P- and S-wave reflections from anomalies in the lowermost mantle

Abstract

Seismic P and S waves recorded at the GRF array in Germany are used to study the inhomogeneous structure of the boundary layer D″ at the base of the mantle. The use of the seismic array allows the detection of small-scale anomalies in the lowermost mantle. The lateral resolution attainable is about 100 to 400 km, i.e. more than 10 times better than with tomographic methods. The analysis of 13 years of GRF broad-band array data yields 255 events with high signal-to-noise ratio which are used to map the lowermost mantle. 74 of these events show anomalous P waves (PdP) which arrive 3–6 s after the direct P wave and have a slowness 0.7–0.8 s/° smaller than the slowness of direct P. Other events close to the events with PdP do not have such an anomalous phase. Using slowness, traveltime, amplitude and waveform information it is demonstrated that PdP is caused by an anomalous lower mantle velocity structure below the turning point of the P wave. If the pp phase (i.e. a P wave first reflected at the free surface near the source) is used together with the P phase, distinct and well-sepaarted P-velocity anomalies can be determined under the Nansen Basin, the Kara Sea and northern Siberia. The areas of the bounce points of PdP in the lower mantle have a lateral extension of about 100 by 200 km, but this is not the size of the anomaly, since the resolution of P waves at 1 Hz in this depth is 130 km by 260 km (Fresnel zone). The accuracy to which the depth of the reflector (2612 km under the Nansen Basin and 2605 km under northern Siberia respectively) can be determined for 1-D models is ±10 to 20 km. The velocity contrast at the lower mantle discontinuity is about 3 per cent ±1 to 1.5 per cent. Areas which have velocity fluctuations smaller than about 1 per cent can not be detected as anomalous areas. 2-D models of the anomalies reveal the range of adequate models and possible trade-offs. If the lateral extension of the anomaly is about 7° the reflector has to be 40 km deeper than in the 1-D model. If the dip of the reflector in the lowermost mantle is only about 1.5° it is difficult to resolve if the reflector is tilted towards the source or towards the receiver. For the anomaly under the Nansen Basin deviations from the great circle path are observed for PdP, indicating 3-D effects. Below the three anomalies the core-mantle boundary (CMB) can be located by PcP with a depth in agreement with standard earth models under northern Siberia has a region where P- and S-velocity anomalies coincide, but also a region where only an S-velocity anomaly but no P-velocity anomaly is observed. This results in changes of the Poisson ratio from +5.9 per cent to −4.8 per cent (±1.5 per cent) across the discontinuity in the lowermost mantle for the regions studied. The analysis of the S waves (S, SdS and ScS) reveals two S-velocity anomalies in the lowermost mantle under northern Siberia. The first anomaly coincides with the P velocity anomaly under northern Siberia and can be explained by a 1-D model with a reflector depth of 2610 km ± 15 km and a velocity contrast of 2.3 per cent. The second S-velocity anomaly is in an area where no P-velocity anomaly can be detected. The corresponding 1-D S-velocity model has a reflector depth of 2575 km ± 15 km and an S-velocity contrast of 2.6 per cent. The smallest structures that can be resolved with the S waves in this depth are about twice as large as for the P waves, i.e. 230 km by 460 km (Fresnel zone). The joint analysis of P and S waves therefore shows a region with a P-velocity anomaly together with weak indications for an S-velocity anomaly (Nansen Basin) and a second region with a P but no S anomaly (Kara Sea). The lowermost mantle under northern Siberia has a region where P- and S-velocity anomalies coincide, but also a region where only an S-velocity anomaly but no P-velocity anomaly is observed. This results in changes of the Poisson ratio from +5.9 per cent to −4.8 per cent (±1.5 per cent) across the discontinuity in the lowermost mantle for the regions studied.