Mehmet C. Tanis

Diving-wave Refraction Tomography And Reflection Tomography For Velocity Model Building

Summary Trinidad and Azerbaijan offshore areas are strongly affected by shallow gas anomalies which greatly attenuate seismic signals. Building velocity models in such areas with shallow water depths and gas can be a difficult task. Here we present two alternative ways to build reliable velocity models in the presence of shallow gas; one that is suitable to very shallow (<100m) and poor data quality areas and the other for deeper water depths. In the first instance, we make use of Diving-Wave refraction tomography method to build shallow velocity models offshore Trinidad and Azerbaijan. Previous use of this method has been limited to processing seismic data to produce a shallow velocity model to determine static corrections in time processing. Our success is in using the velocity model derived from Diving-Wave tomography as a starting model for reflection tomography in depth processing. We show that Diving-Wave method is a robust technique that produces reliable near surface models in the presence of gas and in areas with low signal to noise ratio. In the second case, we show that where data has reasonable offset to work with, reflection tomography can produce fairly accurate and high fidelity velocity models that can be further improved with iterative migration velocity analysis. As a result, depending on available data quality, either Diving-Wave derived shallow velocity model or reflection tomography derived model can be used to improve the ultimate product from iterative pre-stack depth migration and reflection tomography.

An integrated workflow for imaging below shallow gas: A Trinidad case study

Trapped shallow gas severely impacts signal to noise ratio and makes accurate overburden velocity model estimation difficult, since surface P-wave data may be heavily attenuated and further contaminated with both surface related and internal multiples. Additionally, when the seabed is very shallow (<100m), reflection tomography is ill conditioned for reliably estimating near surface velocities due to the lack of short offset traces and the impact of multiples on deeper reflection events. The problems associated with these issues require an integrated approach to imaging ranging from improving data quality to estimating a reliable shallow velocity model, and integration into a robust global velocity model that will adequately image below gas. Building upon the technical strengths of both companies, BP and WesternGeco (WG) undertook a technical collaboration project to evaluate the fundamental causes of data quality degradation in Greater Cassia, offshore Trinidad and to implement appropriate existing solutions to improve imaging below shallow gas. As a product of this collaborative effort, in this paper we present an integrated workflow on real data showing that a targeted demultiple method combined with multiple passes of refraction and reflection tomography is an effective way to deal with data quality and imaging problems below shallow gas.

Finite‐difference migration of 3‐D seismic data with a parallel algorithm

We designed and implemented a production-scale 3-D poststack reverse-time migration RTM) algorithm with ll finite-difference solution to the fu acoustic wave equation on the Cray T3D massively parallel processors. The algorithm is based on finite-difference operators that are accurate to second order in time and fourth order in space and uses a novel approach for the stability condition. We applied this algorithm to several data sets from salt/subsalt plays in the Gulf of Mexico and compared the results with the outputs obtained from a 3D prestack Kirchhoff migration and poststack split-step Fourier depth migration. We show that our implementation of RTM produces good quality images comparable to 3-D prestack depth migration around salt structures but runs in only a few hours on the T3D. We also make a comparison between RTM and split-step Fourier migration on a T3E.

Applications of accurate 3-D poststack depth migration

Oil companies are increasingly exploring in areas of complex geology and strong velocity contrasts, where stacking and time migration can fail to produce interpretable images of hydrocarbon‐bearing structures. Wave‐equation‐based depth migration is routinely being applied to 3-D data both before and after stacking to improve upon time‐migrated data. Most depth‐migration algorithms rely on Kirchhoff and phase‐shift approximations to the acoustic wave equation. Kirchhoff approximation relies on ray‐tracing, which fails as rays propagate past the critical angle, and phase shift assumes a constant velocity within depth layers.