The question at hand is how far out the Forums members should be allowed to discuss storms - in other words, if a storm appears on the GFS model 15 days out, can we talk about it?

For the vote, I created this graph, which shows how far each Forecast Models [JessePedia] (from our AccuWeather.com Professional site) predicts into the future.

*Technically the ECMWF goes out 15 days, but few meteorologists, and no Forum members, have access to it.

Note that, after you back up to 10 days into the future, you have a few models to go on. Down to 3 or 4 days? Plenty of different forecasts.

The arguments swing the gamut, but basically people are concerned with "Wishcasting" which is hanging your hat on a long-range GFS prediction (usually of a hurricane landfall or blizzard), when it's likely that forecast will change considerably. With no other models going out as far as the GFS, there are no other "opinions" on the forecast so people declare it the truth.

I stated my opinions in the thread when I created the poll. I do not believe that we should censor people's thoughts on the weather, and 15-day forecasts are routinely discussed intelligently by AccuWeather.com meteorologists, so there's no reason that we couldn't on the Forums. You can even get point forecasts on AccuWeather.com out to 15 days.

The GFS may present some false positives, and swings wildly in the long-range, especially for point forecasts (see my White Christmas post last year) but it's proved that it shouldn't be ignored many times in the past, for example 12 days ago, it was talking about the snow event that will occur in the Northeast tonight.

What do you think? Post on the Forum thread, or if you're too lazy, leave me a Comment below.

Comments (4): Jim:

I think that 7 days should be the limit. Yes, the models go out further, but the forcast always changes as more data comes in. Sure, I'd love to have two feet of snow in Richmond, VA; I hope for it every year, but it is very unlikely. The last time we had a whopper of a storm (or a real winter for that matter) was early in January in 1996, 12 years ago! Thanks.

Posted by Jim | October 22, 2008 10:05 AM Scott:

It's always a long winter with the GFS. While it has had its moments, I don't trust a lick of what it says until around 4 days out.

Posted by Scott | October 22, 2008 9:53 AM Dominic Froio:

Jesse:

Forecasts are usually accurate 2 or 3 days out.15 days shouldn't even be considered. Forecasters are already predicting heavy rain in my area (Marlton NJ) for Saturday through Sunday morning. I've seen such forecasts in the past 3 or 4 days out where not even a drop of rain feel.

Dominic

Posted by Dominic Froio | October 22, 2008 8:45 AM Billy:

As a meteorology student, I think the GFS model is going to be fun to watch this winter. As Henry said on the early indications of the NE snowstorm, it may be a long winter with the GFS. I think the GFS should not be trusted this winter 15 days out.

Posted by Billy | October 21, 2008 10:23 PM