Too Complex Not to Fail: UT Professors Analyze the Gulf Oil Spill



  • Organizational behavior and complex structures have not kept pace with complex technology
  • You can't study complex systems by simply breaking them apart and putting them back together
  • Systems in nature have self-regulating features that cannot be predicted
  • Organizational behavior and complex structures have not kept pace with complex technology
  • You can't study complex systems by simply breaking them apart and putting them back together
  • Systems in nature have self-regulating features that cannot be predicted

“Horrible things happen when complex technologies and procedures overtake humans, who service the technologies falsely assuming complete control.”

Thus begins a briefing paper [3] on the Gulf oil spill prepared by Dr. Tad W. Patzek for the Energy and Environment Subcommittee of the U.S. House of Representatives.

His analysis is also highlighted in this video presentation Oil in Troubled Waters: Dr. Tadeusz W Patzek - Where Do We Go From Here? [4]

“I argue that organizational structures and human behavior have not kept pace with the complex technologies we–the engineers and scientists–have created,” he writes. “Given the structural changes in the industry, academia, and government, this tragedy has been at least 20 years in the making.”

Too Complex for Humans to Understand

Patzek concludes that given the current work on the BP well blowout, there is nothing in the science and engineering of the tragedy that baffles experts. Rather, the bafflement comes from the complex interaction between the drilling procedures and the vast natural forces involved in the deepwater drilling environment.

“It seems that the human inability to grasp and execute the complex steps of a deepwater drilling procedure led to the tragic outcome,” he reports. “This general failure of organizational structures should also be understood in the context of complexity.”

He points to the complexity involved in splitting a large system into smaller parts, studying each part in separation, and then attempting to put the parts together to build a whole. Unlike a clock, which can be taken apart and reassembled, a system in nature contains complex autonomous features–emergent behavior–that cannot be predicted by studying its parts in isolation.

“It is not true … that a frog can be dissected into parts and when these parts are reassembled the same living frog jumps off the table.

“Complexity is an essential feature of deepwater petroleum and natural gas production systems,” Patzek writes. “Our ability to predict the future behavior of complex living and inanimate systems is never perfect or complete.”

At the Mercy of the Butterfly Effect

Patzek’s conclusions are confirmed by Dr. Edward Anderson, a professor and specialist in systems dynamics modeling at the McCombs School. “In a complex system a small disturbance or misstep can have huge, game-changing consequences,” he says. Anderson refers to the classic example stated by Prof. Edward Lorenz, a climatologist at the Massachusetts Institute of Technology, who postulated that “the flap of a butterfly’s wings in Brazil [can] set off a tornado in Texas.” This is the source of the term “butterfly effect,” which has since seeped into popular culture.

“A nasty corollary of the butterfly effect, shown to be true over and over again, is that the behavior of complex systems are difficult for the human mind to predict,” he says.

Pinpointing Failure in Hindsight

Complexity is also a problem analysis issue. Dr. John C. Butler, academic director of the Energy Management & Innovation Center, warns against hindsight bias, which may lead managers, regulators and others to over simplify the causes of a complex event such as the BP well failure.

“In experimental studies, humans have been shown to believe it is easy to predict an event after that event has occurred,” he says. “We sometimes call this postdiction, which basically means that we tend to reflect on the past and say ‘we should have known.’”

Butler is quick to say this doesn’t mean BP didn’t make mistakes. “Perhaps they did, but first we need to let all of the facts come out before we make judgments. Even if we have all the facts, it’s still hard to interpret that information as if we were in the past.”

Adopting a NASA Approach to Oil Exploration

Patzek suggests more research is needed on understanding the behavior of complex systems in reservoir exploration and production, “an area which has been neglected because of insufficient resources.”

He points to the fact that massive federal funding was used by NASA to arrive at the procedures for handling unexpected extreme events, while the human and ecological impacts of a serious accident in deepwater oil production far exceeds those of even the most serious disaster in space.

Anderson concurs with the NASA analogy. “Currently, parts of a large project such as the BP project are parceled out to suppliers and sub-suppliers much like pieces of Humpty-Dumpty were parceled out to all the King’s men after he fell from the wall,” he explains.

“Someone needs to guide all of these parts back together into a coherent, safe, and effective whole. That is the purpose of systems integrators. Fortunately, a model for such systems integration already exists at NASA and other aerospace agencies. Some firms such as HP have actually adopted such a model and use it to reduce the number of disruptive butterfly effects. Oil industry firms can too.”