Abstract
Geomorphologically, alluvial deposits in river systems are expected to be older on higher terraces than on the lower terraces. However, loess deposits of aeolian origin may also occur on the surface of terrace systems and as seen in the Chinese Loess Plateau (CLP). Such loess is continuously, rather than episodically deposited. This study presents numerous dating results of loess cover on the Hanjiang River terrace system in the southern Qinling Mountains, an atypical loess deposit belt outside of the CLP. We name this “Hanjiang Loess”. Results indicate that the Hanjiang Loess deposited on the high, middle, and low terraces are also the oldest, old, and most recent, respectively, apparently analogous with alluvial expectations. We thus propose a modified depositional hypothesis, whereby terrace loess reflects the deposition of aeolian loess and other material, subsequent fluvial reworking. This depositional hypothesis should also be applicable to loess deposits on the river terraces in the Xiashu Loess, Quaternary Reticulate Red Clay, and Quaternary Red Clay in southern China, and other atypical loess areas outside of the CLP.
I. Introduction
Loess deposit in China is wide and thick, and the developed loess deposit belt (DLDB) lies between 34° and 45° N latitude (Figure 1). In this range, the Chinese Loess Plateau (CLP) contains a continuous record of terrestrial sedimentation since the Pliocene to the latest Holocene. The classic loess–paleosol sequence records the palaeo-environment changes and the climates dominated by the winter and summer monsoon (An et al., 2001; Ding et al., 1998; Liu, 1985).
Map showing the lines between the developed loess deposit belt (DLDB) and atypical loess deposit belt (ALDB), which is between approximately 30° and 34° N latitude in Southern China, the investigated Hanjiang loess in Hanzhong and Yunxian Basins, and other loess and atypical loess sections in Luochuan on the Chinese Loess Plateau, Xiashu loess at Yizheng, Quaternary reticulate red clay at Xuancheng, and Quaternary red clay at Jiujianng form the central area toward southeast China.
The loess–paleosol sequences span the entire Quaternary period. Moreover, underlying the Quaternary sequences is the Tertiary red clay. The wind-blown loess and Tertiary red clay in Northern China have a source in the Gobi desert and other deserts in Central Asia (Liu, 1985; Lu et al., 2010; Zhang et al., 2003). Numerous studies clarified the source, transportation, deposition, palaeo-environment, and palaeo-climate recording of loess deposit on the CLP (e.g., An et al., 2001; Ding et al., 1998; Guo et al., 2002; Liu, 1985; Nie et al., 2015).
Atypical loess (i.e., loess-like) deposit belts cover a large area that lies between approximately 30° and 34° N latitude in Southern China and between approximately 45° and 49° N latitude in Northern China (Liu, 1985) (see Figure 1). A few studies have also focused on atypical loess deposit from the CLP, such as loess-like deposits in the lower and middle reaches of the Yangtze River in Southern China, which began to form in the middle Pliocene (e.g., Hu et al., 2010; Qiao et al., 2003; Xiong et al., 2002). Other areas include the eastern Inner Mongolia in northeastern China, which began in the middle Pliocene (Zeng et al., 2011); Xinjiang Province in northwestern China, which began in the Holocene (Chen et al., 2016); and in the Tibetan Plateau in southwestern China in the Holocene (Zhang et al., 2015).
The Qinling Mountains lie to the south of the CLP and act as a barrier to the southeastward transport of aeolian dust. Nevertheless, aeolian silt being transported through the Qinling Mountains is deposited between approximately 30° and 34° N latitude in Southern China and forms the atypical loess deposit belt (ALDB). The grain size of the atypical loess is finer and the sedimentation rate of such loess is lower than that in the central CLP. The color of this loess is considerably reddish brown because the atypical loess sediment has been gradually affected by oxidation, weathering, and pedogenesis from Central to southeast China (Sun et al., 2017).
In the middle and lower reaches of the Yangtze River, several special names of the atypical loess deposit (e.g., “Xiashu loess” and “Quaternary red clay”) were given to distinguish them from the classic loess on the CLP. The Xiashu loess is mainly distributed along the Yangtze River approximately north of 31°N latitude and is of aeolian dust origin (Hu et al., 2010). The Quaternary red clay is in the southeastwards of the middle and lower reaches of the Yangtze River. One part of the Quaternary red clay is sedimentary with aeolian dust characteristics (Hu et al., 2009), while the other part is derived from the underlying parent rocks. Although only a few geologists investigated the atypical loess deposit in the middle and lower reaches of the Yangtze River (e.g., Hu et al., 2009, 2010; Qiao et al., 2003; Xiong et al., 2002; Zhang et al., 2007), the basic understanding of the source, transportation, deposition, palaeo-environment, and palaeo-climate recording of the atypical loess deposit southward of 34° N latitude (Qinling Mountains) of China remain unclear.
Atypical loess deposits at the Hanzhong and Yunxian Basins along the Hanjiang River in the southern Qinling Mountains are between the classic loess on the CLP and Xiashu loess in the middle and lower reaches of the Yangtze River. The deposit is called “Hanjiang loess” and is of aeolian dust origin as well (Sun et al., 2012, 2017). Several features (e.g., color, thickness, grain size, and sources) gradually changed and are comparable with the classic loess at Luochuan on the CLP, Hanjiang loess in the current study, Xiashu loess, Quaternary reticulate red clay, and Quaternary red clay from Central to southeast China (Figure 2). We used the investigations in the Hanjiang loess as a basis to attempt to understand the depositional process of the ALDB between approximately 30° and 34° N latitude of Southern China.
Photos of the loess and atypical loess sections that form the Luochuan loess (a), Hanjiang loess at Hanzhong in the current study (b), Xiashu loess at Yizheng (c), Quaternary reticulate red clay at Xuancheng (d), and Quaternary red clay at Jiujianng (e).
II. Geographical setting
The Hanjiang River terrace system is composed of alluvial and aeolian deposits along the flanks of river valleys. The lower part of the terrace system consists of fluvial deposits (clays, silts, sands, and gravels), and the upper part comprises loess deposit. Most studies focused on using alluvial terrace structure as a geodetic marker to infer tectonic, climatic, and environmental changes (Gao et al., 2016; Hu et al., 2017; Pan et al., 2009). By contrast, we focus on the aeolian loess deposit on the Hanjiang River terraces.
The Qinling Mountains (QLM) features an average elevation of 2000–3000 m above sea level, and it is regarded as the southern boundary of the CLP (Liu, 1985). In southern QLM, the loess deposit is thin and atypical. Thin loess deposits and associated paleosols occur widely on the southern slopes of the QLM along the Hanjiang River terraces (e.g., Guo et al., 2013, 2015; Liu, 1985; Sun et al., 2012, 2016; Wang et al., 2014; Zhang et al., 2012).
Aeolian loess sediments varying in thickness from 2 to 20 m overlie the fluvial sediments on low and high terraces along the Hanjiang River. Aeolian loess sediments can also reach up to 40 m at a large plain area. Although the sediment grain size is finer and the sediment accumulation rate is lower than those of the deposits on the CLP (Lu et al., 2007; Sun et al., 2012, 2016; Zhang et al., 2012), the loess deposit is significantly affected by post-depositional weathering and pedogenesis, resulting in sequences of alternating loess and paleosol layers, which can be correlated with the CLP (Sun et al., 2012, 2016, 2017).
Researchers also investigated aeolian loess along the Hanjiang River. Lei et al. (1999, 2000) reported the loess and paleosol deposition at Shangzhou Basin along the Danjiang River, which is a branch of the upper Hanjiang River. Recently, several reports have checked, identified, and dated the loess–paleosol sequence at Hanzhong and Yunxian Basins through detailed sedimentological analysis (Bian et al., 2014; Guo et al., 2013, 2015; Pang et al., 2015).
III. Ages of loess cover on the Hanjiang River terraces
Using previous research (Huang and Qi, 1987; Shen, 1956; Yang and Ma, 1987) and our investigations from accurate elevation measurement using a Trimble R8 GNSS System as the basis, new studies reveal the presence of six individual terrace levels (T1–T6) from low to high in Hanzhong Basin and five individual terrace levels (T1–T5) in Yunxian Basin (Sun et al., 2017). We dated several loess sections on the Hanjiang River terrace system from low to high.
We used optically stimulated luminescence including single aliquot regeneration optically stimulated luminescence (SAR-OSL), thermally transferred optically stimulated luminescence (TT-OSL), and post-infrared infrared stimulated luminescence (post-IR IRSL) dating methods and dated the samples from the Hejialiang (T2), Yaochangwan (T3), Longgangsi 2 (T4), and Longgangsi 3 (T5) sections in Hanzhong Basin and Houfang (T2) and Yuelianghe (T3) loess sections in Yunxian Basin (Sun et al., 2012, 2017). We employed paleomagnetic measurements at Yaochangwan (T3), Longgangsi 2 (T4), and Longgangsi 3 (T5) in Hanzhong and Wolonggang (T2) loess sections in Yunxian Basin (Sun et al., 2016, 2017).
3.1. Hanzhong Basin
On the first Hanjiang River terrace, a thin dark-gray sandy soil is found at the top and represents a modern cultivated land surface. No dating age is obtained there.
On the second terrace, the overlying aeolian loess cover is approximately 2–8 m thick. At the Hejialiang site, a paleosol was correlated with S1 paleosol from the CLP on the basis of two TT-OSL ages of 76.4 ± 5.0 and 86.3 ± 6.4 ka (Sun et al., 2012). The age range of the Hejialiang loess section is smaller than ∼100 ka (Table 1).
Site, depth, terrace, position (top, middle, and bottom of the loess section), dating methods, and age ranges of reference loess sections in Hanzhong Basin along the Hanjiang River.
TT–OSL: thermally transferred optically stimulated luminescence; CLP: Chinese Loess Plateau; pIR IRSL: post-infrared infrared stimulated luminescence.
On the third terrace, the loess cover is 5–20 m thick. At the Yaochangwan site, the loess section is 15 m thick, with four loess units (“L”) and five paleosol complexes (“S”) that are correlated with the sequence from Luochuan. Based on the TT-OSL ages, the palaeomagnetism dating result, and the magnetic susceptibility record, the following loess/paleosol units are recognized: S1, L2, S2, L3, S3, L4, S4, L5, and S5 (Sun et al., 2012). The age range of the Yaochangwan loess section is ∼ 90–600 ka (Table 1).
On the fourth terrace, the aeolian loess cover is 2–5 m thick. K-feldspar pIR-IRSL saturated ages indicate that the top part of the loess section at Longgangsi 2 are already older than ∼0.2 Ma, and the paleomagnetic data indicate that the loess at this site is younger than 0.78 Ma. The paleosol complex layer represents a combination of paleosol layers corresponding to S6 and S7 recorded in the central CLP sequence (Sun et al., 2017). The age range of the Longgangsi 2 loess section is ∼600–700 ka (Table 1).
On the fifth terrace, the Longgangsi 3 loess section is approximately 8 m thick. Based on the saturated K-feldspar pIR-IRSL age at the top part of the loess section, the stratigraphic position of the Brunhes/Matuyama boundary, Jaramillo subchron, and Cobb Mountain subchron, together with the magnetic susceptibility stratigraphy, the loess–paleosol sequence at Longgangsi 3 corresponds to the following succession: S8, L9, combined S9–S10–S11–S12, L13, S13, L14, S14, and L15 (Sun et al., 2017). The age range of the Longgangsi 3loess section is ∼700–1200 ka (Table 1).
The sixth terrace is located roughly 70 m above the modern riverbed. The aeolian layer is only 2–5 m-thick. No dating age is obtained there.
3.2. Yunxian Basin
The first terrace is found along both sides of the river. The age range of the Qianfang loess section on the first terrace is younger than ∼25 ka (Pang et al., 2014) (Table 2).
Site, depth, terrace, position (top, middle, and bottom of the loess section), dating methods, and age ranges of reference loess sections in Yunxian Basin along the Hanjiang River.
OSL: optically stimulated luminescence; CLP: Chinese Loess Plateau; pIR IRSL: post-infrared infrared stimulated luminescence.
On the second terrace, the loess is less than 10 m thick; we found and dated Dishuiyan and Houfang loess sections on the terrace. The pedostratigraphy of the loess section at Dishuiyan is recognized as L1, S1, and L2 in sequence, and the pedostratigraphy of the loess section at Houfang is recognized as L1, S1, L2, and S2 in sequence (Li et al., 2014, 2016; Sun et al., 2016). The age range of the Dishuiyan and Houfang sections is ∼25–185 ka (Table 2).
On the third terrace, the loess cover is approximately 5–10 m. K-feldspar pIR-IRSL saturated ages indicate that the top part of the Yuelianghu loess section on this terrace is already older than ∼200 ka (Sun et al., 2017) (Table 2).
On the fourth terrace, the aeolian deposit is 5–10 m thick and contains a thin loess unit and a paleosol complex. The pedostratigraphy of the loess section at Xuetangliangzi section is recognized as S8 and L9 (Guo et al., 2015; Sun et al., 2016). The age range of Xuetangliangzi is ∼700–1000 ka (Table 2).
On the fifth terrace, we found a Wolonggang loess section and dated it using a combination of detailed magnetostratigraphic analyses and pedostratigraphic correlation with the Luochuan loess–paleosol sequence on the CLP. The pedostratigraphy of the loess section is continuous as L9, S9, L10, S10, L11, S11, L12, S12, L13, S13, L14, S14, and L15 in sequence (Sun et al., 2016). The age range of Wolonggang loess cover on the fifth terrace is ∼900–1200 ka (Table 2).
IV. Uncommon deposition
Based on dating results, loess sections are from L15 to S8 with age range ∼1200–700 ka on the fifth terrace, from S7 to S6 with age range ∼600–700 ka on the fourth terrace, from S5 to S1 with age range ∼600–90 ka on the third terrace, and younger than S1 with age younger than ∼10 ka in the Hanzhong Basin (Table 1). Loess sections are from L15 to L9 with age range ∼1200–900 ka on the fifth terrace, from L9 to S8 with age range ∼1000–700 ka on the fourth terrace, from S2 to L1 with age range ∼185–25 ka on the second terrace, and younger than ∼25 ka on the first terrace in the Yunxian Basin (Table 2).
The same holds true for the fifth, fourth, third, and second terraces with different time ranges in Hanzhong and Yunxian Basins. An interesting and uncommon phenomenon appears. On high terraces, the loess deposited only during an old time range; on the middle terraces, the loess deposited during a middle time range; on the lower terraces, the loess deposited during a younger time range (Figure 3). Loess covers on each terrace are discontinuous when only a part of the whole section, but they are approximately continuous from higher to lower terraces in sequence as a whole section.
Setting of loess depositional process model based on the initial depositional rate, redepositional rate, initial erosional rate, true depositional rate, erosional rate, and current depositional rate on the Hanjiang River terrace system.
Although this phenomenon has never been discovered on the CLP, we do not think it is an isolated case. In certain terrace systems on the CLP, this phenomenon is not obvious, but younger loess layers (e.g., L1and S1) on higher terraces are thin and thick on lower terraces (Gao et al., 2016; Pan et al., 2005, 2009), which are caused by erosion. When the erosional rate is adequately large on the CLP, the thin loess layer on higher terraces should have eroded, but the thick loess layer on the lower terraces remains although it is probably thinner than before.
V. Discussion
5.1. Loess depositional hypothesis
Similar to deposition, erosion is a universal geomorphic characteristic. Some investigations had focused on loess erosion on the CLP (Dai, 1980; Hong et al., 1990; Jing and Chen, 1983; Zhao et al., 2002). Surface erosion is estimated as 14–34% (Li and Lu, 2010) of the deposition on the CLP. To the south of QLM, deposition is much smaller than the north, whereas the erosion is much larger than the north, inferring from rainfall. The average annual rainfall of Hanzhong and Yunxian Cities is ∼800–900 mm (Zhang et al., 2010; Zhu et al., 2009), whereas the average annual rainfall of the CLP is ∼400–500 mm (Mu et al., 1992; Zhao and Yang 2012).
The loess erosion on the terraces may be caused by rainfall, wind, humidity, vegetation cover, gravity, relative altitude, slope, and small surface area. Human activities can also destroy the loess deposition on the surface of every terrace, but it is not a universal factor. At the initial stage, the aeolian dust deposit on the surface of every terrace is loose. The plates of the Hanjiang River terraces are small and even with a slope. Then, for example, rainfall (overland flow) can easily erode the loose deposit on higher terraces and move it to lower terraces or into the river (Figure 4).
Supposition of depositional (brown block), erosional (gray block), and historical erosional (blank block) processes that occur on the Hanjiang River terrace system in contrast with those on the Chinese Loess Plateau (CLP). (Color online only.)
Therefore, erosion is a crucial factor. Some researchers regard erosion as an independent process. On the contrary, we propose that erosion and deposition are integrated. At least, erosion and deposition are synchronous in a long-term scale (e.g., during glacial and interglacial scales). We put forward a simple depositional hypothesis regarding the loess deposition along the Hanjiang River. In a long-term range, when deposition is smaller than erosion, it leads to erosion; when deposition equals erosion, it results in deposition or erosion stopping; when deposition is larger than erosion, deposition occurs.
5.2. Loess depositional process model on Hanjiang River terraces
The deposition on the Hanjiang River terrace system can be explained by this simple hypothesis: (a) On the highest terrace, no loess cover is found, and only gravels are left. The erosional rate is larger than the deposition rate. (b) On the fifth terrace, the age range of the loess section is ∼1200–700 ka in Hanzhong Basin. From ∼1200 to ∼700 ka, the deposition is larger than the erosion; thus, deposition occurs. Since ∼700 ka, no loess deposition is found on this terrace, indicating that the deposition rate equals the erosional rate. (c) On the fourth terrace, the age range of the loess section is ∼700–600 ka, which is the same case as that from ∼700 to ∼600 ka. The deposition is larger than the erosion, thus leading to deposition. Since ∼600 ka, no loess deposition occurs on this terrace, implying that the depositional rate equals the erosional rate. The condition is similar on the fifth, fourth, third, second, and first terraces in different time ranges.
Summarizing the rule, when Hanjiang River is downcut and a new T1 terrace formed, deposition decreases or stops on the old T1 terrace and deposition occurs and increases on the new T1 terrace at the lower part. Lower new T1 terraces are always the maximum depositional terraces (Figure 3). On higher terraces (T2, T3, T4, T5, and T6), deposition decreases, stops, or slight erosion happens in sequence.
The initial depositional and erosional rates on every terrace should be equal. Deposition only happens at lower terraces because the true depositional rate contains the initial deposition rate and redepositional rate (Figure 2). The changing of the current depositional rate only depends on the changing of the redepositional rate. The redepositional rate is mainly contributed from the erosional rate by wind and rainfall from higher terraces. Therefore, the redepositional rate on T1 contains the erosional rate that may come from T2, T3, T4, T5, and T6; on T2 from T3, T4, T5, and T6; and on T3 from T4, T5, and T6. The largest redepositional rate is on T1, then T2, T3, and so on in sequence.
The current depositional rate is high on low terraces and low on high terraces. Using the initial depositional rate, redepositional rate, true depositional rate, initial erosional rate, and erosional rate as a basis, we set up a loess depositional process model for the current depositional rate on Hanjiang River terraces (Figure 2).
Considering the true depositional rate and equal erosional rate on every terrace (Figure 2), the current depositional rate on the Hanjiang River terraces can be explained as follows: (a) on the lower (T1) terrace, the current deposition occurs when the true depositional rate is larger than the erosional rate; (b) on higher terraces (T2, T3, T4, T5, and T6), the current deposition decreases, stops, and slight erosion happens from low to high terraces in sequence when the true depositional rate is slightly larger, equal, and then slightly smaller than the erosional rate (Figure 3).
5.3. Loess deposition on and out of the CLP
On the CLP several erosional cases have been found, such as a 4–5 ka hiatus in typical loess deposits, indicating that the loess record is not continuous over millennial time scales (Lu et al., 2006), erosional surfaces found in loess strata on the CLP (Deng and Yuan, 2001; Xia, 1999), and estimated surface erosion (Hong et al., 1990; Jing and Chen, 1983; Li and Lu, 2010). Erosion also occurs on the CLP, and it is a universal process. Based on the exceptional case on the Hanjiang River terrace system, we reassessed the erosion on and out of the CLP. We propose that the deposition on the CLP is a result of an integrated process including both deposition and erosion. On the CLP, the depositional rate is much larger than the erosion rate, thus leading to thick deposits to several hundreds of meters.
In the Hanjiang River Range, the depositional rate is slightly larger than the erosional rate. Thus, the thickness of the Hanjiang loess is less than that of the classic loess on the CLP and is below 40 m. Further to the southeast in the middle and lower reaches of the Yangtze River, the depositional rate is just slightly larger than the erosional rate. The thickness of the Xiashu loess is less than that of the Hanjing loess and is under 30 m. Further south in the area approximately 31° N latitude, the depositional rate is slightly larger than or even equal to the erosional rate. This situation results in the thickness of the Quaternary reticulate red clay and red clay with aeolian dust characteristics to be below 20 m. In the area southward at approximately 30° N latitude, the erosional rate is larger than the depositional rate, thereby resulting in the occurrence of strong erosion. Therefore, aeolian dust deposits were considerably few.
Evidently, the integrated depositional process gradually decreases as the depositional rate decreases and the erosional rate increases from Central to southeast China. Consequently, the atypical loess deposition gradually thinned. Furthermore, loess and loess-like deposits in Europe and other areas, where the depositional rate of aeolian dust is slightly larger than the erosional rate, should be substantially similar to the ALDB between 30° and 34° N latitude of Southern China.
We used the loess depositional process on the Hanjiang River terraces to infer that all depositions on the river terraces approximately southward of 34° N latitude (Qinling Mountains) of China are in similar conditions. The aeolian deposits on the high, middle, and lower terraces are old, during the middle time range, and young, respectively. This phenomenon is beneficial for the age assessments of all archaeological sites along the rivers southward of 34° N latitude (QLM) of China. Archaeological sites buried in the aeolian deposit on the higher terraces should evidently be older than that on the lower terraces (Sun et al., 2012, 2016, 2017).
VI. Conclusions
We provide the following conclusions based on the detailed investigations of atypical loess deposition along the Hanjiang River in southern QLM. 1. The deposition of loess on the CLP and atypical loess, such as Hanjiang loess, Xiashu loess, Quaternary reticulate red clay, and Quaternary red clay, out of the CLP in southern China and in most areas of the world should be the result of an integrated process that includes deposition and erosion. 2. The depositional process model of the atypical loess along the Hanjiang River indicates that the deposition of aeolian origin is older on the higher terraces and younger on the lower terraces. 3. The deposition of aeolian origin on the river terraces, in approximately southward 34° N latitude (QLM) of China, and in any place in the world where the depositional rate is just limited or slightly larger than the erosional rate, should follow the depositional process model on the Hanjiang River terraces. Further investigation and evidence are necessary to support the depositional hypothesis and model.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Natural Science Foundation of China (grant numbers 41572155 and 41690111) and the Global Change Program of the Ministry of Science and Technology of China (grant number 2016YFA0600503).