Avoid drilling into salt

From time to time we see cases where a well targeting an up-dip sand layer in close proximity to a salt body is actually tagging salt instead of the target sand layer or alternatively drilled into the sand layer but much lower into the structure than expected.

In many of these cases the supporting seismic image is of good quality, so what can go wrong?

When constructing a velocity model for a salt-vicinity exploration program, we normally give a lot of attention to the salt body itself. We run several iterations to accurately define the salt/sediment boundary. Defining the salt body accurately is very important but sometimes we forget that in many cases the spatial location of the salt image depends not on the salt velocity alone but on the velocity and anisotropic model of the surrounding sediments. In the case shown in the images here, the exploration targets are both the sand layers around the dome as well as the subsalt structure.

The sedimentary section around the dome is anisotropic. Therefore, in order to correctly position the salt flanks on a migrated PSDM volume, a correct anisotropic model needs to be constructed. The anisotropic model in this case is a TTI model, consisting of delta, epsilon and dip fields. If one of these fields (which are input to PSDM) is not accurate enough, the salt image will be mis-positioned on the resulting PSDM volume. 

Which of these parameters is more crucial to correctly image the salt flanks and surrounding sediments at the correct location?

The two images shown here can help answer this question. To generate the images above, a TTI anisotropic model was constructed and simulated using an anisotropic TTI Forward Time Modeling (FTM) application. The simulated data was then migrated using Reverse Time Migration (RTM) application, once with 50% error in the dip field and then with 50% error in the delta and epsilon fields. The yellow lines in the images above show the true position of the salt body and surrounding sedimentary section layers. As we can see, an error in delta and epsilon fields will result in a more degraded image in the subsalt section as well as greater mis-positioning of the salt body on the migrated PSDM volume. Close examination of these results show the discrepancy in imaging. In the case of a significant error in the delta and epsilon fields used in a TTI model, the resulting PSDM will be under migrated. The salt flanks are not migrated deep enough, resulting with a salt body which is larger than the true salt body. This means that if we plan a well to target a salt/sediment termination based on this image, the well will be drilled too low on the structure, and away from its target. The steeper the dip of the sedimentary layers the greater the error in positioning. 

The conclusion from this example is very simple. When constructing anisotropic models for imaging of salt flanks and the salt vicinity, we need to ensure that the anisotropic model of the sedimentary section is as accurate as possible. This will ensure that the image of both salt flanks and the sedimentary layers truncating against the salt are not only imaged correctly, but just as important, are correctly positioned in space and depth.

Tools and workflows for construction of anisotropic models is one of the subjects that will be discussed in detail in the upcoming 4-day depth imaging course that will take place the first week of October in London: The Practice and Theory of Depth Imaging

In the course we will review the theoretical definition of various anisotropic models, will learn about velocity and anisotropy estimation tools used in the industry and will explain model building and depth imaging workflows designed for the construction of accurate anisotropic models.