October 25, 2009

The Disappearing Windstorm


Yesterday, a few of you were noting the potential for a major windstorm later in the week, with the National Weather Service GFS (Global Forecasting System) model indicating an amazingly deep low pressure system right off our coast. The graphic (sea level pressure-solid lines) is shown above for you amusement (132 h forecast starting 5 AM Saturday). SCARY! Hurricane-force winds offshore! Gales in the Strait of Juan de Fuca! The TV stations would have field day if this was true.

But one thing forecasters learn quickly is to restrain themselves that far out and to look carefully at run to run continuity. Is this forecast stable? Did several model runs (forecasts) show the same feature. Did the forecasts of several modeling systems produce the same solution? For this storm, each run has been very different. Want to see?

Here is the next run starting 12-h later...much less scary--and farther north.



And the run starting 12-h later (this morning at 5 am, 108 hr forecast)... its nearly gone.


Or the run starting a few days (Thursday, 5 AM) before the scary one....the low is weaker and far inland.
The bottom line is that forecasters have little confidence in such erratic predictions that far out. Generally forecasts settle down when one is less than about 108 h out...but not always. The current situation is a particularly tricky one, since some of the energy of a tropical storm in the western Pacific is being injected into the midlatitudes...and that inevitably decreases the reliability of forecasts. The technical terms is extratropical transition...or ET.

20 comments:

  1. It is pretty interesting watching the discrepancies between each models forecast. There are two storms in my recent memory that the models did seem to handle pretty well, even though they were many days out. The first was the Hanukkah eve windstorm and the most recent was the cold snap last December. Last week I wrote down the extended forecast for this upcoming week from King5.com to see how accurate/lucky they are. It will be very intriguing to compare their weather for this week to what actually occurs.

    ReplyDelete
  2. I've been watching that low disappear also, Cliff. At first, when that run that you showed came in, I was ecstatic. Granted, it might of been a bit to strong to handle, but I love when weather gets a bit crazy.

    To bad it's almost non-existant now. Maybe mother nature is saving us this time (and her energy), to gear up for a big 'kabang' sometime in the near future. ;)

    ReplyDelete
  3. How did "weather is my life" manage to post comments on your blog again? He is a sneaky one!

    ReplyDelete
  4. Very valid points. That's why I trust this blog for setting the facts straight before creating mass hysteria. Also, since we all have easy access to the extended GFS run, we all saw it. That's why I like reading the NWS long term discussion because they know stuff we don't. They mention the EURO alot, and tend to rely a little more on that model in the future.

    ReplyDelete
  5. GFS Models are certainly in general agreement about the low being well to our north, up until tonights 18z which shows 994mb low coming into Western Wa early Fri morning. Even GEM model shows the low, but farther inland, north of us and over south-central BC for during the Fri. EURO shows far northerly track as well. So we might see windy conditions, and then again, we might not. We`ll see.

    ReplyDelete
  6. This is great fun to see as well as educational. Par for the course, which is why I keep coming back here. I really like to see what the "it" weather or other topic is. Much better reading than many other sites out there no matter the genre.

    ReplyDelete
  7. How far can atmospheric models go? Is there a limit? Or is this just a question that drives poor researchers further into their beer when they get together.... Personally, I thought it was amazing when the modeling folks pushed things out to 72 hours.

    ReplyDelete
  8. All,
    The European Model is the best (embarrassing to admit as an American). Roughly a day more predictability. ..cliff

    ReplyDelete
  9. ah man I was really looking forward to that. 954 mb low right off our coast? That'd be something. Bring it in 300 miles closer too... haha 30 mb PDX BLI gradient?
    But of course there are far more forecasted major windstorms than actual ones

    ReplyDelete
  10. Besides better computer modeling of the future, I'm just as impressed by the work being done for local, short term forcasting such as the 4km domain. But where is the line between human input and computer output. I live close to yakima, and my dad called freaking out about the big rain! I looked up what he heard, rain likely late tonight 60%, Rain 70% tomorrow, chance of showers rest of week. I mean, is this a human or computer forecast? Surface low way up in Alaska, west flow in all levels. Just by the way the clouds behaved today, Lewis and Clark would know about this pronounced rainshadow.

    ReplyDelete
  11. This sorta leads into a question that I wanted to ask Cliff.

    On Friday night, the NWS forecast called for 100% chance of rain on Sunday (Detailed Point Forecast). On Saturday afternoon, the forecast had changed to 30% chance of rain. This isn't the first time that I have seen a NWS forecast like this.

    What does "100% chance" mean when the forecast can go from 100% to 30% in 12 hours, 36-to-24 hours before the forecasted period?

    ReplyDelete
  12. From Yakima tonight the horizon to the NW is lighting up. With completely clear skies, we can even see the lightning from the convergence zone this far away!

    ReplyDelete
  13. Tipo,
    Basically the NWS altered their forecast based on updated model guidance and observations, suggesting the system was coming in slower. So you can go from 100% to 30%!..cliff

    ReplyDelete
  14. Yes, I saw some cloud to cloud lightning tonight to the north.

    ReplyDelete
  15. We had multiple bursts of pretty big (pea to chickpea sized) hail tonight in East Bellevue, and I heard thunder more than once during my daughter's horseback riding lesson on Hollywood Hill around 7:30 pm.

    And dang, it's COLD... 43F but it feels really raw since I haven't had the natural gas turned back on for the season yet -- our house is chilly and the car, with its "free" heat, is my refuge when running errands.

    I'll be glad to have the heat on even though we sort of took bets on whether or not we could hold out on it until Thanksgiving. My answer today is HELL NO. Turn up the heat, baby!!

    We're ready for wind, though. I filled up the gas cans for the generator with fresh juice over the weekend. :-)

    ReplyDelete
  16. Thanks for the answer to my forecast question, Cliff.

    I understand what happened with the NWS model, but what I was asking was what a forecast calling for 100% chance of anything is supposed to mean.

    Given that these models aren't perfect, how could you ever get a 100% chance forecast?

    ReplyDelete
  17. Pretty quite here yesterday evening as far as weather goes, but hear a rumble of thunder off to my south late last night. Did get quite a bit of rain yesterday morning though and picked up 0.78". My monthly rain so far is 5.01"

    ReplyDelete
  18. I was driving from Oak Harbor to Bellingham last evening from bright sunshine to one of the blackest daytime skies I've ever seen. Scanning to the north and west I noted a lighter ground to sky cloud to the west. As I was asking myself "Tornado?" it lit up with halos of lightning. Further down the road I encountered piles of small hail on the road. I saw quite a bit of lightning activity to the west.

    Is there any chance there was a funnel cloud in that area? I see nothing in the news. The storms looked rather severe from my perspective but by the time I got to Blanchard Mountain on Chuckanut Drive I only encountered light rain.

    ReplyDelete
  19. Cliff, where I said "West" I meant east. Can you edit that? Thanks!

    ReplyDelete
  20. Maybe your next posting will be "The Disappearing Rainstorm."

    ReplyDelete

Please make sure your comments are civil. Name calling and personal attacks are not appropriate.

Should I Let AI Write My Blogs?

I spend a lot of time thinking about and writing my blogs.   Could AI apps do the work for me?   Could any of you tell the difference? Well,...