Using Three-Component Data for Seismic Interferometry Studies at the Kylylahti Mine, Eastern Finland

The reflection seismic surveying method is useful when conducting mineral exploration in the crystalline bedrock because of its good depth extent and resolution. However, the traditional experiments with active sources are expensive and difficult to carry out, especially in remote areas or in conservation areas where mineral exploration is limited due to environmental reasons. Recently, a number of theoretical advances have proven that passive soundings utilizing ambient seismic noise can provide new opportunities for seismic imaging and contribute to data generation for reflection seismic surveys, without the need for explosive or vibratory sources. One of the most promising new methods is seismic interferometry (SI), where the impulse response between two receivers is reconstructed by correlating their signals with each other. COGITO-MIN is a joint project between the University of Helsinki, the Geological Survey of Finland, Polish Academy of Sciences, and industrial partners with the aim of investigating and developing new cost-effective seismic exploration methods in the crystalline bedrock. Within the framework of the project, a passive seismic experiment was carried out in which 45 three-component geophones were deployed for a month in the vicinity of the polymetallic Kylylahti Mine in Polvijärvi, northern Karelia, where the mining operator is the Swedish metal company Boliden. The original purpose of these geophones was to collect data suitable for detecting underground cavities related to underground nuclear explosions. The institute that collected the data was CTBTO (Comprehensive Test Ban Treaty Organization) whose task is to monitor the treaty in the pre-ratification stage. The purpose of this Master's thesis was to develop an SI workflow for the three-component data and to investigate the method's performance in an area where local geology is known after nearly 40 years of exploration and consequent mining operations. The specific scientific objectives of the thesis are (1) to demonstrate the usefulness of collecting three-component data in conjunction with or instead of single-component data, (2) to assess the noise-based SI methods used in previous studies and to improve their stability in the crystalline bedrock, and (3) to investigate the possibilities of SI from an operational perspective. Seismic velocities obtained through laboratory measurements were merged with geological and density models of the target area provided by Boliden. The resulting velocity and density grids were then used as the basis for waveform modelling, and the results from SI were validated against them. The starting point for SI was the noise-driven approach where 'each sample matters'. The interferometric workflow is built on the Seismic Unix suite together with self-written algorithms that are based on theoretical evaluations. SI is followed by an imaging workflow, which provides the basis for the reflectivity profiles. The thesis work focuses on five components of the Green's tensor and the vertical, radial and transverse component of the impulse response. With the horizontal components, one can access the S-wave patterns in addition to the P-waves. As a specialty, the so-called sign bit normalization (SBN) method was also tested. The technique involves destroying much of the amplitude information of the original seismograms by only retaining the sign bit of each sample. According to the results outlined in this thesis, SBN can make it easier to image the weak reflectors of the subsurface. This type of seismic interferometry seems particularly suitable for the early stage of mineral exploration, where the explorer does not yet fully understand the target they are studying. The most important advantage of seismic interferometry, however, is its cost effectiveness, and its potential for reducing risks for the environment.