How unusual is the current low PNW snowpack?

While Amar’s analysis of past snowfall history is interesting, it should not be construed to imply a most likely outcome of this year’s snowpack.  What it does show is that this winter is similar to winters that have been observed in the past.  If you believe the climate modelers, what has been normal or typical in the past will become very abnormal in the near future. Correlations made now with past weather events will become meaningless in predicting the future.  It is quite possible that within the next few years we will begin to notice a very different regime of temperature and snowfall in the Cascades. That change could already be happening.

[size=small](emphasis added)[/size]

Is that really what the climate modelers are saying? Not really, or at least that's a severe exaggeration of what most (scientifically peer-reviewed, published, etc.) climate modelers are saying. But I suspect you know that already, and are just being hyperbolic and alarmist (although some degree of alarm over anthropogenic climate change is probably warranted).

With apologies (since I dislike it when others do this), I've slightly reworded the last 4 sentences to give a more reasonable view that better reflects current climate scientists' consensus:
"If you believe the climate modelers, what has been normal or typical in the past will become very abnormal less typical or normal in the near future over the next century. Correlations made now with past weather events will become meaningless less reliable in predicting the future.  It is quite possible that within the next few years decades we will begin to notice a very different slightly but increasingly different regime of temperature and snowfall in the Cascades. That change could already be happening."

My main point is that, especially considering the (very limited) 17 year set of data I presented, the effects of climate change are unlikely to be significant or detectable in regards to predicting likely snowpack outcomes for this season. The effects due to smaller-time-scale variability such as PDO (period of decades), ENSO (period of years), and MJO (period of months) are much more significant for this season and the next several too than that of climate change.

Amar's post made me want to make some graphs to take a look at the trend visually.

(numbers indicate years since 1900, thus year 113 was last year).

In general, mid-december snowpack (snow water equivalent) is correlated with April snowpack.

I like it -- I was originally thinking of making a plot very similar to this and calculating linear correlation coefficients and all that, using long-term data from Paradise (since 1916), Crater Lake (since 1919), Holden Village (since 1962), perhaps a few other mountain sites where I could get long-term data (50+ years) from the National Climatic Data Center (unfortunately there are very few such sites) -- but then I just decided to take a look at the NWAC's CLISNO tables, since those sites are most relevant for skiers in this region.

Although there is an obvious correlation in the graph, it is fairly weak (i.e. the lines are fairly flat in slope, and the y-intercepts are strongly non-zero) at all except Harts Pass. Looking at the data points pretty much supports the conclusion I reached in the original post above. Taking just the Paradise data (most of the uppermost data points), April 1 snowpacks in the near-normal range (say 65-85" of SWE) are found in seasons which have December 15 snowpacks ranging from 9" to 37" -- that's a vast range of December 15 snowpacks, covering the full gamut from far-below-normal to far-above-normal, all of which produced near-normal April 1 snowpacks! (On a side note, it's worth mentioning that the Paradise SNOTEL site is at 5130 ft, significantly lower than the Paradise NWAC site and where the NPS manually records daily data, 5420 ft.) The effect is less easy to discern for the other sites shown in the graph, and the correlation is stronger than for Paradise in each case, but near-normal April 1 snowpacks still occur over a wide range of December 15 snowpack depths at those sites too.

I would interpret this as saying that often our heaviest snowfalls will come after the new year. However, once we run out of winter (March 1), we essentially will only maintain the snowpack, rather than have any more chance to build it.

Interestingly, in sites to the east of the crest, early season correlations appear to have a stronger influence on the late season.  At Hart's Pass,  even Nov 16 snowpack appears to correlate with May snowpack.  This may be because it is the coldest (receiving air from the BC interior), and highest site I considered.  This actually might also be true for Stampede Pass, which also seems to have more persistent correlations than Stevens Pass.

We definitely don't run out of winter on March 1, at least not at reasonable elevations in the Cascades. At most sites above about 4000 ft in the WA Cascades (i.e NOT Snoqualmie in particular), 5000 ft in the OR Cascades, and 6000 ft in the CA Cascades, the month which has the highest ratio of average snowfall to average precipitation is March! Not January or February! (Several years ago, I analyzed data from many dozens of sites from SW BC all the way to the southern Sierra Nevada, calculating snowfall to precipitation ratios for the season and each winter month as a function of elevation and location -- very interesting data and plots, too much to present here now.) I suspect that cooling ocean temperature in the North Pacific throughout the winter contributes to this effect, along with a reduced likelihood of Pineapple Express systems in March and later. The average seasonal snowpack maximum is around April 1 for those sites (4000+ ft WA Cascades, 5000+ ft  OR Cascades, 6000+ ft CA Cascades), even trending as late as mid-April for some higher sites and more northerly sites.

Not sure what to think about Harts Pass and other east side locations, that is interesting that early-season correlations appear to be stronger. I've spent much less time studying snowpack trends at those sites, even compared to places like southern OR and northern CA, since I personally tend to ski on the east side much less often than I do anywhere along the volcanic crest of the range all the way south to California.

What type of skis work best when skiing in graphs?

I looked around for raw data concerning basin-wide SWE measurements but none is available in an easy-to-use format. This data archive may have the raw data in there as it is used to assemble some of the NSIDC products but frankly I'm not ambitious enough to wade through it myself:
ftp://sidads.colorado.edu/DATASETS/NOAA/G02158/

While poking around I did find some fun products though that I was not previously aware of. I think a lot of folks will get a kick out of this:
www.nohrsc.noaa.gov/nsa/js_animate.html?...thwest&ts=24&large=1


I wish that was in .GIF from so I could just have it display on this page! It is SNOTEL data combined with remote sensing via airplane flyovers to recreate basin-wide snow depth analysis, then animated for your viewing pleasure. If you go to the parent site you can see other years or other basins.  Visualization of the La Nina year of 2007-2008 is fun and so is the 2008-2009 lowland snow storms.
www.nohrsc.noaa.gov/nsa/index.html?regio...nth=6&day=19&units=e

Addressing a comment Amar made and that a lot of other people notice I'm sure: with the exception that it is a pass which makes it climatologically unique, Snoqualmie pass is beginning to, and likely will slowly continue to, transition from the kinds of conditions that make snow a sure-thing with most winter systems to an area that only gets snow during favorable conditions. This is the difference between alpine and lowland systems around here, and that SNOTEL site is right at the transition elevation at this point and is very sensitive to minor fluctuations in temperature. It will be a fascinating site to watch in the decades to come because taken at face value it looks like Snoqualmie ski area could suffer the same fate as poor Mt. Pilchuck, but then it has cold eastern flow that protects what snow it does get.  As a very brief summary, back when I was investigating SNOTEL data a lot more seriously the trends and research suggested that as our climate slowly begins to warm the trend will be towards a slowly rising regional snow line, but that is hydrologically balanced by an increase in percipitation (more snow at higher areas).  The correlations were weak and my understanding is dated though, so take that statement with a grain of salt.

Finally, I'm curious Amar, do you work with this data professionally/academically or do you just have a keen personal interest? Most folks don't know of or throw out acronyms like ENSO, SST, and PDO

What type of skis work best when skiing in graphs?

I'd bring my fatties for a couple feet of fresh data.

What type of skis work best when skiing in graphs?

rax skis due to the tight quarters.
b

Amar's post made me want to make some graphs to take a look at the trend visually.

Very cool. I'm still thinking about how to best interpret your correlation plots. One thing I might suggest: set the red/green scale to span [-1,1]. The location of black as off-zero understates the apparent correlation. Do your tools permit an on-the-fly generation of a significance measure to go along with the plot? Which software are you using?

(Also, for the color-blind, there may be a superior color palette choice. I haven't thought about it in the context of this sort of plot).

Thanks for bringing more data analysis to the party .