If we consider the maximum number of possible interpretations to a seismic image, given the conceptual uncertainty of the data, there is a question of how best to narrow down the interpretational possibilities. One method would be to assign each possible concept with a probability, but given the possible permutations and inherent bias already in the data from acquisition and processing, this method raises further questions of how best to assign probabilities. It’s also likely to generate the same, or most likely, solution and sometimes we might want to think outside the box.
Experiments investigating seismic interpretation have shown that particular techniques have a strong influence on an interpreter’s ability to get the correct answer in model data. Notable amongst the techniques is evidence of genetic or evolutionary thought processes. Evidence of these evolutionary thought processes include drawing cartoons of, writing about, or annotating elements of the sequential evolution of the imaged geology. In fact it doesn’t matter how the interpreter evidences the thought processes, only that they have checked their interpretation is genetically or evolutionarily viable. The use of evolutionary sketches or text forces an interpreter to reconstruct, even in a simple way, the geometric evolution of the geology over time. By showing how the geometries evolve, an interpreter demonstrates that the final geometry in the interpreted model is valid and that the interpretation is plausible. For structural geologists this process is more formally known as cross-section balancing or restoration and is used to validate the geometrical arrangement of picked faults and horizons. But the method works for other disciplines, too.
In an experiment investigating how experts (in this case defined as structural geologists) interpreted a structurally complex synthetic seismic image, we found that only 10 percent showed evidence of having thought about the structural evolution of the imaged geology. Those that did were three times more likely to get the correct interpretation (Bond et al 2012). The few participants that used the technique, but failed to get the correct answer, proved in their own evolutionary sketches that their final interpretations were incorrect, as they were unable to generate the final geometry.
The message is really to make sure that the picks you have made make sense. It is all very well to focus on the data when interpreting, but the interpretation has to be geologically plausible. It would seem that simple validation techniques are infrequently used and they may well be the key to understanding which of the possible interpretations of a dataset are plausible. So next time you are stuck on an interpretation, grab the coloured pencils.
Bond, C E, R J Lunn, Z K Shipton, and A D Lunn (2012). What makes an expert effective at interpreting seismic images? Geology 40, 1, 75–78.