Meant as a general reference resource more in-house, essentially, .. Interested in the idea more myself, and with the "GFS" upgrade set to take place coming up in January (?), these ideas with also, connected to the subject and where looking at the more extensive speculation more model/s related more recently with quite a bit of commentary regarding ability / credibility of whichever of them, I've thought to perhaps initiate a thread more specifically dedicated to the theme.
— What's you're, slant and view where looking at whatever model's performance. ? / One set against another. ? / Least, or most reliable, and why. ?
.. In addition to whatever appreciation more personal that might be tacked in here, there are also different more basic accuracy checks of many of the various different models' capabilities I know. I've seen a few of these in the past, if I can't point to any more immediately here. .. For my own part here further for fun, I'll be going back over some of what's been said (again.) more recently about whichever models' worth as different people of have seen / see the idea, and dropping in some different "quotes" in here.
Model "Pros" and "Cons" ...
Posted 27 December 2014 - 05:37 PM
- Bedonbye likes this
Posted 27 December 2014 - 06:57 PM
".. Is Numerical Weather Prediction One of Mankind's Greatest Achievements?"
A recent commentary on model generated forecasts graphically depicted, by "Cliff Mass", Professor of Atmospheric Sciences, University of Washington.
Posted 28 December 2014 - 02:40 PM
The ".. History of numerical weather prediction", .. according to "Wikipedia".
http://en.wikipedia.org/wiki/History of numerical weather prediction
Posted 04 January 2015 - 02:23 PM
The talk of stuff beyond 10 days seems like mumbo-jumbo until it gets closer... too many things can change. .And then the details usually still ruin it for us. .We are in the heart of the golden Jan 1-10 period everyone talked about for the last month. .That did not work out well.
Posted 04 January 2015 - 02:27 PM
Posted 04 January 2015 - 02:28 PM
That really is a bit silly. The models have improved tremendously in the 4-7 day period over the last few years.
They obviously still struggle with the details, but if the models agree on a region-wide arctic blast inside 7 days, it's probably going to happen. Snow is much more fickle and based on the details so I can understand your skepticism there.
Posted 04 January 2015 - 02:34 PM
Gotta say, it's an even better feeling to just let it roll. Does it really matter to have an extremely vague idea of what may happen weather-wise 400 hours from now?
Posted 04 January 2015 - 02:38 PM
Yeah, the recent "fail" was not really a model fail. More of some of us weenies getting our hopes up. There was never good agreement on anything more than what happened. I will say that I was pretty conservative in my expectations, I was expecting something along the lines of January 2013, Dec 2010. But sadly even my expectations were to high.
Posted 04 January 2015 - 07:54 PM
As much as we think we understand the atmosphere there is much we do not understand about the complex processes that result in our ultimate weather. Computer modeling has made tremendous leaps the past twenty years and has gotten better in the mid ranges but is currently having a difficult time coming to a consensus with the current pattern change. This is one reason the NWS does not go all out when one or two models show cold/snowy weather more than a week ahead. One, because it is generally outside of their forecast time frame and second its the PNW, west of the Cascades and it is not typical winter weather for us.
The thing about extended weather forecasting, and I am by no means an expert, is you never get there. By that I mean everyday you get another bite at the apple and get to revise your forecast plus one more day. As we have seen, especially during the winter months, it is difficult enough to just forecast anything let alone snow. Snow at most of our locations is probably one of the most difficult aspects to forecast as it often under marginal conditions or worst. Then out of no where we wake up to a surprise event not forecast but because the marginal conditions just barely favored us.
The trend right now is to move the ridge closer to the west coast and we all know what that means for the weather at our various locations, and the progression looks sensible. Can the models change the next few days and retrograde the ridge back west a bit..absolutely. Only time will tell. Until that happens lets enjoy the weather for what it is and what ultimately takes place.
When I was a kid I lived for snow. The first thing I would do 'every' morning was go to the window and look outside to see if it had snowed during the night and was disappointed if it hadn't. Didn't want my parents to take us on vacation as I didn't want to miss any potential snow, and in my mind I knew it would snow if I wasn't around. As you grow older you put it all in perspective and realize it isn't as important as you once thought. Don't get me wrong I would seeing it snow and get cold and I still get excited when I see the models point in that direction. You just learn to take emotion out of a forecast and use common sense, statistical probability, and climatology into consideration. Hope that makes sense.
Merry Christmas everyone....
Posted 05 January 2015 - 04:08 AM
.. Relative, to what I've said suggested above, I've also been looking at the idea that the models, "may" just "perhaps" be underestimating the pace of cold's—looked at as a whole—movement more eastward from this point through to the first few days of January. Which would certainly tend to figure in to the actual sets up more kinetic as things move more forward.
Posted 05 January 2015 - 04:09 AM
If the models can't figure out this weekend... the end of next week is sort of meaningless.
Posted 05 January 2015 - 04:19 AM
Posted 05 January 2015 - 04:22 AM
A general focus on the ECMFF model performance, for a 10 or so day period: .. with an actual toggle comparison (Gif loop animation.) showing its capabilities at 168 hrs / 7 days out, with more 0hr initialization inputs depicted daily, where looking at the 850 / 500 millibar height levels.
Posted 05 January 2015 - 09:52 PM
"A Major Advance for Numerical Weather Prediction in the U.S."
Posted 20 January 2015 - 07:15 PM
I Like to do what the Weather channel will do and compare the Euro models and the GFS models. I think that the GFS models are accurate, but you know weather will change quickly all the time, and then the models will completely change
Posted 20 January 2015 - 07:52 PM
.. The comparison is a pretty standard idea. .And the GFS model has just recently been upgraded.
— Welcome to the forum John.
- johnstr8 likes this
Posted 27 January 2015 - 11:30 PM
I'm beginning to think the GFS may win the GFS / vs ECMWF & UKMET battle. Could the GFS be a better model now?