Comparison of short- and long-periodP-wave travel times and arrival angles

Summary Using about 100 earthquakes recorded on both the short-period and long-period matched three-component seismograph systems at Umeå (UME) in northern Sweden, relativeP-wave travel times (LP-SP) are measured and arrival angles (incidence and azimuth) are determined from relative amplitudes ofP. Long-period onsets appear to register consistently later than short-period ones by a couple of tenths of a second. This figure is of the same order as the uncertainty in T _LS (difference between LP and SP onset times), but still is shown to be of some statistical significance. The effect is believed due, at least in part, to the relative inability of the long-period seismometer to visibly register an initial shorter-period lower-amplitude focal phase from some earthquakes.The dip on crustal interfaces implied by the short-period azimuth and incidence-angle anomalies is 7 to 15 times greater than that implied by the long-period anomalies. This effect is believed due to undulation or irregularity of, foremost, the Moho, as well as other crustal boundaries. The wavelength of, say, Moho undulations would seem to be limited above and below by considerations of the dominant long-period wavelengths (100 km) and of the diameter of Moho sampled by the short-period waves (50 km). The possible undesirable effect of the crustal layering on the components of surface motion, and thus on the arrival angles, is discussed and cautiously dismissed, due to the absence of sedimentary layering, in favour of the above interpretation.