Reliable Earthquake Magnitudes for Induced Seismicity (REMIS)

Lead Research Organisation: University of Leeds
Department Name: School of Earth and Environment


The Bowland Shale, England, contains ~1,300 trillion cubic feet of shale gas, a recoverable resource worth hundreds of billions of pounds over coming decades. If this resource is exploited, it must be done so using hydraulic fracturing ('hydrofracturing'), where fluids are injected at high pressure into the rocks to create fractures through which the gas can move. This process has transformed the USA into a gas exporter and dramatically reduced prices worldwide. Nevertheless, hazards arise from this process: hydrofracturing, alongside mining and carbon capture and storage, may induce earthquakes, which if large enough, can cause shaking at the Earth's surface that leads to damage. To mitigate the risk of such earthquakes to people and infrastructure, regulators demand that operations halt if earthquakes above a certain magnitude occur in a 'traffic light' system. However, existing methods used to characterise earthquakes do not account for the possible range of magnitudes, meaning that there will be cases where operations are incorrectly permitted to continue (or are halted) based on random variation or bias in the earthquake parameter estimates. 'False alarms' would lead to millions of pounds of lost income whilst damage from unexpected seismic events would be equally costly-and it is not even known which of these outcomes is rendered more likely by errors in earthquake magnitudes. Recent work shows that errors in event locations may be many times the stated uncertainties, directly impacting earthquake magnitude estimates. More broadly, earthquake magnitudes and locations estimated routinely by geological surveys worldwide suffer from similar trade-offs.

In this technology-led proposal, we propose a new method to estimate jointly the seismic velocities of the subsurface and the locations of observed microearthquakes while varying attenuation, by using recordings from an array of seismometers at the surface. Such an arrangement is advantageous for cost and speed purposes, though is limited by uncertainty in the properties of the subsurface between the earthquakes and the stations. This fully non-linearised approach allows for the first time to calculate true tradeoffs between earthquake parameters and subsurface properties, yielding true joint probabilities that an event occurs in a certain location, and above a certain magnitude.

We will apply the method to several existing datasets. One contains the magnitude 2.3 event at Preese Hall, Lancashire, which halted the testing of hydraulic fracturing in May 2011, and so is a direct recording of what might be expected in future. Another is a set of mining-induced events which were recorded at the New Ollerton coal mine in Nottinghamshire, and serves as an excellent analogue for future industrial deployments in the UK.

We will also test our ability to image magma chambers beneath three volcanoes, in Bolivia and Ethiopia, using our method applied to available data.

By comparing our results to those from existing methods, we will show where bias is present in traditional techniques. Our method, having been validated in several ways, will serve as a useful ground truth against which we may compare methods which do not fully account for the linked distribution of subsurface velocities and event magnitude.

The overarching objective of this proposal is to develop a new method to better image the Earth and enable the creating of specific, testable hypotheses of Earth processes and structure. However, a paired, integral objective is to devise new recommendations to improve monitoring and high-value decision-making for the future of induced seismicity in the UK and worldwide. We will use the results of the work packages we describe to construct specific, probabilistic thresholds for future 'traffic light' monitoring systems, and benefit regulators, operators and the public.

Planned Impact

The wider beneficiaries of this research are:

1. Governments and regulators who wish to minimise the risk from earthquakes which are induced by hydraulic fracturing for the production of shale gas, especially in the UK. We will produce detailed recommendations for a probabilistic set of criteria for monitoring induced earthquakes using arrays of seismometers. This will be an improvement over the current system which does not take into account the distribution of likely earthquake magnitudes, and will be enabled by our new method. Hence they will benefit by minimising risk to public infrastructure, people and their property; such an improved regime will also minimise the chance that operators might be able to take advantage of any bias in the system to ensure they can continue despite creating dangerous earthquakes. Governments will also benefit from tax revenues if shale gas producers operate more efficiently as a result of an improved monitoring regime.

2. Producers of shale gas, especially in the UK. They will benefit in two ways. Firstly, if regulators adopt our recommendations, they will have the benefit of a reliable set of criteria which are objective and allow them to plan longer-term. These criteria will be published openly. Secondly, they may make use of this new technology to infer properties about the reservoir which they manage to produce hydrocarbons more effectively. We will work with industry partners (see letters of support) to incorporate our methods into their workflows.

3. The general public. Under an improved regime for monitoring of induced seismicity, they will benefit from greater transparency and confidence that biases in traditional methods cannot be used to the detriment of regulators or operators. This will help ensure that the public and their property are put at a minimised level of risk. The recommendations we will publish as accessible documents will be designed for the general public.

4. Operators of seismic monitoring facilities globally. They will benefit from our new methods since they enable full probability distributions to be calculated, and thus probabilistic hazard assessments can now include true uncertainties in earthquake location and magnitudes. Furthermore, the velocity models retrieved in the new inversions will give information about the tectonic setting where dangerous earthquakes are expected, or possibly reveal locations which have no historical seismicity, but have the potential for large events. By making our codes open source and our models available freely, we will help public bodies who are tasked with assessing natural seismic hazard around the world.
Description There are several findings from this award:

1. We have developed a new way to examine seismic recordings of ground shaking caused by small earthquakes to understand what happens in the subsurface. In particular, we have shown that it is possible to retrieve uncertainties (error bars) in the seismic wave speed and earthquake location using these data. For the first time, we have shown that is is possible to incorporate the arrival time information of seismic waves from earthquakes and waves obtained from the natural noise recorded at the same stations, and do so taking account of the full probability distribution of each parameter. This is known as nonlinearised probabilistic tomography, and the software is open-source and available at

2. One field in which our new approach is important is in the monitoring of microearthquakes in the production of energy, such as in geothermal power and hydrofracturing to retrieve hydrocarbons. These processes have the potential to cause larger, damaging earthquakes and must be monitored. In the UK and many other places, this is done using measurements of the earthquakes' local magnitude (sometimes called the Richter scale), a measurement which depends on exactly where the earthquake. Until recently, it has been assumed that our exact understanding of the seismic wave velocity of the subsurface does not affect how accurately we can measure magnitudes a great deal. In our work, we show that in fact velocity uncertainty plays a major role in inaccuracy of magnitude estimates. Any error in such estimation may mean that earthquakes which should have been noted to have been above a safety threshold were incorrectly missed, leading to an increased risk of a larger event, due to chance. Likewise, earthquakes may have been incorrectly labelled as 'too large' when in fact their magnitude is below the safety threshold. Our work shows a new way to account for how well we know the velocity of a region, using the methods developed in (1).

3. When we apply our new methods in (1) and (2) to the real-world case of hydraulic fracturing at the Preston New Road shale gas site, we find that indeed there were earthquakes which seem to have been incorrectly labelled as 'safe', and that if our approach had been used to monitor the industrial processes taking place, it is likely that activities would have been stopped on safety grounds earlier. It is possible that this may have delayed or prevented the large earthquake which later occurred.

4. Another example of the need for our work is in the understanding of volcanoes. We investigated earthquakes taking place beneath the Aluto volcano in Ethiopia, which is also the site of the country's only geothermal power plant. Using earthquake and ambient noise data, we were able to create an image of the wave velocity inside the volcano. Unlike previous work, our probabilistic approach revealed a single region of low velocity which is likely to be the source of the magma which erupted as recently as a few thousand years ago. It appears to be bound on one side by a crustal fault, where the earth has been stretched apart to form the East African Rift. For the first time, we are able to give a rigorous range of values for the amount of molten material which may be present using our probabilistic imaging, not just a single estimate. This is useful to assess how risky it is to live and work near the volcano, in case it may erupt again, and also to assess how best to produce geothermal power at the site.
Exploitation Route 1. The software and methods produced by this research are openly available for use and will be used by other researchers and professional practitioners who want to make images of the subsurface from seismic data and understand the uncertainty in their models.

2. Our new approach to the monitoring of earthquake magnitudes to regulate industrial processes which cause earthquakes may be taken forward by the Oil & Gas Authority in the UK and regulators in other jurisdictions. It provides a way to incorporate uncertainty into such monitoring systems whilst allowing regulators or governments to define the level of risk which is acceptable to them and society.

3. Our results at Aluto geothermal power plant and volcano may be used by other researchers to better understand how magma (molten rock) and fluids are stored beneath volcanic edifices, and how the volcanic and geothermal systems are related. This will also help geothermal power production worldwide, as the scientific understanding can be applied when relevant. Our imaging methods can also be used to help provide estimates of the total power potential with an objective way to produce uncertainties on these values.
Sectors Energy,Environment

Description Mantle Circulation Constrained (MC2): A multidisciplinary 4D Earth framework for understanding mantle upwellings
Amount £604,291 (GBP)
Funding ID NE/T012684/1 
Organisation Natural Environment Research Council 
Sector Public
Country United Kingdom
Start 05/2020 
End 06/2024
Description UKRI Covid Allocation Fund
Amount £69,526 (GBP)
Organisation University of Leeds 
Sector Academic/University
Country United Kingdom
Start 12/2020 
End 07/2021
Title MCTomo 
Description Seismic 3D joint velocity tomography and earthquake location software, using Bayes theorem to provide posterior probability distributions for velocity and earthquake parameters, given prior information on each. 
Type Of Technology Software 
Year Produced 2020 
Open Source License? Yes  
Impact MCTomo has now been used in several scientific articles, including those as outputs from this project. 
Title Multichannel Coherency Migration software 
Description The Multichannel Coherency Migration (MCM) is a method of searching recordings of ground motion from seismographs to detect the occurrence of microearthquakes. It is particularly suited to finding earthquakes whose signals are especially small using few seismic stations. 
Type Of Technology New/Improved Technique/Technology 
Year Produced 2019 
Impact The MCM software is being used by the Roman national seismic monitoring agency ( 
Title SeisModels.jl 
Description SeisModels.jl is Julia software for representing and manipulating seismic models of the Earth and other planets. 
Type Of Technology Software 
Year Produced 2020 
Open Source License? Yes  
Impact SeisModels.jl is used by several other open-source software packages and has been used for teaching and research.