INTERNATIONAL SOCIETY FOR SOIL MECHANICS AND GEOTECHNICAL ENGINEERING This paper was downloaded from the Online Library of the International Society for Soil Mechanics and Geotechnical Engineering (ISSMGE). The library is available here: https://www.issmge.org/publications/online-library This is an open-access database that archives thousands of papers published under the Auspices of the ISSMGE and maintained by the Innovation and Development Committee of ISSMGE. The paper was published in the proceedings of the 7th International Conference on Earthquake Geotechnical Engineering and was edited by Francesco Silvestri, Nicola Moraci and Susanna Antonielli. The conference was held in Rome, Italy, 17 – 20 June 2019. Earthquake Geotechnical Engineering for Protection and Development of Environment and Constructions – Silvestri & Moraci (Eds) © 2019 Associazione Geotecnica Italiana, Rome, Italy, ISBN 978-0-367-14328-2 Seismic swarms in South Limburg (The Netherlands): Tectonic or induced as coal mining lagging effect? C. Sigarán-Loría Royal HaskoningDHV, Nijmegen, The Netherlands S. Slob Witteveen+Bos/Cohere Consultants, Amsterdam/Amersfoort, The Netherlands ABSTRACT: Tectonic earthquakes in The Netherlands are usually related to the extensional graben along the Peelrandbreuk in South Limburg. The strongest earthquake recorded in is the M 5.8 earthquake in Roermond on 13 April 1992. In Germany induced earthquakes are recorded related to on-going coal mining activities in the Ruhr area with magnitudes mostly smaller than M 3.0. There is a strong correlation in time and space between the seismic events and mining activity. Other mechanism that induce seismicity are known, such as waste water injection from the shale gas activities in America, some reservoirs from hydropower projects as well due to the increase of water pressure along the stressed faults will reduce the normal stress and thus the friction along the fault and this could induce fault movement, resulting in an earth- quake. Another mechanism may be the increase in mass due to the rising ground water. This increase in mass could be an additional driving force behind fault movement and thus the devel- opment of seismicity. Two earthquake swarms have occurred at the Southern edge of the former coal mining area, around the village of Voerendaal. The first swarm occurred in 1985 and the second swarm at the end of 2000 until the beginning of 2001. These two swarms could be related to the rising mine water. A temporal and spatial analysis on groundwater level devel- opment, seismic data, ground uplift and fault data to determine a possible relation between the occurrence of two earthquakes swarms around Voerendaal and the mine water level increase was performed. The study looked at two possible mechanisms that could have triggered the earthquake swarms: (1) exceedance of the critical state of the active faults and (2) the increase and shift in mass due to ground water level rise as a driving (energy) source. 1 INTRODUCTION The Netherlands has a long history of underground coal mining. Coal mining gradually decreased in the sixties and stopped entirely in the early seventies of the last century. For a long period, the mines were drained, to ensure dry working conditions which resulted in a lowering of the ground water table. Since the closure of the mines, the groundwater pumping decreased in stages. This has resulted in a step-wise rise in groundwater, which is now levelling towards the original level before the coal mining started. There is tectonic and induced seismicity in The Netherlands. The induced seismicity is related to reservoir compaction due to gas extraction in the North of The Netherlands. The tectonic occurs in the south-east of the country, related to extensional movements along the Roer Valley Graben, which is an extension of the Rhine Valley Graben structure. An example is the ML 5.8 Roermond earthquake of 13 April 1992, strongest registered tectonic event that caused some damages. There is induced earthquakes related to coal mining activities in western Germany, close to the border with The Netherlands. In the Ruhr area, every year about 1400 seismic events with magnitudes up to ML 3.0 are measured. There is a strong correlation in time and space between the seismic events and mining activity. Significantly less earthquakes occur at the weekend when there are no mining activities (Bischof et. al., 2005). 5019 Seismicity can be triggered as well by raise in the groundwater level. These earthquakes are often generated by faults that are under shear stress due to regional tectonics or due to past mining (or other human-related) activities. An increase of water pressure along a stressed fault can reduce the normal stress and friction along the fault and this can induce the fault move- ment. The rising ground water as earthquake trigger can be understood as well as an energy shift. The increase in mass or mass shift could be an additional driving force of fault move- ment or seismicity trigger. In South-Limburg, two earthquake swarms have occurred at the southern edge of the former coal mining area, around the village of Voerendaal. The first swarm occurred in 1985 and the second swarm at the end of 2000 until the beginning of 2002. It is unclear if these two swarms can be related to the rising mine water or are purely tectonic. The coal was mined from Carboniferous rocks, at varied depths ranging from 0 to 800 m below ground surface (De Vent and Roest, 2013). The top level of the Carboniferous rocks has a downward trend towards the northwest. These rock formation is dissected by various faults, which are part of the southern boundary of the Roer Valley Graben. This research aimed to identify a possible relationship between the increase in groundwater and seismicity in South Limburg and determine its possible impact on the existing seismic risk in the region. To understand and identify possible relations, the following input data were compiled: – Earthquake catalogue; – Ground water level measurements; – Ground deformation from interferometric synthetic aperture radar (InSAR) data; – Scientific literature from the region and other induced seismicity cases. The approach followed to analyse the data comprised different steps: – Documentation of reference cases; – Geological and tectonic setting and seismicity; – Identification of swarms or anomalous trends in the seismicity catalogue; – Multivariate analysis (temporal variation in mine groundwater, seismicity, ground uplift); – Stress-state analysis at different depths of the active faults to evaluate slip potential; – Energy-balance analysis of the system. 2 REFERENCE CASES In Germany and Belgium, there is a similar history of coal mining as in The Netherlands. In the former coal mine region Campine in Belgium mines have been flooded, but no earth- quakes have been measured (Rosner, 2015). In Germany, parts of the Ruhr and Saar mining areas are being flooded. In the Ruhr region, no flooding-induced seismicity has been observed so far (Fritschen, 2015). Active mining in the Saar area came to an end in June 2012. Since May 2013 mines have been flooded and on 15 September 2014 the strongest flooding-induced seismic event so far was measured with M2.7 (Saarland, 2014). Another example of flooding-induced seismicity in France is the Gardanne coal mining region in the Provence. Seismicity measurements started in 2008 (Matrullo et al., 2015) and seismic events were measured in November 2012 and in December 2014 (M>2.5). Goldbach (2009) describes experiences with flooding-induced seismicity in South African gold mines. Here mining- induced seismicity was measured in the active part of the mine and flooding-induced seismicity in the abandoned part of the mine. Comparing mine-water level increase with seismicity, it was found that the majority of events took place about 14 months after the start of flooding. 3 TECTONIC SETTING AND SEISMICITY 3.1 Tectonic setting, faults The former Dutch coal mining district is situated at the southwestern boundary of the Roer Valley Graben, south of the Feldbiss Fault. At a regional scale, the Roer Valley graben is part 5020 Figure 1. Active faults and top surface level of the Carboniferous (bed)rock. of the Rhine Valley Graben. Along the Roer Valley Graben, the governing active faults are the Feldbiss Fault and Peel Boundary Fault. These faults are active since late Oligocene (Geluk et al., 1994) and have an estimated slip rate of 0.05–0.01 mm/year (Giardini et al., 2013) and are dominantly normal with a small strike-slip component in depth (Dost & Haak, 2007). Locally, the active faults found in the area correspond to the southern boundary of the Roer Valley Graben: the north-west-south-east faults Feldbiss, Heerlerheide, Benzenrade and the west-northwest Kunrader Fault. These have a steep NE dipping direction and tens of meters vertical offset in the geological formations (Dinoloket, DGMdiep v4.0). The source mechanism is normal but Camelbeeck (1994) reports a strike-slip movement in the Kunrader fault. Dost & Haak (2007) propose two hypotheses: decoupling of the crust, where the upper crust is more brittle and presence of deeper faults with different orientation. Until now, the seismicity registered in the Limburg area has been related to tectonics only. The Dutch seismological survey (KNMI) reports the nearest induced seismicity only in the north of The Netherlands or the western part of Germany due to gas extraction and coal mining. The largest tectonic earthquakes registered in the area occurred in April 13, 1992 in Roermond (The Netherlands) and in March 14, 1951, in Euskirchen (Germany), 80 km SE from Roermond. Both earthquakes had a magnitude ML=5.8 and an intensity Mercalli VII, (Dost & Haak, 2007). The Roermond earthquake and most of the seismicity from the region is associated to the Peel Fault, the northern boundary of the graben. The Kunrader Fault, at the south-western side of the graben, has associated seismic swarms during 1985–1986 and 2000–2002 (Dost & Haak, 2007).