Network News

X My Profile
View More Activity
The new Washington
Post Weather website
Jump to CWG's
Latest Full Forecast
Outside now? Radar, temps
and more: Weather Wall
Follow us on Twitter (@capitalweather) and become a fan on Facebook
Posted at 11:00 AM ET, 02/16/2010

Snoverkill forecast: Good overview, details messy

By Jason Samenow

Weather Checker

* Slow snow meltdown: Full Forecast | Weather Wall *
* Blizzards and climate change | Must-see photos/videos *
* News, traffic & storm coverage: Local home page | Get There *

Weather Checker is an outsider's analysis of CWG's forecast accuracy. See previous Weather Checker posts.

By Jamie Yesnowitz

There had to be significant fatigue setting in on the CWG in the wake of Snowmaggedon, when the opportunity to forecast Snoverkill presented itself. Though CWG correctly communicated the big-picture risks of the second crippling storm in less than a week, it was a hair late in sounding the alarm about Snoverkill, and some of its more detailed predictions did not materialize.

Vote in our poll, and then keep reading for more of the Weather Checker's critique of CWG's Snoverkill forecast...

The initial predictions from Brian Jackson's SLCB on Sunday morning, just after the end of Snowmaggedon, had the timing of the onset of the storm on track (Tuesday afternoon or night). The "iffy" accumulation estimates offered in this forecast (25% chance of less than 1 inch, and 85% chance of 8 inches or less) clearly did not verify. Neither did Dan Stillman's forecast Sunday night, which tweaked the accumulation numbers a little bit (15% chance of less than 2 inches and 75% chance of 8 inches or less) -- although it did highlight the potential for "significant accumulation."

Early Monday morning, Jason Samenow's forecast pinpointed a possibility that points north and east of Washington would receive the most snow, but such snow was only characterized as moderate. There was no appreciable change to accumulation amounts (15% chance of less than 1 inch and 80% chance of 8 inches or less). Jason did include a most likely amount of 3-6 inches for the immediate metropolitan area, with more amounts north and east, and less amounts south and west.

snover-kill-amounts-nws.gif
Snowfall accumulations reported by the National Weather Service for the Feb. 9-10, 2010, Snoverkill Blizzard.

By late Monday morning (about 30 hours before the storm started), CWG's forecast started moving in the right direction. Jason indicated a "best bet" of 5-10 inches in the metro area, and 8-14 inches north and east to Baltimore, with 3-6 inches for points south and southwest of Fairfax County. CWG also correctly identified the potential for "serious snow/wind" and for the first time highlighted the potential for blizzard conditions.

The 3:30 pm team forecast came in with the first map with medium confidence. The map tweaked the late-morning forecast slightly, and included a swath of 8-16 inches covering the northern tier of suburbs including practically all of Montgomery County, a little piece of Fairfax County, skirting D.C. proper, and points north and east of D.C.; 5-10 inches for other parts of the metro area; and less than 6 inches with mixed precipitation to the south and extreme east of the city.


CWG snowfall forecast issued 11 p.m. on Feb. 8, 2010.

CWG updated this map later in the evening. The updated map extended the 8-16 inch zone to the upper half of Washington, most of Arlington, and other points in nearby Northern Virginia. Importantly, CWG did not modify its accumulation after that point, unlike the National Weather Service and many media outlets whose forecasts wavered significantly.

Early Tuesday morning, CWG released its storm timeline, predicting temperatures of 26-31 degrees throughout the snowstorm, an initial period of snow Tuesday afternoon and evening (this verified), a snow /sleet line that would impact Washington and points south and east for a time on Tuesday night, and heavy snow/wind late Tuesday night into Wednesday, that would taper to flurries on Wednesday afternoon.

The temperature forecast did not verify during the second half of the storm (though it did for the first half). In the heart of the blizzard conditions that overtook the area on Wednesday, temperatures actually dropped to 20 degrees at Reagan National and Dulles airports at 1 p.m., and 21 at Baltimore-Washington International at 4 p.m. Temperatures actually moderated back toward 30 at night after the storm ended.

As for mixed precipitation impacting the District and points south and east Tuesday night, that forecast came true, although the mixed precipitation clearly moved into south-central Montgomery County (further north than forecast), hearing the classic pings from my vantage point for several hours that evening. CWG correctly predicted the precipitation would let up for a time late Tuesday night and that the snow and wind would pick up in earnest early Wednesday morning.

However, the endgame of Snoverkill was missed. Several bands of snow, enhanced by blowing snow that had previously fallen, along with gale-force gusts, raked the Washington area beyond the late-afternoon time frame forecast. The last flakes did not depart the region until 10 p.m. Wednesday night.

In looking at the reported accumulation totals from the area, while the accumulation predictions verified in many locations, particularly on the updated map (issued Monday night), there were some exceptions. BWI, contained in the 8-16 inch zone, actually received over 19 inches of snow, with portions of Baltimore, Carroll and Frederick counties receiving as much as two feet. IAD verified on both maps issued by CWG with a snowfall of over 9 inches, as did most areas south and west of the District. However, accumulations of more than 6 inches were recorded in portions of Calvert, Charles and St. Mary's counties, areas in which CWG forecasted more mixing, and less snow.

Beginning more than a day before the storm, CWG clearly captured many of the big-picture aspects of this complex storm correctly, including the overall evolution. However, they missed certain details and kind of let the storm sneak up on them, probably because they were still recovering from the last one.

It's been a tough winter season to say the least, and though the challenge of forecasting a major winter event must always be of interest to the CWG meteorologists, I can't imagine them, or many others, really wanting to see any more of the white stuff until December.

Jamie Yesnowitz has been interested in the weather since he rooted for school-closing snowstorms while growing up in Brooklyn and East Rockaway, N.Y. After graduating from Dartmouth College with a bachelor's degree in economics and government, his focus on the accuracy of weather predictions took hold when he moved to Coral Gables, Fla., to attend the University of Miami School of Law. Class was scheduled to begin on August 24, 1992. Hurricane Andrew had other ideas, however, shutting down the school for weeks. But what stuck in Jamie's mind was the final unpredicted swerve of the eye that saved those living in Miami and points north, and completely devastated areas about 20 miles south of Miami.

Undeterred by the hurricane, Jamie ultimately served as editor-in-chief of his law school newspaper, and earned both a juris doctorate and master's degree in taxation. Following law school, Jamie practiced corporate and securities law in New York before shifting to the state and local tax consulting world. Jamie moved from New York to the Washington area in 2003, and he is presently a state and local tax senior manager at a major accounting firm. Jamie lives in Potomac with his wife, Sandra, and their two daughters, Sarah and Carly.

By Jason Samenow  | February 16, 2010; 11:00 AM ET
Categories:  Snowmageddon, Weather Checker  
Save & Share:  Send E-mail   Facebook   Twitter   Digg   Yahoo Buzz   Del.icio.us   StumbleUpon   Technorati   Google Buzz   Previous: Forecast: Slow snow meltdown
Next: PM Update: Continued cold; A few Wed. flakes?

Comments

Interesting analysis. While CWG may not accurately forecast some of the details, CWG forecasting was more consistent and ultimately more creditable than most other forecasts, e.g., re: accumulation amounts, the NWS was yo-yo'ing up and down and then back up before and during the storm.

Manitoba Maulers, which this storm was at least partly, are toughies to forecast. CWG earned at least an A-. NWS gets 3 waffles.

Posted by: JerryFloyd1 | February 16, 2010 11:16 AM | Report abuse

Last Wednesday's blizzard was under-forecast by most services--though I wonder if AccuWeather picked up on it. There should have been a BLIZZARD WARNING for the entire Metro area as early as Tuesday afternoon.

As I posted about a week ago, this series of storms is the worst I've seen in Washington. On a lifetime basis, the only worse storms I can remember are the December, 1968 back-to-back thundersnow near-blizzards in Eau Claire, WI [the dates were Dec. 19, 1968 and Dec. 22/23, 1968 if memory serves me right], and the near-zero blizzard in Eau Claire aoa the night of Jan. 15, 1972. In that storm, the temperature was no warmer than the single digits to low teens, and wind chills [a 25+ mph wind with higher gusts and whiteout conditions] were well below zero. In addition, my parents and grandparents used to tell stories about the great Armistice Day Blizzard on Nov. 11/12, 1940. This may have been thje greatest snowstorm of the twentieth century in the Upper Midwest. The later "Edmund Fitzgerald" storm was more of a wind-driven rain event south of Lake Superior, as I recall.

Posted by: Bombo47jea | February 16, 2010 12:10 PM | Report abuse

I agree with the A- for the CWG--well done, guys--but I don't agree with the 3 waffles for NWS. NWS was reasonably conststent with 10-20 for the region, and that verified pretty well, I'd say. At the height of the storm on Wednesday morning, the combination of very heavy snow and strong wind was as intense I can remember seeing anywhere. I experienced a period with visibility of about 5 or 6 parked-car-lengths, which has to be less than 100 feet.

Posted by: doubtingdavid | February 16, 2010 12:13 PM | Report abuse

I don't know that I've ever read such a thorough post-storm analysis of a weather-predicting team/organization. Since every update is documented and time-stamped, it's fascinating to see the evolution of the forecasts. I would love to see the same analysis applied to each and every local weather outlet (Channel 9,7, etc) as well as the big players online (weather.com, accuweather, wunderground). I wonder how they would stack up against CWG. Iffy at best, I'm thinking.

Posted by: benhere | February 16, 2010 12:39 PM | Report abuse

CWG has definitely gotten all the snow predictions/accumulations more on the money than the local TV weathermen. Perhaps CWG should replace them??? One member to each channel!

Posted by: rwalker66 | February 16, 2010 12:42 PM | Report abuse

Fyi... A late addition to this post: A reader poll on CWG's forecast performance has been added before the jump. Please vote!

Posted by: CapitalWeatherGang | February 16, 2010 1:13 PM | Report abuse

I think they deserve an A. Also I agree with the comment above about the NWS. They did very well on the 2nd storm. They said 10-20 inches and here in central Moco we got like 16 inches or more during the 2nd half of the 2nd storm.

Posted by: celestun100 | February 16, 2010 1:59 PM | Report abuse

I agree with the positive NWS comments re: waffles. On the whole, they did very well with the storm in my opinion. They pulled the trigger early and hit it hard when many others were still gathering themselves after the first event. There may have been some number shifts that were ultimately not necessary but they have a very big audience who sometimes demands too much accuracy town by town.

Posted by: Ian-CapitalWeatherGang | February 16, 2010 2:25 PM | Report abuse

OK, the NWS gets an A, i.e. a jug of Grade A maple syrup for their waffles before and during Snowverkill.

Posted by: JerryFloyd1 | February 16, 2010 2:47 PM | Report abuse

You know what? I had something I was prepared to write, but then I realized that Bombo covered my thoughts perfectly. That was the most intense winter storm I've ever seen in DC proper. That 3rd storm separated itself from the other two in my mind, big as the others were. Many will remember how early the December storm came, or how much snow fell during Snowmaggedon, but I will always remember the howling winds and true whiteout conditions of #3. I think ALL media outlets should've been warning of a true blizzard at least a full day before; the data was right in front of us as far as what to expect.

Posted by: curtmccormick | February 16, 2010 3:01 PM | Report abuse

I thought CWG was good. When you say "late" catching up on Snoverkill, I think that perhaps just a little slower than they could have. But overall, I thought CWG was a good source of data/prediction for me. I'd go to NOAA.gov and then CWG for my data, anyday. Good work guys!

Posted by: lacoatrack | February 16, 2010 3:03 PM | Report abuse

Some points made in the review require clarification. No single probability statement can ever be said to "not verify", as for example was said of "The "iffy" accumulation estimates ....of 25% chance of less than 1 inch, and 85% chance of 8 inches or less".

Yes, it was more likely (85%) that amounts would be 8 or fewer inches, but it was also true there was a 15% chance there would be more. The only way to know whether a given probability forecast was or was not reliable is whether over a large number of cases, the amounts exceeded 8 inches 15% of the time.

CWG clearly recognized there was a quantifiable measure of uncertainty and properly communicated that information. And, it was able to discriminate between this storm and the much higher confidence in forecasts for Snowmaggedon.

An understandable source of some confusion is when the forecast snowfall map - expected distribution of single valued amounts - is viewed independently of the probabilities. A possibly more informative approach would be a map of the area that showed, for example, the distribution of probabilities of amounts greater than 8 inches.

Posted by: SteveT-CapitalWeatherGang | February 16, 2010 3:13 PM | Report abuse

i heart cwg!

Posted by: madisondc | February 16, 2010 3:15 PM | Report abuse

madisondc:

"i heart cwg!"

Please translate!

Posted by: SteveT-CapitalWeatherGang | February 16, 2010 3:23 PM | Report abuse

CWG gets an "A" in my gradebook; accuracy within what can be reasonably expected with clear explanations on what is known and what isn't. I even learned a thing or two. Good work y'all.

Posted by: tom21 | February 16, 2010 6:07 PM | Report abuse

Steve T - maybe during this lull you can write a post describing the scientific method (maybe an algorithm or two?) that CWG employs to reach its probability forecasts. The idea that your method is inherently unverifiable (you will never see a "large number of cases" for a major event, and no two events will ever be directly comparable) seems to suggest it is of little practical value. (And as noted before, it's not analgous to a "40% chance of showers.") As for communicating uncertainty, CWG already states multiple times whether it is a low confidence or medium confidence or high confidence forecast. Is that in addition to the probability forecast - so, for example, it was a "low confidence" 85% chance of 8" or less? Doesn't that seem a little silly?

Posted by: manatt | February 16, 2010 6:46 PM | Report abuse

manatt

I'll be addressing some of the great questions you raise in a post later this week. Please remind me to respond in a comment after that post if I missed something

Posted by: SteveT-CapitalWeatherGang | February 16, 2010 7:32 PM | Report abuse

I'd rather revisit the annual winter season forecasts that several of the big TV'ers always tout every fall and see how close any of them came. Predicting the past two biggies was pretty much a no-brainer, considering the usual complex weather systems we get here. Darn! I always forget to write those predictions down.

Posted by: remick1 | February 16, 2010 8:13 PM | Report abuse

I gave you a B

I think you all did a solid job recognizing the storm quite early but really didn't do as good with the storms following it.

Posted by: Bious | February 16, 2010 11:05 PM | Report abuse

In the cases of Snowmageddon and Snowverkill, I think it is fair to say that the computer models, in of themselves, scored an 'A', while the NWS and the CWG scored 'B+' and 'B-', respectively. Apparently, NWS and CWG analysts simply could not believe what they saw from the models several days in advance of the storms, based on their subjective experiences in forecasting and historical data.

Posted by: Ayrwulf | February 16, 2010 11:48 PM | Report abuse

I think CWG did a great job. Reading their blog posts and accompanying comments kept me well informed of what to expect. So I was able to buy snow-melt and groceries long before such items were in high demand.

By the way, reading CWG is turning me into a weather junkie. I found this blog last year, and all the science fascinates me. I've since tried to find other weather blogs, but there's nothing close. CWG is a local treasure.

Posted by: nlcaldwell | February 17, 2010 12:07 AM | Report abuse

Hi All,

Bob Ryan here on an Olympics late night. It's important to consider the accuracy and effective communication of the forecast for events such as this record winter, but also the long term continuity of what modern meteorology has accomplished . . .and can still accomplish.

I dare say when I was first a "public" forecaster using NAFAX and looking at the barotropic and baroclinic model outputs (only 2 that existed at the time) at 10:55 PM, trying to read the fax charts upside down and plotting obs from a Western Union teletype, I could only hope to have the tools and advances of the science we have today.

Sure, we still, and may for some time more , miss the exact very localized detailed meso-scale features and banding 12-24 hours out of the 2 big events. But as these systems evolved, and we all communicated what was likely to happen, was anyone really "surprised" by these storms? If the CWG, NWS and others miss the timing of the onset of snow by 2 hours or the total accumulation by 4 inches for a 20 inch event is this a "bust"?

One of the great resources of this forum is what we can all learn from you the users about how and why you make decisions based on what is posted here. I strongly believe that the uncertainty and probabilistic nature of predicting the future weather is important to communicate. As the event draws closer, one would hope the uncertainty decreases and confidence increases. That a 3 day "forecast" is not perfect down to everyone's neighborhood is no big deal nor possible, nor should be even "graded". Rather has the message, forecast evolution and continuity been accurately and effectively communicated so that 12-24 hours beforehand (call it critical decision time) you the user make the best weather related decision possible is the question to grade the CWG and all of us on.

I know from experience that human expectations of what is possible in forecasting will always exceed what is scientifically possible. Want an hour by hour forecast 10 days in the future? I know where you can go. Expect it to be accurate? I've also got a bridge. Stay tuned for Monday

Bob Ryan

Posted by: rtryan1 | February 18, 2010 1:18 AM | Report abuse

@Bob Ryan

We're the Capital Weather Gang and we approve your message :)

Posted by: CapitalWeatherGang | February 18, 2010 2:45 PM | Report abuse

The comments to this entry are closed.

 
 
RSS Feed
Subscribe to The Post

© 2012 The Washington Post Company