Teleseismic signal extraction

Abstract

Additive ambient seismic noise in the 0.5 to 2.0 c/s range can be reduced by ‘isotropic’ receiver systems employing horizontal arrays (with 5 to 10 km aperture), or reasonable vertical arrays (with some limitations), probably to teleseismic body-wave noise level (about - 15 dB relative to 1 nm 2 (c/s) -1 at 1 c/s) in reasonable areas. Very large aperture arrays of these receiver systems are expected to detect and extract teleseismic signals below the teleseismic (or mantle P wave) noise level. ‘Signal-generated noise’ resulting from scattering should be reduced by the same approach. Alteration of teleseismic signal waveforms by reverberation at the receiving station is important at some, though probably not all, receiving sites. Appropriate linear filtering which may be dependent on epicentral location should be suitable for adequately reducing this effect. A 0.5 to 2.0 c/s (or broader) representation of the teleseismic signal as it appears in the upper mantle at the receiving location, with additive noise reduced to a hypothetically predictable level, should be achievable at a number of locations by an appropriate combination of isotropic trapped and leaking noise reduction subsystems, built into large aperture directional scanning arrays, with filtering to reduce reverberation. An attempt to exploit the large time × bandwidth product of ‘complex’ teleseisms by coherent processing of signals emanating from different directions in the source ‘radiation pattern’ was unfruitful.