Isolated Frost

9c after a max of 14c. Very cool at the moment, marked drop in dew points as well, current a notably cool 3c DP temperature. Clear with a few clouds and very breezy.

Durham at 13.45c, 0.35c above average. Daytime temperatures will go below average after tonight.

Stockholm / Helsinki / Sapporo / Geneve / Brussels / Lyon / Athens / Los Angeles / Naples / New York City / Montreal / Boston Jan/Feb/Mar/Apr/May/Jun/Jul/Aug/Sep/Oct/Nov/Dec

I think Sebastiaan/Cohen have a strong point; but mid-Sept growth is a positive for me, looking at recent yrs- I think November looks a good bet for early blocking on the AO/Ice/QBO comps- snow advance accompanied by shifting dynamics, and an increasingly weak sun, come the equinox, snow cover becomes increasingly relevant. I think the Strat/Tech threads are becoming increasingly important now- with a big focus on teleconnections and snow/ice.

11c, cloudy, pretty windy- surface wet after earlier rain. Overnight low of 6c.

Sep-Oct-Nov all good; I think this'll promote Nern blocking come Nov.

Another pretty cool day- some cloud in the afternoon, but clear during the morning and evening, a high of 15c. Currently clear skies and feeling very, very nippy outside; 7c.

Nope- just more tabloid rubbish about the MetO, who have really impressed in recent years, and probably have the best forecasters in Europe.

PV, most important feature of our winter in my opinion (amongst many others though), looks very active, which for me is a positive looking at composites. If there is any time you want a strong, active vortex with cold energy just spilling out, it's now. cold arctic high across the Russian arctic should help the ice recover, and large snow events across eastern Siberia and arctic Canada likely; also a potent N/NW flow across Iceland and Britain sharp -AO if you've ever seen one

Would love a cold and frosty November-December period with some snow, but I think I'll be let down- plenty of hard frosts will do. Jan and Feb can have big snowfalls and glumness/darkness to match my mood when I go back to school, and then March onwards, warmth and sunshine por favor.

Much of the next 7 days is dominated by mP airstreams; with temperatures near or just below average, especially across northern areas. Plenty of rain and sleet/snow for this chart in Scotland; the first time the -5c 850 isotherm has appeared on the North Sea stamp- the chance of high pressure building on later, but this weekend, and then an interlude next week look particularly cool, wet, and windy for most.

My composites say November cooler, December average-mild, January cooler and February much cooler.

I don't like spiders, flies, slugs, whatever. Major trust issues with them.

Looking at AO modelling; we seem to be spiralling into a potentially neutral-negative September. Accompany that with the August oscillation- and we're looking at an... Aug Neutral-Sep Neutral (1984, 1985, 1990, 1993, 1997, 2000, 2002, 2003, 2007, 2008) Aug Neutral-Sep Negative (1996) Aug N - Sep N - Oct P (40%) 1985, 1990, 2008 Aug N - Sep N - Oct N (40%) 1984, 2000, 2007 Aug N - Sep N - Oct N-ve (10%) 1993, 1997, 2002, 2003 Aug N - Sep N-ve - Oct P (4%) Aug N - Sep N-ve - Oct N (4%) 1996 Aug N - Sep N-ve - Oct N-ve (2%) Based on current composites, I'd say 85/86, 90/91, 08/09 are the most likely winter composites to come up, and then 84/85, 00/01, 07/08.

November looks interesting even when you take out the projected +ve heights to the NE and nearby polar jet activity- though it is much more conductive to more N/NE rather than E/SE flow if October does come out as +ve AO month. I think I'd go 40-60 for the October CET to be below average, and 70-30 for the November one to be. I think December will be pretty dry, but all signals are for a pretty average/slightly milder start to the season.

FWIW, based on my more updated August-September-October projections and modelling; November looks to be cooler than average.

Milder than of late, 9c. Cloudy and a touch breezy.

Indeed, I'm sure Chiono, Lorenzo and many others will have plenty of technical insight on the upcoming winter as well though.

The first projected snowfall for lowland Svalbard tomorrow, with a veil of cloud bearing rain and snow moving from the SW during the day, by the late afternoon, Longyearbyen should be seeing its first snow of the long winter ahead. Could be quite a snowy evening up there for them- you can track it here: http://longyearbyen.livecam360.com/flash/main.php

My AO experiment, round 1: based on the August reading, September and October projections- this is my VERY preliminary winter forecast. My 3 composites on August, September and October values are the only 3 of the past 30 years, and are all pretty consistent. Note the 3 composites are 65% likely (August definite, September 80% and October 50%, by mid-Oct I can make a much clearer forecast). December about 40% confidence as of September 10 Milder and drier than usual- with the north likely to be wetter, and western areas slightly cooler with a more +ve height outlook. January about 30% confidence as of September 10 Slightly cooler, potentially much cooler at times with +ve heights building to the NE, and a negative tilted polar jet- the likelyhood of battleground situations increase with an increasingly influential continental block against the atlantic. February about 15% confidence as of September 10 Confidence is much lower at this point- but on composites and projections, an extreme month with E winds prominent, a strong +ve height anomaly to the north, and -ve anomaly to the south, with temperatures much colder than the average, especially the further south and east you are. That's my very early attempt of a forecast, I'll update it in a months time (if September doesn't turn out like expected), and then make a final one in early November. Mild and dry start, increasingly cold and active as we move in to 2014 are the early thoughts. Composites that include Feb 86, Feb 91 and Feb 09 definitely make you think.

1990 is cropping up here! I think that the possibility of a strong late season event is strong based on pure arctic composites; good to see that the QBO strengthens that up; will have a read of some of these papers on it I think.

As I said, it's just an experiment on Arctic Oscillation, I used the composite of neutral August (which has occurred), and neutral September looking likely on the basis of -0.5 > 0.5 as neutral values. If you have a look on the in-depth technical thread, here http://forum.netweather.tv/topic/76236-in-depth-model-discussion-and-summaries/?p=2784257, you can see the September neutral / October +ve composite; which I'm sure you won't mind!

My prediction for October and November- based purely on Arctic Oscillation; I will do a pre-winter forecast for Nov-Jan next month and then a full winter forecast in November. This is just an experiment on the factors of the AO- Going for a slightly cooler month, especially in the south, and drier, especially across the north during October. Strong GPH block link-up between Atlantic and Arctic, accompanied by Europe troughing- a cooler, drier month with frosts likely. What a month this looks like it could turn out to be! Big +ve anomaly to the north/north-east, and very cold temperatures for Scandi/NE Euro- warmer across much of Europe and UK though with S flow and -ve anomaly to the SW of UK. Definitely interest from me, large cold segment towards Scandi provides encouragement for cold pooling later on. Drier, cooler October, frosts and sunny days. Turning wetter and milder in November, as the continent roars up a beast.

September neutral and October neutral Alternatively, if September stays on course and October goes neutral as well... have yourself a Euro high and a half. Stream of mild S/SW'lies with big -ve heights o/ Greenland +ve heights o/Europe... coldies look away! September neutral and October negative ...or, October goes negative AO wise, and we see a pretty wet and stormy winter in general; but with northern blocking around, the jet drifts south and we're in for cold. Generally, if September is neutral (which is likely), and October is neutral - expect a wet but mild winter. If September is neutral and October negative - expect a mixed winter, with some wet and mild and some cooler and drier weather. And if September is neutral and October positive - expect a wet and stormy start, and then something potentially special in January or February. Will definitely make a winter forecast once the values for Sept and Oct come out, will be interesting to see whether these composites based on only one value can be so consistent and/or reliable. I'll do the September positive-October... and September negative-October... composites in a bit. And maybe do some monthly ones later on (for Nov, Dec, Jan/Feb).

Right then, now here's one for you. Some of you will LOVE this I'm sure. Say, September ends neutral (likely), and October ends up positive (we don't know yet). 4 years have done that, 1985, 1986, 1990 and 2008. Anything? 4 years in the last 30 have had a neutral September AO and a +ve October AO, 85, 86, 90 and 08. February 86, January 87, February 91 and February 09. Probably some of the most extreme snow/cold events of the last 30 years have all occurred during this scenario, which is a really weird turnout! It appears as though if September stays on course and the Arctic has +ve AO during October, we could end up w/ another big event. One of the more interesting composites about... maybe a wet and stormy December followed by a continental blast in January or February?