Virtual Sources of Body Waves from Noise Correlations in a Mineral Exploration Context

Abstract

Abstract The extraction of body waves from passive seismic recordings has great potential for monitoring and imaging applications. The low environmental impact, low cost, and high accessibility of passive techniques makes them especially attractive as replacement or complementary techniques to active-source exploration. There still, however, remain many challenges with body-wave extraction, mainly the strong dependence on local seismic sources necessary to create high-frequency body-wave energy. Here, we present the Marathon dataset collected in September 2018, which consists of 30 days of continuous recordings from a dense surface array of 1020 single vertical-component geophones deployed over a mineral exploration block. First, we use a cross-correlation beamforming technique to evaluate the wavefield each minute and discover that the local highway and railroad traffic are the primary sources of high-frequency body-wave energy. Next, we demonstrate how selective stacking of cross-correlation functions during periods where vehicles and trains are passing near the array reveals strong body-wave arrivals. Based on source station geometry and the estimated geologic structure, we interpret these arrivals as virtual refractions due to their high velocity and linear moveout. Finally, we demonstrate how the apparent velocity of these arrivals along the array contains information about the local geologic structure, mainly the major dipping layer. Although vehicle sources illuminating array in a narrow azimuth may not seem ideal for passive reflection imaging, we expect this case will be commonly encountered and should serve as a good dataset for the development of new techniques in this domain.