Generation of High Resolution Seismic Catalog Associated With the Production Phase 2021 - 2022 at the Balmatt Geothermal Site

The Balmatt geothermal doublet, developed and managed by VITO (Flemisch Institute of Technological Research), targets the fractured Lower Carboniferous Limestone reservoir in the Campine Basin at the depth of 3000 m to 4000 m. The development of the project started in 2015 and the operation began in 2018. The geothermal plant consists of two active wells, one injection well and one production well. The geothermal production had to be suspended after the occurrence of a stronger M_L 2.2 event on the 23^rd of June 2019 which triggered a red alert status on the local traffic light system (TLS). Production was then resumed in April 2021, following an extension of the seismic monitoring network and an update of the TLS. Activities were suspended again in November 2022 after another strong M_L 2.1 event was induced. Thanks to the network extension, current investigations aim at understanding in detail the main structural features (active faults) and hydromechanical processes involved in the generation of such larger events which will contribute to improving seismic forecast possibilities for future monitoring operations. Here we present insights into ongoing data processing to create a high resolution unbiased (complete) seismic catalog providing the basis for future interpretation of the spatio-temporal and energetic behavior of seismicity towards different production settings. Our current work focuses in particular on the development of an automatic detection routine based on continuous data of the deep borehole sensor (installed at the depth of 2052 m) by combining a machine learning based automatic events detection algorithm and template matching method. The events detection in the continuous data is complicated by the periodic malfunctioning of the sensor and the presence of aseismic noise which leads to the large number of false events detection. To address this issue and to minimize the number of false detections, we employ frequency and amplitude analysis of the seismic data. Secondly we analyze source attributes of the detected events which involve source mechanism inversion and source parameter determination as well as clustering analysis and constraining source location for noisy small magnitude events. Further more the comparison between the production data (injection pressure, temperature, volume etc.) with the results from the seismic analysis will provide us with better constrain on the hydromechanical characteristics of the reservoir and the relation between the geothermal operations and seismicity at Balmatt geothermal site.