When mud volcanoes sleep: Insight from seep geochemistry at the Dashgil mud volcano, Azerbaijan
Ketika gunung lumpur tidur: Ditinjau dari geokimia rembesan di gunung lumpur Dashgil, Azerbaijan
A. Mazzini a,*, H. Svensen a, S. Planke a,b, I. Guliyev c, G.G. Akhmanov d, T. Fallik e, D. Banks f
a Physics of Geological Processes, University of Oslo, Box 1048, 0364 Oslo, Norway b Volcanic Basin Petroleum Research, Oslo Research Park, 0349 Oslo, Norway c Geology Institute Azerbaijan, Husein Avenue 29A, Baku, Azerbaijan d Moscow State University, Faculty of Geology, Vorobjevy Gory, Moscow 119992, Russia e Scottish Universities Environmental Research Centre, Rankine Avenue, East Kilbride, Glasgow G75 0QF, Scotland, UK f School of Earth Sciences, University of Leeds, Leeds LS2 9JT, UK
Abstract
The worlds >1500 mud volcanoes are normally in a dormant stage due to the short duration of eruptions. Their dormant stage activity is often characterized by vigorous seepage of water, gas, and petroleum.
However, the source of the fluids and the fluid–rock interactions within the mud volcano conduit remain poorly understood.
In order to investigate this type of activity, we have combined satellite images with fieldwork and extensive sampling of water and gas at seeping gryphons, pools and salsa lakes at the Dashgil mud volcano in Azerbaijan.
We find that caldera collapse faults and E–W oriented faults determine the location of the seeps. The seeping gas is dominated by methane (94.9–99.6%), with a d13C (& V-PDB) in the _43.9 to _40.4& range, consistent throughout the 12 analysed seeps.
Ethane and carbon dioxide occur in minor amounts. Seventeen samples of seeping water show a wide range in solute content and isotopic composition. Pools and salsa lakes have the highest salinities (up to 101,043 ppm Cl) and the lowest d18O(& V-SMOW) values (1–4&). The mud-rich gryphons have low salinities (<18,000 ppm Cl) and are enriched in 18O(d18O ¼4–6&).
The gas geochemistry suggests that the gases migrate to the surface from continuously leaking deep-seated reservoirs underneath the mud volcano, with minimal oxidation during migration.
However, variations in gas wetness can be ascribed to molecular fractionation during the gas rise. In contrast, the water shows seasonal variations in isotopic composition and surface evaporation is proposed as a mechanism to explain high water salinities in salsa lakes.
By contrast, gryphons have geochemical signals suggesting a deep-seated water source. These results demonstrate that the plumbing system of dormant mud volcanoes is continuously recharged from deeper sedimentary reservoirs and that a branched system of conduits exists in the shallow subsurface.
While the gas composition is consistently similar throughout the crater, the large assortment of water present reflects the type of seep (i.e. gryphons versus pools and salsa lakes) and their location within the volcano. Our data highlight the importance of a carefully planned sampling strategy when the target is water geochemistry, whereas the methane content and isotopic composition is relatively independent of the particular seep morphology.
1. Introduction
Piercement structures, such as mud volcanoes (MV), are common in many sedimentary basins with compressional settings, like in the South Caspian Basin, Trinidad, Mediterranean ridge and Indonesia (e.g. Jakubov et al.,1971; Barberet al.,1986;Cita etal.,1996;Diaet al., 1999; Isaksen et al., 2007).
The activity and behavior of mud volcanoes can be essentially classified as: eruptive (the periodicity and cyclicity of these violent events depend on the overpressure generated at depth), dormant/sleeping (interval in between eruptions that is characterized by no seepage, or microseepage or focused seepage of fluids and sediments), and extinct (no signs of erupted fluids or solid is documented in historic time).
Note that we do not consider isolated gryphons as mud volcanoes unless there is a documentation of their association with eruptions.
Mudvolcano eruptions are mainly driven by release of thermogenic methane generated at depths often greater than 10 km. The eruptions may be triggered by seismic activity and associated pressure waves or fracturing.
As the eruptions commonly last less than 2 days (e.g. Aliyev vet tal., ,2002; ;Mukhtarov et tal., ,2003), ,and as these events represent hazards, ,most studies are normally conducted during the dormant stage. .A notable exception is the Lusi mud volcano in nIndonesia, ,where boiling mud erupted in May 2006 and is still ongoing in October 2008 (Mazzini et tal., ,2007). .
The seeping methane from mud volcanoes to the atmosphere can have an impact on the carbon cycle and it has been estimated to contribute more than 6–9 9Mt yrto the atmosphere (Etiope and Milkov, ,2004). .
Water geochemistry of seeping fluids reveals complex subsurface fluid migration (Dia et al., 1999; Castrec-Rouelle et al., 2002; Planke et al., 2003; Deville et al., 2006).
These studies suggest multiple fluid sources; however, the processes ongoing during the dormant stage still remain poorly understood.
In this paper we present the results of extensive fieldwork and geochemistry of water and gas seeping from the Dashgil mud volcano.
The aim of this paper is to assess the spatial distribution of fluid composition within a mud volcano. T
his represents an important step to estimate the environmental impacts of dormant MVs and to understand the contributions of deep and shallow fluids in the subsurface plumbing system.
2. Mud volcanoes in the South Caspian Basin
The Caspian Basin is known for its hydrocarbon fields, and represents the region with the highest abundance of continental and offshore mud volcanoes (Jakubov et al., 1971).
This is mainly due to three factors: (1) rapid Quaternary infill of one of the worlds deepest sedimentary basins (up to 2.4 km/106 years); (2) diffuse methane generation in deeply buried clay units; and (3) compressional tectonics leading to anticline traps, and frequent seismicity that possibly triggers eruptions (Inan et al., 1997; Nadirov et al., 1997; Guliyev et al., 2004; Mellors et al., 2007).
In this thick and under-compacted basin, hydrocarbon generation and maturation is still ongoing, particularly in the deeply buried (8.5–11 km) Maikop Formation (Fowler et al., 2000).
Jakubov et al. (1971) documented the intimate relationship between mud volcanoes, petroleum reservoirs, and structural traps (e.g. anticlines).
The feeder channels for the mud volcanoes, nor-mally rooted below the reservoir levels (commonly at 1–3 km
depth), act as pathways for fluids during the eruptions and possibly during the dormant stage (Planke et al., 2003).
The processes at various levels of the MVs, i.e. roots, reservoir, and shallow system, still remain poorly understood.
Some dormant MVs (e.g. Dashgil, Bahar, Pirekeshkyul, Guaradag) show vigorous seepage activity, whereas others show only diffuse methane seepage (e.g., Lokbatan and Koturdag).
3 Methods
Quickbird satellite images with RGB true color view and 0.5 m resolution were acquired during January 2006 over the Dashgil mud volcano.
Fieldwork was performed in late September 2005 and in January 2006 including in situ temperature measurements and sampling of water, gas, and oil.
A total of 17 water samples and 12 gas samples were collected, covering the main areas with visible seeps at Dashgil.
Two Thales navigation Mobile Mappers GPS were used respectively as rover and fix stations to get precise coordinates of sampling stations (Table 1).
Cations were analysed using a Dionex ion chromatograph and anions were measured on a Varian Vista ICP-MS, at the University of Leeds, UK.
Gas composition was analysed at the Institute for Energy Technology, Kjeller, Norway, with a Hewlett Packard 5890 Series II GC equipped with Porabond Q column, a flame ionization detector, a thermal conductivity detector and am ethylizationunit.
The accuracy on gas composition is better than 3% (1sigma) for all compounds.
The carbon isotopic composition of the hydrocarbon gas components and CO2 were determined by a GC-IRMS system.
Repeated analyses of standards indicate that the reproducibility of d13C values is better than 1& PDB (2 sigma). The values are reported as d13C (V-PDB).
AP2003 isotope mass spectrometer and the composition with a Micro-mass Optima isotope mass spectrometer. Isotope analyses on water were performed at the SUERC, East Kilbride, Scotland. The values are reported as d& relative to V-SMOW.
4. Results
Dashgil MV is located on Cape Alyat w60 km southwest of Baku. Dashgil is situated on the crest of the Dashgil fold together with several other mud volcanoes (Koturdag, Bahar, Bahar Satellite, and Delianiz) along a E–W trend (Fig. 1).
A reconnaissance work on w10 MVs showed that Dashgil has the largest variety of seeps well accessible for sampling, and is thus an ideal target for detailed studies. Previous studies on Dashgil have focused on the morphology and the distribution of the seeps in the crater (Jakubov et al., 1971; Hovland et al., 1997). More recently Planke et al. (2003) targeted water seeps in order to identify the sources of seeping water.
However the limited number of water samples and absence of gas data in the latter study led us to a re-examination of the dormant stage activity at Dashgil with a comprehensive water and gas sampling program.
Dashgil has a smooth ‘‘pie’’ morphology with a low elevation (w90 m above sea level), and a 200 m wide central crater. The eruptive mud breccia flows cover an area of w5.5 km2.
Based on the new interpretations of s atellite images, corroborated by ground observations, we have compiled a new geological map of Dashgil (Fig. 2A,B).
Three distinct mud flows are present, but additional flows may be partly covered and eroded. Our results are supported by the eruption database, where five known eruptions have taken place since the late 1800’s.
The most recent eruption occurred in 1958; while in 2001 higher flow rates of mud were witnessed from gryphons in the north-western part of the crater (Aliyev et al., 2002).
The eruptions typically as fo ra few days, characterized by methane-driven bursts and burning gas and petroleum flares. Evidence of former fire flares is present to the south-east of the crater where sinter cones ridges are elongated in E–W direction (Fig. 3B).
Seepage of fluids and mud occurs continuously, with different dynamics, at closely spaced sites forming gryphons, pools and two salsa lakes (Hovland et al., 1997; Planke et al., 2003). In this paper we refer to these seepage structures defining them as:
· _Gryphons: positive features with a conical shape that resemble to a miniature scale of a mud volcano. Gryphons have an average height of 2–3 m and from their craters gas, water, oil and mud are continuously erupted and burst in different density and amount. These structures are normally gathered in the central part of the crater.
· _Pools: subcircular seepage features without or with low elevation that can be isolated or, more commonly, distributed at the feet of the gryphons. The size of the pools varies from few centimeters up to 1–2 m. At these sites water is continuously released together with gas and minor amount of fine grained sediment.
· _Salsa lakes: these are veritable lakes that can reach the diameter of 30 m and the depth of 10 m. Here large amount of gas and water are vigorously venting allowing these lakes to last despite the significant evaporation. Usually a limited amount of mud is seeping.
The new mapping shows that E–W oriented faults control the direction of elongated subsidence zones, as they are coinciding with the preferentially eastward trending mud breccia flows. Furthermore, the faults and the locations of subsidence affect the position of the seepages at Dashgil (Figs. 3 and 4).
The seeping dry gas is methane dominated (94.9–99.6%, Table 2) with traces of ethane (<0.4%) and CO2 (0.3– 5.1%). The d13CCH4 values are clustered between _43.9 and _40.4& (Fig. 5A) while d13CCO2 shows a wider compositional range (_23.5 and þ3.2&, Fig. 5B).
Results show that the seep gas composition is uniform in the crater, whereas higher C1/C2þand gradually decreasing d13CCO2 values are measured outside the crater (Fig. 5A,B). C3þhydrocar-bons have been detected in extremely low concentrations.
We cannot exclude the presence of Ar in the seeping gas but the methodology used does not allow distinguishing between oxygen and argon signals.
The seeping water from the pools, salsa lakes, and gryphons shows a wide range in solute content. The 17 water samples collected at Dashgil are dominated by Na and Cl, with most samples having Cl concentrations between 7000 and 30,000 ppm (Fig. 6A, Table 3).
The gryphons expel water with lower Cl contents (8738– 14,467 ppm) while the water-dominated pools and salsa lakes are at most hypersaline (as high as 101,043 ppm in one sample). One
Fig. 3. (A) Detailed satellite image of the crater region framed in Fig. 2B; (B) main seepage and geological features observed in the crater region. Profile A-A0shows a section through the crater highlighting the position of the main seepages.
Fig. 4. Examples of the main seeping features observed in Dashgil MV. (A) Gryphons field inside the crater; (B) salsa lake A where two main venting sites are observed; (C) small gryphon where dense mud continuously flows; (D) small pool where gas and water vigorously seep.
sample of gryphon water from 2002 had a chlorinity of 23,900 ppm (Planke et al., 2003).
Salsa lake A has Cl ¼13,340 ppm (versus 11,457 ppm in 2002), salsa lake B has Cl ¼28,458 ppm (versus 27,168 ppm in 2002), thus demonstrating relatively constant lake compositions in the 4 year time span.
There are, however, discrepancies in minor element composition when the 2002 and 2006 data are compared (e.g., Br, B, Sr, SO4; Table 3).
Also, trace elements like B and Li are enriched in the gryphons and pools compared to in the salsa lakes (Fig. 6B). Note that the seep water chlorinities are generally higher than in the Caspian Sea water (5650 ppm Cl), but comparable to the nearby Dashgil oil field production water from w2000 m depth (12,469 ppm Cl) and the offshore Guneshli oil field production water (11,502 ppm Cl, Planke et al., 2003). Cl/Br ratios ran between 197.5 and 297 (Fig. 6C), while Mg/Ca varies between 0.2 and 6.1.
Combined oxygen and deuterium isotope analyses of the seeping waters show that values are clustered in a region where d18O varies between 1.1 and 5.9&, and dD varies between _4 and _48& (Fig. 6D).
The most 18O enriched waters are from the gryphons. Furthermore, the water samples from gryphons, pools, and salsa lakes define three groups in d18O-dD space (Fig. 6D). All samples with dD lower than _30 & are from pools.
When comparing the September and January sampling of salsa lakes A and B, we find consistent higher Cl (by 1043 and 2994 ppm respectively) and d18O (by 1.1 and 2.3 & respectively) during the fall.
The two most 18O enriched samples were collected in September from gryphons, but other samples do not display a clear seasonal trend. A water sample from the Koturdag mud volcano was sampled during January 2006 from a frozen pool, and has a dDof _72.0& and a d18Oof _8.8&.
These values are taken as representative for the isotopic composition of water entering the Dashgil salsa lakes during winter.
5. Discussion
Besides being a well exposed and easily accessible mud volcano, there are several other aspects that make Dashgil an interesting natural laboratory to study dormant mud volcanoes.
First, the long time span since the last major eruption (50 years in 2008) makes a new event likely to occur in the near future.
Moreover recent studies suggest gradually increasing fluid pressure around the salsa A at Dashgil MV suggesting a future violent eruption (Kopf et al., 2009).
Hence it is important to monitor the seep activity. Further-more, Dashgil has a high number of seeps with varied morphology.
The newly acquired satellite images presented here, combined with previous mapping, allowed new interpretations on the flow chro-nology and identified structures (e.g. faults) controlling the seep distribution.
Our mapping shows that the Dashgil seeps are located either in the crater or along faults zones. Gryphons and associated pools are mainly confined to the crater.
Interestingly, most pools are located on the periphery of the gryphons apparently as satellite seeps. We interpret this as a result of ongoing subsidence around the continuously growing gryphons. It is suggested that this collapse causes small fractures connected with the gryphon conduit.
Over-pressured fluids rise along the gryphon conduits and the excess gas is partially expelled along these fractures. This allows the fluids to mix with shallow meteoric water, finally appearing on the surface as pools where only water and gas are seeping (Fig. 7C).
More pools are present along the caldera collapse structures and the E–W trending faults aligned along the anticline axis (Figs. 3 and 7). Asimilar trend is observed for the two salsa lakes on the eastern flank of Dashgil.
The lakes and the crater are generally aligned along the E–Woriented fault that accommodates along fold axis also Bahar, Bahar Satellite, and Delianiz MVs (Figs. 1 and 7A,B).
There are three main common sources of seeping water at dormant mud volcanoes:
(1) pore water (usually marine) entrapped during the fast burial of the source sediments;
(2) water at Dashgil may thus represent mixtures of the three groups, mineral-bound waters expelled during clay mineral diagenesis (e.g. but may also be modified by surface processes (diluted by rain and/ illitization of clay minerals usually occurring at depths 2–5 km); or concentrated by evaporation and dissolution of salt crusts). and
(3) shallow meteoric water. The geochemistry of the expelled Furthermore, seasonal variations in the isotopic composition of rain and snow can periodically alter the water composition in the seeps resulting in differences in the d18O and dD as observed comparing summer and winter sampling.
In the following section we discus the geochemistry of the three main seep systems at Dashgil: (a) gryphons; (b) pools; and (c) salsa lakes.
A deep source of water (from the dehydration of clays) feeding the gryphons is consistent with low salinities, high d18O values (from mineral–water interactions) and high Mg/Ca ratios, (typically higher than 0.5) (Savin and Epstein, 1970; Lavrushin et al., 2005).
This hypothesis is further supported by the constant expulsion of dense and clast-rich mud breccia and by constant temperature in the gryphons throughout the year.
These new results support the initial interpretations by Planke et al. (2003). The shallow meteoric fluids are most likely bypassed during the powerful rise of over-pressured mud and gas along the impermeable and defined conduits of the gryphons. This mechanism allows negligible contamination from shallow gas and water.
The salsa lakes are located outside the crater at lower altitude, hence meteoric fluids may flush these sites and affect the chemistry.
The overall high salinity recorded in the salsa lakes, and the hypersaline fluids collected in the pool AZ05A30 close to salsa A, represents a paradox since the relatively low d18O (compared to gryphons) is not indicating in situ evaporation. Two hypotheses can be formulated based on the data.
Hypothesis 1: The salsa lake water is sourced from a deep saline reservoir. This is however not supported by some of the other geochemical characteristics like low B, SO4, and Li contents relative to the gryphons.
Hypothesis 2: The salsa lakes water geochemistry is controlled by evaporation. In theory, the salsa lakes should be affected by evapo-ration since the water volume is limited and little water is physically expelled over their margins. An essentially meteoric origin of the water, possibly mixed with lesser volumes of water rising from deeper levels together with the gas, may explain the low concentrations of typical fluid–rock interaction enriched elements (Li, B).
The absence of an evaporation signature in the salsa water in comparison with the gryphons can be explained by different source ompositions etween the gryphon and salsa lake waters. Surface water, rain and snow at Dashgil will fall along the GMWL, and we regard the melt water from snow at the Koturdag summit (dDof _72.0& and a d18Oof _8.8&; Table3) as representative for the water feeding the Dashgil salsa lakes.
Evaporation of this source water will result in isotopic compositions close to that of the gryphons even though their sources are different.
The salinity of the salsa lakes may additionally be derived from In situ evaporation in seeps has been documented at hydro-dissolution of halite crusts near the summit.
These are observed thermal seeps in the Imperial Valley (CA, USA) where pools have nearby gryphons and result from in situ evaporation outside the salinities higher than neighboring gryphons Svensen et al. (2007).
At Dashgil the higher d18O and Cl recorded in the salsa lakes during (average of 287 by mass) than the gryphons (average: 215) and pools the warmer season, reflects the seasonal evaporation of the meteoric component to the seeping fluids.
Fig. 7. (A) NW–SE section of Dashgil MV. Vertical axis not to scale. The marked locations in and around the crater represent respectively. (1) The gryphon field inside the crater; (2) diffuse seepage along the outer fault margin; and (3) salsa lakes. Symbols in the stratigraphy: PT ¼Productive Serie-sandstones; S ¼Sarmatian-shales; TC ¼Tarkan–Chokrak-shales/sandstones; M ¼Maikop-shales; (B) magnification of area framed in image A highlighting the collapse controlled by faults that act as preferential pathways for deeper fluids seepage. Seepages outside the crater show stronger d13CCO2 depletion and higher amount of CH4. At large salsa lakes deep fluids and shallow meteoric fluids converge and mix; (C) interpreted plumbing system of gryphon-pool complex based on field observations and gas/water analyses. Overburden of the gryphons causes collapse and fractures through which the deep fluids migrate, mixing with shallow meteoric waters. At gryphon sites evaporation is likely to have a limited influence as gryphons contain dense mud and differ morphologically (e.g. from pools) ‘‘isolating’’ the fluids inside the crater and in the internal chambers. d18O values support a confined seepage of fluids through the feeder channel allowing a bypass through the intervals charged with meteoric fluids.
5.2.3. Pools
Unlike the gryphons whose conduits are self-sealing by the a convective circulation of fluids is expected as also observed in the dense mud, the pathways feeding the pools are shallower and salsas in Trinidad (Deville and Guerlais, 2009). allow meteoric fluids to mix with the deeper fluids (Fig. 7C). This mixing is causing lower d18O values.
As seepage at pools occurs with different rates and vigor they tend to have a broad variety in composition. One of the pools contains hypersaline brine (pool near Salsa A) without being more 18O-enriched than the salsa lakes, which rules out evaporation as an explanation for the high solute content in this sample. It remains uncertain how common hypersaline seeps are at Dashgil and how the high salinity is derived.
The water analyses reveal a large spread of results in the isotopic and solute composition. This spread is remarkable considering that the samples described are collected from a single mud volcano.
This highlights the difficulty in making broad statements and conclu-sions about water sources when interpreting data from few sampling stations, and some of the Planke et al. (2003) interpre-tations have been modified with the new data.
This challenge is illustrated by results from e.g. Lavrushin et al. (2005) who draw regional conclusions based on a limited number of water samples collected from a large number of mud volcanoes in the Taman Peninsula, Georgia, Azerbaijan, and Turkmenistan. When comparing the variations in water d18O-dD of the Dashgil seeps and the Lavrushin et al. (2005) data, we observe that the Dashgil water geochemistry covers most of the range in water geochemistry from mud volcanoes from the mentioned areas.
This emphasizes the need for sample collection throughout the year, the need of sampling a wide range of seep structures, and the difficulty in making general statements about the controlling factors for water geochemistry between different geographic areas
As methane is typically the main gas seeping at mudvolcanoes, its purity and isotopic composition may be used to elucidate the activity in the feeder channel.
For example, previous studies on offshore mud volcanoes show that low d13CCH4 values and low amount of C2þhomologues can be interpreted as evidence of a significant input of biogenic methane produced in the shallow subsurface.
This suggests a reduced flux of deep fluids reaching the surface and a rather diffuse shallow microbial activity (e.g. Mazzini et al., 2004).
In contrast, moderate 13C depletion and higher amount of C2þhomologues are commonly interpreted as thermogenic deep-rooted gas that rises rapidly towards the surface (e.g. Blinova et al., 2003). See Etiope et al. (2009b) for a more extensive explanation of the global statistics of gas seeping from mud volcanoes worldwide.
Comparison between gas sampled from the Dashgil MV and that from the neighboring oil fields (Katz et al., 2002)gives insight about the mechanisms of gas migration. Katz et al. (2002) show that numerous of the reservoir gases from the South Caspian were not generated in situ and have been altered and/or represent mixed source hydrocarbons.
The d13CCH4 isotopic signatures of the gas seep-ing at Dashgil are similar to those from deeper oil field gas (Table 2). Like also pointed outby Katzetal.(2002), our result ssuggest that most of the deeper-sited thermogenic mature (?) gas migrates from depth greater than 3 km and that there is a negligible contribution from shallow biogenic methane. However the d13CCH4 of Dashgil oil field is slightly lower than the neighboring reservoirs (Table 2) suggesting a small biogenic input. Similarly to what pointed out by Etiope et al. (2009a) our data also suggests that isotopic fractionation related to microbial oxidation is not significant. Yet the seeping gases (Fig. 5A, Table 2) show dramatically lower amounts of the C2 component and presence of C3þonly in some cases and anyhow in negligible amounts (i.e.concentrationsbelow0.01%).
This could be interpreted as gradual reduction of the reservoir gas wetness due to microbial alteration (e.g. James and Burns, 1984). Alternatively the lack of significant C2þconcentration has been interpreted as a result of molecular fractionation not related to microbial activity but to the mechanical and solubility processes during fluid migration (e.g. Deville et al., 2003; Mizobe et al., 2007; Etiope et al., 2009a). Interestingly the comparison the d13C of methane and ethane (Fig. 5C) shows a linear trend that can be explained by increasing source rock maturation in the direction of increasing 13C.
The data also shows differences between the gas samples from the crater and the gas seeping on the edge or outsidet he main crater (Fig. 3). Combining C1 and C2 isotope values (Fig. 5D) we observe that the difference d13CC1– d13CC2 remains constant when comparing the values of Dashgil oil field (_17.9&) and the seeping gas (_16.7& of main cluster).
This data supports our interpretation based on the variations on C1þhomologues, that is that the isotopic frac-tionation is not particularly high at this location. d13CCO2 data show that seeps located outside the Dashgil crater (Fig. 5B), and thus away from the main feeder channel, have the most depleted values with higher C1/C2þ.
This could be due to a localized production of light CO2 via oxidation of methane occurs at shallow depths along the faults in the outskirts of the crater (Fig. 7B), or, most probably, produced by decomposition of organic matter (e.g. Jenden et al., 1993; Hunt, 1996). However other studies have suggested that the d13CCO2 can have significant variability over time, possibly related with changing water solution–dissolution processes (Pallasser, 2000; Etiope et al., 2009b).
The presence of samples with d13CCO2 as high as þ3.2& could suggest a contribution of anaerobic biodegradation of petroleum (e.g. Pallasser, 2000). Moreover Etiope et al. (2009b) pointed out how the combined CO2 and C2þisotopic analyses are a powerful tool to confirm the presence of such biodegradation.
These authors showed that in other mud volcanoes in Azerbaijan (Airantehian, Bozdag Gobu, Shikazdgirly) moderately depleted C2 and C3 isotopes actually support biodegradation and are consistent with positive d13CCO2 values.
In our Dashgil study case, the absence of isotopic data of heavier hydrocarbons (e.g. C3þ) does not allow such more complete approach.
5.4. Modeling the post eruption seepage
While the water results give insights mainly a bout the shallow plumbing system, gas data can be used to understand the largescale seepage mechanism of the volcano.
Overall our gas analyses show consistent deep fluids seeping at Dashgil. Two alternative hypotheses can be suggested to explain the seepage along the conduit (Fig. 8).
The first hypothesis assumes a gradual faulting and collapse of the crater.
This collapse facilitates the connection between the roots and the crater of the volcano allowing the continuous release of deep fluids during the dormant stage.
A second alternative suggests that a gradual sagging of the fluidized sediments occurs at the end of each eruption event. This mechanism allows the remaining fluids in the fluidized sediments to be squeezed out at seepage sites. This combination of collapse and compaction of the
6. Conclusions
Detailed mapping of the dormant Dashgil mud volcano shows that at least three separate mud breccia flow units are present on the low relief flanks.
The flows can be traced to the present day summit crater, where vigorous gas seepage is taken place through gryphons and pools.
In addition, many other seeps are associated with faults, emphasizing the importance of post eruptive subsidence for where seeps are located. Since the last main eruption occurred in 1958 the subsidence close to the summit must be active and the seep plumbing system stationary.
Sampling and geochemical analyses of 12 individual gas seeps and 17 water seeps have resulted in new interpretations about the controlling factors for seep geochemistry and the subsurface plumbing system during the dormant stage:
· _Gas analyses show a similar composition of the seepages inside the Dashgil crater, while gas expelled along the caldera collapse faults have higher C1/C2þ and corresponding decreasing d13CCO2. Here, a more prominent decomposition of organic matter is suggested. Overall a deep source of methane is detected throughout the sampled stations.
· _The water geochemistry highlights that different water sources and reactions occur at gryphons, pools, and salsa lakes. Gryphons have a signature of deep-rising fluids, while pools and salsa lakes show imprint of meteoric fluids and a solute content increased by in situ evaporation.
· _Based on combined field and geochemical observations, we suggest that an interconnected and intricate plumbing system is present in the near subsurface of the mud volcano.
· _The high variety in water composition between narrowly spaced seeps emphasizes that representative end-member data is difficult to obtain without a full-scale sampling program. This has implications for mud volcanoes that are difficult to monitor and sample (particularly for offshore mudvolcanoes),especially if only a fraction of the seeps have been sampled.
Acknowledgements
M. Efendiyeva, C. Aliyev, O. Barvalina, E. Poludetkina, M. Hov-land and H. Rueslåtten, GIA, GOC, Statoil Hydro are thanked for the fruitful discussions and their support during the fieldworks. We gratefully acknowledge support from a Center of Excellence grant and a PETROMAKS grant to Anders Malthe–Sorenssen from the Norwegian Research Council. The paper greatly benefited from the input of G. Etiope and two anonymous reviewers.
References
Aliyev, A., Guliyev, I.S., Belov, I.S., 2002. Catalogue of Recorded Eruptions of Mud Volcanoes of Azerbaijan. Nafta Press, Baku, 87 pp.
Barber, A.J., Tjokosapoetro, S., Charlton, T.R., 1986. Mud volcanoes, shale diapirs, wrench faults and melanges in accretionary complexes, eastern Indonesia. American Association of Petroleum Geologists Bulletin 70, 1729–1741.
Blinova, V., Ivanov, M.K., Bohrmann, G., 2003. Hydrocarbon gases in deposits from mud volcanoes in the Sorokin Trough, north-eastern Black Sea. Geo-Marine Letters 23 (3–4), 250–257.
Castrec-Rouelle, M., Bourles, D.L., Boulegue, J., Dia, A.N., 2002. Beryllium geochemistry constraints on the hydraulic behavior of mud volcanoes: the Trinidad Island case. Earth and Planetary Science Letters 203 (3–4), 957–966.
Cita, M.B., Ivanov, M.K., Woodside, J.M., 1996. The Mediterranean Ridge diapiric belt. Special Issue. Marine Geology 132, 273. Deville, E., Guerlais, S.H., 2009. Cyclic activity of mud volcanoes: evidences from Trinidad (SE Caribbean). Marine & Petroleum Geology 26, 1681–1691.
Deville, E., Battani, A., Griboulard, R., Guerlais, S.H., Herbin, J.P., Houzay, J.P., Muller, C., Prinzhofer, A., 2003. Mud volcanism origin and processes. New insights from Trinidad and the Barbados Prism. In: Van Rensbergen, P., Hillis, R.R., Maltman, A.J., Morley, C. (Eds.), Surface Sediment Mobilization. Special Publication of the Geological Society, London, pp. 475–490.
Deville, E., Guerlais, S.-H., Callec, Y., Griboulard, R., Huyghe, P., Lallemant, S., Mascle, A., Noble, M., Schmitz, J., 2006. Liquefied vs stratified sediment mobilization processes: insight from the South of the Barbados accretionary prism. Tectonophysics 428 (1–4), 33–47.
Dia, A.N., Castrec-Rouelle, M., Boulegue, J., Comeau, P., 1999. Trinidad mud volca-noes: where do the expelled fluids come from? Geochimica et Cosmochimica Acta 63 (7–8), 1023–1038.
Etiope, G., Milkov, A., 2004. A new estimate of global methane flux from onshore and shallow submarine mud volcanoes to the atmosphere. Environmental Geology 46, 997–1002.
Etiope, G., Feyzullayev, A., Baciu, C.L., 2009a. Terrestrial methane seeps and mud volcanoes: aglobal perspective of gas origin. Marine and Petroleum Geology 26, 333–344.
Etiope, G., Feyzullayev, A., Milkov, A., Waseda, A., Mizobe, K.C.H.S., 2009b. Evidence of subsurface anaerobic biodegradation of hydrocarbons and potential secondary methanogenesis in terrestrial mud volcanoes. Marine & Petroleum Geology 26, 1692–1703.
Fowler, S.R., Mildenhall, J., Zalova, S., Riley, G., Elsley, G., Desplanques, A., Guliyev, I., 2000. Mud volcanoes and structural development on Shah Deniz. Journal of Petroleum Science Engineering 28, 189–206.
Guliyev, I.S., Huseynov, D.A., Feizullayev, A.A., 2004. Fluids of mud volcanoes in the Southern Caspian sedimentary basin: geochemistry and sources in light of new data on the carbon, hydrogen, and oxygen isotopic compositions. Geochemistry International 42 (7), 688–695.
Hovland, M., Hill, A., Stokes, D., 1997. The structure and geomorphology of the Dashgil mud volcano, Azerbaijan. Geomorphology 21 (1), 1–15.
Hunt, J.M., 1996. Petroleum Geochemistry and Geology, second ed. Freeman, 743 pp.
Inan, S., Namik Yalcin, M., Guliyev, I.S., Kuliev, K., Feizullayev, A.A., 1997. Deep petroleum occurrences in the Lower Kura Depression, South Caspian Basin, Azerbaijan: an organic geochemical and basin modelling study. Marine & Petroleum Geology 14, 731–762.
Isaksen, G.H., Aliyev, A., Barboza, S.A., Plus, D., Guliev, I.S., 2007. Regional evaluation of source rock in Azerbaijan from the geochemistry of organic-rich rocks in mud-volcano ejecta. In: Yilmaz, P.O., Isaksen, G.H. (Eds.), Oil and Gas of the Greater Caspian Area, 55. AAPG Studies in Geology, pp. 51–64.
Jakubov, A.A., Ali-Zade, A.A., Zeinalov, M.M., 1971. Mud Volcanoes of the Azerbaijan SSR. Atlas (in Russian), Azerbaijan Academy of Sciences, Baku.
James, A.T., Burns, B.J., 1984. Microbial alteration of subsurface natural gas accumulations. America Association Petroleum Geologist Bulletin 68 (8), 957–960.
Jenden, P.D., Hilton, D.R., Kaplan, I.R., Craig, H., 1993. Abiogenic hydrocarbons and mantle helium in oil and gas fields. In: Howell, D.G. (Ed.), The Future of Energy Gases. (US Geological Survey Professional Paper 1570). United States Govern-ment Printing Office, Washington, pp. 31–56.
Katz, B.J., Narimanov, A., Huseinzadeh, R., 2002. Significance of microbial processes in gases of the South Caspian basin. Marine and Petroleum Geology 19 (6), 783–796.
Kopf, A.J., Stegmann, S., Delisle, G., Panahi, B., Aliyev, C.S., Guliyev, I., 2009. In situ cone penetration tests at the active Dashgil mud volcano, Azerbaijan: evidence for excess fluid pressure, updoming, and possible future violent eruption. Marine & Petroleum Geology 26, 1716–1723.
Lavrushin, V., Dubinina, E., Avdeenko, A., 2005. Isotopic composition of oxygen and hydrogen in mud-volcanic waters from Taman (Russia) and Kakhetia (Eastern Georgia). Lithology and Mineral Resources 40, 123–137.
Mazzini, A., Ivanov, M.K., Parnell, J., Stadnitskaia, A., Cronin, B.T., Poludetkina, E., Mazurenko, L., van Weering, T.C.E., 2004. Methane-related authigenic carbon-ates from the Black Sea: geochemical characterisation and relation to seeping fluids. Marine Geology 212 (1–4), 153–181.
Mazzini, A., Svensen, H., Akhmanov, G.G., Aloisi, G., Planke, S., Malthe-Sorenssen, A., Istadi, B., 2007. Triggering and dynamic evolution of the LUSI mud volcano, Indonesia. Earth and Planetary Science Letters 261 (3–4), 375–388.
Mellors, R., Kilb, D., Aliyev, A., Gasanov, A., Yetirmishli, G., 2007. Correlations between earthquakes and large mud volcano eruptions. Journal of Geophysical Research 112, B04304.
Mizobe, M., Waseda, A., Tanaka, K., 2007. Geochemical characteristics of natural gases from mud volcanoes in Tokamachi City, Niigata Prefecture. Japan. Geo-science Union Meeting, May 2007, Abstract Y239–P011.
Mukhtarov, A.S.h., Kadirov, F.A., Guliyev, I.S., Feyzullayev, A., Lerche, I., 2003. Temperature evolution in the Lokbatan mud volcano crater (Azerbaijan) after the eruption of 25 October 2001. Energy Exploration & Exploitation 21 (3), 187–207.
Nadirov, R.S., Bagirov, E., Tagiyev, M., 1997. Flexural plate subsidence, sedimentation rates, and structural development of the super-deep South Caspian Basin. Marine & Petroleum Geology 14, 383–400.
Pallasser, R.J., 2000. Recognising biodegradation in gas/oil accumulations through the d13C compositions of gas components. Organic Geochemistry 31, 1363–1373.
Planke, S., Svensen, H., Hovland, M., Banks, D., Jamtveit, B., 2003. Mud and fluid migration in active mud volcanoes in Azerbaijan. Geo-Marine Letters 23, 258–268.
Savin, S.M., Epstein, S., 1970. The oxygen and hydrogen isotope geochemistry of clay minerals. Geochimica et Cosmochimica Acta 34 (1), 25–42.
Svensen, H., Karlsen, D.A., Sturz, A., Backer-Owe, K., Banks, D., Planke, S., 2007. Processes controlling water and hydrocarbon composition in seeps from the Salton Sea geothermal system, California, USA. Geology 35 (1), 85–88.