30" of snow lost in 21hrs @ Snoqualmie?

So I'm looking at telemetry sites and I notice the dramatic totals at Snoqualmie .
It appears that 30" of snowpack have been washed away in the last 21 hours.  Do you suppose that this is correct?  The temps have been hovering around 40 with plenty of precip but still it seems a bit much to me.

I think that could be correct.
Very depressing!

With the amount of light and fluffy under hard and heavy, combined with rain... I doubt 30" of snow melted away, but I wouldn't doubt that the snowpack compressed by that much.

In a former life working at a East Coast ski area it was not the rain we feared as much as rain combined with wind.  I don't know the exact physics involved but the moist air blowing over the snow surface could destroy a slope overnight compared to just heavy rain.  From the telemetry data the wind was blowing fairly steady with some fair gusts so while the 30" seems high, settling of the snow combined with the wind driven rain could explain the loss.   Of course I really hope we did not lose that much snow!

I propose a ban on all "we need a rain event to take care of the layers" comments on TAY. After I post this message I am going to google DIY Ark plans.

Unfortunately this is correct. Much of the 30" is due to massive settlement, or densification. As of ~2200 hrs 1/7 nearly 10.50" rain has fallen at the Snoqualmie site. A full-depth pit at the study plot revealed a 30+cm layer of slush at the base of the snowpack! -note that the site is flat and inhibits good drainage.

I also think that the 30" decrease in snowdepth is almost entirely due to settlement (compaction) and not melting. It's a common misconception (or is it fear, for skiers?) that when rain falls on a snowpack, it must be melting the snow. The fact is that rain, especially when it falls within a few degrees C of the freezing point, is almost entirely incapable of melting snow.

What, you might say? No way...  But the simple reason (if you remember from science class) is the huge latent heat of fusion , which is the amount of heat (= energy) needed to melt a given quantity of snow or ice into liquid. For water, this energy for melting is roughly the same amount of energy as required to warm the same quantity of water by 80 degC, nearly to the boiling point.

So to just barely melt a kilogram of snow at the freezing point (0 degC), we'd have to put it in a pot and pour 1 kilogram (=1 liter) of 80 degC water onto it, and that water would be cooled all the way down to 0degC in the course of melting the snow. If we decided to try using 5 degC (= 41 degF) water, we would need 16 liters of it (80/5) in order to melt the same kilogram of snow. Using 1 degC (34 degF) water, we'd now need an astonishing 80 liters of it to melt the snow.

Now back to snowpack and rainfall: it's easier to think in terms of inches of snowdepth or precip instead of kilograms, so let's switch to that. How much rain fell at Snoqualmie Pass over 2 days, about 10" at close to 41 degF, right? That amount of rain would only be capable of melting 10/16 = 0.6" of snow-water equivalent (SWE). The 74" of snowpack prior to the rain probably had a density of roughly 25%, so it contained about 18" of SWE, and the rain could only have melted 0.6" of that, a fairly insignificant amount.

What about melting due to the actual warm air temperature? Again, air has a very hard time melting snow, even more so than rain because the heat capacity of air is about 4.2 times less than that of water. Therefore 4.2 times as much mass of air is needed to perform the same melting as a given mass of water. So melting a kilogram of snow using air at 5 degC (41 degF) would require 66 kilograms of it (330/5), which is a volume of about 55 cubic meters (almost 2000 cu ft), or the volume of a 12 x 20 ft room with 8 ft ceilings. And all that air would be cooled right to 0 degC in the process of melting 1 measly kilogram of snow. It's easy to see why strong winds are key to melting snow rapidly, because you've got to keep a continual fresh supply of 40 degF air coming through if you're going to melt any significant snow. With calm winds, the air right above the snow is quickly cooled to 0 degC, and it just sits there, incapable of melting any snow at all. The warmer air farther above is insulated from the snow by the denser now-cooled air, and without wind (or solar heating) the cold dense air can't move out of the way.

So two days of 40 degF temperatures probably melted even less of the snow than did 10" of rain at that temperature, but trying to calculate this is not easy. Let's just say it's lost a total of about 1" of SWE out of 18". With the current 44" depth and now 17" of SWE, that gives a density of about 40%, which makes perfect sense. That's the same as the typical springtime density of a well-consolidated maritime snowpack. Instead of having an unusually light-and-fluffy snowpack as we did before this storm, it's now settled to more typical Pacific Northwest density. But very little of it, only a few %, has actually been lost due to melting.

But how can we verify that any of these calculations are even close to correct? We need some way of knowing the actual SWE in the snowpack. Thankfully, even though NWAC doesn't use them, the NRCS SNOTEL network has a SWE sensor at every one of its sites (it's a rubber pillow filled with antifreeze, set flush to ground level, with a pressure sensor to measure the weight of the snowpack sitting on it). We just need to find a nearby SNOTEL site at a similar elevation, about 3000 ft: let's pick Cougar Mountain SNOTEL at 3200 ft (NOT the same as the Cougar Mountain near Issaquah), which has decent data (no big gaps or missing values like most other nearby sites) and this map shows how close it is to Snoqualmie Pass. I grabbed the last 7 days of hourly data and plotted it:



The green line shows that about 10" of precip have fallen the last 2 days, with temps (cyan line) around 40-43 degF. The blue line shows the rapidly decreasing snowdepth over the past 2 days, while the black line is the water content of the snow (SWE). Even though the depth has decreased from 50" to 37", the SWE has barely decreased at all, dropping only from about 17" down to 16". Meanwhile, the density (violet line) at this site has increased from 31% to 44% over the same time. Which illustrates numbers comparable to those calculated above, verifying that the basic point of this post is true: The snowpack at 3000 ft in the Central Cascades has settled during this Pineapple Express, but not melted in any significant way. Despite all the rain, only a few percent of the pre-existing snowpack has melted, well under 10% loss even at this fairly low elevation.

(Note that this SNOTEL site is west of the crest and isolated from any very cold easterly pass flow like Snoqualmie Pass gets, so it makes sense that the Snoqualmie Pass site would have had a much lighter-density snowpack prior to the rain, and thus settled more during the rain. Maybe John, i.e. Stimbuck, has the actually density numbers from his pit?)

Hey Amar,

Very well explained & supported - and logged at 3am. Your inner geek should be proud!

-Dave R

...now help me reconcile the Tinkham site.

Great post, Amar!

You should change your TAY login to Ask_Mr_Science.

A rare instance (these days) of science helping to dispel the blues.

Nicely done, Amar. I'd not thought about it in those terms before, but after your third sentence, I was sold. Data made it all the more convincing. How'd you learn about the inner workings of the snow measurement equipment?

In other news:

Fascinating post, Amar!  Isn't there also the possibility of some mechanical erosion of the saturated snowpack surface by runoff at inclined sites?

I'll get some densities later today when we work through the numbers, but it looks like the Snotel is right on track.

We also utilize a snow lysimeter that allows us to measure the outflow form a sample of the snow pack. A quick glance last night looked like we received 10"+ precipitation and around 3.5" of water left the snow once drain channels were established. Not too surprising considering the dry snow we had.

I'll run through those numbers as well.

$

What about melting due to the actual warm air temperature? Again, air has a very hard time melting snow, even more so than rain because the heat capacity of air is about 4.2 times less than that of water. Therefore 4.2 times as much mass of air is needed to perform the same melting as a given mass of water. So melting a kilogram of snow using air at 5 degC (41 degF) would require 66 kilograms of it (330/5), which is a volume of about 55 cubic meters (almost 2000 cu ft), or the volume of a 12 x 20 ft room with 8 ft ceilings. And all that air would be cooled right to 0 degC in the process of melting 1 measly kilogram of snow. It's easy to see why strong winds are key to melting snow rapidly, because you've got to keep a continual fresh supply of 40 degF air coming through if you're going to melt any significant snow. With calm winds, the air right above the snow is quickly cooled to 0 degC, and it just sits there, incapable of melting any snow at all. The warmer air farther above is insulated from the snow by the denser now-cooled air, and without wind (or solar heating) the cold dense air can't move out of the way.

But what about the heat of condensation when the water vapor in moist, warm air condenses on cold snow? If condensation is possible, that should be a much bigger transfer of heat than conduction. And if I understand what the dew point signifies, if the dew point is above freezing then condensation on the snow will be occurring?

And is that an African swallow, or an European swallow? Laden or unladen? How would it grip the coconut, by the husk?

Amar's post was sure helpful.
For the area east of the pass check out Larry R's website for Sasse Ridge snowpack and weather, more graphs etc
www.larryscascaderesource.com/
Jane

And is that an African swallow, or an European swallow?  Laden or unladen?  How would it grip the coconut, by the husk?

Ha Ha, very funny. I love Monty Python.
But seriously I am in awe of Amar's post. We got some serious bad ass snow experts on this web site!

Here is a cool breakdown of the various mechanisms of snowmelt and their relative effectiveness:

www.cnrfc.noaa.gov/publications/Rain_on_snow.ppt

Amar Andalkar for the win!
Well explained, I learned something new today.
Always a bonus.

Thanks for the kind comments.  I'll try to answer some of the questions.

...now help me reconcile the Tinkham site.

That's a tough one, it appears as though the density has not increased significantly because the SWE is dropping roughly in concert with the snowdepth.  My guess is that this is a damaged or malfunctioning SWE sensor, looking at its sensor history log the Tinkham site has a recent history of SWE sensor troubles in late 2008, which they apparently did try to repair.

The SWE sensors can be very finicky and behave in seemingly strange ways at times, since the pressure of the snowpack on the pillow is not a simple function. The sensor is very slow to adjust to new loads, such as a heavy snowfall, often taking 1-2 days to reach equilibrium. It is also affected by ice layers in the snowpack, which can act as bridges spanning across the pillow and reducing the weight felt by it. Rainwater can sometimes pool above the pillow (depending on local geometry), causing a temporary increase in the reading until the water drains away off to the side. In addition, they are sometimes prone to leaks which ruin all subsequent data as the antifreeze slowly drips out.  The Tinkham pillow may be leaking??  Here's an example of a SWE sensor showing a spike, comparing data at Cougar Mountain (updated version of plot above) and nearby Meadows Pass.  There's an obvious strange spike in the Meadows Pass SWE, which is slowly going away over the course of a day. This looks like it could be caused by water pooling and then slowly draining.




Other than the weird spike, both of these sites are fairly consistent with other.

How'd you learn about the inner workings of the snow measurement equipment?

Reading, lots of reading, and being obsessed by snow and snowfall. Here's some useful links with info:

www.id.nrcs.usda.gov/snow/siteinfo/typical_snotel.html , nice photos of SNOTEL sites
www.webs.uidaho.edu/epscor/snotel/Teache...de/teachersguide.htm , click on "Selecting A SNOTEL Site to Adopt"

Fascinating post, Amar!  Isn't there also the possibility of some mechanical erosion of the saturated snowpack surface by runoff at inclined sites?

Don't know. Clearly we see big runnels in the snow surface after major rain-on-snow events, some of which may be caused by erosion, but probably more so by drainage channels within the snowpack. I would guess that surface erosion is only significant when rain falls on a very hard firn or ice surface (such as bare glacial ice), since a seasonal snowpack is usually too porous for water to flow along the surface and erode it. Small streams are often seen to be eroding into the surface of most glaciers or permanent snowfields during late summer.

But what about the heat of condensation when the water vapor in moist, warm air condenses on cold snow?  If condensation is possible, that should be a much bigger transfer of heat than conduction.  And if I understand what the dew point signifies, if the dew point is above freezing then condensation on the snow will be occurring?

I'm not sure how much this contributes. The heat of vaporization (condensation) for water is almost 7 times as great as the heat of fusion, so every millimeter of water condensing onto the snow surface could potentially melt almost 7mm of SWE from the surface of the snowpack. But even 1mm of condensation is a large amount, far more than even the most dew-drenched field has on a fall morning. I don't know how much vapor condenses onto the surface of a snowpack under normal rainstorm conditions, and a quick search didn't turn up any good leads for finding the answer. Maybe someone else knows?

I don't have any numbers to put science up like Amar, but from my East Coast experience when we really watched every inch of snow melt away, fog was our biggest enemy. We might have heavy rains for days and find that our snowpack had only lost a half inch. One warm foggy day, though, and we could lose several inches. That seems to support the idea that condensation is a major contributor to SWE loss.

I don't have any numbers to put science up like Amar, but from my East Coast experience when we really watched every inch of snow melt away, fog was our biggest enemy.  We might have heavy rains for days and find that our snowpack had only lost a half inch.  One warm foggy day, though, and we could lose several inches.  That seems to support the idea that condensation is a major contributor to SWE loss.

snow sucking fog, a skier's worst nightmare...

An overview of the snow densities puts the current snowpack layers in the upper-30 to mid-40% range.

Storm total through this evening was 14.71", with 11.38" as rain.

The snow lysimeter measured 3.57" water leaving the snowpack.

A few inches new snow this evening on Snoqualmie, rather calm as the storm passes. Wicked windy over here in Ellensburg, looks like Mission took a hit from the wind.

snow sucking fog, a skier's worst nightmare...

That's gotta be a bad horror film...

Yesterday at Summit West and Alpy the snow was realy dense and heavy. Around 3PM it started snowing pretty hard, and between 3PM and 5:30 we got about two and a half inches of beautiful snow. I went home then, so I'm not sure how much snow they got in total.

A friend of mine just sent me this link www.flickr.com/photos/wsdot/sets/72157612242253091/show/ which illustrates some of the damage from this weather event.

This topic is a couple months old, however I just read the below quoted entry and it reminded me of Amar's post:

"I am therefore compelled to beleive that the snowey mountains yeald their warters slowly, being partially effected every day by the influence of the sun only, and never suddonly melted down by haisty showers of rain."

-Meriwether Lewis-     
Friday June 14th 1805, Missouri River, near present day Great Falls, Montana    {quote and spelling}

Sunny and crowded like buffalo at Alpental this morning.

I also think that the 30" decrease in snowdepth is almost entirely due to settlement (compaction) and not melting.


Years ago I mentioned this notion to some friends--most were non-believers.... Thanks for the wonderful academic explanation Amar!!! Fantastic!!!!!!!




(or is it fear, for skiers?) that when rain falls on a snowpack, it must be melting the snow. The fact is that rain, especially when it falls within a few degrees C of the freezing point, is almost entirely incapable of melting snow.

What, you might say? No way...  But the simple reason (if you remember from science class) is the huge latent heat of fusion , which is the amount of heat (= energy) needed to melt a given quantity of snow or ice into liquid. For water, this energy for melting is roughly the same amount of energy as required to warm the same quantity of water by 80 degC, nearly to the boiling point.

So to just barely melt a kilogram of snow at the freezing point (0 degC), we'd have to put it in a pot and pour 1 kilogram (=1 liter) of 80 degC water onto it, and that water would be cooled all the way down to 0degC in the course of melting the snow. If we decided to try using 5 degC (= 41 degF) water, we would need 16 liters of it (80/5) in order to melt the same kilogram of snow. Using 1 degC (34 degF) water, we'd now need an astonishing 80 liters of it to melt the snow.

Now back to snowpack and rainfall: it's easier to think in terms of inches of snowdepth or precip instead of kilograms, so let's switch to that. How much rain fell at Snoqualmie Pass over 2 days, about 10" at close to 41 degF, right? That amount of rain would only be capable of melting 10/16 = 0.6" of snow-water equivalent (SWE). The 74" of snowpack prior to the rain probably had a density of roughly 25%, so it contained about 18" of SWE, and the rain could only have melted 0.6" of that, a fairly insignificant amount.

What about melting due to the actual warm air temperature? Again, air has a very hard time melting snow, even more so than rain because the heat capacity of air is about 4.2 times less than that of water. Therefore 4.2 times as much mass of air is needed to perform the same melting as a given mass of water. So melting a kilogram of snow using air at 5 degC (41 degF) would require 66 kilograms of it (330/5), which is a volume of about 55 cubic meters (almost 2000 cu ft), or the volume of a 12 x 20 ft room with 8 ft ceilings. And all that air would be cooled right to 0 degC in the process of melting 1 measly kilogram of snow. It's easy to see why strong winds are key to melting snow rapidly, because you've got to keep a continual fresh supply of 40 degF air coming through if you're going to melt any significant snow. With calm winds, the air right above the snow is quickly cooled to 0 degC, and it just sits there, incapable of melting any snow at all. The warmer air farther above is insulated from the snow by the denser now-cooled air, and without wind (or solar heating) the cold dense air can't move out of the way.

So two days of 40 degF temperatures probably melted even less of the snow than did 10" of rain at that temperature, but trying to calculate this is not easy. Let's just say it's lost a total of about 1" of SWE out of 18". With the current 44" depth and now 17" of SWE, that gives a density of about 40%, which makes perfect sense. That's the same as the typical springtime density of a well-consolidated maritime snowpack. Instead of having an unusually light-and-fluffy snowpack as we did before this storm, it's now settled to more typical Pacific Northwest density. But very little of it, only a few %, has actually been lost due to melting.

But how can we verify that any of these calculations are even close to correct? We need some way of knowing the actual SWE in the snowpack. Thankfully, even though NWAC doesn't use them, the NRCS SNOTEL network has a SWE sensor at every one of its sites (it's a rubber pillow filled with antifreeze, set flush to ground level, with a pressure sensor to measure the weight of the snowpack sitting on it). We just need to find a nearby SNOTEL site at a similar elevation, about 3000 ft: let's pick Cougar Mountain SNOTEL at 3200 ft (NOT the same as the Cougar Mountain near Issaquah), which has decent data (no big gaps or missing values like most other nearby sites) and this map shows how close it is to Snoqualmie Pass. I grabbed the last 7 days of hourly data and plotted it:

The green line shows that about 10" of precip have fallen the last 2 days, with temps (cyan line) around 40-43 degF. The blue line shows the rapidly decreasing snowdepth over the past 2 days, while the black line is the water content of the snow (SWE). Even though the depth has decreased from 50" to 37", the SWE has barely decreased at all, dropping only from about 17" down to 16". Meanwhile, the density (violet line) at this site has increased from 31% to 44% over the same time. Which illustrates numbers comparable to those calculated above, verifying that the basic point of this post is true: The snowpack at 3000 ft in the Central Cascades has settled during this Pineapple Express, but not melted in any significant way. Despite all the rain, only a few percent of the pre-existing snowpack has melted, well under 10% loss even at this fairly low elevation.

(Note that this SNOTEL site is west of the crest and isolated from any very cold easterly pass flow like Snoqualmie Pass gets, so it makes sense that the Snoqualmie Pass site would have had a much lighter-density snowpack prior to the rain, and thus settled more during the rain. Maybe John, i.e. Stimbuck, has the actually density numbers from his pit?)

This topic is old, but I was reminded of it reading Cliff Mass's blog this morning.

It has another good explanation of the dangers snow-sucking fog pose to the snowpack compared to sun or rain.

At least there's still good surfing and climbing to be had, eh!?

Speaking of...the port was sweet this morning. Got shacked on my 1st wave. Surfing is alot more fun than touring anyhow