Two years ago, on January 26, 2015, a pipeline exploded near Follansbee, WV. There were no injuries: it happened in a rural forested area, and no one was around. But that particular accident does have lessons for us as we contemplate what it would mean to have the Dragonpipe (Mariner East 2 pipeline) running through our neighborhoods. The Follansbee pipeline has several key features in common with the Dragonpipe. It remains the best available example of what a Dragonpipe failure might look like.

Like the Dragonpipe, the Follansbee pipeline was carrying “natural gas liquids”—highly volatile gasses, produced as a byproduct of natural gas fracking, and compressed into liquid form. The pipeline was the same size as the largest of the three Dragonpipes (20 inches in diameter) and was new, having been installed just over a year previously. Although the rural setting was different from our suburban neighborhoods, we can get some clues from the Follansbee accident concerning two key issues: the mechanism that could cause the Dragonpipe to fail, and the type of damage that could occur if it did.

The cause of the Follansbee explosion. The Pipeline and Hazardous Material Safety Administration (PHMSA) investigated the cause of the Follansbee explosion. In their report, the primary cause of pipeline failure was the bending of the pipe, due to subsidence (ground giving way) underneath it. The bending stressed the pipe and caused it to break at a weld.

In investigating the local geology, PHMSA found that the subsidence was related to an old mine at the site that was presumably slowly collapsing and causing the ground above it to drop.

Could something similar happen here? As far as I know, there aren’t many old mining sites around here. But there is a much more widespread threat. It is called “karst geology”. Karst is the geologist’s term for a type of soluble rock (typically limestone). In karst areas, groundwater gets into cracks in the rocks and slowly, over the centuries, enlarges them and turns them into underground caverns and rivers. If these underground spaces (referred to as “voids”) get large enough, the weight of the ground above them can cause them to collapse. If they are close to the surface, that results in a sinkhole.

The geology of Chester County is dominated by karst. (There is comparatively little of it in Delaware County.) Southeastern Pennsylvania’s “great valley”, through which Route 202 runs, is composed of karst. Exton is located in the center of that valley. Anyone driving through Exton on Route 100 is familiar with the steep drop going into town and the steep climb coming out the other side. That valley exists, in large part, due to the gradual collapse of the underlying karst over the millennia. And the process continues today. There are hundreds of sinkholes in Chester County  and new ones continue to occur from time to time.

So subsidence (possibly resulting in a sinkhole) is a real and widespread issue in the Exton area. It is due to basic geology, not old mines, but if it took place under the Dragonpipe, it could cause the type of bending stress and pipeline failure that caused the Follansbee explosion. That is the first lesson of Follansbee. Sunoco needs to tell us how the Dragonpipe will be protected from a failure like that.

The damage at Follansbee. The extent of damage from an explosion like the one at Follansbee is highly variable, depending on factors such as how far underground the pipe is, whether the gasses leaked for a while before exploding, and how quickly the flow in the pipeline is shut off once the leak is detected. Those factors limit the information we can glean from a single incident. Still, there are some things we can learn from Follansbee.

The Follansbee explosion occurred on the morning of January 26, and the ensuing fire could not be extinguished until the evening of the following day. That is expected for an explosion of this type: the explosive material in the pipe must simply burn itself out. In the Follansbee case, two remotely-controlled valves were closed from the company control room when the pressure drop from the rupture was detected. That isolated a section of pipeline about 21 miles long. Then, when crews were able to get to the area, they closed manual valves, further limiting the flow to a stretch of about 3 miles. After that, nothing could be done except to wait for the fire to burn out.

If a similar rupture occurred with the Dragonpipe, it is not clear how long a stretch of pipe could be isolated. Sunoco has apparently decided not to build a planned valve station in West Goshen, and I am not sure where the other valve stations will be and which ones will be remotely controlled. We know that they were planned to be several miles apart. (There is information about valve locations buried in some of the paperwork that Sunoco has filed with the DEP. Perhaps a reader can help unearth that information.)

The nearest structure to the Follansbee explosion was a house about 2,000 feet away. It survived, but its vinyl siding was badly melted. We have to assume that there would be similar thermal damage at the same distance in the case of a Dragonpipe explosion.

Follansbee siding damage 1-27-18
Thermal damage to vinyl siding and shutters at 2,000 feet.

There were no structures in the immediate vicinity of the explosion, but five acres of trees were destroyed. We can assume that a similar area of destruction could occur in a Dragonpipe explosion, in which case several dozen houses might be destroyed.

The Follansbee pipeline was 12 feet underground at the point of rupture. The Dragonpipe, in locations where it is being constructed by trenching, is generally just 4 feet underground. (In areas where horizontal directional drilling is used, the Dragonpipe may be much farther underground. At valve stations along the line, it is above ground.) A rupture at a shallower depth would be likely to damage a larger area than the Follansbee explosion did.

Follansbee rupture site 1-27-18
The crater from the Follansbee explosion. The pipeline was 12 feet underground at this point.

It could happen here. The Follansbee explosion is an example of what could happen here. The pipe is the same size and carries the same materials as are planned for the Dragonpipe. A rupture here could create just as large an explosion as in Follansbee (or larger, if it should happen at a point where the pipe’s depth is shallower). The karst landscape that the route goes through means that subsidence may well trigger a problem, as it did in Follansbee.

Unfortunately, the consequences here could be far worse than in Follansbee. In parts of the pipeline route in Delaware and Chester Counties, the population density is such that a Follansbee-type explosion would almost certainly result in many deaths.

Whether Sunoco has the legal right to put our lives at risk so that the company can profit from the production of plastic in Europe is a matter for the courts to decide. But even if it is within the law, it is morally wrong and the pipeline must be stopped.