NOAA Long Range Winter Outlook 2013-2014

[vortmax]

I've been thinking about this lately and wanted to spark some discussion around the accuracy of NOAA's long-range Winter Outlook (specifically for this winter). Based on the data released in late November (http://www.noaanews.noaa.gov/stories2013/20131121_winteroutlook.html), why did they not even remotely see signs of this extreme -EPO pattern developing? If the Winter Outlook is meant to "support local and state governments in their effort to plan for public needs during the winter", then they really don't appear to be helping much with such an inaccurate forecast. Would it be better not to release a Winter Outlook (with multiple revisions) until more accurate ones become possible? Or maybe only release an outlook/alert if abnormal/extreme patterns appear likely to develop. 

 

With that said, I do have a BS in Met so I'm not completely ignorant of how difficult it is to forecast a couple of days in advance, let alone months. I'm just trying to understand the major disconnect for this season. Thanks.

[RCNYILWX]

I've been thinking about this lately and wanted to spark some discussion around the accuracy of NOAA's long-range Winter Outlook (specifically for this winter). Based on the data released in late November (http://www.noaanews.noaa.gov/stories2013/20131121_winteroutlook.html), why did they not even remotely see signs of this extreme -EPO pattern developing? If the Winter Outlook is meant to "support local and state governments in their effort to plan for public needs during the winter", then they really don't appear to be helping much with such an inaccurate forecast. Would it be better not to release a Winter Outlook (with multiple revisions) until more accurate ones become possible? Or maybe only release an outlook/alert if abnormal/extreme patterns appear likely to develop.

With that said, I do have a BS in Met so I'm not completely ignorant of how difficult it is to forecast a couple of days in advance, let alone months. I'm just trying to understand the major disconnect for this season. Thanks.

If you check their verification plots back to the mid 90s, I counted 2 or maybe 3 outlooks that verified well and obviously this winter will not be one of them. One of the issues i believe is that they're too reliant on a superensemble of long range guidance as opposed to looking at things like trying to predict the predominant state of the AO or EPO or tropical forcing mechanisms aside from ENSO. It's become clear that in the case of the extreme -EPO that it's been prevalent this winter at least in part due to the pool of very warm water in the eastern NPac. Another issue is the outlook is not really a forecast but a probability scheme. But if you just go by the spatial coverage of the enhanced chances for above and below normal temperatures, their outlooks have very rarely verified. As someone who is involved in climate research at my WFO and also on the CR climate team, either CPC should overhaul how they do their outlooks or stop doing them altogether.

[Mitchel Volk]

The science is just not there yet. For my office I released a winter outlook right after Thanksgiving saying above normal temps and below normal snow for NYC. By the next week I realized I would be totally wrong and did a 180 and forecast blow normal temps with above normal snow. I noticed the EPO as well it has been a cold November and the medium range models look cold for the next two weeks. So using persistance, oscillations, and waiting until the winter has started and the models I revised my outlook. I all ways use a disclaimer when I do long range forecasts. The only reason I do them is that people want it, I think it will be years before we can make any accurate seasonal forecasts.

[PSUBlizzicane2007]

Yeah I can't even think of a winter outlook that verified well.  Sometimes I almost feel as if, once they have a forecast with their usual methods, they should just report the opposite of what their methods produce, and it will be more accurate.

[wildweatherman179]

I'm not sure what they were thinking. I've been laughing about a lot of them since late October. I posted my forecast to Facebook for all my friends to see on November 3 I think and so far it has been incredibly close. Need February to bring some snow to NE Oklahoma and it will be a spot-on forecast.

[JBG]

Yeah I can't even think of a winter outlook that verified well.  Sometimes I almost feel as if, once they have a forecast with their usual methods, they should just report the opposite of what their methods produce, and it will be more accurate.

I do remember the forecast for the epic 1995-6 winter being generally accurate. Most have bombed, most notably 2010-11 which was forecast to be typical La Niña garbage amd turned out to be epic.

[Jonger]

I have a feeling we are going to start seeing more blank maps of the country with equal chance for temps and precip. Any forecast beyond 10 days is asking for trouble.

[WEATHER53]

Noaa does what most do who have a <50% success rate-they try to throw in all the indexes plus come up with a new one each yer. Too many ingredients. Ask KA, Zwyts or me what the winter will be and 75%+ you will get the right answer. Nature shows her hand in a general sense as the year progressess, you can learn how to read that and carry it over into the winter but only if you dump most of the indexes.