Chevron records show the large, McKittrick-area oil leak that has shone an unflattering light on Kern County petroleum production probably originated with an idle well being worked on at the same time the company was injecting high-pressure steam just 360 feet away, a combination that industry people say should not have been performed simultaneously in such close proximity and which possibly contributed to the release.
The San Ramon-based oil producer told state regulators in a recent written analysis that a well it was using to put steam into the Cymric Oil Field was not switched from injections to production mode until 7½ hours after the company noticed oil seeping to the surface at 5:30 a.m. on May 10.
Observers within the industry said that timeline suggests steam injection activity was happening at the same time Chevron had opened up and was "re-abandoning," or resealing, a well idled in 2004. The routine re-abandonment activity in this case began May 7, according to Chevron's records, which do not indicate the start of what is typically a dayslong steam injection process.
The problem with steaming a well near concurrent work on another well, people familiar with local oil fields say, is that there's a chance steam will make its way through uncharted channels underground before coming to the surface in an area not outfitted to receive oil.
Several people interviewed said Chevron should have "shut in" — meaning turned off — the steam injection well that state maps show lies 360 feet from the surface of a well the company blames for several thousand barrels of oil ending up in a dry creek bed during a series of uncontrolled releases near McKittrick.
"I definitely would say they need a 600-foot shut-in radius if they are doing a re-abandonment," said Bakersfield geologist Burton R. "Burt" Ellison, former district deputy at the California Division of Oil, Gas and Geothermal Resources, the state's primary oil regulatory agency.
Others, noting the complexity of subsurface conduits in western Kern oil fields, said it's hard to say what a safe distance would have been in this case, and that additional nearby wells may have played a role in the leak. But they still questioned the wisdom of steaming so close to a well undergoing work.
"It’s probably too close," Bakersfield independent oil producer Ken Hunter said. "It's not definite one way or another ... given the fact that steam goes into a crevice, and instead of it going on ... uniformly, it goes through a crack."
A local engineer familiar with the "cyclic steaming" technique used in this case, and the area where it was being performed, agreed to review Chevron's accident report and speak about it on the condition of anonymity because the person works for a company whose clients include Chevron. The engineer concluded Chevron's report on its steaming activity near and during the re-abandonment job represents "a smoking gun" suggesting the company did not follow its own best practices before the accident.
For its part, Chevron denies steam played any role in the uncontrolled releases that have continued intermittently at the site since mid-May. Preliminarily, it has pointed to a pressure test, performed during a root-cause analysis ordered by regulators, as the most likely cause of the mess. Regulators did not specifically order Chevron to pressure-test the well as part of the RCA.
The company said the nearest well to have undergone steam work during the re-abandonment work was actually 597 feet from the broken well's underground perforations.
"The distance between surface locations on a map is 360 (feet) but this is not relevant for determination of a steam block," the company said by email Friday. "Steam block" refers to a procedure for halting steam injections to part of an oil field.
Chevron said it has performed extensive diagnostic testing of wells near the leak, including temperature surveys, nitrogen testing for well-bore integrity and steam testing. It said steam was injected in four wells near the leak for about 24 hours and that this had no apparent impact on the seep, "which indicated that these wells and steam were not the root cause."
"We then pressure-tested the abandoned well with water, which led to the reactivation of the seep. These results indicated that the root cause was the abandoned well," the company stated.
Ellison, the former DOGGR official, said Chevron appears to have worked in good faith to determine the cause of the problem. But he was skeptical of its claims that steam played no role in the accident, saying he would have called for a steam-block radius of as much as 1,000 feet around the re-abandoned job.
Re-abandoning an oil well involves drilling out what is essentially a concrete plug before doing work that ends in a new concrete seal.
Ellison's theory is that another well not mentioned in Chevron's analysis, which he said runs laterally through the same area, may have been the real culprit allowing steam to carry oil to the site of the surface releases, hundreds of feet from both the well that was being steamed and the one being worked on.
"It's a convenient explanation but not really plausible in my mind," he said of Chevron's report.
DOGGR did not respond to a request for comment on Chevron's report, which it has judged an inadequate explanation of the cause of the releases.
The uncontrolled release of oil and water at the site is a form of "surface expression" that, though forbidden under DOGGR regulations, has happened periodically in the history of Kern County oil production.
In this case, no one was reported injured and no surface water or drinkable groundwater was placed at risk. The area is in an unpopulated part of western Kern with no development other than oil facilities.
What makes this surface expression especially significant is its size and visibility. More than 13,000 barrels of black fluid have come up from the ground, about a third of which is oil, meaning more than 190,000 gallons of crude came up from the ground onto the surface.
The accident, along with striking photos of black fluid in a dry creek bed, has attracted notice as far away as San Francisco. It has come at a time when environmental activists up and down the state, many of them more concerned about climate change than they are knowledgeable about technical oil processes, are calling for an end to California oil production.
"Obviously, we're in a very visual world, and so … (the releases) do not put the business in a good light, so to speak," Hunter said.
But he added the situation isn't as bad as it appears.
"These are the types of things that can happen in an industrial site. And this is an industrial site,” he added. "There's going to be no long-term damage to the environment."
DOGGR has issued two notices of violation against Chevron relating to the releases. The agency has also ordered Chevron to stop steam injection work within 1,000 feet of the abandoned well believed to be at fault. About 20 injection wells in the area have been shut in, while nine idle production wells have been reactivated as a way of reducing underground pressure in the area.
The agency also told Chevron in May to prepare a monitoring and prevention plan for review by the division no later than Nov. 20. The plan was to include a surveillance system or a pressure and flow monitoring system "that will give adequate warning to prevent surface expressions."
DOGGR also ordered the company to map the leak, "including cracks, fissures and sinkholes related to underground injection work," as well as plans for restricting access to the area and training for people working in the area.
This story has been corrected to reflect that regulators did not specifically request the pressure testing Chevron blames for the oil releases.