And the winner is....Buckinghorse! Yay!

The Verlot number was from the late 60's and needless to say stuck with me as I'm a bit of a weather junky.

Perhaps there's a glitch with Buckinghorse as well. 18 miles from Hurricane Ridge which had only half an inch. Granted Buckinghorse is closer to the windward side, but that's a huge difference. Anyone know how these snotel sites could misread like that?

Yes, I was surprised also with the Pigtail data as White Pass hadn't seen that much on NWAC data. Also, White Pass was pretty much on the southern edge of the rain train for most of the storm, so I'm proud to say I think Buckinghorse wins.
...
Note that there was a 24 hour period where Buckinghorse scored a brilliant 12.4".

As I mentioned in my previous post above, the Buckinghorse "data does look a bit fishy during the first few hours from 4am to 8am, when over 7" of precip was recorded. That's a very unusually high rainfall rate, so it may be a glitch." These can occur as follows:

The SNOTEL sites use a storage precipitation gauge (see www.wcc.nrcs.usda.gov/factpub/sect_4b.html ), basically a 12" wide steel cylinder tall enough to hold an entire year's worth of precip (over 250" in wet years at some sites like June Lake) and extend above the maximum snowdepth too (so gauges are up to 32 ft tall). There is enough antifreeze in it to keep the whole mixture from freezing even as the rain dilutes it, and a layer of oil on top to prevent evaporation. A pressure sensor at the bottom measures the depth of the fluid mixture, with a calibration curve correcting for the changing density of the mixture as rainfall dilutes it.



The top of the gauge has a ring of metal feathers surrounding it, a so-called "alter shield at the orifice to reduce wind effect on precipitation catch" ( www.wcc.nrcs.usda.gov/factpub/ads/adsintro.html ). However, unfortunately the gauge's catch is always less than 100% for snow if its windy. Neither the gauge nor the alter shield are heated at all, most SNOTEL sites are too remote with no AC power for that.

Anyway, heavy snowfall and riming can build a mass of snow and ice on this alter shield, or just inside the orifice of the gauge, eventually enough to block the entire orifice of the gauge with a plug. From www.id.nrcs.usda.gov/snow/data/geninfo/delta_explain.html :
"The plugged precipitation gage situation is fairly common for some high elevation sites that normally plug up in heavy snow years or during a single intense snow storm. Sometimes the plug may freeze inside the gage and remain frozen (i.e., inoperable) for 2 or more months until springtime. When the plug finally drops or melts in we may see 5 - 20 inches increase in one day."

So that probably is what happened at Buckinghorse, the heavy rain melted the plug (whether it was full or partial) and chunks of snow and ice fell into the gauge, suddenly raising the level by 7" in only 4-5 hours -- it is almost certain that 7" of rain did not fall in 4-5 hours.

In contrast, the NWAC sites use an entirely different type of gauge, a tipping bucket rain gauge , which has a funnel-like collector that dumps collected precip onto a small seesaw bucket. The bucket is designed to tip when exactly 0.01" of precip has funneled from the collector, which records a single blip on an electronic switch. This is why the NWAC gauges read in 0.01" increments. The whole thing is usually heated in order to melt falling snow (NWAC data will note that "precip gauge not heating" if the heater is offline for some reason). A similar plug of snow and ice can build if the heater is offline, which will then melt later at some point due to warming or the heater being fixed, causing very high erroneous precip readings during those hours.

Anyway, this type of gauge is much more accurate than the storage precip gauge. It's pretty much standard for all airports and weather offices, but it requires a lot more maintenance to keep it running, so it's not as well suited for truly remote sites. It is immune to some of the errors that affect SNOTEL precip gauges, but can have other quirks of its own too.

Buckinghorse (I wish I could tell exactly where it is but suffice it to say it is closer to the SW slope of the Olympics near the head of the Dungeness)

Buckinghorse is 18 miles due south of Hurricane Ridge. I created a Google map with locations of all SNOTEL sites in WA, using a high-resolution data file from NRCS with very accurate lat-lon coordinates which are good to a few meters (much more accurate than the vague coordinates listed on each SNOTEL site's webpage). I dumped the coords into an online program at www.batchgeo.com to produce this map: All Washington SNOTEL sites

Mouse over each site to see its name. And mouse over "Map" in the upper right corner, then switch to "Terrain" view instead of the standard map.

I made a similar page for all Oregon SNOTEL sites and all Wyoming SNOTEL sites .

For California, this map shows all SNOTEL sites , but does not include most of the snow telemetry sites in the state since those are run by California Cooperative Snow Surveys -- I may add those to the map if I can find a list of accurate coordinates for them.

Haven't done the other states yet, but I may at some point.

Amar, thanks for the data from St. Helens from 2006. I wonder about rainfall amounts in the central Cascades from the October event somewhere around that same year. (The event that washed out the west side of Mt. Maude - it's possible that that happened as long ago as 2003.) Despite a run of just a couple of thousand feet, that one cut a 25' deep trench near 6500' on Maude's west side!

The October event must have been 2003, which set Seattle's all-time daily rainfall record of 5.02" on Oct 20, 2003 ( www.seattlepi.com/local/144779_weather21.html ). The heaviest rainfall from that storm was in the Cascades north of Stevens Pass, with the flood destroying all roads near Glacier Peak (still NOT fixed 8 years later, thanks USFS). So it's probably the same event which hit Maude.

Unfortunately there are very few SNOTEL sites anywhere near Glacier Peak or Maude, the four closest got less rain than Seattle:

Miners Ridge (6110 ft, 9 miles NE of Glacier Peak): 4.1" on Oct 20, 2003, 0.3" on the day before and after
Lyman Lake (5980 ft, 11 miles NE of Glacier Peak): 4.1" on Oct 20, 2003, 0.2" on the day before and after
Stevens Pass (3950 ft, 25 miles S of Glacier Peak): 4.0" on Oct 20, 2003, 0.2" on the day before and 0.3" on the day after
Pope Ridge (3590 ft, 15 miles SE of Maude): 2.6" on Oct 20, 2003, 0.1" on the day after

Precip amounts must have been much greater farther west on Glacier Peak, but there's no way to know without data. And the Miners Ridge SNOTEL site was discontinued as of August 2010, so now there's even less data.

In my heart of hearts Buckinghorse still got 12.4" of rain in 24 hours.

Amar, thanks for the satellite map showing the location of Buckinghorse between Buckinghorse and Godkin creeks and 7 miles west of Mt. Anderson.

That would be a pretty wet area but is rain shadowed a bit by Muncaster Mountain and with westerly flow by Mt. Christie and the mountain at the head of Delabarre Creek (don't know it's name as it is not on the map I'm using.) The point is that there are places in the Olympics that get even heavier precipitation than Buckinghorse, yet Buckinghorse is already up to 123" on this (calender?) year.

From observation there seem to be two separate flow patterns that enhance precipitation in the Olympics:
1) S to SE flow in advance of a low pressure system - the Skokomish drainage benefits the most here but so would the upper Quinault and Duckabush area; hence the glaciers on Mt. Anderson. The southern slope of the eastern Olympics definitely has more snowpack than the northern edge of the eastern slope and the Skokomish is the most flood-prone river in Washington.
2) The more obvious SW to W flow that enhances precipitation up the S fork of the Hoh (Mt. Olympus) and also up the Queets and Quinault. The Queets is noteworthy not only for the glaciers on Olympus but also for the glaciers that obviously existed near Dodwell-Rixon Pass, which despite being only around 5000' surely shows evidence of significant glaciation as recently as the 1800's. The channeling these river valleys provide must be responsible for a good deal of precipitation in those areas. That channeling would also enhance precipitation on Mt. Anderson.

Anyway, if Buckinghorse has 123" of precipitation already, Olympus must be having one heck of a winter.

There is a computer modeled precipitation map of Washington that was created 20 years or more ago I believe by Pam Speers (Pam Speers-Hayes) as part of her graduate school work. If anyone knows whether this map is available online it would be cool to see it posted.




Regarding October 20, 2003: The precipitation doesn't seem like enough to cause that amount of damage but it must have been very intense.

Thanks again.