Many thanks for taking the trouble to do that analysis of the CFS runs, Blizzard of 78. It's really appreciated.

I've been following the runs of CFS version 2 recently and comparing them with version 1. (Version 2 will be taking over next month.)

Version 1: http://www.cpc.ncep.noaa.gov/products/analysis_monitoring/lanina/ensoforecast.shtml

Version 2: http://www.cpc.ncep.noaa.gov/products/CFSv2/CFSv2seasonal.shtml

The 2m temps are hard to compare between the two versions, as version 1 anomalies are compared to the 1981-2008 temps, while those in version 2 refer to the 1999-2010 temps.

Generally speaking, though, the 2m temps forecasted for the UK in Dec., Jan. and Feb. seem to be milder in the new version 2 than in the old version 1.

There's no certainty that mild will win out, of course. For one thing, I'm always suspicious these days when calculations get tweaked and, hey presto, turn out with a milder outcome, particularly for winter. After all, it was the adjustment for "background warming" that led some of the UK Met Office winter forecasts astray, until they stopped publishing them.

Originally Posted by: TomC

Originally Posted by: Solar Cycles

There's no certainty that mild will win out, of course. For one thing, I'm always suspicious these days when calculations get tweaked and, hey presto, turn out with a milder outcome, particularly for winter. After all, it was the adjustment for "background warming" that led some of the UK Met Office winter forecasts astray, until they stopped publishing them.

Indeed, better to leave the data alone without adjusting for faux warming. I do believe the MetO's probability charts are free from such adjustments.

The probability charts are calculated by running the met office unified model which means they start from current conditions hence they include the climate change to date. Obviously future climate change is irrelevant to the forecast for the next season. The other key factor is when you talk about cold, average and mild is what averaging period you refer to using those terms.

[/quote]When you say include does that mean the 1970-2000 average, or with the dreaded AGW manipulated factored into it. If it's the former that makes sense, if the latter, well

Originally Posted by: TomC

Originally Posted by: Solar Cycles

Originally Posted by: TomC

Originally Posted by: Solar Cycles

There's no certainty that mild will win out, of course. For one thing, I'm always suspicious these days when calculations get tweaked and, hey presto, turn out with a milder outcome, particularly for winter. After all, it was the adjustment for "background warming" that led some of the UK Met Office winter forecasts astray, until they stopped publishing them.

Indeed, better to leave the data alone without adjusting for faux warming. I do believe the MetO's probability charts are free from such adjustments.

The probability charts are calculated by running the met office unified model which means they start from current conditions hence they include the climate change to date. Obviously future climate change is irrelevant to the forecast for the next season. The other key factor is when you talk about cold, average and mild is what averaging period you refer to using those terms.

When you say include does that mean the 1970-2000 average, or with the dreaded AGW manipulated factored into it. If it's the former that makes sense, if the latter, well

I don't understand your comment. I mean the current global temperatures, ice cover etc. I don't see what it has to do with AGW other than its obviously warmer.  You don't initialise a model from the average temperature over 30 years but using the measured conditions on the day you intitailise it. The model then gives a predicted temperature distribution and from that you calculate anomalies from the baseline you chose. Obviously that baseline will be higher if you use more recent averages.

I think the key here is that the input to the model is the state of the atmosphere as it is observed from satellite radionsndes and surface observations at the time it is initialised not some average conditions.