Roberts

Journal of the Geological Society, London, Vol. 168, 2011, pp. 1013–1030. doi: 10.1144/0016-76492010-158

Structural controls on mud volcano vent distributions: examples from Azerbaijan and Lusi, east Java

KATIE S. ROBERTS1;2*, RICHARD J. DAVIES1, SIMON A. STEWART3 & MARK TINGAY4

1CeREES (Centre for Research into Earth Energy Systems), Department of Earth Sciences, Durham University, Science Laboratories, South Road, Durham DH1 3LE, UK

2 Present address: Hess Services Ltd., London, WC2N 6AG, UK

3 Institute of Petroleum Engineering, Heriot–Watt University, Edinburgh EH14 4AS, UK

4 Tectonics, Resources and Exploration (TRaX), Australian School of Petroleum Geoscience, University of Adelaide,

Adelaide, SA, 5005, Australia

*Corresponding author (e-mail: KSRoberts@Hess.com)

Link to Picaca Web for previes slide shows of Figure caption:

https://picasaweb.google.com/hardiprasetyo9/StructuralControlsOnMudVolcanoVentDistributionsExamplesFromAzerbaijanAndLusiEastJava?locked=true#

ABSTRACT

Structural mapping, nearest neighbour and two-point azimuth statistical analysis of mud volcano vent distributions from nine examples in Azerbaijan and the Lusi mud volcano in east Java are described.

Distributions are non-random, forming alignments subparallel to faults within anticlines, ring faults, conjugate faults and detachment faults; this finding confirms a spatial relationship and supports a model for subsurface flow along these features as well as showing fractionation at depth.

As fracture and fault orientations are related to structures such as anticlines and the in situ stress state they are therefore predictable.

We use vent distributions in Azerbaijan, where the structural geology is well constrained, to propose what controls the distribution of 169 vents at the Lusi mud volcano.

This mud volcano system shows evidence for initial eruptions along a NE–SW trend, parallel to the Watukosek fault, changing to eruptions that follow east–west trends, subparallel to regional fold axes.

Our analysis indicates that regions east and west of the Lusi mud volcano are more likely to be affected by new vents than those to the north and south, owing to probable onset of elongate caldera collapse within a 10 km diameter of the central vent.

Introduction

Mud volcano systems are a dynamic type of piercement structure that are integral components of many sedimentary basins globally (e.g. Kopf 2002).

However, because of the ephemeral nature of the fluid flow the structural pathways exploited during their intrusion are poorly defined. Fine-grained sediment and fluid can be transported from depths exceeding 8 km (Kopf 2002) resulting in eruption at the surface producing edifices that can potentially reach 25 km3 in volume (Davies & Stewart 2005).

The driving force is generally considered to be excess pore fluid pressure (e.g. Davies et al. 2011) but how mud is entrained and the pathways for this fluid are poorly understood.

It has been proposed that folds, faults and fractures may be some of the dominant controls on fluid migration pathways during the intrusion of mud volcano systems (e.g. Morley 2003; Roberts et al. 2010) but this has not been tested.

Analysis of surface vent patterns and their spatial relationships to these structures represents one type of test, which has successfully been carried out for igneous volcanoes but not for their mud volcano counterparts.

Alignments of point-like geological features such as volcanic cones, hydrothermal vents, fluid expulsion pipes and springs have often been shown to follow underlying structures, such as dykes, faults or joints (e.g. Hammer 2009; Moss & Cartwright 2010; Paulsen & Wilson 2010).

Igneous vent patterns have been studied in great detail both on Earth (Lutz 1986; Wadge & Cross 1988; Connor 1990; Hammer 2009; Paulsen & Wilson 2010) and on extraterrestrial bodies (Bleacher et al. 2009).

Alignments of point-like geological features such as volcanic cones, hydrothermal vents, fluid expulsion pipes and springs have often been shown to follow underlying structures, such as dykes, faults or joints (e.g. Hammer 2009; Moss & Cartwright 2010; Paulsen & Wilson 2010).

Igneous vent patterns have been studied in great detail both on Earth (Lutz 1986; Wadge & Cross 1988; Connor 1990; Hammer 2009; Paulsen & Wilson 2010) and on extraterrestrial bodies (Bleacher et al. 2009).

Time-dependent changes in igneous vent distributions have never been considered, mainly because these changes tend to occur over longer time scales than mud volcanoes, over thousands to millions of years (Paulsen & Wilson 2010).

Here we use eight mud volcanoes in Azerbaijan to statistically analyse the distribution of vents and relate these to well-exposed geologically mapped structures such as folds, faults and fractures (Hovland et al. 1997; Guliyev et al. 2000; Planke et al. 2003).

These examples provide confidence in assessing the controls on vent distributions that can then be applied to the Lusi mud volcano in east Java (Fig. 1)

This has continuously erupted since 2006, displacing 13000 families, but the structural geology that could be influencing vent locations is not well constrained.

At the time of writing Lusi has 169 vents, which erupt and sometimes ignite without warning in the surrounding densely populated area (Tingay 2010).

Our analysis allows for a general assessment of the pathways for fluid and gas in mud volcanoes. In the case of Lusi, it provides a better understanding of where vents are more likely to occur in the future.

Mud volcanoes and vent populations

A mud volcano system includes an intrusive domain containing the feeder complex; a source domain, which comprises the source of water, gas and mud; and an extrusive domain, which is dominated by the eruptive edifice (Stewart & Davies 2006; Roberts et al. 2010).

Mud volcano eruptions can be violent with quiescent inactive stages during which eruptions from multiple, small vents are the dominant process (Guliyev et al. 2000).

Dormant mud volcano edifices can have anything from one to thousands of vents of differing morphologies erupting varying compositions of mud, water and gas (Table 1; Fig. 2).

Vent types include cinder mounds, which are the result of 100% gaseous phase eruptions. Mud breccia flows consist of 90–100% mud that has the rheology of a Bingham body (Iverson 1997).

Salses contain suspensions of ,30% mud particles in water and gryphons have a composition of 30–90% mud particles in water that are either thixotrophic or shear thinning (Yassir 1990, 2003; Mueller et al. 2010).

Herein the pattern of vents on edifices is termed ‘vent populations’. The area that is most densely populated with vents is termed the active vent zone (Roberts et al. 2010).

Geological settings: Azerbaijan

Azerbaijan’s mud volcanoes probably form as a result of rapid sedimentation during the last 5.5 Ma, tectonic compression, the presence of a thick overpressured mudstone (Maykop Formation) at 5–8 km depth and hydrocarbon maturation (Davies & Stewart 2005; Evans et al. 2008).

All the mud volcanoes are located along or near the crests of anticlines and most are thought to have initiated in the Pliocene (c. 3.5 Ma; Narimanov 1993; Yusifov & Rabinowitz 2004).

The mud volcano systems may also incorporate fluids rising from below the Maykop Formation (Kopf 2002; Hovland et al. 2006).

The region has undergone 2.4 km of tectonic subsidence since 5.5 Ma (Allen et al. 2003).Several kilometres of sediment accumulated during the Pliocene and have subsequently been folded, with these structures having a dominant NW–SE fold axis orientation (Allen et al. 2003; Yusifov & Rabinowitz 2004).

Geological settings: Lusi, Sidoarjo, east Java

This mud volcano erupted in the east Java basin in May 2006 (Davies et al. 2007; Mazzini et al. 2007). During the Eocene, NE– SW-oriented rift basins formed and filled with continental clastic rocks that host both source rock and productive reservoirs (Kusumastuti et al. 2002).

In the Oligocene to early Miocene east– west-trending normal faults formed (Kusumastuti et al. 2002; Istadi et al. 2009).

Carbonate platforms developed on some palaeo basement highs. Carbonate reefs are located beneath the Lusi mud volcano and have an east–west orientation (Kusumastuti et al. 2002).

Compression during the late Miocene–Pleistocene resulted in inversion associated with east–west-trending fault movement (Istadi et al. 2009). This produced the east–west orientation of the anticline structures (Istadi et al. 2009).

Subsequent Pliocene– Pleistocene sedimentation consisted of an eastward-prograding mudstone-dominated volcaniclastic wedge derived from the Java volcanic arc (Istadi et al. 2009).

The mudstone of the Pleistocene Kalibeng Formation is overpressured at 900–1870 m depth at Lusi (Istadi et al. 2009).

This is the source of the mud that makes up the solid fraction of the erupted liquid mud (Mazzini et al. 2007).

The water is most likely to have been sourced from the Miocene carbonates (2833–3500 m; Tanikawa et al. 2010; Tingay 2010; Davies et al. 2011) with a contribution from the remobilized Upper Kalibeng Formation (Davies et al. 2007).

Some fluids may also be sourced from shallow aquifers in the Pleistocene Pucangan Formation at 280–900 m depth (Tingay et al. 2008).

New vents form frequently and several have ignited causing injury. For example, the Porong highway, near the Lusi mud volcano, developed metre-long cracks leaking methane on 2 July 2010, with the highway surface increasingly sloping toward the mud embankments used to limit the spread of the mud.

New vents form frequently and several have ignited causing injury. For example, the Porong highway, near the Lusi mud volcano, developed metre-long cracks leaking methane on 2 July 2010, with the highway surface increasingly sloping toward the mud embankments used to limit the spread of the mud.

It has been predicted that it will take 26 years for the flow rate to reduce to 10% of its initial rate (Davies et al. 2011). Therefore more vents will form and the subsidence (Abidin et al. 2008; Istadi et al. 2009) will continue for decades.

Database and methods : Structural mapping

Mapping of vent populations was carried out using a handheld global positioning system (GPS) receiver, with a positional accuracy of 5 m (Azerbaijan) and 5–12 m (Lusi data, courtesy of Badan Penanggulangan Lumpur Sidoarjo (BPLS)).

Bedding, fracture and fold orientations were plotted using GEOrient software onto stereographic and rose projections.

The GPS coordinates and corresponding structural data were integrated in ArcMap software.

The coordinate system for these data was input using spheroid WGS 1984.

Vents were classified as either gryphons, salses, cinder mounds, mud plugs or pools (Fig. 2; Table 1; Hovland et al. 1997; Guliyev et al. 2000; Mazzini et al. 2009).

Each is marked onto satellite imagery with different symbols (Fig. 2). The potential spatial relationships between folds, faults and fractures and vent populations should be clear, as exposure is 60%. Structural data from outcrop and 2D seismic coverage for Lusi mud volcano are limited.

Two statistical approaches, adapted from igneous vent systems, are used to characterize spatial patterns within vent populations.

At igneous vent systems these techniques have revealed that magmatic volcanic vents often form clusters and define alignments at several scales from tens of metres to over 1000 km (Bleacher et al. 2009; Paulsen & Wilson 2010). As the GPS accuracy is 5 m, vent alignments have 5 m accuracy.

TThe nearest neighbour technique (Clark & Evans 1954) tests randomness in spatial distributions by calculating the ratio of the observed mean distance to the expected mean distance for a hypothetical random distribution to determine whether the points are clustered.

A ratio of unity is a random distribution and a ratio of, 1 is clustered; the nearer to zero the more clustered the distribution.

This analysis was carried out using ArcGIS, which measures the distance from every vent point to its nearest neighbouring vent point.

The two-point azimuth technique (Lutz 1986; Bleacher et al. 2009) is used as a measure of the significance of alignments between vents.

The technique quantitatively identifies trends within vent populations and has been widely used in studies on the structural geology of igneous volcanoes (Wadge & Cross 1988; Connor 1990; Bleacher et al. 2009).

The azimuths of line segments that connect each vent to all other vents east of its location were calculated (Bleacher et al. 2009).

Only points to the east of each vent were measured so as not to duplicate any measurements.

Database and methods: Statistical analyses

Two statistical approaches, adapted from igneous vent systems, are used to characterize spatial patterns within vent populations.

At igneous vent systems these techniques have revealed that magmatic volcanic vents often form clusters and define alignments at several scales from tens of metres to over 1000 km (Bleacher et al. 2009; Paulsen & Wilson 2010).

As the GPS accuracy is 5 m, vent alignments have 5 m accuracy.

The nearest neighbour technique (Clark & Evans 1954) tests randomness in spatial distributions by calculating the ratio of the observed mean distance to the expected mean distance for a hypothetical random distribution to determine whether the points are clustered.

A ratio of unity is a random distribution and a ratio of, 1 is clustered; the nearer to zero the more clustered the distribution.

This analysis was carried out using ArcGIS, which measures the distance from every vent point to its nearest neighbouring vent point.

The two-point azimuth technique (Lutz 1986; Bleacher et al. 2009) is used as a measure of the significance of alignments between vents.

The technique quantitatively identifies trends within vent populations and has been widely used in studies on the structural geology of igneous volcanoes (Wadge & Cross 1988; Connor 1990; Bleacher et al. 2009).

The azimuths of line segments that connect each vent to all other vents east of its location were calculated (Bleacher et al. 2009).

Only points to the east of each vent were measured so as not to duplicate any measurements.

Histograms of azimuth values (08 ¼ north, 908 ¼ east, 1808 ¼ south) were produced with 108 bins.

Peaks in the frequency distribution of the azimuths result from preferred formation of vents in response to structural controls (Bleacher et al. 2009).

In this study the ‘dominant’ alignment refers to the azimuthal trend with the highest frequency of azimuths.

Sub alignments include smaller peaks in azimuth frequency less significant than that of the ‘dominant’ alignment.

Different vent types are separated and the azimuth alignments of each of the vent fluid types displayed (i.e. mud, water and gas) are analyzed.

The ‘overall’ azimuth alignments, which include all vent types for each volcano, are also plotted to identify larger scale influences on vent alignments of the whole edifice.

On each of the graphs ‘Y’ indicates the orientation of the anticline axis in the region, ‘X’ the orientation of any faulting measured during mapping and ‘A’ any anomalous values that may be the result of external factors, such as human influences (e.g. loading induced fluid flow around manmade dams (e.g. Londe 1987)).

This contrasts with the extrusive features seen at Koturdag B and C, at which gryphons, salses and breccia pipes are present (Fig. 2a and b).

Koturdag B has a high concentration of salses, 0.2–5 m in diameter, compared with the increased concentration of 1–2 m high gryphons found at Koturdag C (Fig. 3a).

Koturdag C is located at 100 m higher elevation than Koturdag B and has twice as many gryphons.

In contrast, Koturdag B has twice as many salses as Koturdag C. Both Koturdag B and C edifices have long axes that align with the anticline axis at 1308N (Fig. 3a).

The orientations of fractures in the area are subparallel to the anticline axis at 130–1408N with another peak at 908 to this, at 0508N (Fig. 3b).

When including all the vent positions along Alyaty Ridge as a whole the observed frequencies of azimuths derived from the two-point azimuth technique show preferential alignment in the direction of 120–1308N (Fig. 3c).

Koturdag B (Fig. 3d) and C (Fig. 3e) share this dominant 1308N trend.

Koturdag B also shows a peak in salse alignment at this orientation whereas Koturdag C shows a peak in gryphon alignment (Fig. 3d and e).

Observations :

Alyaty Ridge

This ridge is an anticline that extends for 12 km in a NW–SE orientation and hosts several mud volcano systems. Koturdag A has a single, 240 m diameter caldera on its summit, which is 500 m to the north of the anticline axis (Fig. 3a).

The most recent mud breccia flow has been continuing for 50 years and is currently extruding mud breccia from a 20 m wide vent at a rate of 2–6 cm day1.

The flow has areas of oxidized mud breccia and cinder, which are the result of escaping gases igniting mud during eruptions (Fig. 2c; Hovland et al. 1997; Guliyev et al. 2000).

The 20 m wide vent has a 1 m high gryphon 5 m away from it.

Kichik Kharami mud volcano

This is 1 km to the south of a NW–SE-trending anticline axis and is roughly circular in plan view (Figs 1a and 4a, b).

Minor amounts of mud are being expelled in the form of salses, although a 1.2 km long mud flow to the south of the feeder complex is evidence for a significant eruption of mud breccia within the past few hundred years (Fig. 4a).

The salses have a circular arrangement at the centre of the volcano (Fig. 4b), and at 100 m from the centre of the volcano they orient themselves in NW–SE and NE–SW linear trends (Fig. 4b).

Fractures are dominantly arranged subparallel to the anticline axis at 120–130^oN with a set perpendicular to this at 03^oN.

There are also smaller fracture alignments at 10^oN and 16^oN, which form two planes, each at roughly 3^o to the fold axis (Fig. 4c).

The dominant azimuthal frequency at Kichik Kharami is 130–140 ^oN (Fig. 4).

There are also secondary alignments, for example at 09 ^oN (Fig. 4), which do not share a common orientation with any structures in the area.

The salses show a dominant azimuth subparallel to that of the strongest fracture orientation at 13^oN.

Pirsaatadag mud volcano

This is located on the axis of a NW–SE-trending anticline and has an elliptical shape, the long axis of which is aligned with the anticline axis at 15 ^oN (Figs 1a and 5).

Minor amounts of mud are being expelled in the form of salses and pools.

The mud volcano is heavily eroded so exposure of country rock at its centre allows easy measurement of structures. The active vent zone of the volcano is offset to the northwestern end of the edifice and displays a slight circular arrangement of vents at its centre (Fig. 5a).

The dominant azimuthally frequency at Pirsaatadag is 18 ^oN (Fig. 5c and d); however, there is also a high azimuth frequency subparallel to the anticline axis at 15 ^oN. There is a lack of azimuths at 09 ^oN.

Akhtarma-Karadag mud volcano

Akhtarma-Karadag crops out along an ENE–WSW-trending anticline axis and is also elongate parallel to this anticline axis (Figs 1a and 6a).

The active vent zone on the summit is found at the western end of the edifice (Fig. 6a).

It has three eruptive compositions: cinder mounds, salses and gryphons.

There are three cinder mounds at the western edge of the mud volcano (Fig. 6a), only 1 m in height and diameter.

The salses are towards the centre of the edifice and have a maximum diameter of 10 m.

The main concentration of gryphons is closer (c. 6 m) to the cinder mounds.

There are also numerous dormant gryphons (Fig. 6a).

The two-point azimuth technique shows a dominant azimuth frequency for gryphons and salses at 03 ^oN whereas the pools tend to align at 12 ^oN (Fig. 6b).

When including all vent types, the dominant alignment can be seen to be at 07 ^oN, which does not align with the anticline axis oriented at 09 ^oN (Fig. 6).

Dashgil mud volcano

Dashgil mud volcano This is on the crest of the Dashgil fold (Fig. 1a), which is 6– 8 km long, 3.5–4 km wide and trends in an east–west direction.

The active vent zone is offset to the western end of the edifice (Fig. 7a).

There is a concentration of gryphons, 2–3 m in height, clustering at the centre of 200 m diameter crater to the west of the volcano (Fig. 7a).

A 200 m long row of 2–3 m high, 4–5 m wide cinder mounds trends in an east–west direction.

These are found only in the southeastern section of the volcano and form a sharp, straight boundary to the edge of the active vent zone.

Dashgil also has two salses 20–30 m in diameter on its summit in the eastern portion of the mud volcano. These are composed of several bubbling centres.

There is also a small cluster of dormant gryphons in the northern section of the volcano.

Both the combined and separate vent type two-point azimuth results show that the dominant orientation in this system is at 050 oN with sub-orientations at 10 ^oN and 17 ^oN (Fig. 7c).

When separating different vent types from each other three ‘peaks’ in azimuth frequency can be seen for both gryphons and salses at 06 ^oN, 11 ^oN and 17 ^oN, whereas pools only have one dominant trend at 060 oN (Fig. 7b).

Durovdag mud volcano

Durovdag volcano has a wide spread of vent azimuth frequencies, which is also seen on a smaller scale at the centre of Kichik Kharami volcano (Fig. 4).

The dominant orientation in this system is at 1608N with sub-orientations at 10 ^oN and 02 ^oN (Fig. 4).

Durovdag mud volcano The crest of the volcano is dominated by gryphons and salses, which are , 2 m in height (Figs 1a and 8a).

There is a concentration of gryphons at the northern end of the volcano, with an average vent spacing of 5 m.

Owing to the unstable nature of this region separate readings could not be taken and so the area has been considered as one large vent in the statistical analysis.

The majority (92%) of the remaining vents on the summit are found around the outer edge of the mud volcano, forming an 800 m diameter ‘ring’ (Fig. 8a).

The vents also align at tens of metre scale, along linear conjugate paths within this ‘ring’ zone.

Durovdag volcano has a wide spread of vent azimuth frequencies, which is also seen on a smaller scale at the centre of Kichik Kharami volcano (Fig. 4).

The dominant orientation in this system is at 16^oN with sub-orientations at 10^oN and 02^oN (Fig. 4).

Lusi mud volcano, east Java

The Lusi edifice is 3.4 km by 2.6 km in areal extent (Fig. 9). The main active vent is 100 m in diameter and located at the centre of the edifice (Fig. 9).

The first seven vents at Lusi formed roughly aligned in a NE–SW direction during first week of eruption (29 May 2006; Mazzini et al. 2007).

This increased to 34 in November 2006, also in a NE–SW orientation.

A fracture hundreds of metres long and tens of centimetres wide was observed a few days after the eruption, which also had a NE– SW orientation (Mazzini et al. 2007).

This was interpreted as being the Watukosek fault, which is interpreted to cross the area(Fig. 9; Mazzini et al. 2007).

Most of the early ‘sandy’ eruption sites discussed by Mazzini et al. (2007) were buried during the second week of June 2006, by the mud erupting from the main vent.

New smaller vents started erupting in November 2006 1 km to the SW of the main crater (Mazzini et al. 2007).

Currently there are 169 active vents (BPLS) although not all vent occurrences can be documented owing to limited access to the majority of the edifice and because some are short lived.

The vents near the main central vent had a roughly concentric pattern (Fig. 9a) whereas vents further away are closer to the observed faults in the region (Fig. 9a).

Newer vents occur further away from the central vent and are now clustering close to the Kendensari River to the west of Lusi (Fig. 9b and c).

These eruption sites erupt gas or suspensions of ,20% mud in water.

The two-point azimuth data for Lusi mud volcano (Fig. 9) show the vent distribution in 2006, a few months after it first erupted, compared with the vent distributions seen in 2009 and 2010.

In 2006 the dominant azimuth frequency is WNW–ESE (10^oN), with two smaller trends at 06^oN in a NE–SW orientation and 12^oN in a NW–SE orientation (Fig. 9a).

There is also a large spread in azimuth frequencies apart from the dominant trends (Fig. 9a).

In 2009 there are two dominant azimuthal trends at 01^oN and 1808N with two less dominant trends at 10 ^oN and 12^oN (Fig. 9b).

The 06^oN NE–SW azimuthal orientation of vents in 2006 has now decreased in frequency.

In 2010 this trend continues with the decreasing influence of the 06^oN and 12^oN alignments and increasing frequency of alignments at 01^oN, 18^oN and 10^oN (Fig. 9c).

Discussion:

Alignment along anticline crests.

Dominant trends at each of these volcanoes are also at 130 oN, showing that both the mud volcano systems as a whole and the metre-scale vent populations align in the same orientation as anticline axes. Fluids are most probably taking advantage of pathways produced by increased compressive shear failure in the anticlinal cores, and outer arc crestal faulting along the anticlines (Ramsay & Huber 1987).

The folding has brought the overpressured Maykop Formation to a shallower depth in the subsurface and allowed thickening of these strata in the anticlinal hinges (Allen et al. 2003).

This, as well as the unloading of the anticlines during exhumation onshore and decreased overburden load, would decrease the force needed for the overpressured Maykop Formation to overcome the vertical stress and the tensile strength of the overburden (Magara 1981; Yassir & Bell 1996).

Alignment along anticline crests.

These factors significantly increase the potential for the mud– water–gas mix to travel to the surface and erupt along these planes of weakness (Yusifov & Rabinowitz 2004).

Mud volcanoes also tend to become elongate in the direction of the anticline axis, as seen for many of the examples in this study (Figs 3–8).

Elongation of edifices is also seen in igneous volcanoes and is generally parallel to the maximum horizontal stress (Nakamura 1977; Paulsen & Wilson 2010).

This is attributed to formation of vents along feeder dykes that orient parallel to the maximum stress and open perpendicular to the minimum horizontal stress (Hmin; Paulsen & Wilson 2010). In mud volcano systems and their vent populations this is not the case, as they all extrude along or subparallel to anticline axes, which form perpendicular to the maximum horizontal stress (Hmax; Fig. 11).

This is to be expected, as the ‘source’ of the fluids and any mud chambers feeding the edifices would also become elongate perpendicular to the maximum horizontal stress (Fig. 11).

The result of this is that vent populations on mud volcano edifices provide a good indicator of both palaeo- and current regional stress regimes.

Alignment with fractures.

Fault and fracture networks can act to either enhance or prevent fluid flow depending on their relative permeability compared with that of the surrounding country rock (Aydin 2000; Eichhubl & Boles 2000; Faulkner et al. 2011).

When faults and fractures have high permeabilities they are able to act as pathways allowing fluids to utilize them as a conduit to the surface (e.g. Sibson 1996; Faulkner et al. 2011).

A prominent characteristic of mud volcano systems is high fluid pressures, which may result in the formation of hydrofractures and shearing producing open fractures and dilatant faults (e.g. Aydin 2000).

By comparing vent alignment orientations with structures mapped in close proximity it is possible to identify which fault and fracture systems have the highest permeability in a certain region.

The cinder mounds on Dashgil are found only in a discrete elongate zone and so probably form when gas venting from the mud volcano feeder complex travels along a pre-existing fault plane (Fig. 10a).

This faulted zone may intersect a mud chamber that has separated phases of gas, water and mud within it.

Periodically the pressure in this chamber would become high enough to overcome the tensile strength and minimum horizontal stress, producing new hydrofractures in a similar way to fault-valve behaviour, allowing fluids to erupt at the surface as discrete events (Sibson 1990, 1992).

This fault may even be an anticline crestal fault, as the cinder mounds can be seen oriented in an east–west direction similar to that of the Dashgil Fold (Fig. 7a).

Kichik Kharami mud volcano is similar to Durovdag at its centre, with a 10 m diameter ring of salses forming along a circular collapse structure. However, 100 m out from the centre, the salses are aligned in rows in NW–SE (160 oN) and NE–SW (1008N) directions (Fig. 4d and e).

These orientations are coincident with the orientation of shear fractures (Fig. 4c) found on anticline flanks (Ramsay & Huber 1987) and both occur at 308 from the anticline axis orientation of 1308N.

This implies that these have the highest permeability compared with other structures in the region (Fig. 10b).

Dashgil also displays these fault arrangements (Fig. 7), with ‘peaks’ in both gryphon and salse azimuths occurring at 060 oN and 170 oN fracture orientations occurring at 608 to the anticline axis orientation (1108N).

Detachment fault alignment.

Both active and extinct gryphons and salses on the Akhtarma-Karadag mud volcano align along a linear offset that can be traced around the summit of the volcano, which is interpreted here as a detachment fault (Figs 6 and 10e; Roberts et al. 2011).

Pressure ridges of sediment can be seen at the centre of the detachment fault, suggesting that the mud volcano appears to be failing to the NE (Fig. 6a).

This movement is confirmed by the presence of plants being torn across the head of the detachment fault and en echelon fracturing.

This is again supported by the two-point azimuth statistics, which show that the vents have a dominant orientation similar to that of the detachment fault at 070 oN (Fig. 6c).

Ring fault alignment.

Durovdag displays clear alignment of its vents, with 92% of the gryphons and salses erupting around the periphery of the edifice (Fig. 8).

This alignment is to be expected for a caldera collapse system (Stewart & Davies 2006; Evans et al. 2008), with the majority of the vents forming a ‘ring’ around the outer edge of the mud volcano (Fig. 10c).

These fluids are taking advantage of ring faulting that is forming as a result of the gravitational collapse of the mud volcano. This distribution is displayed as a large spread of alignments on the two-point azimuth histograms, as well as showing the slightly more dominant anticline axis alignment (1608N) and less dominant alignments that may be caused by fracture alignments (1008N and 0208N; Fig. 10c).

On a metre scale vents align in a conjugate pattern similar to shear fracturing on anticline limbs (Ramsay & Huber 1987).

These metre-scale alignments occur around the trace of the kilometre-scale ‘ring’ fault itself (Fig. 8a).

It is likely that these metre-scale conjugate vent alignments formed first aligning with the pre-existing anticline fractures.

After this, caldera collapse initiated and formed the more recent ring fault alignments, which then overprinted the conjugate alignments to produce the dominant azimuth frequency.

The concentration of gryphons to the north of the volcano indicates that there may be a large mud chamber beneath this area.

Distributions: fractionation of vent eruptive phases

Dashgil and Akhtarma-Karadag both produce three eruptive compositions: gaseous (cinder mounds), watery mud (salses) andviscous mud–water mix (gryphons).

They also show a similar spatial distribution of erupting fluid types. Dashgil is dominated by gryphons on its westerly side, salses to the east and cinder mounds to the south of the active zone of the edifice (Fig. 7a).

Akhtarma-Karadag has cinder mounds in the most westerly section, 5–10 m from an area of gryphons at the centre of the active zone, and then salses at the easterly end of the volcano (Fig. 6a).

From these observations it is possible to ascertain that these three phases must be separating at depth and travelling to their points of eruption via different pathways.

This has been noted by others in past studies at Dashgil mud volcano (Mazzini et al. 2009).

Mazzini et al. (2009) found that the water geochemistry highlights different water sources and reactions that occur at gryphons, pools and salses.

Gryphons have a signature of deep-rising fluids, whereas pools and salses show the imprint of meteoric fluids and a solute content increased by in situ evaporation (Mazzini et al. 2009).

When integrating this with the observations it can be assumed that gryphons may be fed directly from a mud chamber in the main feeder complex of the mud volcano at depth, whereas salses and cinder mounds are most probably sourced from shallow, smaller chambers that remain ‘stagnant’ for periods of time, allowing them to interact with the surrounding meteoric fluids.

The azimuth frequencies for each vent type show that gryphons and salses often display common orientations, indicating that they may share similar fluid flow pathways in the subsurface (Figs 3, 6 and 7).

Pools show no correlation with other vent types, in agreement with Mazzini et al. (2009), who suggested that these are only shallow fluid flow pathways that are not influenced by regional structure (Figs 3, 6 and 7).

Time-dependent changes, Lusi mud volcano

The dominant 010 oN and 180 oN vent azimuth orientations seen at Lusi in 2009 may result from local loading by the earth dams in this region, which share this alignment.

Loading would allow focused fluid flow in this orientation (e.g. Londe 1987) or could be the result of sampling bias, as these locations are more readily reported by residents of Sidoarjo.

Because of these possible influences we focus on the second most dominant azimuth orientations, which change from NE–SW to east–west vent alignments.

During the first eruptive phase in 2006 vents were aligned in a NE–SW orientation (Fig. 9a) at c. 308 to the present-day maximum horizontal stress (1, Hmax) orientation of NNE– SSW (Mazzini et al. 2009; Sawolo et al. 2009; Tingay et al. 2010).

This is consistent with fluids initially travelling up the highest permeability paths, which were optimally oriented for sinistral shear failure in a strike-slip faulting stress regime, and originating from the Miocene carbonates at 2800 m depth (Fig. 12a; Davies et al. 2008; Tingay 2010).

Analogous to this is the formation of Miocene shale dykes along faults in the Jerudong Anticline of Brunei (Tingay et al. 2005).

It is well documented that faults can transmit significant volumes of fluids when active (Barton et al. 1995; Sibson 1996) especially if they have a higher permeability than the surrounding country rock.

However, this in no way indicates that reactivation of the fault triggered the initial eruption (see Davies et al. 2011).

The second phase of eruption during 2009 showed the east– west (1008N) vent azimuth alignments becoming even more prominent, and these became increasingly dominant in 2010 (Fig. 9c).

Evolution of these vent populations has occurred in only 9 months and implies that the fluid pathways themselves are developing during a similar time period.

It also suggests that an important east–west-trending, regional-scale anticlinal structure influences the feeder system architecture, reducing the importance of the local NE–SW fault.

The fact that so few vents are erupting the same fluid as the main Lusi vent, and that most are thought to be very shallow rooted, implies that the source for the main vent and the smaller vents may differ.

One preferred interpretation is that many of the water eruptions are coming from 290–900 m deep aquifers (Tingay et al. 2008) that have become faulted owing to subsidence, resulting in seal breakage and fluid flow.

From studies of mud volcanoes in Azerbaijan it is possible to make the assumption that older feeder systems naturally take advantage of pre-existing structures in the region (Fig. 12b).

It is proposed that this change in orientation occurred as a result of a drop in pore fluid pressure in the system once the majority of the main source of overpressured fluid had been erupted.

The decreased pore fluid pressure was lower than the tensile strength and minimum principal stress required to keep the original hydrofractures open, resulting in closure of these pathways and a decrease in their permeability (e.g. Jolly & Lonergan 2002).

A ring-like arrangement of vents is observed around the main central vent at Lusi (Fig. 9a) and is similar to the pattern seen at Durovdag (Fig. 8).

This could indicate that a ring fault has formed as a result of subsidence in the region owing to the evacuation of large volumes of fluid from depth (Fukushima et al. 2009).

Abidin et al. (2008) and Fukushima et al. (2009) both used time-lapse synthetic aperture radar interferograms from 1 year after the start of the Lusi mud eruption in May 2006 to show subsidence over an ellipsoidal area of 12 km2 centred on the main eruptive vent.

Depletion of material and decrease of fluid pressure at depth were described as being the dominant cause of the subsidence.

Fukushima et al. (2009) found that deflation of an oblate spheroid lying shallower than 1 km explains the observed displacements.

This observation is supported by the 2010 azimuth data (Fig. 9c), which show a wider spread of azimuth trends than seen in 2009.

Mode of formation: 2008–present.

The Miocene carbonates are proposed as the primary source of the fluids driving the eruption from the main vent (Davies et al. 2007, 2008; Tanikawa et al. 2010).

However, other studies have suggested that the shallower Upper Kalibeng clays are the source of the majority of the fluids (Mazzini et al. 2007).

We speculate that from 2008 to the present, subsidence up to 10 km away from the main vent, resulting from the evacuation of large quantities of remobilized mud (Abidin et al. 2008), may have been accommodated by the reactivation of the east–west-trending crestal faulting along an anticlinal structure.

During reactivation these faults would have breached aquifers located in the Pucangan Formation (280– 900 m depth; Fig. 12b; Tingay et al. 2008).

Overpressured fluids from these aquifers would use these high-permeability, reactivated faults as conduits to the surface.

This is supported by the relatively low height of eruptions (1–3 m) at satellite vents, indicating modest overpressure and that the pore fluid is not hydraulically connected to the source of fluid for the main vent, where eruptions can be tens of metres in height.

The Miocene carbonate deposits also trend in an east–west orientation (Carter et al. 2005) and so subsequent subsidence in the vicinity of the reefal mounds could also result in localized reactivation of preexisting east–west faults.

Almost none of the satellite vents are erupting mud; indeed, a very large number are erupting fluids consisted of methane, CO2 and a mixture of thermogenic and biogenic hydrocarbons (Mazzini et al. 2007; Sawolo et al. 2009).

As the main vent continues to remobilize mud from the Kalibeng Formation (900–1870 m), this will load the surface and more subsidence will occur, resulting in more faulting, aquifer breaches and new vent formation.

When the system evolves further an elongate caldera collapse could develop, similar to the Porong collapse identified 8 km to the NE of Lusi (Fig. 12c; Istadi et al. 2009).

This will produce a vent azimuth distribution similar to that seen at Durovdag mud volcano; indeed, the 2010 vent azimuth histogram is already exhibiting ring fault distribution, to a greater extent than in previous years.

Differences between mud volcanism in Azerbaijan and at Lusi

It should be noted that the alignments seen at Lusi may differ from those in Azerbaijan as it is almost certainly not a naturally occurring mud volcano.

The temperature of the fluids erupting at Lusi are around 70–100 o C (Tingay 2010) whereas mud volcano fluids in Azerbaijan are classically around 10–20 oC (Guliyev et al. 2000).

This may be due to relatively rapid fluid ascent rates at Lusi compared with those in Azerbaijan, where fluid flow pathways have been present for thousands of years.

Lusi had an average mud and fluid flow rate of 64 000 m3 day1 over the first 3 years (Istadi et al. 2009; Tingay 2010), differing dramatically from most naturally occurring mud volcano systems.

In Azerbaijan flow rates of only a few tens to hundreds of cubic metres per day occur, but occasionally there are eruptions that are short-lived (1–14 days) and extremely violent (100,000–1 000,000 m3 day1; Tingay 2010).

When compared with mud volcanoes in Azerbaijan, Lusi is an extremely rapidly evolving system, but this does not mean that the structural influences will differ, and ultimately regional structure will govern both areas.

Conclusions

The orientation of regional folds, faults and local metre- to kilometre-scale fractures, detachment and ring faults are the key control to the vent patterns in the mud volcanoes studied here.

The most dominant vent orientations occur subparallel to anticline axes, causing elongation of the volcanic edifice perpendicular to the regional maximum horizontal stress.

If later detachment or ring faulting forms this will overprint the original subparallel anticline crestal faulting.

Zonation of eruptive phase types also occurs, implying that there is some form of fractionation beneath the edifices in either one large chamber or a network of smaller linked chambers.

The composition of the fluids being erupted and alignment of vents along anticline axes are significant as these characteristics will dictate how the edifice itself will accrete over time.

Mud volcano alignments can occur on a range of scales from metre-scale vents that erupt along crestal fractures to the 1–4 km systems that align along anticline axes.

Lusi mud volcano is an example of how fluid flow pathways evolve through time from a localized kilometre-scale fault zone and hydrofracture system in 2006 to exploiting pre-existing pathways in the larger regional anticlinal structural control in 2009 and 2010.

This evolution is likely to continue along this trend and in a similar ring fault style to that seen in Azerbaijan, which could have major implications for the local population.

It can be predicted that the flux of fluid flow up east–west-oriented structures at Lusi will be more important than NE–SW faulting, and that as more subsidence occurs in the region more hazardous vents will form, eventually producing multiple ring fault alignments and ultimately elongate caldera collapse up to 10 km away from the main vent, as seen in paleo-collapse structures in the region.

References

Abidin, H.Z., Davies, R.J., Kusuma, M.A., Andreas, H. & Deguchi, T. 2008. Subsidence and uplift of Sidoarjo (East Java) due to the eruption of the Lusi mud volcano (2006–present). Environmental Geology, doi:10.1007/s00254008-1363-4.

Allen, M.B., Vincent, S.J., Alsop, G.I., Ismail-Zadeh, A. Flecker, R. 2003. Late Cenozoic deformation in the South Caspian Region: Effects of a rigid basement block within a collision zone. Tectonophysics, 366, 223–239, doi:10.1016/S0040-1951(03)00098-2.

Aydin, A. 2000. Fractures, faults, and hydrocarbon entrapment, migration and flow. Marine and Petroleum Geology, 17, 797–814, doi:10.1016/S02648172(00)00020-9.

Barton, C.A., Zoback, M.D. Moos, D. 1995. Fluid flow along potentially active faults in crystalline rock. Geology, 23, 683–686, doi:10.1130/00917613(1995)023,0683:FFAPAF.2.3.CO;2.

Bleacher, J.E., Glaze, L.S., et al. 2009. Spatial and alignment analyses for a field of small volcanic vents south of Pavonis Mons and implications for the Tharsis province, Mars. Journal of Volcanology and Geothermal Research, 185, 96–102, doi:10.1016/j.jvolgeores.2009.04.008.

Bonini, M. 2007. Interrelations of mud volcanism, fluid venting, and thrust– anticline folding: Examples from the external northern Apennines (Emilia-Romagna, Italy). Journal of Geophysical Research, 112, B08413, doi:10.1029/2006JB004859.

Bonini, M. 2008. Elliptical mud volcano caldera as stress indicator in an active compressional setting (Nirano, Pede-Apennine Margin, Northern Italy). Geology, 36, 131–134, doi:10.1130/G24158A.1.

Carter, D.C., Mandhiri, D., et al. 2005. Interpretation methods in the exploration of Oligocene–Miocene carbonate reservoirs, offshore Northwest Madura, Indonesia. In: Proceedings, Indonesian Petroleum Association, Thirtieth Annual Convention and Exhibition. Jakarta, Indonesia, 30 August– 1 September 2005. IPA05-G-003, 179–216.

Clark, P.J. Evans, F.C. 1954. Distance to nearest neighbour as a measure of spatial relationships in populations. Ecology, 35, 445–453, doi:10.2307/ 1931034.

Connor, C.B. 1990. Cinder cone clustering in the TranMexican volcanic belt: Implications for structural and petrologic models. Journal of Geophysical Research—Solid Earth, 95, 19395–19405, doi:10.1029/JB095iB12p19395.

Davies, R.J. Stewart, S.A. 2005. Emplacement of giant mud volcanoes in the South Caspian Basin: 3D seismic reflection imaging of their root zones. Journal of the Geological Society, London, 162, 1–4, doi:10.1144/0016764904-082.

Davies, R.J., Swarbrick, R.E., Evans, R.J. Huuse, M. 2007. Birth of a mud volcano: East Java, 29 May 2006. GSA Today, 17,4–9.

Davies, R.J., Brumm, M., Manga, M., Rubiandini, R., Swarbrick, R.E. Tingay, M. 2008. The East Java mud volcano (2006 to present): An

earthquake or drilling trigger? Earth and Planetary Science Letters, 272, 627–638.

Davies, R.J., Mathias, S.A., Swarbrick, R.E. Tingay, M.J. 2011. Probabilistic longevity estimate for the LUSI mud volcano, East Java. Journalofthe Geological Society, London, 168, 517–523, doi: 10.1144/0016-76492010-129.

Devlin, W.L., Gogswell, J., et al. 1999. South Caspian Basin: young, cool, and full of promise. GSA Today, 9, 1–9.

Eichhubl, P. Boles, J.R. 2000. Focused fluid flow along faults in the Monterey Formation, coastal California. Geological Society of America Bulletin, 112, 1667–1679, doi:10.1130/0016-7606(2000)112,1667:FFFAFI.2.0.CO;2.

Evans, R.J., Stewart, S.A. Davies, R.J. 2008. The structure and formation of mud volcano summit calderas. Journal of the Geological Society, London, 165, 769–780, doi:10.1144/0016-76492007-118.

Faulkner, D.R., Jackson, C.A.L., Lunn, R.J., Schlische, R.W., Shipton, Z.K., Wibberley, C.A.J. Withjack, M.O. 2011. A review of recent developments concerning the structure, mechanics and fluid flow properties of fault zones. Journal of Structural Geology, 32, 1557–1575, doi:10.1016/ j.jsg.2010.06.009.

Fukushima, Y., Mori, J., Hashimoto, M. Kano, Y. 2009. Subsidence associated with the LUSI mud eruption, East Java, investigated by SAR interferometry. Marine and Petroleum Geology, 26, 1740–1750, doi:10.1016/ j.marpetgeo.2009.02.001.

Geological Survey of Indonesia 1963. Geological map of Djawa and Madura. Geological Survey of Indonesia.

Guliyev, I.S., Feizullayev, A.A. Belov, I.S. 2000. All about mud volcanoes.Geology Institute, Azerbaijan Academy of Sciences, Baku, Azerbaijan.Hammer, Ø. 2009. New statistical methods for detecting point alignments. Marineand Petroleum Geology, 35, 659–666, doi:10.1016/j.cageo.2008.03.012.

Hovland, M., Hill, A. Stokes, D. 1997. The structure and geomorphology of the Dashgil mud volcano, Azerbaijan. Geomorphology, 21, 1–15, doi:10.1016/S0169-555X(97)00034-2.

Hovland, M., Fichler, C., Ruesla˚tten, H. Johnsen, H.K. 2006. Deep-rooted piercement structures in deep sedimentary basins—manifestations of super-critical water generation at depth? Journal of Geochemical Exploration, 89, 157–160, doi:10.1016/j.gexplo.2005.11.056.

Istadi, B.P., Pramono, G.H., Sumintadireja, P. Alam, S. 2009. Modeling study of growth and potential geohazard for LUSI mud volcano: East Java, Indonesia. Marine and Petroleum Geology, 26, 1724–1739, doi:10.1016/ j.marpetgeo.2009.03.006.

Iverson, R.M. 1997. The physics of debris flows. Reviews of Geophysics, 35,245– 296, doi:10.1029/97RG00426.

Jackson, J., Priestley, K., Allen, M.B. Berberian, M. 2002. Active tectonics of the South Caspian Basin. Geophysical Journal International, 148,214– 245, doi:10.1046/j.1365-246X.2002.01588.x.

Jolly, R. Lonergan, L. 2002. Mechanisms and controls on the formation of sand intrusions. Journal of the Geological Society, London, 159, 605–617, doi:10.1144/0016-764902-025.

Kopf, A. 2002. Significance of mud volcanism. Reviews of Geophysics, 40, 1005, doi:10.1029/2000RG000093.

Kusumastuti, A., Van Rensbergen, P. Warren, J.K. 2002. Seismic sequence analysis and reservoir potential of drowned Miocene carbonate platforms in the Madura Strait, East Java, Indonesia. AAPG Bulletin, 86, 213–232.

Londe, P. 1987. Malpasset dam failure. Engineering Geology, 24, 295–329.

Lutz, T.M. 1986. An analysis of the orientations of large scale crustal structures: A statistical approach based on areal distributions of pointlike features. Journal of Geophysical Research, 91, 421–434, doi:10.1029/ JB091iB01p00421.

Magara, K. 1981. Mechanisms of natural fracturing in a sedimentary basin. AAPG Bulletin, 65, 123–132.

Mazzini, A., Svensen, H., Akhmanov, G.G., Aloisi, G., Planke, S., Malthe-Sørenssen,A. Istadi, B. 2007. Triggering and dynamic evolution of the LUSI mud volcano, Indonesia. Earth and Planetary Science Letters, 261, 375–388, doi:10.1016/j.epsl.2007.07.001.

Mazzini, A., Svensen, H., Planke, S., Guliyev, I., Akhmanov, G.G.M., Fallik,

T. Banks, D. 2009. When mud volcanoes sleep: insights from seep geochemistry at the Dashgil mud volcano, Azerbaijan. Marine and Petroleum Geology, 26, 1704–1715, doi:10.1016/j.marpetgeo.2008.11.003.

Morley, C.K. 2003. Outcrop examples of mudstone intrusions from the Jerudong Anticline. In: Van Rensenbergen, P., Hillis, R.R., Maltman, A.J. Morley, C.K. (eds) Subsurface Sediment Remobilization. Geological Society, London, Special Publications, 216, 381–394.

Moss, J.L. Cartwright, J. 2010. The spatial and temporal distribution of pipe formation, offshore Nambia. Marine and Petroleum Geology, 27, 1216–1234, doi:10.1016/j.marpetgeo.2009.12.013.

Mueller, S., Llewellin, E.W. Mader, H.M. 2010. The rheology of suspensions of solid particles. Proceedings of the Royal Society of London, Series A, 466, 1201–1228, doi:10.1098/rspa.2009.0445.

Nakamura, K. 1977. Volcanoes as possible indicators of tectonic stress orienta

tion—principle and proposal. Journal of Volcanology and Geothermal Research, 2, 1–16, doi:10.1016/0377-0273(77)90012-9.

Narimanov, A.A. 1993. The petroleum systems of the South Caspian Basin. In: Dore´, A.G., Augustson, J.H., Hermanrud, C., Stewart, D.J. & Sylta, O. (eds) Basin Modeling Advances and Applications. Norwegian Petroleum Society, Special Publications, 3, 599–608.

Paulsen, T.S. Wilson, T.J. 2010. New criteria for systematic mapping and reliability assessment of monogenetic volcanic vent alignments and elongate volcanic vents for crustal stress analyses. Tectonophysics, 482, 16–28, doi:10.1016/j.tecto.2009.08.025.

Planke, S., Svensen, H., Hovland, M. Banks, D.A. 2003. Mud and fluid migration in active mud volcanoes in Azerbaijan. Geo-Marine Letters, 23, 258–268, doi:10.1007/s00367-003-0152-z.

Ramsay, J.G. Huber, M.I. 1987. The Techniques of Modern Structural Geology: Folds and Fractures. Academic Press, New York.

Roberts, K.S., Davies, R.J. Stewart, S.A. 2010. Structure of exhumed mud volcano feeder complexes, Azerbaijan. Basin Research, 22, 439–451, doi:10.1111/j.1365-2117.2009.00441.x.

Roberts, K.S., Stewart, S.A., Davies, R.J. Evans, R.J. 2011. Sector collapse of mud volcanoes, Azerbaijan. Journal of the Geological Society, London, 168, 1–12, doi:10.1144/0016-76492010-115.

Sawolo, N., Sutrioni, E., Istadi, B.P. Darmoyo, A.B. 2009. The LUSI mud volcano triggering controversy: Was it caused by drilling? Marine and Petroleum Geology, 26, 1766–1784, doi:10.1016/j.marpetgeo. 2009.04.002.

Sibson, R.H. 1990. Conditions for fault-valve behaviour. In: Knipe, R.J. & Rutter, E.H. (eds) Deformation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 15–28, doi:10.1144/ GSL.SP.1990.054.01.02.

Sibson, R.H. 1992. Implications of fault-valve behaviour for rupture nucleation and recurrence. Tectonophysics, 211, 283–293, doi:10.1016/0040-1951(92) 90065-E.

Sibson, R.H. 1996. Structural permeability of fluid-driven fault-fracture meshes. Journal of Structural Geology, 18, 1031–1042, doi:10.1016/01918141(96)00032-6.

Stewart, S.A. Davies, R.J. 2006. Structure and emplacement of mud volcano

systems in the South Caspian Basin. AAPG Bulletin, 90, 771–786.

Tanikawa, W., Sakaguchi, M., Wibowo, H.T., Shimamoto, T. Tadai, O. 2010. Fluid transport properties and estimation of overpressure at the Lusi mud volcano, east Java Basin. Engineering Geology, 116, 73–85, doi:10.1016/j.enggeo.2010.07.008.

Tingay, M. 2010. Anatomy of the ‘Lusi’ mud eruption, east Java. ASEG Extended Abstracts, 1–6, doi:10.1071/ASEG2010ab241.

Tingay, M., Hillis, R.R., Morley, C.K., Swarbrick, R.E. Drake, S.J. 2005. Present-day stress orientation in Brunei: a snapshot of ‘prograding tectonics’ in a Tertiary delta. Journal of the Geological Society, London, 162, 39–49, doi:10.1144/0016-764904-017.

Tingay, M., Heidbach, O., Davies, R.J. Swarbrick, R.E. 2008. Triggering of the Lusi mud eruption: Earthquake versus drilling initiation. Geology, 36, 639–642, doi:10.1130/G24697A.1.

Tingay, M., Morley, C.K., King, R.E., Hillis, R.R., Hall, R. Coblentz, D. 2010. The Southeast Asian Stress Map. Tectonophysics, 482, 92–104, doi:10.1016/j.tecto.2009.06.019.

Wadge, G. Cross, A. 1988. Quantitative methods for detecting aligned points: An application to the volcanic vents of the Michoacan–Guanajuato volcanic field, Mexico. Geology, 16, 815–818, doi:10.1130/0091-7613(1988)016 ,0815:QMFDAP.2.3.CO;2.

Yassir, N. 1990. The undrained shear behaviour of fine-grained sediments. In: Knipe, R.J. & Rutter, E.H. (eds) Deformation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 399–404, doi:10.1144/GSL.SP.1990.054.01.36

Yassir, N. 2003. The role of shear stress in mobilizing deep-seated mud volcanoes: geological and geomechanical evidence from Trinidad and Taiwan. In: Van Rensbergen, P., Hillis, R.R., Maltman, A.J. Morley, C.K. (eds) Subsurface Sediment Mobilization. Geological Society, London, Special Publications, 216, 461–474, doi:10.1144/GSL.SP.2003.216.01.30.

Yassir, N. Bell, J. 1996. Abnormally high fluid pressures and associated porosities and stress regimes in sedimentary basins. SPE Formation Evaluation, 11, 5–10.

Yusifov, M. Rabinowitz, P.D. 2004. Classification of mud volcanoes in the South Caspian Basin, offshore Azerbaijan. Marine and Petroleum Geology, 21, 965–975, doi:10.1016/j.marpetgeo.2004.06.002.