Author: Prof.Dr. Horst Rüter

In an article that attracted a lot of press attention Keranen et al. (2014) take a closer look at the strong increase in seismicity in Oklahoma and therefore also in the central and eastern USA, something that has been talked about in publications for some time. The title alone [Sharp increase in central Oklahoma seismicity since 2008 induced by massive wastewater injection] draws a connection between this increase with wastewater injection. The article describes disposal wells as ‘potentially responsible’. This also means that there is no direct link with fracking, i.e. with the stimulation of oil and gas deposits and no link with the extraction of gas and oil. Throughout the article the term ‘wastewater’ is used without stating whether this is wastewater from the oil/gas industry; however, it can be assumed that this is the case.

The authors regret that pressure data (of the injection pressures) are rarely accessible, even though only these would be able to establish a causal link between injections and seismicity. The authors use model calculations of possible spread of pressure as a substitute.

The observed quake swarms are very carefully researched and their statistical information presented. An astonishing result is that the many thousands of disposal wells in Oklahoma are almost exclusively aseismic, i.e. at least do not produce any noticeable events. Only the four wells with the highest injection rates are said to be responsible for causing 20% of the entire seismicity in the US in the years 2008 to 2014 and around 45% of seismicity (> M3) in the eastern states. The largest of these quakes (M 5,7) in 2011 near Prague, Oklahoma,is said to ‘probably’ be linked to these injections. Overall this means a 40times increase in annual seismicity compared with the years 1976 to 2007.

The article focuses on quake swarms near the town of Jones, which are therefore also called ‘Jones swarms’. For these, hydraulic model calculations are presented, with four of the main wells having 4 million barrels/month (250 l/sec) injected into them. The detailed localisation of the quake swarms shows that although these are in the neighbourhood (up to 35 km distance) of the injection wells, they are also linked to well-known fault systems like the Wilzetta Fault and the Nemaha Fault. The reason why here in particular such extreme seismicity occurred is the coming together of high injection rates with the presence of active fault systems.Purely empirically (i.e. without considering a connection with rock mechanics) the occurrence of events is contrasted with modelled spread of pressure.This shows that even pressure increases of only 0.07 MPa (0.7 bar) can trigger an event. At least here events were localised in areas where the pressure modelling (no data were available) had exceeded this value.This value agrees well with the values mentioned elsewhere in the literature.

Summary and conclusions:

1. The article does not establish a link between quakes and drilling, fracking or the extraction of unconventional gas. Given the hundreds of thousands of such activities we can continue to assume that they have little significance as the cause of induced seismicity.
2. A link between the extreme rise in seismicity in Oklahomawith wastewater injection (disposal wells) is regarded as likely.
3. Thousands of disposal wells do not cause noticeable events. The observed extreme rise in seismicity is merely due to 4 disposal wells.
4. These 4 disposal wells do not only have extremely high injection rates, they are also located in the immediate vicinity of known recently active fault systems.
5. Wastewater injection is thus generally possible without generating seismicity (many thousands of examples also in Oklahoma)
6. Disposallocations must be explored carefully paying particular attention to mapping fault systems and information about the stress field.
7. Injection tests must be conducted prior to the operational injection. Only locations with a high rate of injectivity are suitable.If these tests produce unfavourable results, the drilling must be abandoned.
8. Disposal locations must be seismologically monitored.
9. The injection must be stopped (response profile) if there is a sign of increased seismicity. The location will then be unsuitable, and the drilling has to be stopped.
10. It is difficult to see why the 4 injection wells in Oklahoma were used (or are still being used?) for so many years despite the observed extreme seismicity.