scholarly journals Clustering characteristics of gas-extraction induced seismicity in the Groningen gas field

2020 ◽  
Vol 221 (2) ◽  
pp. 879-892
Author(s):  
A G Muntendam-Bos
Keyword(s):  
Statistical Method ◽  
Induced Seismicity ◽  
Time And Space ◽  
Gas Field ◽  
Statistical Probability ◽  
Aftershock Sequences ◽  
Seismicity Rates ◽  
Background Seismicity ◽  
Intermediate Time ◽  
Groningen Gas Field

SUMMARY The Groningen gas field in the north of the Netherlands is one of the largest gas fields in the world. Since the early 1990s induced seismicity has been recorded. The largest magnitude event observed so far was a Mw = 3.6 event at the town of Huizinge in 2012. The risk posed by the induced events urged the necessity to build comprehensive seismological models capable of explaining the spatial-temporal distribution of the recorded seismicity and evaluating the regional seismic hazard and risk. The link between the occurrence of the seismicity and pressure depletion due to the production of the gas has been firmly established. However, the construction of comprehensive seismological models as well as hazard assessment is complicated by the fact that it is difficult to distinguish between induced and clustered events (events triggered by stress transfer of preceding, neighbouring events). This paper explores the contribution of clustered populations (i.e. aftershocks) to the Groningen induced seismic catalogue based on a statistical methodology in the time–space–magnitude domain. Specifically, the distributions of space–time distances between pairs of nearest-neighbour earthquakes, referred to as cluster style, is analysed. The cluster style of the Groningen induced seismicity is found to be very diffuse and characterized by a very low proportion of fore-/aftershock sequences and swarms (∼5 per cent) and a large proportion of repeater events (∼10 per cent). In contrast to human-induced seismicity in other regions, the background seismicity rate of Groningen is very low. Temporal variations in background seismicity rates can be related to changes in fault loading rates induced by gas production. Furthermore, a significant amount of independent, coincidental events (events occurring very close in time, but long distances apart) are observed. As the large gas field is fully connected, loading of the faults occurs roughly simultaneously throughout the field. Hence, the statistical probability of events occurring very close in time, but spatially far apart is significantly larger than in areas of fluid-injection induced seismicity The significant amount of repeaters and coincidental events cause an overabundance of events at intermediate time- and space-distances. This is further enhanced by the larger location errors in the catalogue increasing the estimated space-distance for non-relocated events. The diffusivity due to this overabundance of events at intermediate time- and space-distances, and the low-proportion of true fore-/aftershocks renders the statistical method used incapable of deriving a proper mode-separation value. However, this is not unique to this method. Any statistical method aimed at resolving two populations will break down if one of the populations analysed is too small. Hence, it is advisable to use caution when distinguishing fore-/aftershocks sequences or swarms for induced seismicity where the relative proportion of clustered events may be significantly lower than for tectonic events. In addition, given the small proportion of clustering and the general uncertainty in earthquake statistics, the results of this paper indicate that a distinction for earthquake risk modelling in Groningen is unnecessary.

scholarly journals Using machine learning for model benchmarking and forecasting of depletion-induced seismicity in the Groningen gas field

2021 ◽  
Author(s):  
Jan Limbeck ◽  
Kevin Bisdom ◽  
Fabian Lanz ◽  
Timothy Park ◽  
Eduardo Barbaro ◽  
...  
Keyword(s):  
Machine Learning ◽  
Support Vector Machines ◽  
Induced Seismicity ◽  
Statistical Significance ◽  
Support Vector ◽  
Effective Sample Size ◽  
Gas Field ◽  
Seismicity Rates ◽  
Vector Machines ◽  
Groningen Gas Field

AbstractThe Groningen gas field in the Netherlands is experiencing induced seismicity as a result of ongoing depletion. The physical mechanisms that control seismicity have been studied through rock mechanical experiments and combined physical-statistical models to support development of a framework to forecast induced-seismicity risks. To investigate whether machine learning techniques such as Random Forests and Support Vector Machines bring new insights into forecasts of induced seismicity rates in space and time, a pipeline is designed that extends time-series analysis methods to a spatiotemporal framework with a factorial setup, which allows probing a large parameter space of plausible modelling assumptions, followed by a statistical meta-analysis to account for the intrinsic uncertainties in subsurface data and to ensure statistical significance and robustness of results. The pipeline includes model validation using e.g. likelihood ratio tests against average depletion thickness and strain thickness baselines to establish whether the models have statistically significant forecasting power. The methodology is applied to forecast seismicity for two distinctly different gas production scenarios. Results show that seismicity forecasts generated using Support Vector Machines significantly outperform beforementioned baselines. Forecasts from the method hint at decreasing seismicity rates within the next 5 years, in a conservative production scenario, and no such decrease in a higher depletion scenario, although due to the small effective sample size no statistically solid statement of this kind can be made. The presented approach can be used to make forecasts beyond the investigated 5-years period, although this requires addition of limited physics-based constraints to avoid unphysical forecasts.

scholarly journals Statistical physics models for aftershocks and induced seismicity

2018 ◽  
Vol 377 (2136) ◽  
pp. 20170397 ◽  
Author(s):  
Molly Luginbuhl ◽  
John B. Rundle ◽  
Donald L. Turcotte
Keyword(s):  
Induced Seismicity ◽  
Statistical Physics ◽  
Aftershock Sequence ◽  
Induced Earthquakes ◽  
Aftershock Activity ◽  
Gas Field ◽  
Natural Time ◽  
Parkfield Earthquake ◽  
Groningen Gas Field

A standard approach to quantifying the seismic hazard is the relative intensity (RI) method. It is assumed that the rate of seismicity is constant in time and the rate of occurrence of small earthquakes is extrapolated to large earthquakes using Gutenberg–Richter scaling. We introduce nowcasting to extend RI forecasting to time-dependent seismicity, for example, during an aftershock sequence. Nowcasting uses ‘natural time’; in seismicity natural time is the event count of small earthquakes. The event count for small earthquakes is extrapolated to larger earthquakes using Gutenberg–Richter scaling. We first review the concepts of natural time and nowcasting and then illustrate seismic nowcasting with three examples. We first consider the aftershock sequence of the 2004 Parkfield earthquake on the San Andreas fault in California. Some earthquakes have higher rates of aftershock activity than other earthquakes of the same magnitude. Our approach allows the determination of the rate in real time during the aftershock sequence. We also consider two examples of induced earthquakes. Large injections of waste water from petroleum extraction have generated high rates of induced seismicity in Oklahoma. The extraction of natural gas from the Groningen gas field in The Netherlands has also generated very damaging earthquakes. In order to reduce the seismic activity, rates of injection and withdrawal have been reduced in these two cases. We show how nowcasting can be used to assess the success of these efforts. This article is part of the theme issue ‘Statistical physics of fracture and earthquakes’.

Probabilistic Moment Tensor Inversion for Hydrocarbon-Induced Seismicity in the Groningen Gas Field, the Netherlands, Part 2: Application

2020 ◽  
Vol 110 (5) ◽  
pp. 2112-2123 ◽  
Author(s):  
Bernard Dost ◽  
Annemijn van Stiphout ◽  
Daniela Kühn ◽  
Marloes Kortekaas ◽  
Elmer Ruigrok ◽  
...  
Keyword(s):  
Induced Seismicity ◽  
Moment Tensor ◽  
Moment Tensor Inversion ◽  
Gas Field ◽  
Double Couple ◽  
Recent Developments ◽  
Inversion Procedure ◽  
Seismic Sections ◽  
The Moment ◽  
Groningen Gas Field

ABSTRACT Recent developments in the densification of the seismic network covering the Groningen gas field allow a more detailed study of the connection between induced seismicity and reactivated faults around the gas reservoir at 3 km depth. With the reduction of the average station distance from 20 km to 4–5 km, a probabilistic full-waveform moment tensor inversion procedure could be applied, resulting in both improved hypocenter location accuracy and full moment tensor solutions for events of M≥2.0 recorded in the period 2016–2019. Hypocenter locations as output from the moment tensor inversion are compared to locations from the application of other methods and are found similar within 250 m distance. Moment tensor results show that the double-couple (DC) solutions are in accordance with the known structure, namely normal faulting along 50°–70° dipping faults. Comparison with reprocessed 3D seismic sections, extended to a depth of 6–7 km, demonstrate that (a) most events occur along faults with a small throw and (b) reactivated faults in the reservoir often continue downward in the Carboniferous underburden. From non-DC contributions, the isotropic (ISO) component is dominant and shows consistent negative values, which is expected in a compacting medium. There is some indication that events connected to faults with a large throw (>70  m) exhibit the largest ISO component (40%–50%).

The Productivity of Cascadia Aftershock Sequences

2021 ◽  
Author(s):  
Joan Gomberg ◽  
Paul Bodin
Keyword(s):  
Pacific Northwest ◽  
Subduction Zones ◽  
Geological Survey ◽  
Clear Correlation ◽  
Seismogenic Fault ◽  
Intermediate Depth ◽  
Aftershock Sequences ◽  
Intraslab Earthquake ◽  
Seismicity Rates ◽  
Background Seismicity

ABSTRACT This study addresses questions about the productivity of Cascadia mainshock–aftershock sequences using earthquake catalogs produced by the Geological Survey of Canada and the Pacific Northwest Seismic Network. Questions concern the likelihood that future moderate to large intermediate depth intraslab earthquakes in Cascadia would have as few detectable aftershocks as those documented since 1949. More broadly, for Cascadia, we consider if aftershock productivities vary spatially, if they are outliers among global subduction zones, and if they are consistent with a physical model in which aftershocks are clock-advanced versions of tectonically driven background seismicity. A practical motivation for this study is to assess the likely accuracy of aftershock forecasts based on productivities derived from global data that are now being issued routinely by the U.S. Geological Survey. For this reason, we estimated productivity following the identical procedures used in those forecasts and described in Page et al. (2016). Results indicate that in Cascadia we can say that the next intermediate depth intraslab earthquake will likely have just a few detectable aftershocks and that aftershock productivity appears to be an outlier among global subduction zones, with rates that on average are lower by more than half, except for mainshocks in the upper plate. Our results are consistent with a clock-advance model; productivities may be related to the proximity of mainshocks to a population of seismogenic fault patches and correlate with background seismicity rates. The latter and a clear correlation between productivities with mainshock depth indicate that both factors may have predictive value for aftershock forecasting.

Sign in / Sign up

Export Citation Format

Share Document