The Associated Press story says that in their report on the huge Eaton fire, “Los Angeles County fire officials” say it “was caused by electrical arcing on [Southern California Edison’s] out-of-service tower …”
Looks like “out-of-service” doesn’t mean what you would think. What does it mean, here?
Maybe it wasn’t out-of-service before, but it is now?
This New York Times article says
For months, the utility and lawyers for customers had suspected that an inactive line might have been responsible for the Eaton fire. Even when out of service, the lines, sometimes referred to as “zombie” lines, can act as a conductor, picking up energy from active high-voltage lines nearby.
There was a story once about a farmer who had a fence running next to a power line. He got a shock from his fence. When he wired it to a light bulb, it lit up. The power company got upset but there was not much they could actually do to stop him.
Still, coupling enough power to actually cause arcing seems like a stretch.
It’s definitely possible, and has happened. One good paper on the topic:
It’s more commonly an issue with worker safety rather than wildfires though. There’s even an IEEE guide on worker safety around de-energized lines that can have induced voltages or currents: IEEE SA - IEEE 1727-2013/Cor 1-2015
More on the physics of how this happens: https://www.pacificorp.com/content/dam/pcorp/documents/en/pacificorp/wildfire-mitigation/responses-issued/2026-2028-wmp-responses/Attach_OEIS_7.6-2.pdf
There are pros and cons to grounding the unused conductors, or letting them float completely disconnected. Grounding them results in induced currents, letting them float (ungrounded) results in induced voltages; both can lead to problems.
Excellent. Had no idea.
And this is why I love The Dope. 
A hundred volts or so is a definite worker safety issue. Seems like it would take thousands to cause arcing.
Initiating a spark across a clean air gap takes kilovolts, but that’s not how these fires start. Ignition is mostly from heat, a sustained current of just tens of mA can start a fire, and induced currents on dead lines are up to an amp (with grounded conductors forming a loop). For example, dry vegetation with that current pushed through it, a failing connection that turns into a high impedance at a single spot where the power is concentrated, etc. This is the kind of low-energy, high impedance fire threat that residential arc-fault breakers guard against at a house. Also, striking an arc across a gap needs kV; drawing one at a corroded or separating connection needs almost none - just like an arcing 120V switch. An ungrounded line really does sit at kilovolts though - that first paper showed 2 kV of induced voltage from a 115 kV energized line. Also, during a fault on the adjacent energized circuit, the induced current jumps far above the steady state numbers.
That is an interesting article on deenergized power lines. The authors ought to get up to speed on transmission line theory though. The basic physics should be the same as the digital electronics world has been dealing with for years now. The dimensions are a little different. I take it that there was coupling between a big and still energized transmission line and a deenergized smaller distribution line.
The dry manzanita in the foothills of the San Gabriel Mountains seems like it would be a good insulator. It would not be easy to get it to conduct 10 mA for an extended period. Once it does start burning though, you do have a serious fire on your hands.
For those that didn’t see it at the time, here is video from a service station camera showing the sparks that started the fire.