AN INVESTIGATION OF THE UPPER MANTLE BENEATH NW AUSTRALIA USING A HYBRID SEISMOGRAPH ARRAY

A large array of short‐period, portable seismographs was operated in the Northern Territory of Australia for 3 months in 1986 in order to record earthquakes in the island arcs to the north and east of Australia. The array consisted of 18 digital recorders, 28 analogue tape recorders and the 20 stations of the permanent Warramunga (WRA) seismic array. An unusual aspect of this experiment was the variable station spacing and apertures of the different elements of the hybrid array. The WRA and portable digital arrays had station spacing of 2.2 km and 5–13 km, respectively, and allowed confident identification of phases returned from the upper mantle, whilst the analogue array had station spacing of 40 km and allowed us to track mantle phases across its 500 km aperture. Seismograms from 17 shallow earthquakes (mb 4.3–5.1) in the Indonesian arc are used to investigate the P‐wave velocity structure of the upper mantle beneath NW Australia. We combine seismograms from these events into a composite record section covering ranges from 1000 to 2600 km. Strong phases following the first P‐wave arrivals in this composite record section clearly indicate the presence of significant structure in the upper mantle. Lateral heterogeneity in the upper mantle causes the timing and amplitudes of mantle phases to vary across the array and among earthquakes. In order to minimize the effects of lateral heterogeneity, we filter and stack the data and concentrate on features in the data that are seen for many individual seismograms and for several earthquakes. We calculate WKBJ and reflectivity seismograms in order to construct a vertical velocity profile that fits the observed traveltimes and waveforms in an average sense. Our preferred model NWB‐1 includes second‐order velocity discontinuities of 3.7 per cent near a depth of 200 km, 6.1 per cent near 400 km and 3.7 per cent near 620 km in order to reproduce the amplitudes of the later phases. The low amplitudes of the first arrivals in the range from 1600 to 2200 km require either a low‐velocity zone below 230 km depth or a low velocity gradient between 230 km and the 400 km discontinuity. Model NWB‐1 is smoother than some models that have been previously proposed for this region which may have mapped lateral heterogeneity into vertical velocity profiles. Copyright © 1990, Wiley Blackwell. All rights reserved