The capabilities of three-component (3-C) and array stations for estimating azimuth and slowness are compared for short-period P-type phases recorded at the NORESS array. For vertical array data, azimuth and slowness estimates are obtained from broadband frequency-wavenumber (f-k) analysis. For 3-C data, polarization analysis is performed. The data processing is automated, using arrival time and dominant frequency information from the NORESS Bulletin. Independent determinations of azimuth and/or slowness, obtained from locations in the NEIS or regional network bulletins, are used as reference estimates. Over 100 events are analyzed, both teleseismic and regional. They were selected from a variety of distances and azimuths, and cover a wide range of signal-to-noise ratios (SNR).
The capability of 3-C stations for azimuth and slowness estimation critically depends on SNR. For SNR below a threshold of ∼2, the scatter in the estimates is very large for both parameters, and the slowness of teleseismic events tends to be overestimated. Also, the results are site-dependent within the NORESS array. The array measurements obtained with the broadband f-k method are not significantly affected by noise at the levels of SNR considered. For events with sufficient SNR, both methods compare well, and only a slightly better performance is observed with the f-k method.