January 11, 2017

Major Snow Bust Around Portland

Although weather forecasts are immeasurably better today than even a decade ago, occasionally we still have major forecast failures (or "busts" as they are known in the field).  And no prediction is more difficult than lowland snow forecasts west of the Cascade crest.

Picture courtesy of Justin Sharp

On Tuesday, National Weather Service forecasts were consistent with model predictions, suggesting 1-4 inches in the Portland Metro area.    The reality, unfortunately, was very different:  as much as 12-14 inches in some locations.   The snow totals provided by the Portland Weather Service Office shows the story (see below), with a broad swath  of heavy snow from Salem, Oregon to Vancouver, Wa.  The eastern side of Portland got particularly hard hit.

The snow was associated with a low pressure system that crossed the southern Oregon coast and headed eastward VERY SLOWLY south of Portland.   To illustrate this, below is a short-term forecast map for 10 PM Wednesday, showing sea level pressure (solid lines), lower tropospheric temperatures (shading), and near surface winds (barbs).  With the low south of Portland, cold air moves into the northern Willamette Valley from the north and through the Columbia Gorge. Furthermore, there was a zone of confluent flow just north of the low (called a deformation zone), which produces a band of upward motion and precipitation.  It was a combination of cold and upward motion that produced the snow.

A radar image at 5:30 AM Wed. morning shows the modest precipitation with this low and the center of the circulation (curved radar echoes) between Salem and Portland.

Forecast models indicated snow, but nearly enough over Portland.  Here is the 24hr snow total forecast ending 4 PM Wednesday for the forecast started at 4 PM Monday. A few inches over Portland, but much more to its west and southeast.

Modest errors in the structure, position, and movement of the low pressure area and the associated deformation zone resulted in this forecast error.  

And even the ensemble forecasts (running the models many times) had difficulty with this prediction.  For example, the NWS SREF forecast initialized at 1 AM Tuesday for snow accumulation at Portland Airport (which got around 7 inches) showed a lot of uncertainty, with totals ranging from zero to 8 inches and an ensemble average of 1.5 inches. 

These ensemble forecasts indicated a possibility of heavy snow, but suggested it was improbable.

Can we do better in such a difficult event?   I believe so.  But a big issue is that the system came off the Pacific Ocean, where we have less near-surface information.   A coastal radar, such as the one recently placed on the Washington coast (the Langley Hill radar), would undoubtedly have helped the short-term forecast (0-12 hr), allowing folks more time to prepare.  Improvements in our models and how we use observations (data assimilation) could also have helped.

Folks in Oregon need to tell their political leadership that placing a powerful weather radar on the central Oregon coast is important and requires priority.  Quite honestly, it could pay for itself many times over in even one storm like this.

The Oregon Coast Needs One of These!

Northwest Weather Workshop

The Northwest Weather Workshop, the region's main gathering to discuss all aspects of Northwest weather, will take place on March 3-4, 2017 in Seattle at NOAA Sand Point.  There will be a special session of communicating forecast uncertainty during the first day.  More information on the meeting, as well as registration details, are  found at: https://www.atmos.washington.edu/pnww/

If you are interested in giving a talk at the meeting, please send me a title and short abstract by February 1.


  1. Cliff... after a series of notable forecast BUSTS this year ( the fall Puget Sound windstorm that didn't occur, the latest SE snowstorm that left Atlanta with 0" but was being hyped for weeks but then ended up surprising NY City and Southern NE with unforeseen impact, and this Portland snowstorm episode) I think forecasters need to start including the probabilities of events NOT occurring as well as ( what I assume to be) is the most probable outcomes. In that cases referenced above, given the variables needing to occur I would think the Atlanta snow and the fall Puget Sound non-wind storm would have high BUST probabilities, so perhaps not lead to post forecast animosity but this Portland snow event would still have had a low BUST probability.

  2. I was going to say much the same thing as Stefan. You do a wonderful job of communicating the probabilities for weather/number geeks, but the TV news forecasters and newspapers don't really try to do the same for their audiences. The audience expects simple yes/no answers, and the forecasters try to deliver it — to the detriment of their audience and their own reputations.

    No matter how much better our technology gets, there will always be some unknown factors, especially around here.

    Some people have trouble assimilating information that seems like "too much math" (or with a lot of technical jargon), but it wouldn't be terribly difficult to come up with some standard graphic formats (most people can read and understand a bell-curve chart, for example) and use those to illustrate the predictions *and* probabilities. Train people to ask not "will it snow?" or "how much will it snow?", but "what are the chances of heavy snow?"

  3. But this also points to the need for early and continuing K12 instruction in applied math. Some schools do this really well and offer models at the middle and high school level for (1) supplementing regular math classes with applied math labs,and (2) integrating math applications across science, social studies and some art classes.

  4. Vancouver, BC had a big bust on New Years Eve when 0-2" were forecasted and a deformation zone also set up and anywhere from 0-10" of snow fell.

  5. Agree with Stefan and Sandra completely. If the reason for the warnings is for the public safety, that is all well and good. But either underhyping or overhyping is bad as well. I think the biggest issue with the NWS is the lack of updates, or the inability to update whenever it is needed. You can check your local weather on NOAA, but they seem to only do ACTUAL updates twice a day, which is woefully unhelpful in winter. Weather is tough to predict and there are many variables to consider. Models are so sensitive that one thing just a hair off can throw the whole model off down the line. I mean winter is massive challenge. I believe most of us could accurately predict the weather from May to September in Eastern Washington. Aside from passing rain showers and the possibility of T-storms after very hot days, we all know that the temps are going to be from 70-95 in general. The sun is going to be bright and strong, with very few clouds. The winter? With all the variables taken into account, winter is a massive forecast cluster-you know what!

    Cliff: Do you think La Nina is done with, or will be in the near future? I have been following the SSTs for months in that region and it has not been much below normal at all, like right on the line, or in the neutral range.

  6. Portland forecast is for 1 to 4 inches and they get 8 to 12.

    Seattle gets 1 to 2 inch forecast and we end up with zero, 3 or 4 times in a row.

    Lame. :)

    This was by far the biggest 'bust' of a winter as far as snow goes, at least from what I saw in the eastside/renton and south areas.

  7. Cliff,

    Thanks! I agree and completely support the need for a coastal radar on the Oregon Coast. As a member of a Service Assessment Team for the Pacific Northwest Storms of December 1-3, 2007 event, I campaigned hard to include a recommendation for radar on the SW Washington Coastline. http://www.weather.gov/media/publications/assessments/pac_nw08.pdf

    The recommendation made it in (in spades), despite quite a bit of opposition by some outside the team. If you want to be in the kitchen, sometimes you have to take the heat. I'm retired now, so it's safe to talk about it. :-)

    So I say, keep campaigning. It helps to have some non NWS support (ie local DC Representative) chiming in as well. Satellite and ocean buoys are not sufficient. You need coastal radar to help assess precipitation approaching from overwater areas.

  8. Been a huge winter for Oregon though. I remember lots of years where we got little to nothing and there was quite a bit up north. Just depends on the year.

  9. It's sad that in a winter like this, where the pattern was such that cold(ish) continental air was continuously pouring into our area or poised just north of here for six weeks, that we can't even get a single decent snowfall out of it, nor a decent cold snap, where Seattle itself was in the low teens and highs in the mid-20's at best. This used to happen when I was growing up here. Now we're lucky to get into the mid-20's at night and not hit 40 in the day. That is the new normal for "cold" weather. Sigh.

  10. Could not agree more regarding the urgent need for another forecasting radar on the coast, but this is a city (along with the state leadership) that thinks nothing of spending millions of taxpayer dollars on cleaning up Portland after a week's worth of dangerous rioting (which they could have easily contained), yet can't find anything in the coffers to fill up giant potholes that have formed after years of deletrious maintenance of our roads. They don't give a rat's behind about the infrastructure until it's almost at catastrophic levels. Sadly, the near - hurricane that could've done serious damage to the region less than a few months ago did nothing to change that widespread attitude.

  11. Well put Richard. I know the ingredients must be just right to make a decent snow a reality but you are right about having the cold air in place but not the overrunning moisture to make snow a reality. I distinctly remember various occasions where rain is falling, then a wind shift to the North and the rain changes to snow. We would hear the reports...it is snowing in Everett, it is snowing in Northgate...and so on further south as the changeover takes place. With snow so infrequent here, I hate to say that I can remember most of the snow events here quite well...almost too well...from the initial snow event...then the period of being clear and cold and then the eventual warm-up where the warm Pacific air finally wins out. Always a disappointing time for a snow lover!

  12. Perhaps though the next 1-5 years may bring many more events like this years. We are extremely over due. I doubt we'll see snow again this entire winter in the metro, but our time will come quite soon.

  13. Glad to see there are others who follow snow events so closely around and share in the disappointment when we miss out! I agree with Richard and others that cold air is nice but without snow, a bit of waste! And we have had plenty of cold air... I have followed just about every snow event for about the last 40 years and this one is one of the harder ones to watch since had the low moved in just 75 miles further North we would be singing and dancing! Watching the radar Tuesday night was painful as the deformation band set up right over PDX but stayed south. I told my wife then, that this was going to be an epic event for them as the traffic cams proved it out. I know in time we will get ours, although the long range for Feb. is looking less cold with our trough moving east, at least according to the Euro Ensembles. So those who like mild in Feb. will be happy. Hey, at least it's been dry and sunny, much better than a warm rain in my book! ;)

  14. I agree about the need for a coastal radar. I remember the campaign to get one on the Washington coast, and as a good citizen I wrote to my representatives and senators. Maria Cantwell responded to my mail and then followed up later with another when it was announced we could get one. I would say a similar campaign is needed in Oregon.


New Podcast. Why do we have showers and sunbreaks during the spring? And a favorable Mother's Day forecast.

The visible satellite today is a spring-time classic, convective showers with lots of sun breaks over the Pacific and sections of western Wa...