How good are geologists?
Feature Articles, Feb  19  2010 (Digital Energy Journal)

- In a test of almost 500 geological interpreters, conducted by the University of Glasgow, only 21 per cent could correctly identify a geological structure of an inversion from the seismic image, said Dr Alan Gibbs, director of Glasgow based Midland Valley Exploration, speaking at the Jan 20-21 Exploration, Technology and Business conference in London.

Dr Alan Gibbs, director of Glasgow based Midland Valley Exploration

With the help of post-doctorate researcher Dr Clare Bond, Midland Valley created a geological structure on a computer, then asked GX Technology, a seismic company to show what the seismic data would look like if rock with that geology was surveyed (synthetic seismic). It then put the seismic in front of 500 professional geoscientists and asked them to interpret it.

Only 21 per cent got the correct answer – inversion. Other geoscientists gave a wide range of different responses, including a thrust (26 per cent), salt / mud diapirism (8 per cent), extension (6 per cent). This could indicate that in the real world, geologists are correct only 21 per cent of the time.

35 per cent of respondents considering themselves “experts” got the right answer – so many experts got the answer wrong, and many non experts got it right.

Midland Valley then analysed the methods used – and discovered that the more different techniques that had been used to try to answer the question, the more likely they were to get the right answer.

In fact, every respondent who used 5 or more techniques got the correct answer, such as drawing horizons, drawing sticks, looking for features, putting in annotation, descriptive writing or trying to work out how the rock ended up as it was (evolutionary thought).

And of these methods, the most successful one was evolutionary thought (94 per cent of people using evolutionary thought got the correct answer). This compares to 45 per cent who drew horizons, and 10 per cent who drew sticks. The numbers were similar for both people who considered themselves “experts” and “non experts” – suggesting that people’s effectiveness depends on their technique, not whether or not they consider themselves to be an ‘expert’.

“You don’t need software – you just need to think about it,” he said. “Everybody can be more effective.”

Readers who are concerned that geologists might only be accurate 21 per cent of the time can take comfort from similar studies of doctors, showing that doctors are about the same amount of the time, he said.

And when people are correct, it is surely often more because of luck than expertise.

Midland Valley operates as something of a geological troubleshooter, being called into companies who have discovered that the rock they are drilling through is very different to what they expected from their models.

Evolutionary thinking

Evolutionary thinking is about trying to work out how the rock below the surface became the way it is now. Pieces of rock which can be hundreds of millions of years old have been moved, by tectonic forces, into their present day shape and position.

If it not always clear how the subsurface reached the shape it is today in a model. This lack of understanding could indicate a serious problem in your model – and if you can understand how all the pieces got to their present condition, a large amount of uncertainty has been removed.

Working out how the rock evolved to (what you think is) its current structure can be done in your head, or by cutting up paper, or (more commonly today) using 3D computer models.

Work process

In many ways modern work processes do not support the types of geological working which are most likely to bring the right results, he said. Geologists are also not encouraged to carefully work out different scenarios, just to “stick it on your PowerPoint and drill it,” he said.

“Every model is incorrect – the question is just at what level they are incorrect”? he said.

A better working environment would involve more peer review and peer challenge – with more people looking at the same data, coming from different approaches, and discussing their ideas.

Many oil companies have implemented robust processes to make sure every model is fully assessed before decisions are made with it. But they might be better off putting their efforts into nurturing a work environment where people are creative and take an interest in each others’ work, so that each model is looked at in a number of different ways.

This is how things used to be like, he said. “When I first joined the oil industry, the industry had large teams of people – and most companies had big internal training programs.

“But over the last few years it has tipped downwards. We have been forced in most companies to cut down the number of people who are around. So the amount of time for mentoring people, peer review and peer debate has reduced.”

“We also have people working on computers almost entirely. Computers are very powerful but the ‘Nintendo generation’ tend to believe what they see on computer.”

Universities are also rewarded by the research they do on what is new and exciting – not by how well they teach people fundamental geological skills, he said.

Even if someone has good academic geological qualifications, they might not necessary have good core geological skills, if they have been mainly working on specific projects for their research, he said.

And of course motivation is important. “Motivating people so they come to work enthusiastic and work as a team is actually difficult,” he said. [A good working environment is when] people chat to each other on a team basis. People have the flexibility to be creative. When someone’s up against a time deadline the rest of the team help that person – a whole pile of stuff – it’s very very difficult and I’m sure the bigger your organisation the more difficult it gets,” he said.