Optimized Experimental Network Design for Earthquake Location Problems: applications to geothermal and volcanic field seismic networks

• Published in: Journal of Volcanology and Geothermal Research
• Main Authors: Toledo Zambrano, T., Jousset, P., Maurer, H., Krawczyk, C.
• Format: Article in Journal/Newspaper
• Language: English
• Published: 2020
• Subjects: Reykjanes; Iceland
• Online Access: https://gfzpublic.gfz-potsdam.de/pubman/item/item_3755918; https://gfzpublic.gfz-potsdam.de/pubman/item/item_3755918_6/component/file_5002437/3755918.pdf

We constructed a network optimization scheme based on well established survey design tools to design and qualify local and regional microseismic monitoring arrays dedicated for geothermal exploration and volcano monitoring. The optimization routine is based on the traditional minimization of the volume error ellipsoid of the linearized earthquake location problem (D-criterion) with the twist of a sequential design procedure. Seismic stations are removed one by one to obtain networks for constraining the locations of multiple hypothetic earthquakes with varying local magnitudes. The sequential approach is simple and allows the analysis of benefit/cost relations. Cost curves are computed for all hypothetic events to reveal the minimum optimal number of stations given specific design experiment objectives. The scheme is first demonstrated on three test design experiments. Later, we use the routine to augment an existing seismic network for monitoring microseismicity in a geothermal field in NE Iceland (Theistareykir). The resulting 23 station network would become the backbone of a reservoir behavior and exploitation activity study. Hypothetic event locations and magnitude relations are taken from a previous regional seismicity study and coincide with geothermal injection and production areas. Sensitivities are calculated with a known 1D velocity model profile using a finite-difference back-ray tracer, and body wave amplitudes are computed from known local magnitude relations. Finally, expected earthquake location accuracies are calculated via multiple Monte Carlo experiments. The design routine is later used to qualify an existing seismic network located in SW Iceland (Reykjanes). The seismic array is reduced to strategic positions, and benefit and expected accuracies are quantified to observe whether costs could have been optimized had a previous network design experiment been performed. Overall, we explore quick and flexible tools for designing and qualifying networks for many applications at various scales.