Accuweather winter forecast is out for 2011-12

[M Conder]

What is the difference between "we don't offer an opinion" and "cannot be determined"?

The difference is that science went into the forecast and they "admitted" that they couldn't improve on climatology in that period. "We don't offer an opinion" sounds like they weren't even interested in investigating it.

In my opinion, there is nothing wrong with forecasting climatology. But, if you are going to criticize AccuWeather for doing it, then the NWS should be criticized for forecasting climatology or "no opinion."

Indeed, but that is not what you claimed. You insinuated that the NWS was somehow worse because it was not even bothering to forecast climatology, which is not the case.

 

[William Monfredo]

Also, William, Mr. Bastardi has left AccuWeather. He does not produce their seasonal forecasts.

Also, since my brother leaves in Chicago, I will be interested in seeing if he "wants to move" after this winter as AccuWeather's Long-Range Meteorologist Josh Nagelberg claims.

Patrick, thanks for your update on Bastardi. I followed him extensively on the Accu-Wx Pro site for a couple of years, and he presented himself reasonably well IMO, even when he only had analogues to rely upon. But, yes, I too know people in Chicago, and I am wondering if a forecast of their departure will pan out.

 

[Mike Smith]

The difference is that science went into the forecast and they "admitted" that they couldn't improve on climatology in that period. "We don't offer an opinion" sounds like they weren't even interested in investigating it.



Indeed, but that is not what you claimed. You insinuated that the NWS was somehow worse because it was not even bothering to forecast climatology, which is not the case.

I claimed no such thing. Please do not read something into my words that isn't there.

My point is, and has been, is that there is nothing intrinsically wrong with forecasting climatology nor is it wrong (in fact, as I scientist I like it) to say, "we don't know" or "we don't offer an opinion."

The point I am trying to make is that the NWS forecasts climo and no one says a word. AW does it and StormTrack lights up. Just asking for a little consistency.

 

[Patrick Marsh]

The point I am trying to make is that the NWS forecasts climo and no one says a word. AW does it and StormTrack lights up. Just asking for a little consistency.

The difference is, I don't recall anyone making a special post about the NWS forecast. Furthermore, the forecast you provided is the forecast for May, June, and July of next summer -- a 7.5 month lead time. Where is AccuWeather's forecast for next summer? Compare apples to apples.

If you are going to call people out and ask for consistency, I would hope that you would be consistent on what you are comparing.

 

[David Wolfson]

To my eyes, it's pretty remarkable the extent to which this winter forecast is busted so far. The GFS continues a persistent zonal flow into-mid January with little or not arctic air incursions into the US and storms limited to the Northeast. The usual Siberian cold pool isn't evident into January, and Europe remains above freezing through the period. Uncharacteristic east oceanic highs sit off both the US west coast and Europe.

Still plenty of time, of course, to return to climatology, but the December forecast is definitely history. Any thoughts?

 

[Jim Cross]

I found this the other day on the Chicago NWS office web page. http://www.crh.noaa.gov/images/lot/newsletter/Rest_of_Winter_12_Outlook.pdf It explains how the effects of the North Atlantic Oscillation (NAO) and Artic Oscillation (AO) impact the winter season.

 

[David Wolfson]

Very interesting indeed. I gather from doing some reading, though, that it's not well understood whether the NAO/AO is causative of or just correlated with the circulation patterns. When AccuWeather made the winter forecast the AO was solidly negative, and it seems to rather abruptly be going positive. The "why" doesn't seem to be at all clear.

I gather that most research has been statistical, and that variables with a significant lead signal, e.g. snow/ice cover, SST, don't appear to explain much. Perhaps it has been done, or is now being done; but if I were a researcher an NCEP I would find it irresistible to set up a Bayesian approach -- running out the GFS, likely even beyond 384hr and twiddling certain suspect inputs to generate distributions specifically of the AO.

Sorry if this gets into the weeds, but its interesting I think.... FWIW.