Mud volcanoes top hazards for future Azeri operations

World's greatest concentration threat to drilling, installations

First phase production from AIOC's Chirag/Azeri/Guneshli development passed off without incident last November. But more severe tests may lie ahead as activity moves into other areas of this southern Caspian Sea acreage.

The area under development lies on a steep northwest to southeast anticline ridge, with water depths dropping from 100 meters to 400 meters. This is also an area known to overlie mud volcanoes, with shallow faulting, soil slumping, slope failure, and weak, poorly consolidated soil strata. What is unclear is how future phase drilling and facilities installation might be impacted.

AIOC classifies this part of the Caspian as moderately active in seismic terms - somewhere between seismicity levels in San Francisco and the North Sea. Substantial earthquakes have been recorded within 50 km of the production sharing agreement (PSA) acreage. In addition, shallow gas, over-pressured water-bearing sands, and mud volcano breccia at depth could all present severe hazards to drilling operations.

For future phases of the project to succeed, these risks need to be quantified and the patterns and mechanisms associated with the geohazards must be fully understood. Potentially, there could be six drilling and production centers on the PSA involving perhaps two new platforms plus floating production units. Pipeline export routes must also be scrutinized carefully, in view of the slope instability problems.

Being conscious of the development's high international profile, the AIOC consortium launched a comprehensive study of the area's geohazards last November, due to be completed some time this summer. Firm conclusions are needed for the next phase of construction to proceed as planned through to 2002.

Regional survey

For the recent first development phase, Fugro-Geoteam in Aberdeen reinterpreted all existing data within the PSA area, integrating this with AIOC's newly acquired, high resolution 3D seismic data. Fugro-Geoteam subsequently performed bathymetric and stratigraphic surveys using HR and UHR analogue and digital techniques. "We're interested especially in the soil's upper 200 meters," said Chris Golightly, AIOC's senior geotechnical engineer at a Fugro presentation in London last fall.

The Phase 1 mapping/pattern recognition work is also being incorporated into the current study, which splits into two parts - geohazards facing engineering facilities, and geohazards affecting drilling and well planning.

Coordinating consultant for the current program is Fugro Engineers in The Netherlands, which has conducted nearly all the PSA's survey and geotechnical work to date. This includes platform and well site geophysical surveys and geotechnical studies of gravity cores across the PSA acreage. Three further sub-consultants have been contracted to analyze the individual types of geohazard anticipated.

The current study's initial phase may already be concluded. This comprises detailed geophysical interpretation and mapping of geohazards followed by pattern recognition - the observation and categorization of the mechanisms and patterns of mud volcanoes, shallow faulting, and slope failures. The drilling study is determining the occurrence of shallow gas, deeper faulting, mud volcano breccia, soil anomalies and overpressured water.

Risk assessment

A second, three-month analytical phase will include risk assessment of the observed patterns and mechanisms of the three main geohazards with the aim of ascertaining design criteria, safe stand-off distances and, ultimately, a review and reconstruction of the existing "traffic light" geohazard identification map.

The third phase of the study will be a two-month reporting and overview exercise, the aim being to advise AIOC on the constraints to future development. A key element will be conversion of the geohazard identification map into a facilities risk map, allowing as much as possible of the seabed to be freed up for development.

This would also impact directional drilling, Golightly claimed: "Through correct analyses, the wireline costs saved could more than cover the cost of this study." A final report will detail a geohazard avoidance strategy and define related design criteria and design approaches for facilities engineering.

Mud volcanoes

Future work will be split into packages based on the three main geohazards - mud volcanoes, slopes, and faults. The world's greatest concentration of mud volcanoes is found in Azerbaijan - 220 out of a global total of 350. "We have seven in our area alone," said Chris Golightly, "with the land mud volcanoes known to erupt on a return pattern of perhaps 100-250 years, and with mud flows up to 5 km long. Soviet scientists in the past tried to evaluate the reasons - sunspots, varying levels of the Caspian Sea - to me, it seems to be a random effect."

However, Professor Barry Voight of the Department of Civil Engineering, University of Pennsylvania, is attempting to identify a pattern. Voight is one of the few specialists in assessing volcanic hazards from a geomechanics perspective, and is familiar with landslide dynamics pertaining to volcanic eruptions.

The dilemma for oil companies in areas of the Caspian where mud volcanoes exist - AIOC is not the only group affected - is that the overlying sediments are lightweight. According to Golightly, " They don't give enough resistance to the upwelling of diapirs underneath. The result may be a diatreme, which can generate a very high velocity gas plume up to 800 meters high, which will flame and burn people 1 km away." Drilling too close to these mud volcanoes could have catastrophic consequences, and is recognized as a key part of the study.

Clearly, this also presents a distinct risk to offshore platforms and associated facilities along the Apsheron anticline. Mitigation of the risk can be achieved via siting constraints and appropriately engineered designs, but this depends on a thorough understanding of the nature of the risk from mud volcanoes, including potential seabed disruption, the occurrence and extent of submarine mudflows and return periods for mud volcano activation.

AIOC's scope of investigation for mud volcanoes includes:

Confirmation, by comparison with borehole data and SOCAR interpretation maps, of the extent of surface mud flows identified from regional geophysical survey work
Determination of geotechnical properties of mud volcano deposits
Mapping of deep (100 meter) palaeo MV breccia
Compilation and review of literature on mud volcanoes, leading to a database enabling integration of SOCAR/GIA and western data, ideas and interpretations relating to mud volcanoes
Study of geotechnical properties, physical processes, relation of diapirs to faulting, hazards relating to mudflows, ground disruption and gas emissions
Analysis and determination of nature and extent of seafloor disturbance (faulting and settlement) and flow hazards around PSA area mud volcanoes.

Slope instability

There is extensive evidence of past slope instability within the PSA survey area. The bathymetry data reveal that slope failures at the south end of the anticline cut back into the ridge, with severe faulting in many cases up to the mudline zone.

Types of slope instability include major and minor mudflows emanating from mud volcanoes, secondary failure on fault scarps and slump scars, deep rotational slides, and large scale slope failure associated with the main and older slumps.

The Norwegian Geotechnical Institute (NGI) in Oslo is responsible to AIOC for assessing slope stability hazard and risks in the PSA area. As background, it can call on the results of 250 basic geotechnical boreholes conducted by the Soviets, plus another borehole drilled recently by Fugro. These reveal that the soil is virtually all clays, with no sands and very poorly consolidated.

"All the geotechnical properties of these clays vary due to the climate at the time of deposition," Golightly said, "which alternated between aridity and humidity. The low level of consolidation is a result of extremely high past sedimentation rates in the southern Caspian."

NGI will be looking at the potential for shallow and deep slides induced by triggers such as gas, seismicity and erosion. "If we place platforms there, or wellhead structures in deep water with tiebacks up the slope, how stable will this arrangement be?" The Institute will investigate this for the whole length of the ridge, allocating strength parameters for the different areas and recommending a safe stand-off distance from any potential failures. Its main scope of actitivity will comprise:

Interpretation of geophysical and borehole data to identify possible glide planes, critical slopes and cross-sections
Evaluating possible release mechanisms such as earthquake loading, shallow gas, gas hydrates, mud volcanoes, erosion and over-steepening
Identifying and selecting design geotechnical profiles for each cross-section identified
Static stability analysis and dynamic (earthquake) stability analysis
Post-earthquake analysis of deformed slopes
Probabilistic analysis of selected slopes, applying known variations in earthquake acceleration, soil strength and model uncertainties
Determining potential slide scenarios.

Faulting

Faulting is prevalent along large parts of the Apsheron Ridge. Five different fault types have been identified in the PSA area - tectonic faults, gravity failures, faults associated with mud volcanoes, growth faults and constant offset faults. Some are clearly tectonic in origin, others are caused by near-surface and deep-seated gravity failures. In other cases the origin is uncertain, a problem exacerbated by the fact that faults of different types can occur in the same area (for instance, tectonic faults acting as planes of weakness along which subsequent gravitational failures occur).

Sub-consultants for the fault displacement hazard study are Geomatrix of San Francisco. This is a consulting firm specializing in seismic and engineering geology and seismicity. It has special expertise in identifying faults that might pose a threat to facilities, and can assess the probable activity, frequency and magnitude of surface fault ruptures.

Main aim of this study will be to develop a map of tectonic faults (or zones of faults) with the potential for surface fault rupture, as well as corresponding displacement hazard curves showing the relationship between the displacement size and the frequency at which it will likely be exceeded.

Factors to be input into the fault displacement hazard model will include:

Tectonic setting and structural models of fault kinematic relationships
Fault type classification
Location and three-dimensional geometry of faults and associated folds
Cumulative displacement of various geologic marker horizons and the ages of the displaced units
Potential fault rupture lengths and/or rupture areas
Estimated maximum earthquake and earthquake recurrence models, plus the average/maximum displacement incurred per event
Continuity and age of undisplaced strata.

Under the agreement with SOCAR, an ex-Soviet jacket was refurbished for the first phase production in 120 meters of water, with new topsides installed. But alternate facilities are permitted for future production phases - possibly an "earthquake tower" design comprising a slender tower/deck on a square jacket, said to react well to seismic loading. FPSOs are also in the frame. These offer the advantage of quick disconnection.

An AIOC reservoir modeling team in Baku may also investigate the chances of Ekofisk-type subsidence, but since full re-injection will take place, this should not be a major issue, says Golightly.