My research lately has focused around faults, especially on leakage of fluids or gases through or along faults. As part of this I was lucky to attend a conference in southern France that focused on just this and still I struggled a bit when I had to explain what a fault exactly is and what it does to some first years on their first field trip (ever). So I thought why not try to explain faults in a short and simple blog entry (lets see how that works out!).
What are faults and how do they form?
Imagine a cubic volume of rock. Any rock will do, but for the purpose of a better understanding and illustration imagine a layered (sedimentary) rock that may look like this:
Now imagine the volume of rock is located somewhere in the earth’s crust. In the crust rocks are always exposed to certain stresses. Just think about the amount of rock that overlays your volume of rock and how much pressure that puts on your volume of rock! And if you consider other sources of stress like plate-tectonics or volcanic activity is becomes quickly clear that stress affects the volume of rock from all sides, like this (where σ is stress and x, y, z the directions from which the stress affects the volume of rock):
If all three principal stresses are the same not much will happen to the rock (at least at shallow depths). However, if the stress from one direction is higher than the others (e.g. σz > σx >σy ) fractures will form. The direction of the fractures depends on the principal stresses and physical parameters (i.e. stiffness) of the rock itself. For example fractures may form in the dotted rock but not in the grey rock (due to differences in the shear moduli and/or Young’s moduli). Additionally, old fractures may influence how new ones form.
The formation of fractures is then followed by movement of the rocks along the fracture planes to accommodate the stress that affects the rock. Once there has been displacement along the fracture plane it is called a fault plane! Depending on the motion along the fault plane one can differentiate between three major kind of faults: Normal faults, reverse faults and strike-slip faults (nice animations can be found here!). The displacement often happens rapidly and can release a lot of energy – an earthquake appears! The size of faults is very variable, from only a few meters long and cm-scale displacement (picture) to 100s of kilometers long faults with a throw of several kilometers (e.g. the Great Glen Fault).
As most models the one above is highly simplified. For example the fault can be classified into a fault core zone and a damage zone: The fault core is the “actual” fault, the zone where most displacement appears. The damage zone is a zone in the surrounding rocks adjacent to the fault core zone where the fracture density is very high. The fault core is often filled with crushed rock material (cataclasis) or, depending on the rock type that has been displaced, with clay-rich material. To make things more exiting a fault often consists of several core zones that overlap spatially.
Faults and leakage
Now that we have a good understanding of what a fault actually is (I hope you do now!), we can come to the more exiting part. I am studying natural CO2 reservoirs and here faults can play an important role as they can either act as high permeability pathways or as barriers (seals) for fluids. The same applies for faults in oil and gas reservoirs and thus the characteristics of faults have been studied for several decades. While the sealing capacity of a fault is important for the oil and gas industry, it is generally accepted that it is not totally bad if a fault is leaking because that just means that the oil or gas can be found somewhere else (probably a few hundreds meters away). However, for CO2 storage sites it is essential that no CO2 leaks from the reservoir to higher rock units or even the surface as the idea for the storage sites is to contain the CO2 for at least 10.000 years in order to reduce the impact on climate change (and imagine the public perception!). Thus I am (trying at least) building a model with which the sealing properties of faults can be quantitatively assessed.
In order to assess the structure of a fault and its fluid flow properties one has to understand the juxtaposed rocks, possible clay-rich material or cataclasis in the fault core and the amount of throw the fault has. This holds the first challenges: A fault is normally identified on a seismic section and depending on the quality of the seismic data and the experience of the interpreter the outcome may vary (significantly).
Seismic section (left) and interpreted faults and horizons (right). From Underhill et al. (2009).
Even if the fault is picked correctly there will be an error when the seismic depth (two way travel time in seconds) is transferred into “normal” depth (meters). Thus a good approach for a quantitative fault seal analysis should account for these uncertainties.
The interpreted seismic section gives us an idea on how big the throw of the fault is and thus what kind of rocks can be found on both sides of the fault. This means that so called juxtaposition seals can be easily identified: In the picture below the grey sections are rocks that are unpermeable and have a low porosity (seals) and the white sections are permeable rocks with a good porosity (reservoirs). If the horizons are displaced along a fault there are juxtaposition seals/barriers (A) where no fluid can flow across the fault and places where fluids can flow right through the fault (B) if the properties of the fault core allow it.
While this seems straight forward there are some pitfalls: In addition to the uncertainties described above (using seismic data), the properties of the reservoir and sealing rocks are often derived from wells that may be located several hundred meters away from the fault. This means that we have to extrapolate the rock properties (thickness, porosity, permeability…) for the rocks that actually influence the fluid flow at the fault. In order to minimize the error of this extrapolation a good understanding of the way the rock properties may change is necessary and a statistical approach is very helpful.
After the juxtapostion of the strata on both sides of the fault has been calculated (with the corresponding errors of course!), the next tricky part is to predict what the fault core looks like: Is there some kind of clay-rich material within the fault that might act as a barrier for fluid flow? Luckily many scientists have already tackled that question and have come up with algorithms that predict the properties of a fault core if the strata on both sides and the throw of the fault are known! However, often these algorithms are only used for one particular case (e.g. best/worst case juxtaposition) and thus the “standard” approach for fault sealing is more qualitatively than quantitatively. Thus I am planning on using all the uncertainties involved to create a statistically solid (and hopefully quantitative) model.
More on faults and fluid flow will follow soon! Hope you will now that you have seen how fascinating it is look forward to the next entries about it 😉