Delineation of the target fault-zone for Koyna scientific deep drilling by accurate location of microearthquakes

Seismicity induced by impoundment of water reservoirs originates from the response of the reservoir to initial filling as well as water level changes. The Koyna area in western India is the most prominent site of reservoir-triggered seismicity. Seismicity started to occur shortly after the impoundment of the Shivaji Sagar Lake in 1962 and continues until now. On 10th December 1967, an earthquake with a magnitude of M 6.3 occurred near the site of the Koyna dam and was felt for a radius of 700 km. It was among the greatest earthquakes of historical time in the Indian peninsula, claimed at least 180 lives and injured over 1,500 people. More than 80% of the houses were damaged in the township. A better understanding of the earthquake triggering by impoundment of the Koyna reservoir is indispensable to avoid future catastrophes and will additionally pave way for understanding earthquakes genesis in general. The drilling of 10 exploratory boreholes started in December 2012; each borehole is drilled to a depth of about 1200 m to 1500 m. Eight boreholes are equipped with seismometers. A pilot hole has been completed down to a depth of 3 km and will be equipped with a geophone string and temperature sensors. In collaboration with the International Continental Scientific Drilling Program drilling of a deep main hole is planned. Precise earthquake locations are a prerequisite for the design of the trajectory of the planned deep borehole observatory as the borehole must intersect the plane of an active fault. In addition to the borehole network, a surface network consisting of 23 broadband stations has been installed in the area. The results of the project can be summarised as follows: 1) NORSAR licensed an in-house developed software with state-of-the-art processing algorithms for automatic seismic event detection to NGRI as well as a leading-edge software for 3-D raytracing. For both, on-site training was held for which NORSAR?s researchers visited NGRI facilities in India. 2) Parameters that produce the best results for the detection and location of seismic events in the Koyna region were investigated. Using NORSAR?s software, the number of detected events compared to the previous visual inspection procedure, could be increased threefold. 3) Difference in event location uncertainties were analysed in relation to using only surface records, only borehole records or a combination of both. Surface stations provide a better constraint for the estimation of epicentres, whereas the incorporation of borehole data improves depth estimates. 4) To improve location uncertainties further, we applied a relative location method based on waveform similarity, improving the definition of the orientation of the geological structures responsible for the seismicity as well as uncovered the existence of two potential fault planes being seismically activated. Such information will be valuable for the planning of the borehole trajectory. 5) Both ray paths and full waveforms were modelled employing several 1D velocity models from literature to enhance understanding of the recorded seismograms. We note the influence of the overlying basalt thickness, the influence of the source depth on recorded wave phases, the receiver depth effect on incidence angles as well as recorded amplitudes and the potential importance of diving waves. Especially events with shallower source depth show large differences in their waveforms for different velocity models and might therefore be of specific value for tomography.

NGRI obtained NORSAR's in-house software allowing to switch from manual to automatic data processing. The introduction of state-of-the-art seismic processing methods is anticipated to further the understanding of the geological structures that are being seismically activated and to provide useful information for the planning of the deep borehole observatory. We believe that the establishment of a common base of trust could only be achieved through long-term personal relations that facilitated the crossing of cultural differences between India and Norway through frequent discussions. The project contributed towards international efforts towards managing hazards posed by induced/triggered seismicity through fundamental understanding of reservoir-triggered seismicity. The management of seismic hazards is indispensable for safe exploitation of water reservoirs in the Maharashtra region, India's leading state in terms of agriculture as well as its largest power generating state.

Induced earthquakes have been observed as a result of the impoundment of reservoirs, oil production, or the injection of fluids into underground formations. The Koyna area in western India is the most prominent site of reservoir-triggered seismicity. Seismicity started to occur shortly after the impoundment of the Shivaji Sagar Lake in 1962. The 10 December 1967 M 6.3 earthquake is the largest triggered event so far in this region, but more than 22 events with M > 5 and 200 events with M > 4 have occurred since 1967. The drilling of ten exploratory boreholes started in December 2012; presently, six boreholes are equipped with seismometers at depths of 1.2 - 1.5 km, two more will be installed. In 2017,one pilot hole has been completed down to a depth of 3 km and will be equipped with geophone strings.One additional pilot hole will be realized in near future. In collaboration with the International Continental Scientific Drilling Program (ICDP) drilling of a main hole is planned. In addition, a surface network consisting of 20 broadband stations has been installed in the area. Monitoring induced seismicity provides unique in-situ information about subsurface pressure variations and precise microseismic event locations can be used to map tectonically active faults and fluid migration patterns. While microseismic event locations are nowadays obtained routinely, a particular challenge is the correct estimation of source parameters. The principle advantage of borehole seismic data is that a broader frequency band of the recorded wavefield can be used for analysis due to recordings of higher frequencies. In addition, geophones in a borehole are placed closer to the target zone and allow monitoring at substantially reduced noise conditions. We will focus on automatic methods to locate microseismicity on continuous data from the borehole seismometers. Potential patterns of microseismicity close to the detection threshold will extend the range of earthquake observation and might define areas of concentrated deformation or weakness zones.