EAGE/SEG Research Workshop - Frequency Attenuation and Resolution of Seismic Data 2009

We present a new method for broadband marine acquisition and processing. A 3D shallow towed-streamer spread is deployed, designed to optimize the mid- and high-frequency parts of the bandwidth. In addition, data are simultaneously acquired from a small number of deeper towed streamers. The depth of these deeper streamers is optimized for the low frequencies such that the combined overall bandwidth is enhanced. Because the deep streamers only provide the low-frequency part of the bandwidth, we can more sparsely sample these data enabling efficient acquisition scenarios as fewer streamers are required. A 3D case study using this new acquisition method was acquired off the NW Shelf of Australia. The streamer spread consisted of six shallow streamers towed at a depth of 6 m and two deeper streamers (below shallow streamers 2 and 5) towed at a depth of 20 m. The resulting data exhibit both high resolution and deep penetration for subsalt and sub-basalt imaging, for example. In addition, inversion for acoustic impedance, imaging, and velocity model building, also benefit from the broadband result. Data acquired in this way are more robust to poor weather conditions than conventionally acquired data.

From the digital communication theory, we present some considerations to quantify signal characteristics and the relation with the theoretical resolution. However, this theory was basically developed on narrow bandwidth signals and transposed to broadband seismic signal. This has some implication and also gives the carrier frequency to be considered as a characteristic parameter. ‘Detection’ is also included in the discussion as the ability to identify and hopefully characterize thinner geological bodies. One can also consider tuning effect to identify the key parameters. We show that these two approaches give the equivalent conclusion about the main role played by the bandwidth and the carrier frequency to be considered together. If a larger bandwidth could be the processing challenge to drive the seismic imaging to higher resolution and/or better detection, it is necessary to spend time on controlling what kind of bandwidth broadening is sought after and how it is obtained. It is useless to reach broader spectrum while amplifying noise. Implicit techniques must involve checking the level of noise to ensure reliability. But when enlarging the bandwidth cannot be the solution, other way can be used to make signal in proper condition to deliver relevant layer properties.

In this paper, I introduce the intrinsic attenuation for periodically layered structure to investigate how the intrinsic attenuation interferes with the scattering due to layering. The presence of intrinsic attenuation results in the stretch of reflection and transmission responses in frequency domain. I show that the anomalies in energy loss follow the stop-bands in frequency scale which are wider for higher acoustic contrast between the layers.

The experiments have been conducted to investigate the effect of strain amplitude on the compressional and shear wave velocity, Vp, Vs, attenuation, Qp-1, Qs-1 and dynamic Poisson’s ratio, Vd in quartz samples of three types: the intact quartz, fractured quartz and smoky quartz and sandstone. The measurements were performed using the reflection method on a pulse frequency of 1 MHz with changing strain in the range 0.3 < e < 2.0 m-strain under a confining pressure of 10 MPa and at ambient temperature. The anomalous inelastic behavior occurs owing to the change in value of the strain-amplitude. The evident strain-amplitude effect takes place only on Q-1, Vd - parameters. P- and S- wave velocities are little sensitive (~1%) to changing strain amplitude. The greatest decrease in Q-1 (to 16%) takes place in the smoky quartz whereas in other quartz, effect is practically absent. Amplitude variation leads to the modification of relaxation spectrum. The broadening of peaks width with increasing strain amplitude that takes place for P- wave attenuation in the smoky quartz. Such behavior of attenuation can be explained by a joint action of viscoelastic and microplastic mechanisms in the material with defects.

Measurement of seismic attenuation is affected by spectral interference, which limits the accuracy of standard spectral ratio regressions. Furthermore, heterogeneities and anisotropic effects above the zone of interest complicate the choice of a reference event. We introduce a prestack Q-inversion (PSQI) algorithm which operates in the tau-p domain such that the reference event changes with the slowness of the trace considered. By inverting the spectral ratio data for all of the prestack traces simultaneously, the influence of the interference peaks and notches is reduced. The result of the PSQI approach is a more accurate and precise attenuation measurement. We show the justification for our approach using synthetic data, and then demonstrate the method on a shallow reservoir 3D seismic survey.

This study deals with the extension of the frequency-domain full-waveform inversion/modelling (FWI) from the acoustic to the viscoacoustic case in application to the wide-aperture seismic data recorded in the Polish Basin. Attenuation was accounted for by introduction of complex velocities. Two inversion strategies were tested: (i) coupled inversion in which Qp and Vp were simultanously inverted and (ii) decoupled inversion in which Vp model was taken from the acoustic inversion and remain fixed during the inversion for Qp. The latter approach produced preferred attenuation model. Together with some a priori information from well-log data and laboratory measurements we interpret recovered high attenuation anomalies in terms of fluid saturation and rock permeabilities.

Knowledge of key fracture properties at reservoir level such as their dominant orientation is important for successful hydrocarbon exploration since it can help enhance oil recovery. A common method to determine fracture orientations is by means of the analysis of amplitude variations with offset and azimuth (AVOA). Unfortunately inferred results are not unique and a 90 ambiguity in fracture orientations can exist – even if solely a single major trend is present. Like amplitudes, attenuation also exhibits azimuthal variations in the presence of fractures and/or cracks. The analysis of attenuation variations with offset and azimuth (QVOA) thus holds complementary information to AVOA inversion results. This paper discusses a methodology used to extract variations of attenuation on a 3D multi-azimuth sectored dataset. This methodology is also applicable to both multi-azimuth walkaway VSPs.

Conventional imaging processes for a situation with a gas cloud do not offer satisfactory solutions. Due to the complex wave propagation through the anomaly and the transmission imprint on the reflections from below this area, the image below the anomaly is usually not properly recovered. The main idea is that the reflection response of the complex area - e.g. the gas cloud - including its coda, carries detailed information on the gas cloud properties that can be obtained and translated into a transmission correction operator. With this multi-dimensional deconvolution operator the transmission effect from the overburden can be removed from the reflections below these complexities, after which true amplitude images of the reflections below this overburden can be obtained. For the 1.5D situation the inversion procedure is shown to provide satisfactory results. Furthermore, for the 2D case, the feasibility of a full waveform deconvolution process will be illustrated with synthetic data from a complex gas cloud model.

The quantitative use of the seismic amplitude information during the interpretation is a key point for many prospect evaluations and almost all reservoir characterization studies. The local amplitude information of interest is always affected by a series of signal attenuators along the propagation of the incident and reflected wavefield which are highly dependent upon the geological context, structural shape, lithologies and fluids. These many causes of attenuation are often pragmatically treated by a combination of a few well-known tools: spherical divergence compensation, surface-consistent or volumic time & frequency-dependent compensations. Cases where severe amplitude attenuation effects cannot be treated using usual approaches are a serious issue, in particular for seismic characterization of reservoirs. In this paper, a methodology adapted to the study of the impact and relevance of specific attenuation processes is presented; Alternative mechanisms and tools for a quantitative assessment of these processes are proposed. This methodology is illustrated through a field case study, using vertical incidence VSP and 3D surface data. The seismic amplitude attenuation analysis uses the VSP data to validate the modeling of measurable effects and to propose possible causes for the remaining strong attenuation, such as diffuse gas.