Introduction

The understanding of deep-water depositional systems has advanced significantly in recent years (Posamentier and Kolla 2003). As exploration has focused on deep-water areas, more and more multi-channel high resolution 2D and 3D seismic data were acquired in deep-water settings (e.g., Beaubouef and Friedmann 2000; Posamentier et al. 2000; Posamentier and Kolla 2003; Yuan et al. 2009). The 3D seismic data afford an excellent view of the deep-water depositional environment, in some instances with vertical resolution down to 2–3 m (Posamentier and Kolla 2003). Gravity-induced across-slope sedimentation processes, such as slides, slumps, debris flows, and turbidity currents, are important and predominant on continental slope domains (Yuan et al. 2009). Through compiling from Dott (1963), Nardin et al. (1979), Posamentier and Kolla (2003), Moscardelli et al. (2006), Moscardelli and Wood (2007) and Bull et al. (2009), the terminologies used here and their seismic reflection characteristics are as follows: (1) slide is defined as shear failure along discrete shear planes without internal deformation or rotation and high-amplitude, continuous reflections; (2) slump is defined as shear failure accompanied by rotation with complicated internal deformation, and low- and high-amplitude chaotic reflections; (3) debris is defined as shear distributed throughout the sediment mass and low-amplitude, semitransparent chaotic reflections, and its basal erosional surface is usually characterized by grooves; (4) turbidity flow is supported by fluid turbulence, and is usually characterized by parallel moderate- to high-amplitude reflections with channel structures.

Many authors have documented the repetitive cycles of mass-transport deposits (MTDs; slide, slump, and debris flows deposits) and channel systems including distributary channel complexes (DCC), leveed channel complexes (LCC), and avulsion channel complexes (ACC) in the Gulf of Mexico (e.g., Beaubouef and Friedmann 2000; Posamentier et al. 2000), Amazon Fan (Pirmez et al. 1997; Manley and Flood 1998), Trinidad and Tobago (Brami et al. 2000; Moscardelli et al. 2006), Nigeria, and Indonesia (Posamentier and Kolla 2003) and attributed them to sea-level change. This depositional pattern is important in understanding the depositional system on slope areas, and is also useful for hydrocarbon exploration.

Turbidity current channels, such as central channel (central canyon), and mass transport deposits have been documented in the Qiongdongnan Basin (Lin et al. 2001; Su et al. 2009; Yuan et al. 2009; Wang et al. 2009; Zhu et al. in press). However, the cyclic depositional pattern observed in this study represents a new aspect in deep-water sedimentation of the South China Sea. This study focuses on the characterization of the seismic facies, geometries, and internal structures of the main depositional elements on the southern slope of Qiongdongnan Basin. It encompasses the evaluation of the potential influences of sea-level fluctuations, sedimentation rate, tectonics and palaeo-seafloor morphology on sedimentation types and their depositional architecture.

Geological setting

The Qiongdongnan Basin lies offshore south Hainan Island with an arc-like shape, and trends NE–SW. The basin is located between 108°51′E and 114°41′E, 17°00′N and 18°50′N (Fig. 1). It is a 290 km long and 181 km wide basin with a total extent of 6,000 km² (Sun et al. 2010). Its western border is marked by a strike-slip fault zone continuing into the Yinggehai Basin (Xie et al. 2006; Wu et al. 2009). The study area is situated on the southern slope of the Qiongdongnan Basin in a present water depth of 1,230–1,300 m, with a seafloor topography dipping NE (Fig. 1).

Fig. 1
figure 1

a Sketch of the northern South China Sea (modified after Schimanski and Stattegger 2005; Yuan et al. 2009); b expanded map of the study area on the slope of the Qiongdongnan Basin

Full size image

The Qiongdongnan Basin is a Tertiary basin formed by lithospheric stretching during the Paleogene, followed by thermal subsidence since the Miocene (Zhou et al. 1995; Xie et al. 2006). Its sedimentary fill is composed of 2 mega-sequences spanning the rifting period (53.5–23.3 Ma) and post-rifting period (23.3 Ma to Quaternary; Wu et al. 2009), respectively. The depositional environment in the study area was initially transformed from lacustrine to marine conditions and later from neritic to bathyal, starting in Eocene times and continuing to the present (Xie et al. 2006). During the Cenozoic, two main sources provided sediment, namely the South China Block from Oligocene to Miocene and the Vietnam Uplift from Miocene to recent times (Yao et al. 2008; Yuan et al. 2009). Tectonic movements ceased in the post-rifting period (Zhou et al. 1995; Sun et al. 2003). However, a polygonal fault system is well developed in the Mid-Late Miocene succession (Sun et al. 2009, 2010). A turbidity current channel system developed in the study area in the Pleistocene (Yuan et al. 2009). The deposition rate gradually increased since the Miocene time, up to 0.7 mm/year in the Quaternary (Sun et al. 2009). The Red River strike-slip fault, which is one of the major tectonic lineaments in Southeast Asia and is the western boundary of Qiongdongnan Basin, was active as a left-lateral strike-slip fault from Eocene to Miocene; however, it changed its direction of movement and became right-lateral after 5.5 Ma (Morley 2002). Subsequently, the Qiongdongnan Basin was undergoing rapid subsidence (Zhang and Hao 1997).

Data and methods

The primary source of data for this study comprises an industrially acquired 3D reflection seismic dataset. It was acquired by China National Petroleum Corporation (Sun et al. 2010) and covers about 2,800 km2, with a CDP (common depth point) interval of 12.5 m and line spacing of 25 m. The data were loaded into work stations and interpreted with the Geoframe software of Schlumberger. The dominant frequency of the data is approximately 60 Hz for the interpreted intervals. We applied an average seismic velocity of 1,700 ms−1 to calculate the resolution in the area of interest, giving a maximum vertical resolution of ca. 7 m and horizontal resolutions of ca. 10 m, respectively.

Horizon-time structure maps were generated from identified and interpreted horizons. Additionally, dip and coherency attribute cubes were extracted from the 3D data. The coherency cube is based on comparison of waveforms across adjacent traces, and gives apparent continuity to discontinuous features, such as faults, edges, and channels (Hart 1999). Dip and coherency cubes are useful for enhancing the appearance of laterally continuous geological features that may otherwise be overlooked (Dalley et al. 1989), such as channels and grooves in the study area, and avoiding interpretation bias (Bull et al. 2009). Through utilizing a combination of time-structure maps, vertical seismic sections, cube attributes and amplitude attributes, the identification of channels, grooves, folds and thrusts were significantly improved.

Three-dimensional seismic observations and interpretation

Based on the three-dimensional seismic reflection characteristics and internal structures, the interpreted succession is subdivided into six units (Units 1–6) by mapping seven surfaces throughout the cube, namely Horizons A–F and the seafloor from base to top (Fig. 2).

Fig. 2
figure 2

Stratigraphic subdivision of the studied sedimentary slope succession. The succession could be subdivided into six seismic units (Units 1–6) bounded at base and top by prominent seismic horizons (Horizons A–F) and the seafloor, based on -reflection seismic characteristics and internal structures. For location see Fig. 1b

Full size image

Unit 1

Observation

Unit 1 marks the base of the interpreted succession, and is bounded by Horizon A and B at its base and top, respectively (Fig. 2). Horizon A is a high-amplitude non-continuous event, and there are widespread onlap and truncation along it. Unit 1 can be subdivided into two sub-units: Unit 1A and Unit 1B, according to their seismic facies and configuration. Unit 1A represents the lower part of Unit 1, which only developed in the northwestern part of the study area (Fig. 3), and it is characterized by low- to -high amplitude, non-continuous to chaotic seismic reflections. The sediments seem to have been rotated along a basal erosional surface (Horizon A, Fig. 3b). The Unit exhibit faulted areas of extension and compressional zones with folds, and thrusts (Fig. 3c). The trend of the extensional faults is nearly NE–SW (Fig. 3d). In the northwestern part of the study area, Unit 1B forms the upper part of Unit 1, and is characterized by semi-transparent, subparallel seismic reflections (Fig. 3a, b). It becomes thinner from northwest to southeast and in places it is missing (Fig. 3a, b, c).

Fig. 3
figure 3

a The un-interpreted version of (b); b representative seismic profile for Unit 1. The black square is the location of (c). For position of seismic line see Fig. 1b; cline drawing interpretation of Unit 1; d dip attribute map of Horizon A. MTDs (mass transport deposits, the red dashed line marks the boundary) and extensive faults are labeled. For location see Fig. 1b

Full size image

Interpretation

The Horizon A is interpreted as major sequence boundary (the boundary between lower Pliocene and upper Pliocene) on the basis of widespread onlap, truncation and borehole correlation (e.g., Ya 35-1-1 on Fig. 1), and its age is determined to 2.9 Ma (Lu and Zhang 1995; Gong and Li 1997; Xie et al. 2008). The complicated internal structures (folds, thrust and extensional faults) and chaotic seismic reflections indicate Unit 1A is a slump. The extensional area is interpreted as head domain, and the fold and thrust area is interpreted as toe domain (Fig. 3c). Unit 1B is interpreted as debris flow based on is semi-transparent, subparallel seismic reflections.

Unit 2

Observation

Unit 2 is bounded by Horizon B and Horizon C (Fig. 2). Horizon B is a very high-amplitude, continuous seismic reflection event. The internal configuration of Unit 2 is characterized by moderate-amplitude, continuous seismic reflections. The sediment seems to be slight onlap from north to south (Fig. 4).

Fig. 4
figure 4

a Un-interpreted version of (b); b representative seismic profile of Unit 2 and Unit 3. The black square is the location of (c). For location see Fig. 1b; cline drawing interpretation of Unit 2 and Unit 3; d dip attribute map of Horizon C. MTDs (mass transport deposits), and grooves (the purple dashed line delineates the boundary) are labeled. For location see Fig. 1b

Full size image

Interpretation

Because we could not identify any special structure (e.g., channels, grooves, folds and thrusts) in Unit 2 and along Horizon B, and its homogeneous seismic appearance (nearly constant moderate amplitude, Fig. 2) we interpreted the Unit as hemipelagic–pelagic drape. However, onlap termination in the northern part of Unit 2 may indicate the occurrence of turbidities.

Unit 3

Observation

Unit 3 is characterized by low-amplitude, semi-transparent to transparent, subparallel seismic reflection (Fig. 4a, b). It is separated by Horizon C and D from Unit 2 and 4, respectively. In the northern part of the study area its character changes: sediment thickness increases, seismic facies becomes chaotic and many folds and thrusts appear (Fig. 4). Some grooves develop along Horizon C in the attribute map (Fig. 4d).

Interpretation

The northern part of Unit 3 is interpreted as a slump. The grooves are the representative features of basal mass transport deposits. They imply debris flow processes, where the flow direction is parallel to trend of the grooves (Bull et al. 2009). Together with the semi-transparent to transparent seismic reflection pattern, we interpret the rest of Unit 3 as debris flow deposit.

Unit 4

Observation

Unit 4 is characterized by predominant high-amplitude, continuous seismic reflections. There are several leveed channels identified in the seismic profiles and dip attribute maps within Unit 4 (Fig. 5). The leveed channels are meandering, and they are 120–350 m wide and trend nearly W to E (Fig. 5b, c). The base of the leveed channels exhibits high amplitude reflection.

Fig. 5
figure 5

a Dip attribute map of Horizon D, with some identified leveed channels. For location see Fig. 1b; b un-interpreted version of (c); c a representative seismic profile shows the leveed channels within Unit 4

Full size image

Interpretation

The Unit 4 is interpreted as turbidity flow deposits, according to the high-amplitude, continuous reflections and meandering structure in the attribute map. The high amplitude reflection at the base of leveed channels indicates coarse-grained sediments at their base.

Unit 5

Observation

Unit 5 has semi transparent to transparent or chaotic seismic reflections (Fig. 6b, c). Horizon E is a low- to moderate-amplitude, continuous seismic event, with many grooves developed on it (Fig. 6a, b, c). These grooves are long (in some instances longer than 15 km) and linear in plan view. They tend to diverge from SW to NE and E (Fig. 6a). These erosional scours, in some instances more than 10 m deep and over 200 m wide, characterize the base of these semi-transparent to transparent or chaotic deposits. The thickness of Unit 5 is variable, and decreases from SW to NE; giving the Unit an external wedge-shaped geometry (Fig. 6d, e).

Fig. 6
figure 6

a Dip attribute map at the base of Unit 5 (Horizon E) illustrating long, continuous, linear grooves, which show down-slope divergence in map view. For location see Fig. 1b; b un-interpreted version of (c); c seismic profile across Unit 5. Erosional scours at the base of this unit are locally greater than 10 m deep and 200 m wide; d un-interpreted version of (e); eseismic profile along grooves, which shows that the thickness of Unit 5 decreases down-slope

Full size image

Interpretation

A distinctive attribute of debris flow deposits, regardless of whatever their geometries are, is the pattern of grooves that commonly characterizes the basal surface upon which they are deposited (Posamentier and Kolla 2003; Posamentier and Walker 2006). Therefore, the semi-transparent to transparent seismic reflection characteristics and long linear grooves indicate Unit 5 is a typical debris flow deposit.

Unit 6

Observation

Unit 6 is dominated by a moderate- to high-amplitude, parallel and continuous seismic reflection pattern. However, in places additional reflection patterns occur. The first one is a low-amplitude, V- or W-shaped pattern, whereby the flanks are usually higher than the center (Fig. 7b, c). These seismic reflection features are highly sinuous in plan view (Fig. 7a, marked LC). The incised base of this seismic pattern usually penetrates into Unit 5 and in some instances it reaches its base (Fig. 7b, c). This seismic pattern usually has a narrow base and becomes wider upwards (Fig. 7b, c). The second reflection pattern is a low-amplitude interval embedded into the high amplitude Unit 6. They form small V- or W-shaped incisions (Fig. 7a, b, c, marked DC). In the dip attribute map, the DC is a distributary at the end of LC, and many low-sinuosity depositional bodies diverge downslope (Fig. 7b, c, nearly from W to E and NE). There is third but minor pattern of low-sinuosity in plan view (Fig. 7a, marked AC), it is usually related to LC. It is normally U-shaped in the seismic image; however, its amplitude is generally higher than that of the other two patterns (Fig. 7b, c). The thickness of Unit 6 is homogeneous, except where the turbidity current channels (especially the LCC) occur. The DCC and ACC developed in a limited time interval, ceasing below the modern seafloor. However, the first pattern (LCC) continues to develop until the ocean bottom (Fig. 7d). But its sinuosity is lower, becoming linear basinward with a SW to NE trend (Fig. 7d).

Fig. 7
figure 7

a Dip attribute map at the base of Unit 6 (Horizon F) illustrating leveed channels (LC, between red dashed lines), low-sinuosity distributary-channels (between blue dashed lines), and avulsion channels (AC). For location see Fig. 1b; b the un-interpreted version of (c); c a representative seismic profile showing the reflection characteristics of leveed channels, distributary channels, and avulsion channels; d time structure map of the seafloor which shows the leveed channel well-developed at the modern seafloor. For location see Fig. 1b

Full size image

Interpretation

According to the seismic reflection profiles and plan view attribute maps, three channel complexes, LCC (the first seismic-reflection pattern), DCC (the second seismic-reflection pattern) and ACC (the third seismic-reflection pattern), are identified. The LCC has high sinuosity and a leveed structure (Fig. 7a, b, c); the DCC consist of low-sinuosity depositional bodies with a fan-like distribution (Fig. 7a). The ACC usually are considered to be caused by relatively infrequent, large-volume, high-velocity flows (e.g., Pirmez et al. 1997), and the degree of channel instability versus peak-flow volumes decides where an avulsion would occur (Kolla 2007). The avulsion node usually occurs on the outside of a meander bend (Fig. 7a, marked AC1). This location favors the probability of an avulsion occurring due to flow stripping, but may also represent the zone of highest channel instability (Armitage 2009). The dominant channel infill is fine-grained sediments deduced from the low-amplitude seismic facies (Fig. 7b, c), and only the ACC may be infilled by coarse-grained material (high-amplitude seismic reflections).

Discussions

Vertical stacking patterns

Three primary depositional systems—mass transport deposits (MTDs), turbidity flow deposits, and deep-water drape—are identified from the youngest strata on the southern slope of Qiongdongnan Basin, according to the three-dimensional reflection seismic characteristics and internal structures. Two types of MTDs, slump and debris, are developed in the study area, based on the classification by Nardin et al. (1979) and Moscardelli et al. (2006). The slump occurred in two units, Unit 1A and the northern part of Unit 3. Unit 1A is a small-scale but well developed slump. However, the scale of the slump within Unit 3 is undefined because it extends out of the data, and only the toe domain could be identified (Fig. 4).

Three units (Unit 1B, Unit 3 and Unit 5) are composed of debris-flow deposits. However, the degree of groove development is variable, from well-developed in Unit 5 to poorly-developed in Unit 3. This variability can be attributed either to differences in sediment caliber and the size of indurated blocks within the respective flows or differences in the degree of induration of the substrate across which the flows travel (Posamentier and Kolla 2003). The common divergence of grooves down-system and the thickness variability from SWW to NEE suggest divergent flow vectors in an unconfined setting and a flow direction from SWW to NEE.

Unit 4 and Unit 6 are predominated by a turbidity current channel system, which includes leveed channel complex (LCC), distributary channel complex (DCC) and avulsion channel complex (ACC). The LCC developed both in Unit 4 and Unit 6, and they are usually of high sinuosity in the attribute maps (Figs. 5a, 7a). The DCC only developed in Unit 6, and it is distinctive with leveed channels in plan view (Fig. 7a). The DCC has a fan-shaped map pattern at the distal end of a feeder channel complex (Fig. 7a). The ACC developed only in Unit 6, and generally with low sinuosity in plan view (Fig. 7a) and steep boundaries in seismic profile (Fig. 7b, c). The avulsion channel usually originated from LC (Fig. 7a, marked AC1 and AC3) and DC (Fig. 7a, marked AC4).

There are no special internal structures in Unit 2, and the seismic-reflection pattern is similar to those of highstand condensed sections (Posamentier and Kolla 2003). Therefore, we identify Unit 2 as hemipelagic–pelagic drape equivalent to a highstand condensed section.

A remarkable characteristic of the stratigraphy and depositional history of the southern slope of Qiongdongnan Basin is the cyclic nature and the repeated stacking patterns that occur. We could identify three cycles in the study area. Mass transport deposits (slump or debris-flow deposits) are found at the base of each cycle and are interpreted to be usually associated with a sequence boundary. These mass transport deposits are commonly overlain by interpreted turbidite successions or a hemipelagic–pelagic drape. For Cycle I, the slump deposits lie at the base and are overlain by debris-flow deposits within Unit 1, finally covered by a deep-water drape (Unit 2). In Cycle II, the slump laterally grade into debris-flow deposits within Unit 3, and then they are overlain by turbidites of Unit 4, which is mainly composed of leveed channel complexes (LCC). In Cycle III, the turbidites (including LCC, DCC, and ACC) of Unit 6 directly overlie the debris-flow deposits of Unit 5 (Fig. 8).

Fig. 8
figure 8

The vertical depositional stacking pattern in the study area. For details see text

Full size image

Possible controlling factors

Many factors can control deep-water depositional style and associated features, such as sea-level fluctuation, sediment supply, seafloor topography, tectonics, bottom-current activity, sand-to-mud ratio, steep slope, etc. (Pirmez et al. 1997; Badalini et al. 2000; Beaubouef and Friedmann 2000; Brami et al. 2000; Posamentier and Kolla 2003; Wynn et al. 2007; Armitage et al. 2009). However, the vertical stacking pattern in the study area is slightly different from the characteristics concluded by Posamentier and Kolla (2003). Turbidity flow deposits often develop in highstand system tracts, similar to what Carvajal and Steel (2006) documented. Four factors, as follows, may control this vertical stacking pattern.

(1) Sea-level fluctuation. Sea-level change plays an important role in controlling deep-water depositional style (Pirmez et al. 1997; Hernández-Molina et al. 2002; Posamentier and Kolla 2003). Sea-level fluctuations were frequent in Quaternary time (Haq et al. 1987), and the largest sea-level drop was about 150 m in this period (Aharon and Chappell 1986). Horizon A marks the boundary between the upper Pliocene and the lower Pliocene, according to the stratigraphy derived from wells in the northern Qiongdongnan Basin (sequence boundary S27 in Gong and Li 1997 and Xie et al. 2008; Hao et al. 2000), and it is comparable to the sequence boundary at 2.9 Ma, documented by Haq et al. (1987). Horizon C separates the Pliocene from the Quaternary (1.6 Ma, sequence boundary S20 in Lu and Zhang 1995 and Gong and Li 1997; Hao et al. 2000). The turbidits in the study area determined by Yuan et al. (2009) are equivalent to Cycle III. They proposed that the turbidites formed during the Mid-Pleistocene Revolution (MPR, 0.9–0.92 Ma) which is very close to the last significant sea-level drop (at 0.8 Ma) documented by Haq et al. (1987). Therefore, the three depositional cycles are apparently associated with the last three cycles of main sea-level fluctuation since 2.9 Ma proposed by Haq et al. (1987). Because of insufficient well data, we cannot go further. However, based on the previous sequence stratigraphy studies (Lu and Zhang 1995; Gong and Li 1997; Xie et al. 2008), the vertical depositional patterns and comparison with other similar case studies (e.g., Gulf of Mexico, Amazon Fan, and Trinidad and Tobago), the influence of sea-level fluctuation on cyclic deposition in the study area is believable.

(2) Tectonic reactivation and differential subsidence. The study area has been in its post-rifting stage since Miocene time (15.5 Ma; Xie et al. 2006), and the tectonic movements ceased in the study area after that (Zhou et al. 1995; Xie et al. 2006). However, the Red River strike-slip fault in the western boundary of the study area reactivated during the Quaternary (Schimanski and Stattegger 2005). The dextral slip rate is estimated to be 1-5 mm/yr, according to measurements of the onshore Red River strike-slip fault (Allen et al. 1984). The reactivation of Red River strike-slip fault caused frequent seismicity which has been documented in China (Zhang et al. 1993), Vietnam (Tsai et al. 1999) and, the Qiongdongnan and Yinggehai basins (Wei and Chung 1995). Tectonic reactivation and related seismicity can cause slope instabilities by seismic shaking, changing pore pressure and creating zones of weakness.

The Qiongdongnan Basin has experienced rapid subsidence since 5.5 Ma (Zhang and Hao 1997). We can also deduce this from the thick Quaternary sediments in the central subsidence belt which is just north of the study area (Xie et al. 2006). Moreover, the northern Vietnam shelf is uplifting in Holocene time, caused by the Red River strike-slip fault movement (Schimanski and Stattegger 2005). Therefore, the rapid relative subsidence in the central subsidence belt which is just north of the study area, and the relative uplift of the Vietnam shelf which is just southwest of the study area will steepen the slope. The steep slope and the reactivated strike-slip fault may have triggered slope instability and slope failure, leading to the generation of slumps, debris flows, and turbidity current flows.

(3) A high sedimentation rate. The build-up of excess pore pressure in fine-grained sediments as a result of rapid sedimentation is a potentially important factor in controlling the instability of continental slopes (Zhu et al. in press). The sedimentation rate increases quickly during the last 3 m.y.; it is 400–800 m/m.y. in the northern Qiongdongnan Basin and Yinggehai Basin (Zhu et al. in press), and 50–100 cm/ky (500–1,000 cm/ky for short periods during the early Holocene) on the central Vietnam shelf (Schimanski and Stattegger 2005) in the western South China Sea. The collision of India with Eurasia and the Himalayas would have caused great erosion and abundant sediment were carried into the South China Sea by small mountain rivers and larger rivers (e.g., Pearl River and Red River) whether sea level is high or low (e.g., Schimanski and Stattegger 2005; Xie et al. 2008; Zhu et al. in press). In addition, this high sedimentation rate is also linked to a climate change. Zhang et al. (2001) suggested that disequilibrium states of frequent and abrupt changes in temperature, precipitation and vegetation are effective in causing erosion, but only after 4–3 Ma, as global climate deteriorated. We infer that this rapid increase in sedimentation rate oversteepens the slope deposits and induces large-scale sediment failures on the upper slope, which, in turn, generated large gravity-controlled mass movements. So to some degree, the high sedimentation rate may be more important than the sea-level fluctuation in controlling the special depositional styles such as the turbidity flow deposits developing in the highstand system tract.

(4) Paleo-seafloor morphology. The paleo-seafloor morphology plays a primary role in controlling the general distribution and facies architecture of the overlying sediments (Armitage et al. 2009). It would have influenced the flow direction of the downslope deposits. The slope dips from SW to NE, so most of the debris flows and turbidity flows (e.g., these in Units 1, 3, 5 and 6) trend from SW to NE (Figs. 5a, 6a, 7a).

Implications

The continental slope is an important and active depositional area. It plays an important role in understanding source-to-sink processes (Sømme et al. 2009), and now many studies, such as those on MTDs, channel systems, etc., are focused on this area (e.g., Manley and Flood 1998; Brami et al. 2000; Moscardelli et al. 2006; Sømme et al. 2009). This study discovered new important depositional cycles using high-resolution 3D seismic data. These depositional cycles should be widespread phenomenon in deep-water margin settings worldwide, and may also be prevalent in ancient slope successions. Therefore, our study provides a valuable contribution to the understanding of depositional processes on slope areas.

Through other studies in the Campos Basin, offshore West Africa, and Gulf of Mexico, researchers (e.g., Pettingill and Weimer 2002; Dailly et al. 2002) found that channels in deep-water areas could be excellent reservoirs. MTDs are prone to being cap rocks (Moscardelli et al. 2006; Armitage et al. 2009), because of their mud-dominated sediments and homogeneous physical properties. Hence, the observed cyclic deposits may be an excellent group for deep-water hydrocarbon accumulation. A few turbidite channels in the study area (e.g., these in Unit 4) are infilled by coarse-grained sediments and fall into the style referred above. However, the low amplitude facies within most of the channels, especially those in Unit 6, indicates that the infill is fine-grained (Fig. 7b, c). Therefore, these channel deposits in our study area are not comparable to those described by many researchers (e.g., Pettingill and Weimer 2002; Dailly et al. 2002; Posamentier and Kolla 2003; Armitage et al. 2009). This implies that in general not all channel deposits serve as good reservoirs. It is indispensable to have a good understanding of the depositional facies and stacking pattern, respectively, to gain a better-informed reservoir model and especially to reduce exploration risk. Moreover, the Qiongdongnan Basin is a petroliferous basin, and the mass transport deposits in the shallow surface must be assessed for pipeline design and drilling plans.

Conclusion

Three main depositional elements, namely MTDs (including slump and debris flow deposits), turbidity flow deposits (including DCC, LCC and ACC), and deep-water drapes are identified embedded in six units, according to 3D seismic data (both seismic profiles and attribute maps). The slumps usually reveal chaotic seismic reflections and are internally folded and thrusted. The debris flow deposits are characterized by transparent to semi-transparent up to chaotic seismic reflection pattern and pronounced grooves at their bases. The unique grooves indicate the flow direction and are commonly divergent down-system. The channel system deposits normally have parallel seismic reflections, but within the channels they are usually transparent and different from the host successions. The DCC, LCC and ACC are fan-like, high-sinuosity, and low-sinuosity, respectively, in attribute maps. The three main depositional elements are cyclic in vertical relationship. Three cycles are defined in the study area. Each cycle contain two units, and usually begins with MTDs (slump and/or debris) at the base, and then followed up by turbidity flow deposits or a deep-water drape. These vertical stacking patterns are mainly controlled by sea-level fluctuation and a high sedimentation rate. Our results show that the three cycles may be comparable to the last three sea-level fluctuations since 2.9 Ma recognized by Haq et al. (1987). Moreover, other factors, such as tectonics, differential subsidence, and paleo-seafloor morphology, etc., maybe also contribute to their formation. We assume that this study provides new insight into the understanding of deep-water depositional systems, and may help by the estimation of exploration and engineering risk.