Elementary, my dear Watson

By Peter Duncan, PhD
Founder & Co-Chairman, MicroSeismic, Inc.

I think it was sometime in the summer of 2004 that I was visiting a friend of mine, Larry Oldham, in Midland when I first appreciated the power of microseismic monitoring. At that time Larry was CEO of Parallel Petroleum, a publicly traded E&P company he had founded 20 years earlier. Larry and I had worked together over the years, including drilling many wells together. Over coffee I asked Larry if he had ever microseismically monitored a frac. Larry got kind of excited, went over to his desk where, after shuffling some papers around, he found a CD. It turned out he had just received the data on the CD, the result of monitoring a frac of an Abo well in New Mexico. He popped the CD in his pc and started the playback. What I saw literally blew me away. Here on the screen I was watching a time lapse representation of the progression of the frac away from the well in the form of dots representing the hypocenters of the microseismic events created as the rocks fractured. I was watching a movie of the cracking of the reservoir rocks more than a mile below the earth’s surface. I was in geophysical heaven. I asked Larry what he had done with the movie. “Well, I showed it to my Board and they were fascinated”, was his reply. So I asked him if he were going to monitor any more wells and he said, “Why should I? I already have the movie.”

And therein, in a nutshell, is the biggest hurdle we have faced in growing the frac monitoring market. To a geophysicist the elegance of “seeing” a representation of the frac growing in time is enough to justify the effort and expense of doing the monitoring. But the CEO of an oil producer needs to get more than a science buzz out of his spend, he needs to see either more hydrocarbons or less drain on the budget. As another engineer put it when I told him we had reduced hypocenter uncertainty by 50%, “So what?” Time and time again as we have made advances in our ability to image the stimulation more accurately or more completely, the geophysicists have proclaimed amazement but the engineers have been unimpressed.”So what?”

Over the past few years we have taken the analysis of microseismic data through a number of developments. We started with the location of the events which we found we could image more accurately and more efficiently most of the time with surface or near surface arrays. Next we realized that surface arrays allowed us to unravel the focal or failure mechanism of the rock fractures. This led to better estimates of subsurface stress magnitudes and directions which in turn allowed us to represent the stimulation as an array of scaled and oriented fracture planes rather than simple dots. Next we found we could upscale the network of fractures into an enhanced permeability model. These models are unscaled unless we can calibrate through a history match with production, but nevertheless represent the relative performance of the frac job to a high degree of accuracy. However, while the answers have been ever more elegant and complete they have still elicited a “So what?”

Finally, as a last step, we have found that we can drop the enhanced permeability model into a production simulator and predict the production of a given stage, well or pad. We have demonstrated the truth in these predictions by successfully matching our predictions to the producing history of wells we have monitored. At last we have reached a result that is not met with a “So what?” Most of the time an operator must wait at least 6 months before he knows if he has drilled a P10 or a P90 well. We now believe we can give a reliable adjudication of that metric within a few days of finishing the completion. Our observations are no longer reported in terms of “the number of events”, or the half length of the frac, or the SRV in acre-feet. Rather, we conclude that this well or stage will produce 50% more hydrocarbons than that other well or stage, and with proper calibration we can even state that result in BOE.

Lately I have been watching a TV series starring Benedict Cumberbatch, entitled “Sherlock”. It is one of several instances of the current resurgence of interest in the Sherlock Holmes franchise. What is it that makes this quirky detective so enduringly popular? Certainly he seems to see everything. His observation skills are legion. But then his assistant Watson is no slouch at observation either yet he never seems to “connect the dots”. It remains to Holmes to relate these observations to outside events, to turn data into actionable knowledge that solves the case. “Elementary, my dear Watson” is what he says, but it is anything but elementary. Microseismic monitoring has been all about elegant observations. “Seeing” the rocks break is fascinating but it is not enough to justify another dataset, watching another episode if you will. Predicting production, now that is actionable knowledge and while not elementary it may prove fundamental to making microseismic monitoring truly valuable.