The North Sea Lg-blockage puzzle

Abstract

SUMMARY The North Sea Lg blockage for wave paths across crustal graben structures is a well-established observational fact. Analysis of such observations implies that Lg blockage takes place in graben areas associated with sedimentary basin formation and crustal thinning. These intriguing observations have triggered many theoretical studies aimed at highlighting specific Lg loss mechanisms, albeit so far with only moderate success. Our approach to this problem is to simulate seismic wavefield propagation through the crustal waveguide using 2-D finite-difference techniques. The graben structures are known in detail from oil exploration works in the North Sea, which has enabled us to use realistic crustal models in our Lg synthetics. In the most extreme model tested, the crystalline crust thickness beneath the graben amounted to only 5 km, while the overlying sedimentary pile is nearly 10 km thick. At the base of the crust in the graben area the Moho is elevated nearly 10 km. This model has similarities to the oceanic crustal waveguide, where total Lg blockage is claimed for path lengths exceeding 100 km. The synthetic wavefields are displayed in terms of snapshots, semblance velocity analysis and time-space rms amplitudes. The dominant structural Lg loss mechanisms are the delay of the Lg waves in the thick sediments, Lg-to-Rg conversions (scattering) by lateral heterogeneities in the sediments, and S-wave leakage out of the crustal wave-guide and into the upper mantle. A fraction of these upper-mantle S waves return to the crust and appear as Sn coda. Observationally, strong Sn phases of long duration are often associated with weak Lg phases and vice versa. Our synthetics produced Lg amplitude decay amounting at most to 6-10 dB, while observational data imply blockage amounting to 15-20 dB. The latter is equivalent to a Pn-Lg magnitude difference of nearly one magnitude unit. The main outcome of this study is therefore that Lg-wave propagation is very robust and that a dominant blockage effect associated with intrinsic attenuation, that is Q values of the order of 100 at 2 Hz for a path length of minimum 100 km, is necessary to conform to observations.