How can we define 'true' hydrocarbon anomalies? (2 of 3)

In my previous article ‘When is a seismic anomaly not anomalous?’ I talked about how, without sufficient information or data, it can be difficult to define which seismic anomalies are anomalous in the context of exploring for hydrocarbons. I suggested that defining what is anomalous requires a benchmark, or the norm to understand the deviations from it. So how can we do that?

One way is to think of the benchmark or norm in seismic amplitudes as the background shale trend and brine responses in our seismic data. The only 'anomalies' we are really concerned with are ones generated from oil or gas responses, so if we can understand the shale and brine responses, we will have a better chance of identifying true hydrocarbon anomalies as deviations from these trends. 

A method I have found useful to do this is one introduced to me by DownUnder Geosolutions that uses probabilistic depth trends1. Although the referenced article focusses more on uncertainty through a QI workflow, it lends itself nicely to understanding what is regarded as anomalous. Depth trends are generated from well data of elastic properties for all potential lithologies or facies in a basin, and then stochastically modelled to create probability density functions (PDFs) and understand the background shale and brine (normal) seismic responses, and hydrocarbon (anomalous) seismic responses for a given depth – an example of a possible outcome for gas is shown above. These modelled responses can then be scaled to the seismic (very important calibration step) and compared to our observed seismic at prospect locations to understand whether amplitudes are indeed anomalous or not. This comparison can be made in both the relative, reflectivity domain (as in the image above) or in the absolute domain if an inversion has been conducted.

I quite like this method as it starts and ends with the geology. It is the geology from drilled wells that drive the method and even though we look for geophysically derived seismic anomalies, it is the geology that causes them. If we don't have this prior data then it can be difficult to understand what seismic amplitudes are anomalous. 

When I was thinking about this, it reminded me of a comment I heard from an experienced geophysicist (apologies if you recognise this – I do not remember your name!) a few years ago at a Rose & Associates DHI Consortium conference2:

“In wildcat exploration, the AVO method is only useful when predicting lithology. Use of amplitudes for fluid discrimination only become useful when bringing in spatial constraints such as flat spots and clear amplitude conformance to structure”.

I think this is a great comment that highlights the difficulty in defining what our true hydrocarbon anomalies are without the data to create benchmarks and calibrate too. I do believe however that with careful QC of good seismic and well data, and calibration to the wells, it can be possible to predict fluid type. Despite this, without the mentioned spatial constraints greater uncertainty will still exist.

Now we’ve defined our seismic anomalies, how can we evaluate them in the context of specific prospects to help build confidence in our models? In my final article* on seismic amplitudes I’ll describe a method I’ve found useful in analysing and, as importantly, communicating the robustness of amplitudes.

As always, please feel free to like, share, or comment on any of my material, and I will try to respond to any feedback or questions when I can. 

*Article 3 is now published at the following link: Making sense of seismic amplitude variation

1. Lamont, M. G., Thompson, T. A. and Bevilacqua, C. (2008). Drilling success as a result of a probabilistic lithology and fluid prediction – a case study in the Carnarvon Basin, WA. APPEA Journal, 48, 31-42.
2. https://www.roseassoc.com/dhi-consortium/