OR/18/029 Research
| Baptie, B. 2018. Earthquake seismology 2017/2018. British Geological Survey Internal Report, OR/18/029. |
Is Earthquake activity in the northern British Isles driven by glacio-isostatic recovery?
Seismicity in northwest Scotland appears to be clustered around a number of large, steeply dipping major faults that strike either NE-SW or NW-SE suggesting that earthquake activity across the region is driven by reactivation of such fault systems by deformation associated with first-order plate motions rather than deformation associated with glacioisostatic recovery.
A number of authors have suggested that the main cause for earthquake activity in northern Britain is deformation associated with glacio-isostatic recovery. This is mainly based on the correlation between the spatial extent of the seismicity in northwest Scotland and the region of maximum ice thickness during the last glacial maximum.
Detailed analysis of spatial distribution of observed seismicity suggests that most clusters of earthquake activity are associated with steeply dipping faults that strike approximately NE-SW or NW-SE. For example, the Great Glen fault, the Strathconon fault and the Kinloch Hourn fault. Assumpçao (1981)[1] suggests that the proximity of the hypocentre calculated for the 1974 Kintail earthquake to the Strathconon Fault, along with the agreement between the NE-SW strike of the fault and one of the calculated fault planes, provides evidence that the earthquake took place on this fault. Similarly, events such as the Oban earthquake of 1986 and the Inverness earthquakes in 1816, 1890 and 1901 could be associated with reactivation of the Great Glen fault.
Similarly, focal mechanisms determined for instrumentally recorded earthquakes consistently show strike-slip faulting with N-S compression and E-W tension, which results in either left-lateral strike-slip faulting along near vertical NE-SW fault planes, or right-lateral strike-slip faulting along near vertical NW-SE fault planes.
These trends match the recent geological history of the large-scale fault structures in the British Isles where Alpine-related compression has driven faulting. In addition, the strain rate field calculated from continuous Global Positioning System measurements also exhibits predominantly left-lateral strike-slip loading along a NE-SW trend.
These results suggest that earthquake activity across the region is driven by reactivation of favourably oriented, steeply dipping fault systems by deformation associated with first-order plate motions rather than deformation associated with glacio-isostatic recovery.
Ground motions for the South Wales earthquake of 17 February 2018
We suggest that the large difference between the moment magnitude (4.0 ± 0.2 Mw) and the local magnitude (4.6 ± 0.4 ML) is a result of the relatively high stress drop for the earthquake. This also results in higher recorded peak ground accelerations for the earthquake than those predicted by commonly used ground motion prediction equations.
We determined a moment magnitude by modelling the source displacement spectra, using the spectral fitting method of Ottemöller and Havskov (2003)[3], where the seismic moment, M0, and the corner frequency, fc, are determined using a grid search. A value of 4.0 ± 0.2 Mw was determined from the 15 observations. The spread in the moment magnitudes measured at each station is significantly less than for the measured local magnitudes. We find values for the source radius, r, and stress drop, Δσ, of 0.4 ± 0.1 km and 11.1 ± 7.8 MPa, respectively. The large uncertainty in the stress drop reflects the station-to-station variability of the corner frequency measurement.
The large difference between the moment magnitude and the local magnitude may be a result of the relatively high stress drop, since differences between moment and local magnitude have been observed from other high stress-drop intraplate earthquakes of a similar size (Carreno et al., 2008[4]; Ottemöller and Sargeant, 2010[5]).
We measured peak ground accelerations (PGA) on all three component sensors at distances of up to 360 km from the epicentre. The maximum observed PGA is 9 cm/s2 recorded at station OLDB (Oldbury), approximately 97 km from the epicentre.
We compare the observed PGA with PGA modelled with Akkar et al (2014)[6] using moment magnitudes of 4.0 and 4.3 Mw for both a rock site and a soft rock site (NEHRP (US National Earthquake Hazards Reduction Program) classes B and C). In all four scenarios, we used a source depth of 7.5 km and strike-slip faulting.
For a magnitude of 4.0 Mw, on a rock site, most of the observations are outside the ±1σ bounds. Increasing the magnitude or changing the site conditions to NEHRP Class C improves the fit, but, in general, the GMPE still underestimates the observed PGA values. We repeated this analysis for two other GMPEs (Campbell and Borzognia, 2014[7]; Chiou and Youngs, 2014[8]) and find very similar results.
We suggest that the high stress drop may also result in higher recorded peak ground accelerations for the earthquake than those predicted by commonly used ground motion prediction equations used for seismic hazard assessments.
Earthquake triggering potential
An investigation of links between subduction earthquakes in Mexico since 1978 (Segou and Parsons, 2018) suggests that the magnitude 8.1 Chiapas earthquake of 8 September 2017 did not trigger the magnitude 7.1 Puebla earthquake near Mexico City on 19 September 2017. Instead, extensive postseismic deformation following the magnitude 7.5 Oaxaca earthquake in 2012 appears to have critically stressed the Puebla rupture.
In September 2017, two damaging earthquakes hit Mexico posing the question: Are the two earthquakes linked? The first earthquake occurred offshore the state of Chiapas, in the southwest of Mexico on 8 September with a magnitude of 8.1. It was followed 11 days later by a magnitude 7.1 event in Puebla State in central Mexico. Although the latter was 600 km away from Mexico City, it caused significant casualties and major damage near the capital. It occurred during planned earthquake drills marking the anniversary of the devastating 1985 M8.0 Michoacan earthquake.
To answer the question about potential links between the two events, we investigated previous links between subduction earthquakes in Mexico since 1978 by assessing the dynamic and static triggering potential along this subduction zone. Our results show that the magnitude 8.1 Chiapas earthquake on 8 September did not trigger the magnitude 7.1 Puebla earthquake, rejecting the hypothesis of any link between them.
Looking back at the recent deformation history of the subduction zone on the Pacific coast of Mexico, we find that the extensive post-seismic deformation following the magnitude 7.5 Oaxaca earthquake in 2012 critically stressed the Puebla rupture. Similarly, we find that a magnitude 7.2 earthquake off the coast of southwest Mexico in 1993 was the most likely prompt for the magnitude 8.1 event on 8 September. We also find several other links during the past 40 years that repeat this pattern.
More generally, subduction zones worldwide pose a significant threat to coastal and other communities and megathrust-related events, such as the magnitude 9.0 Tohoku earthquake in 2011, are often characterized by complex deformation histories and intriguing patterns of coseismic slip at offshore locations.
A recent collaboration between BGS and the Disaster Prevention Research Institute in Kyoto (Japan), funded by a RCUK-DPRI Kyoto Research Grant, has focused on an investigation of the Kumamoto earthquake sequence in Japan, in 2016. This study compared aftershock occurrence following both crustal and subduction events. The results show that local stress heterogeneity in Kyushu Island controls the geometry of aftershock ruptures.
A similar approach has been applied to the magnitude 7.2 Baja, California earthquake in 2010, which occurred at the continental collision between the America and Pacific tectonic plates near the well-known San Andreas Fault. The results suggest that modelling elastic deformation at the time of the earthquake is not enough and in order to achieve a realistic Earth representation our geophysical parameters, such as principal stress axes, should include the pre-seismic history of the location. These findings have significant implications for the previously long-standing approach to modelling stress changes, since our extensive statistical testing shows they perform poorly in comparison with the innovative total stress method.
In the future, we hope to extend the application of this approach and integrate it fully with rate-and-state models of aftershock forecasting in other high- seismic hazard locations of the world, focusing on big cities with high city growth rates.
Earthquakes following the 8 September event are plotted, with most clustering around that mainshock. Inset panels compare the locations and frequency of seismicity in the vicinity of the 19 September event; there was actually a reduction in the local earthquake rate, lending little support for a dynamic triggering response.
Improving event detection and location
We have tested a number of automatic phase picking algorithms so that these can be included in our data acquisition to improve near real-time detection and location capability.|}Very dense networks of seismic stations, such as that at the Vale-of-Pickering, that are designed to monitor very small earthquakes present a novel set of challenges in earthquake detection. For example: stations are often very close together, so noise may be coherent on several stations at once; separate phases may be very close together; the requirement to detect very small earthquakes means that the signal-to-noise ratio may be poor. In addition, earthquakes may occur in rapid succession during hydraulic fracturing, and these need to be located quickly and reliably to make effective operational decisions.
We have tested a number of automatic phase picking algorithms to assess their suitability for near real-time detection and location using a dense network. The algorithms that we tested are as follows: STA/LTA (Trnkoczy, 2012[9]); Carltrig STA/LTA (Johnson et al., 1995[10]); recursive STA/LTA (Withers et al., 1998[11]); Z-Detect (Withers et al., 1998[11]); Akaike Information Criterion (AIC), (Kitagawa and Akaike, 1978[12]); the FBPicker (Lomax et al., 2012[13]); and, the Kurtosis Picker (Saragiotis et al., 2002[14]).
Each trigger has several parameters that need to be jointly optimised. This was done by trying many different combinations for each trigger and ranking them based on the number of known phases found, with some consideration to the number of false triggers.
We used 63 events from a sequence of over 300 mining induced earthquakes at Thoresby Colliery, New Ollerton (Verdon et al., 2017[15]) to test the different detection algorithms. This gave a total of 362 manually picked P-wave arrival times on seven local stations.
We tested each algorithm using only short windows of data around known events, rather than scanning long continuous records. This allowed us to check the accuracy of automatic picks by comparing them against manual picks, as well as to assess the number of missing and false detections. We consider a pick good if it is within 1 second of the manual pick for that station. Picks made more than 1 second from a manual pick were considered ‘bad’. We also consider the time taken for the algorithm to scan the data for each event.
| Algorithm | Number good | Number bad | Time per event |
| Basic STA/LTA | 307 | 441 | 0.002 sec |
| Carltrig STA/LTA | 299 | 154 | 0.3 sec |
| Recursive STA/LTA | 335 | 40 | 0.001 sec |
| Z-detect | 141 | 239 | 0.02 sec |
| AIC picker | 312 | 41 | 6.5 sec |
| FBPicker | 321 | 83 | 0.7 sec |
| Kurtosis picker | 315 | 48 | 12.6 sec |
Apart from the Z-detect algorithm, all of the pickers detected more than 80% of the picks found for these events manually.
This means that, they would have detected all of the events if three station triggers were required for a detection. However, we find significant differences in the number of bad picks and in the time taken to calculate the characteristic function. The latter is an important consideration for real-time detection and location, and both the AIC picker and the Kurtosis pickers are unsuitable for real-time processing for this reason. The number of bad picks is important because too many bad station picks increase the chance of false event triggers significantly. Both the basic STA/LTA and the Carltrig STA/LTA have many more bad picks and so are not as good a choice. This leaves the FBPicker as implemented by Lomax et al. (2012)[13] and the recursive STA/LTA. The latter was quicker for this test and found slightly more good picks and slightly less bad ones. The difference is not significant but this algorithm is also very simple to implement and is the algorithm chosen to carry forward to the next step, which is automatic event location.
Funding expenditure
In 2017–2018 the project received a total of £734K, including a contribution of £464K from NERC. This was matched by a total contribution of £270K from the customer group drawn from industry, regulatory bodies and central and local government. The funding we receive from government, currently via NERC, is increasingly targeted. The reduction in NERC funding is primarily a consequence of specific targeting on Official Development Assistance (ODA). However, in 2017/2018 we were also awarded £57K for ODA projects.
The projected income for 2018–2019 is slightly less than that received in 2017–2018, mainly as a result in further reductions in NERC funding. This reflects a reduction in NERC funding for BGS in general. The NERC contribution for 2018–2019 currently stands at £420K, but we hope to increase this through applications for additional funding through the year. The total expected customer group contribution currently stands at £289K. We have also been awarded £111K for ODA projects.
Total spending in 2017/2018 was approximately £830k, slightly more than the project income.
References
- ↑ Assumpcão, M. 1981. The NW Scotland earthquake swarm of 1974. Geophys. J. Roy. Astr. Soc. 67, 577–586.
- ↑ Heidbach, O, Tingay, B, Barth, A, Reinecker, J, Kurfeß, D, and Müller, B. 2010. Global Crustal Stress Pattern Based on the World Stress Map Release 2008. Tectonophysics, 482(1–4), 3–15.
- ↑ Ottemöller, L, and Havskov, J. 2003. Moment magnitude determination for local and regional earthquakes based on source spectra. Bulletin of the Seismological Society of America, 93, 203–214.
- ↑ Carreno, E, Benito, B, Solares, J M M, Cabanas, L, Giner, J, Murphy, P, Lopez, C, Del Fresno C, Alcalde, J M, Gaspar-Escribano, J M, Anton, R, Martinez-Diaz, J, Cesca, S, Izquierdo, A, Cabanero, J G S, and Exposito, P. 2008. The 7 June 2007 Mb Lg 4.2 Escopete earthquake: An event with significant ground motion in a stable zone (central Iberian Peninsula), Seismological Research Letters, 79, 820–829.
- ↑ Ottemöller, L, and Sargeant, S. 2010. Ground-Motion Difference between Two Moderate-Size Intraplate Earthquakes in the United Kingdom. Bulletin of the Seismological Society of America, 100, 4, 1823–1829.
- ↑ Akkar, S, Sandikkaya, M A, and Bommer, J J. 2014. Empirical ground-motion models for point- and extended-source crustal earthquake scenarios in Europe and the Middle East. Bull Earthquake Eng., 12, 1, 359–387. doi.org/10.1007/s10518-013-9461-4.
- ↑ Campbell, K W, and Bozorgnia, Y. 2014. NGA-West2 Ground Motion Model for the Average Horizontal Components of PGA, PGV, and 5% Damped Linear Acceleration Response Spectra. Earthquake Spectra, 30, 3, 1087–1115.
- ↑ Chiou, B S J, and Youngs, R R. 2014. Update of the Chiou and Youngs NGA Model for the Average Horizontal Component of Peak Ground Motion and Response Spectra. Earthquake Spectra, 30, 3, 1117–1153.
- ↑ Trnkoczy, A. 2012. Understanding and parameter setting of STA/LTA trigger algorithm, in New Manual of Seismological Observatory Practice 2 (NMSOP-2), IS 8.1, 20pp.
- ↑ Johnson, C, Bittenbinder, A, Bogaert, B, Dietz, L, and Kohler, W. 1995. Earthworm: A Flexible Approach to Seismic Network Processing. IRIS Newsletter, 14, 2, 1–4.
- ↑ Jump up to: 11.0 11.1 Withers, M, Aster, R, Young, C, Beiriger, J, Harris, M, Moore, S, and Trujillo, J. 1998. A comparison of select trigger algorithms for automated global seismic phase and event detection. Bulletin of the Seismological Society of America, 88 (1), 95–106.
- ↑ Kitagawa, G, and Akaike, H. 1978. A procedure for the modelling of non-stationary time series. Ann. Inst. Stat. Math., 30, 351–363, Part B.
- ↑ Jump up to: 13.0 13.1 Lomax, A, Satriano, C, and Vassallo, M. 2012. Automatic picker developments and optimization: FilterPicker: A robust, broadband picker for real‐time seismic monitoring and earthquake early warning. Seismological Research Letters, 83 (3), 531–540.
- ↑ Saragiotis, C, Hadjileontiadis, L, and Panas, S M. 2002. Pai-s/k: A robust automatic seismic P phase arrival identification scheme. IEEE Transactions on Geoscience and Remote Sensing, 40(6), 1395–1404.
- ↑ Verdon, J P, Kendall, J M, Butcher, A, Luckett, R, and Baptie, B. 2017. Seismicity induced by longwall coal mining at the Thoresby Colliery, Nottinghamshire, UK. Geophysical Journal International, 212 (2).