Eastern Snow Density, Snowfall Measuring Methods

[Tony Crocker]

As we have discussed before, simple measurement of snow water content (total snow divided by total water) in the East are contaminated by rain events. It is necessary to utilize daily records to determine incidence of snow density. Fred Lavenberg has provided me with at least 25 years of records from the Mansfield Stake, for which I have used December-March data. I have compared these to November-April data from 22 years of Mammoth ski patrol and also the last 6 years posted online from Alta.

The data indicates that Northern Vermont snow density is by far the most volatile. The most frequent events have snow density lower than the coastal but higher than the intermountain regions of the West. The super light snow events of <5% density are at the same frequency as Alta (7%). But as we know on the other end 14% of December-March precipitation is of the pure liquid variety, with another 10% mixed.

 

[J.Spin]

That’s a really informative analysis Tony. It’s far more useful than simply total snow divided by total water, and it’s actually a necessity for ski locations that receive any sort of liquid precipitation if one wants to get a sense for typical snow densities.

I do have one very large concern about the validity of the Mt. Mansfield data though. If the snowfall numbers used to calculate snow density are derived from the evening snow accumulations that get reported from the stake, then the Mt. Mansfield data are not likely to be comparable to the other two locations, nor are they a very good representation of what skiers are going to encounter on the slopes at the start of the ski day. If the analysis uses the stake snowfall numbers, and those numbers accumulate to the 230 annual inches you have posted on your website, then we know that the snowfall used in the analysis is a far cry from reality. This type of calculation would inflate the resulting water content numbers substantially. Using Stowe’s annual snowfall number of 333 inches, and assuming that the 10% H2O peak number shown in your data is the mean of a roughly normal distribution, that would center Stowe’s distribution at ~6.9% H2O, or presumably the 7% H2O bracket in your table. That value actually makes much more sense than the current peak at 10% H2O.

To add some practical weather perspective on the issue, the rule of thumb used by forecasters to estimate snow totals from synoptic storm precipitation around here in the valleys is 10 or 12 inches of snow to 1 inch of water. That means average snow in the 8.3% H2O to 10.0% H2O range. That’s for the valleys, and as far as I know, that’s the rule of thumb they even use down on the coastal plain in southern New England. So, if 8.3% H2O to 10.0% H2O is the average snow density for those types of locations, even they (assuming a relatively normal distribution) would have a peak of frequency at, or assuming an average of 9% H2O snow, below the peak of frequency shown on your table for Mt. Mansfield. Since most valley locations, especially those well away from mountains, lack both the colder temperatures of the mountains, and the mesoscale upslope/orographic events and upslope/orographic portions of storm cycles that often produce extremely dry snow, then the average snow density in places such as the higher elevations of the Northern Greens should be much lower than it is for the typical valley locations in New England.

I’ll provide two additional arguments to suggest that the actual average snowfall density for the snow falling on the upper elevations of Mt. Mansfield (especially when the analysis is focused on just December through March) is even lower than the ~7% H2O number calculated above.

1) While we know that the annual snowfall number that results from summing the snow totals from the Mansfield stake is depressed due to the collection method, the snowfall deficit is likely not depressed equivalently across all storms. The 100+-inch deficit in snowfall is most likely derived disproportionally from the storms with the driest snow, the upslope/lake-effect style snow that settles very quickly after falling. Denser snow (such as we’ve seen in some of the storms we’ve had this spring) does a lot less settling during the day than fluffier snow. With the snowfall deficit being derived disproportionately from the driest snow, most of the correction in snowfall numbers must be made there, resulting in most of the drop in snow density being obtained there. That is going to push the mean/mode for snow density further downward.

2) This second modification will depend on how the 230-inch annual snowfall number that you have on your site was derived, but if it is for the entirety of November through April, then the differential between the stake numbers and the actual snowfall is going to be a bit more than 103 inches. This is because Stowe stops recording snowfall when they close around the middle of April. The additional snow contributing to the increased differential is probably not going to be super dry most of the time since it is the second half of April, so the differential between what was recorded at the stake and what would have been recorded by the resort is not going to be as huge, but if anything it will push the snow density number still lower.

With regard to correction number 1 above, I guess the perfect way to do it would be to devise an equation that would gradually factor in the 100+-inch loss of snowfall due to settling in a linear (or perhaps non linear if that is how settling behaves) manner with respect to snow density. I’m not sure if there are any empirical data on the rate of snow settling with respect to density, so barring that, one could simply choose a cutoff value, a point at which snow is dense enough that it is assumed to settle minimally after falling, and distribute the 100 inches of missing Mt. Mansfield snowfall proportionately to all the categories in the table below that value. In people’s experience, what density of snow marks the point at which settling becomes negligible? I’m going to throw out a value of ~12% H2O based on experience, but I’d be interested in hearing what other people think because it’s not something I think about too often in my snow analyses.

So, based on the above analyses, in Tony’s table format I would put the peak for Mt. Mansfield’s snow somewhere below 7% H2O, presumably in the 6% H2O to 7% H2O range with the additional two corrections made above, although it’s hard to say just how dramatic correction #1 will be. Looking at where the peak of Alta’s snowfall density is located in the table, I would say that Mt. Mansfield’s peak would be in a similar location, but probably a bit denser. I’d also argue that if in the initial analysis Mt. Mansfield and Alta were running around the same frequency for super light snow events of <5% density, then Mt. Mansfield actually receives a higher frequency of such events compared to Alta with the refined analysis, especially since those super light snow data should be corrected to the greatest degree in the Mt. Mansfield analysis. That would fit in well with the volatility aspect that Tony Mentioned for Mt. Mansfield’s distribution of snow density.

With the greater snow density, I suspect it’s not quite as critical at Mammoth as it is on Mt. Mansfield and at Alta, but how are the snowfall data at the ski areas collected? Are 24-hour snowfall totals obtained from a single snow and liquid collection in the morning, or are they summed from two (or three of four) collections during the day? Obviously the analysis will be best (apples to apples) if the three ski areas collect their snowfall data in the same manner, because there can be substantial settling effects with regard to the drier snow. I’m sure we can find out from Powderfreak exactly how Stowe’s data are collected.

-J

 

[Tony Crocker]

I should have reminded readers (as I now do on my website http://bestsnow.net/eastnet.htm) that the Mansfield Stake data is collected at 4PM vs. the very early morning nearly everywhere else. Given my estimate of Jay at 336, I would guess that the comparable number at Stowe is close to but unlikely any more than 300. But JSpin's point is well taken that the 4PM measurements would overstate density relative to areas that measure in the morning. I'm not sure if I agree about a big density difference in density between valley and mountain snow. Once you take the rain factor out I think it boils down to mainly a quantity of precipitation difference from orographic uplift. This is perhaps my western bias. Mammoth's altitude confers many advantages over Tahoe and the PNW, but snow density is not one of them. The 1979-1995 Westwide Network data showed Alpine Meadows and Stevens Pass having similar 13% water content numbers as Mammoth.

Here's the chart again with Mansfield daily densities multiplied by 80%

While this may be overall more accurate, I suspect the "4PM measuring bias" may affect some storms/densities more than others.

 

[Harvey]

Guys... I'll be honest I don't totally get all this, but it is very cool.

Here's a dumb question. Density is higher (more settling?) at 4PM because .... most snow falls overnight? The impact of the sun?

Also are you quoting an average density for each area? If so how is that figured?

 

[J.Spin]

Here's a dumb question. Density is higher (more settling?) at 4PM because .... most snow falls overnight? The impact of the sun?

It's not a dumb question, it might be the core of what makes the snowfall numbers obtained from the Mt. Mansfield stake so low (just 230 inches of snowfall are reported at the stake annually according to Tony's website, when Stowe itself reports 333 inches at a similar elevation, Jay Peak to the north reports 355 inches, Bolton to the south reports 312 inches from a lower elevation, and even Smuggler's Notch to the northeast/east reports 288 inches). I don't believe there's any real relationship around here with regard to snow falling overnight vs. during the day, the low snowfall accumulations from the stake are thought to arise from the fact that the snow sits there during the warmest part of each day before it gets measured, vs. being measured in the early morning before whatever sun/warming hits. The only other thing I can think of that would produce the low numbers (vs. other areas in which snow is collected at 24-hour intervals) would be if the accumulations are simply taken as the difference observed in the snowpack instead of the accumulation on a clean, hard, non-compressible surface like a board. Unless there is a very solid crust on the snow, measuring accumulations as simply the increase in snowpack each day would also result in low numbers because the underlying snow is compressed as the new snow falls on top of it. Looking at the Mt. Mansfield data from each storm, it sometimes seems as though that could be the case, but at other times it doesn't. I wonder if Powderfreak has ever watched how the stake guys do their analyses. I guess there's always the chance that snowfall for the stake report is just collected in a spot that is somehow sheltered and doesn't receive much snow.

-J

 

[Tony Crocker]

the low snowfall accumulations from the stake are thought to arise from the fact that the snow sits there during the warmest part of each day before it gets measured, vs. being measured in the early morning before whatever sun/warming hits.

That has been my assumption. Yes, it would be good for powderfreak to get up there and find out some details and whether our assumptions are correct.

Also are you quoting an average density for each area? If so how is that figured?

That is usually done by taking total water divided by total snow using seasonal or monthly data, as I did for my original 1995 Powder magazine feature: http://bestsnow.net/snoqlnet.htm. Rain is so pervasive in eastern winter monthly weather data that in order to get at the question of density of just snow you have to analyze daily data.

I have never asked for daily data but occasionally it has been provided to me. I just stored it even though I had no immediate use for it. 3 years ago I realized the daily data would be useful to get at the frequency of powder days. The eastern density issue intrigued me after seeing much pictorial evidence from powderfreak, JSpin etc. that some of that snow is pretty light and dry. As with many statistical questions, this is one where averages provide inadequate information and you need to see the whole distribution to understand what's happening.

 

[Harvey]

What Jspin said makes TOTAL sense to me. The snow stake is an annual total.

The snow keeps piling up ON TOP of other snow, not a board like the one J has in his yard.

Which I assume you sweep off every time you measure.

So unless it's boilerplate - you get (Seinfeldian?) SHRINKAGE. :-D

Does it matter that the guys measuring the snow stake probably aren't in marketing?

EDIT: Jspin I tried to feed your blog on Harvey Road, but it doesn't seem to work. The message I got was that the feed was not enabled?

 

[J.Spin]

the low snowfall accumulations from the stake are thought to arise from the fact that the snow sits there during the warmest part of each day before it gets measured, vs. being measured in the early morning before whatever sun/warming hits.

That has been my assumption. Yes, it would be good for powderfreak to get up there and find out some details and whether our assumptions are correct.

I sent out an email to Powderfreak yesterday to let him know about this thread, but delivery failed, and that's the only email I've ever used in our communications. So I sent him a private message through EasternUSwx.com, and I'm hoping that will get through. I'm sure he'll be very interested in Tony's analysis, and although I believe Powderfreak is usually on the mountain in the morning, while the stake guys are on the mountain in the afternoon/evening, he may have some insight into their methods.

-J

 

[Tony Crocker]

I just sent an e-mail to my source at the Mansfield Stake, referring to this thread and following up on the questions raised here. I knew JSpin would be all over this. :lol: That's what peer review is all about.

 

[bobbutts]

Cool chart.

After the snow falls it seems to behave differently too.
One thing I noticed when living in Tahoe is how the snow would maintain a soft creamy consistency even following multiple freeze-thaws. In NE the same weather seems to produce hardpack.

How much does atmospheric pressure affect the snowpack?
How much does humidity affect the snowpack?
Is the lower pressure and humidity at higher elevations a significant factor in the softer western conditions?

 

[J.Spin]

What Jspin said makes TOTAL sense to me. The snow stake is an annual total.

The snow keeps piling up ON TOP of other snow, not a board like the one J has in his yard.

Which I assume you sweep off every time you measure.

I actually don’t know that accumulations of new snow at the stake are reported on top of the old snow, it’s just one theory to explain why the numbers might come out so low. I’ve seen quotes from Powderfreak on how even getting just a double digit (as in 10 inches) accumulation of snow at the stake is pretty infrequent. To think of the number of times during the winter season that Mt. Mansfield/Stowe gets more than 10 inches of snow during the span of an entire 24 hours, rarely getting double digit snow accumulations at the stake is very strange. Perhaps it is all settling of the snow that’s doing it, but at times it almost seems like there has to be more to it than that.

Measuring snow accumulations off of a hard surface, and starting fresh by clearing the accumulated snow after the measurement is an important concept in snow collection. One wants the accumulations to be on a level playing field so to speak, without influence from the compressibility of the snowpack below (which can vary a lot).

Snowboard clearings are typically done every 6, 12, or 24 hours, depending on how the setup is monitored. I try to measure the snow on my board and then clear it every 6 hours, but most of the time my intervals end up being 12 hours (6:00 A.M. and 6:00 P.M. clearings) because the other potential clearing times fall during the middle of the ski/work day and the middle of the night. If I’m around and awake though, I’ll try to hit those intermediate clearings, especially if there’s a big storm in progress. Some information about my setup and procedures is on my Waterbury snowfall page.

One other note on clearing and measuring off of snowboards with regard to your comment above – you can actually measure the snow whenever you want, it’s only the clearing that needs to be done at certain intervals. Typically I measure just before I clear the board, but below is an example of why one may want to measure at a different time or multiple times. The goal according to the NWS snow measurement guidelines, is to report the greatest, or total, accumulation since the last observation/clearing, and I’ve pasted in some of their text with an example below:

“Remember, you want to report the greatest accumulation since the last observation. If snowfall occurred several times during the period, and each snowfall melted either completely or in part before the next snowfall, record the total of the greatest snowdepths of each event and enter in your remarks "snowfall melted during the OBS period". For example, three separate snow squalls affect your station during your 24-hour reporting day, say 3.0, 2.2, and 1.5 inches. The snow from each event melts off before the next accumulation and no snow is on the ground at your scheduled time of observation. The total snowfall for that reporting 24-hour day is the sum of the three separate snow squalls, 6.7 inches, even though the snow depth on your board at observation time was zero. Snow often melts as it lands. If snow continually melts as it lands, and the accumulation never reaches 0.1 inches on your measuring surface, snowfall should be recorded as a trace (T) and record in your remarks that the "snow melted as it landed".”

So while 90% of the time I simply measure the snow on the snowboard, and then clear it, if I happen to be around during the middle of a collection cycle and see the way that things are melting, I’ll do my best to record that highest accumulation before it melts. During most of the season that’s not an issue, since it’s cold enough that there isn’t any melting, but a great example of the process is from the Mother’s Day storm a couple weeks back. I woke up and found 0.2 inches of snow on the board, cleared it at 6:00 A.M., and then another 0.2 inches fell. Not long after that, the warming of the day melted those 0.2 inches. But then, much heavier snowfall came in and beat out the warming of the day, leaving an additional 0.6 inches on the board (all before clearing). The end result was 1.0 inches of snow accumulation from that event at that location, even though we never had 1.0 inches of snow accumulated at any one point.

Does it matter that the guys measuring the snow stake probably aren't in marketing?

It might, in that they are presumably impartial. Personally though, now that I actually partake in observations myself and appreciate how much snow falls (especially compared to how much the casual observer thinks actually falls) combined with the fact that observations from Bolton, Stowe, Smuggler’s Notch, and Jay Peak, taken by folks like Powderfreak, nicely corroborate each other with a beautiful trend of snowfall increasing to the north, and falling off a bit to the east as topography would suggest, I find it hard to believe that everyone could somehow be in cahoots to inflate snow totals. If we can accumulate over 200 inches of snow at our house as we did in the winter of 2007-2008, then there’s little doubt that the mountains can do far better than that.

EDIT: Jspin I tried to feed your blog on Harvey Road, but it doesn't seem to work. The message I got was that the feed was not enabled?

I actually wasn’t aware that I had a blog, but fill me in and I’d be happy to try setting one up. I would assume that it’s possible to set one up at my web page?

-J

 

[Tony Crocker]

Measuring snow accumulations off of a hard surface, and starting fresh by clearing the accumulated snow after the measurement is an important concept in snow collection. One wants the accumulations to be on a level playing field so to speak, without influence from the compressibility of the snowpack below (which can vary a lot).

The Mansfield Stake snowfall is described as being measured from a "bucket." So I think it's OK in terms of being off a hard surface. Whether a bucket would produce different measurements from a board is is a different question, could depend upon size and material composition of the bucket and wind effect.

Snowboard clearings are typically done every 6, 12, or 24 hours, depending on how the setup is monitored....... The goal according to the NWS snow measurement guidelines, is to report the greatest, or total, accumulation since the last observation/clearing

In my conversations with Mt. Baker after NOAA certified their 1998-99 snowfall record I was told that NOAA accepts snowfall counts that are measured at least once but not more than twice per day. We do know the Mansfield Stake is measured only once a day, so that is a source of understatement but again we don't know by how much.

observations from Bolton, Stowe, Smuggler’s Notch, and Jay Peak, taken by folks like Powderfreak, nicely corroborate each other with a beautiful trend of snowfall increasing to the north, and falling off a bit to the east as topography would suggest

I'm inclined to agree with JSpin's comment about consistency among the ski areas, aside from my adjustment for Jay to be consistent with historical data. We know the Mansfield Stake is impartial and internally consistent over a very long time frame, but its measuring method is not the same as most ski areas, East or West.

How much does atmospheric pressure affect the snowpack?
How much does humidity affect the snowpack?
Is the lower pressure and humidity at higher elevations a significant factor in the softer western conditions?

Not sure about the first question, but the answer to the 2nd and 3rd questions is yes, a lot IMHO. We've all seen these reports, most recently from admin last weekend, about undisturbed new snow in the West "drying out." I've even observed rain crust (Island Lake Snowcat near Fernie, 2003) break up and dry out over about a 300 vertical foot range over the course of 3 days. Then there is preservation of dry winter snow through sustained drought, and on north facing steeps through warm weather. As most of you know California precipitation is volatile, yet I've seen winter snow conditions on nearly half the mountain at Mammoth after an entire month with no new snow. And I've seen that same winter snow hold up on Snowbird's steeps in March when it's been warm enough to ski in a T-shirt. By contrast my 3 March days in northern Vermont hit the first warm-up after 2+ months of below freezing weather. On day 1 ~50F at Stowe low angle groomers in the sun (Perry Merrill, etc.) were already sticky by midday. Day 2 ~60F at Mad River was similar to a typical late May to June day at Mammoth in snow surface, somewhat lucky as that's how I like it for mogul skiing. And when the temps went back into the 30's on day 3 at Jay naturally most of the snow turned to boilerplate.

The altitude/humidity are the likely reasons for East vs. West snow preservation, because as noted above the Vermont snow is actually lower in water content than the Sierra's when it's fresh. At some eastern areas you can argue that a lot of the snow being manmade contributes to the hardpack. But my recent experience at Sun Valley suggests that even manmade snow can preserve well with favorable altitude/humidity.

 

[bobbutts]

I found this chart
http://wiki.xtronics.com/index.php/Pres ... sion_Table
that says air pressure at
820' is 14.26 psi
4101' is 12.64 psi
8202' is 10.83 psi
11483' is 9.54 psi

So there's a significant difference even within the elevation range at a single resort with enough vertical.. how much that affects the snowpack, I have no clue. Since so many other factors change with elevation in the real world it would seem hard to isolate just the effects of pressure.

 

[Tony Crocker]

I think it's mostly humidity. I do notice at Mammoth in spring that wind results in more sublimation of snow into the air and less melting and refreezing on the surface. Lower air pressure could also cause more sublimation, but I wouldn't know how important a factor that might be.

 

[J.Spin]

The Mansfield Stake snowfall is described as being measured from a "bucket." So I think it's OK in terms of being off a hard surface. Whether a bucket would produce different measurements from a board is a different question, could depend upon size and material composition of the bucket and wind effect.

OK, that clears things up a lot; at least we now know that measurements are taken off a hard surface so that snowpack compressibility is not an issue. However, your concern about potential differences between snow collection on a board, and snow collection in a bucket is warranted. I have firsthand experience with that issue through simultaneous collection of snow on snowboards and in a rain gauge (bucket). I do not have the exact same type of gauge/bucket that is used on Mt. Mansfield, which, according to what Fred from the Mansfield stake crew said in the email you sent me, is that standard NOAA issue stainless steel unit, pictured below:

Those gauges are 24 inches high as Fred indicated, and have a top opening that is 8 inches in diameter. Unfortunately, in terms of direct comparison, my rain gauge is not that type, it is the standard 4-inch rain gauge used by CoCoRaHS (although also by NOAA etc.). Both gauges are described in a document from NOAA. If you look at that document, you’ll see my type of gauge referred to as an 11-inch rain gauge, which I believe refers to the height of the gauge, not the diameter of the opening. While the opening on my gauge is a bit smaller than those NOAA ones, so not a perfect comparison to how theirs might perform, once I installed it and the snow started to fly, it was quickly evident that I was not going to be using it to get accurate measurements of snowfall. I have observed at least a few issues with trying to measure snowfall in a rain gauge:

1) The first issue is the effect of wind reducing the amount of precipitation entering the gauge. If the precipitation is falling straight down, one would expect it to fall in the gauge and accumulate just as it would on other flat surfaces (barring effects of the rim of the gauge, which I will describe in issue #3 below). Now imagine some wind, and the precipitation coming in at a 45-degree angle as it falls to the opening of the gauge. This precipitation that comes in at that angle is still technically falling on the same amount of surface area created by the circular opening at the top of the gauge, so one would expect to get the same amount of precipitation into the gauge. In practice, I don’t think this is quite the case for snow due to effects of the edge of the gauge, but again I’ll talk about that in #3 below. Take the angle of the precipitation to the extreme however, and imagine a very strong wind in which the precipitation is essentially passing horizontally past the gauge; no precipitation is going to enter the gauge in that situation. There is obviously going to be a continuum between how much precipitation falls into the gauge when the precipitation falls perfectly vertically (hopefully all of it) and perfectly horizontally (essentially none of it) with an increasing loss of captured precipitation skewed toward those angles closer to horizontal. So, the stronger the wind is, the more precipitation that is going to be missed by the gauge. Now let’s focus on snow, because that seems to be where the major trouble lies in terms of wind. In terms of collecting rain, which is quite dense compared to most snow, it’s going to take quite a bit of wind to have that rain falling nearly horizontally (think hurricane etc.). But with snow, especially the dry upslope snow that we often get in parts of Northern Vermont, it doesn’t actually take much of a breeze to get that snow falling at a substantial angle. Thus, even a little wind is going to mess with the ability of one of these gauges to collect snowfall accurately, and the amount of snow caught is going to be reduced. Obviously NOAA is aware of the issues with wind and attempting to catch snowfall in these types of gauges, as you’ll see in the linked document above. They talk about it more on pages 2 and 3, but already on page 1 they state “Wind shields can be used in areas where snow constitutes 20 percent of the total catch.” The bottom line is that the more wind there is, the more precipitation that is likely to be missed with a rain gauge, and it is an especially big problem if the precipitation that you are trying to catch is snow.

2) The second issue I’ve seen is related to the first. With wind, not only is the accumulation of the snow going to be reduced, but as you’ll see in one of the images I have below, if snow is falling into the gauge under the influence of wind, the snow is not likely to accumulate evenly within the gauge, and that’s going to provide inaccurate readings of snowfall. Now, depending on how one tries to measure that potentially slanted/inconsistent accumulation of snow in the gauge, it could mean reporting more or less snow than what actually fell, but it’s certainly going to produce error.

3) Relative to my gauge, on which the walls of the cylinder are probably ¼” thick and flat on top, this third issue is probably less of a factor in the metal gauges, because the walls are thin. However, it is still a cause for concern in any collection unit with walls. The issue here is snow collecting atop the wall of the cylinder. This may seem totally insignificant, but strange things happen when you deal with upslope snow. I have seen my gauge clog after just a few inches of fluffy snow have fallen. All that is required to create an issue is for a few dendrites land on the rim of the cylinder. Then, they start grabbing onto their friends that are falling in close proximity, including those trying to fall into the cylinder near its edge. The accumulation of snowflakes on the rim of the gauge grows upward, outward, and inward, and due to the ones growing inward, before you know it the whole thing can be clogged and snow is not getting into the cylinder at all. I’d argue that NOAA may be aware of this phenomenon as well, as on page 4 of that linked document they state, “The collector’s upper edge must remain “sharp” to get an accurate catch and excess paint may round this edge.” As I mentioned above, this may not be a very big issue with the big, 8” metal gauges because they have a thin wall, but it is definitely something to watch out for if one is trying to measure snow accumulations inside containers. When it comes to dry, upslope snowfall composed of certain flakes, I’ve literally seen the snow accumulate sideways (not due to the wind, but simply due to something seeding snow accumulation and the dendrites attaching to each other). A couple of sample images are shown below. The first image is an example from the snowboard, showing how the snow is actually accumulating out into space beyond the edges of the board. For perspective on how far outward that snow accumulation grew, the depth of snow on the board is 4.3 inches, so I’d say the snow grew outward a couple of inches. Unfortunately I wasn’t calculating snow densities at that point, but just from looking at the picture you can tell that it’s very fluffy snow. The second image below is from a small metal pole that I had out for measuring snow depth. The top of the pole is a flat surface probably 1 cm or less in diameter, and all it took was a tiny bit of accumulation to seed the growth of that cotton ball on the side, and that fluffy ball of snow is a couple of inches in diameter. If that 1 cm surface of a metal rod can form an accumulation like that from just 4.3 inches of snow, imagine what could happen over the entire surface of a rain gauge. Unfortunately I don’t think I have a web picture of my gauge closed off with snow, but I’ve seen it happen to varying degrees on more than one occasion. I think it’s easy to see how it could happen, and with the dendrites closing in from all sides, it happens quite quickly. Additional pictures and details from that snowfall can be found in my report from April 5th, 2007.

4) My fourth point is simply another one of accuracy. It’s much harder to get an accurate measurement of the fallen snow if you have to reach down into one of these opaque-walled gauges with a ruler, and then try to look down on the ruler at an angle from above. Since the type of gauge I have is transparent, that actually makes it more practical, but one of these opaque gauges is certainly going to be less accurate than looking direction across at your ruler on an open, elevated surface like a snowboard. That’s why an elevated snowboard is so nice, although if you’re willing to bend over and stick your head right down to snow level, you can also get a good reading off of a ground-based snowboard. With a snowboard I can easily measure new snow accumulations to tenths of an inch, which is presumably more difficult from inside an opaque gauge. I could see reaching into the gauge and marking with your hand the depth of the snow on the ruler, then pulling the ruler out to read it, although one would still have to worry about disrupting delicate accumulations of snow to get the most accurate reading. The stake measurements of snow are done only to the nearest inch as far as I know, so this may not be a huge issue. An example of just how delicate these accumulations can be, is shown below in an image I took while measuring off my snowboard:

Now I’ll show a comparison of how snow accumulated in my gauge and on my snowboard during the same snowfall. This example occurred during the 2008-2009 snowfall season, in a snowfall event that took place from December 19th through December 21st, 2008. The weather details can be found in the link above for that snowfall event, and for those that want to see all the pictures of the snow and the resulting skiing, those are found in the December 20th, 2008 skiing trip report. On Friday evening, December 19th, I measured the first accumulation from that storm, which turned out to be 7.9 inches of 5.1% H2O snow. I happened to post pictures of the way the snow had accumulated both on the snowboard and in my rain gauge, and it’s a perfect example of how superior the snowboard is to the gauge in terms of accurate snow measurement. In the first image below, we see how the snow accumulated on the board in a fairly square shape. As Powderfreak says, when the snow stacks straight up like that (instead of forming a trapezoid like sand, dirt, rocks, and various other substances would), it’s a sign that you’ve got nice dendrites and “dry”, high-quality powder. Many types of smaller flakes, plates, needles, granules, graupel, sleet, etc. will accumulate in various trapezoidal formations like the image I linked above, but to really get snow that stacks straight up you need those fluffy dendrites as shown below in the picture of the snowboard from that evening:

As I’ve mentioned before, our house is in a very sheltered location that doesn’t get much wind, and that’s important for accumulating fluffy snow. If wind is present, accumulations can be blown away, or flakes can potentially even get smashed into other objects and broken up as they are falling. The accumulation shown on the snowboard above formed with little wind as usual, but if you look at the accumulation in the rain gauge below, you can see what even the tiniest bit of wind can do to snow accumulations in there. The rain gauge is mounted on the boy’s swing set, only about 50 feet away from the snowboard, so there should be little difference in the wind there. As a control for the wind though, you can also tell that it was very light by the way that the snow accumulated evenly on top of the bar of the swing set itself. As the snow accumulates to enough depth, the accumulations on the bar can certainly affect what goes into the gauge, but clearly the gauge did not collect a nice flat accumulation of snow, even with very minimal wind:

In the past I have also placed one of my snowboards over near the gauge to compare how it worked in that location, and confirm that I’m getting similar amounts of snow around our property, and the accumulated snow in that location was the same as I get on my main snowboard.

I’ll finish with another thought, and that is about total precipitation. If a location is trying to document the total precipitation that falls over the course of a year and a significant portion of that precipitation is snow, relying solely on a cylindrical gauge may not be the best way to go. In snow events, I have compared the amount of liquid equivalent collected by my gauge, and the amount of liquid equivalent collected on my snowboard(s) many times. In the case of the snowboard, the liquid is obtained by taking a core sample from the stack of snow, melting it down, and measuring the water that is there (this is how the snow density is calculated). On some occasions with snowfall, the gauge has caught the same amount of liquid as the board, but most of the time, the amount of liquid collected by the gauge falls well short of what was collected on the snowboard. The gauge just doesn’t collect all the snow most of the time, even with very light wind. In mixed precipitation events (or rain events) the gauge is necessary, and does a great job, but those are situations where the precipitation is denser and likely falling closer to vertically.

After the clear deficiencies I’ve seen in how my cylindrical gauge collects snow, I always use my snowboards to collect snow and get core samples to determine the liquid equivalent and snow density. I’m glad to hear that the accumulations at the stake are done on a hard surface, although it’s interesting to find out that a cylinder gauge is being used for snow collection. It’s possible that this method could be contributing to part of the deficiency in observed snowfall.

-J

 

[Tony Crocker]

Impressive analysis. I believe the long-standing question about Mansfield Stake snowfall understatement has probably been answered. I presume JSpin has sent this information to Fred.

Just a few comments:
1) You said that use of the bucket vs. a board understates water content as well as snowfall. But I get from your comments that you think it understates snowfall more. So the density measurements are probably not as far off as the snowfall measurements.
2) With regard to wind, is it not possible that gusting winds could blow previously accumulated snow off a snowboard while that snow would remain inside a container? At your Waterbury site you have been doing measurements over a long enough period of time to observe that board measurements are nearly always higher than container measurements. But you are in a very calm location, safe to say less wind than near the Mansfield Stake.
3) Another obvious question is that since the bucket is only 24 inches high, the rare larger event will be understated. I looked up Valentine's Day 2007 and to no surprise found 24 inches reported Feb. 14 with none on the prior or next day. Fred defends that number as reasonable because the snow stake depth measurement also increased by exactly 24 inches from that storm, but I have my doubts given the widespread reports of twice that much at other Vermont mountain locations, as well as 26 inches at Burlington airport. In our discussion it also came up that the stake is probably better sheltered than the bucket.

Despite caveat #2 above, and the valuable insights from so many years of consistent and impartial data, it seems likely that the Mansfield Stake measuring methodology is not in line with most weather stations. JSpin's NOAA reference specifically recommends that daily precipitation measurements be done in early morning. And it is very likely that the vast majority of high snow locations use snowboards and not containers to measure snowfall.

I would also be interested in JSpin's opinion of snowfall measurements on Mt. Washington. How could they possible get accurate measurements with the consistent winds up there? Even ski area marketing types don't try to maintain snow data records from exposed/windy locations, and it's safe to say Mt. Washington would be more difficult than anyplace with lift service.

 

[J.Spin]

Impressive analysis. I believe the long-standing question about Mansfield Stake snowfall understatement has probably been answered.

Not having worked with the 8” rain gauges myself, or seen what the specific wind patterns might be at their location (I don’t even know whether or not they have a wind shield for their setup or feel that they need one), it’s difficult to say how much their gauge-based snow measurement system is affecting the snow totals compared to what would be obtained with a snowboard. Ideally, to see if and how much the gauge method of snow collection is reducing snow totals compared to snowboard-style measurement, one would want to collect snowfall on a snowboard alongside their gauge, using the same collection time/interval for both systems. But, with the information at hand, my current hypothesis is that the discrepancy between the snow totals obtained in association with the stake, and those obtained by Stowe (and other local ski resorts) is likely due to a combination of 1) the afternoon instead of morning measurement time, and 2) snow collection in a gauge instead of on a snowboard or other hard, flat, open surface. There is also the chance that 3) the specific location of the gauge on the mountain plays a factor, but that’s hard to say without knowing what sort of drifting they typically see. If the gauge used for the snowfall measurements is up near the stake though, that should be a pretty nice spot to catch whatever snow falls on the higher elevations of Mt. Mansfield. For their analyses, Fred stated on May 27th in your email discussion that “…Stowe measures (I double checked this today) the snow fall by having the groomers go to different places on the mountain and then average the readings.” Getting measurements from multiple spots is a great way to go in terms of averaging out the effects of wind or location issues, but I don’t know much about their method beyond that. I’m sure Powderfreak could provide more details on how Stowe makes their snowfall measurements.

I presume J.Spin has sent this information to Fred.

I haven’t been in direct contact with Fred yet, since I saw that you gave him the link to this thread and hopefully he is following along here. It would be great if we could have the discussion all in one place for archival purposes.

You said that use of the bucket vs. a board understates water content as well as snowfall. But I get from your comments that you think it understates snowfall more. So the density measurements are probably not as far off as the snowfall measurements.

I’m suspect that you are referring to part of what I wrote in the penultimate paragraph of my last post: “On some occasions with snowfall, the gauge has caught the same amount of liquid as the board, but most of the time, the amount of liquid collected by the gauge falls well short of what was collected on the snowboard.”. As I look at that statement now in the context of your above comment, I can see how that might be read, but I didn’t mean to imply that there are any differences in the density obtained from the snow collected by the board vs. the density of snow obtained by the gauge. I just wanted to point out that when the gauge and snowboard are compared head to head on each snowfall, the gauge always collects less snow, and thus less liquid equivalent. I’ve always assumed that the snow density is roughly the same (since it’s the same snow falling only 50 feet away). Because of the difficulties in actually getting a nice reliable stack of snow within the gauge though, I’ve never even tried to use it to calculate snow density. I’ve just melted down the snow in the gauge to get the amount of water there, and compared it to the amount of water obtained from melting a core sample off the snowboard. On occasion, they will be the same (meaning that the gauge in that instance was actually able to capture exactly the same amount of snow as the board). That’s what one would expect if the gauge caught snow perfectly. But those instances (perhaps zero wind and snow comprised of small flakes that don’t aggregate together) seem to be pretty rare. Most of the time, the gauge just does not catch all the snow that fell.

Hopefully that clears up what I meant, but your comment made me realize something with regard to the snow density calculations you brought up. If the gauge collection method is a reason for part of the deficiency observed in the snowfall data that go along with the stake reports, that process itself shouldn’t need much correction for snow density. The gauge just catches less snow than a board or other open area, but as I discussed above, it’s all the same snow. You might catch less snow with the gauge, but if you melt it down it should contain essentially the same amount of liquid as if that amount of snow fell on a flat surface like a snowboard. There’s still density correction that needs to be done due to the snow settling that takes place through the day vs. a morning measurement. And, I don’t know if this is the case, but if the snow is falling into the cylinder at an angle due to a bit of wind, it’s possible that there could be breaking of delicate dendrites relative to what happens when the flakes land on the soft surface of the snow itself on a snowboard. I haven’t tested this because other factors make the accumulation in my gauge an issue, but I would suspect some loss in the snow loft in the gauge if this was taking place. If everything is falling straight down, the snow density in the gauge should be the same as the snow on the board as I said above, but disruption of the crystals is a possibility if they are crashing into a hard surface like the side of a metal cylinder before landing.

With regard to wind, is it not possible that gusting winds could blow previously accumulated snow off a snowboard while that snow would remain inside a container?

That’s absolutely the case with an elevated snowboard, but much less of an issue for the traditional ground-based snowboards. However, that’s one of the reasons that I use both kinds of boards as I discuss on my snowfall measurements page. The elevated board is convenient for measuring snowfall with a ruler, and can negate the effects of warm ground on the fringes of the season, while the ground-based snowboards are more protected from the wind, and when they are sitting on the existing snowpack they provide some cooling to prevent melting of the fallen snow if it gets warm. Barring extremely windy areas where everything simply gets scoured away, ground-based boards are not going to have their snow knocked off, because it is protected by all the snow that has fallen around it. With this protection of course comes burying of the snowboard, so one needs to flag the snowboards on the ground or they can be very difficult or impossible to find. A gauge/container is excellent for preventing the loss of snow due to wind, but if the spot is prone to such wind, I worry about the accuracy of the snow collection in the container to begin with. At our location, it’s rare that I ever have a stack of snow on the elevated snowboard ruined by wind, but it has happened a couple of times when just the right gust has interacted with very fragile, fluffy snow. Sometimes it will just affect a corner of the snow stack, which is no big deal. But, I often check my ground-based snowboards to confirm what I’m getting off the elevated board. I have to keep up with the ground-based snowboards anyway, as they need to be cleared and set back atop the snow, or else they will disappear.

Another obvious question is that since the bucket is only 24 inches high, the rare larger event will be understated. I looked up Valentine's Day 2007 and to no surprise found 24 inches reported Feb. 14 with none on the prior or next day. Fred defends that number as reasonable because the snow stake depth measurement also increased by exactly 24 inches from that storm, but I have my doubts given the widespread reports of twice that much at other Vermont mountain locations, as well as 26 inches at Burlington airport. In our discussion it also came up that the stake is probably better sheltered than the bucket.

I’d argue that an increase of 24 inches at the stake is a virtual guarantee that more than 24 inches of snow fell in that event. For 24 inches of snow to fall and result in a 24 inch increase in snowpack, that would have to mean that there was zero settling of the new snow and zero compression of the underlying snowpack at the stake. I wasn’t recording snow density back when we had the Valentine’s Day Storm, but it certainly wasn’t super dense snow that wasn’t going to compress. In fact, as I discuss in my report of The Valentine’s Day Storm, we could actually hear the snow settling while we shoveled. I recorded 29.2 inches of snow from the event at our house, and as you mentioned, Burlington recorded 26 inches of snow from the event and the local ski areas reported around four feet of snow. A 24-inch accumulation on Mt. Mansfield really isn’t consistent with any of those other reports, not is zero settling and zero snowpack compression at the stake very likely.

I would also be interested in J.Spin's opinion of snowfall measurements on Mt. Washington. How could they possibly get accurate measurements with the consistent winds up there? Even ski area marketing types don't try to maintain snow data records from exposed/windy locations, and it's safe to say Mt. Washington would be more difficult than anyplace with lift service.

I’ve always heard that due to the strong winds, they simply do the best they can in terms of measuring snow atop Mt. Washington, and the accuracy seemed to be somewhat in question. However, there is a video on YouTube showing one of the guys up there carrying what looks like the cylinder from an 8” rain gauge in the snow, so that must be at least one method. They also test a lot of new types of equipment for measuring snowfall there; poking around on the internet I found several pieces of text about the testing. It would be interesting to talk with them and see what their latest thoughts are in terms of the accuracy of their snowfall numbers.

-J

 

[Tony Crocker]

With regard to the original topic here of Eastern snow density as measured from the Mansfield Stake:
1) The snowfall understatement due to container vs. snowboard understates water content similarly, so density measurements are not likely distorted by this problem.
2) The 4PM measurement would provide correct water content but due to settling of snow would understate snowfall AND density vs. an early morning measurement.

I get the impression JSpin thinks the container issue is more important in terms of underestimating snowfall. So the density distribution lies between the 2 alternatives I posted but perhaps somewhat closer to the first one. I note from JSpin's last post that where the wind is a big problem on Mt. Washington, he thinks they use a container.

 

[J.Spin]

I’ve seen quotes from Powderfreak on how even getting just a double digit (as in 10 inches) accumulation of snow at the stake is pretty infrequent. To think of the number of times during the winter season that Mt. Mansfield/Stowe gets more than 10 inches of snow during the span of an entire 24 hours, rarely getting double digit snow accumulations at the stake is very strange.

I wanted to follow up on this statement, because Tony mentioned it in an email discussion with Fred Lavenberg, who is involved with the data collection at the Mt. Mansfield Stake. Although I’d never picked up on it myself, Powderfreak, who is fairly attentive to what is reported at the stake has mentioned the following:

February 24th, 2006

March 16th, 2006

At the Mt. Mansfield Stake data archives on SkiVT-L, one can pull up the daily snowfall reports associated with the stake, and here are the numbers of days with double digit snowfall indicated for each of the past four seasons since we’ve been back from Montana:

’06-’07: 1
’07-’08: 3
’08-’09: 0
’09-’10: 1

So essentially those data indicate that the upper elevations of Mt. Mansfield pick up greater than 10 inches of snow in a 24-hour period roughly once per season. We’re not even talking about a foot of snow here; these are just the number of times per season that the report hit 10 inches. I’d think for anyone that has spent much time skiing on Mt. Mansfield or has even kept up on the trip reports that come out of there, that sort of frequency seems ludicrous.

Clearly the data that are archived at SkiVT-L (at least the 2005-2006 season) aren’t perfect though for whatever reason – the two events that Powderfreak mentioned in his text above don’t seem to come up in the SkiVT-L data (those data indicate 0.00 inches for 2/24/2006 and 4.00 inches for 3/16/2006) and there is only one double digit report for the ’05-’06 season. Also, there appear to be entries in there to the hundredth of an inch, which doesn’t make sense for snowfall. I’ll have to ask Powderfreak and Wes Wright about what’s going on there, but based on Powderfreak’s comments about the frequency of double digit snowfall reported at the stake, it seems quite low compared to what people are actually finding on the mountain.

-J

 

[powderfreak]

What an awesome discussion everyone! I finally found this discussion after a PM from J.Spin on ESUWX Forums but I will chime in with some thoughts now that I've read through the thread.

First off, great information from Tony regarding the densities from the various climates. I'll get to my thoughts on snow densities later on... there are a lot of topics covered in this thread and I'd like to touch on them all.

Ski Resort Marketing and Snowfall Totals

As many know (but many also don't know) I am employed by the Mount Mansfield Company (Stowe Mtn Resort) and work in the Marketing Department as one of the two winter-time Snow Reporters. One of the fun things about being a snow reporter is that while you are technically a marketing employee, you are based out of Mountain Operations which is filled with fun but no-nonsense people who are only concerned with the day-to-day operations of the resort. Wind, temperatures, snowfall, precipitation, it is all very important to be as close to spot-on as possible with this information when dealing with mountain operations because it actually has an impact on the work that needs to be done that day (ie. how extensive do we need to groom, are snowmaking needs temporarily alleviated on some trails, etc). There is no "marketing" tweaks done to this information and the same information disseminated to Mtn Ops is posted on the Stowe website. Now, snowfall measurements on a mountain that is several miles wide with elevations ranging from 1,300ft-3,700ft is not an exact science but we are trying hard to get it as close to accurate as possible. The snowfall differences around the mountain can be striking at times and it should be noted that our measurements are done on Mansfield and not Spruce Peak which I have seen can get significantly less snow than Mansfield.

I have been skiing at Stowe, and on Mansfield, since the fall of 2003 (so I still only have 7 years of weather/snowfall experience on the hill). I bought my first seasons pass there when I moved up to Burlington to go to college at UVM...and I purchased my pass before ever skiing there; I was just told that Mansfield was the mountain to ski by some knowledgeable UVM grads. Needless to say, after my first season there I was hooked as Mansfield is unlike any other mountain in the northeast in terms of terrain/snowfall combo. I do like Jay Peak a lot, but IMHO it is not Mansfield. I am not talking about the respective resorts on these mountains, but the mountains themselves. The size and scope of the skiable terrain on Mansfield stretches over miles and can land you in 4 different towns. I also think the history of skiing on Mansfield (first ever ski patrol, CCC trails, etc) is unmatched in the east and its long history is also why there are great ski trails (backcountry and inbounds) all over the mountain. My love of the mountain is eventually what brought me in touch with Stowe Mtn Resort marketing... basically through years of posting pictures, weather forecasts, and trip reports from Mansfield all over the internet. I tell you this because although I work for the marketing department in the winter, I am a skier and weather/science guy first.

Now, I've heard some comments that ski area marketers like to inflate snowfall totals...though oddly this is only really inferred when talking about East Coast resorts (I'm not sure I've ever heard anyone question a Western mountain's snowfall). People think that ski area marketing is the reason for the discrepancy between the stake's 230" total and the resort's 330" average total (which should actually fall thanks to last season's paltry 246" total; however 31.5" was recorded at the stake after we closed which is not included in our operating season total). Stowe's listed snowfall is a 10-year average and anyone that has been paying attention knows that the 2000-2010 was overall quite favorable with over 300" most seasons, including a couple 400 inchers (2001 was over 400" and 2007 ended up right near 400"). 2008 and 2009 saw 330-350" and 2010 was 246" (by far the lowest snowfall total since I started skiing Stowe in 2003). The seasonal snowfall total is the summit snowfall total (base area is probably as much as 100" less being a full 2,000 vertical feet lower).

As far as marketing interfering with snowfall totals... that is flat out wrong. In my 2 years of being a snow reporter there has NEVER been any pressure from upper management to report more snow. Not even a hint of it. Our snowfall totals are generated by me and my colleague personally measuring snow around the mountain and that data is supplemented by members of the ski patrol and grooming departments (grooming helps largely with the 6am upper mountain readings; I personally measure only base area snowfall for the 6am report but then measure all over the mountain throughout the ski day to get an idea of accumulations). Also, members of the ski patrol have been trained by the scientists at the Mount Washington Observatory in how to measure snow in harsh, windy climates. From this training, ski patrol has set up snow study plots at various elevations, much like resorts out west, and we keep detailed daily logs in an excel spreadsheet of weather conditions and snowfall. There are two main plots that include a fixed snow stake for measuring snowpack (listed as our natural snow base depths on the stowe.com/report conditions page) and a plywood board for measuring new snow. One plot is near 3,000ft in the woods near the High Road/Perry Merrill intersection off of the Gondola. Another is down at near 1,600ft behind Barnes Camp (old building next to the parking area at the winter-time road closure of RT108) just off the Snowcat work road that goes between Spruce Peak and Mansfield.

I used to think that resorts inflated snowfall totals before moving up to northern Vermont and working at Stowe. This is obviously talked about greatly in the ski industry and there is a general assumption that the less average snowfall a mountain receives and the closer it is to a large population center, the more incentive there is for them to inflate their totals so skiers do not pass by those areas to go further north to the historically more snowy spots. Anyone that frequents places like Jay Peak, Smuggs, Stowe, Bolton, MRG, or Sugarbush knows that these places do actually get a lot of snow. Snowfall averages at these areas are on par with many inter-mountain west resorts. Everyday I compare our snowfall with Smuggs' snowfall (Stowe's neighbor on the other side of the Notch) and its generally about the same. In 2008-2009, Smuggs recorded 353" while at Stowe we recorded 349"... which I found pretty remarkable that independent parties measured nearly identical amounts of snow between the two resorts that are only a 1-2 miles apart. As J.Spin has mentioned, I really doubt there is a collaborative effort between resorts to mislead the public with regards to snowfall totals...and it is probably not by chance that snowfall averages decrease in a very organized fashion from Jay Peak south to Killington.

Lastly, there is a social element to snow reporting... the Stowe locals would crucify me if I inflated the totals. I know a ton of the local crowd that frequents that mountain and they know who is measuring, so my name is on the line. I can most certainly see how people think/assume ski areas inflate totals, but I would almost say that a lot of times we are actually quite conservative with the amount of snowfall on the hill. I got called out more than a couple of times last season for under-reporting snowfall. Even if I suspect the actual amounts might be higher than my measuring spots are showing for whatever reason, I will always go with what I actually measure.

http://list.uvm.edu/cgi-bin/wa?A2=i...4CA9396E52&Y=braaten.scott@GMAIL.COM&P=119929
then
http://list.uvm.edu/cgi-bin/wa?A2=i...4CA9396E52&Y=braaten.scott@GMAIL.COM&P=125875

Mansfield Stake Data

Ahh, the stake. So much lore and yet it begs so many questions. The stake is only useful in my mind for a comparison of snow depth from year to year and not actual snowfall (though I may be biased because I do believe the northern VT mountains, Bolton to Jay, do receive 300" per year on average). Sometimes the stake measurement is spot on, especially if a storm is just winding down or it was snowing during the daytime. A wind-less storm is usually well reported but wind rarely stops at 4,000ft in the low level jet. Other times the amount of "NEW" snow reported is just flat out wrong and not even close to what actually happened on the hill. In Mountain Operations, we frequently get a kick out of how low the "NEW" snowfall reported is when compared with what we have actually seen on the mountain. Even the groomers joke about how they just groomed the deepest "trace" anyone's ever seen. I believe local topography, wind, and collection method has a lot to do with this.

If Tony is right and they use a cylinder to measure snow up there, it would explain a lot because in high winds a bucket is going to under-catch significantly. This would make sense because sometimes you will see the actual snow depth increase MORE than what they report in the new snow column. Wind is the major factor up there and this is likely the large reason for the snowfall discrepancy between the Stake and what actually falls on Mansfield. In fact, wind might result in less total snowfall measured day-to-day because wind effects density and loft... more on this later.

Another large part of the discrepancy in snowfall is the practice of one measurement per day at a fixed time. At the very least, you should measure snowfall right when the snow actually stops falling but NWS guidelines allow for up to 4 cleared measurements per day (1 every 6 hours). When measuring at the stake, I've noticed that since they only report once a day at 4pm, they do not report snowfall if it fell and melted within that 24 hour period. If 4" falls and then is washed away by rain, you still record the 4" as snowfall. Oftentimes in marginal temperature profiles, nighttime is when you are going to be able to accumulate snow before precipitation goes back to rain during the daytime. I can remember two storms this winter when it snowed several inches of wet heavy snow overnight on the front-end of the system, before turning to rain during the daytime as warm air surged northward. One storm we had 5" at the summit at 6am which was washed away during the course of the day and the stake showed it as 1.5". The other storm we recorded 4" and the stake showed a trace at 4pm. I would also wonder about just how diligent the engineers are about getting a good snowfall measurement. If it has been snowing hard for a day straight but with high winds, I wonder just how much effort they put in to get a good measurement or if they just check the bucket and record that without thinking about it (such as the 24" on Valentines Day 2007 which seemed like a very low estimation in my mind).

The stake does have one awesome redeeming value and that is comparing snow depths between winters. However, I have found that there isn't always a direct relationship between snow depth and the amount of actual snowfall that has fallen. For example, we have gotten spoiled the past 5 winters (basically the past 10 winters, too) with largely above normal maximum snow depths as you can see in the graph below. 2005-2006 was the worst in terms of snow depth as seen by the red line.

Last season was the most odd though, as we reached 100" of settled depth about a month earlier than the historical max, even though the ski area only had recorded 200" of seasonal snowfall. That is a 2:1 snowfall to depth ratio. I like to compare snowfall to depth because it gives an idea on the amount of settling and thus the type of snow falling (works best out west where snow that falls doesn't melt or thaw out). Usually, at Stowe we have closer to a 3:1 total snowfall to depth ratio (likely higher than it should be due to rain/ice events). The reason that Mansfield hit 100" of depth with only 200" of snowfall this season was primarily because of a few big Sierra Cement storms, coupled with no real big thaws. Mid to late February we particularly got hit hard with massive dense snowfalls that totaled near 4 feet at the summit with 3-5" of liquid falling on the mountain. That storm was primarily needle flakes (flake type effects density greatly) and a moist needle at that with an isothermal column just a touch below freezing.

Snowfall Densities

I have noticed in casual observation that intermountain west ski areas usually have a 3:1 or even higher snowfall to depth ratio. I bet you any Utah or Colorado or Wyoming resort that gets to 100" of settled base has received 300"+ of snowfall to get there while the coastal mountains out west are probably closer to a 2:1 snowfall to depth ratio (ie. Kirkwood with 500" of snowfall and a 250" base, while Alta sees 500" and only has 140" on the ground). Tony's analysis makes sense to me in that our East Coast average snowfall density isn't as low as the inter-mountain west, but it also isn't as heavy as the coastal ranges.

I do think the average density of snowfall varies greatly among East Coast resorts as resorts that rely on purely synoptic snowstorms (like a nor'easter), such as the Berkshires, Poconos, Catskills (though some Lake Effect reaches them) and portions of the Whites, will often end up with a near 10% snow density. While synoptic snowfall does vary in density, it is documented that snowfall ratios in nor'easters on the east coast averages near 10-12:1 regardless of elevation. A lot of that is a function of flake type and wind (nor'easters often come with a lot of wind).

Wind is a huge factor in snowfall measurements, as well as densities, and this is often vastly over-looked. First off, think of what a good low-water content snow is made up of? Mostly air with a little bit of substance, as the big spindly arms of a dendrite crystal cause them to stack on top of each other, trapping air in the snowpack, like potato chips do in a bag. In a way, the density is basically a measure of how efficient the snowflakes are at accumulating. Wind can drastically alter snow accumulation efficiency and I'm betting that the majority of low density snowfalls occur with relatively little wind and that those flakes were formed at a very favorable snow-growth temperature (more on flake types and temperatures below).

What would be a very fluffy snowfall with 4% water-content in a calm environment can turn into a 10% water-content snowfall in a windy environment because flakes become wind-packed together with no loft. Good dendrites that would produce high ratio snowfall become fragmented and shatter from collision in a windy environment, thus losing their loft-building characteristics. This brings me back to the Mansfield Stake and the local topography on Mansfield... most of our upslope snowfall (generally a fluffy snow in mid-winter) comes from a strong NW wind flow lifting moisture over the 4,000ft ridge. Stowe Mountain Resort occupies the eastern side of the mountain, which in some respects is sheltered from the wind (while Stowe is a windy ski area, it still pales in comparison to the wind experienced up on the ridge of Mansfield). It is no surprise that the deepest snow on the mountain is found in the wind-sheltered areas because those flakes are able to build loft and thus accumulate to a deeper depth. I often find that our deepest snowfall measurements are not at the very top of the mountain, but down closer to 3,000-3,300ft in elevation and out of the wind (ie. the Rim Rock, Cliff Trail, Bypass, ND Glades region). In some situations, I've even seen calm conditions near the base area at 1,500ft allow for a greater snowfall measurement than up above 3,500ft, even though more precipitation has likely fallen at the summit...the snow has just been packed-out by the wind up there.

As I alluded to earlier, flake type is also a primary factor in snowfall densities and this has more of an effect on density than even surface temperatures in most cases. With good dendrites, you can still get a relatively fluffy snowfall at temperatures near 32F...however, it can be 5F outside and if you are seeing needles as the snowflake type, you can rest assured the density will not be much better than 10%. Needles produce absolutely no loft and thus trap no air in the snow; if dendrites are like potato chips, then needles are like grains of rice.

In order to get the best loft building dendrites, you want the maximum upward motion (lift) in the atmosphere to occur at temperatures between -12C and -18C (roughly 0F to +10F). This is what meteorologists refer to as the best snow-growth zone. After dendrites, plates are the next favorable flake type, while anything resembling a needle is significantly less efficient at accumulating. Given a wind-less environment, you can expect 20-40:1 ratios from pure dendrites and that falls to <10:1 ratios in small needles. Here's a chart that shows what flake type you can expect to form given the temperature of the atmosphere in the area of strongest lift.

I have sort of written off surface temperatures in this discussion but they, of course, have something to do with ratios...though maybe not as much as people seem to think. Temperatures in the snow growth region are much more important and surface temperatures only effect densities once they exceed 32.0F. If you are below freezing at the surface, the only things that will effect the density is wind and temperatures aloft. It is not uncommon to get a wet snow with cold surface temperatures if the temperatures aloft are warming...likewise you can get a fluffy snow with temperatures of 32-33F if temps aloft are quite cold. In synoptic storms, ratios generally increase throughout the storm...starting off dense and ending fluffy. The cause is warming temperatures aloft in the first segment (southerly wind component) of a synoptic storm will generally cause lower ratios as those snow-growth region temperatures creep up toward freezing; while on the backside of a storm system ratios generally go up as the wind shifts to the north and the snow-growth region cools back towards that -12C to -18C temp range.

I am certainly rambling now and will end this post by repeating that snow measuring is not an exact science. There are a lot of factors at play but all we can do is try our best, in the name of science, to get the most accurate measurements that we can.

I look forward to continuing this discussion and again, thanks for the great points raised in this thread by Tony and JSpin.

-Scott