On the evening of May 31, 2021, a leak occurred at Sunoco’s Boot Road pump station (next to the Boot Road overpass over Route 202 in Chester County). We can learn a lot about emergency response procedures and preparedness from reviewing what happened that evening.
We don’t know when the leak began, but it was detected by a hydrocarbon alarm system within the pump station around 9:45 p.m. The system alerted a pipeline controller in a remote office, who shut down the pump and shut valves to isolate the leaking area. The Chester County Emergency Management System was notified. This sequence of events is contained in Sunoco’s “Accident Report” to the Pipeline and Hazardous Materials Safety Administration (PHMSA), filed on June 30. The report does not specify what the leaking gas was—it is described as “NGL”, which can mean ethane, propane, or butane. They all form large, ground-hugging, explosive clouds.
A series of problems. What happened once local authorities were notified is not described in Sunoco’s report. It was a series of missteps that illustrate what’s wrong with the current emergency response plans. I lay out some of those problems below. But first, let me say that much of the narrative that follows is pieced together from reports by residents and local officials, and I can’t guarantee that it is trustworthy. That in itself is a problem: if residents are to feel safe, they need a full and open accounting of this event. There is no such accounting.
- Uncertainty around the extent of the leak. Because of a sensor in the building that houses the pumping station, Sunoco knew that there was a leak, and perhaps the company knew the concentration of the gas inside the building. But it evidently did not know anything about the presence or extent of a flammable cloud outside the building, generated by the leak. (Or, if Sunoco knew, it apparently did not inform local first responders. When subsequently asked about external monitoring, they were evasive.) If the possible extent of a flammable cloud is unknown, that’s a problem, because the correct action to take depends on how extensive the cloud is.
- Inappropriate first-responder action. That lack of information contributed to the rash decision by a first responder to drive up to within 150 feet of the pump station. If there had been a flammable cloud, the first-responder’s vehicle would have ignited it.
- Lack of notification. Most residents in the area had no idea what was going on. Boot Road and Route 202 were both closed down, but no one was notified about the cause. Chester County Department of Emergency Services later said they would have gone door-to-door notifying residents if there had been “a real problem”. But how did they know this was not “a real problem”? It was apparently “real” enough to shut down traffic on the region’s busiest highway. And if they did decide to notify residents, how would they do so? Would they enter a vapor cloud? Would they ring doorbells (an ignition source)? Would they use the “Ready Chesco” emergency system? It is phone-based, but the use of phones, including cell phones, is something PHMSA specifically warns against, because they are also potential ignition sources.
- Inappropriate resident response. Residents were not told what to do. Should they evacuate? Should they shelter in place? At least two residents took matters into their own hands. They jumped into their cars and drove out of the neighborhood. That was the worst possible response, because it would ignite the flammable cloud if there were one.
- Impossibility of following Sunoco’s guidance. If residents had known of the leak (which they didn’t), what would the appropriate response have been? The Sunoco-issued guidance is to head uphill and upwind, on foot. Trying to actually do that would have presented serious problems. First, the pump station is near the top of a hill. Anyone in the vicinity who headed uphill would have been going toward the leak. Second, the evening was basically calm. (Area airports reported a breeze of 5mph or less.) How would a resident know which way was upwind? And exactly these conditions are particularly dangerous because flammable clouds flow downhill and reach their greatest extent in calm conditions.
We were lucky this time: the leak was inside the pump station, where Sunoco had equipment that could detect and report it; and it was confined to the pump station, so local residents were not endangered. Sunoco was able to direct the air in the pump station and the remaining gas in the line to an emergency flare, where it burned off harmlessly. But what if the leak had been more serious and had not been inside the pump station building?
The cause of the leak. In Sunoco’s report to PHMSA, the cause of the leak is described as a “pump seal leak”. According to the report: “Investigation determined that the seal failed due to the inboard seal set screws loosening which allowed the rotating seal to axially shift causing the seal faces to contact resulting in damage to the faces and o-ring seal.” And what caused those screws to loosen? Sunoco’s report says it was “excessive vibration”.
Those statements are not reassuring. What caused the axis of the rotating seal to shift? Did that shift cause the “excessive vibration” or was it the result of the vibration? Could it be that an explosion in this same pump station less than two years earlier was the root cause? The next time this happens, will the problem be confined to the pump station building?
The bigger issue. Apart from this particular event, the problems listed above point to a more general issue. The emergency response procedures did not work. Local residents were not safe. For the most part, this is not fault of local first responders; it is due to Sunoco’s failure to provide a workable plan for emergency response.
The next leak may not be near a Sunoco sensor. What happens then?
We know that pipelines leak. No matter what precautions are taken, sooner or later they leak, and Mariner East has already experienced small leaks. Someday, it will experience a large one. When that happens, will there be an emergency response plan in place that actually protects the public? If we can’t answer that question, the pipeline should be shut down until we can. Lives are at risk.
Bring your questions to the County Commisioners’ meeting. The monthly “sunshine meeting”, where matters are discussed prior to being voted on two days later, is on Tuesday July 20, 10:00-10:30 a.m. You can attend in person, via Zoom, or on the phone.
Details for how to participate are here:
Our County Commissioners have made statements about ensuring the safety of the pipelines, but they have failed to act. The Boot Road leak is a clear example of the danger. So are the massive problems in the wetlands near the library, where multiple sinkholes are occurring in the midst of active pipelines. So is the undermining of the library building itself by Sunoco’s work (as documented by a report the County itself commissioned).
This is your chance to let the Commissioners know that this is an urgent matter. Ask them to convene a public meeting to unearth the details of the Boot Road leak and the failure of emergency response. The Boot Road event was a warning; if nothing is done, the next leak could be a catastrophe.
From the accident report: Local time operator resources arrived onsite: 21:40 or 9:40 pm. This happened in PA so this would be PA local time. Fire Chief Grant Eberhard arrived onsite around 10:30 pm and was there before the Sunoco/Energy Transfer keyholder.
Local time & date of accident: 5/31/21 at 21:26 or 9:26 pm
(Is this PA or TX time?)
Local time & date of shutdown: 5/31/21 at 20:46 or 8:46 pm
Looks like they shut down the pump BEFORE there was an accident.
Local time operator identified failure: 5/31/21 at 20:26 or 8:26 pm
The timeline doesn’t look right.
You’re right. Field 4 of the report–“Local time (24-hr clock) and date of the Accident”–must have been entered incorrectly.