Hello folks,

I've been doing a bit of research into the last two winter predictions from the CFS model.

Have looked at predictions from August and September 2011 and 2012 to see whether the charts CFS is showing now for winter 13/14 should be taken seriously.

http://www.gavsweathervids.com

See what you think.

The final and official autumn 2013 forecast will be issued next Sunday.

---

Cheers Gav

Interesting to see CFS hasn't a bad record .

Let's hope the States don't steal all the cold eh?

---

To me it looks like one of those winters similar to 78/79 where both side's of the Atlantic would be cold.

What I find quite surprising (and impressive) from CFS V2 forecast for the past two winters is that they were very different winters (11/12 largely zonal, Atlantic dominated - 12/13 blocked, cold, wintry, etc... especially later in the winter) And it appears that both times it had these very different patterns reasonably well predicted as early as August.

I wasn't expecting that at all, to be honest.

 

---

Thanks Gav. Very interesting.

I do find the CFS to be quite useful for long term trends.

Whistling in the dark?  thats a phrase i have never heard before.  Is it similar to pissing in the wind?

---

 

---

The Glosea4 model has updated for August....very interesting. Basically very supportive of those CFS v2 anomalies in your vid Gavin. HP anomaly situated Scandi-Iceland-Greenland.

 

---

Interesting, anyone know what Glosea4 predicted for last Winter?

---

has madden issued his first 'harshest winter in a century' prediction yet?

I still only rate the CFS V2 model as the Glosea4 model is just as erratic as the rest, but it's all about trends from here on so I'll keep an eye on the next update from this model now. Some cracking charts still being shown on the CFS CC.

Thanks.   That will actually appearing at the end of my autumn forecast video which will be released on Sunday, but don't tell anybody. 

---

I wouldn't normally do this analysis in August, but the weather's not interesting enough at the moment and this thread invites me to go for it :

Analysis of the CFS long-range output for Winter 2013/2014 as of 21st August 2013

http://www.cpc.ncep.noaa.gov/products/people/wwang/cfsv2fcst/htmls/glbz700e3Mon.html

CFS has been trending away from high-latitude blocking in November towards what is a very zonal setup with above average heights to the south and below average between the UK and Iceland. Near average heights over Greenland mean that a few toppling highs bringing back-end northerlies from storms could occur.

December looks like a month of transition, during which the Atlantic starts to dive south… but it’s impossible to infer when. There seems to be a signal for strong HP to build north through the UK (above average heights over us), which would coincide with the Atlantic diving south. The position of the negative height anomalies suggests that the change occurs sooner rather than later. It’s worth noting that some individual runs have developed full-blown high-latitude blocking for much of December.

January has seen a lot of epic blocked runs and this is reflected in the average chart. It looks like the model still has quite a number of runs blasting the blocks away, but with anomalously high Arctic pressure a common theme, as it is for December. The majority take the jet stream well south of normal and Europe looks to have some very disturbed weather at times.

For February, the Pacific anomaly catches the eye. Who knows what’s driving that one…? The signal for a southerly tracking jet remains, but there is no indication regarding the Arctic pressure pattern, accompanied by only subtle hints of blocking near the UK.

Overall, CFS is showing a winter loaded with potential, and recently the number of epic runs for the Dec-Jan period has been notable.

---

 

http://www.bbc.co.uk/blogs/paulhudson/posts/whats-behind-the-coldest-winter

---

Thanks for that interpretation, SC. But I have a problem with trying to tease anything out of those CFS z700 anomaly charts which you posted. It's important to remember that they do display anomalies and not actual high or low pressure. 

What I mean is that a positive anomaly to the north of the UK, for instance, might still indicate low pressure, but of a shallower depth than average or perhaps a low-pressure system centred over Iceland rather than Greenland.

---

Ill say ive used the CFS since about 2009, and its always been scarily accurate. I remember using it to predict the winter of 2010 in October. 

---

http://www.meteociel.fr/modeles/cfsme_cartes.php?ech=4&mode=3&carte=0&run=10

 

If the above(as showing now) came off in January,we would certainly have something to talk about!

I know what you mean, and in the end some assumptions have to be made, for example that zonal conditions in winter don't tend to involve weak troughs and shallow ridges. It comes down to trying to interpret what the model has to project in order to produce those anomalies, using various typical pressure patterns taken from past experience. Unusual patterns are almost impossible to forsee even just one month in advance.

In my experience, a weak +ve anomaly usually suggests a period or two of high pressure in a given region (based on what I've seen verify) while a strong anomaly points towards a month dominated by a blocking high.

The CFS model varies a lot from run to run (November and December keep swinging around from epic blocking to very zonal), so we end up looking for what keeps cropping up more often than not. Usually that means just one or two areas of key interest, for example a lot of recent runs have shown a spell of extensive high-latitude blocking sometime between mid-November and towards the end of that month.

...then comes the probable bias that models tend to show. CFS often seems to either flatten a blocking high with lightning speed or go completely nuts with a raging jet tracking well south of normal and generating truly sensational charts. In many ways the 'average of outputs' charts are by and large a reflection of how often CFS is going one way or the other, with not much in the way of laid-back, non-energetic charts... yet they have proved to be a remarkably effective forecasting tool over the past couple of years or so.

---

As we've seen in previous winters, I think the behaviour of the stratosphere will be key. This may be especially pertinent given that ENSO values look like they may well be close to neutral. Last year we saw a signifficant warming event towards the end of December and into January which preceded a signifficant pattern change from the zonal conditions we had been experiencing up until that point. Even the BBC weather website had a video explaining the effects of a warming stratosphere on the troposphere.

 

The CFS seems heavily in favour of a -AO at the current time. This is still true even when more zonal charts are being thrown up (as it's possible to have a -AO and a +NAO). I think if this -AO pattern comes off we stand a good chance of seeing some very cold spells this year, regardless of the CFS positioning of troughs around our locale.

 

Having the GLOSEA4 model in agreement with the CFS is also a plus. This is suggestive of mean SLP being higher than average from Scandi across to Greenland, centred on Iceland. As good as you could hope for if you're after some wintry bouts of weather this coming winter season.

---

Thanks for that. It all makes sense and I look forward to any further analysis you offer us from the CFS charts.

Good point about nondescript, average charts appearing less often than, er, average in the CFS output. This is a problem, because the long-range forecast most likely to verify is, of course, one which predicts an average season. And CFS unfortunately steers you away from that.

For that reason my attempts at a long-range forecast for winter are still stuck in the Dark Ages: just predict some mild, wet spells and some cold, dry spells and you won't go far wrong, though, quite rightly, nobody would be interested. The one exception I'd make is that to get an epic cold spell in the UK the record of the last century, inadequate though it is, suggests that the Pacific Decadal Oscillation needs to be in its approx. 30-year negative phase. Epic cold spells are still unlikely, mind you, but they at least become possible. As the PDO has been negative for a few years now, then I get as interested as anyone when the CFS runs throw up plenty of high-lat blocking, since the background signal lends the scenario some plausibilty. I just don't expect it, that's all.

Thanks again.

---