State-of-the-Art of Monitoring Methods to evaluate Storage Site Performance

 
1. INTRODUCTION
 
2. MONITORING TECHNIQUES
 
3. MONITORING CONCEPTS – STATUS QUO
 
4. SETTING UP A SITE-SPECIFIC MONITORING PLAN
 
5. CONCLUSIONS AND RECOMMENDATIONS
 
6. REFERENCES
 
 
2.1 CO2 plume migration in the storage reservoir
 
2.2 Surface uplift
 
2.3 Induced seismicity and mechanical reaction of overburden
 
2.4 Faults
 
2.5 Abandoned wells
 
2.6 Overlying and adjacent aquifers
 
2.7 Freshwater aquifers
 
2.8 Near surface eco-compartments
 
 
» 2.1.1 Seismic reflection
 
2.1.2 Gravity
 
2.1.3 Geoelectrics and electromagnetics
 
2.1.4 Well Logging / Wireline Logging
 
2.1.5 Satellite interferometry and other techniques for surface movem...
 
 

2.1.1 Seismic reflection

In seismic measurements surface sources (e.g. dynamite, vibrating machines or air gun arrays for onshore and offshore use, respectively) are utilised to generate downward propagating elastic waves that are reflected from subsurface features and return to the surface where they are recorded by ground motion sensors (geophones), resulting in a three-dimensional view of the subsurface. In the case of a 3D survey, a regular 2D grid of surface sources and sensors is deployed. The recorded data are combined to produce a 3D image of the subsurface. The seismic survey provides an initial baseline that can be compared to changes in subsequent seismic surveys to create a time lapse image of CO2 plume migration and to detect significant leakage or migration of CO2 from the storage site. Surface seismic techniques provide detailed spatial resolution of CO2 distribution, but are less sensitive than well-based methods and, therefore, may require the presence of large volumes for detection of CO2 (Monea et al., 2008).

Lumley (2010) describes various aspects of seismic monitoring for CO2 injection: The effectiveness of the seismic monitoring depends on the properties of the pore fluid (including the CO2) and the compressibility of the dry rock frame. If the dry-frame compressibility is low, i.e. the rock is stiff, the seismic measurements will not easily sense the properties of the pore fluid. When injecting into a depleted hydrocarbon reservoir, depleted oil with low solution gas-oil ratio (GOR) will give more favourable conditions for seismic monitoring of CO2 injection than depleted oil with high GOR. The presence of residual hydrocarbon gas in the pores will furthermore provide less favourable conditions for seismic monitoring (cf. Picotti et al., 2012).

In addition, the effectiveness of seismic monitoring depends on the nature of the seismic acquisition set-up, in particular on the temporal frequency content of the data. This influences both subsurface resolution and the sensitivity for detection of gas or fluids. In order to achieve high resolution, it is necessary to record the high frequencies; however, high-frequency signals are also attenuated more quickly (which limits depth penetration) and more susceptible to effects of reverberation and scattering.

The injection of CO2 alters the compressibility and the density of the reservoir fluid, which has several effects on the seismic response. Firstly, the injection changes the velocity of the seismic waves, which affects the time required for a seismic wave to pass through the reservoir. In seismograms, this can be observed, for example, as time shifts in waves reflected from layer boundaries below the reservoir, and this is a valuable tool for quantifying the amount of injected CO2 (Fig. 2-2). In order to calculate the CO2 layer thickness from the time shift, the velocity must be estimated, e.g. by using assumptions about porosity and CO2 saturation (Chadwick et al., 2004).

Fig 2_2

Fig. 2-2: Seismic attribute maps from time-lapse measurements during the project CO2SINK at Ketzin, Germany (Ivanova et al., 2012). The grey symbol marks the injection borehole. Left panel: Normalised time-lapse amplitude at the level of the reservoir, showing an amplitude anomaly due to the injected CO2. Right panel: Time shift of a reflection below the reservoir caused by a velocity pull-down effect due either to partial CO2 saturation in the reservoir or to a pressure increase.

Secondly, the injection-induced changes in the reservoir have an effect on the amplitude and the frequency content of the reflected waves. By comparing seismic amplitude and frequency maps from measurements carried out before and after injection, it is possible to track the CO2 migration with high lateral resolution (e.g. Chadwick et al., 2004, Ivanova et al., 2012). Volume estimates can also be derived by assuming relationships between reflection amplitude and CO2 layer thickness, also requiring that additional assumptions are made about porosity and saturation (Chadwick et al., 2004). It is also possible to combine the time shifts and the amplitudes to derive volume estimates, e.g. by using the time shift to estimate the thickness of the CO2 layer and the amplitude to estimate saturation (Ivanova et al., 2012).

Several recent studies on CO2 storage reservoirs (Rabben and Ursin, 2011; Rubino and Velis, 2011) utilise the amplitude variations of the reflected seismic wave as a function of incidence angle. This approach has been used for a long time in the hydrocarbon industry through amplitude versus offset (AVO) analysis, and there are different classes used to distinguish reservoirs based on the AVO characteristics. For a saline aquifer environment, the injection will cause a much smaller change in the S-wave velocity than in the P-wave velocity, and this has effects on the variation of reflection amplitude with incidence angle. Rabben and Ursin (2011) applied amplitude versus angle (AVA) analysis to seismic data from Sleipner to estimate seismic reflection coefficients, which ultimately can be used to calculate the mass of injected CO2. Numerical studies by Rubino and Velis (2011), again with focus on Sleipner (cf. Section 3.3.1), indicate that it may be possible to obtain reasonable thickness estimates for CO2-bearing layers having a thickness of only a few meters using AVA analysis.

The most established seismic method for detailed mapping of CO2 migration is 3D seismic reflection measurements, or rather 4D when carried out in time-lapse mode. Numerous 3D/4D surveys have been carried out in connection with CO2 injection, both on land and offshore (e.g. Arts et al., 2004; juhlin et al., 2007; Urosevic et al., 2011). For seismic time-lapse measurements it is important to achieve high repeatability. A useful procedure for assessing the similarity of two or more time-lapse data sets is to use repeatability metrics (cf. Kragh and Christie, 2002). Poor data quality can considerably reduce the detection sensitivity (presence of noise and/or non-repeatable acquisition patterns). Therefore, the same seismic recording parameters should be used for the baseline and repeat surveys. The shot points and geophones should be placed at approximately the same locations for all measurements. Also, the source of the seismic signal should preferably be the same. In land measurements, the position of the groundwater table affects the seismic response, and therefore all measurements should ideally be carried out at the same time of the year. Even after taking precautions to ensure that the data acquisition is carried out correctly, it is necessary to apply careful data processing in order to enable a comparison of the various datasets (Bergmann et al., 2011).

Seismic 2D surveys (i.e., seismic measurements carried out along profiles), are much cheaper than 3D measurements. Sometimes 2D land measurements can be better in resolving structure, e.g. thin layers, than 3D measurements acquired with similar instrumentation. By using a suitable arrangement of 2D profiles it can thus be possible to monitor CO2 injection, although it can be difficult to know exactly where to place the profiles and there is a risk to miss the CO2 plume.

The usefulness of seismic measurements varies depending on several factors, e.g. the geometry of the layer boundaries and the physical properties of the rock matrix and the pore fluid. The applicability of seismic methods needs to be assessed when selecting techniques for site characterisation and monitoring, and in some settings seismic methods will not work well. When selecting techniques, it is also necessary to consider the environmental impact of the seismic data acquisition, e.g. the potential damage caused by using dynamite charges or Vibroseis trucks, clearing the vegetation to install geophones or building new roads to transport equipment and personnel.

Seismic borehole measurements can provide higher resolution data than surface measurements, although the lateral coverage is in general more limited. Cross-hole measurements, using a combination of borehole sensors, potentially have very high resolution in a limited volume of the subsurface. It is also common to use a combination of borehole sensors and surface signal sources, commonly referred to as vertical seismic profiling (VSP). VSP provides valuable information about the geological structure of the subsurface and is one of the best techniques to study seismic anisotropy. In VSP exploration, the seismic energy is produced via a surface source at or near a borehole. By using a receiver array in the well it is possible to record both the downgoing and upgoing seismic waves. One big advantage of the method is the ability to correlate the upgoing, or reflected waves, directly to the layer boundaries. The VSP method also produces full volumetric images of the subsurface structures around the well with improved seismic resolution in comparison to surface seismic methods.

The common use of VSP is to depth-correct a seismic survey, i.e. to bind surface seismic (2D or 3D) to well logs and stratigraphy information. VSP can be implemented in a "walk-away" fashion to monitor the footprint of the plume as it migrates away from the injection well. In the walk-away VSP configuration the sources are arranged on radial profiles around the injection well in order to create an offset at the surface as the receivers are held in a fixed location. In connection with the CO2 injection at Ketzin, a seismic experiment has been carried out using a number of seismic sensors buried in shallow boreholes (at depths of around 50 m), below the groundwater table. Preliminary results show that it is possible to image the CO2 reservoir with high resolution, by avoiding the degradation of the seismic signal when passing through the highly attenuating dry overburden. (Ivandic et al., 2012).