| Abstract |
Earthquake hypocentres constitute a unique source of information for understanding the physical processes at the origin of earthquakes, describing the subsurface and quantifying earthquake seismic hazard. For example, it is the primary attribute of an earthquake without which other characteristics such as occurrence time, seismic moment, magnitude or focal mechanism cannot be determined. Tectonic interpretation or fault identification using the spatial distribution of earthquakes is also the purpose of many studies. However, earthquake location errors exist and need to be properly quantified for reliable result interpretation. Therefore, in this work, the impact of several factors on the absolute location of earthquakes in a reservoir was investigated. The methodology, which used state-of-the-art techniques, consisted in relocating synthetic hypocentres under hypotheses different from the data modelling step, and comparing the initial and relocated hypocentres. Hence, the effects of the p- and s-wave onset time uncertainties and inaccuracies were examined as well as the effects of the velocity model uncertainties and inaccuracies. In particular, we looked at the location errors driven by using a 1d velocity model instead of several 1d or 3d velocity models. The analysis was applied to the rittershoffen geothermal field, where seismicity was induced between 1000 m and 5000 m, under the seismic monitoring conditions existing during the chemical and mechanical stimulations of the well grt1. The 3d analysis of the location uncertainties and inaccuracies covers a zone of a couple of kilometres around the open-hole section of the well. An overview of the different areas and networks used during our study are shown on figure 1. |