Summary of the 5th Anniversary of LUSI: A Symposium sponsored by Humanitus

Prof. Dr. Stephen A. Miller Geodynamics/Geophysics, University of Bonn, Germany

The 5th year anniversary of the LUSI mud eruption brought together scientists, engineers, and government officials to discuss the current state of understanding of this system and propose future research and mitigation strategies. Despite continued interest from the public about what triggered LUSI, the scientists focused on understanding this system as it now exists and did not dwell on its initiation. The triggering debate may re-emerge in the future, however, because of the implications of the discoveries and research presented.

The broad range of topics covered by the 13 presentations, spanning the thousand-kilometer (tectonic) scale to the nanometer scale of LUSI mud particles, made it clear that significant progress has been made in understanding this system. Much more work is needed, but an important common thread was found in almost all talks that sets a solid foundation for future studies: Namely, LUSI appears to be a new-born hydro-thermal system, a geologic rarity that offers many possibilities for both assessing its future and providing a platform for basic research in hydro-thermal systems and links to the magmatic plumbing of arc volcanism.

In this summary, I attempt to synthesize what is currently known into a coherent story and offer my own personal assessment on the state of LUSI.

Probably the most significant new finding about the LUSI system was offered by Mazzini from his geochemical studies. His studies provide strong evidence that LUSI is a newborn hydrothermal system, thus suggesting a direct link to the nearby Arjuno-Welirang volcanic complex. If these ideas are verified in (essential) future studies, then this has major implications for the future behavior of LUSI, possible geothermal energy exploitation, and likely some fundamental advances in our understanding of hydro-thermal systems and volcanic plumbing. The chemical analyses by Mazzini of the erupting mud and waters show that identifying LUSI as a mud volcano is probably a misnomer because  properties  of  this  system are  quite  different  than  those  typically called mud volcanoes. Traditional mud volcanoes are cold, with low geothermal gradients, and infrequent in eruptions. LUSI, on the other hand, has very high geothermal gradients, is CO2-dominated, and is a very long-   lasting and geyser-like system. Mazzinni suggests LUSI be re-designated as a hydro-thermal system, placing it with other similar systems such as Cosos, CA., Salton Sea, CA., Yellostone, and Iceland. This change in identity is significant because it implies that LUSI’s plumbing system is directly linked with the adjacent volcanic arc, a large step forward in our understanding of how LUSI might have been triggered and where it is likely to go from here. Mazzini reports from geochemical analysis that the initial dominant gas from LUSI was microbial methane (CH4) of biogenic origin, which gave way to thermo-genic  origin  after  2006.  Further  analyses  of  helium  and  carbon isotopes are consistent with a high temperature reaction indicating that the fluid source is mantle-derived, implying a fluid pathway from a source much deeper history than assumed to date. Mazzini suggests that the source of fluid is at a depth of about 4400 m (the Ngimbong source rock), which is consi stent with the need for an external fluid source and which also absolves the Kujung limestone formation of guilt as the external fluid source. Mazzini proposes a new   scenario   for   this   system   whereby   magmatic   intrusion   from  the neighboring Arjuno-Welingong provides a direct heat source that promotes kerogen decarbonization and dehydration (the fluid source), and is consistent with the observed high heat flow and CO2-dominated system with the presence of mantle helium. An attractive aspect of this model is that the observations  suggest  a  switch  from  high-pressure,  high-temperature conditions to high temperature, low-pressure conditions, implying rapid transport over large distances and thus fracture-dominated flow. The results of Mazzini allow an upscaling of the problem to a larger context; namely, the role played by the adjacent volcanic arc.

Observational support for this hypothesis is that LUSI changed her behavior in the last year to that of a geyser system, with semi-periodic eruption cycles and the eruption of mostly water with each cycle. Geyser behavior, and the reduction in extruded mud volumes, points to a (probably vast) deep source of fluid and some structural stability of the mud layer. LUSI is still dominated by eruptions in the central crater, but subsidiary vents are also identified, emitting smaller eruptions and with their own cycles. Documenting and quantifying these cycles would be a very important dataset for constraining the underground plumbing system, and the potential for documenting this was beautifully demonstrated by Lorc and colleagues with a time-lapse sequence recording thermal and visual images from the new observation tower overlooking LUSI.

Further observational evidence for a deep fluid source was described by Tanikawa with his important finding that the fluids are Lithium-enriched. Tanikawa showed that the Lithium concentrations are near that of the Uyuni salt flats in Bolivia, where 50-70% of the world’s supply of this important element is found. Not noted in the talk, but speculated here, is that the Lithium   rich brines in the Bolivian salt flats may have also resulted from a tectonic- scale hydrothermal system, which (if true) would further corroborate Mazzini’s suggestion that LUSI be designated a hydro-thermal system instead of a mud volcano. The simple explanation of Lithium-enriched fluids is that at low temperature, lithium is absorbed into clay minerals, while under hydrothermal conditions thermal desorption releases Li+ into the hydrothermal fluids. In an appeal for turning LUSI from disaster to opportunity, Tanikawa proposes mining the Lithium from the fluids for economic gain to the community.

Besides the lithium story, Tanikawa et. al. also presented a very physics- based approach to this addressing this problem, including laboratory experiments and modeling to try and understand the mechanisms by which overpressures developed in this sedimentary basin. The sedimentary history and  lithologic  properties  are  important  because  they  control  the  flow properties of the system, both in terms of the fluid overpressure at depth and the permeability structure which had trapped the overpressure. Tanikawa showed an intriguing plot where the lithology of LUSI deviates substantially from the typical compaction curves (that is, the loss of porosity with depth) of sedimentary basins. He showed that the Kalberg formation, presumably the source of the mud, has significantly more porosity than expected from a compacting system, and that thermal expansion of the entrained fluids results in a fluid pressure close to that of the rock pressure. Such a condition means that the rock in this layer is mechanically weak (or even zero strength in shear if fluid pressure equals rock pressure), and that even a minor input of external fluid could destabilize or trigger the system. Tanikawa showed that mud viscosity decreases substantially as a function of the solid fraction, with the effect that an influx of fluid would reduce the solid fraction and allow the mud to flow much more easily at the consequently much lower viscosity.

(Author note: The identification of a high porosity-overpressured region of the Kalberg formation also provides an avenue for study not yet proposed. Namely, this condition is amenable to a scenario for the development of porosity waves; that is, solitary waves of fluid-filled porosity that propagate towards the surface at a velocity that depends primarily on the size of the porosity perturbation and the bulk viscosity of the entraining medium. If porosity waves are a viable triggering mechanism, then this may provide an alternative, or complementary mechanism for the initiation of LUSI.)

The equations governing flow and pressure diffusion are dominated by the permeability of the medium, the viscosity of the fluid, and the rate at which new fluid (either external or from local dehydration) is brought into the system. Reports from the laboratory experiments put good constraints on these dominant properties. Tanikawa reaches the same conclusions as Mazzini is the sense that an external source of fluid is needed to explain the water/mud ratios, and that the fluids must be deep because of the high concentrations of Lithium.

In other geochemical studies, Hilairy Hartnett and colleagues used geochemical results of LUSI fluids for comparison with other mud volcanoes and surface waters to trace the origin of the fluids driving LUSI.   Their conclusion that ‘it is complicated’ nonetheless provided some important clues for the origin of the fluids, and seems consistent with the Mazzini model of a deep origin. Without getting into details, the essence of their finding is that the fluids are hot, quite different from sea water, and from hydrogen and oxygen isotopes, very similar to the older mud volcanoes studied. This indicates that whatever fluids are driving LUSI are probably the same as that driving ancient volcanoes, with a link to magmatic processes a viable culprit. The results of clay mineralogy show that LUSI has a higher percentage of clay minerals that the older Kalang Anyar mud volcano (which has substantially more quartz), but the implications of this are not clear. Hartnett et. al., propose from trace element  analyses  that  the  clays  in  LUSI  are  altered  volcanics.  This observation   places   another   morsel   on   the   table   about   the   complex interactions between volcanism, deep heat sources, and mud volcanism. Additional work is needed to model the high temperature water-rock interactions, which is also suggested by the high temperature desporption of Lithium reported by Takinawa, and the deep source proposed by Mazzini.

Guntoro reports that the presence of Deuterium provides yet more evidence for a magmatic origin for the LUSI fluids, while the 70% water and 30% mud (and  very  high  flow  rates)  are  atypical  for  conventional  mud  volcanoes. Guntoro  provided  a  comprehensive  summary  of   mud  volcanism,  and discussed LUSI’s location in the proper larger scale context of Indonesian (Java) tectonics. The contradicting observations of H2S gas likely derived from the Kujung formation, while that formation itself was not penetrated by the BP- 1 well, suggests that the Kujung formation was penetrated from below (not from above) by a deep source of fluid. All of the observations, from gravity anomalies to fluid mud ratios to the presence of Deuterium lead Guntoro to conclude that LUSI is a result of typcial sedimentation processes in back-arc basins. In back-arc basins, fast sedimentation rates trap and bury fluids that increase in pressure as additional sediments are supplied above. The observation that the mud and fluid come from two different sources implies that a viable triggering mechanism is the influx of fluids from below and into the the Kugung formation. In this scenario, the fluids up H2S along the way, which then drives the mud from the Upper Kaliberg towards the surface.

Re-designating LUSI as a hydrothermal system and not a mud volcano was not the only suggestion about current terminology for this system. Tingay and collegues suggest that the oft-named Kujung carbonate formation, the alleged   source  of external water driving LUSI, is 23 Ma, whereas their proposal as the most likely candidate for a highly over-pressured source is the Porong-1 formation, which is about 7Ma younger than Kujung. (The conflicting source of external fluid, identified as a deep source by Mazzini/Tanikawa/Guntoro, and the Porong-1 formation by Tingay, needs to be further clarified in future studies). Another significant change is Tingay’s suggestion that the volcano- clastic sands just above the carbonates should be re-named to reflect that they are much more likely to be volcanics of igneous origin. This claim arises from analyses of the cuttings from the drill core, which consisted of andesite, dacite, and welded tuffs. The original designation as volcanic sands, he suggests, arose simply because they looked like sand. They looked like sand, he says, because it was so difficult to drill into that the initially hard and competent igneous rock was ground down to sand-sized particles. This re- designation is not only semantics because volcanic sands are very permeable and the description of flow through them is described by Darcian flow through a porous (20-25% porosity) medium. Igneous volcanics, however, have much lower porosity (<9%), and fluid flow through them is fracture dominated. That is, fracture-dominated flow would involve localized flow through highly permeable fractures in an otherwise no-flow solid, an important distinction for describing and modeling of high-pressure driven fluid flow. Tingay also used geo-mechanical arguments to show that the main NW -SE direction of the primary fault through which LUSI erupts is fully consistent with that predicted by Andersonian theory for strike-slip faulting in a predominantly NS-directed maximum horizontal tectonic stress field. Furthermore, the observed later development of a NW -SE trending system is also consistent for slip on the antithetic strike-slip faults subjected to the same far-field loading. Identifying and constraining the orientation of the maximum horizontal stress is essential for any further numerical modeling studies of the hydro-mechanics of this system. The important questions to answer, according to Tingay, are 1) where is the source of the water? 2) what is the detailed structure of the subsurface fracture network providing the fluid pathways, and 3) how has the structure evolved, and how is it evolving now? He suggests future studies to answer these questions, including 1) 3D/4D seismic studies (that is, a 3D seismic study in different time windows); 2) magnetotellurics to identify fluids and their connected   pathways   at   depth;   3)   additional   geochemical   studies;   4) monitoring wells to get better constraints of two very important physical properties (permeability and porosity) that control flow, and 5) the deployment of tiltmeters that record deformation at the surface.

Kadurin et. al. use a geophysical network (called a polygon) to detect seismicity and from which they can construct 3D structures. This team has previously constructed a 3D image of the subsurface of LUSI by interpreting seismic reflection profiles, and identified a number of faults, diapirs, and what they claim are ancient mud structures at depth (with the implication that LUSI   is a geologic structure that has erupted in the past). The polygon they propose to set up for LUSI would provide much-needed information about the structures at depth, and similar polygons installed in Russia and around Istanbul demonstrates their capabilities in leading such an initiative. If this system is installed in conjunction with the proposed 3D seismic campaign, then a wealth of data will be obtained and much learned about the inner workings of this system.

Future attempts to model LUSI will also require constraints on LUSI’s thermal properties. Lorc and colleagues showed results of their thermal studies, and it was (rightfully) made clear that more observations are needed of geyser frequency and the associated thermal field. Thermal imagery, with a planned installation in August 2011, will be instrumental in quantifying the temperature field of the erupting water, and with it a measure of flow rates and thermal structure. A time-lapse movie showed clearly that continuous monitoring via video or thermal imaging can be processed to produce a time-series of eruption frequency, with particular attention to the geyser response to nearby or distant earthquakes. This data would be critically important for determining any links between the behavior of LUSI and varying degrees of ground shaking.

The response of LUSI to ground shaking has always been at the core of the debate between a drilling trigger or a Yogyakarta earthquake trigger.  The work of Amanda Clarke and colleagues showed that both Merapi and Semeru volcanoes responded to the Yogyakarta earthquake with an increase in heat flux in  the  days  following the  earthquake.  It  is  clear  that  searching (and finding) remotely sensed thermal perturbations to ground motion and other tectonic processes is a powerful approach, and will certainly result in many interesting future observations. She pointed out (as is also often pointed out by those calling for a drilling trigger for LUSI), that the static stress changes are miniscule, while the dynamic stress changes, although larger, are still only on the order of tens of kPa. Nevertheless, the correlation between the response of the open systems of Merapi and Semeru to the Yogyakarta earthquake is strong, so a physics-based explanation is needed but not yet identified.

Clarke et. al. used a different study to try and relate LUSI to other systems triggered by distant earthquakes. She presented results from their study of distant triggered seismicity in response to the M6.4 Oaxaca, Mexico earthquake. For its size and location, this earthquake was unspectacular, yet many (small) earthquakes were triggered in North America (a few thousand km to the north). Interestingly, the best correlation between the Oaxaca earthquake and triggered seismicity was with young volcanism, which are hydrothermally altered  and  fluid-rich; precisely the description  of  LUSI as   concluded by Mazzini. The results from the Oaxaca study are consistent with the first observed triggered seismicity study following the 1992 Landers earthquake, which showed that triggered seismicity correlated strongly with geothermal and hydrothermal systems (e.g. Long Valley, CA, and Cosos, CA). These are very interesting and important results, and which should be further pursued, both in understanding the physics behind this correlation, and future targeted observations to determine the ubiquity of such correlations.

These new findings can be put into the larger context of the regional tectonics, and an extensive report of the tectonics of the region presented by Sukandar Asikin put LUSI into the context to which it belongs: Namely, a natural system as a result of the complex geodynamics of the region. This is not to say this provides evidence one way or the other about what triggered LUSI, only that LUSI exists for purely natural reasons. The details of the presentation are too extensive to summarize here, but the thrust of this talk can be distilled as follows: Subduction of the Australian Plate beneath the Sunda plate produces uplift, volcanism, and back-arc spreading (where LUSI resides). Spreading results in subsidence in the back-arc, into which eroding sediments quickly pile into the basin. If sedimentation is faster than the entrained fluids can escape, buried layers of overpressured and underconsolidated form at depth. As  they  are  buried  and  cooked  through  the  geotherm,  they  transform to natural gas, and at higher temperatures and pressures, to kerogen and finally to oil. This is why the oil companies were prospecting there in the first place, because typically mud volcanoes are a good indicator for trapped gas and oil because they mark regions of the sedimentation processes amenable to fossil fuel production. The source of mud that is LUSI is one of the overpressured and underconsolitated layers that result from large-scale and long-term tectonic processes.

Evidence for natural mud volcanoes in the region is well-established and supports (in part) the idea that this is a natural system. Using a different tack, A. Satyana put LUSI into a historical context through the lens of empires and their rise and fall. He presented possible connections to the fall of ancient Javanese kingdoms and neighboring erupting volcanoes. Although fascinating speculation in its own right, it is speculation, and the only direct relevance to the current LUSI eruption is that, like the neighboring geologic evidence of ancient mud volcanism, large mud volcanoes are ubiquitous in the region. The interesting aspect of this is that historical eurptions occurred spontaneously (without drilling), lending evidence that LUSI would have ultimately happened with or without an assist from local drilling.

Numerical modelling of LUSI to date is still fairly rudimentary, with a lot of room for advances and more sophisticated models. Modeling studies from two presentations by Davies et. al. and Rudolph et. al., aimed to assess from a   probabilistic standpoint how long LUSI might erupt. Both models used a technique called Monte-Carlo simulations, where you basically roll dice for assigning material properties to the unknowns, constrained within reasonable limits. These types of studies are useful for getting probabilities because thousands of simulations are performed with some outcomes more likely than others. The model of Davies assumes a model of draining he Kujing limestone fluid source through a vertical pipe, and uses the observed volume at the surface as a rejection/acceptance criterion for including the model result into the statistical analysis. Personally, I found it disconcerting that less than 400 simulations out of 10,000 model realizations met the acceptance criterion of matching extruded volumes. To me this indicates that the conceptual model is flawed, but nevertheless, the results of this study predicted that about 26 years is the answer with the highest probability. Davies pointed out that this is probably an overestimate, and future observations will help constrain future model realizations.

Rudolph et. al. presented results from a model where mud also escaped through the simplified geometry of a circular pipe, but the attractive aspect of their approach is that hydro-mechanical coupling and an H2O-CO2-CH4 equation of state was included. In this model, an external fluid source is needed to drive the system, but shear failure of the boundary between liquid mud and solid mud was included to allow mechanical erosion of the source region, and thus a growing cavity that influence both subsidence and a perpetually increasing source of mud from the layer. Also using a Monte-Carlo approach, Rudolph investigated the probability of two mechanisms to stop the eruption, either caldera formation or a fluid pressure lower than that needed to drive the eruption. He pointed out that caldera formation (failure along a shear plane from the chamber to the surface), may not stop the eruption but only that their model is no longer applicable. A range of outcomes were discussed, and for most sets of parameters the eruption duration was short and the longer eruption durations tended to reduce caldera formation. Personally, I like their modeling approach as a basic start, but it needs to be made more sophisticated before any meaningful (or useful) outcome probabilities can be deemed reliable.

An interesting corollary to Randolph’s results, and piecing it together with Mazzini’s ideas, is that the mud source has stabilized, but the system keeps erupting because of the much larger fluid sources implied by a very deep source. That the system has transformed to a geyser-like system with little or no mud being erupted, is good news for preventing the formation of a caldera and therefore good news for subsidence. Subsidence occurs in response to the  loss  of  material  at  depth,  and  a  caldera  could  form  by  stress concentrations at the boundaries of the eroding mud source. If the stress state has  stabilized  because  the  geometry  of  the  extruded  volume  at  depth redistributes  stress  to  the  more  competent  layers  above  and  below  the source, then the system is structurally sound (at least for now), and subsidence should wane.

For the people of Sidarajo, one needs to ask ‘is there any light at the end of this dark tunnel?’. Sudarman suggested that there might be substantial opportunities for geothermal prospecting that would certainly benefit the local community. If, in fact, LUSI is a new-born hydrothermal system with a large supply of hot water and linked to the Arjuno-Welirang complex, then a real possibility exists to develop LUSI as a green energy resource. Much development would need to be done, but he argues that it is certainly worth pursuing.

In the best of worlds, one could imagine a growing local economy based on locally harvested Lithium batteries to store green energy produced in a geothermal complex, while researchers from around the world welcome LUSI as a geologically rare new-born hydrothermal system to study the inner workings of Earth.