Geophysical signatures associated with fluid flow and gas hydrate occurrence in a tectonically quiescent sequence, Qiongdongnan Basin, South China Sea


Corresponding author: X. Wang, Key laboratory of Marine Geology and Environment, Chinese Academy of Sciences, Qingdao, China.
Email: Tel: +86 532 82898543. Fax: +86 532 82898544.


Interpretation of high-resolution two-dimensional (2D) and three-dimensional (3D) seismic data collected in the Qiongdongnan Basin, South China Sea reveals the presence of polygonal faults, pockmarks, gas chimneys and slope failure in strata of Pliocene and younger age. The gas chimneys are characterized by low-amplitude reflections, acoustic turbidity and low P-wave velocity indicating fluid expulsion pathways. Coherence time slices show that the polygonal faults are restricted to sediments with moderate-amplitude, continuous reflections. Gas hydrates are identified in seismic data by the presence of bottom simulating reflectors (BSRs), which have high amplitude, reverse polarity and are subparallel to seafloor. Mud diapirism and mounded structures have variable geometry and a great diversity regarding the origin of the fluid and the parent beds. The gas chimneys, mud diapirism, polygonal faults and a seismic facies-change facilitate the upward migration of thermogenic fluids from underlying sediments. Fluids can be temporarily trapped below the gas hydrate stability zone, but fluid advection may cause gas hydrate dissociation and affect the thickness of gas hydrate zone. The fluid accumulation leads to the generation of excess pore fluids that release along faults, forming pockmarks and mud volcanoes on the seafloor. These features are indicators of fluid flow in a tectonically-quiescent sequence, Qiongdongnan Basin.

Geofluids (2010) 10, 351–368


Fluid leakage in marine sediments is a widespread phenomenon in many basins (Shaw et al. 1997; Taylor et al. 2000; Berndt 2005; Cartwright 2007; Gay et al. 2007). Fluids can migrate from the trap to the surface by several different leakage mechanisms in consolidated sediments; for example, by fracture flow (Roberts & Nunn 1995), Darcy flow and diffusion (Kroos & Leythaeuser 1996). Structural surfaces along salt diapirs, faults, pockmarks, gas chimneys and polygonal faults can create pathways for fluid migration (Meldahl et al. 2001; Deptuck et al. 2003; Xie et al. 2003; Hansen et al. 2005; Mayall et al. 2006; Cartwright et al. 2007; Gay et al. 2007; Løseth et al. 2009). These associations with structural features indicate that specific tectonic environments can be responsible for focused fluid flow and fluid escape at the seafloor (Abrams 1992; Orange et al. 1999).

Geologic indicators of fluid leakage include pockmarks and gas chimneys. Pockmarks generally appear in unconsolidated fine-grained sediments as columnar, cone-shaped circular or elliptical depressions ranging from a few meters to 300 m or more in diameter and from 1 to 80 m in depth (Solheim & Elverhoi 1993; Baraza et al. 1999). Pockmarks are known to occur on the continental slope accompanying gas hydrate and in association with slides and slumps along the mid-Norwegian continental margin (Bünz et al. 2003; Hustoft et al. 2007). Gas chimneys are vertical zones of fluid flux showing acoustic turbidity or complete wipe-out in seismic data. The shape of gas chimneys can vary from diffuse shadows, funnels, pipes to distinct cigars (Løseth et al. 2009). Because gas chimneys are associated with low-velocity anomalies, velocity sag (pull down) are often observed underneath the chimneys. The anomalies are interpreted to be associated with the upward movement of free gas or fluid existing above either tectonically induced structures or high pressure reservoirs. When gas chimneys extend to the seafloor, mud volcanoes, pockmarks and craters (Ginsburg 1998; Ligtenberg 2003) or mound-like vents (Dai et al. 2008) are formed on the seafloor.

Gas hydrates often occur in areas characterized by fluid flow features. In the Lower Congo Basin, structural surfaces along polygonal faults, salt diapirs, faults, and chimneys create pathways for fluid migration both laterally and vertically, which can become trapped by the impermeable gas hydrate stability zone leading to the occurrence of a bottom seismic reflector (BSR; Gay et al. 2004, 2006a,b; Gay & Berndt 2007). Numerous examples of layer-bound polygonal fault systems have been described worldwide from 3D seismic data (Cartwright 1994; Cartwright & Lonergan 1997; Cartwright & Dewhurst 1998; Lonergan et al. 1998). Because of the low permeability and strong capillary forces of fine-grained sediments (Clennell et al. 1995; Ruppel 1997), vertical migration of fluids of Pliocene-Recent fine-grained sediments cannot occur at a rate sufficient to explain the observed seafloor seeping structures in the shallow sediments in the Norwegian margin (Gay & Berndt 2007). In particular, polygonal fault systems developed in the shallow sequences are suspected to lead to pore fluid escape and allow deeper fluids to migrate through the fine-grained cover (Cartwright & Lonergan 1996; Hansen et al. 2005; Gay & Berndt 2007). Mud diapirs (volcanoes) often release extensive fluid that is associated with gas hydrate occurrences in the Blake Ridge (Dillon et al. 1983; Taylor et al. 2000; Hornbach et al. 2008), in the Gulf of Mexico (McManus & Hanor 1993; Dai et al. 2008; Hutchinson et al. 2008), in the Southern Okinawa Trough (Xu et al. 2009) and on the continental margin of India (Satyavani et al. 2005; Calvès et al. 2008). Surface expressions and geometry of mud volcanism are highly variable and depend primarily on the origin of the fluid and solid phase (parent beds) (Kopf 2002; Calvès et al. 2010) and the fluid flux has a more profound effect on mud volcano formation (Brown 1990). Milkov (2000) discussed the possible significance of gas hydrate dissociation in mud volcano processes, while gas hydrate may form temporarily when the methane passes through the gas hydrate stability zone.

Understanding fluid flow processes in the Qiongdongnan Basin (QDNB) first requires understanding the tectonic evolution and depositional environment. The QDNB characterized by a major change in tectonic environment at 5.5 Ma. Before 5.5 Ma, during the Miocene, seafloor spreading occurred in the South China Sea (SCS) and extensive normal faults were formed (Clift & Sun 2006). These older normal faults can act as pathways for the upward migration of deep fluids through the sedimentary column. The QDNB does not appear to have been significantly tectonically active since 5.5 Ma (Clift & Sun 2006), so almost all normal faults do not extend to the seafloor. Fluid migration from older to the younger sediments and the gas hydrate stability zone is difficult to understand based on the old seismic data for the limitations of low-resolution two-dimensional (2D) seismic data. Gas hydrate and its relationship to gas chimney and fluid potential have been studied using high-resolution 2D seismic data (Dong et al. 2008; Wang et al. 2008). The role of polygonal faults in the migration of fluid flow and the formation of gas hydrate and gas and oil has been identified by three-dimensional (3D) seismic data shot in recent years (Wang et al. 2010; Wu et al. 2008). Gas fields, for example the Y13-1, Y13-4 and Y13-6 were discovered in the shallower water depth and the gases were composed predominantly of CH4 (85–90%), with relatively low CO2 and N2 contents (Huang et al. 2003; Zhu et al. 2009). The sedimentation rates is high (up to 1.2 mm/year) and the geothermal gradients are high (39–41 C/km) in this basin. Rapid sediment loading and associated undercompaction, led to overpressure development throughout the QND basin (Zhu et al. 2009). Regional geological research indicates that the deep-water region shares the same hydrocarbon source kitchen as the shallow-water area, thus the hydrocarbon potential of the deep-water zone had to be abundant (Zhu et al. 2009). The purpose of this study is (1) to identify the flow pathways associated with shallow and deep fluids in the context of sedimentary formation; (2) to describe different structural controls on fluid flow in a lateral and vertical sense; and (3) to discuss fluid flow and its roles in the occurrence of gas hydrate in the QDNB.

Geological setting and related work

The north slope of the SCS is a passive continental margin formed prior to, or during, the middle Oligocene-marly Miocene opening of the so-called central basin (32–17 Ma; Briais et al. 1993; Clift et al. 2002). The structure of the northern margin has been extensively studied in the course of oil exploration and during geophysical investigations of the amount of crustal extension during the opening of this basin (e.g., Clift & Lin 2001; Clift et al. 2002; Clift & Sun 2006). Since 1999, high-resolution multichannel seismic surveys have been carried out for gas hydrate resource studies in the northern SCS by the Guangzhou Marine Geological Survey (GMGS). Since 2001, these studies have been expanded to include geological, geophysical, and geochemical investigations for gas hydrate in the northern region of the SCS. The migration of natural hydrocarbon gas, biogenic gases, thermogenic gases or oil, interstitial water, or a combination of these has been studied in several basins in the SCS (Chen et al. 2006; He et al. 2008), especially in the context of fluid flow in the Yinggehai Basin (Xie et al. 1999, 2001, 2003). Evidence of gas hydrate occurrence in the northern SCS was reported by Guo et al. 2004; Wu et al. 2005, 2007; Liu et al. 2006; Wang et al. 2006; and Yang et al. 2008;. More recently, in 2007, gas hydrate has been recovered in the Shenhu area on the north slope of the SCS (Zhang et al. 2007). BSRs have been identified in the Xisha Trough (Wu et al. 2005), in the Pearl River Mouth basin (Guo et al. 2004) and in the Taixinan Basin of the northern SCS (Wang et al. 2006).

Qiongdongnan Basin is a Cenozoic passive continental margin basin that lies at the junction of the Eurasian, Pacific and Indo-Australian plates. The study area is in the southern part of the QDNB (Fig. 1), southeast and east of the Yinggehai Basin and the Vietnam Uplift, south of Hainan Island, and west of the Xisha Uplift. The QDNB is an overpressured basin formed by 10 sags (He et al. 2008). Four of the sags, the Songnan, Ledong, Yanan and Baodao sags are rich in hydrocarbons, particularly the Ledong sag. The structural evolution of the basin can be divided into two stages: an Eocene-Oligocene rift and a Neogene – Quaternary post-rift thermal subsidence (Ru & Pigott 1986; Zhu et al. 2009). In response to the structural evolution, two sets of petroleum systems (upper and lower) containing both oil and gas have been identified, the dominant environment for sedimentary deposition changed from paleolakes in the early stage, through a bay, to an open-marine setting in the late stage (He et al. 2008; Zhu et al. 2009). These two systems are separated by the cap rock of the Meishan Formation (15.5–10.5 Ma: Wu et al. 2008; Hao et al. 2005). The lower petroleum system occurs in syn-rift sequences within half-grabens which developed under extensional tectonics. The 10 sags in QDNB (D1–D10) are underlain by half-grabens of the lower petroleum system. The upper petroleum system is formed during the post-rifting stage and the Miocene Sanya, Meishan, and Huangliu Formations are characterized by neritic-bathyal calcareous mudstones and neritic shelf sandstones (Zhu et al. 2009). A dense array of steeply dipping faults (60–90°) has been imaged in high-resolution profiles in the Pliocene and Quaternary strata of the Ledong sag. The interval appears as chaotic and blank seismic facies interpreted to be a result of hydrofracturing (Xie et al. 1999).

Figure 1.

 (A) Area of exploration in the northern part of the South China Sea. BSRs, deformational zone and chaotic seismic facies zone distributions and sags in the Qiongdongnan Basin (QDNB), (D1) Ledong Sag. (D2) Lingshui Sag. (D3) Songnan Sag. (D4) Baodao Sag. (D5) Songdong Sag. (D6) Songxi Sag. (D7) Yabei Sag. (D8) Yanan Sag. (D9) Beijiao Sag. (D10) Huaguang Sag. (B) Location map of the study area, showing gas chimneys identified from 2D seismic data [pink triangles for present data and blue triangles for earlier data (Wang et al. 2008; Sun et al. 2009)] and a turbiditic paleochannel by Yuan et al. (2009).

Eight sequences have been recognized based on drilling results and eight seismic reflectors have been confirmed from synthetic seismograms. The five reflection horizons used in this study (T2, T3, T5, T6 and Tg) have ages of 1.9, 5.5, 15.5, 23.3, and 53.5 Ma based on drilling results in Ying-Qiong basin, respectively (Chen et al. 1998; Wu et al. 2009a,b; Yuan et al. 2009a,b). Reflector T6 is a regional unconformity that divides the entire stratigraphic column into syn-rift and post-rift sequences (Chen et al. 1998; Wang et al. 1998).

Data and methods

The multichannel 2D seismic lines 0101, 0104, EE′ and FF′ used for this study were shot by the GMGS in 2004 and in 2006, respectively. Seismic lines 0102, 0103, 0105, AA′, BB′, CC′ DD′ and 3D seismic data were shot by the China National Petroleum Corporation (CNPC) in 2005. The seismic lines were shot using a 3000-m long streamer with 240 channels (trace interval 12.5 m) and a coherent air gun source with a total volume of 8 × 20 inch3 shooting every 25 m. The sampling interval was 1 ms. The streamer depth was 8 m and the source depth 5 m. The 3D seismic data cover approximately 40 km × 35 km, which were processed by CNPC. The bin spacing in the in-line and cross-line directions were 12.5 and 25 m, respectively. They have a 2-ms vertical sampling rate. The seismic volumes were processed to zero phases and the average frequency content of the full-stack seismic data was 40–45 Hz.

The data are of high quality, after processing by the Institute of Processing, GMGS. The processing sequences included prestack filtering, true amplitude recovery, noise attenuation, deconvolution, velocity analysis (first), residual statics, velocity analysis (second), multiple attenuation, dip moveout (DMO) velocity analysis, DMO stack and noise attenuation. P-wave velocities were estimated from stacking velocities using the VelMod package (Jason Geosciences Workbench 2003; Wang et al. 2006).

Well Ya35-1-2 was drilled in 1995 in water depths of about 200 m within the Ledong sag, along seismic line 0101. Total depth driller is 5281.9 m. The Ledong Formation has a total thickness about 2491 m. It consists of thick blue mudstone and the interbedded yellow siltstone. The Yinggehai and Huangliu Formations consist of blue mudstone and yellow sandstone. The Meishan Formation mainly consists of Yellow fine sandstone and interbed between blue mudstone and yellow siltstone (Fig. 2). Well logs such as spontaneous potential, acoustic logging, resistivity log, density log, gamma ray log, porosity log were performed for oil and gas.

Figure 2.

 Stratigraphy from an industrial well Ya35-1-2 in the northern Qiongdongnan Basin. The well suggests the seismic reflector sequences (T), lithology (interbedded refers to the interbed of mudstone and sandstone) and general age of the sediment.


Seismic expression of fluid migration

Seismic line 0101(Figs. 3, 4) shows a high-amplitude reflection at about 250 ms two-way-travel time (TWT) sub-parallel to, and below, the seafloor. It has reversed polarity compared with the seafloor reflection (Fig. 3c). A significant reduction in amplitude is observed above it (Fig. 4a). P-wave velocities above this high-amplitude reflection are 1950 m/s and decrease immediately below it to about 1550 m/s (Fig. 4b). On the basis of its shape, polarity, velocities, and association with blanking, this reflection is interpreted as a BSR. The BSR occurs primarily in the northern portions of the study area and is patchy and discontinuous.

Figure 3.

 (A) Seismic profile 0101 showing an example of reserved polarity of the BSR (C) and slope failure with chaotic seismic facies and the main deformational features. (B) Interpreted profile, showing evidence of gas chimneys and pockmarks. The profile through from the listric fault at the head. (C) BSR shows reversed polarity comparing to the seafloor. (D) The front of the failure (sub-vertical frontal fault), which is further detailed in Fig. 11. In addition, shown is the location of well Ya35-1-2, which is further detailed in Fig. 2.

Figure 4.

 (A) Part of seismic profile 0101 showing gas chimneys below the BSR. (B) The P-wave velocity obtained from the stack velocity by Jason software (2003). See Fig. 3a for the profile location.

The low velocity in the underlying sediments is consistent with the presence of free gas below the BSR causing a decrease in the acoustic velocity of sediments. Free gas can also cause acoustic anomalies or zones of acoustic turbidity such as high-amplitude, disrupted, incoherent reflections, or pull-down effects. These zones of acoustic turbidity are interpreted as gas chimneys. Several new gas chimneys have been identified based on the seismic profiles in the study area (Fig. 1b) (Wang et al. 2008; Sun et al. 2009). These acoustic anomalies are caused by the migration of fluid from deeper reservoirs (Meishan layer, 15.5–10.5 Ma; Figs. 3, 4). The gas chimneys are generally present underneath the BSRs.

A polygonal fault system was shown from the 3D seismic survey using coherence slices along time horizons (Fig. 5a). The polygonal faults are characterized by numerous, closely spaced normal faults, which have small offsets (5–30 m). The faults affect a zone of sediments about 700 m thick. Three distinct types of mounded structures have been identified as resulting from focused fluid/gas migration on the Gjallar Ridge, offshore mid-Norway (Hansen et al. 2005). The mounded structures caused the polygonal faults geometry to change from a polygonal pattern to radial pattern, which suggest that the growth of the mounded structures was coeval with or preceded the development of the polygonal faults. Reflections below the mounded structures can be observed to bend upwards and are characterized by discontinuous, low-amplitude reflections resembling a gas chimneys or mud diapir (Fig 5b). The overburden contains a stack of high-amplitude events and discontinuous reflections. The low amplitude is present at TWT 2980. The deformation style is related to fluid, and possibly sediment, injection. The 2D seismic lines 0102 and 0103 show evidence of the polygonal faults identified in 3D seismic data (Figs. 6c, 7c). Seismic stratigraphic ties to Well Ya35-1-2 show that the polygonal faults occur primarily within a Pliocene and Pleistocene mudstone layer (5.5–1.9 Ma) that is about 800 m thick. Some of the polygonal faults may also extend into the deeper Miocene layer (Fig. 7c). Figure 2 shows that three sets of sandy layers with a total thickness greater than 200 m occur in the middle Miocene and Pliocene units. The polygonal faults and gas chimneys might have acted as pathways for upward fluid migration and the sandstones have been the preferred conduits for lateral fluid migration from the deeper to the shallower sediments.

Figure 5.

 (A) Seismic expression of polygonal fault systems through coherence time slice at 2980 ms TWT in 3D seismic survey; mound structures with low coherence zones in elliptical or circular shape are identified in the polygonal faults developed level. (B) Detailed expression of the mound structure in seismic profile. Zero indicates the coherence between adjacent trace is low and one is high.

Figure 6.

 (A) Seismic profile 0102, showing paleochannel, faults, gas chimneys, detailed section of the seismic profile (insert), showing the different seismic facies nearby the bright spot. (B) Interpreted profiles, showing evidence of gas chimneys below BSR and the change of seismic facies between the northern QDNB and the southern QNDB located over the Pliocene turbiditic channels (Yuan et al. 2009a.b). The arrows indicate the migration of fluids. (C) Seismic profile 0105, showing the character of turbiditic channel in the Xisha Trough. The channel incised into the polygonal faults developed formation. Bright spots above the polygonal fault and the channel indicate the fluid upward migration from the deeper levels.

Figure 7.

 (A) Seismic profile 0103, showing the slope failure and volcano. (B) Interpreted profile, showing evidence of numerous polygonal faults (Detailed in Fig. 5). (C) Detailed section of seismic profile 0103, showing the base of the failure and the deformational features zone above the failure. The seismic profile across the failure and faults shows the headwall of the failure. (D) Detailed section of seismic profile 0103, showing a bright spot and a pockmark. Polygonal faults may provide a conduit for fluids upward migration.

Mound structures have been identified with dome-shaped features in QDNB and are interpreted to occur as a result of vertical fluid expulsion (Figs. 8, 9). The structure appears to penetrate the succession of Yinggehai Formation and does not extend to the Ledong Formation (Fig. 8). A dome-like structure was formed between the two levels. Bright spots occur near, or adjacent to, the crest of a mound structure characterized by discontinuous, low-amplitude reflections resembling a mud diapir. The overburden contains high-amplitude event and discontinuous reflections resembling gas chimney (Fig. 8). Another group of mound structures are observed as relatively small domes and are restricted to a Pleistocene interval with localized amplitude anomalies. Enhanced reflection is more pronounced and parallels the overlying formation, which has high-amplitude, high frequency and continuous reflections (Fig. 9).

Figure 8.

 Seismic profile EE’ showing a mud diapir and high-amplitude anomalies at the flank and above the diapir. Bright spots and pull-down seismic reflection at the top and flank of diapir indicate the presence of fluids.

Figure 9.

 Seismic profile FF’ showing the mounded structure between the continuous reflection and the chaotic reflection; enhanced reflection below the continuous reflection indicates the gas/fluids were trapped.

Seafloor pockmark and potential gas hydrates

In the study area, a large pockmark occurs on the high-resolution 2D seismic line 0104 Fig. 10). The pockmark is about 860 m in diameter and located in about 1376 m water depth. A small mud volcano is shown in the pockmark (Fig. 10a, insert). High seismic amplitude anomalies underlying the sediment of the pockmark indicate active fluid flow below it. The turbidite channel identified by Yuan et al. (2009a,b) in the adjacent zone is characterized by a low-amplitude reflection and has a measurable length, while the pockmark is characterized by discontinuous high-amplitude reflection and only is identified in seismic line 0104. A BSR is interpreted at 250 ms TWT below the seafloor, and is parallel to the seafloor. An increase in P-wave velocity and frequency is also observed above the BSR (Fig. 10). Near the location of the pockmark, several zones of acoustic turbidity occur at depths greater than the BSR. These are best seen in instantaneous frequency plots of the seismic data, in which high and low frequencies are represented by yellows and blues respectively (Fig. 10c). In the acoustic turbidity zone, the frequency and the P-wave velocity is lower than the identical strata (Fig. 10c,d). Moreover, reflection planes of the primary layer are preserved within the gas chimneys. This can be contrasted to mud diapirs, where the primary layers are remobilized. We interpret the vertical zone of acoustic turbidity as gas chimney because we observed the bright anomalies in the seismic amplitude at the flanks and above gas chimneys were observed (Fig. 10a). The pockmark, BSRs, and gas chimneys are indicators of fluid flow from deeper formations. The presence of pockmarks reveals that some fluids have been able to escape to the seafloor (Fig. 10b). The presence of BSR shows that other fluids are accumulated in the gas hydrate stability zone to form gas hydrate. Gas hydrate modifies the elasticity of the sediments and decreases the permeability of the sediments. Free gas will be trapped below it.

Figure 10.

 (A) Seismic profile 0104, showing the acoustic anomalies below a large pockmark (insert) and a BSR; the broken line shows the depth of gas hydrate stability zone (GHSZ). (B) Interpreted profile showing gas chimneys, which can act as pathways for fluid flow migration from deeper sediments. (C) The instantaneous frequency of seismic profile 0104. Low-frequency is present in gas chimney zone. (D) P-wave velocity of seismic profile 0104 in rectangle; high velocity shows the presence of gas hydrate above the BSR and low velocity in gas chimney zone.

Submarine slope failure

The seismic data show evidence of an unreported slope failure in the central and southeastern part of the study area. The head scar is easily identified from seismic lines AA′, BB′ and DD′ (Fig. 11). Near the intersection of seismic lines AA′ and BB′, the listric fault at the head of the failure can be traced down to an area of chaotic seismic character; this listric fault is called the East headwall (Fig. 11b,c). Seismic line 0101 shows the extent of the slide on a dip line across the basin (‘chaotic seismic facies’ labeled on Fig. 3). Seismic lines 0103 and DD′ show the sidewalls of the failure (Figs. 7, 11d). The internal structure of the slide shows low-amplitude, discontinuous reflections near its base, although there are also some somewhat more coherent, high-amplitude reflections near its top. The slope failure formed a deformational structure zone with a thickness of 400 m. Based on the 3D and 2D seismic data, we have interpreted the chaotic seismic facies (green line) and the deformational features (purple line) within the failure (Fig. 1b). This failure is named the Qiongdongnan slide.

Figure 11.

 A close-up view of the headwall and sidewall of the failure from several seismic profiles and the frontal fault of failure, showing sharper imaging of headwall and sidewall of the failure. TF marks top of failure, whereas BF marks base of failure.

Although the BSR is not visible in the slide, it can be identified in the unaffected sediments near the slope failure. Seismic lines AA′ and DD′ show amplitude anomalies such as acoustic blanking (wipe out) at TWT 300 and 250 ms below the seafloor respectively (Fig. 11b,d). The relationship between the slide and the presence of gas hydrate is not clear. Gas hydrate in the sediments prior to the slide has probably been removed or dissociated by the changes in pressure and temperature associated with the mass movement.

A large amount of volcanic activity occurred southeast of the Xisha Uplift (Fig. 7). The volcanism was most active during the Pliocene-Recent within the oceanic basin along previous faults (Yan et al. 2006). The volcanic activity could have caused steeper slope angles and instabilities, which in turn could trigger the slope failure.

Changing seismic facies

The seismic facies of the Lower Pliocene and younger deposits change from south to north. Seismic lines 0101 (Fig. 3) and 0102 (Fig. 6) show that the seismic amplitude, reflection continuity, and frequency are quite different between the south and north at the level of horizon T3 (base Pliocene). The seismic facies near Hainan Island in the north are characterized by medium-amplitude, medium-frequency and semi-continuous reflections, while the seismic facies near Xisha and the Vietnam Uplifts to the south are characterized by high-amplitude, high-frequency, and continuous reflections. The surface of change is called a phase change surface. The bright spot in Fig. 6a (insert) may be evidence for gas migration by this pathway. The phase change incises into the base of the Pliocene and roots in an underlying turbidite paleochannel (Fig. 6). The paleochannel is about 570 km long and 4–8 km wide, which has been drilled by well Ya35-1-2 (Yuan et al. 2009a; b). Most of the BSRs identified in the QDN basin are located above the paleochannel (Fig. 1). Polygonal faults are formed in the Miocene deposits and bright spots and enhanced reflections occur at, or above, the sediments of the polygonal fault and the paleochannel. The geometry of channel changes greatly. It is U-shaped in the Ledong sag, while it is V-shaped in Changcang sag (Yuan et al. 2009a; b). The mound structure is also identified on the seafloor. The seismic facies are characterized medium-amplitude and continuous reflections in the seismic profile (Fig. 6c). The phase change surface and polygonal faults act as pathways for the upward migration of gas in the low permeability mudstone layer.


The source of the fluids is most likely thermogenic gas from the petroleum system that occurs within deeply buried syn-rift sediments in the shallow-water depth (Huang et al. 2003; He et al. 2008; Zhu et al. 2009). Recent geochemical analyses by other researchers on 127 cores in the QDNB and Xisha Trough in the deep-water were treated with acid-digestion analysis. The results show that methane contents range from 10.7 to 243.5 μl/kg. Methane to ethane + propane ratios vary in the range 10–30. Carbon isotopic analyses show methane δ13C values of approximately −43.8 to −26.6‰. The isotopic values of these cores indicate that the gas is thermogenic (Zhu et al. 2008) and support the conclusion that the deep-water region shares the same gas source as the shallow-water area (Zhu et al. 2009). Faults, vertical fracture pipes, diapirs, sand injection, hydraulic fractures, and polygonal faults have been widely recognized as primary conduits for fluid migration (Hurst et al. 2003; Parnell & Schwab 2003; Cartwright 2007; Hurst & Cartwright 2007; Løseth et al. 2009), although how stratigraphic unconformities, structural discordances, and diagenetic deformation serve as pathways for fluid migration is still poorly understood (Campbell 2006).

Qiongdongnan Basin was not tectonically active after 5.5 Ma, although episodic expulsion of fluids have been discussed in the adjacent Yinggehai Basin and this basin (Xie et al. 2001, 2008; Wang & Huang 2008). The evaluation of low-amplitude gas hydrate in the Qiongdongnan Basin requires a thorough understanding of the parameters that control the migration and accumulation of fluids. Diapirs, polygonal faults, gas chimneys, pockmarks and mud volcanoes may all be the pathways for thermogenic gas migrating to the seafloor.

Vertical fluid migration through breakup unconformity and faults

Discontinuities or unconformities are more effective pathways for fluid migration than simple diffusive seepage through the sedimentary column in consolidated sediments (Abrams 1992; Brown 2000). The breakup unconformity along the South China margin has been identified by the penetration of syn-rift sediment at ODP Site 1148, which documented the change from extension and rifting to seafloor spreading and the rapid change in the rate of sediment accumulation at 28 Ma (Clift et al. 2001). The breakup unconformity during the Miocene and Oligocene divides the sediment into upper and lower megasequences. The upper sequence has the characteristics of post-rift thermal subsidence, containing few fractures, whereas the lower sequence records syn-rift tectonics, marked by well-developed fault systems and half-graben basins (Clift & Lin 2001). Y13-1 is the largest gas field in China’s offshore region. The gas and condensate in the field are trapped in a composite structure developed along the eastern flank of the Yacheng uplift, with unconformity-truncated, fault, partly dip-controlled closure (Xie et al. 2008; Zhu et al. 2009).

Some gas blow-out zones and hot fluid expulsion from deeper sediments have been observed around No. 1 faults and adjacent diapir areas. No. 1 fault as a main syn-depositional marginal fault is associated with an accumulation of coarser clastic sedimentary rocks. Fluids from overpressured systems migrate firstly through the coarse clastic rocks and then upwards along the fault (Xie et al. 2003). Therefore, the presence of Ya13-1 gas field adjacent to the No. 1 fault suggests that this fault is one of the main fluid migration pathways. Consequently, the breakup unconformity and faults are commonly considered to be conduits through which fluids migrate vertically from the source rock (in the lower sequence) into the shallower reservoirs (the upper sequence). The sandy layers in well Ya35-1-2 (Fig. 2) may behave as both a conduit for lateral migration and as a reservoir.

Migration through polygonal fault and gas chimneys

Polygonal fault systems are widely developed in many continental margin basins, abyssal basins, and some foreland and intracratonic basins (Cartwright & Dewhurst 1998), although their genesis is still poorly understood. Many experts have proposed mechanisms for the development of polygonal fault systems such as gravitational loading, syneresis and density inversion (Cartwright et al. 2003), overpressure development and downslope gravitational sliding (Henriet et al. 1989) and volumetric strain (Cartwright & Lonergan 1996; Goulty 2002; Goulty & Swarbrick 2005; James et al. 2006). Polygonal faults have been initiated during early compaction of mud-dominated sediments. The instantaneous loading by an overlying debris-flow deposit does not only increase the vertical effective stress beneath it, leading the formation of polygonal faults but also reactivates the pre-existing polygonal faults (Gay & Berndt 2007). In this context, the polygonal faults will be a major pathway for fluids, but fluids are preferentially driven along the triple-junction of three hexagonal first-order faults. The mud diapirs or gas chimneys are formed at the interconnected network zones and allow deeper fluids to migrate upward, reaching the shallower zone (Hansen et al. 2005; Gay et al. 2006; Gay & Berndt 2007). The chimney or diapir strongly affected the plan form of the polygonal faults to change from polygonal away to radial (Hansen et al. 2005). Faults related to tectonic movement are not developed in the late Pliocene to Recent strata in the QDNB (Figs. 3, 6, 7). The younger strata are fine-grained mudstones with interbedded sands from the top of Pliocene to the present (Fig. 2). Numerous polygonal faults and gas chimneys have been identified in the 2D seismic lines and 3D seismic data in the Pliocene-present sediments (Figs. 5, 6). These polygonal faults can act as conduits for the upward migration of gas and fluid (Bünz et al. 2003; Trincardi et al. 2004; Gay et al. 2006a,b; Gay & Berndt 2007; Hustoft et al. 2007; Sun et al. 2009). Bright spots are present at a stratigraphic level above the polygonal faults and at the flank of gas chimneys, which indicates the vertical and lateral seepage of fluids (Fig. 10). Pockmarks are formed where polygonal faults reach the seafloor (Sun et al. 2009). Fluids coming from the deeper levels are trapped in the turbiditic channels and then migrate upward along polygonal fault, gas chimneys, and mud diapir to the shallower level (Figs. 6, 7).

Evidences of slope failure

Submarine slope failure is a fundamental process in the formation of sedimentary basins. The slope failure can be represented by several features including: sub-vertical frontal faults, fault scars, listric faults, imbricate faulting or thrusting, chaotic seismic units, and discontinuous seismic reflections (Davies & Clark 2006). Slumps and channels are created by downslope erosional flows (Orange et al. 2003). In this study, we use the criteria proposed by Solheim et al. (2005) to identify the slide. One is the existence of a headwall, as well as sidewalls, which can be readily identified in seismic profiles and which form part of a mappable slide scar. The other criterion is the presence of seismically homogeneous or chaotic deposits in the slide scar. Based on the 3D seismic data, polygonal faults have been identified in the layer below the failure. Deep thermogenic fluids produced within the Paleogene source rocks can migrate to the Sanya Formation (shallow open marine mudstone and sandstone interbeds) and Meishan Formation (carbonate) along faults that developed during the syn-rift period (Xie et al. 2008). The presence of polygonal fault in the Meishan and Huangliu Formations is a preferential pathway for fluid migration upward (Fig. 6c). Once the fluids migrate to the slope failure, they can escape to the seafloor through an active fault or a frontal fault in the deformed strata (Fig. 7). The fluids can migrate laterally to the adjacent formation along the glide plane (base of failure).

Many factors, such as gas hydrate dissociation, seismic loading, rapid sediment accumulation and under-consolidation, changes in sea level, seepage, earthquake, volcanic island processes, and overpressure, among others can cause slope failure (Dugan & Flemings 2000, 2002; Locat & Lee 2002; Sultan et al. 2004; Flemings et al. 2006, 2008; Dugan 2008). Large-scale slope failure in the Gulf of Cádiz was triggered by upslope salt withdrawal and downslope diapirism that induced by extensional tectonics (Maestro et al. 2003). The primary mechanism of slope failure is because of internal forcing (overpressure, weakening), rather than by downslope erosional flow (Orange & Breen 1992; Orange et al. 2003). Dugan & Flemings (2000, 2002) and Flemings et al. (2008) have presented a model to predict fluid flow and submarine landslides. Rapid sedimentation of low permeability material causes overpressure in young sedimentary basins like the Gulf of Mexico, which is a dominant factor in triggering slope instability. Although the rate of sediment accumulation is high from the Pliocene to the present in the region south of Hainan (Clift & Sun 2006), the rate of sediment accumulation is low in the southern Qiongdongnan Basin (He et al. 2008) near seismic lines 0101,0103, AA′, BB′ and DD′ (Figs. 3, 7, 11) where the slide is mapped. Therefore, seismic loading is unlikely to have triggered the slope failure in the Qiongdongnan Basin. Earthquakes are common mechanism triggering slope failure, but these would be expected to cause multiple slumps along the margin (Canals et al. 2004) rather than a single large slide, as identified in the QDNB data set. The occurrence of gas hydrates and their possible relationship with slides have been discussed earlier (Bouriak et al. 2000; Bünz et al. 2003; Brown et al. 2006; Davies & Clark 2006; Gee et al. 2007). The gas hydrate stability zone depends on pressure/temperature conditions and its stability is sensitive to changes in both water temperature and sea level (Dickens et al. 1994). BSRs indicative of gas and gas hydrate occur in this basin except within the slope failure zone. With known overpressure in the basin (He et al. 2008), the changes in pressure or temperature could cause gas hydrate to dissociate into free gas, which could weaken the sediments and trigger slope failure.

Fluids (water or gas) were expelled during the formation of polygonal faults, which can promote slope failure by the lubricating effect generated by the expelled fluids. A large amount of volcanic activity occurred southeast of the Xisha Uplift (Fig. 7), which could have caused steeper slope angles and instabilities. We propose that the dissociation of gas hydrates, the formation of polygonal faults and volcanic activity are the key factors in triggering the large scale slope failure in this area.

Gas hydrate and depth of gas hydrate stability

The issue of whether the BSR is caused by free gas beneath the hydrates is a controversial issue (Hyndman & Spence 1992; Singh et al. 1993). ODP 164 and 146 indicate that free gas is present underneath BSR (Paull & Matsumoto 2000). Recent insights from Gulf of Mexico show that gas hydrates can exist without a BSR (Hutchinson et al. 2008; Ruppel et al. 2008; Shedd et al. 2009) in sand layers. Waite et al. (2009) suggested that low gas hydrate saturation (20%) would be accompanied by only a negligible or modest increase in sediment velocity and BSR can not be seen. Detailed analysis of amplitude anomalies related with gas/fluid migration in 3D seismic reflection data was used to identify gas hydrate in the Black Sea and the Northern Indus Fan when there is no BSR (Calvès et al. 2008; Wagner-Friedrichs et al. 2008). There are two ways to estimate the depth of gas hydrate stability: by considering velocity measurements, which can provide a direct estimate of depth, and by considering the pressure–temperature equilibrium associated with the base of gas hydrate stability.

Using the first method, the average P-wave velocities above the observed BSR in the seismic data are 1700 m/s. The depth of BSR is located at about 250 ms TWT below the seafloor, which is about 212.5 m.

Using the second method, the depth to the base of hydrate stability can be determined from the seafloor temperatures, thermal conductivity, heat flow and pressure–temperature stability conditions for methane hydrate. Numerous studies have been made on heat flow based on the BSR depth (Yamano et al. 1982; Davis et al. 1990; Hyndman et al. 1992; Ganguly et al. 2000; Kaul et al. 2000). Comparison between measured and BSR-derived heat flow values have been obtained in several margins, such as Cascadia (Ganguly et al. 2000), Makran accretionary prism (Kaul et al. 2000) and Xisha Trough of the SCS (Wang et al. 2005; He et al. 2009).

Two measured seafloor temperatures near the seismic lines 0101 and 0102 are about 3.074 and 4.284°C, respectively. The average measured seafloor temperature is 3.679°C.

Wu et al. (2007) established an empirical expression giving the average thermal conductivity from ODP Site 1148:


where k is the thermal conductivity in W/(mK) and z is the depth below the seafloor in meter. This relationship was also developed by Davis et al. (1990) using the simple geometric mean. The thermal conductivity value is in agreement with the measurements (1.0 W/mK) in the sediments in the Xisha Trough of the SCS (He et al. 2009).

The depth of the BSR (Zbsr) was calculated using the simple conductive heat transport relation:


where H is heat flow in mW/m2; Tbsr is the temperature at BSR depth in K, which was determined from the pressure–temperature stability conditions for methane hydrate (Dickens & Quinby-Hunt 1994). The heat flow value in seismic profile 0101 and 0102 is about 79 mW/m2 (Clift & Lin 2001; Shi et al. 2003). The depth of the base of hydrate stability can be calculated using Eqs. 1 and 2, which yields a depth of about 246 m. This depth is about 34 m greater than the depth to the BSR estimated from velocity measurements, but is within the uncertainties of both methods. The high amplitude anomalies above the BSR were caused by the presence of gas hydrate in the sediments, while the bright spots below BSR indicated the presence of free gas.


Based on high-resolution 2D and 3D seismic data, fluid pathways have been identified in fine-grained younger sediments, a tectonically-quiescent sequence of the Qiongdongnan Basin. The pre-Miocene syn-rift sequence has a well-developed faults system, breakup unconformity and permeable sand layers that act as pathways for the migration of thermogenic fluids containing gas from underlying source rocks. The post-Miocene, postrift sequence has very-low-permeability mudstone sediments. These mudstone sediments formed the regional caprock and trapped the hydrocarbons generated from the Paleogene strata. The polygonal faults developed in the Meishan and Huangliu Formations are preferential pathways for fluid migration from deeper levels. Gas chimneys are also identified above or within the polygonal faults developed sediments. The high-amplitude reflections and slightly elevated velocities observed above gas chimneys at the approximate depth to the gas hydrate stability zone indicate the presence of gas hydrate in the sediments. The presence of bright spots above or at the flanks of gas chimneys indicates lateral and vertical migration of gas/fluid in the Qiongdongnan basin.


We are grateful to thank Dongdong Dong for the suggestion of geological setting in Qiongdongnan Basin. We are grateful to Dr. Shaoming Lu for improving earlier version of the manuscript and Dr. Deborah Hutchinson for improving the English of the last version. We wish to thank Zhiliang Qin for his input regarding slope failure mechanisms. We wish to thank Brandon Dugan, Peter Clift and Joe Cartwright for having shared their research papers regarding this topic. I want to thank editor Richard Worden and the anonymous reviewers’ suggestions and comments to help revising the papers. Our research is supported by the National Basic Research Program (2009CB219505), the CAS Knowledge Innovation Program (KZCX2-YW-229), the National Natural Science Foundation of China (40706026) and the Knowledge Innovation Program of the Chinese Academy of Sciences.