FLIEDNER, MORITZ M., ROBERT S.WHITE, and JUERGEN FRUEHN, University of Cambridge

Abstract: Determining the Structure of Basalt Flows with Wide-Angle Seismic Data

Introduction

High-velocity basalt flows create a major obstacle to imaging lower-velocity underlying sedimentary structures. The highly reflective top of the basalts scatters much of the incident seismic energy; short-period ringing, simple and peg-leg multiples obscure weak sub-basalt reflections with similar move-out; the high-velocity basalt layer absorbs preferentially the higher frequencies in the incident wavelet, degrading the achievable resolution of a sub-basalt image; and strong ray-bending may distort the sub-basalt image.

Recording the seismic wave-field at longer offsets (with super-long arrays or by two-ship marine acquisition) may improve the chances for sub-basalt imaging in several ways: it allows the recognition of sub-basalt low-velocity layers by their shadow-zone effect on the wide-angle wave-field (step-back from the first arrivals in the basalt flows to the sub-sedimentary basement arrivals); higher reflection amplitudes may be recorded near the critical distance; multiples created in sedimentary layers above the basalt will be absent at wide-angles; travel-time data from wide-angle arrivals allow an improved migration-velocity model to be constructed; there is the possibility of identifying arrivals in the pre-stack gathers for selective imaging.

Influence of basalt velocity structure on wide-angle wavefield

In order to assess what information can be extracted form the wide-angle wave-field produced by a layer of basalt flows and the underlying structure, we calculate synthetic reflectivity seismograms with basalt velocity structures based on well-log and geologic mapping data. Due to the low resolution of the low-frequency seismic energy that penetrates through the basalt and the convergence in time of seismic arrivals with increasing offset, even greatly simplified versions of the realistic velocity model produce similar wide-angle wave-fields. It ought therefore to be possible to derive a seismic velocity model of the basalt from wide-angle travel-times and amplitudes that is sufficiently accurate for the migration of sub-basalt events. Whereas travel-time modeling and inversion by ray-tracing of refracted and reflected arrivals to derive velocity models is a well-established technique, modeling the amplitudes of these arrivals is rarely used for that purpose.

In wide-angle data, the first arrival from the basalt flows is easy to identify and its amplitude can be modeled as a function of the elastic velocity and density structure of the basalt layer. The velocities and densities of the overlying sediments are usually well known, as is the average compressional-wave velocity of the basalt (from seismic travel-times). Limited shear-wave data from bore holes indicate that the V p /V s -ratio of basalts varies mostly in a narrow band between 1.8 and 2.0 and varies little throughout the layer. Densities can be estimated from established velocity-density relations. Intrinsic attenuation in basalts has been found to be low (Q > 150). Since most of the effective attenuation in basalt flows is kinematic, it is thus correctly modeled in a reflectivity synthetic seismogram.

The simplest, but not very reliable way to estimate the thickness of the basalt layer is to use the (gradient-dependent) penetration depth of the seismic ray at the termination of the first basalt arrival (the range at which the step-back to the basement arrival occurs). A more.direct way is the tracing of the base-basalt wide-angle reflection if it can be identified. An independent method is provided by modeling the amplitude-versus-offset behavior of the first arrival. For simple gradient-layer models the first-arrival amplitude reaches a maximum, the location of which depends on the thickness and gradient of the basalt layer (Figure 1). The velocity gradient also determines the width and height of the peak and (for a given intrinsic attenuation) the slope of the AVO curve. Modeling real data typically requires more than a single gradient layer (Figure 2).

Imaging of sub-basalt events by pre-stack depth migration

When a good velocity model is available, it becomes possible to migrate sub-basalt data into an image where otherwise faint events are boosted because of the higher amplitudes of the wide-angle data and the reduced contamination with multiples. It is necessary to select carefully the parts of the wide-angle wave-field that contribute to the stack. This selection will be guided by the first-order structural interpretation from the wide-angle (ray-tracing) velocity analysis. The synthetic data example (Figure 3a) demonstrates the information that can be recovered under ideal circumstances (perfect velocity model). In this case it is possible to distinguish unequivocally between primary and other events in the stack, even though this 1-D example does not allow a discrimination by a criterion like dip that is usually available in real data. The field example (Figure 3b) is a single CMP from a fairly one-dimensional area with a velocity structure similar to the one used in the synthetic. It shows that under more realistic conditions, the near-vertical migration alone contains less useful sub-basalt information than the noise-free and perfectly migrated synthetic; the contribution from selected wide-angle data (inset on third panel of Figure 3) is hence more important, and highlights arrivals not seen on conventional migrations.

Figure 1. Basalt first-arrival AVO curves from synthetic shot gathers (reflectivity method). A high-velocity layer of varying thickness and velocity gradient overlies a constant low-velocity layer.

Figure 2. Real data example of modeled basalt amplitude. The starting model was the result of ray-tracing (dashed velocity-depth curve where it deviates from the final model).

Figure 3( a) First two panels are details of a synthetic gather. Traveltime is reduced by 5000 m/ s. Third panel shows 1- D pre-stack depth migrated stack of the synthetic data overlaid by the velocity model; for the inset only selected wide-angle data were migrated. (b) First two panels are details of wide- angle gather acquired on the North- Atlantic volcanic margin. Inset in depth-migrated third panel contains selected wide- angle data only. SF sea floor, TB top basalt, BB base basalt, B basement, 1M first basalt multiple, 2M second basalt multiple.