Volume backscattering functions and optical extinction coefficients are computed for eight suggested major cloud models using the Mie theory for optical wavelengths of 0.488 micro, 0.694 micro, 1.06 micro, 4.0 micro, and 10.6 micro. Results show that there is no clear advantage of one wavelength over another for improving cloud transmission; however, backscattering is significantly reduced at the longer wavelengths. Variations in the optical properties of clouds are also discussed and calculations summarized to indicate the effects of cloud thickness, inhomogeneity, and geographical location on the backscatter function and extinction coefficient.

S U M M A R YWe model seismic wave propagation from intermediate depth earthquakes in a subduction zone using a 2-D Chebyshev pseudospectral method. Particular attention is directed to the influence of a deep, low-viscosity subduction channel on top of the plate contact where metamorphic rocks may be exhumed by forced return flow. The study is motivated by observations of complicated dispersive and high-amplitude P-and S-wave trains in the forearc of the Hellenic Subduction Zone. The basic model is a subducted slab with a thin oceanic crust forming a lowvelocity layer. Our model setup closely follows recent results on the structure of the Hellenic Subduction Zone obtained from receiver functions and surface wave studies. They exhibit an abrupt change of the dip of the downgoing slab at about 70 km depth. The subduction channel is modelled as a thin, wedge-shaped layer of intermediate seismic velocity above a slower oceanic crust and below a faster overlying mantle wedge. We also look into the effects of a continuous phase transition from basalt to eclogite in the subducted oceanic crust and near-surface crustal structures. In all models, wave propagation is characterized by dispersive guided channel waves trapped in the low-velocity subducted crust. They produce high-amplitude arrivals in the forearc. A fast guided wave train (gP) originates from the direct P wave and a slower one (gS) from the direct S wave. Guided waves are radiated into the overlying mantle where the dip of the slab is abruptly changing. Seismogram sections for models without a subduction channel typically show two spatially separated guided wave trains, one following the oceanic crust and one travelling more steeply towards the forearc high. A subduction channel above the plate contact enhances the radiation effect of gP waves at the slab bend due to the weaker velocity contrast and inhibits the separation of the wave trains. In models with additional near-surface crustal structures the wave field is dominated by reverberations. However, guided waves remain discernible in seismogram sections because of their high amplitudes. When a basalt to eclogite phase transition is considered, guided waves are preferentially generated in the presence of a subduction channel. A pronounced gP-wave train develops particularly for sources inside this channel. On the contrary, guided waves are hardly distinct for sources below the subducted crust.

S U M M A R YWave propagation in coal seams is numerically modelled in order to identify approaches towards the reconnaissance beyond the heading face of an advancing coal mine roadway. Complete synthetic wavefields including P-SV body waves and Rayleigh-type seam waves are calculated using a Green's function approach for simple, laterally homogeneous models and a parallel elastic 2-D/3-D finite difference modelling code for more realistic geometries.For a simple three-layer model the wavefield within the seam is dominated by a fundamental Rayleigh seam mode symmetrical with respect to the centre of the seam on the vertical component and antisymmetrical on the horizontal component. If the seam contains an interleaved dirt band with higher velocities and density, higher modes dominate the wave propagation, depending on the thickness of the dirt band. Wave propagation in laterally inhomogeneous coal seam models with disturbances like seam ends, faults, thinning, washouts and seam splitting is strongly influenced by the type of disturbance. Amplitudes of seam waves reflected from these disturbances strongly depend on the fault throw and the degree of thinning or washout. In some cases, conversion to higher modes can occur. In all investigated models, those Rayleigh seam wave phases are preferably reflected, which have frequencies above the fundamental mode Airy phase. Lower frequency phases are preferably transmitted. However, seam waves are not reflected from a seam splitting disturbance. Thus a detection of seam splitting with reflected seam waves appears to be impossible. FD computations for 3-D models containing an ending tunnel parallel to the seam and a source beyond the heading face of the tunnel show that seam waves are converted into Rayleigh waves at the tunnel face. They propagate along the surface of the tunnel and interfere with the seam waves propagating beside the tunnel. This effect has to be taken into account for subsequent treatment of experimental data, where the locations of sources and receivers are restricted to a small seismic layout in the vicinity of the tunnel. As tunnel surface waves have slightly lower frequencies than seam waves, it may be feasible to separate tunnel waves from the seam wave reflections, particularly because higher frequency phases of the seam wave are preferably reflected at seam disturbances. Polarization analysis showed, that the elliptically polarized Rayleigh-type seam waves in the vertical-radial plane can be distinguished from Rayleigh tunnel waves propagating on the sidewall of the tunnel adjoining the coal layer with elliptical polarization in the radial-transversal plane.

No abstract