International Journal of Rock Mechanics and Mining Sciences
Large earthquakes during hydraulic stimulations at the geothermal site of Soultz-sous-Forêts
Introduction
Since 1987, a European geothermal project has been developed at Soultz-sous-Forêts (Alsace, France) in the western part of the Upper Rhine Graben [1]. This area is characterized by an extensional tectonic regime and shows a high temperature gradient anomaly [2]. The crystalline basement is covered by about 1400 m of Cenozoic and Mesozoic sediments. Analysis of cores recovered during drilling and of borehole geophysical logs has shown that the granitic massif is highly fractured [3]. The main orientation of the fracture system is NNW–SSE. Petrographical studies have concluded that the majority of the fractures are hydrothermalized and exhibit a slight, natural permeability [4].
The aim of the geothermal project is to extract the heat stored in the deep and hot granitic rocks through the forced circulation of water between injection and production boreholes: the cold water is injected into the rock mass through the open-hole section of the injection well, exchanges heat with the rocks while circulating in the fracture network, and is then pumped up to the surface via production wells. For this purpose, four deep wells have been drilled since 1987: GPK1, drilled to about 3600 m depth; GPK2 initially drilled to 3800 m depth, then deepened to 5000 m depth; finally GPK3 and GPK4 also reaching 5000 m depth. At this depth, the temperature is around 200 °C. The boreholes GPK2–4 are expected to form a future geothermal triplet (one injection well and two production wells). They are now used for the study and development of the deep geothermal reservoir. They are aligned in a N170° direction, and the distance between the bottom of two adjacent wells is about 600–650 m.
Although the fracture system is naturally permeable, the permeability values and the connectivity between the wells and the fracture network on one side, and local permeable cells within fractures on the other side, are too low to reach an optimized exploitation of the geothermal resource. One of the major goals of the geothermal project is to enhance the hydraulic properties of the reservoir. This is usually achieved through massive hydraulic stimulation tests: water is injected at high flow rates into the granite in order to increase the pressure in the deep reservoir. Hubbert and Rubey [5] first showed that an increase of the pore fluid pressure tends to decrease the effective normal stress on fractures and, as a result, triggers earthquakes. Thus large-scale injections are likely to induce shearing on pre-existing joints that are favourably oriented within the prevailing stress field. Two important effects are expected: permeability and connectivity enhancement, and generation of an intense microseismic activity.
Section snippets
Stimulations
Several stimulation tests were performed at Soultz-sous-Forêts in order to develop the reservoir at 3600–3800 m depth [6], [7], but in this paper we will only focus on the last four stimulations of the deep reservoir at 5 km depth. Hydraulic parameters (flow rates and pressures) of the injections are shown in Fig. 1. In 2000, GPK2 was stimulated at flow rates up to 50 Ls−1 for a total injected volume of 25 000 m3. GPK3 was stimulated in 2003 at similar flow rates but with short peaks at 60 and 90 Ls−1
Seismicity
For each stimulation test, the induced microseismic activity was monitored by a surface seismological network, installed by the “Ecole et Observatoire des Sciences de la Terre” from the Strasbourg University (Fig. 2). In 2000, 18 temporary seismological stations, composed of one component (1-C) and three components (3-C) sensors were set up. Since 2003, a permanent network of nine stations has been deployed around Soultz-sous-Forêts to continuously monitor the seismic activity. For the 2003
Temporal evolution
In Fig. 3, Fig. 4 we show, for each of the four stimulations, the origin time of each seismic event plotted against the wellhead pressure (heavy black line), and the flow rate (thin black line).
For 2000, the seismicity mainly concentrates in the first four days, after which the time between consecutive microearthquakes increases. A large event (M=2.5) occurred on the fourth day of injection, while the flow rate was at 50 Ls−1 and the wellhead overpressure at 13 MPa. The second largest seismic
Spatial evolution
The location of the microearthquake occurring in 2003–2005 is performed by a hypoinv-like algorithm [9], which was modified to take into account variations in Vp/Vs ratio and the 3D configuration of the seismic network. The 1D velocity model contains five layers whose characteristics have been determined thanks to different log data and a calibration shot made in the 1988 and 1989 [10]. We consider the location to be valid if the root mean square between the observed and predicted arrival times
Focal mechanisms
We have constructed the focal mechanisms of most seismic events presented in this study. For each seismic event, the number of polarity observations depends on the year of the experiment since the seismological network was not identical for all the hydraulic tests. The number of polarities is as follows: at least 15 polarities of P wave for 2000, at least 17 polarities for the 2003 experiment, eight polarities for the events of the 2004 and 2005 experiments. The focal mechanisms have been
Source dimensions
Since several years, it has been argued that the seismicity in Soultz-sous-Forêts has a particular source property. Results summarized in Michelet et al. [14] from the automatic study of the spectral content of the acceleration recorded by the downhole seismological stations show that the invariant is not the stress drop but the source dimension. Therefore, a large seismic event would imply a huge stress drop. This point is unusual and contradicts several studies for such a range of magnitude
Data
The main problem for this kind of study comes from the sensor and from the acquisition chain involved in the recording of the data. The transfer function of the whole chain has to be precisely known. We have thus chosen to use only the data from a velocimeter CMG-3 T manufactured and calibrated by Güralp. This device has a constant response for the frequencies involved, typically from 1 to 90 Hz. The sampling rate is 200 points per second. Considering the work of Abercrombie [15], from the range
Empirical Green's function (EGF)
The separation of the source characteristics from the path and site effects is one of the main problems of the spectral analysis. One method for removing the path and site effects from the recording of a given event consists to choose a smaller event as an empirical Green function and deconvolve its seismogram from the seismogram of the original event [30].
Actually a recorded seismogram is the result of the temporal convolution of three different operators: the Green function, which includes
Interpretations
The occurrence of large magnitude events at the beginning of the 2003 experiment indicates the existence of a structure large enough to accommodate these events. The ultrasonic borehole images (UBI) of the well [31] show that there is a highly fractured zone at 4.7 km depth which strikes 167° and dips 70° (Aki convention). The width of the damaged zone in the well allows us to consider that this feature is a multihectometric or a kilometric structure. Fig. 11 points out that the location of this
Conclusion
The occurrence of large late induced seismic events has been observed in many fields where injection of fluid has been involved ([39], [40], [41] among others). Their link with the hydraulic activity is no more argued but their mechanic is still poorly understood. We present the analysis of the M⩾1.4 events during the hydraulic activities on the geothermal field of Soultz-sous-Forêts from 2000 to 2005. The results from the spectral analysis performed on the data recorded by a 200 m depth
Acknowledgement
We acknowledge the anonymous reviewers for their fruitful comments. This work is a part of the Jean Charléty's PhD thesis supported by the ADEME (Agence de l’Environnement et de la Maitrise de l’Energie). We also thank the EEIG “Heat Mining” for providing us with the hydraulic data and for fruitful discussions.
References (41)
- et al.
Analysis of macroscopic fractures in granite in the HDR geothermal well EPS-1, Soultz-sous-Forêts, France
J Volc Geotherm Res
(1996) Triangle diagrams: ternary graphs to display similarity and diversity of earthquake focal mechanisms
Phys Earth Planet Inter
(1992)- et al.
European HDR project at Soultz-sous-Forêts, general presentation
Geotherm Sci Technol
(1991) - et al.
Hydrothermic studies in the Hot Dry Rock project at Soultz-sous-Forêts
Geotherm Sci Technol
(1991) - et al.
Comparative analysis of direct (core) and indirect (borehole imaging tools) collection of fracture data in the Hot Dry Rock Soultz reservoir (France)
J Geophys Res
(1997) - et al.
Mechanics of fluid-filled porous solids and its application to overthrust faulting
Geol Soc Am Bull
(1959) - Baria R, Garnish J, Baumgärtner J, Gérard A, Jung R. Recent development in the European HDR research programme at...
- Gérard A, Baumgärtner J, Baria R. An attempt towards a conceptual model derived from 1993–1996 hydraulic operations at...
- Cuenot N, Dorbath C, Dorbath L. Analysis of the microseismicity induced by fluid injections at the Hot Dry Rock site of...
- Klein FW. Hypocenter location program-HYPOINVERSE: users guide to versions 1, 2, 3 and 4. US Geological Survey Open...
Seismic studies on the HDR site of Soultz-sous-Forêts (Alsace, France)
Seismic source parameter evaluation and its importance in the development of an HDR/EGS system
Earthquake source scaling relationships from −1 to 5 ML using seismograms recorded at 2.5-km depth
J Geophys Res
Attenuation and source properties at the Coso Geothermal Area, California
Bull Seismol Soc Am
Source parameters of injection-induced nicroearthquakes at 9 km depth at the KTB deep drilling site, Germany
Bull Seismol Soc Am
Theoretical basis of some empirical relations in seismology
Bull Seismol Soc Am
Parameters and a magnitude moment relationship from small earthquakes observed during hydraulic fracturing experiment in crystalline rock
Geophys Res Lett
An introduction to mining seismology
Cited by (106)
A 2D FDEM-based THM coupling scheme for modeling deformation and fracturing of the rock mass under THM effects
2022, Computers and GeotechnicsInduced and triggered seismicity by immediate stress transfer and delayed fluid migration in a fractured geothermal reservoir at Pohang, South Korea
2022, International Journal of Rock Mechanics and Mining SciencesIdentification and analysis of barriers for harnessing geothermal energy in India
2022, Renewable EnergyFast migration episodes within earthquake swarms
2023, Geophysical Journal InternationalScaling Microseismic Cloud Shape During Hydraulic Stimulation Using In Situ Stress and Permeability
2023, Journal of Geophysical Research: Solid Earth