Evolution model of a modern delta fed by a seasonal river in Daihai Lake, North China: determined from ground-penetrating radar and trenches

Abstract

While deltas fed by seasonal rivers are common in modern sedimentary environments, their characteristics remain unclear as compared to those fed by perennial rivers. This study identifies a small delta discharged by a seasonal stream flowing into Daihai Lake, in northern China, which is driven by ephemeral and high-energy flood events. Detailed 3D facies architecture was analyzed using ground-penetrating radar (GPR) and sedimentary logs from outcrop and trenches. Four types of radar surfaces, including truncations of underlying inclined strata, weak reflections, and depositional surface of downlap and onlap, were identified. Six radar facies (high-angle oblique-tangential, low-angle subparallel, gently plane parallel, plane-parallel, chaotic, and continuous strong reflection) were identified based on distinctive reflections, including amplitude, continuity, dip, and termination patterns. Five depositional units (Unit A to E) were documented from proximal to distal delta. Seasonal discharge signatures include significant grain-size decrease over short distance, abundant Froude supercritical flow sedimentary structures, poorly developed barforms, and small-scale scour and fill structures. Records of lake-level and sediment budget were evaluated over the past 60 years. In highstand stage (1960–1980), amalgamated channel (Units A and B), and delta front (Unit C) were deposited. In slope stage (1980–1996), the lower deposits (Units A, B, C) were eroded by Unit D with a distinct truncation surface. In lowstand stage, most eroded sediments bypassed the incised channel and accumulated in the distal part, in which a new depositional unit was formed (Unit E). The model demonstrates that deltas fed by seasonal rivers tend to accumulate large amounts of sediments carried by high magnitude floods within short periods.

Keywords

delta evolution / seasonal discharge / ground-penetrating radar / sedimentary architecture / Daihai Lake

Introduction

Deltas are land forms developed by the deposition of sediments at the mouth of a river as it reaches a standing body of water, i.e., lake or ocean. Deltaic reservoirs host approximately 30% of all of the world’s oil, coal, and gas (^{Tyler and Finley, 1992}) and as such, have been extensively investigated. Studies have shown that the flow dynamics in deltas play a significant role in terrain features (^{Reimnitz, 2000}) and the sedimentologic complexity of associated deposits strongly controls hydrocarbon recovery (^{Tyler and Finley, 1988}, 1992; ^{Ori et al., 1991}; ^{Tan et al., 2017}).

Delta facies models have been described in literature mostly focused deltas fed by rivers with constant discharge under limited climatic settings (^{Bhattacharya, 2006}) and alternatively, as fluvial deposits with varying discharge; claims which have been largely ignored thus far (^{Fielding et al., 2009}, ²⁰¹¹). As such, delta deposits formed by seasonal rivers have not received much attention despite their distinct sedimentary characteristics with different hydraulic processes (^{Demko, 2015}; ^{Fricke et al., 2015}; ^{Ventra et al., 2015}). Nevertheless, deposition affected by varied discharge has gained more interest in recent decades (^{Alexander and Fielding, 1997}; ^{Billi, 2007}; ^{Fielding et al., 2009}, ²⁰¹¹; ^{Allen et al., 2013}, ²⁰¹⁴; ^{Plink-Björklund, 2015}; ^{Ventra et al., 2015}; ^{Lowe and Arnott, 2016}; ^{Wang, 2018}). However, most of these studies focused on seasonal fluvial deposits whereas only a few detailed studies were conducted on deltas supplied by seasonal rivers. Such deltas are characteristic of high-energy and rapid deposition at the river mouth (^{García-García et al., 2011}). Abundant Froude supercritical flow sedimentary structures (^{Ventra et al., 2015}), upslope migrating bedforms (^{Dietrich et al., 2016}), and decimetre-scale couplets of coarser- and finer- grained flood event deposits (^{Gugliotta et al., 2016}) have been recognized as links to variable discharge.

Ground-penetrating radar (GPR) provides high-resolution reflection data from shallow subsurface, and has been commonly used to depict the architectural features of both modern and ancient depositional systems (^{Corbeanu et al., 2001}; ^{Best et al., 2003}; ^{Lunt and Bridge, 2004}; ^{Olariu et al., 2012}). In this study, GPR and trenches were excavated to resolve the 3D architecture of deltaic deposits built by a seasonal river (Bantanzi stream) on the northwestern margin of Daihai Lake, northern China (Fig. 1). The Bantanzi stream is significantly influenced by monsoon precipitation and experiences seasonal floods. In addition to previously documented variable discharge signatures in deltas, this study identifies characteristics including: 1) significant grain-size decrease over short distance, 2) poorly developed barforms, and 3) small-scale scour and fill structures. Furthermore, five depositional units were reconstructed showing a complex evolution by integrating lake-level and sedimentary records.

Study site

Daihai Lake is located in Inner Mongolia, northern China, at the margin of the area affected by the East Asian monsoon. This hydrologically closed basin was formed during the Late Pliocene to the Early Pleistocene (^{Li, 1972}) and filled with Quaternary alluvial deposits derived from the erosion of surrounding terrain. The lake is located at an elevation of 1219 m above sea level (m a.s.l.), covering an area of 70.6 km² with a maximum water depth of 16 m. The lake is gradually shrinking due to the arid climate and human activities, such as agriculture. The mean annual precipitation is 406 mm, with approximately 262 mm, or almost 65% of annual precipitation, occurring during the summer months under the effect of the Asian monsoon, which may promote flash flood events.

Daihai Lake is surrounded by many rivers, including the Bantanzi, Sandao, Muhua, Tiancheng, and Buliang, all of which form deltas (Fig. 1(a)). The Bantanzi River is seasonal river that originated in the mountains to the northwest. The streambed of the Bantanzi remains completely dry for most of the year and only yields a sufficient volume of runoff in response to monsoon rainfalls. The river consists of headwater tributaries, the main channel, and a terminal flow expansion without distributary systems (Fig. 1(b)). Its most recent active main channel scoured and eroded older deposits, and thus, its longitudinal section is suitable for sedimentological comparison with GPR profiles. Furthermore, generated floods have high kinetic energy on relatively steep slopes of 2°. The Bantanzi delta formed by the river is approximately 1.2 km long and 400–600 m wide. The Quaternary deposits mainly consist of medium- to coarse-grained sands and pebbles, with some cobbles.

Dataset and methods

To identify the internal radar characteristics and architectures, GPR surveying and stratigraphic logging along the cut bank section and trenches were carried out in the field. GPR data were collected by a TerraSIRch SIR 3000 GPR system with a 400 MHz antenna that provided high-resolution imageries up to depths of about 3.0 m. A total of five GPR profiles (3.93 km) were designed to cover the river-delta system in parallel and perpendicular to the feeder channel. In this study, the GPR Line_2 profile (1070 m) was selected to illustrate the sedimentary architecture and evolution of the delta system (Fig. 2).

GPR data processing involves frequency filtering, down-the-trace, trace-to-trace stacking, constant-velocity migration, and gain adjustment using Reflexw version 5.6. Mean velocity of 0.096 m/ns was used for time-depth conversion, and was determined by correlating sediments in trenches, the cut bank section, and GPR profile. Topographic correction along GPR lines was performed using a global-positioning system (Trimble GEO XT 2008) with a horizontal and vertical accuracy less than 1 m (^{Shan et al., 2015}).

As a geophysical method, GPR is based on the transmission of high-frequency electromagnetic signals and reception of energy reflected back from the subsurface (^{Bristow and Jol, 2003}; ^{Baker and Jol, 2007}). To link GPR results with sedimentologic records, a total of 12 stratigraphic logs (T5‒T16) were measured and compared with the GPR profiles to make sedimentary interpretations (Fig. 2). Radar facies and surfaces were identified based on the reflection characteristics (e.g., amplitude, continuity, termination patterns, and internal configuration). The interpreted radar profiles were used for architecture analysis. In addition, the annual records from 1960 to 2015 of lake-level and sediment input of the Daihai Lake were collected from both literature (Yu et al., 2013) and Google historical images to reconstruct the evolution of the Bantanzi delta.

Results

Ground-penetrating radar

Radar surfaces are identified as bounding surfaces and represent depositional breaks or unconformities in the sedimentary sequence (^{Neal, 2004}; ^{Baker and Jol, 2007}). Radar facies consist of sets of reflections with distinctive amplitude, continuity, dip, and shape that represent the bedding and internal structure of a sedimentary facies (^{Neal, 2004}; ^{Baker and Jol, 2007}). Four radar surfaces (Rs) are identified and subdivided into depositional (Rs 1-1 and 1-2) and erosional groups (Rs 2-1 and 2-2). Six radar facies (Rf) are separated by radar surfaces and subdivided into inclined (Rf 1-1 and 1-2), plane (Rf 2-1 and 2-2), and irregular (Rf 3-1 and 3-2) groups (Fig. 3).

Radar surface

Radar surface 1-1 (Rs 1-1) is a sharp erosional contact that separates the underlying inclined reflections (Rf 1-1) and the overlying low-angle subparallel reflections (Rf 1-2). This surface is defined by a change of geometry from reflections above and below, rather than a single continuous reflection (Fig. 3). Depending on the thickness of the transition zone, this surface locally produces high amplitude and either continuous or dispersed reflection. The lateral extent was typically 50–100 m (Fig. 3, Rs 1-1). This surface is well developed in distal profiles close to the paleo-shoreline. Trough surfaces overlying the inclined reflections indicate clinoforms eroded by channels. Rs 1-1 is interpreted as the transition zone between delta plain and front.

Radar surface 1-2 (Rs 1-2) is less sharply defined as erosional radar surface that overlies subparallel discontinuous reflections (Rf 1-2). The thickness varies between 0.5 m and 1.0 m, and the lateral extent ranges from 50 m to 80 m. This radar surface was predominantly found in proximal profiles associated with coarse and disorganized sediments. These less sharply defined erosional surfaces truncating weaker reflections below suggest the erosion of older deposits by channels in the proximal part.

Radar surface 2-1 (Rs 2-1) is characterized by the overlying steeply lakeward-dipping reflection downlap onto a high-to-moderate amplitude and continuous reflection (Fig. 3, Rs 2-1). This radar surface primarily occurs under the paleo-shoreline in the distal part and is laterally associated with Rs 1-1. The continuous high-angle downlap surface (Rs 2-1) that occurs above the flat reflections suggest delta front progradation on the delta toe. This radar surface primarily occurs at the base of the delta front.

Radar surface 2-2 (Rs 2-2) comprises a high amplitude and continuous reflection onlapped by horizontal or low-angle reflections (Fig. 3, Rs 2-2). The underlying reflections are concordant with Rs 2-2, which dips lakeward at a high angle and progressively flattens upwards becoming sub-horizontal lakeward. This radar surface typically occurred at the most distal part of the depositional system, and was only found in distal trenches. The onlapping horizontal or low-angle reflections, and the underlying, high-angle lakeward-dipping reflections indicate that new delta sediments were deposited directly on the former delta front.

Radar facies

Radar Facies 1-1 (Rf 1-1) is characterized by high amplitude, continuous, and parallel reflections, which dip lakeward at a high angle and downlap obliquely to tangentially onto radar surface Rs 2-1 (Fig. 3, Rf 1-1). This radar facies was well developed above radar surface Rs 2-1 and appeared to correspond to the delta front, which prograded into the lake.

Radar Facies 1-2 (Rf 1-2) is characteristic of high amplitude, discontinuous, and low-angle subparallel reflections. It has less continuity and lower dip angles as compared to Rf 1-1 (Fig. 3, Rf 1-2). This radar facies was predominantly found in proximal coarse-grained sandstone deposits. The gentle lakeward-dipping reflections were interpreted as downward accretions and internal erosion in the proximal channels of the delta plain.

Radar Facies 2-1 (Rf 2-1) is characterized by high–moderate amplitude, insignificant continuous, and parallel reflections (Fig. 3). This radar facies is well developed in the distal region relative to Rf 1-2 and is composed of medium- to coarse-grained sandstone deposits. The gentle plane reflections were most likely interpreted as sand sheet filled in the distal channel of the delta plain.

Radar Facies 2-2 (Rf 2-2) consists of sets of parallel reflections at low amplitude reflections with high continuity (Fig. 3). This radar facies occurred directly beneath the Rs 2-1 surface or Rf 1-1 facies, which is dominated by fine-grained sands and muds. The high continuity indicates that fine sediments were deposited at the delta toe.

Radar Facies 3-1 (Rf 3-1) is characterized by extremely low amplitude and discontinuous reflections, which appear chaotic and weak (Fig. 3). This radar facies is typically developed in the proximal region, which consists of mostly disorganized conglomerates. It is interpreted as the coarsest sediments deposited in the proximal channel.

Radar Facies 3-2 (Rf 3-2) appears as very high amplitude, continuous, and strong reflections at the top of GPR profiles (Fig. 3). It is interpreted as having the youngest deposits independent of any barriers without signal attenuation.

Radar profiles

The radar profile was separated into five continuous profiles to illustrate the detailed internal structure of the delta (Fig. 4). All the GPR profiles have a maximum two-way travel time of 50 ns corresponding to a depth of ca 2.8 m.

The most proximal GPR profile (0‒200 m) shows the characteristics of channelized units (Fig. 4(a)). Stacked channel patterns are characterized by irregular basal erosion surface (Rs 1-2). Internal features are predominantly low-angle cross-stratification (Rf 1-2), plus chaotic radar facies (Rf 3-1) and gentle plane reflections (Rf 2-1), which represent very coarse-grained and unorganized sediments in proximal and planar laminated sandbodies, respectively. Two stacked channel units (Unit A and B) were identified in the proximal part, and interpreted as upstream channels in the fluvial-delta system. The lower channel unit (Unit A) was partially eroded by the uppermost deposits (Unit B).

The following GPR profile (200‒400 m) comprises vertically stacked delta front (Unit C) and small-scale channel units (Unit A) separated by Rs 1-1 (Fig. 4(b)). Small-scale channels have a high width-to-thickness ratio and interpreted as extensions from the proximal channels. The internal structures of the channels in Unit A are predominated by gentle plane and parallel radar facies (Rf 2-1). Sheet (Rf 2-1) or low-angle strata (Rf 1-2) in Unit C is interpreted as lobe deposits in delta front. This area on the radar profile was considered to be the transition zone from channel to lobe.

The next connected GPR profile (400–600 m) was distinguished by inclined reflection patterns that gradually transition from low-angle (Rf 1-2) to high-angle (Rf 1-1) lakeward (Fig. 4(c)). The inclined strata are about 120 m long and 1.5 m thick, which indicates that the clinoforms were deposited by delta front progradation. The lower boundary of the clinoforms is a downlap surface (Rs 2-1). This clinoform (Unit C) was eroded by high amplitude, continuous, and strong reflections (Unit D) with a distinct unconformity surface (Rs 1-1).

The following GPR profile (600‒800 m) has the steepest slope corresponding to the recent topography in slope area (Figs. 2 and 4(d)). The upper Unit C (600‒660 m) is featured with low-angle to high-angle lakeward-dipping reflectors (Rf 1-2 and Rf 1-1) and the lower Unit C (660‒800 m) is dominated by parallel and plane reflections (Rf 2-2). In the distal slope (~720 m), a distinct lakeward-dipping onlap surface could be identified (Rs 2-2), which separates Unit C below and Unit E above. This phenomenon of two distinguished reflectors can be interpreted as two depositional cycles induced by significant climate change.

The connected GPR profile (800‒1070 m) was located at the most distal part of the depositional system (Fig. 4(e)). Unit E is characteristic of high-angle reflectors (Rf 1-1) and parallel, plane, and weak reflections (Rf 2-2), which are interpreted as delta front and delta toe deposits, respectively.

Sediments in trenches

Twelve trenches (T5‒T16) along the GPR profiles were used to determine the sedimentary characteristics of the delta from the proximal to distal zones and to ascertain the relationship between radar reflections and sedimentological records (Fig. 5).

The deposits in the Bantanzi delta ranged from cobbles (T5‒T7), to pebbly coarse-grained sands (T8‒T10), coarse- to medium-grained sands (T11‒T13), to fine-grained sands and dark grey muds (T14‒T16) (Fig. 5). The wide variation in grain size within 1 km indicates that the carrying flows were strong and declined rapidly, which is attributed to the hydrodynamic characteristics of flash floods (^{Grimm et al., 1995}; ^{Powell et al., 2001}; ^{Rhodes et al., 2013}).

Sedimentary structures in this system include a suite of structures produced by upper-flow regimes with rapid sedimentation rates, which characterize flood events caused by seasonal precipitation. Planar laminations dominated proximal pebbly deposits (F1 in Fig. 5), which are interpreted as upper plane bed (^{Allen, 1984}; ^{Bridge and Best, 1988}; ^{Best and Bridge, 1992}). Low-amplitude, convex-upward laminations with frequent internal truncations commonly occurred in sandy sediments (F2 in Fig. 5, and Fig. 7 in ^{Tan et al. (2018)}). This sedimentary structure is regarded as antidunes formed by supercritical flows (e.g., ^{Fielding, 2006}; ^{Lang and Winsemann, 2013}; ^{Lang et al., 2017a}, ^b). The channels occasionally featured backset stratifications (F3 in Fig. 5; and Fig. 8 in Tan et al. (2018)), which are classic features generated by supercritical flows with hydraulic jumps (e.g., ^{Schmincke et al., 1973}; ^{Fralick, 1999}; ^{Fiore et al., 2002}; ^{Fielding, 2006}). Abundant asymmetrical and symmetrical ripple-like bedforms were identified in the delta front (F4 in Fig. 5). Delta toe deposits were primarily comprised of lenticular, flaser, and wavy bedding-like heterolithics (F5 in Fig. 5) in the lower area. These sedimentary structures in the delta front and delta toe have a possible genesis of supercritical flow. Spinewine et al. (2009) generated centimetre-scale bedforms in a delta front under a supercritical flow regime, which have similar morphology and internal features to those found in the Bantanzi delta. A detailed discussion about the facies produced by supercritical flow has been systematically organized in ^{Tan et al. (2018)}. In addition to these upper flow regime structures, soft-sediment deformation and vertical burrow were found in the delta front, suggesting rapid sedimentation and turbulent environment at the mouth of the river (^{Postma, 1984}; ^{Martinsen, 1989}; ^{Callot et al., 2009}; ^{Buatois et al., 2012}; ^{Pisarska-Jamroży and Weckwerth, 2013}).

Two types of sedimentary sequences could be identified from the trenches including fining-upward (T5‒T10) and coarsing-upward (T11‒T16) sequences, which indicate channel and delta front genesis, respectively (Fig. 5).

Individual radar profiles were connected to form one continuous profile to determine the architectural relationships. Five depositional units (Units A‒E) could be identified from both the continuous radar profile and trench correlation (Fig. 6). Lithologic surfaces from trenches could be correlated well with the corresponding radar surfaces.

Deposits of Unit B, which are dominated by cobbles and pebbles, overlie the fine sediments of Unit A. The surface between Unit A and B indicate that the former channel was eroded by the next one. Unit C, with high-angle clinoforms in the delta front, directly underlies proximal channels. The unconformity between Unit D and underlying units A, B, and C was very clear in the trenches and radar profile, indicating a shift of allogeneic controls. In the distal part, trenches (T15‒T16) less than 1.1 m were dug due to the shallow depth of the water table. The deposits of Unit E comprised sandy and muddy sediments.

Discussion

Architectural elements

The GPR profile and trench correlation were integrated for architecture analysis of the Bantanzi delta. Two essential architectural elements, channels and clinoforms, were identified with distinct features (Fig. 7).

Cut and fill structures with lenticular or wedge shapes were interpreted as main channels in the delta plain. According to internal fills, the channel could be subdivided into two types. Type I channels are filled with low angle downstream accretions with lagged gravels on the erosional base, and commonly occur in the proximal part (Fig. 4(a)). The erosional surface and internal fills correspond to Rs 1-2 and Rf 1-2, respectively. The thickness of the Type I channel ranged from 0.5 m to 1.5 m, with a length of up to 50 m, which can be attributed to channels with perennial discharge in the small scale. Type II channels are characterized by undulated and shallow based erosional surface, which can be identified as Rs 1-1 on the GPR profile. The channel was filled by planar or low angle aggradational packages that appear as Rf 2-1 on the GPR profile. The geometry of Type II channels, with length and thickness of 180‒260 m and 0.4‒1.2 m, respectively, is much flatter compared to that of Type I channels.

Clinoforms are formed by deposition with the progradation of the delta front in the slope. Two subdivisions of clinoforms could be distinguished according to the dip angle. Low-angle clinoforms underlying the channel element appeared at the transition from channels to high-angle clinoforms. They extended 200‒300 m, with a thickness ranging 0.4‒1.0 m. This element represents the expansion of sediment discharge on the ‘shelf’. High-angle clinoforms, which extend lakeward up to 120 m with a thickness ranging 1.2‒1.5 m, could be easily recognized from the radar profile (Rf 1-1) and cut bank exposure.

Variable discharge signatures

The architectural elements of deltas fed by seasonal rivers differ from ‘conventional’ deltas supplied by perennial rivers (^{Johnson and Graham, 2004}; ^{Gani and Bhattacharya, 2007}; ^{Ahmed et al., 2014}). ^{Fielding (2006)} defined upper flow regime (UFR) sheets, lenses, and scour fills as indicators of a strongly seasonal paleoclimate that involves a pronounced seasonal peak in precipitation and runoff. With the investigation of the architecture of the Bantanzi delta, we expand the criteria for identifying records of seasonal discharge, which may help interpret ancient successions.

The distinct features include:

1) Significant grain-size decrease over short distance, indicating rapid flow deceleration—a characteristic of flood events (e.g., ^{Chakraborty and Ghosh, 2010});

2) Abundant sedimentary structures attributable to supercritical flow preserved in all architectural elements, indicating high-magnitude floods generated by abrupt increases of discharge (^{Billi, 2008}, ²⁰¹¹; ^{Ventra et al., 2015});

3) Poorly developed barforms in the channels instead of laminated aggradational packages, which have been suggested as a characteristic of rivers with seasonally variable discharge (e.g., ^{Rhee and Chough, 1993}; ^{Eji Uba et al., 2005}; ^{Billi, 2007}; ^{Hampton and Horton, 2007}; ^{Fielding et al., 2009}; ^{Hulka and Heubeck, 2010}; ^{Allen et al., 2014});

4) Small-scale cut and fills formed by surges in areas indicating hydraulic jump induced by unstable non-uniform flows at higher Froude numbers (^{Karcz and Kersey, 1980}; ^{Cartigny et al., 2014}).

Evolution model: response to lake-level change and sediment supply

Lake-level change

From 1960 to 2015, the area of Daihai Lake has gradually decreased from 160.12 to 70.6 km² due to the arid climate and agricultural activities (Figs. 8 and 9). The lake area expanded and reached the upper terrace in the 1970s, but then decreased in size over the next several years (personal communication with local residents).

Records of the Daihai Lake water levels taken from 1960 to 2015, and collected from Yu et al. (2013) and ^{Huang and Jiang (1999)}, show a persistent decline since 1960. According to the paleo-shoreline position, three stages are identified: highstand stage, slope stage, and lowstand stage (Fig. 10). The ‘highstand stage’ (from 1960‒1980) is the period during which the lake-level was maintained above the 1225 m a.s.l. breakpoint (Figs. 9(a) and 10). During the highstand stage (1960‒1980), the lake level ranged from 1225 to 1226 m a.s.l. with minor changes. From year 1980 to 1996, the lake-level fell from 1225 to 1221 m a.s.l. across the slope between the upper and lower terraces. The slope stage experienced a relatively rapid fall rate of ~0.25 m/yr which can be considered a forced regressive stage in the sequence stratigraphy concept. After 1996, the lake-level continued to decline across the lower terrace from 1221 to 1219 m a.s.l. with a relatively slow rate of ~0.1 m/yr. In the lowstand stage, the upper terrace is expected to be deeply incised, with an accumulation of sediments in the lower terrace.

Sediment supply

There are no specific sediment input data for the Bantanzi River; however, the total sediment input data of the Daihai Lake from 1960 to 1995 were obtained from previously written literature (Fig. 9(b)) (^{Huang and Jiang, 1999}; Yu et al., 2013). Hence, some semi-qualitative recognition for the Bantanzi River could be obtained from the total input discharge data.

The average sediment input of the Daihai Lake during the highstand stage (1960‒1980) reached 0.9 × 10⁸ m³ in contrast to 0.45 × 10⁸ m³ during the slope stage (Fig. 9(b)). The sediment input after 1995 was estimated from the relevant rainfall and evaporation data, as the data after 1995 is missing in previous literature. Through field observation in recent years, streams around Daihai Lake were found to remain dry for most of the year and then suddenly generated flash floods during the rainy seasons. Moreover, due to the increased evaporation and inconsistent rainfall (Fig. 9(b)), the total input discharge during the lowstand stage is believed to remain low, which is estimated at 0.25 × 10⁸ m³ on average.

Evolution model

The evolution of the fluvial-deltaic system was reconstructed using a combination of the analysis of lake-level change, sediment supply and sedimentary architecture from the radar profile, and trench correlation (Fig. 11). Despite the lack of direct chronological data, we attempted to assign a relative chronological framework based on the known lake-level records.

In the early stage, the lake level remained at a relative highstand and then slowly declined from 1226 to 1225 m a.s.l. between 1960 and 1980. The vertical lake level ranged from 1225 to 1226 m, covering the area from T11 to T13 (yellow area in Fig. 10) in the upper terrace, implying the paleo-shoreline shifted its position between these three trenches in the early stage. The moving paleo-shoreline, combined with distinctive high-angle clinoform architecture in the corresponding section (Fig. 6(d)), and the presence of lacustrine shells in the inclined strata (Fig. 5), indicate that delta front sedimentation (Unit C) occurred in this area. This is supported by the sufficient sediment supply (specify the number) present during this stage (Fig. 11(a)). This thick deltaic system was mainly formed during a flood event, due to the increased deposition of sediment. Furthermore, the gap between each flood event was of benefit towards the accumulation of sediment.

During the slope stage (1980‒1996), the lake-level declined across the slope area at a rapid rate with decreasing sediment supply. The forced regression caused former deposits of Units A, B, and C to erode, Unit D to be deposited, with the formation of unconformity in between (Figs. 6 and 10(b)). Although the average annual discharge in the slope stage (0.45 × 10⁸ m³) was half that of the highstand stage (0.9 × 10⁸ m³), the thickness of Unit D was 0.3–0.5 m, one third that of the lower units (1–1.5 m). This is interpreted as a result of lower flood event frequency during the slope stage when Unit D was deposited. The highstand stage experienced several high input discharge years that significantly contributed to the thicker deposits due to violent floods (Fig. 9(b)). This phenomenon of the reduction in flood events can also support the theory that Unit D was deposited during the slope stage.

In the lowstand stage (1996‒2015), the upper terrace was deeply incised and the active channel, at a depth of ~1.8 m, was created. Most eroded sediments bypassed the incised channel and accumulated in the distal part, in which a new depositional unit was formed (Unit E) (Fig. 11(c)). Several abandoned splay lobes aligned with the active channel in the lower terrace can be identified from Google images (Figs. 2 and 10). These suggest the periodicity of flood events and confined deposition in the lower terrace during the lowstand stage. However, the sediment volume of Unit E is not typically compared to that of the former units due to the significant decrease in sediment supply during this stage.

The evolution of the Bantanzi delta was reconstructed according to the architecture and semi-quantitative lake-level change and sediment supply. In addition, there was an accumulation of a ca. 2 m thickness of sediments over a 50-year period. As compared to deltas fed by perennial rivers, those supplied by seasonal rivers tend to accumulate large amounts of sediments carried by violent floods within a short period. The great sedimentation rate is ascribed to the extreme variability in the discharge pattern. This recognition may give a new perspective for future study of ancient flood deposits.

Conclusions

Radar surface, radar facies, and sedimentary architecture of the Bantanzi delta in Daihai Lake were defined by GPR data and trenches. According to the fundamental principle of seismic stratigraphy, four radar surfaces were identified in the radar profile, which could be summarized into depositional and erosional groups. Three groups of six different radar facies (inclined, plane, and irregular) were identified based on distinctive reflections including amplitude, continuity, dip, and termination patterns.

Two essential architectural elements, channel and clinoform, were identified with distinct features, which could be subdivided into four types according to their internal characteristics. With the architecture study of the Bantanzi delta, the criteria for identifying records of seasonal discharge could be expanded. The criteria include a significant decrease in grain size within a short distance, abundant supercritical flow of sedimentary structures, poorly developed barforms, and small-scale cut and fill structures.

Five depositional units were distinguished from the GPR profile and trench correlation. An evolution model of this depositional system was reconstructed based on the analysis of lake-level variation, sediment supply, and the following depositional units. The model demonstrates that deltas supplied by seasonal rivers tend to accumulate large amounts of sediments carried by violent floods within a short period. Finally, we conclude that seasonal flood events play an important role in building deltas fed by seasonal rivers.

References

[1] Ahmed S, Bhattacharya J P, Garza D E, Li Y (2014). Facies architecture and stratigraphic evolution of a river-dominated delta front, Turonian Ferron Sandstone, Utah, U.S.A. J Sediment Res, 84(2): 97–121

[2] Alexander J, Fielding C (1997). Gravel antidunes in the tropical Burdekin River, Queensland, Australia. Sedimentology, 44(2): 327–337

[3] Allen J P, Fielding C R, Gibling M R, Rygel M C (2014). Recognizing products of palaeoclimate fluctuation in the fluvial stratigraphic record: an example from the Pennsylvanian to Lower Permian of Cape Breton Island, Nova Scotia. Sedimentology, 61(5): 1332–1381

[4] Allen J P, Fielding C R, Rygel M C, Gibling M R (2013). Deconvolving signals of tectonic and climatic controls from continental basins: an example from the Late Paleozoic Cumberland Basin, Atlantic Canada. J Sediment Res, 83(10): 847–872

[5] Allen J R L (1984). Parallel lamination developed from upper stage plane beds; a model based on the larger coherent structures of the turbulent boundary layer. Sediment Geol, 39(3‒4): 227–242

[6] Baker G S, Jol H M (2007). Stratigraphic analyses using GPR. Spec Pap Geol Soc Am, 432: 1‒181

[7] Best J L, Ashworth P J, Bristow C S, Roden J (2003). Three-dimensional sedimentary architecture of a large mid-channel sand braid bar, Jamuna River, Bangladesh. J Sediment Res, 73(4): 516–530

[8] Best J, Bridge J (1992). The morphology and dynamics of low amplitude bedwaves upon upper stage plane beds and the preservation of planar laminae. Sedimentology, 39(5): 737–752

[9] Bhattacharya J P (2006). Deltas. In: Posamentier H W, Walker R G, eds. Facies Models Revisited. Special publication of Society for Sedimentary Geology, 84: 237–292

[10] Billi P (2007). Morphology and sediment dynamics of ephemeral stream terminal distributary systems in the Kobo Basin (northern Welo, Ethiopia). Geomophology, 85(1‒2): 98–113

[11] Billi P (2008). Bedforms and sediment transport processes in the ephemeral streams of Kobo basin, Northern Ethiopia. Catena, 75(1): 5–17

[12] Billi P (2011). Flash flood sediment transport in a steep sand-bed ephemeral stream. Int J Sediment Res, 26(2): 193–209

[13] Bridge J S, Best J L (1988). Flow, sediment transport and bedform dynamics over the transition from dunes to upper-stage plane beds: implications for the formation of planar laminae. Sedimentology, 35(5): 753–763

[14] Bristow C S C, Jol H M (2003). An introduction to ground penetrating radar (GPR) in sediments. Geol Soc Lond Spec Publ, 211(1): 1–7

[15] Buatois L A, Santiago N, Herrera M, Plink-Björklund P, Steel R, Espin M, Parra K (2012). Sedimentological and ichnological signatures of changes in wave, river and tidal influence along a Neogene tropical deltaic shoreline. Sedimentology, 59(5): 1568–1612

[16] Callot P, Odonne F, Debroas E J, Maillard A, Dhont D, Basile C, Hoareau G (2009). Three-dimensional architecture of submarine slide surfaces and associated soft-sediment deformation in the Lutetian Sobrarbe deltaic complex (Ainsa, Spanish Pyrenees). Sedimentology, 56(5): 1226–1249

[17] Cartigny M J B, Ventra D, Postma G, Van Den Berg J H (2014). Morphodynamics and sedimentary structures of bedforms under supercritical-flow conditions: new insights from flume experiments. Sedimentology, 61(3): 712–748

[18] Chakraborty T, Ghosh P (2010). The geomorphology and sedimentology of the Tista megafan, Darjeeling Himalaya: implications for megafan building processes. Geomorphology, 115(3–4): 252–266

[19] Corbeanu R M, Soegaard K, Szerbiak R B, Thurmond J B, Mcmechan G A, Wang D (2001). Detailed internal architecture of a fluvial channel sandstone determined from outcrop, cores, and 3-d ground-penetrating radar: example from the middle cretaceous ferron sandstone, east-central utah. AAPG Bull, 85(9): 1583–1608

[20] Demko T (2015). Evidence of supercritical flow during deposition in delta front clinothems of the Eocene Sant Llorenç del Munt fan delta, Ebro Basin, Spain. In: Proceedings of 31st IAS Meeting of Sedimentology, Kraków, Poland, 150

[21] Dietrich P, Ghienne J F, Normandeau A, Lajeunesse P (2016). Upslope-migrating bedforms in a proglacial sandur delta: cyclic steps from river-derived underflows? J Sediment Res, 86(2): 113–123

[22] Eji Uba E, Heubeck C, Hulka C (2005). Facies analysis and basin architecture of the Neogene Subandean synorogenic wedge, southern Bolivia. Sediment Geol, 180(3–4): 91–123

[23] Fielding C R (2006). Upper flow regime sheets, lenses and scour fills: extending the range of architectural elements for fluvial sediment bodies. Sediment Geol, 190(1‒4): 227–240

[24] Fielding C R, Allen J P, Alexander J, Gibling M G (2009). Facies model for fluvial systems in the seasonal tropics and subtropics. Geology, 37(7): 623–626

[25] Fielding C R, Allen J P, Alexander J, Gibling M R, Rygel M C, Calder J H (2011). Fluvial systems and their deposits in hot, seasonal semiarid and subhumid settings: modern and ancient examples. In: Davidson S K, Leleu S, North C P, eds. From River to Rock Record: The Preservation of Fluvial Sediments and their Subsequent Interpretation. SEPM special publication, 97: 89–111

[26] Fiore J, Pugin A, Beres M (2002). Sedimentological and GPR studies of subglacial deposits in the Joux Valley (Vaud, Switzerland): backset accretion in an esker followed by an erosive jokulhlaup. Geogr Phys Quat, 56(1): 19‒32

[27] Fralick P (1999). Paleohydraulics of chute-and-pool structures in a Paleoproterozoic fluvial sandstone. Sediment Geol, 125(3‒4): 129–134

[28] Fricke A T, Sheets B A, Nittrouer C A, Allison M A, Ogston A S (2015). An examination of Froude-supercritical flows and cyclic steps on a subaqueous lacustrine delta, Lake Chelan, Washington, U.S.A. J Sediment Res, 85(7): 754–767

[29] Gani M R, Bhattacharya J P (2007). Basic building blocks and process variability of a Cretaceous delta: internal facies architecture reveals a more dynamic interaction of river, wave, and tidal processes than is indicated by external shape. J Sediment Res, 77(4): 284–302

[30] García-García F, Corbí H, Soria J M, Viseras C (2011). Architecture analysis of a river flood-dominated delta during an overall sea-level rise (early Pliocene, SE Spain). Sediment Geol, 237(1–2): 102–113

[31] Grimm M M, Wohl E E, Jarrett R D (1995). Coarse-sediment distribution as evidence of an elevation limit for flash flooding, Bear Creek, Colorado. Geomorphology, 14(3): 199–210

[32] Gugliotta M, Kurcinka C E, Dalrymple R W, Flint S S, Hodgson D M (2016). Decoupling seasonal fluctuations in fluvial discharge from the tidal signature in ancient deltaic deposits: an example from the Neuquénn Basin, Argentina. J Geol Soc London, 173(1): 94–107

[33] Hampton B A, Horton B K (2007). Sheetflow fluvial processes in a rapidly subsiding basin, Altiplano plateau, Bolivia. Sedimentology, 54(5): 1121–1148

[34] Huang Q, Jiang J (1999). Analysis of water level descent in Daihai Lake. J Lake Sci, 11(4): 304–310 (in Chinese)

[35] Hulka C, Heubeck C (2010). Composition and provenance history of late Cenozoic sediments in Southeastern Bolivia: implications for Chaco foreland basin evolution and Andean uplift. J Sediment Res, 80(3): 288–299

[36] Johnson C L, Graham S A (2004). Sedimentology and reservoir architecture of a synrift lacustrine delta, southeastern Mongolia. J Sediment Res, 74(6): 770–785

[37] Karcz I, Kersey D (1980). Experimental study of free-surface flow instability and bedforms in shallow flows. Sediment Geol, 27(4): 263–300

[38] Lang J, Brandes C, Winsemann J (2017a). Erosion and deposition by supercritical density flows during channel avulsion and backfilling: field examples from coarse-grained deepwater channel-levée complexes (Sandino Forearc Basin, southern Central America). Sediment Geol, 349: 79–102

[39] Lang J, Sievers J, Loewer M, Igel J, Winsemann J (2017b). 3D architecture of cyclic-step and antidune deposits in glacigenic subaqueous fan and delta settings: integrating outcrop and ground-penetrating radar data. Sediment Geol, 362: 83–100

[40] Lang J, Winsemann J (2013). Lateral and vertical facies relationships of bedforms deposited by aggrading supercritical flows: from cyclic steps to humpback dunes. Sediment Geol, 296(14): 36–54

[41] Li H (1972). The formation of Daihai Lake and its topographical features. Journal of Beijing Normal University (Natural Science), 1: 98‒110 (in Chinese)

[42] Lowe D G, Arnott R W C (2016). Composition and architecture of braided and sheetflood-dominated ephemeral fluvial strata in the Cambrian–Ordovician Potsdam Group: a case example of the morphodynamics of early Phanerozoic fluvial systems and climate change. J Sediment Res, 86(6): 587–612

[43] Lunt I A, Bridge J S (2004). Evolution and deposits of a gravelly braid bar, Sagavanirktok River, Alaska. Sedimentology, 51(3): 415–432

[44] Martinsen O J (1989). Styles of soft-sediment deformation on a Namurian (Carboniferous) delta slope, Western Irish Namurian Basin, Ireland. Geol Soc Lond Spec Publ, 41(1): 167–177

[45] Neal A (2004). Ground-penetrating radar and its use in sedimentology: principles, problems and progress. Earth Sci Rev, 66(3‒4): 261–330

[46] Olariu C, Bhattacharya J P, Leybourne M I, Boss S K, Stern R J (2012). Interplay between river discharge and topography of the basin floor in a hyperpycnal lacustrine delta. Sedimentology, 59(2): 704–728

[47] Ori G G, Roveri M, Nichols G (1991). Architectural patterns in large-scale Gilbert-type delta complexes, Pleistocene, Gulf of Corinth, Greece. In: Miall A D, Tyler N, eds. The Three-Dimensional Facies Architecture of Terrigenous Clastic Sediments and Its Implications for Hydrocarbon Discovery and Recovery – Concepts in Sedimentology and Paleontology. Society for Sedimentary Geology, 3: 207–216

[48] Pisarska-Jamroży M, Weckwerth P (2013). Soft-sediment deformation structures in a Pleistocene glaciolacustrine delta and their implications for the recognition of subenvironments in delta deposits. Sedimentology, 60(3): 637–665

[49] Plink-Björklund P (2015). Morphodynamics of rivers strongly affected by monsoon precipitation: review of depositional style and forcing factors. Sediment Geol, 323: 110–147

[50] Postma G (1984). Slumps and their deposits in delta-front and slopes. Geology, 12(1): 27–30

[51] Powell D M, Reid I, Laronne J B (2001). Evolution of bed load grain size distribution with increasing flow strength and the effect of flow duration on the caliber of bed load sediment yield in ephemeral gravel bed rivers. Water Resour Res, 37(5): 1463–1474

[52] Reimnitz E (2000). An overview of the Lena River Delta setting: geology, tectonics, geomorphology, and hydrology. J Coast Res, 16(4): 1083–1093

[53] Rhee C W, Chough S K (1993). The Cretaceous Pyonghae Basin, southeast Korea: sequential development of crevasse splay and avulsion in a terminal alluvial fan. Sediment Geol, 83(1‒2): 37–52

[54] Rhodes R, Goodwin K, Johnson J P (2013). Transported and surface grain size changes during experimental flash floods. In: Proceddings of American Geophyiscal Union Fall Meeting, San Francisco

[55] Schmincke H U, Fisher R V, Waters A C (1973). Antidune and chute and pool structures in the base surge deposits of the Laacher See area, Germany. Sedimentology, 20(4): 553–574

[56] Shan X, Yu X, Clift P D, Tan C, Jin L, Li M, Li W (2015). The ground penetrating radar facies and architecture of a Paleo-spit from Huangqihai Lake, North China: implications for genesis and evolution. Sediment Geol, 323: 1–14

[57] Spinewine B, Sequeiros O E, Garcia M H, Beaubouef R T, Sun T, Savoye B, Parkers G (2009). Experiments on wedge-shaped deep sea sedimentary deposits in minibasins and/or on channel levees emplaced by turbidity currents. Part II. Morphodynamic evolution of the wedge and of the associated bedforms. J Sediment Res, 79(8): 608–628

[58] Tan C, Yu X, Liu B, Qu J, Zhang L, Huang D (2017). Conglomerate categories in coarse-grained deltas and their controls on hydrocarbon reservoir distribution: a case study of the Triassic Baikouquan Formation, Mahu Depression, NW China. Petrol Geosci, 23(4): 403–414

[59] Tan C, Yu X, Liu B, Xu L, Li S, Feng S, Tang Y (2018). Sedimentary structures formed under upper-flow-regime in seasonal river system: a case study of Bantanzi River, Daihai Lake, Inner Mongolia. Journal of Paleogeography (Chinese edition), 20(6): 929–940 (in Chinese)

[60] Tyler N, Finley R J (1988). Reservoir architecture−a critical element in extended conventional recovery of mobile oil in heterogeneous reservoirs. AAPG Bull, 72(2): 255

[61] Tyler N, Finley R J (1992). Architectural controls on the recovery of hydrocarbons from sandstone reservoirs. In: Miall A D, Tyler N, eds. The Three-Dimensional Facies Architecture of Terrigenous Clastic Sediments and Its Implications for Hydrocarbon Discovery and Recovery. Special Publication of SEPM, 3: 1–8

[62] Ventra D, Cartigny M J B, Bijkerk J F, Acikalin S (2015). Supercritical-flow structures on a late carboniferous delta front: sedimentologic and paleoclimatic significance. Geology, 43(8): 731–734

[63] Wang J (2018). Fluvial Fan Architecture, Facies, and Interaction with Lake: Lessons Learned from the Sunnyside Delta Interval of the Green River Formation, Uinta basin, Utah. Dissertation for Doctoral Degree. Colorado School of Mines, 1–166

[64] Yu X H, Li S L, Tan C P, Xie J, Chen B T, Yang F (2013). The response of deltaic systems to climatic and hydrological changes in Daihai Lake rift basin. Journal of Paleogeography, 2(1): 41–55

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag GmbH Germany, part of Springer Nature