Rayleigh wave focal spot imaging of the Kylylahti ore deposit

  

Seismic imaging has traditionally been applied to far-field signals using earthquake tomography and, in more recent decades, ambient noise tomography. However, with modern large-N seismic arrays it has been demonstrated that structural observations at sub-wavelength distances are possible utilizing the surface wave focal spot imaging method. The focal spot is the zero-lag correlation amplitude field derived from dense array noise correlations. One advantage compared to noise tomography is the ability to analyze waves at length scales that are large compared to the array size, which enhances the depth resolution. The lateral resolution of focal spot imaging also improves at short distances when the station density is high. The focal spot is the time domain representation of the spatial autocorrelation field (SPAC). We can thus use established analytical SPAC methods to constrain local Rayleigh wave phase velocity estimates from the focal spot shape. Here, we investigate the effectiveness of the method in the context of critical raw material exploration. We apply focal spot imaging in the Kylylahti polymetallic mining area, hosting sulphide ore deposit in ophiolite-derived rock assemblage in the Finnish Outokumpu belt, located in eastern Finland. Our study utilizes a dense array dataset acquired in August-September 2016 within the ERA-MIN COGITO-MIN project. Our passive focal spot imaging extends the original objective of the project to produce structural imaging of the Kylylahti area using 2D seismic reflection profiles, 3D body-wave reflection seismic interferometry, and sparse 3D active-source survey. The array consisted of 994 stations that covered an area of 10.5 km² and featured lines spaced 200 m apart with 50 m receiver spacing. Each of the 994 stations consisted of six co-located 10 Hz vertical-component geophones and a data logger. Seismic noise was recorded at 500 Hz for 20 hours per day over a 30-day period, generating approximately 600 hours of passive seismic data. As part of the project, numerous active shots were conducted, which we found to disturb the noise records. To enhance the quality of the noise correlations by ensuring better compatibility with the diffusivity assumption of the ambient wavefield we remove the project-related shot data from the noise records using cataloged shot information. We decimate the daily records to 125 Hz and apply standard noise tomography pre-processing steps. We calculate the cross-correlations of the 1-hour time windows and stack them linearly to create 30 s maximum lag cross-correlations. The noise correlations are filtered using a Gaussian filter around the center frequency and the focal spot is parameterized with the SPAC Bessel function model to estimate the Rayleigh wave phase velocity. We present phase velocity maps produced at different periods, which allow us to characterize the geology of the Kylylahti mine area in the top kilometer. The effectiveness and accuracy of this method is demonstrated by comparing our results with previous active and passive imaging results.