Abstract

A common problem faced by explorationists in interpreting normal faults from seismic data is identifying the proper location of the fault plane and the impact of fault shadows on potential misinterpretations. In the time domain, normal faults can produce fault shadows. This is apparent even in environments with simple velocity fields. Fault shadows are caused by rapid lateral velocity changes across the fault plane. These lateral velocity variations are not always properly accommodated by time processing. The result, commonly referred to as a fault shadow effect, can include a range of artificial time structures and complexities below the fault that can be variously interpreted as additional faults, gas chimneys, or even potential prospects. Both modeled and real data are used to illustrate the potential impact of the fault shadow in time domain processing and the potential of depth domain processing to resolve these issues.

A model has been built using a simple velocity field with a high-angle normal fault. A variety of different types of ray tracing experiments were conducted, including zero offset and full offset. Full-offset ray tracing was conducted to generate synthetic CDP sorted traces. Three different processes were applied to these synthetic traces, i.e., post-stack time migration (Post-STM), pre-stack time migration (Pre-STM), and pre-stack depth migration (Pre-SDM). The comparison of these different processes show that the fault shadow still has a significant expression in the time domain processing, both Post-STM and Pre-STM, while the structure and imaging below the fault are properly resolved by Pre-SDM.

A sample 2D seismic line from Irian Jaya is used to demonstrate how Pre-SDM handles the fault shadow problem. Irian Jaya has a variety of complex geological conditions with large scale faulting that are best resolved by Pre-SDM. Normal faults produce fault shadows in the original Post-STM and Pre-STM. In this dataset, we also deal with a significant velocity inversion below the Miocene that is associated with a change from the thick Yawee carbonate back to siliciclastic sediments. This velocity inversion cannot be properly resolved by time-domain processing. However, these problems have been corrected with an accurate velocity model in depth. The velocity model was built using the layer-by-layer depth interval velocity modeling technique and iterative Pre-SDM. Applying the more accurate velocities, the fault shadow problem is resolved and the faults and sediment reflectors are both well imaged and correctly placed.