February 11, 2018

The Strange Case of the Quinault Blow Down: The Ultimate Solution

It is time to put the facts together and to explain the mystery.   Using high-resolution modeling, theory, studies in other locations, and available observations, we will attempt to solve the compelling scientific puzzle of the massive fall of old-growth and other trees on the north side of Lake Quinault during the early morning of January 27th.

Sherlock Holmes made use of a 7% solution of a certain drug to prepare for such cases.   I will make use of more appropriate drug to heighten my mental prowess, one used by many scientists in our region:   a tall Starbucks coffee.

Science is great fun, particularly for a difficult case like this.   Using limited observations and knowledge of basic physical principles, we attempt to explain natural phenomenon.     The enjoyment of an intellectual puzzle and a detective story.

And when the pieces come together, and when we gain an understanding of something that no one has understood before, the feeling is one of satisfaction and even elation.   A feeling that once experienced, becomes addictive.    A reason why many of us love being scientists.

So as Sherlock would say: "the game is on".  And I will describe my chain of logic, starting with known facts and then examining each possibility until we determine the most probable cause.

Let's review the facts

1.  A large tree fall occurred on the north side of Lake Quinault around 1:30 AM on January 27th. Many of the trees were old-growth, or at least, very large.

2.  The tree fall area was quite limited in size:  perhaps a half-mile on a side and extending from the lake toward the crest of about 2500 ft. 

3.  Several of the trees snapped off and this can only be explained by very strong winds (certainly at least 60-70 mph).

4.  The trees fell to the south and thus the winds must have been from the north.

5.  None of the limited surface observation locations in the area reported any winds even close to those needed to topple the trees.  For example, a site just across the Lake only reported light winds during the tree fall.

5.  The strong winds could NOT have been the result of microburst associated with a thunderstorm or strong convection.   Weather radar showed no such feature and the lightning detection network had no strikes in the region.

6.  An occluded front was approaching the coast at the time of big winds and tree fall.

The first question you should ask was whether the approaching weather system had strong northerly (from the north) winds associated with it.  Or even northerly winds at all.

We know that the surface winds with system did not have strong northerly winds  from the surface weather stations of the region.  But if there were northerly winds aloft, there would be the possibility of a downslope windstorm, as northerly winds accelerated down the slope north of the Lake.

Fortunately, there are sufficient observations to answer this question.  NOAA Earth Systems Research Lab (ESRL) maintains a device called a radar-wind profiler at Forks, Washington  (up the coast a bit) that is capable of determining the wind and temperatures aloft in real-time.  Here are the wind observations for 0000 UTC 27 January through 0000 UTC 29 January from the surface to 9 km above the surface.  Winds are shown by the typical wind barbs and are color coded .  The incident in question occurred at approximately 0930 UTC 27 January.  A blow up at the critical time is shown below as well.  Note that the front came in from the west and hit Forks before Lake Quinault.

The Bottom Line:  No hint of northerly flow during the period in question.  There were southeasterly winds at low levels, with increasing southerly and southwesterly winds aloft.

But we have another observing asset as well:  the Langley Hill radar near Hoquiam.  This radar is a Doppler radar and provides wind information aloft.  Specifically, it provides the radial wind component--the speed of precipitation (and the air it is in) towards or away from the radar.

Here is the Langley Hill radial velocities from the lowest scanning angle at 0927 UTC.  Green and blue indicate flow towards the radar, yellow/red/orange the opposite.  Not easy to read without experience.   But to my practiced eye, the radar suggests southeasterly winds of up to around 30 knots at low levels, turning to southerly and then southwesterly aloft.    Consistent with the profiler.  No northerlies

So if the air coming in off the Pacific was from the southeast east or south, where did the powerful northerlies come from?  Perhaps the comments pushing a secret government project,  aliens, or a meteorite strike were on to something.

Or perhaps not.  There is no evidence of any space object reaching the earth in this region (I checked).  And there IS a possible meteorological explanation:  a rotor circulation associated with a strong mountain lee wave.

But first some atmospheric rotor 101.  If fairly strong winds are approaching a mountain crest, they can undergo wavelike undulations in the lee of the barrier.  A situation in which air surges down the mountain and then suddenly rises up, followed by potentially more down and up motions.  If the wave has sufficient amplitude, a rotor can form underneath the wave, with flow moving in the opposite direction from the flow approaching the mountains.  You see why this is interesting...here is a way to get northerly flow when the general flow is southerly.

Mountain lee waves can increase in amplitude as the winds approaching the mountain strengthen.  But they can also amplify if there is a stable layer near the mountain crest, or if there is what is called a critical level above the crest level.  A stable layer is one where air temperature does not cool rapidly with height, and a critical level occurs when the wind component perpendicular to the mountain reverses direction.  These features help trap and amplify the low-level wave energy, producing stronger waves and stronger rotors.

But it is even better than this.  A large rotor can in turn break down into highly intense subrotors that can have strong winds associated with them.  Two colleagues of mine, James Doyle of the Navy Research Lab of Monterey and Dale Durran, a fellow faculty member at the UW, did a very nice paper showing the results of an ultra high-resolution simulation of these critters.  Here is a vertical cross section across the lee slopes mountain that shows the rotor and subrotors (indicated by the red colors).
And there have several observational studies of rotors, including the T-REX (the Terrain-Induced Rotor Experiment) project and intense studies near the Hong Kong Airport.

Could these conditions have occurred during the early morning hours of January 27th?  I think the answer could be yes.

As the offshore front approached, the wind approaching the crest to the south of the lake increased (see topographic map, which indicates the key terrain features and the direction of the flow).

During the period in question, cooler air near the surface (in  the southeasterly flow) was surmounted by warmer air above.   This results in increasing stability above crest level.  And with southeasterlies at the surface and southwesterlies developing aloft, this led to the development of a critical level, where the flow reversed.   So all the factors supporting a strong mountain lee wave and potentially a rotor were in place.

But do any observations suggest such a development? 

The development of a strong wave would result in substantial sinking along the lee slopes of the terrain feature.  Sinking causes warming and pressure falls.  We happen to have a weather observation just to the north of the terrain slopes (located on the south side of Lake Quinault, see map).    Wow...there was a sharp pressure fall around 1:30 AM, just as the big blowdown occurred (see below).  Suggestive.

But we have a tool that Sherlock would be envious of:  high resolution numerical simulations.  Considering the small scale of the blow down, I suspect we would need to run our model (called WRF) with uber-fine resolution (grid spacing of around 100 meters).   The best the National Weather Service models do is around 4-km.  Our UW WRF is 1.3 km.    But for this case, UW graduate student Robert Conrick took WRF down to 444 meters and fellow student Nick Weber has produced some nice graphics.   

So let us see whether we can simulate this event...or at least determine whether we are on the right track.  I am going to show you a series of vertical cross sections, oriented SSE-NNW, that pass over the blowdown site.  Each cross section will have potential temperature (solid lines), wind vectors  in the cross section, wind speed (color shading) and vertical motion (blue for descent and red for ascent).

At 0400 UTC (8PM), you can see wave-like undulations in the temperature, modest downslope on the terrain and some weak northerlies over the blowdown area at low levels.

As 1240 AM (0840 UTC), the flow had strengthened greatly aloft and a rotor was obvious in the lower atmosphere over the valley.
 The rotor strengthens over the northern side of the Quinault Valley at 0850 UTC
And at 0915 UTC (1:15 AM 27 January), all hell breaks loose with huge amplification of the wave pattern, with stronger northerlies at low level, just as they did in reality.
The amplification at this resolution (444 m) was much greater than for the coarser grids (e.g., 1.3 km or 4 km), and I suspect amplification would be far greater if we ran the simulation at 100 m or less.

But we have seen enough, I believe.  The strong winds were not from UFOs, an angry Sasquatch, a microburst from convection, or some errant meteor.

An approaching front produced just the right conditions to produce a high amplitude mountain wave on the upstream ridge, which resulted in a strong rotor that produced powerful reverse flow (northerlies).   As in the research work cited above, a very energetic subrotor was probably produced, and that resulted in a localized area of intense winds as it rotated down to the ground.

Perhaps we will try going down to higher resolution, but I have substantial confidence that the puzzle is solved.  If I were Sherlock Holmes, I would take out my violin.   But my reward, other than the satisfaction of completing a large puzzle, will be to catch up on the Olympics...or to watch one of my favorite TV shows---Air Disasters--but don't tell anyone.


Announcement:  A very interesting free lecture open to the public

The history of cloud seeding to enhance precipitation, and prospects for the future.  Professor Bart Geerts, University of Wyoming

February 15th, Kane Hall, University of Washington Campus, 7:30 PM
For information and to register go here:


  1. Beautifully presented! Great fun. Thanks!

  2. Slightly beyond my comprehension, but I appreciate the effort. I stayed at the Quinault Lodge a few days after the blowdown and was hoping to get a better explanation than aliens (or maybe not). My husband and I had a wind farm in California years ago and I know the wind does some very strange things. Thank you!

  3. We own property out in Randle, Wa and discovered a massive tree felled just like these. Roit ball still intact and buried with the break about 12 feet above the ground surface. No rot visible in the stump portion. It is an absolute mystery as to why this tree snapped at its base.

  4. Very interesting, Thanks. But aren't we talking about the same thing that brought down Galloping Gertie? Very different geometry and no harmonics in the Quinault case, but the same rotation.

  5. Impressively deduced and diagrammed. Why doesn't this happen every time the wind blows at this place?

  6. I think I've witnessed something similar to this at a much smaller scale. I have a property over in super windy Cle Elum. The property has a gentle slope up from west to east. On the east side of the property is high bank waterfront that drops about 80 feet down to the Teanaway river. The wind will blow across the property and then curl back toward the 80 foot cliff. I figured this out when someone fishing near the river dropped a cigarette butt and started a fire. I could see the phenomenon in the smoke. The fire burned for at least 45 minutes in a small area because the wind kept trapping it on the hillside and wouldn't let it spread. The fire department was able to put it out quickly. Pretty fascinating and thanks for the research!

  7. Great mystery series, Cliff! Well done!

  8. Aliens being the obvious answer and one that you still haven’t disproven!... I will accept this as an interesting if far fetched answer for the time being.

    Really enjoyed this and it was very educational... more please! This is exactly why I enjoy your site.

  9. Do the winds in the simulation generate enough force to snap the trees in this manner? I suspect someone from the forestry or engineering departments could provide you the force calculations for those trees. If those calculations match the wind speeds in your simulation, then you'd have an essential piece of corroborating evidence.

  10. Could the moderate southeasterly flow (30-40 knots) really create northerly rotor winds of 70-80 knots? My unpracticed eye can't tell whether or not that is indicated by your simulations. Is there a conceptual explanation for how this magnification of winds could occur? Thanks for the interesting post!

  11. To answer John's question, the unique part appears to be the combination of the strong inversion and strong shear right at the ridge crest as the warm/occluded front passed over. In our detailed observations in OLYMPEX in 2015-16 we never observed this shear/stability combination despite seeing several strong warm fronts move though.

    That said, power outages are common in this area so I wouldn't be surprised if this phenomena is to blame for some of them.

  12. Wow, thank you violinist Sherlock Holmes! Very thorough research and explanation.It's educational and I've never seen or heard of such a phenomenon, with large old trees snapping off like matches. The element wind is mysterious, for we can't see it.

    Here in Devon UK, in an area with high hills, I experience the motion of wind by sensing it on my face and hearing it move from a distance, towards me. It's got to deal with highs and lows of landmass. Very different from the wind in Holland, my birth-country, where most of the land is flat.

    I expect you will appreciate looking into the work of Steinbacher https://www.youtube.com/channel/UCj9L8_UjybuMiVBYDtml5hg

  13. Good explanation of the probable cause!

  14. I see 35 knots indicated at the bottom of the rotor in the last chart. So the sim doesn't seem to show enough speed for tree breakage with your assumption of 60 or so knots.
    I agree with your deductions but seeing red at the bottom of that rotor, or evidence that trees fall with less wind than conjectured, would be nice. You know, so I can let this go.
    Thanks for your thoughts and writing. I've learned much.

  15. @knainak Don't worry, we are running the model at a higher resolution and will produce a more detailed scientific analysis.

  16. A couple of things;

    1) was this what is called colloquially a "Williwaw"? A localized concentrated increase in wind, down flowing from a prominent terrain feature all unrepresentative of the prevailing free winds is exactly that although I've never heard a technical explanation for it.

    2) would a similar modelling of theoretical pressure differences through an elevation cross-section illustrate the question of how the "wind burst" can be so much greater than the norm?

  17. I am reading the autobiography of Sir Stanley Hooker who worked on so many early jet engines for British companies - largely Rolls Royce. It seems like they dealt with such problems inside the super and turbo chargers and fans in those engines too - fascinating - and not so "Elementary my dear Watson."

  18. Great job and thank you for the followup. I just fail to see how this damage was caused by the weather phenomenon you describe. I don't have the education nor the aptitude to dispute your conclusion, scientifically. Having said that, I've worked in that particular forest my entire life. I was out on the ground during the 2007 storm and spent the better part of that winter cleaning up the downfall timber it left behind. I've seen micro-bursts, I've seen hurricane damage. I've spent my entire career fighting wildfires and dealing with damaged trees the nation over. I have serious doubts as to what caused this incident. Not because your data is wrong or incomplete and not because your hypothesis doesn't hold water. But looking at the pictures and your assessed wind speeds. There is more than meets the eye.

  19. Rather than a brief one-directional 60kn wind, is it not possible those trees were initially subjected to a 35kn upslope wind that suddenly reversed direction, causing the trees to spring back and over?

  20. Microburst, tornado.... We need a term us lay folk can knowingly call this event. Leewindrotor? Sounds all wrong.

  21. Would be interested in your take on proppsed budget cuts at NOAA and their reported impact on NWS forecasting services:


  22. Nice piece of sleuthing work, Cliff! Some of the respondents above seem to miss the point of the exercise. It was not to obtain definitive proof that a specific phenomenon was to blame, but to build a strong enough case to draw a scientifically principled conclusion. I think this was done and that the conclusion in this case is a very, very convincing one.

  23. I would suggest that the wind that snapped these trees was not the only significant cause. I think saturated soils combined with a significant and rapid temperature change caused hydraulic pressure to build unevenly within the trunks causing them to be ready to fracture with much lighter winds than it would take under more normal conditions.

    Mike Mullen
    Anacortes, WA

  24. Any comment about this, Professor Mass?


  25. Good sleuthing. Sounds like the most likely solution to me, but I have a couple questions. 1) Why is the model cross section taken up the river from the lake and not right where the blowdown occurred (if I am interpreting the small maps correctly)? 2) 35 or 40 mph is the max I see on the model cross section. I understand it's just a model, and is suggestive of a rotor that could have been stronger, but it would have had to be a lot stronger. I'd say the 60-70 mph suggested as needed to take those trees down is on the low side if anything. Curious if the 100m grid shows anything different.

  26. The very cold weather forecast for next week will be devastating for the fruit industry of Central Washington State.

  27. yeah right. What a load of crap to cover up the REAL reason: a Directed Energy Weapon was used to create wind or even to destroy these magnificent trees. They are testing the weapons for use in WA like they were used in CA to create fake 'forest fires'. Washington BEWARE! Wake up people !!

  28. Perhaps there's another explanation? In the opening to the write up and fine deductive reasoning, it appeared that the cause of 'strong winds' was already in the cards.

    I find the notion that the trees were not necessarily 'knocked down' in toto, but were rather snapped off several feet above the ground. Don't you find this unusual?

    What force might cause this? Strong enough to snap the trees, but localized enough to snap them in a specific location. Wind, while being localized over the tight geographic area, would effect the tree in its entirety, no? The leaves and branches catching this unusually strong wind system and knock the tree down from its root structure...

    Perhaps there's another explanation? I wonder...are we so certain that the seismic event was the result, and not necessarily a symptom of, the underlying CAUSE?

    As Ever

  29. I'd be very interested to know what the Navy growler jets were up to that night.

    The blowdown is curiously close to one of the training sites: https://drive.google.com/file/d/1hMl_fr05o6TXazHkxijTDL3K0dfAXjuO/view?usp=sharing


    There are several environmental organizations that were vehemently opposed to the Navy getting a permit to do testing in this area, and I'm sure they would be happy to know if this testing is causing environmental damage.

    If anyone is seriously investigating this, I'd love to know where it ends up: martydogger@gmail.com

  30. Thinking about the intense subroutines you're proposing under the main rotor - wouldn't those turn the wind southerly, again? Or am I misunderstanding the dynamic?

  31. For years I have periodically seen this same blow-down evidence along the central and No. BC coastline, but never with a 1/2 mile per side perimeter. It is usually closer to 1/8 mile with the same type of directional blow-down. I also thought some sort of rotor activity was involved. What is perplexing in comparing this with your model info, is that often there is no ridge elevation seemingly high enough or localized for a lee side mountain wave. (I fully understand about mountain wave and have 2 good stories from my aviator days; but I digress!)
    Secondly, it takes more velocity to destroy those trees than what is being discussed. I go with Stihl44 on this one. Given their location, those trees have seen at least 80 knot incidents over many years, and probably more, without destruction. However, they could possibly be stressed for prevailing winds from a different direction and thus weaker for northerlies(?). Hoping for more data on actual velocity at the time of destruction. Thanks for publishing such detailed data.

  32. The lake is in a geologically active zone. An earthquake, or more likely, an avalanche on the opposite shoreline, created a rogue wave that toppled the trees where it struck. If the shoreline included a wedge or V shape, this would have amplified the height of the wave. And, naturally, it would have toppled the trees in the direction away from the lake...a far simpler explanation than all the hooey about aliens, etc. (Occam's razor) This phenomenon has been documented in remote Alaskan glacial bays when a large glacier calves. Incidentally, a published article explains how NYC is at risk from a similar threat from a geologically unstable island across the Atlantic. La Bufadora, located south of Ensenada, demonstrates the profound effect the shape of a shoreline can have on incoming waves/swells. One of the early explorers of the mouth of the Colorado river discovered this phenomenon when a large bore swept away all of his camping gear on the beach.

  33. When my child was very young I would talk with the parent of another child before school started. The other dad was a young Chief from an Olympic peninsula tribe and he once told me of the special knowledge/training he received regarding mysterious winds that happen on the coast. The tribes have been aware of these unusual events for a long time and tribal lore/beliefs have been shaped and influenced by these mysterious wind events.

  34. For Dave,it's complicated and interesting.

  35. I'm going with "it was a saquach fight". Did anyone look for large footprints?

  36. Same day, mid morning, the wind had already knocked out power here in Shelton, WA. I opened the upstairs curtains and found small chores to do in a room that points southwest. The wind was coming as usual from the southwest, across 10 acres of newly logged acreage and the newly revealed valley. But suddenly the house was being hit by branches from the east and south side. Big branches, rapid and eyeball high my second story view, coming now from the south, where there was only a few large cedars southeast of me. Then in less than a moment, as I was already looking out the window frozen and curious, the big second-growth fir trees, in my view, facing the southeast, were falling towards me. I hollered and ducked down. The trees were all uprooted, and there is good explanation... shallow root system, high water tables, fresh logging. So not that interesting, except. No one else around us lost a tree, and that powerful wind change (or swirl?), that was wild. I remember the storm that tore through (the backside of Hoquiam?), breaking off trees in a swath like a monster walked through, but this tiny strange burst was our private monster. One tree went across our shop, a couple missed the front porch by inches, another clipped another outbuilding, also missing me my perch. Same day, around the 9:30 10:00 AM time.


  37. Enough said -




Please make sure your comments are civil. Name calling and personal attacks are not appropriate.

The Weather Regimes of Summer

 Weather patterns tend to get "stuck" for extended periods and we have certainly seen such persistent conditions this summer.    W...