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 10:30 AM ET, 02/22/2011

Snow totals fall short of forecasts: why?

By Jason Samenow

Snow totals reported by National Weather Service.

As a snow enthusiast and someone who strives to make the most accurate forecast possible, this morning's lack of snow in the immediate metro area was a bitter disappointment. From northern Montgomery county to the south and southwest, our snow total predictions were off by about 1-3". Of course, in many of these locations 1-3" was about all the snow that was forecast meaning very little (or even no) snow. To the north and to the east, we got the snow totals mostly right, but most of us don't live there (alas, forecasting for Baltimore would've been so much easier for this one ) ...

CWG's snowfall prediction.

So what went wrong?

* We got more sleet than forecast and even some freezing rain
* We got less precipitation overall than forecast

The writing was on the wall between 9 and 11 p.m. last night when the radar showed most of the heavy action concentrated north of Montgomery county, with just light bands of precipitation to the south. I also noted that the forecast models had backed off precipitation totals, only forecasting around 0.1" liquid equivalent (1" or snow or less) after 1 a.m. That was really significant, because earlier model runs had suggested 0.2" or more after that time.

Particularly from D.C. and points south, the after 1 a.m. window was important for snow accumulation because that's when the models showed the atmosphere sufficiently cold for snow rather than sleet. When amounts during that window were cut in half, I knew our snow total forecast was in trouble.

But even after 1 a.m., sleet continued to mix in with the snow in some spots due to a stubborn warm layer aloft, further cutting back totals.

We did mention the fact that there was a realistic possibility we'd get more sleet, less snow, and less precipitation overall in our forecast yesterday - so the fact the storm underperformed our "most likely" forecast shouldn't have come as a total surprise. Consider the following three separate statements in our midday update:

How confident are you in your forecast? We have medium to high confidence there will be disruptive, accumulating ice and snow tonight, but just medium confidence in amounts. This is because there will probably be significant variability in amounts from north to south. Slight shifts in the band of the heaviest precipitation as well as the timing of the arrival of cold air and hence the changeover from rain to sleet to snow could easily change how much snow and ice different areas receive.

The timing of the sleet to snow transition is tricky because of a warm layer aloft which may favor a several hour period of sleet before the precip mixes with and changes to snow within an hour or so of midnight...(or even a little later in the southern suburbs).

The biggest bust or "fizzle out" potential is just south of town where temperatures will be mildest and more sleet than snow cold fall. But even there, temperatures should drop below freezing and slick spots are likely. Also, there's some chance the storm loses some if its steam as it heads east and the precipitation crosses the mountains - so it wouldn't be a shock if totals ended up on the low side of estimates even north of town.

Unfortunately for snow lovers, what we said could go wrong did go wrong. It's just been that kind of winter. Maybe karma for last year?

Finally, in fairness to us, we did correctly predict the evolution of temperatures during the storm, and that there would be icy accumulation on roads that would make for a slick commute pretty much everywhere. We also were right in predicting less snow than the National Weather Service.

By Jason Samenow  | February 22, 2011; 10:30 AM ET
Categories:  Latest, Recaps, Weather Checker, Winter Storms  
Save & Share:  Send E-mail   Facebook   Twitter   Digg   Yahoo Buzz   StumbleUpon   Technorati   Google Buzz   Previous: Forecast: A couple cold days follow icy storm
Next: Joe Bastardi resigns from AccuWeather


Copying a comment here that I just made in the previous post... The upper level energy needed to get good vertical motion - and thus heavy enough precip to really cool the atmosphere and give us more snow than sleet - never looked that great on the models for this storm. That's one reason why our forecast hedged lower than most, but wasn't low enough. -Dan, CWG

Posted by: CapitalWeatherGang | February 22, 2011 10:38 AM | Report abuse

I was sad that this storm was pretty much a bust in Centreville VA.
But it was fun to pretend I was a comet this morning, as my car billowed a tail of icy sleet particles behind me. Sleet is quite different from snow, and as we were reminded last night, it is much noisier.

So is that it for this winter? March is always full of surprises. Always in motion is the weather future.

Posted by: FIREDRAGON47 | February 22, 2011 10:39 AM | Report abuse


First, let me just say I'm a BIG fan of yours. But, I'm really sorry, I must admit, this was a big bust. We only got a coating to a half inch here in NE Loudoun; well below the predicted 2 - 4".

I must say, however, that you guys were quite a bit more accurate than some other forecasters such as the NWS. I will continue to follow CWG on a day-to-day basis regardless.

I will also continue to recommend CWG to my colleagues, friends, family, etc.

Thanks for all the hard work :) !


Posted by: BobMiller2 | February 22, 2011 10:46 AM | Report abuse

It didn't, Dan. If you go by the checklist to get much snow in the metro area this one never had "the look" at 500mb at least -- with everything passing north. There were some solid runs at 700mb which seemed to indicate we'd get better lift around midnight to help cool the column, but it always looked like someone was going to miss out--I guess it was us. Guidance did not do terribly well 'til late yesterday on the southern end of higher precip amounts either.

Posted by: Ian-CapitalWeatherGang | February 22, 2011 10:47 AM | Report abuse

The poll doesn't work for me... it just shows "results," but with no results in it. I tried in chrome and firefox.

Posted by: megamuphen | February 22, 2011 10:52 AM | Report abuse

Your forecasts were pretty much spot on for us here in Gaithersburg...

Posted by: dak3 | February 22, 2011 10:54 AM | Report abuse


I'm having the same issues. Taking the poll down until I can figure out the problem.

Posted by: Jason-CapitalWeatherGang | February 22, 2011 10:55 AM | Report abuse

I'd vote for a "somewhat bust." You did cover your bases, but probably not enough. Also, I do think your forecast was much more accurate than NWS, for what it's worth. And I still think CWG offers the best forecasts for the DC area. Live and learn for next time.

Posted by: jms12 | February 22, 2011 10:57 AM | Report abuse

I was disappointed- but that disappointment came last night before I went to bed at 11. I read CWG's latest update so this morning was not a surprise.

Posted by: Snowlover2 | February 22, 2011 10:58 AM | Report abuse

I have a related question. The past few weeks I have noticed that AccuWeather's forecast high for DC is averaging 5-7 degrees less than the actual high. I wonder why.

Posted by: TheChileanPresidentIsMuchBetterRespondingToDisastersThanObama | February 22, 2011 11:01 AM | Report abuse

The fact that you dedicated an entire article to discuss why the forecast was off shows how CWG is in a class by themselves!! LOL Love you guys.

Posted by: MsMacMama | February 22, 2011 11:10 AM | Report abuse


Last week there were several days when high temps went higher than models indicated. Sometimes models underdo temperatures when we have west winds (which we had a lot last week). A lot of forecasters recognize this, but psychologically it's tough to forecast 70+ temps in February so sometimes we don't go high enough.

Posted by: Jason-CapitalWeatherGang | February 22, 2011 11:13 AM | Report abuse

DCA is up to 9.9" for the year. I like round numbers. Let's get at least another .1" inch.

Posted by: JerryFloyd1 | February 22, 2011 11:14 AM | Report abuse

When I went to bed last night, there was a layer of snowy sleet out there. When I woke up, not much had changed. I was surprised but not disappointed. The kids went to school and all is back to normal. I'm ready for spring.

Posted by: veronica7 | February 22, 2011 11:31 AM | Report abuse

No one is GOD. Don't be mad that you, Tony Perkins, Sue Palka, and the rest have been completly wrong this year.

I would suggest you give a wider range, instead of being so specific. And you may even want to be hyper-specific (IE Upper/Lower Montgomery County, etc...).

Posted by: 4thFloor | February 22, 2011 11:39 AM | Report abuse

Has the 18z NAM earned anyone's respect yet?

Posted by: sgustaf1 | February 22, 2011 11:52 AM | Report abuse

do you think when all this melts we'll have flooding problems like these?

i would have been perfectly fine if yesterday's build-up hadn't happened. at first (i.e., over much of the weekend) we weren't expecting anything but rain and i had made peace with that crappy outcome. i kept touting the "back-end" possibility of an inch or two...(boy...that seems like a lot in retrospect...) but then the nam shifted... the gfs followed - they all aligned, apparently, with the euro. then we got the crystal ball, the "team forcast" (that ring-around-the-roses graphic cracks me up) and the accumulation map... ugh... it's the roller-coaster that gets me... like charlie brown, i was running toward the ball: i already had plans for those cute little 3" i was supposed to get.... dang.

next time you see people talking up a storm just hours ahead of time that you think doesn't have "the look", please let us know.

Posted by: walter-in-fallschurch | February 22, 2011 12:10 PM | Report abuse

Let's call it a bad weekend of forecasting. As late as Sunday afternoon forecasters were saying the high Monday was going to reach 67 and any precipitation would end early Monday morning. I planned a golf outing based on that forecast. Well, it was raining and cold (mid-40s) Monday afternoon. The outing was a disaster.

Posted by: ShovelPlease | February 22, 2011 12:18 PM | Report abuse

"Hedged lower than most"

LOL... CWG called for 2-4" I got 0.03 LMAO.

I dont know how many times i said it this week all over this blog, this would be a non event for DC metro.

Problem was the cold air came in faster than most of you thought and that actually hurt totals as it was cold/dry air. The trend will always rule out models and what happened last night is the trend around DC this year.

When you live by the models (model hug), you die by them along with your credibility. The trend shouldve been your friend, not stupid models!

Posted by: KRUZ | February 22, 2011 12:52 PM | Report abuse

meant to say quarter of an inch in my above post.... well off from CWG;s 2-4 inches.

Posted by: KRUZ | February 22, 2011 1:00 PM | Report abuse

A couple good names for this storm for the DC metro would be EURO fail or maybe GFS fail or my favorite, NAM fail!

The sun has officially melted all the (barely) quarter inch of snow outside. Time to put on my snow boots :/

Posted by: KRUZ | February 22, 2011 1:15 PM | Report abuse

1" sleet/snow accumulation this morning at 1800 ft east of Front Royal

wxRisk seemed to be on the right track with this event for Virginia

Posted by: spgass1 | February 22, 2011 1:34 PM | Report abuse

Can't help but believe that Prince Georges County is really second guessing their decision to close schools today.

Posted by: mbrumble | February 22, 2011 1:47 PM | Report abuse

The predicitions were actually pretty close to spot on in Baltimore and points north. We had about an inch of sleet/ice with three inches of fluffy snow on top.

Posted by: bachaney | February 22, 2011 1:48 PM | Report abuse

Points North and West always get the snow, DC doesnt in these kinds of events. Mets in DC shouldve known better.

Posted by: KRUZ | February 22, 2011 1:53 PM | Report abuse

I am officially bringing out the warm weather clothes after this one. At least I didn't buy a snow blower.....

Posted by: TBAlexandria | February 22, 2011 1:58 PM | Report abuse


You're right we busted, but you're wrong on the reasons. If cold air killed our moisture, how did places that were COLDER around Baltimore and up toward the PA border get 4-6" of snow? We would have even gotten less if it had been warmer because the disturbance would've tracked even farther north! The reasons we didn't get enough snow had to do with the fact it was just barely not COLD enough about 1 mile up AND because the upper level energy was a bit north of what was ideal-- by just about 45 miles.

Posted by: Jason-CapitalWeatherGang | February 22, 2011 2:24 PM | Report abuse


Well if im wrong you then are also calling BOB RYAN wrong, heres what he tweeted....

3 hours ago:
cold dry air really kept amounts on very low side here 4-6'' northern Maryland"

12 hours ago:
"more dry air coming in now from west doesn't look like a whole lot more for metro area 6" Baltimore north snow lovers"

So Jason, you are wrong or Bob Ryan is wrong. I think I'll side with Bob Ryan being correct here, no offense,

Posted by: KRUZ | February 22, 2011 2:30 PM | Report abuse

Did the Potomac River serve as some kind of snow blocker, at least north of DC? On the map above it seems that snow totals in Maryland, while not high, are substantially more than those along the same latitude in Virginia.

I got a taste of the differing gradients of snow on my not very long commute (mostly by Metro) today. At home in Gaithersburg we got about 2 inches; less than anticipated but still enough to completely cover the sidewalks, often in ice. By the time I got to the office in Bethesda there was barely a visible coating.

Posted by: mkarns | February 22, 2011 2:48 PM | Report abuse


I agree with Bob, the dry air hurt everyone and the models even showed that. We took that into account. But the cold dry air helped us as at the same time as it hurt us. If we hadn't had the cold dry air moving in, we wouldn't have had any frozen precipitation period. What ultimately killed us was the exact track of the upper level feature and where the heavy precip set up. The air was colder and drier (lower dewpoints) to our north, but more snow and precip overall fell there.

Posted by: Jason-CapitalWeatherGang | February 22, 2011 2:48 PM | Report abuse

Weather people always make a big deal of predicting snow, or hurricanes at the start of the season or other epic weather events. Usually, it just turns out that they are trying to attract attention and viewers/readers.
Just look at the pictures they post with their predictions. If there is a 30% chance of rain, there is a picture of rain. I guess they don't understand they're saying there's a 70% chance it won't rain. I'm glad airline pilots make fewer mistakes than weather people.

Posted by: dncandme | February 22, 2011 3:35 PM | Report abuse

No biggie, pointy-headed weather geeks, no biggie.

Posted by: FedUpInMoCo | February 22, 2011 3:46 PM | Report abuse

I live in stafford county and there was a 2 hour delay for schools but nothing else.
We had a dusting of snow but some freezing rain. I like snow so was not to happy when I woke up this morning but I heard that March is a snowy month for us so I still have my hopes up for a foot or more!:)

Posted by: andrewjohnsonlevine | February 22, 2011 3:54 PM | Report abuse

walter, we get snow without a classis look from time to time. I thought it was going to snow more in the city than it did... I certainly would not have gone against the forecast here yesterday!

Posted by: Ian-CapitalWeatherGang | February 22, 2011 4:00 PM | Report abuse

dncandme, cut 'em some slack, sheesh, fella!

These so-called "weather people" you speak so unconvincingly about are dealing with ***immensely complex*** and less-than-well-understood systems. They sometimes have to rely on heady things like chaos theory and whatnot, disciplines that are still in their infancy.

Also, they're encouraged to dumb their messages down a few shades (vis-a-vis the animated precip icons) for a lot of different reasons (among them that lay people hate uncertainty and don't like being forced to grok even rudimentary probability/statistics).

Posted by: FedUpInMoCo | February 22, 2011 4:05 PM | Report abuse

KRUZ, I think you are misinterpreting Bob in your effort to troll us. The "dry air" is essentially the dry slot on the bottom side of the low, not cold related... As Jason noted it was clear that would be there, perhaps it ended up north of forecast. Other than that guidance had much less precip east of the mtns than west.. and with a storm moving so fast you're not going to get much higher totals than those who jackpotted got.

Posted by: Ian-CapitalWeatherGang | February 22, 2011 4:08 PM | Report abuse

The fact that you don't get it completely right is not what bothers most of us. But particularly your TV brothers bring on our contempt by boasting about the technology used and therefore implied accuracy. Then we chuckle as they cover their @$$es with a lot of double-talk when it doesn't work out. You're not politicians, can't have it both ways!

Posted by: BD14 | February 22, 2011 4:09 PM | Report abuse

weather forecasting is the only job where you can be wrong 90 percent of the time and still keep your job.

Posted by: dlpetersdc | February 22, 2011 4:16 PM | Report abuse


You get the forecast completely wrong, I quote Bob Ryan and now im a troll?....

Hey, better luck with the next storm!

Posted by: KRUZ | February 22, 2011 4:20 PM | Report abuse

The "bust" and discussion on what went wrong is a red herring. Forecasting precipitation type and difference in amounts of .1" or so is one of the most difficult forecast problems there is. Moreover, the level of accuracy in the details you are quibbling about is in the noise level of what's predictable - insufficient model resolution AND likely beyond the capability of even the most experienced forecaster to distill ahead of time.

The most important point is that all the forecasts provided a justifiable heads up of the probable (without complete certainty) occurrence of frozen precipitation and implicitly (at least) with extremely high confidence this was not going to be a major event (amounts, whatever precip type).

It's like quibbling by economists about why the market went up or down by a couple points, as contrasted with an unexpected very large jump one way or the other.

Even the explanations in hindsight to explain what happened are less than satisfactory and probably unknowable without extensive post analysis and diagnosis.

Posted by: SteveT-CapitalWeatherGang | February 22, 2011 4:22 PM | Report abuse

It was just this past saturday that I said on this very blog that we would all be here on CWG reading the excuses and downplaying the fact that the storm was a complete bust.

And alas, here we are doing just that. Some ppl could see this bust a mile (week) away. Welcome back to DC winter, mets!

Posted by: KRUZ | February 22, 2011 4:32 PM | Report abuse


Nodding head profusely. Really well put.

Posted by: Jason-CapitalWeatherGang | February 22, 2011 4:33 PM | Report abuse


You said:
"The most important point is that all the forecasts provided a justifiable heads up of the probable (without complete certainty) occurrence of frozen precipitation and implicitly (at least) with extremely high confidence this was not going to be a major event (amounts, whatever precip type)."

Any computer generated forecast based off of any of the models (that you use also) would/could do that as well. The Weather Channels 10 day computer generated forecast showed snow for yesterday 10 days out and every single day leading up to the event.

We just expect human weather forecasters to do a better job than the computerized ones. But it just seems to me that they are one in the same. Except the computer generated ones come cheaper to employers and are just as accurate when going by your logic in your quote above.

Posted by: KRUZ | February 22, 2011 4:59 PM | Report abuse

"But particularly your TV brothers bring on our contempt by boasting about the technology used and therefore implied accuracy."

BD14, you know full well that technology should imply precision but never accuracy. Hence, "death by GPS" and similar cr@ppy outcomes from overreliance upon technology:

Posted by: FedUpInMoCo | February 22, 2011 5:01 PM | Report abuse

"We just expect human weather forecasters to do a better job than the computerized ones."--KRUZ

So do you basically mean to say that the lay public should expect meteorologists to inject more "art" into the science of forecasting?
(I.e. informed judgement/discretion based upon less-than-data-intensive intelligence, intuitive thought, and so forth?)

I don't think that ridiculing and taunting meteorologists is the way to get them to trust their instincts better, do you? Are you a meteorologist or climate/weather scientist, by the way?

Posted by: FedUpInMoCo | February 22, 2011 5:14 PM | Report abuse


What I really mean to say is, what makes a human meteorologist any more accurate in giving us a "probable heads up" forecast than a computer generated forecast?

And im not really ridiculing and especially not taunting any mets anywhere. But these are professionals and like any professional. especially who is in the public realm like WAPO website they should be questioned and challenged as they as for.

And what makes me a troll or makes this any more ridiculing than what CWG mets have done to Joe Bastardi. They ridicule him when he's wrong. If writing a whole article up on a fellow pro met and calling him fearless joe and such is ok, why is my questioning not ok, and why would i be a troll for doing so :/

Posted by: KRUZ | February 22, 2011 5:30 PM | Report abuse

KRUZ, just the impression I get from your stream of comments of recent. I guess I disagree with Bob if he is saying it was the cold dry air (from the north) that caused the issues. An argument could certainly be made that dry air from the west along the base of the low track was a problem. That aside, I'd love to take you up on your idea that forecasting with no model guidance is a worthwhile endeavor. Guidance is the first part, there is much more that goes into it than just regurgitating what they show.

Posted by: Ian-CapitalWeatherGang | February 22, 2011 5:35 PM | Report abuse

So Kruz, why are you paying so much attention to CWG if all you need is the raw, unadulterated model output which you can find easily on your own, OR, just take Bastardi's word for it. There is plenty of evidence (if your are concerned enough to look for it yourself) to support the conclusion that objective, experienced forecasters add value, i.e, improve upon, raw model output and do so without unwarranted and egotistical hype.

Posted by: SteveT-CapitalWeatherGang | February 22, 2011 5:50 PM | Report abuse

"It was just this past saturday that I said on this very blog that we would all be here on CWG reading the excuses and downplaying the fact that the storm was a complete bust."

I wouldn't call it a COMPLETE bust. At least SOMETHING fell. (all caps for emphasis)

Posted by: mkarns | February 22, 2011 5:52 PM | Report abuse


"why are you paying so much attention to CWG if all you need is the raw, unadulterated model output which you can find easily on your own, OR, just take Bastardi's word for it."

I pay attention to CWG because I too thought experienced forecasters added value and were more accurate than computer generated forecasts (i am wrong). I thought experienced forecasters could also interpret models (better than comp generated) along with using real world (DC winter) data in their forecasting to get things more accurate, which most of the time is not the case with ANY human mets in the DC area.

And also, CWG is not the only place I come to get a weather forecast, I find amateur blogs to be just as accurate if not moreso.

SteveT, if you dont want CWG to be criticized for botching a weather forecast just put it posting guidelines that no one should do it. If you are asking for me to not return here and/or post here, let me know that as well and I will stop.

And I laughed when Bastardi made a prediction for 3-6 inches and I posted it on here a couple days ago. Ive never taken Bastardi seriously as far as forecasts for DC. I would never take just Bastardis word for it, just the same as I wouldnt take CWG's word for it.

Posted by: KRUZ | February 22, 2011 6:19 PM | Report abuse

What I don't understand (and what is not explained above) is why the heavy precip in the ohio valley withered away before it got to us. When I went to bed there were yellows moving in our direction. When I got up (3:30) there were a few shards of green left moving over us. Did the low weaken that much?

Posted by: eric654 | February 22, 2011 7:03 PM | Report abuse

CWG, you guys were right-on with the forecast for my area! We received less than one inch, and most of the precip that fell overnight was rain. Good job!!

Posted by: david_in_stafford | February 22, 2011 8:20 PM | Report abuse


I'm not SteveT, but please don't stop posting here! I always enjoy reading the things you post! :)


Posted by: BobMiller2 | February 22, 2011 8:43 PM | Report abuse

While it seems like the CWG is better than the rest, I have very little confidence in the "science" when I look at the forecast vs. the actuals for Fairfax for Monday: Sunday evening the forecast for Monday during the day calls for temperatures approaching 70. It was in the low 40's at best on Monday. Then 2-4" of snow in the forecast yesterday evening. We received a trace. It's far too easy to hedge your bets and make a forecast, but also list the 5 other scenarios that could come to fruition...and then say that you said that might happen. That's not really weather forecasting.

Posted by: gerin02 | February 22, 2011 9:02 PM | Report abuse

One thing I liked about your forecast was that the accumulation map was more expansive (showed the far surrounding counties) than some of the others posted this season.

Posted by: spgass1 | February 22, 2011 9:53 PM | Report abuse

Consider this comment a vote for more like this post. Do the candid postmortem. Air the dirty laundry. Figure out what went wrong and what went right. Credibility is built when you're brutally honest.

Posted by: imback | February 22, 2011 10:04 PM | Report abuse

"That's not really weather forecasting," wrote gerin02 with regard to multiple scenarios.

Yes it is. It's 21st century weather forecasting. Instead of a single deterministic weather forecast, we now have probabilistic weather forecasts. Paradoxically, probabilistic weather forecasting isn't due to less accuracy but rather more accuracy and more of a handle on the uncertainty.

Posted by: imback | February 22, 2011 10:18 PM | Report abuse


I'm pretty sure our accumulation map covered the same area this storm as all previous storms this winter. -Dan, CWG

Posted by: CapitalWeatherGang | February 22, 2011 10:22 PM | Report abuse


To be fair, by Sunday midday we were forecasting mid-50s to low 60s for Monday's highs, and by late Sunday evening we were forecasting 50s (which I believe verified very early Monday morning but not during the bulk of the day).

Our MO here is to be honest about the uncertainties associated with a given forecast. That *is* weather forecasting in our opinion. There still comes a point before any potential storm that the public is going to expect us or any forecast outlet to put out a somewhat deterministic forecast. We do that of course, and we own up when it's a bust, no matter how many other possible scenarios we may have discussed. We did bust significantly low on this storm. I don't think I'd call it a total bust - we predicted 1-4" from south to north across the immediate metro area, and ended up with more like 0.5-1.5" - but it was a pretty big bust, even if less so than others. -Dan, CWG

Posted by: CapitalWeatherGang | February 22, 2011 10:40 PM | Report abuse

Dan, I guess that is the map y'all normally use... I might be thinking of a "zone map" Jason used a time or two that was tighter in on DC

Posted by: spgass1 | February 22, 2011 10:52 PM | Report abuse


You're always welcome to post here. We're not trying to be defensive here and stifle debate, but explain our reasoning and objectively talk about what we did right and wrong. We also thought it was necessary to explain why we believed you got the right answer for the wrong reason when you forecast less snow - not to mention explain how you might not be reading the right thing into Bob Ryan's 15 word tweet (which necessarily lacked much detail or elaboration).

Ian is right- try forecasting without the models. They're the best tools we have. Yes, you have to understand the model biases and the nuances of the area you're forecasting and make the right adjustments. That's what we try to do, but every storm is different and complex. No one's been getting these storms exactly right for reasons Steve T discussed-some of the small scale details (like will the heavy snow band be in howard county or montgomery county?) are not very predictable. Finally, if you were to take all of the storms this winter, and evaluate performance across the entire set, I challenge you to find an outlet which has done better than we have. If you say you have your sources that have been consistently more accurate, demonstrate that.

The bottom line is that it's been a very challenging winter for snow forecasting locally. More challenging than any year since we started this blog. Every forecasting outlet has struggled. What we try to do here, which is different, is communicate the full range of scenarios while giving our best bet. We also provide a lot of information about how the storm will impact the region - which you can't get from raw model output. Ultimately, there's more to forecasting than getting it right or wrong, it's about the communication of risk and impacts - which is something we put a lot of effort into. And in terms of how the storm would impact the region, irrespective of the error in amounts, we were more or less right last night (as Steve T indicated).

Posted by: Jason-CapitalWeatherGang | February 22, 2011 11:03 PM | Report abuse


We're all about being open to forecast criticism. We just prefer when it's fair and acknowledges the limitations of the science. Our metro area forecast of 1-3" south and 2-4" north obviously turned out too high. But it was better than some who were forecasting 4-8" where we were forecasting 2-4" (of course, we'd much prefer to do better in absolute terms rather than relative to others).

There were models that were indeed showing substantially higher totals than we predicted, and we went against them in part driven by the "real world (DC winter) data" you speak of and our experience/gut instinct.

You say, "The trend will always rule out models and what happened last night is the trend around DC this year ... When you live by the models (model hug), you die by them along with your credibility. The trend shouldve been your friend, not stupid models!" I've heard others suggest that meteorologists shouldn't base their forecasts on models so much. But models are all we have, really, until a storm starts to take shape and approaches, when satellite and radar and observations come into play. Today's models are pretty incredible - they are the reason we're almost never blindsided by storms like folks were decades ago. A good meteorologist interprets models and tweaks their forecasts based on what they know about a model's bias, how well a particular model handles a particular weather pattern, and their experience and instinct. But at the end of the day it's the models that form the basis of any weather forecast, and that's not going to change. -Dan, CWG

Posted by: CapitalWeatherGang | February 22, 2011 11:04 PM | Report abuse

Of course models are a necessary part of forecasting. I'm by far no WX expert, but I can see that this real world data stuff being touted is just another way of phrasing long-term typical outcomes (climatology) and short-term (ie current season) outcomes ("the trend"). I'm pretty sure both of those (climatology and the trend) are usually referred to in CWG's forecasts, and I'm pretty sure using those alone would result in many busted forecasts. Some of the haterz IMHO before the first big storm of last winter would have shouted bust solely because of DC area winter climatology- and that's outrageous.

At the end of the day, CWG's forecast-bust amounts to 0.3" - 5.0" actual snowfall across the region when 1" - 4" was forecast for the DC area. So we're getting all worked up over a 0.7" discrepancy. Have you ever looked at 0.7" on a ruler??

Posted by: kolya02 | February 22, 2011 11:58 PM | Report abuse

Re much earlier comment by "Chilean" about recent temperature underforecasts, I believe the problem is relying too much upon climatologically watered down MOS (Model Output Statistics) forecast temperature forecasts rather than model 850mb forecast temps. In a well mixed warm season style pattern with a low level westerly wind component in a well mixed atmosphere (like we had last week even tho it was still Feb.), following the dry adiabat on the skew-T chart and adding 25-30F (closer to 30F in summer) to the GFS forecast 850mb temps gives a much more accurate result than MOS. There were quite a few times last summer that the GFS forecast 21C (70F) 850mb temps and DCA recorded @100F sfc temps!

Posted by: buzzburek | February 23, 2011 1:16 AM | Report abuse


i was wondering how long it would take for someone to bring up last winters 70+ inches (El Nino) and try to tie it into a typical DC winter, like this year. No one wouldve shouted bust because most models were agreeing on the storm almost 2 weeks out. Koyla you will not see a winter like last years for another 20+ years in DC, maybe more ;)

Misinterpreting Bob Ryan.... How many words do you need to say:
"cold dry air really kept amounts on very low side"

Either you agree with Bob Ryans statement or disagree.

KRUZ (from sunday)
"no models will verify this thing until prolly sometime tomorrow as the 1st low is overhead. will prolly end up being a nowcast with radar. I still dont think the cold enough for snow air is gonna come in until after low 2 moves by us here in DC. Looking like a snow event north of DC, i think 1 inch accumulation is still far fetched with this much uncertainty in the models.
Posted by: KRUZ | February 20, 2011 4:18 PM"

See, im not saying models arent valuable or shouldnt be used in forecasting, they are a tool in forecasting but not always right, especially when they dont agree. And my point was that when there is as much uncertainty in the models within 48 hours of the event like there was in the models for this storm, from temps, to tracks to precip amounts and even type... Thats when trends come into play, especially for the DC metro area and our dreaded dry slot.

As far as trends go most storms this winter except for the one perfect "thread the needle" storm, all have cut to our north and west or south and east. Early on this winter most were cutting south and then OTS east. Later into winter storms started heading north from our west and either missing or clipping us from the north and/or west.

Storm #1 from sunday tracked to our north which storm #2 would imo inevitably try to follow. I thought the models had the track too far south and I think storm #1 had alot to do with not only the models getting it wrong but also the storm taking the more northerly track.

Posted by: KRUZ | February 23, 2011 1:17 AM | Report abuse

sorry for that half double post

Posted by: KRUZ | February 23, 2011 1:25 AM | Report abuse

great discussion here.

you did that calculation that purported to show CWG's forecast was off by .7". well, that's totally the wrong way to look at it. you can't look at absolute figures, especially for such a small storm. falls church was juuust north of the 1-3" / 2-4" line. i'll consider it on the line and take the average of the two zones' forecasts. based on that, i could have reasonably expected 2.5". i got 1/8" of snow on top of 3/8" sleet - a total of .5".

so... i got 20% of what i was forecast to get. that's a bust. even if you take the average of the low end predictions for falls church, you get 1.5". in that case i still only got 33% of what was forecast. that's a bust too.

no one ('cept possibly KRUZ...) got this forecast right. CWG has "manned up" and "owned up" to that error many times all over the blog yesterday. i was as disappointed as anyone. but after some initial venting, i remembered that weather is fickle and forecasting it is an inexact science/art. that's the way it goes. there's always the next storm...

Posted by: walter-in-fallschurch | February 23, 2011 8:17 AM | Report abuse

Good point, buzzburek. You win a box of doughnuts.

Posted by: imback | February 23, 2011 9:00 AM | Report abuse

The comments to this entry are closed.

RSS Feed
Subscribe to The Post

© 2012 The Washington Post Company