Monday, November 26, 2012

CFS Model Grossly Over-Exaggerates December Forecast

The Climate Prediction center's Coupled Forecast System model, or the CFS model (version 2) has put out a new December forecast that crosses the line between unrealistic and just plain absurd.

The top image is the anomaly forecast for the month of December, and the bottom image is the skill forecast (areas where the model is very confident in its forecast). As you can see, the CFS has temperature anomalies well over 1.5 degrees above normal spreading from the West Coast, through the Rockies and well into the Plains. Temperature anomalies exceeding 1 degree above normal exist in just about every state in the Lower 48, with the possible exception of eastern Florida. The skill map is confident in the East US' warmth and the Plains, but not so much for the Rockies.

This forecast is just plain ridiculous. The model is going off the idea that the persistent Omega Block will continue through the rest of the month, which, as shown in the GFS forecasts below, is likely not going to happen.

The top image shows the GFS Ensembles' 500mb height anomaly forecast for Hour 384. Those of you affluent in the weather model forecasting business know that Hour 384 is the end of the GFS run and generally considered worth nothing that far out. However, when you get ensembles that come along, the idea changes, and suddenly you have 20 different forecasts coming together. The GFS Ensembles believe that the Omega block over the Bering Sea will dissipate by Hour 384, thus ending the NE Pacific trough and moderating the continuous warm shots that the nation has seen and will continue to see for another week or two. The bottom image is the GFS model itself, and the model shows this Omega block being pushed to the west- not the best situation, but good enough so that the persistent trough over the West Coast can leave the area and provoke a positive PNA pattern to set up behind the persistent trough, which has now moved west with the Omega Block.

Now, there is definitely potential for this blocking to stick around in the Bering Sea and make December warm. However, there is one more reason why you should not take the CFS for what it's saying.

This is the CFSv2 model's RMS Error for 2m temperatures in forecasts made in November. What this shows is the gap between forecasted situations by the CFSv2 and observed conditions for that same forecasted timeframe. Basically, this is the difference of those two factors. The DJF period (December, January, February) is shown above, and it appears that forecasts made in November typically verify extremely warm in the DJF period. This not only includes this December forecast, but the other two winter months that we don't particularly care for at the moment.
The CFS' track record is definitely a warm-biased one, meaning it has a tendency to forecast conditions too warm according to this map.

Not only that, but this chart of multiple models' forecasts for the month of December show that the CFSv2 is the ONLY MODEL to show a torch for the first month of winter. The only reason the NMME and IMME models are showing warmth is because of the CFSv2 is averaged into their forecasts.

So please, winter weather lovers, do not get so fired up about the CFS' forecast. In my eyes, it's the least desirable model (next to the Japanese and Canadian models) to use for a forecast. Do you REALLY think that December will bring extreme warmth to the nation??


Anonymous said...

The local weather guys in Ohio are saying above normal temps.the first two weeks of December.I agree 100% with them.This up and down pattern we are in right now is going to be hard to break any way you cut it.

Josh Herman said...

Andrew, why is the model showing such a warm December in your opinion?

Anonymous, other 'models' based on specific cycling weather theory show OH having a warm first half of Dec and then cooling off dramatically the second half.

Anonymous said...

Sorry, Andrew, but I have a hard time not believing what it is saying about the Rockies about the historic hot and dry conditions, this really is a horrible season, winter for us just isn't coming, and it doesn't look like it will EVER come back to us, time to get out of here, I don't want to be here when those deadly fires inevitably come back to Colorado next summer, when there is no doubt in my mind that temperatures will soar to levels they have never been to in all of history, only this time, there wont be any water left to fight them, and what is left of our state will burn fiercely, and by next fall, there may not be anything left worth protecting...goodbye central Rockies. :(

Andrew said...

Josh: The CFS weeklies carry the Omega Block over the Bering Sea through the month of December, hence at least part of the heavy warmth. I believe other parts of the warmth is the bias, and there are other factors that also influence the CFS into its forecast.

Anonymous: The Rockies have survived for centuries upon centuries- why stop now? :)