The Real Story Behind the California Wildfires

There has been a huge amount of media coverage regarding the tragic northern California fires, documenting the terrible loss of life and billions of dollars of damage to buildings, infrastructure, and the economy.  As I write this, the death toll has risen to 41, over 5000 buildings have been destroyed or damaged, and the estimates of the financial loss are in the tens of billions of dollars.

Media stories have blamed the catastrophic fires on many things:  a dry environment after the typical summer drought, unusual warmth the past several months, excessive rainfall producing lots of flammable grass, strong winds, global warming, and  the lack of vegetative maintenance (clearing of the power lineright-of-ways) by the local utility (PG&E).

But none of the stories I have read get at what I believe is the real truth behind this unprecedented, severe, and explosively developing wildfire event:

A unique mountain-wave windstorm produced the strongest winds in the historical record at some locations.  An event produced by the unlucky development of just the right flow regime that interacted with regional mountains to produce extreme winds beyond contemporary experience.

In short, this blog will make the case that the extreme nature of the wildfires was the result of a very unusual weather event, one that our weather models had the ability to forecast and warn about, if only their output were applied more effectively.  The blog also suggests that better use of state-of-the-art weather prediction offers the hope of preventing a similar tragedy.

The Unique Wind Event

Although there have been a lot of media reports about windy conditions, few have described the extreme, often unprecedented, nature of the winds on Sunday night and Monday morning (October 8/9th).   Some have even mocked PG&Es claims of hurricane-force winds, suggesting wind speeds of 30-40 mph.

Let's clarify a few things.  There was a wide range of winds that night, with the strongest winds on ridge tops and on the upper lee slopes of terrain.  Some winds was startling.

For example, at 10:30 PM on 9 Oct 2017 the wind gusted to 96 mph on a 3400 foot peak NE of Geyersville, about 20 miles NNW of downtown Santa Rosa. They reported sustained 74 knots (85 mph).  Those are hurricane force winds (sustained of 64 knots or more).

At the Santa Rosa RAWS station (U.S Forest Service and Bureau of Land Management) at 576 ft elevation, the wind accelerated rapidly Sunday night to 68 mph (see below).

A few miles to the NNW and a bit higher (2000 ft), winds at the Hawkeye site accelerated abruptly to 79 mph.

Unprecedented

What is really amazing about the winds at these sites, was that they were unprecedented:  the strongest winds on record, with records going back to 1991 (Santa Rosa) or 1993 (Hawkeye).  And we are not talking about winds during the fall, but winds any time during the year.  Even during the stormy winter season when powerful storms can cross the region.

At low-levels, the situation was more mixed.  For example, at Napa Valley Airport (36 ft), the sustained winds at 11:15 PM October 9 (37 knots) were the strongest observed (looking back to 2001) at that location from July 1- November 30, while at the Santa Rosa Airport (KSTS) the sustained winds only reached 28 mph, with 40 mph gusts.

So why were the winds so strong and unprecedented at higher levels in the hills?  These winds were key for causing the wildfires to explode and to quickly move into populated regions.  And the winds undoubtedly damaged power transmission lines and thus helped start electrical fires, which may, in fact, have initiated the big wildfire runs.  And why were the lower-level winds less severe?  What can explain such differences?

Mountain Wave/Hydraulic Jump/Downslope Winds

When strong flow interacts with terrain, the air can be greatly accelerated.  The schematics below show you some situations, with air accelerated over and downstream of mountain crests.

Such acceleration is well know in Washington State, with some locations experiencing huge winds (like Enumclaw where winds reached 120 mph on Dec. 24, 1983 while it was calm in Seattle.)

That night (Sunday evening), strong to moderate easterly/easterly flow was approaching the terrain north of San Francisco, something shown by the 6hr forecast of height (like pressure) and winds at 850 hPa (about 5000 ft)--a forecast valid at 11 PM Sunday night (see below).  The strong winds and their orientation was the result of cooler air and high pressure moving into the Northwest during the previous day.

So we had modestly strong winds (30-50 knots) approaching the terrain.  A very favorable situation for strong mountain-wave winds is a stable layer at or just above crest level.  A stable layer can be noted when temperature is steady with height or increases with height (an inversion).   The nearest vertical sounding (radiosonde launched weather sensors) was at Oakland, CA.  The sounding at 5 PM Sunday DOES shows an inversion at the right levels (see plot), roughly between 850 and 800 hPa (roughly 5-6 thousand feet ASL).

A group at the Desert Research Institute runs a forecast model (WRF) at very high resolution (2-km grid spacing).  Here is their 6-h forecast for sustained surface winds at  11 AM Sunday.

OMG...there it is.  You can see the banded structure of strong winds over and immediately downstream of major terrain features, with lower speed winds near sea level.  I inserted a terrain map below...you can see how the wind maxima were oriented the same way as the ridge lines.  And the model reveals something else:  the enormous  horizontal variability of the winds during such events.

Other major modeling systems also predicted the strong mountain-wave winds.  For example, here are the max wind gusts predicted by the NOAA/NWS High Resolution Rapid Refresh Model for 2 AM Monday.  Same banded structure, with gusts near Santa Rosa of 50-55 knot (58-63 mph)

Professor Rob Fovell of the University of Albany completed another high resolution simulation, one initialized at 5 PM on Thursday.  Here are is a vertical cross section though the Tubbs fire that affected Santa Rosa.  You can see the acceleration of winds (sustained) on the slopes.

The predicted winds at the Tubbs site was scary strong, with max winds around 70 mph.

The creation of such downslope mountain-wave type windstorms is very sensitive to the characteristics of the air moving towards the mountains.  You not only need strong approaching flow, but the proper vertical structure of temperature and winds.  Clearly such conditions don't happen often--otherwise similarly strong winds would have occurred before.  There is no reason to expect that such extreme wind conditions were made more probable by global warming.

So I think we can outline what happened Sunday/Monday of last week.

The vegetation was dry after little rain over the summer (quite normal).  The ground vegetation was perhaps drier than normal because the summer had been usually warm (by 1-4 F as shown by the NOAA Western Region Climate Center map for the last 90 days.)

On Sunday afternoon, winds approaching the mountains of northern CA increased, and the vertical structure of an inversion over cooler air was established.  A strong mountain wave/downslope wind event was initiated, bringing winds of 60-90 mph to the crests and upper lee slopes of the regional terrain.  Such winds helped initiate the fires (possibly due to interaction with power lines) and then caused the resulting fires to explode.  The fires, driven by the strong, gusty winds, pushed very rapidly into populated areas.

The good news in all this?  Our models seemed to be able to simulate this event, providing some warning of the imminent wind acceleration.

What could be done with such information?  Much better warnings of a potential blow up?   Shutting off the power to threatened communities?  There are lots of possibilities.   But one thing is for sure:  we can not let this happen again.  And the first step is to really understand what happened, without assuming we know the answer beforehand.

Too many people are suggesting the wildfire event is all about climate change, when it may prove to reflect a severe weather event unrelated to global warming.  Similarly, some of the same folks claimed that the great rainfall with Hurricane Harvey was all about global warming, when a stalled storm was probably more to blame.  Only by knowing the true cause of disasters and acting on that information can we protect people in the future.

Cold Weather and Snow Hits the Pacific Northwest

Snow has hit the Cascade passes and below-normal temperatures have spread over the Pacific Northwest, with temperatures dropping into the single digits in portions of eastern Oregon.

And ironically such cold temperatures are a bad sign for those battling the wildfires north of San Francisco.

To "warm up" this blog, lets start with the latest cam shots at Snoqualmie and Stevens Passes.  White stuff.  Enough to make folks think about the upcoming winter season (which may be a good one because of La Nina).

During the last day, many NW folks have observed frost, with temperatures dropping below freezing on both sides of the Cascades.  The map of minimum temperatures for the 24-h ending 8 AM Saturday (below, click to expand), shows 20s in eastern Washington, with teens and even single digits in the valleys of the high plateau of Oregon.  Klamath Marsh RAWS site east of Crater Lake dropped to 5F and Burns, Oregon set a new record for the date (10F).

The below-normal temperatures were clearly evident at Sea Tac and Pasco, WA--here are plots for the past two days at these locations, with average maxima (purple) and minima (light blue).    At Sea-Tac, Thursday was crazy cold (about 12F below normal) and this mornings low temperatures were clearly the coldest so far this fall.

And many days dropping below normal at Pasco.

The air morning into our region is near record cold, something shown by comparing the incoming temperature at 5000 ft (850 hPa) at Quillayute (Wa coast) with climatology (see below).  The gray dot shows the observation at 5 PM Friday and the blue line shows the record low for the date.

The cold air that moved into the Pacific Northwest is associated with higher pressure, something shown by the 12-h forecast for 5 AM today (Saturday) for sea level pressure (solid lines) and lower atmosphere temperature (color shading).  At the leading edge of the cold/high pressure there is a large change in pressure (pressure gradient) that is associated with strong winds.  Unfortunately, some of that pressure-change zone is now over northern CA, which is revving up the winds, particularly over the northern Sierra.   Not as bad as Sunday/Monday, but enough to bring concerns of reinvigorated fires.

The winds yesterday were mostly northerly (from the north) over northern CA, and that blew the smoke southward toward San Francisco (see MODIS image).

I am working on an analysis of winds during this fire event, particular an evaluation of how unusual they were...stay tuned.