was perennially occupied by water forming a great paleolake and depositing thick lacustrine sediment (Wang et al., 1996). The variation of the lacustrine sediment in color, pollen composition, chemical composition, and grain size indicate that the paleolake level often changed over time.Mo (1991) and Dong et al. (2001) adopted paleoclimate changes to account for these variations, while other researchers attributed these changes to the neotectonic movement (e.g.,Yang, 1987; Li et al., 2004). The actual cause and mechanism for these paleolake regressions and transgressions in the basin are still not well understood; however, a study on the process experienced by these regressions or transgressions could help us better identify these geologic events.

Previous Quaternary studies on the FRGB have typically focused on geomorphic-sedimentary features exposed in some drillings and stream-incised sections (e.g.,Mo, 1991; Wang et al., 1996). Due to the lack of continuous tracing data about the geomorphic-sedimentary features along some transverse sections, information about the specific processes of paleolake regressions or transgressions is lacking. In this paper, we first present some geomorphic-sedimentary survey results to reveal the timeframe and detailed process of the cyclical paleolake regressions or transgressions during the mid to late Quaternary in the basin, and analyze the factors that controlled the process. We then take these new findings and previously published geophysical research results regarding the upper mantle movements underneath the graben into account to discuss the tectonic implication of the paleolake regression-transgression cycles. The aim of the research is to provide further insight into the influence of the tectonic movement on the geomorphic-sedimentary development in the graben basin.

Geographical and geological setting

Geographically, the FRGB is located in the eastern China Loess Plateau, where the climate is dominated by the East Asian Monsoon. The climate is temperate and sub-humid, with a mean annual precipitation of ca. 450‒500 mm (Editorial Board of China’s Physical Geography, Chinese Academy of Sciences, 1985). Most of the precipitation occurs during the warm and moist summer monsoon, whereas the cold and dry winter monsoon causes intense dust storms and loess deposition (An et al., 1991a; Zhang et al., 1994). There are two sub-basins within the graben basin; the Linfen Basin (LB) and the Taiyuan Basin (TB); which are surrounded by both mountainous and highland areas (Fig. 1). The Fen River flows from north to south and cuts across a highland (the Linshi Highlands) linking the two sub-basins hydrologically. Many tributaries of the Fen River also cut into the sub-basins, thus forming and extending the river valley system. The main types of landforms observed in the sub-basins are lake terraces, alluvial terraces, floodplains, and valleys, all of which are typically capped by loess of different sequence types and thicknesses following dust deposition throughout the Quaternary.

The Ordos Massif has been a tectonically stable block in North China since the Paleogene Period, surrounded by a series of active graben basins (Fig. 1). The FRGB is situated to its east. The FRGB, geologically characteristic of other graben basins, is bounded by a series of neotectonically active normal-faults within the surrounding mountains and highlands. It was tectonically affected by the movement of the Tibetan Plateau during the late Cenozic (Zhang et al., 1979; Hu et al., 2012; Gao et al., 2015). Many alluvial terraces extending across these boundary faults are offset laterally and vertically (Yang, 1987). The vertical displacement during the Quaternary of Neogene gravels along some primary boundary faults reached ca. 2500 m, with a lateral displacement of ca. 12.5 km (Hu et al., 2010). In addition, the basin experienced some major earthquakes, e.g., the Hongdong Earthquake in 1303 and the Linfen Earthquake in 1695, each with estimated magnitudes of>8.0 (Research Group of State Seismological Bureau, 1988).

Methods

Loess is an aeolian sediment deposited during the Quaternary. The loess sections in northern and northwestern China are composed of less-weathered, massive, brown loess units and red, maturely-weathered, soil units, forming a loess/paleosol sequence. In the sequence, each of the loess and paleosol units has been assigned a specific, stratigraphic designation (e.g., loess units: L₁, L₂…; paleosol units: S₁, S₂…), and has been dated (Liu, 1985; Kukla and An, 1989; Shen et al., 1994). Each loess and paleosol unit can therefore be used as a distinctive, regional stratigraphic horizon of known age to constrain the ages of some sediments and landforms (Porter et al., 1992).

Under conditions of continuous loess deposition and lake regression and/or transgression, once a lake has receded, some of the lake floor would most likely be exposed subaerially to form a lake terrace, and loess/paleosol units would begin to deposit and develop on its surface. The main characteristic of this kind of terrace is that its base is composed of lacustrine sediment. Therefore, by determining the stratigraphic unit of the lowest loess or the paleosol above it, the time of lake regression occurrence can be inferred. Conversely, if a transgression occurred and lake water flooded part of the land beside the lake, where loess/paleosol units had been deposited or developed, lacustrine sediments would be superimposed on them. Therefore, the topmost loess or paleosol unit underlying the lacustrine sediment can be taken to indicate the time when the lake transgression occurred.

In addition, lake regression or transgression can also be determined as having been rapid or slow, based on changes in the horizontal distribution of the lowest loess or paleosol units overlying the lake terrace from its margin to the center, or of the topmost loess or paleosol units underlying the lacustrine sediment’s margin to the innermost sections. When the lowest loess or paleosol units overlying the lake terrace are the same at the margin as in the center of the lake floor, it can be inferred that lake regression was rapid (Fig. 2(a)). Conversely, when the loess or paleosol units differ between the margin and the center, (i.e., they are younger toward the center of the basin), this indicates a slow lake regression rate (Fig. 2(b)). In contrast, when the topmost loess or paleosol units underlying the lacustrine sediment at both the center and the near-margin of the basin are the same, a rapid transgression is indicated (Fig. 2(c)). Equally, should the topmost loess or paleosol units become younger with proximity to the basin margin, very slow transgression can be inferred (Fig. 2(d)). We used the above models in this study to reveal the processes of paleolake regressions and transgressions.

Some prominent characteristic differences exist between loess/paleosol stratigraphies and lacustrine sediment. Besides color difference, loess/paleosol stratigraphies exhibit no bedding, but lacustrine sediment does; a paleosol layer has pedogenic horizons (e.g., argillic and calcic horizons) (Guo and Fedoroff, 1990), but lacustrine sediment does not. It is therefore easy to distinguish with the naked eye the two types of deposition, even in the field.

The identification of the loess/paleosol units relies on their distinctive physical characteristics (e.g., texture, structure, thickness, color, paleomagnetism, and magnetic susceptibility) as well as their stratigraphic positions within the regional succession. For instance, the Brunhes-Matuyama polarity boundary (B/M) is located in loess layer L₈; those units younger or older display Brunhes normal polarity and Matuyama reversal polarity, respectively (Yue and Xue, 1996). The determination based on paleomagnetism of the B/M boundary in any loess section is an effective way to determine L₈ and other loess/paleosol units. The S₅ layer in North China is of distinctive paleosol with a greater thickness, darker brown color, and thicker clay coating than other paleosol units (Han et al., 1998), and is actually a trial of soils that constitutes an easily identifiable stratigraphic marker. There are normally four other prominent paleosols above S₅. Identifying S₅ in a loess section also helps to distinguish other loess/paleosol units. The S₁ layer is the first prominent reddish-brown palesol below the ground surface.

Prior to our field survey, we analyzed the topographic maps (scale1:50,000) and satellite imagery of the FRGB to determine which valleys we should focus on. Accordingly, we investigated the deeply-incised valleys that extend from the edge of the basin to its center, in order to trace the lacustrine sediment and also the lowest loess/paleosol unit overlying, or the topmost unit underlying, the lacustrine sediment. All paleomagnetism samples extracted were tested at the Institute of Geomechanics (Ministry of Geology and Mineral Resources) of the Chinese Academy of Sciences.

Geomorphic-sedimentary features of the graben basin and the cyclical process of the regressions and transgressions revealed

Geomorphic-sedimentary features in the LB

The LB is a reversed “L” in shape, extending NNE to SSW in its northern sector, with a length of ca. 80 km and a width of up to 40 km, and ENE to WSW in its southern sector, with a length of ca. 60 km and a width of ca. 30 km (Fig. 1).

In its northern sector, there is an extensive lacustrine platform of ca. 20 km in width to the east of the Fen River. The Ju valley, a tributary of the Fen River, incised into the platform and exposed its geomorphic-sedimentary features. Field surveys revealed that the platform is composed of three lake terraces, each with different heights and different loess/paleosol sequences overlying them.

The highest terrace is ca. 4 km wide from east to west, ca. 110 m above the present Fen River. Its eastern boundary is a normal fault which separates it from the Fushan Highlands and defines the eastern margin of the basin (Fig. 3). The base of the terrace is composed of lacustrine sediment, which is laminated grey-green silt and fine sand. The thickness of the exposed lacustrine deposit is ca. 30 m, but its lower content remains unexposed. The loess/paleosol stratigraphy overlying the terrace is up to 30 m thick, and includes five well-developed paleosols. From the far to the near side of the terrace, the lowest loess/paleosol unit, developed directly on the top of the lacustrine sediment, is entirely of the fifth paleosol. This paleosol displays a darker, reddish-brown color and thicker clay coatings than the four others; it can be inferred as S₅ (Fig. 3(a)). At the front of this terrace, S₅ extends and dips toward the lower terrace, and is partially capped by the lower lacustrine sediment, before finally dying out.

The middle terrace is ca. 8 km wide and 60 m above the current level of the Fen River, with laminated silt and fine sand with interlayered clay building up its base. A ca. 15 m-thick loess/paleosol stratigraphy has been deposited on the surface of the terrace, where two prominent paleosols have developed. The lowest loess/paleosol unit above the lacustrine sediment on both the far and near sides of the terrace is the same, and is inferred to be S₂, based on its stratigraphic position in the loess/paleosol succession as well as its magnetic susceptibility feature (Fig. 3(b)).

The lowest terrace is ca. 3 km wide from east to west and lies 45 m above the Fen River’s present level. The terrace base is also composed of laminated clay, silt, and fine sand. There is a ca. 10 m-thick loess/paleosol stratigraphy overlying the lacustrine sediment. The lowest unit from the far to the near side of the terrace is composed of the same paleosol. According to its magnetic susceptibility feature, this paleosol can be inferred as S₁.

In addition to the three lake terraces, there are two well-developed accumulational terraces within the basin, both of which are lower in height than the lowest lake terrace.

In the field we also found other geomorphic-sedimentary phenomena. It is a gentle slope in the transition zone of ca. 2.5 km long between the highest and middle lake terraces. By horizontally tracing the loess/paleosol stratigraphies, which extend from the highest terrace along the valley in this zone, we found that they gradually dip toward the middle terrace, some of which are partially overlain by the younger lacustrine sediment. The exposed sections in the downvalley of this zone show the existence of the S₅ layer, which is overlain by a younger lacustrine sediment. Alternatively, toward the upvalley, it can be identified that L₅, S₄, L₄, and S₃ layers are directly overlain by a younger lacustrine sediment, whereas the L₃, approximately 2 m-thick, is directly covered by younger lacustrine sediment (Fig. 3(c)). The longitudinal section along the valley in this zone shows that the shorter the distance from the highest terrace, the younger the topmost loess/paleosol stratigraphy underlying the lacustrine sediment.

The Emei Highlands is to the south of the FRGB (Fig. 1). Four descending lake terraces are discernible from the highland margins to the center of the basin. The Licun Valley is a long, deeply-incised valley which cuts across these terraces and thus clearly shows their geomorphic-sedimentary differentiation (Fig. 4).

The highest lake terrace is 210‒220 m high above the Fen River and the loess/paleosols overlying the lacustrine sediment are ca. 48 m thick. Of these, the lowest loess/paleosol unit is a reddish paleosol. The second highest terrace is 160‒170 m above the current Fen River, and is overlain by loess/paleosols ca. 50 m thick; the lowest loess/paleosol unit is also a paleosol. Sixty-one paleomagnetic samples were taken from the overlying loess/paleosol stratigraphies in the two terraces, most of which were selected at 80 cm intervals, with a lesser amount taken at 40 cm intervals near the inferred reversal of the paleomagnetic field. The paleomagnetic results show that the B/M boundary is located in a loess layer on the lowest paleosol, indicating that the lowest paleosol unit overlying the two terraces is S₈ (Yue and Xue, 1996) and that both terraces were exposed subaerially almost concurrently. The present-day surfaces of the two terraces are southward-tilting reverse slopes, with normal faults both between them and in front of the second highest terrace (Fig. 4(a)). The two faults extend from east to west marking the boundary between the highlands and the FRGB. The difference in elevation between the two terraces may have resulted from the tectonic movement of the normal fault between them. The S₈ paleosol, located at the front of the second highest terrace, appears to dip toward and extends into the middle terrace, composed of lacustrine sediment, before petering out.

The middle terrace is ca. 110 m above the present Fen River, and is overlain by loess-paleosol stratigraphies ca. 30 m thick. The lowest aeolian sediment in direct contact with the lacustrine sediment is a mauve paleosol layer, with a well-developed manganese membrane. Owing to its significant clayification, mauve appearance and stratigraphic position within the loess/paleosol succession, it can be identified as S₅. At the front of this terrace, the paleosol dips toward the lowest terrace and extends into the lacustrine sediment which is composed of the lowest terrace. A further tracing survey taken downvalley found that S₅ extends nearly to the center of the basin, occurs in the sections near the present Fen River, and is overlain by a 26 m thick layer of younger lacustrine sediment.

The lowest terrace is ca. 50‒40 m above the current Fen River, with marked differences in the sedimentary features of its far side and frontage. An exposed vertical section at the far side of the terrace shows that the underlying sediments (from bottom to top) are old lacustrine, S₅, young lacustrine, S₂, L₂, S₁, and L₁ in turn; the sedimentary stratigraphy at the terrace frontage (from bottom to top) runs through old lacustrine, S₅, young lacustrine, S₂, L₂, younger lacustrine, S₁, and L₁ in turn (Fig. 4(b)), with a layer of celadon lacustrine sediment covering L₂ and underlying S₁.This terrace can therefore be divided into two sub-terraces.

Two accumulational terraces also developed in this sector after the formation of the lowest lake terrace.

In the transition zone between these lake terraces, we found geomorphic-sedimentary phenomena similar to those identified in the northern sector. In the transition zone of ca. 3 km in length between the second highest and the middle lake terrace, a tracing survey found that all the loess/paleosols dip toward the middle terrace, and enter into and underlie its lacustrine sediment. From the downvalley to the upvalley, the topmost loess/paleosol unit underlying the lacustrine sediment can be identified as S₈, L₈, S₇, L₇, S₆, and L₆ in turn, and the lacustrine sediment between the overlying paleosol S₅ and these loess/paleosols were also observed to thin upvalley (Fig. 4(c)). The exposed sections in the transition zone between the middle and the lowest terrace, ca. 5 km in length, showed similar changes in the loess/paleosol stratigraphies that underlay lacustrine sediment, but they are S₅, L₅, S₄, L₄, S₃, and L₃ in turn from the downvalley to upvalley (Fig. 4(d)).

Geomorphic-sedimentary features in the TB

The TB extends from NE to SW, with a length of ca. 105 km and a width of up to 40 km (Fig. 1). A lacustrine platform, named the Zhangbi Platform, is located in its southeastern part-, with a deeply-incised valley exposing its geomorphic-sedimentary features. This platform is composed of three lake terraces, which decrease in height from upvalley to downvalley. Loess/paleosol stratigraphies of different sequences and thicknesses have been deposited on the surfaces of these terraces (Fig. 5).

The highest terrace is ca. 120 m above the present Fen River; its lacustrine sediment is bedded celadon silt and fine sand, with some fragments of snail shells. A nearly 50 m-thick aeolian sediment deposited on this terrace, in which eight paleosols can be identified in both its far and near sides. The eighth paleosol is the oldest aeolian unit and was developed directly on the surface of the lacustrine sediment. At the far side of the terrace, another older paleosol exists beneath the lacustrine sediment. The lacustrine sediment between the two paleosols has a wedge-like shape that thins upvalley, indicating the terrace is near the shore of the paleolake. Paleomagnetic results show that the eighth paleosol overlying the lacustrine sediment is S₈ and that underlying the sediment is S₁₁ (Fig. 5(a)). The middle terrace stands about 40 m lower than the highest terrace. The lacustrine sediment of the terrace base is composed of compacted grey clay and silt, locally interlayered with fine sand. The loess/paleosol sediment overlying this terrace is ca. 30 m thick, and S₅ is extensively developed directly on its surface (Fig. 5(b)).

In the transition zone between the two terraces, we found the same sedimentary phenomenon as displayed in the LB, i.e., the loess/paleosol stratigraphies S₈, L₈, S₇, L₇, S₆, and L₆ are partially capped by lacustrine sediment. This transition zone is a gentle slope and about 2 km long, and the lacustrine sediment between S₅ and older loess/paleosol increasingly becomes thinner toward upvalley. Due to the contact of loess/paleosol stratigraphies with the overlying lacustrine sediment, the typical features of paleosol S₈, S₇, and S₆ are not as apparent and their thicknesses are thinner than normal; however, the pedogenic horizons, such as the argillic and the calcic horizons, are still easily identified in the field with the naked eye. Some small calcareous gravels are also deposited in the lacustrine sediment from the loess/paleosols.

The lowest terrace is about 60 m lower than the middle terrace and is overlain by a loess/paleosol stratigraphy ca. 15 m thick. The lacustrine sediment of this terrace is made up of silt and light-colored fine sand, with numerous snail shell fragments. The lowest paleosol in the overlying loess/paleosol stratigraphy is S₁ and is widely developed on the surface (Fig. 5(c)). We did not find a lake terrace on this platform that corresponds to that in the LB proximally capped by S₂.

In the transition zone between the lowest and the middle terrace, the loess/paleosol stratigraphies S₅, L₅, S₄, L₄, S₃, L₃, S₂, and L₂ all dip toward the center of the basin and are partially covered by younger lacustrine sediment. From the downvalley to the upvalley, it can be easily identified that the topmost paleosol units underlying the lacustrine sediment are S₅, S₄, S₃, and S₂ in turn. The younger lacustrine sediment is loose, consisting of three light-red clay interbeddings. The S₄, S₃, and S₂ distribute toward the lacustrine sediment and seem to be continuous with the clay interbeddings, indicating that they developed or deposited contemporaneously, respectively.

There are, in addition, two younger and lower alluvial terraces along the Fen River.

The cyclical process of paleolake regressions and transgressions as recorded by geomorphic-sedimentary features

The two transverse geomorphic-sedimentary sections in the LB show that several paleolake regressions and transgressions occurred during the mid-late Quaternary. The earliest paleolake regression took place when S₈ began to develop, and the same paleosol unit covering the terrace surface indicates that this regression was rapid. It is determined that the rapid drop of the paleolake level in this regression was about 40‒50 m based on the height difference of the two adjacent terraces. The duration of low lake level after this regression corresponds approximately to that of the formation of S₈. After the development of S₈, a paleolake transgression occurred. This subsequent transgression was very slow and lasted from the early stage of the deposition of L₈ to the later stage of the deposition of L₆, and was composed of the first paleolake regression-transgression cycle in the FRGB. The second paleolake regression occurred when S₅ began to develop, which was somewhat rapid. The subsequent low lake level lasted until the late stage of the development of S₅ when the second paleolake transgression took place. This regression caused the paleolake level to decline more than 60 m and may have left the southern sector of the FRGB empty because S₅ extends nearly to the center of this sector. The second paleolake transgression was also very slow and lasted from the early stage of the deposition of L₅ to the late stage of the deposition of L₃. This is the second paleolake regression-transgression cycle noted in the basin. The third paleolake regression occurred when S₂ began to form. Sedimentary evidence shows a rapid 10 m drop in the paleolake level. During the deposition of L₂, the third paleolake transgression dominated the basin, representing FRGB’s third paleolake regression-transgression cycle. The fourth paleolake regression took place with the beginning of the formation of S₁. The extensively-developed S₁ on the surface of the lacustrine sediment shows that this regression was rapid. Indeed, this regression was so extensive that nearly all the lake water drained resulting in a virtual disappearance of the paleolake, with the Fen River emerging in the lowest part of the basin. During these paleolake regression-transgression cycles, each regression was rapid, and each transgression was slow.

The paleolake regressions and transgressions observed in the LB have also been recorded in the TB, with the exception of the third paleolake regression-transgression cycle.

Figure 6 shows the cyclical process of the paleolake regressions and transgressions in the FRGB during the mid-late Quaternary. Each significant, rapid paleolake regression was followed by a slow paleolake transgression (except the last regression). The later transgression would have raised the lake’s level, but not as high as that of the former lake.

Discussion on the tectonic implication of the cyclical process of the paleolake regressions and transgressions in the graben basin

A lake is a significant water body on the Earth’s surface. Lake basins are differentiated by the type of water loss: inflow or outflow. An outflow lake basin not only has surface water or groundwater inflow, but it also an opening (outlet) to allow for a continuous outflow. Many geomorphic-sedimentary facts show that the LB and TB were outflow basins during the mid to late Quaternary. The drilled lacustrine samples and exposed lacustrine sections in both basins did not contain salty crystalline, indicating the paleolake water had been fresh (Wang et al., 1996). Several alluvial terraces were formed during that time in the Lingshi Highlands and the Caizhuang sectors, indicating a consistent flow of the paleolake water (Hu et al., 2005).

To be an outflow basin, it is clear that the paleolake regression-transgression cycles in the FRGB were not the result of paleoclimate changes. There are three reasons to account for this deduction. First of all, the occurrence of several lake regressions is almost synchronous with the formation of some paleosols, when the paleoclimate was warm and wet and was accompanied by high annual precipitation (An et al., 1991b; Porter and An, 1995) and an increased runoff into the basins. Secondly, the occurrence of each paleolake regression-transgression cycle does not coincide with the paleoclimate changes which are recorded in the Quaternary loess-paleosol sequences. Thirdly, paleoclimate changes between dry and moist stages can affect the volume of the water inflow; an increased inflow will result in an increased outflow, with the alternative also observed—a decreased inflow will result in a decreased outflow. Research has indicated that paleoclimate changes caused lake levels to fluctuate by only 2‒3 m in the FRGB during the mid to late Quaternary (Hu et al., 2005). Therefore, tectonic movement may be the primary factor affecting paleolake regression-transgression cycles in the graben basin.

Neotectonic features in the graben basins around the Ordos Massif

The main neotectonic features in the FRGB and other graben basins around the Ordos Massif include the uplift of the upper mantle, the thinning of the crust, and the extension of the basin (Wang, 1979; Chang et al., 1981; Deng et al., 1982). The depth of the Curie surface and the thickness of the crust in the graben basins are 20‒25 km and 36‒40 km, respectively; while those in the Ordos Massif and surrounding mountains are 30‒35 km and 43‒46 km, respectively (Research Group of State Seismological Bureau, 1988; Bai and Hou, 1994; Gao et al., 2015). The extension rate in a NW‒SE direction in the graben basins was 0.5‒3.1 mm·yr⁻¹ during the Pliocene‒Quaternary (Zhang et al., 1998). The uplift of the upper mantle in the graben basins was also evidenced by the distribution of heat flow and the volcanic activity, e.g., the heat flow in the graben basins was ca. 75.3‒79.5 mW/m², while that in the Ordos Massif was only 46.0‒58.6 mW/m² (Hu et al., 2000; Wang, 2001). The Datong Graben Basin experienced several episodic intensive volcanic eruptions during the mid-late Quaternary (Li et al., 1998; Zhao et al., 2012).

Some additional evidences show that the FRGB and other graben basins slowly subsided during the late Quaternary when L₁ deposited. 1) The top surface of the middle Quaternary lacustrine sediment in the center of the LB and TB is funnel-shaped (Wang et al., 1996). A great number of aggradational alluvial fans developed in the piedmonts beside the Weihe Graben Basin (Han et al., 2001). 2) The archeological sites dating back 2500 years in the LB were naturally buried ca. 1.6 m deep below layers of mud and sand (Liu and Meng, 1975). 3) Pingyao, an ancient city in the TB, was built during the Ming Dynasty ca. 600 years ago. The lower section (ca. 2 m) of the city wall is now buried underground, along with an old stone bridge which is also buried ca. 2.2 m deep. 4) Monitoring the changes in the height of the ground using the small baseline subset InSAR technique shows the Dadong Basin is now slowly subsiding due to tectonic deformation (Yang et al., 2014).

Tectonic implication of the cyclical process of the paleolake regression-transgression cycles in the FRGB

The lake level of an outflow lake basin is controlled by the height of the lake’s outlet, which is normally located in the surrounding mountains or highlands. Lake regressions in the FRGB, which are recorded by the loess/paleosol-overlying-lacustrine sediment, can only take place under the following two circumstances: 1) the lake basin floor rises relative to the lake outlet; or 2) the lake outlet lowers relative to the lake basin floor. As a result, the lake bottom was subaerially exposed allowing for loess/paleosol deposition. Conversely, a transgression, recorded by the lacustrine sediment-overlying-loess/paleosol, will occur if the lake basin floor lowers relative to the lake outlet or if the lake outlet rises relative to the lake basin floor, resulting in flooding and lacustrine deposition. The accumulation of the sediment in the basin will reduce the water storage space, but cannot raise the lake level.

In the FRGB, the occurrence of a regression may mean the graben basin floor uplifted relative to the paleolake outlet in the surrounding mountains or highlands, whereas for a transgression, the floor subsided relative to the paleolake outlet. Tectonically, there are two competing factors for determining the development of the continental graben basins (uplift or subsidence, and the magnitude): 1) the upwelling of the hot mantle material and 2) the thinning of the low-density crust. The subsequent cooling of the uplifted warm mantle also plays some role in the subsidence of the basins (Jarvis, 1984). Depending on which effect is dominant, either uplift or subsidence may occur. Therefore, the regression-transgression cyclical process of the paleolake in the FRGB implies: 1) in the first stage of an episodic tectonic movement, the upwelling of the upper mantle underneath the graben basin may be dominant and resulting in a rapid uplift of the basin floor; 2) in the following stage, the thinning of the crust and cooling of the warm mantle material underneath the graben basin may become dominant and cause the basin floor to subside slowly. The four rapid paleolake regressions during the mid to late Quaternary indicate four occurrences of intensive upwelling of the upper mantle underneath the graben basin. Three slow paleolake transgressions and the basin floor subsidence during the deposition of L₁ (after the last regression) indicate subsequent thinning of the crust and cooling of the warm mantle material after each intensive upwelling of the upper mantle.

In other graben basins around the Ordos Massif, the stepped lacustrine or alluvial terraces and episodic intensive volcanic activity also record these episodic intensive tectonic movements (Xia, 1992; Han et al., 2001; Sun and Xu, 2007; Hu et al., 2012; Ren et al., 2014). These incidents are attributed to the northward or northeastward push of the Tibetan Plateau (Zhang et al., 1979; Li, et al., 1998; Gao et al., 2015).

Because of the intensive stream erosion at lake outlet sites in the surrounding mountains or highlands during each rapid lake regression, the elevation of each outlet became progressively lower. Consequently, the water levels of any younger lakes would have to be lower than those of older lakes, and the elevations of any younger lake terraces would be lower than those of older ones.

Conclusions

1) The lowest loess/paleosol units overlying the lake terraces and the topmost units underlying the lake terraces from the margin to the center of the LB and TB revealed that several paleolake regression-transgression cycles took place during the mid to late Quaternary in the FRGB. When paleosols S₈, S₅, S₂, and S₁ began to develop, the basins experienced four rapid lake regressions. A slow lake transgression followed each lake regression. Three slow lake transgressions took place at approximately the same time as the deposition or development of loess/paleosols L₈ to L₆, L₅ to L₃ and L₂. After the last significant regression, the FRGB was completely drained and the paleolake disappeared.

2) The cyclical process implies that a cause and effect relationship may exist between the paleolake regressions and transgressions in the basin and the tectonic activity in the upper mantle underneath the graben basin. When the upwelling of the upper mantle underneath the graben basin was dominant, the basin floor uplifted rapidly and a rapid lake regression took place. Alternatively, when the thinning of the crust and cooling of the warm mantle material underneath the graben basin became dominant, the basin floor subsided slowly and a slow lake transgression occurred.

3) The four rapid paleolake regressions and subsequent transgressions or subsidence of the basin floor in the FRGB indicate the possible occurrences of four time-intensive upwellings of the upper mantle and the subsequent thinning of the crust and cooling of the warm mantle material during the mid to late Quaternary.

According to the research results, we deduce that the intensive movements of normal faults in the graben basin should be synchronous with the rapid lake regressions when the intensive upwelling of the upper mantle underneath the graben basin was dominant, i.e., episodic intensive movements of the faults during the mid to late Quaternary are likely to correspond to the paleosols S₈, S₅, S₂ and S₁. Thus, future research on the movement history of faults in the graben basin is needed in order to further confirm this tectonic implication of the paleolake regression-transgression cycles.

References

[1] An Z S, Kukla  G J, Porter  S C, Xiao  J L (1991b). Magnetic susceptibility evidence of monsoon variation on the Loess Plateau of central China during the last 130,000 years. Quat Res, 36(1): 29–36

[2] An Z S, Kukla  G J, Proter  S C,Xiao   J L (1991a). Late Quaternary dust flow on the Chinese loess plateau. Catena, 18(2): 125–132

[3] Bai L W, Hou  T G (1994). Deep geophysical exploration on Shanxi Graben System. North China Earthquake Sciences, 12(3): 27–35 (in Chinese)

[4] Bergner A G N ,  Strecker M R ,  Trauth M H ,  Deino A ,  Gasse F ,  Blisniuk P ,  Duhnforth M  (2009). Tectonic and climatic control on evolution of rift lakes in the Central Kenya Rift, East Africa. Quat Sci Rev, 28(25‒26): 2804–2816

[5] Briggs R W, Wesnousky  S G, Adams  K D (2005). Late Pleistocene and late Holocene lake highstands in the Pyramid Lake Subbasin of Lake Lahontan, Navada, USA. Quat Res, 64(2): 257–263

[6] Chang W Y, Li  Y H, Ma  F C, Zong  J Y (1981). On the mechanical mechanism of the formation of graben. Scientia Geological Sinica, 1: 1–11 (in Chinese)

[7] Cukur D, Krastel  S, Schmincke H U ,  Sumita M ,  Tomonaga Y ,  Namik Caǧatay M  (2014). Water level changes in Lake Van, Turkey, during the past ca. 600 ka: climatic, volcanic and tectonic controls. J Paleolimnol, 52(3): 201–214

[8] Delvaux D, Kervyn  F, Vittori E ,  Kajara R S A ,  Kilembe E  (1998). Late Quaternary tectonic activity and lake level change in the Rukwa Rift Basin. J Afr Earth Sci, 26(3): 397–421

[9] Deng Q D, Zhang  Y M, Xu  G L, Fang  F T (1982). The characteristics of neotectonic stress field in China and its relationship with the intra-plate movement. Seismic Geology, 1(1): 11–22 (in Chinese)

[10] Dong Y, Xie  A H, Wei  H (2001). The formation of Dingcun Lake and its environment change. Journal of Shanxi Teacher’s University (Natural Science Edition), 15(4): 70–74

[11] Editorial Board of China’s Physical Geography, Chinese Academy of Sciences (1985). Physical Geography of China: General. Beijing: Science Press, 3 ‒ 20, 74–102 (in Chinese)

[12] Eriş  K K  (2013). Late Pleistocene–Holocene sedimentary records of climate and lake-level changes in Lake Hazar, eastern Anatolia, Turkey. Quaternary International, 302(17): 123–134

[13] Gao G M, Kang  G F, Li  G Q, Bai  C H (2015). Crustal magnetic anomaly in the Ordos region and its tectonic implications. J Asian Earth Sci, 109: 63–73

[14] Guo Z T, Fedoroff  N (1990). Genesis of calcium carbonate in loess and in paleosols in central China. Developments in Soil Science, 19: 355–359

[15] Han H Y, Mi  F S, Liu  H Y (2001). Geomorphological structure in the Weihe Basin and Neotectonic movement. Journal of Seismological Research, 24(3): 251–257 (in Chinese)

[16] Han J T, Fyfe  W S, Longstaffe  F J (1998). Climatic implications of the S5 Paleosol complex on the southernmost Chinese Loess Plateau. Quat Res, 50(1): 21–33

[17] Hu S B, He  L J, Wang  J Y (2000). Heat flow in the continental area of China: a new data set. Earth Planet Sci Lett, 179(2): 407–419

[18] Hu X M, Chen  M J, Wang  D T, Wu  J L, Hu  H C (2012). The sequence difference in the times in the geomorphic-sedimentary evolution in the Fenwei Graben Basins during the middle-late Quaternary and its tectonic significance. Quaternary Sciences, 32(5): 1–10 (in Chinese)

[19] Hu X M, Li  Y L, Yang  J C (2005). Quaternary paleolake development in the Fen River basin, North China. Geomorphology, 65(1‒2): 1–13

[20] Hu X M, Wang  L L, Zhe  J, Lu H L  (2010). Morpho-sedimentary evidence of the Huoshan Fault’s late Cenozoic right-lateral movement in the Linfen Graben. Shanxi Graben System, North China. Frontiers of Earth Science, 4(3): 311–319

[21] Imbrie J, Hays  J D, Pisias  N G, Prell  W L, Shackleton  N J (1984). The orbital theory of Pleistocene climate: support from a revised chronology of the marine δ¹⁸O record. In: Berger J I A, Hays J, Kukla G, Saltzman B, eds. Milankovitch and Climate, Part 1. Dordrecht: Reidel, 269–305

[22] Jarvis G T (1984). An extensional model of graben subsidence—The first stage of basin evolution. Sediment Geol, 40(1‒3): 13–31

[23] Kukla G, An  Z S (1989). Loess stratigraphy in central China. Palaeogeogr Palaeoclimatol Palaeoecol, 72: 203–225

[24] Li Y L, Shi  X M, Fu  J L (2004). Geomorphic transformational event around 1.2 Ma BP in the southern Shanxi Province. Scientia Geographica Sinica, 24(3): 292–297 (in Chinese)

[25] Li Y L, Yang  J C, Xia  Z K, Mo  D W (1998). Tectonic geomorphology in the Shanxi Graben System, northern China. Geomorphology, 23(1): 77–89

[26] Liu C J, Meng  F H (1975). Using archaeological method to investigate the relation between the recent structural movements and earthquankes, taking Linfen Basin, Shanxi Province, as an example. Acta Geophysical Sinica, 18(2): 127–136 (in Chinese)

[27] Liu T S (1985). Loess and the Environment. Beijing: China Ocean Press, 322–358 (in Chinese)

[28] Massaferro J, Recasens  C, Larocque-Tobler I ,  Zolitschka B ,  Maidana N I  (2013). Major lake level fluctuations and climate changes for the past 16,000 years as reflected by diatoms and chironomids preserved in the sediment of Laguna Potrok Aike, southern Patagonia. Quat Sci Rev, 71(1): 167–174

[29] Mo D W (1991). The study on the paleoenvironmental evolution in the Linfen Basin during the Cenozoic. Journal of Peking University (Natural Science), 27(6): 25–29

[30] Porter S C, An  Z S (1995). Correlation between climate events in the North Atlantic and China during the last glaciation. Nature, 375(6529): 305–308

[31] Porter S C, An  Z S, Zheng  H B (1992). Cyclic Quaternary alluviation and terracing in a nonglaciated drainage basin on the north flank of the Qinling Shan, Central China. Quat Res, 38(2): 157–169

[32] Ren J J, Zhang  S M, Meigs  A J, Yeats  R S, Ding  R, Shen X M  (2014). Tectonic controls for transverse drainage and timing of the Xin-Ding paleolake breach in the upper reach of the Hutuo River, north China. Geomorphology, 206: 452–467

[33] Research Group of State Seismological Bureau (1988). Active fault system around Ordos Massif. Beijing: Seismic Press, 77 ‒ 113, 238–246 (in Chinese)

[34] Ross  K A ,  Smets  B ,  Batist  M D ,  Hilbe  M ,  Schmid  M ,  Anselmetti  F S  (2014). Lake-level rise in the late Pleistocene and active subaquatic volcanism since the Holocene in Lake Kivu, East African Rift. Geomorpholgy, 221: 274–285

[35] Shanahan T M, Overpeck  J T, Wheeler  C W, Beck  J W, Pigati  J S, Talbot  M R, Scholz  C A, Peck  J, King J W  (2006). Paleoclimatic variations in West Africa from a record of late Pleistocene and Holocene lake level stands of Lake Bosumtwi, Ghana. Palaeogeogr Palaeoclimatol Palaeoecol, 242(3‒4): 287–302

[36] Shen C D, Yi  W X, Liu  T S (1994). ¹⁰Be records in loess with high resolution and the dating of loess strata. Quaternary Sciences, 14(3): 203–213 (in Chinese)

[37] Sun J M, Xu  L L (2007). River terraces in the Fenwei Graben, Central China, and the relation with the tectonic history of the India Asia collision belt during the Quaternary. Quaternary Sciences, 27(1): 20–26 (in Chinese)

[38] Urabe A, Tateishi  M, Inouchi Y ,  Matsuoka H ,  Inoue T ,  Dmytriev A ,  Khlystov O M  (2004). Lake-level changes during the past 100,000 years at Lake Baikal, southern Siberia. Quat Res, 62(2): 214–222 

[39] Wang N L, Yang  J C, Xia  Z K (1996). The Sediment and Tectonic Landform during Cenozoic in the Fen River Drainage Basin. Beijing: Science Press, 120 ‒ 205, 334–346 (in Chinese)

[40] Wang Y (2001). Heat flow pattern and lateral variations of lithosphere strength in China mainland: constraints on active deformation. Phys Earth Planet Inter, 126(3‒4): 121–146

[41] Wang Y P (1979). The earthquakes in the intra-plate in China and the features of the stress field in the Mesozoic and Cenozoic. Seismology and Geology, 1(3): 1–11 (in Chinese)

[42] Xia Z K (1992). The study on the change of ancient lake shore of Datong-Yangyuan Basin. Geogr Res, 11(2): 52–59 (in Chinese)

[43] Xu X W, Ma  X Y (1992). Geodynamics of the Shanxi Rift system, China. Tectonophysics, 208(1‒3): 325–340

[44] Yang C S, Zhang  Q, Zhao C Y ,  Wang Q L ,  Ji L Y  (2014). Monitoring land subsidence and fault deformation using the small baseline subset InSAR technique: a case study in the Datong Basin, China. J Geodyn, 75: 34–40

[45] Yang J C (1987). The alluvial terraces and neotectonics in the south sector of the Fen River. In: Research Group of State Seismological Bureau, ed. The Tectonic Stress and Crust Structure. Beijing: Geology Press, 132–136 (in Chinese)

[46] Yue L P, Xue  X X (1996). China Loess and Paleomagnetism. Beijing: Geology Press, 58–70 (in Chinese)

[47] Zhang X Y, An Z S, Chen T , Zhang G Y, Arimoto  R, Ray  B J  (1994). Late Quaternary records of the atmospheric input of eolian dust to the center of the Chinese loess plateau. Quat Res, 41(1): 35–43

[48] Zhang Y M, Wang  L M, Dong  R S (1979). Discussion on the changes of Cenozoic tectonic stress field in the eastern part of north China. Seismology and Geology, 1(1): 23–28 (in Chinese)

[49] Zhang Y Q, Mercier  J L, Vergely  P (1998). Extension in the graben systems around the Ordos (China), and its contribution to the extrusion tectonics of south China with respect to Gobi-Mongolia. Tectonophysics, 285(1‒2): 41–75

[50] Zhao H, Wang  C M, Mao  H L, Lu  Y C, Liu  L J, Ji  Y P, Zhao  H M (2012). OSL dating of volcanic baked sediment in Datong area, Shangxi Provience of China. Quaternary Sciences, 32(3): 510–515 (in Chinese)

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg