Could the Northern California Wildfires Have Been Prevented Using Preemptive Power Outages and High-Resolution Weather Forecasts?

The catastrophic fires in northern California are still burning, with the death toll rising to 21 and damage estimates ranging into the tens of billions of dollars.

Is is possible that this tragedy could have been prevented or minimized by cutting the power to threatened areas before the fires started, using the best available weather forecast models to guide decision making?

This question is explored in this blog.

The proximate cause of the explosive fires were discussed in my previous blog.  We started with a very dry landscape, following the typical rain-free summer.  Wind picked dramatically on Sunday evening, with gusts to 40-70 mph over northern CA.  With offshore flow, relative humidities were very low.  And temperatures had been above normal.

But something initiated the fires and did so at multiple locations within a period of a few hours during the late evening on Sunday.  Although there is the possibility of arson, the most probable fire starter was arcing power lines damaged or shorted by falling trees and branches.

There is a history of California wildfires started by falling trees/branches during strong wind events, such as the 2015 Butte Fire near Sacramento that killed two and destroyed 550 homes.  And the there were reports of downed and arcing  power lines Sunday evening prior to the wildfire conflagration.

So if downed or arcing power lines was the key initiator of Sunday/Monday's fires what can we do to lessen the chances of a repeat of the tragedy?

Some media reports have suggested that the relevant power company (PG&E) was not effective in trimming vegetation around its power lines.   Certainly, an effective program of vegetation control around powerlines is essential.  And burying power lines in vulnerable areas would be wonderful, but very expensive (I have seen estimates of 1 million a mile)--but perhaps a reasonable investment.

But perhaps there is something else that can be done, which could provide substantial protection:  cutting the power to regions that are directly and immediately threatened by powerline-induced wildfires.

This is how it would work. 

This approach would only take place in regions and periods in which there are threatening amounts of dry fuels on the ground, making wildfires possible.

Thus, portions of California near vegetated area during the dry season (last summer, early fall) would be candidates.

Only areas with appreciable population would be candidates, thus remote areas would not be considered.

Only when dry atmospheric conditions and high winds are imminent and threatening, would pre-emptive blackouts be considered.  I would suggest using periods when gusts are predicted to exceed 40 mph (35 knots) with relative humidities less than 30% as a potential criterion.  If those conditions are forecast to exist within 6 hours or if they are observed, the power would be cut for the affected areas.

A very promising modeling system for such a purpose is the NOAA/NWS HRRR (High-Resolution Rapid Refresh) high-resolution model, which is run every hour, out to 18 h.    Here are the ten and four hour forecast of wind gusts valid at midnight Sunday/Monday (0700 UTC) for central CA.  Both forecasts are threatening, with predicted winds over 40 mph in many of the areas north of San Francisco where they were, in fact, observed.

Both showed the winds revving up above 40 mph around 8 PM and dropping below that value around 9 AM Monday morning. 

So the pre-emptive black out would run for 11 hours (8 PM to 9 AM) and folks would get a series of warnings that it would occur.   With modern numerical prediction, warnings of a potential blackout would be given a few days before, with a penultimate warning 6 hrs before, and a final warning an hour before.

Yes, there would be some inconvenience, but that would be minor compared to the benefits.  In the present case we are talking about saving roughly two-dozen lives and tens of billions of dollars of economic impacts.  And we haven't even touched on the negative impact on air quality for the heavily populated San Francisco metro area.

And there is the issue of false warnings or the "crying wolf" syndrome.  But a casual look at the climatology at a few locations around the area suggest that this was an unusual event, and that picking a realistic criterion (e.g., 40 mph gusts, August through October only, relative humidity below 30%) would produce very few blackout events.   I will explore this more during the next few weeks.

Is this a crazy idea?  If so, why?