Environmental Sciences Research Institute

Capability building for the geophysical interpretation of the dynamics of CO2 sequestration

An Irish National Geoscience Project funded by the Irish Department of Energy under the Griffith Geosciences Research Programme

A post-doctoral research position will be advertised for this project in the near future.

The University of Ulster Geophysics Research Group will work mainly on induced seismicity

Background & Project Objectives

Capture and geological storage of CO2 provide a way to avoid emitting CO2 into the atmosphere, by capturing it from major stationary sources (such as fossil fuel burning power stations), transporting it – usually by pipeline- and injecting it into suitable deep rock formations. In Ireland, possible suitable formations are likely to fall into two categories (i) Depleted gas fields (ii) Deep unused saline water-saturated reservoirs rocks. Both of these possible options have their respective advantages and disadvantages. Depleted fields have the advantage of significant advanced knowledge of the structures from exploration and production studies. However the full mechanical and chemical effects of depletion are not always well understood in terms of future reservoir performance and the possibility of cap rock breaches from abandoned/poorly plugged wells is usually a significant concern. Saline aquifers are thought to offer the best potential in terms of the world-wide storage of CO2. However they are usually poorly understood in advance as they have not been studied in detail, and require significant investment in terms of characterisation (usually using seismics/pumptests/numerical simulations) prior to any actual CO2 injection. Prior to storage CO2 must be first compressed, usually into a supercritical state (dense fluid). This reduces the volume relative to its gas phase by about a factor of 35. For a geothermal gradient of about 25°C/Km, CO2 must be injected below about 800m depth, to maintain it in this low volume supercritical state. Hence, for the efficient storage of CO2, a reservoir must be deeper than approximately 800m.

Once a potential site at a suitable depth has been identified the actual storage of CO2 is associated with a variety of interrelated considerations. First is the thermodynamics of multi-component mixtures of two- (oil/gas, water/gas) or three- (oil/gas/water) phase systems confined in pore structures of complex morphology, at near critical conditions. Second, the short- and long-term prediction of the fate of the injected gas including risk assessment of the storability in underground reservoirs. The injected liquid CO2 in a deep aquifer may disperse at long distances as a compound dissolved in the flowing aqueous phase, may react with salts of the aqueous phase by creating stable or unstable CO2 hydrates, may form bubbles (free gas phase), may be adsorbed on the solid pore walls, escape from subsurface formations and either migrate up to the earth’s surface or contaminate drinking water supplies, etc. Moreover, the depth and geometrical characteristics of the reservoir and overlying layers need to be precisely defined in order to monitor the extent of the CO2 saturation and possible leakages to overlying aquifers and adjacent formations. All these factors affect the acoustic (seismic) properties of the reservoir rock and their analysis and investigation are absolutely necessary in order to be able to assess, in a predictive way, the feasibility and safety of CO2 storage. Injection can also modify the stress state of the reservoir, which can lead to changes in its performance and potential hazards associated with induced seismicity, which often accompanies injection.

Induced seismicity

It has now been demonstrated in many studies that small (<<1 Bar) spatially extended stress changes can trigger earthquakes which range in energy from micro-seismicity to great M>8 events. This observation is one of the most persuasive arguments for the view that the earth’s crust is (at least) almost everywhere and (at least) almost always at or near failure. Thus forced injection of fluid into a rock mass can be expected to induce small earthquakes even in otherwise seismically inactive areas. This phenomenon is, in fact widely observed at injection sites. This presents both a potential threat and a clear opportunity.

Firstly, any alteration of the stress in a rock mass, such as the effect of filling reservoirs, deep mining or injection of fluids, has attendant seismic hazard. It is indeed surprising that in area with little or no inherent seismic hazard small stress changes due to human activity have triggered induced seismicity. Some of these induced events can be large even in seismically quiet areas. Recent attempts at recovering geothermal heat by fluid injection into hot crustal rocks near Basle in Switzerland had to abandoned when the pumping site experienced a M*** earthquake which caused a small amount of damage in the city. In India, the filling of the Koyna dam in a seismically quiet region triggered a M>6 earthquake. Such a risk requires careful consideration in advance of any CO2 sequestration project.

Secondly, induced seismicity provides methods whereby we can:

1. Image the movement of the fluids at depth,

2. Monitor the stress related variation in permeability associated with cracking

3. Image the fracture network which contributes strongly to the permeability and its anisotropic components. It is clear that the monitoring of induced micro-seismicity will be a vital component of both the testing and monitoring phases of a commercial CO2 sequestration project.

Method 1 is straight-forward and simply involves the precise location of induced events and their space-time migration in response to the injecting fluid. This will involve standard seismic location and relocation algorithms and will require little work as part of this project. Methods 2 will additionally require that the focal mechanisms and moment tensors of the micro-seismic events are calculated and statistically analysed to allow the tracking of the fluid front and volumetric changes associated with its propagation. These methods will require significant work and should allow us both to infer the evoluytion of permability buty also of the stress perturbation field around the fluid front. Method 3 relies on the fact that information concerning the stress perturbations inducing micro-seismicity can give us information about the fracture network on which the seismicity is occurring – this network will contribute strongly to the anisotropic permeability and will therefore feed back into methods 1 and 2. The idea depends on the fact that while the inducing stresses are able to trigger seismicity in such a critical system, it is clear that they very much smaller than the failure stress required to produce new fractures which would be optimally orientated for failure in the new stress field.

This work clearly shows that despite significant changes in the orientation of the principle axes of the stress perturbation tensor, the orientation of the induced micro-earthquakes remains constant to first order – existing fractures act as a structural keel and dominate the orientation of the seismic sources even though they may be far from an optimal orientation with respect to the stress field. The corollary of this is that the focal mechanisms of earthquakes generated by stresses which are significantly smaller than the whole rock failure stress, contain information as to the orientation of this structural fabric. Hence, in the field, the focal mechanisms of induced seismicity can potentially be used to determine the orientations of dominant, fluid flowing fractures.