For the past several months Bob Arrington has deconstructed the official timeline story around the Parachute Creek spill complete with all its odd little foibles and discrepancies. Then he reconstructed it with the advantage of time and some known knowns. What he offers here, in his expert opinion, is a more plausible chain of events.
Guest post by Bob Arrington*
In a general time line format I will relate an exchange I have had with the Colorado Department of Public Health and Environment (CDPHE) that also involved the Colorado Oil and Gas Conservation Commission (COGCC).
This exchange relates to the Parachute Creek Spill by Bargath, an owned subsidiary of Williams Companies, Inc. On the initial spill, the COGCC was the lead state agency investigating and responding to the spill. About a week after I sent a letter, by fax on April 21, 2013 and surface mail the next day, to the Governor, about the COGCC not being the proper agency, the CDPHE was assigned to replace the COGCC. On the 21st, I also sent email copies to the agency directors Urbina (CDPHE) and Lepore (COGCC).
Earlier, on April 4, 2013, at an Energy Advisory Board meeting, there had been presentations by Dave Keylor (Williams) and Matt Lepore (COGCC) about not being able to find the leak source. At that time, I showed a possible leak configuration based on possible cold weather conditions. Keylor responded he didn’t feel that would have been the source, because they had performed several pressure tests on the line. He went on to relate they had only experienced a leak before when a pressure gauge burst, and a crew went out and plugged it, and estimated less than 25 gallons spilled.
Hearing of the burst gauge, I exclaimed, “Ball game over!” and went on to say that could be the source of the entire leak. As we started to discuss the issue, Brent Buss chair of the meeting, interrupted to move on to another member’s question.
I went home and the next day sent Donna Gray, spokeswoman for Williams, a copy of calculations showing how great the leak could be.
Timeline — January 3, 2014 — Description of Events
At the time of discovery, contrary to the above statement, control room workers noticed a remote valve closed unexpectedly, indicated by a pressure spike on monitors in the plant and a short time later a maintenance crew was dispatched to ascertain why as stated below.
What drew attention to the pressure spike? An alarm, or was a worker watching gauges or a recording/real time device? Since the valve is a “fail-safe” designed to close upon system malfunction, usually an “alarm” situation would call worker attention to review a recording system.
Moreover, it is implied the maintenance crew sent out to determine the valve closure reported it as being due to freezing in the air supply operating pipe system. They mention no maintenance on this problem. The question is: how can an ice blockage in an air supply line cause this? The air pressure would still be on the valve holding it open in front of the ice formation and pressure would still be on the back of the ice as long as supply was there. This would be equivalent of “locking” the valve open – NOT causing it to fail-safe close. Not only that, but maintenance workers would not know there was “ice’ in the line unless they took it apart and tried to blow it out. Yet after plugging the gauge leak, they restarted the liquid line flow with no mention of valve operator repair. The statement of clean-up of ground and “restarted the line” is not supported with description of maintenance that was done regarding finding and correcting the valve shut-down. A further question is why was a valve to the gauge left open with nobody at the site doing anything and needing a reading and why, on a remote valve, was it not a under a “lock/unlock” when needed basis?
The most plausible explanation is that the leaking gauge hit a threshold pressure drop for the liquid line, by ever increasing flow over the entire leak period, until it caused the automatic valve closure.
Will Bargath, Williams and/or WPX be fined for this release? [emphasis added]
Based on information provided by Bargath, the release of natural gas liquids occurred due to the failure of a pressure gauge attached to a four- (4) inch pipeline that is used to transport natural gas liquids from the gas plant to storage tanks, prior to sale as a product. The release occurred due to the fracture of a small diameter (about the size of your pinky finger) behind the face/dial of a gauge that was about four inches in diameter. There are no regular maintenance requirements for the inside of a gauge so the failure of the gauge was not the result of poor maintenance or negligence. After determining that the source of the leak, Bargath immediately replaced all gauges of that same model on their pipelines.
The leak in the pressure gauge was discovered just after midnight on January 3, 2013 by maintenance workers responding to reset a valve that had closed automatically due to the cold weather. The maintenance workers noticed the leak coming from the pressure gauge and closed the valve that turned off flow to the gauge, which stopped the leak. The maintenance workers believed that the broken gauge had occurred due to a pressure spike that happened when the automatic valve closed. The maintenance workers shoveled up all of the snow and unfrozen soil that appeared to be impacted by the release and restarted the pipeline.
As noted above, the natural gas liquids leak occurred through a very small diameter pipe at a rate of less than three (3) gallons per minute. The pressure drop and volume reduction caused by the release was not a large enough magnitude to trigger alarms in the automatic sensors that monitor each pipeline at the plant. It was not until after the actual significance of the release was realized, and further investigation of product inventories in March that Bargath realized the failure in the pressure gauge had actually begun on December 20, 2012 and was stopped on January 3, 2013. Again, the Colorado Department of Public Health and Environment (state health department) has no reason to believe that the initial lack of release discovery was due to negligence on the part of Bargath.
The determination that Bargath was not negligent in the cause and non-discovery of the natural gas liquids leak is very important because the state health department does not have any authority to regulate the natural gas liquids in the pipeline (because they are product) or the pipeline itself. The state health department’s authority only begins once the natural gas liquids have reached the ground as a result of a spill, because once a product hits the soil and groundwater it becomes classified as a waste. Disposal of waste without a permit is a violation of Colorado’s solid waste laws, and in this case, also the hazardous waste laws. If the state health department had determined that Bargath intentionally released (disposed) of natural gas liquids, or that significant negligence on Bargath’s part had resulted in the release (disposal), then the state health department would have sought to assign significant fines to Bargath. However, in this case, the natural gas liquids released had a significant value to Bargath that was lost as a result of the release. There would never be a reason for Bargath to intentionally release natural gas liquids – their product – or ignore a known leak.
To date, Bargath staff have done everything possible and required by authorities to contain the leak. Immediately upon receiving direction from appropriate experts, staff began removing hydrocarbons from the soil and groundwater. The ongoing remediation efforts have caused the company to spend a very large amount of money and it will continue to do so until the cleanup is complete. State health department, the Colorado Oil and Gas Conservation Commission (COGCC) and the U. S. Environmental Protection Agency (EPA) staff have overseen every action Bargath has taken to ensure the leak is cleaned up as quickly, efficiently and completely as possible.
Considering the present information, the Hazardous Materials and Waste Management Division of the state health department (the division) itself does not plan to assess Bargath any penalties for the leak. If subsequent information changes, penalties would be reconsidered. Fines could be levied if Bargath does not follow the division’s orders requiring the company to clean up the environment. The division’s decision not to assess penalties under its enforcement authority does not prevent the Water Quality Control Division or the Air Pollution Control Division from independently from imposing penalties. The possibility of fining the company is still under consideration.
The liquid hydrocarbons are kept liquid by pressure and/or low temperature. The outside temperatures during much of the period of the leak were sub-freezing and even below 0° F. Even with 4-inch pipe section isolated, and original liquid in line, the hydraulic pressure would NOT cease as it would for incompressible fluid. This is because the mixture involved gases compressed and chilled to liquid state and is NOT totally incompressible fluid. The original 200 psi fluid would continue to ablate, by vaporization increasing volume in a confined pipe, and maintaining pressure relieving blowdown virtually until the pipe emptied.
It was stated above that the leak was “less than three (3) gallons per minute.” I had provided all parties with calculations that showed a median flow of 22 ½ gallons per minute and a range of 19.9 to 25 gallons per minute. Now with 80% vapor and 20% liquid coming out, 3.8 gallons per minute would be just the liquid at the low end of the total leak rate. Interestingly the above report says the hole was about the size of a person’s “pinky” which is very close to the less than a ¼ inch diameter I used as a minimum feed pipe to the gauge. Now the crew returned the gauge to the plant and any proper procedures would have had an evaluation made by someone competent or trained to do so.
Even looking at a response time of within 10 minutes, of the PRESSURE SPIKE and the gauge bursting that night, would have calculated the worst case of 250 gallons spilled total in 10 minutes and for an incident nothing less than “worst case” should be acceptable to operator or regulator.
January 3, 2013 to March 8, 2013
Bargath had been fined $275,000.00 for failing to implement/or maintain storm water measures to prevent potential pollutants during planned construction. The State notified Bargath in November 2010 of violations and order immediate action that was not taken for 7 months (2011). They created a record of inaction to a problem.
With this in mind, Bargath may have concluded that the spill was in the neighborhood of the 6 bbls (250 gallons above) if they did any reconsideration of the maintenance crew estimates. BUT, time had expired for required reporting, and ground was still frozen, and it would take marshalling of trenchers, booms, and vacuum trucks, and contracting temporary employees.
Nine (9) weeks later, Bargath had such equipment and manpower put together, but to allegedly ascertain pipeline locations for another (expansion) pipeline. Here I say “allegedly” because hydro-trenchers are far more likely to damage pipes than using map confirmation, stream anchor locations, probes and metal detection equipment that is more passive. Instead, they “discovered” the hydrocarbons of the spill and immediately changed mission to ascertaining the extent and removal of hydrocarbons. But it became obvious that what they were collecting even more than 6 bbls.
Contrary to Kent Kuster’s above statement of delay of “3 months,” it was about 9 weeks before remediation started.
The initial trenching was inside the 30 yard distance from the valve set (and down grade), as referenced in paragraph labeled 1. (above), wherein Mr. Kuster said the maintenance crew checked. And to quote another release of information above [emphasis added]: “It was not until after the actual significance of the release was realized, and further investigation of product inventories in March that Bargath realized the failure in the pressure gauge had actually begun on December 20, 2012 and was stopped on January 3, 2013,” which would then bring realization that the spill had been longer and of greater volume than expected even from an expanded maintenance crew time spill of 6 bbls. I would draw attention to “further investigation,” which implies some investigation had occurred on product inventory as they had to realize a time volume of liquids from the plant to the valve from the line was lost during the response time on Jan. 3, 2013.
It is also curious, that Bargath introduced the information of the gauge leak to the COGCC in statements and valve set picture notes, but the inspectors never picked up on the importance of this leak. It was not until the April 4, 2013 Garfield County Energy Advisory Board meeting, when I heard a gauge had burst that I brought forth that the gauge was the leak source and followed the next day with calculations to show the rate of such a leak. Days later, the gauge was acknowledged as the source.
Now being in a situation of again violating rules after a fine for ignoring/delaying compliance to rules, is it possible Bargath chose try to clean-up what it might have considered a “6 bbls” spill only to discover a huge spill had occurred?
Size of the Spill From December 20, 2012 to January 3, 2013
By their estimate from flow meter recordings, they lost 1,150 bbls or 48,300 gal.
By their estimate of 4/10/13, they lost 241 bbls in to the soil (and recovered 142 bbls by April 10) with 80% of all fluid evaporated. That implies the total loss was 1,205 bbls or 55 bbls more than numbers to CDPHE. By April 10, they had recovered 59% of the spill according to their claims.
The number of 1,150 bbls, made above, may have been the number submitted to either COGCC or CDPHE on a report. An amended report (Doc.# 2232360 pg. 3) Form 19 went to COGCC when the first report (Doc.# 2232359 pg.1) did not include ground water being contaminated. But according to Kuster’s letter above the 1,150 bbls, wherever made in a report, would be underreporting if their 4/10/13 estimate was accurate. This latter number, 1205 bbls has been used for correspondence to date and was never challenged by parties.
Further, this 1205 – 241 = 964 bbls (40,488 gallons) represents 40,488 x 0.1337 g/cf = 5,408cf of liquid with a weight density as liquid at about, for a low of 32 (for ethane) and 42 (natural gasoline) lb /cf , approximately 37 lbs/cf, then 5408 cf x 37 lb/cf= 205,504 lb or ~102.7 tons.
Or even using their 1,150 bbls , 909bbls evaporated, 38,178 gals = 5,104 cf x 37 lb/cf = 188,863 lb or ~ 94 tons.
This represents 94 to 102.7 tons of liquid hydrocarbons containing Hazardous Air Pollutants (HAPs) released from 12/20/2012 to 1/3/13 plus what continued to evaporate until clean-up started capturing vapors. That means the year 2013 was starting with a concentrated release of HAPs. The CDPHE, when investigating high birth defects for the end of 2013 in downwind (prevailing) areas DID NOT include this release in their report considerations.
But this massive release could reach every listed location from Meeker to Snowmass. The volume of the released hydrocarbons was 1000 fold, or more, over any single well site and liquids released were the essence of the most harmful.
Consider a rig in position for a 3-month period and what it would have as fugitive emissions. In Garfield County the 2011 emissions for area sources was 55 tons/day per for about 10,000 well locations or about 0.0055 tons/day/well. The 15-day period of spill put about 94 tons/15 days or 6.3 tons/day or 1,000+ times all the well O&G sources in Garfield County. Compare a well putting out 0.0055 tons/day at 10 miles to the spill putting out a 1,000 times that amount/day of 6.3 tons/day at 100 miles!
Such HAPs, as benzene and acetone, can be carried back to ground by precipitation. Kuster states “dispersed into the atmosphere and been carried out of the area.” The proper statement would have been “dispersed into the atmosphere to travel throughout the area of prevailing weather and air currents” where he thinks it was removed an inconsequential distance away from the gas plant. And while CDPHE checked for DBPs, they didn’t check for HAPs and as late as checks were made, contaminants could be long gone for any of the listed chemicals.
Further, this incident alone constitutes a common risk factor happening in the first trimester of 2013 pregnancies, with weather patterns able to carry to every location, and to able to be in both air and water temporarily from December 20th to March. This IS NOT pregnancies happening “well after the spill/release,” but a time coinciding right with the spill time frame. To make a statement by stating the clinics reported these in January 2014, “a year later” only places the spill at the opportune time to be significant.
In conclusion, I opine the due diligence of the regulators was seriously lacking in inspection, knowledge of operation, and observations of discrepancies of activities in the information flowing from this incident. While major focus was applied in ground clean-up reporting and detail, regulatory agencies were just too compliant in accepting hearsay of events and did not display technically competent enough ability to question more thoroughly, determine negligence issues or address possible cover-up, evaluate operation errors, and consider missed reporting deadline.
Excusing, or “satisfying the Department,” the operator on the basis of no economic benefit is non sequitur if the operator was involved in trying to avoid a potential penalty incurred; or because of early operator errors, minimize or quietly address the problem as a side operation. Upgrades, prevention steps, remediation, etc., are expected results and fine(s) expected if acts were intentional or negligent. The Department, on behalf of the people of the state, has the obligation to investigate when there are discrepancies and inconsistencies in the reports, records/logs (or lack thereof), and reconstruction of events that are questionable and are indicators of such intention acts, including later deception of facts, or negligence. Questions have not been resolved by the investigation and reports presented.
Then there is the other problem of the vapor that was leaked. This was dismissed by CDPHE as “dissipated in the atmosphere.” This was not natural gas that is commonly lost from equipment and operations, this was a concentrated mix of heavier and hazardous hydrocarbons. It was at least 1000 times the volume of vapors released at a point source like a rig and it was concentrated for the hazardous elements. This is a significant release of a magnitude much greater than what a rig might release in 15 days of completion activity. These heavier hydrocarbons would not have the buoyancy of methane and some can also be returned to ground and water by precipitation. Weather patterns can carry them up the river valleys in general North, and West and Southerly (in the case of the Roaring Fork Valley) directions. This could spread from Meeker to Snowmass and be the “common” factor of an introduction of many chemicals that would introduce the birth defects that spiked in the 9 months following this release. The CDPHE did not consider this release in their investigation of the birth problems of 2013!
As far as reports of maintenance crew involved, by what has been supplied to me, I would come to the conclusion that the actual events of Jan. 3 would have occurred more like the following:
a) An alarm went off with the valve closing and operator(s) noticed the spike on pressure recording equipment. This would likely have been the “hammer” of a quick closing valve. Such a shutdown would have required a supervisor being notified and in turn, dispatching a maintenance team. With the valve closing, there would have been speculation as to cause, but that cause most likely had to be valve closure due to leak on one side or the other of the valve, and there may have also been a plant valve set closure, and storage tank valve set closure (this to isolate the leaking line to about half the line).
b) The crew would have had to track the line to the mid-valve set (and I go by the fact of their use of nomenclature of valve “sets”) and since the line is underground, they would have been looking for a geyser of vapor coming up and a possible earth disturbance. However, depending on line of sight they may have spotted the vapor cloud at the leak site and proceeded there directly.
c) Arriving on site, they should have seen, heard, and known the exact location and device leaking. Turning off the valve on the gauge line, they brought the leak to a stop. If it was pouring 3 to 4 gallons a minute, they may have even put a 5 gal. bucket under the gauge (which would have given them some method to estimate a one-minute flow time of 3 gallons).
d) Looking around, they could have seen spray and condensate patterns on the snow, or they may have looked as far as they had seen the vapor cloud (30 yards). As far as stating they could not see “hydrocarbons on the ground,” that makes no sense. It could be they couldn’t see a “wetness” and an area of complete snow melt. But what would a clear or even opaque fluid look like (clean motor oil would be difficult to see except as “wet”)? However, one report says they shoveled up soil and removed it (with buckets?) so snow was evidently gone from some area.
e) The supposition, that an ice blockage in the pneumatic system occurred, is not verified in the reports to say they removed any pneumatic lines to clear them. Most control pneumatics use desiccant dryers to remove moisture to prevent such ice blockage from happening and corrosion/malfunction of operators, and if natural gas was the pneumatic fluid, it would have undergone dehydration. It might have been discussed as a possibility because of the cold weather when their response began, but then was quickly dismissed when the vapor cloud was spotted. This appears more as a diversion to explain the realization of the valve closing. However, in a pneumatic hold open situation, as explained above, ice forming would not mean pressure release. It is possible the supposition would have involved an ice rupture (that is what was originally thought possible on the gauge) and such rupture would have been evident with external inspection of pneumatic lines. But once again, there is nothing in the activities to indicate that they did anything about pneumatic lines after seeing the leaking gauge.
f) After removing the gauge (and plugging, because “replacing” would require carrying a spare gauge and not be very likely), they resumed liquid transferring. This indicates they knew they had resolved the problem. They did not report checking the line from valve set to tanks other than by resuming line flow. It would however, make the time of the response and turn off very important to estimating Jan. 3 spillage, because they knew the line from the plant was bleeding down until they turned off the gauge valve. If it took 15 minutes from valve closure to arrival at site, 3 gallons a minute would have spilled a reportable bbl. (~45 gallons). It is a conclusion they could quickly come to that they had spilled more than the “less than 25 gallons,” but it may have been days later when engineers or management people started looking at it. Still, if under the impression Jan. 3 was the only leak time, they could have assumed no more than the volume of the line from the plant to the valve leaked and when the ground thawed, it might be easily cleaned up. Besides, they had already suffered a fine for one infraction and didn’t need another, as a missed timely report might cause.
*Bob Arrington is a retired engineer and the Battlement Mesa citizen representative on Garfield County’s Energy Advisory Board (EAB). He also represents the Grand Valley Citizens Alliance and the Battlement Concerned Citizens.