Teleseismic P coda analyzed by three-component and array techniques: Deterministic location of topographic P-to-Rg scattering near the NORESS array

Abstract

Abstract P -wave coda of 75 sec duration from eight teleseismic events recorded by the NORESS array, Norway, were examined using both array and three-component analysis techniques. The array data have been analyzed using the semblance technique that allows us to determine accurately the time of arrival, apparent velocity, and azimuth of the scattered waves. Coda coherency was found to vary considerably, probably reflecting extended source duration for some events. Slow beam amplitude decay rates support significant coda generation within the source region although source-end scattering could not be separated from the source pulse per se . At the receiver end of the path, we resolved locally scattered wavelets even within highly coherent teleseismic P coda by subtracting the teleseismic array beam from the records. The majority of receiver-end scattering contributions appear to be P -to- Rg conversions occurring in both forward and backward directions from two nearby areas with pronounced topographic relief, namely Bronkeberget (distance ∼ 10 km, azimuth ∼ 80°) and Skreikampen (distance ∼ 30 km, azimuth ∼ 225°). The scattering is multiple in the sense that both primary and secondary phases from the source region contribute to the scattered Rg waves. P -to- S scattering constitutes a significant part of the receiver-end scattering observed later in the P coda, although it is more diffuse than Rg and forward scattering is more dominant. P -to- P scattering is weak and mostly confined to the immediate vicinity of the array. The analysis of individual three-component records shows high sensitivity to interference from locally scattered waves. This interference is probably responsible for a marked decrease in apparent velocity observed for some events 3 to 4 sec after P onset.