Interitus

This was even more noticeable with your previous comparison of Nimes and Ajaccio with the latter influenced by onshore flow from a 23°C Mediterranean. Also noting the very dry Saharan air at altitude then increasingly to lower levels, with a breeze things should become less oppressive.

Largely fine, but of course much cooler.

Look exactly like that from north Cheshire right now, with the same wave pattern - shows just how high up they are. They were good earlier in the week, but with the clearer maritime air that has come in today they are stunning.

There is a vast difference here between the operational and ens in the near term when they should be pretty much identical. Suggests that possibly the values for London are approximated from different grid points limiting the usefulness.

Meant to ask whether you captured these this morning, they were fantastic between 2-3 am viewed from Warrington but wondered whether some cloud cover to the east obscured your view - evidently not.

Interesting feature that has been meandering round Scandinavia and northwest Russia since our deep low a couple of weeks ago, and the partial remnants of a tropopause polar cyclone from Canada prior to this (see https://twitter.com/SimonLeeWx/status/1113003251077656576). It is currently over Belarus and the southern Baltic states and is due to circle the Scandi high before hitting the UK next week in the above charts.

It had no effect whatsoever. The total energy released was less than a quarter of that released by an average hurricane in one second. There was a similar discussion regarding the larger Chelyabinsk event in February 2013 and in terms of atmospheric energetics the values are miniscule, but some still seemed unconvinced. Simply examining temperatures from radiosonde readings from the nearest stations show no deviation at the time of airburst at 25.6 km altitude (roughly the 20mb level).

Have to disagree with this, a degree of climatic persistence is evidenced in the data, as should be expected from the duration of patterns of low-frequency intraseasonal variability. From a quick look at the last 100 years 1920-2019 (arbitrarily chosen, earliest CET values probably poorer quality, and note data should be detrended for a full analysis) there is a weak significant correlation between January and February CET of 0.36 (it was 0.41 for 1919-2018). If randomly distributed then there would be a 25% chance of both months being below average, 25% above average and 50% of them being different. 26/100 years had both months below average so close to random. However, 37% had both months above average - so a milder January is more likely to be followed by a milder February than would be expected by chance. Some of these months though may be very close to the average and not register as particularly extreme, and when looking at months that are 1 standard deviation from normal something interesting is apparent. Only once was a very mild January followed by a very mild February (1990) and it is just as uncommon to flip from very mild / very cold (1945) or very cold / very mild (1983). However, a very cold January is seeming quite likely to lead to a very cold February, occurring 6 times in 100 years (1929, 1942, 1947, 1963, 1979 and 1985) and on almost half the occasions (13 v. cold Januaries). This may be because of the stubbornness of blocking patterns perhaps. Finally, from the linear relationship between the months, using January as a predictor of February has a RMSE of 1.73°C which is marginally better than predicting purely by the February mean, RMSE 1.89°C. The linear relationship may not be stationary though, so its performance may differ over time.

You're taking the p### now!

What went wrong were the selections for the plots! This is the composite mean SLP for the full period - - and here is the mean 500mb geopotential -

No statistically significant correlation between the winter multivariate ENSO index and winter CET, however since winter 1950-1, 7/10 coldest CET occurred with negative ENSO (i.e. tending towards La Nina).

Succinctly put Knocker. Wondering if I am one of those characters? Extenuating circumstance, unforeseen factor, whatever, the point is this excuses the author from any acceptance that, just maybe, possibly, their assessment was incorrect, or even bunkum. With the Jason Furtado tweet, more telling was the reply from Prof. Paul Roundy - So much for that. The major teleconnection as far as UK and western Europe are concerned is of course the NAO. For those interested enough, NOAA ESRL have handy tables of monthly correlations between various teleconnections. Few correlate well with the NAO - https://www.esrl.noaa.gov/psd/data/correlation/table/ The signals are generally far weaker than they are given credit for and are largely swamped by natural variability. NB. There is no issue with long range forecasts, by whatever method, or those attempting them, but it is bemusing when they are reluctant to discuss or debate the ideas on a public forum and view it as a slight or impugnment, bizarre. And not being a "coldie", there is no axe to grind nonsense. Meanwhile, here is a monster chiclet chart of GFS forecasts of u1060 from 09/12/18 to 31/01/19 - Well more a Hovmoller than a true chiclet with the hours shifted to meet in the vertical but this requires discarding data as horizontal resolution is 6 hours, vertical is 3 hours. The low res 12 hourly increments beyond +240 is interpolated to upscale. Running left to right from Dec 9th 00z, analysis at bottom, T384 at top, positive zonal wind red, negative blue, it captures the full development of the SSW which when signaled by the model in long range from the position of an extant vortex, progresses fairly consistently to T0. It highlights though the apparent difficulty to be certain of long range once the SSW is imminent or occurring with frequent fluctuations between vortex recovery and extended SSW conditions - whether this is real or an artefact of the GFS requires investigation, but it also over cooked the strength of the reversal in shorter range which didn't materialise (the deepest blue tones in the centre). However, when the recovery is finally detected, it progresses consistently from long range, and there appears much less variability with the vortex back in place.

On this point, I posted an analysis last September that November and December CETs actually link well to May/June SST in that general area so a potential 5-6 month lead time. Of the 12 coldest SST in May/June the December CET averages over 6°C, including the warmest Decembers on record. 7 CET above 6°C and only one December CET was below average (2009). With this winter now make that 12 from 13 above average and another >6°C - https://www.netweather.tv/forum/topic/90431-uk-first-ground-and-air-frost-watch-season-2018-2019/?page=3&tab=comments#comment-3908027

Yes, it's possibly style to an extent and it's not restricted to teleconnection based posts. But some will not entertain discussion on what is after all a public forum which is conceit bordering on arrogance. And a literal readout of teleconnections are often expected (ignoring the climate change aspect). Some time back, Knocker linked to a Michael Ventrice video in a tweet examining the downstream effects of the MJO and one of the key points he made was it is not a 'black box' (the exact phrase I had intended to use) i.e. propagation through either affecting the jet stream, or by Rossby waves on a great circle isn't a given. This from an advocate and acknowledged expert in the field, and it can be taken as applicable to any other teleconnection one chooses to mention. Anyway to stay on topic, here is a 3D plot of the vortex defined by relative vorticity from today's GFS 12z analysis of the 1-150hPa levels, coloured by temperature. A fairly coherent structure at higher levels, still more fractured below. Temperatures have recovered in the heart of the vortex with the absence of wave activity allowing radiative cooling such that it has reached daily record minima at 10hPa in the MERRA2 data. Now as mean westerly zonal winds have returned throughout, weak wave activity is returning with a wave 1 warming shown on the Asian side and a tell tale 'hole' in the vorticity over the north pacific representing anticyclonic curvature -

Without singling out any poster, there does appear a tendency in some teleconnection based posts to be stated with overreaching authority and certainty which when failing to come to fruition are because of some unforeseen extenuating circumstance (if revisited at all), not that the original assessment was flawed or incorrect. Clearly, certainty of forecasts and teleconnections are always limited.

Is it?

The nature of 'downwelling' has been described by a number of theories but whatever the downward influence on the troposphere doesn't alter the fact that the pattern of tropospheric heights will be always be seen in the stratosphere to some extent.

Of course the trop imprints up into the strat.

Ah, ok it was well after beer o'clock

The closest 30-day analogues of 10mb 60°N zonal wind to 11/01/19 (MERRA2 data) are 20/02/01, 15/01/04 and 20/01/13 - This chart includes the GEOS forecast period to 19/01/19 during which 2004 deviates somewhat, but 2001 and 2013 remain closest analogues. From these dates, the AO chart to 11/01/19 shows closest similarity to 2013 but this then undergoes a couple of weeks in +ve territory for the forecast period while 2001 and 2004 have some fairly negative AO - Similar to the AO, the lower strat 150mb wind is closest to 2013 whereas 2001 and 2004 weakened from stronger values but notable in the near term forecast period is that these analogues all display more marked weakening than 2019. At this pressure level it is not always clear if the effects are working downwards or up, but if downwelling is suspected, it is quite poor at the moment and in the short term.

The GFS parallel has attracted some attention in recent runs with increased meridionality towards the end which can be seen with reversed tropospheric zonal flow at 60°N. However the interesting thing is that this is led from the troposphere to the strat as can be seen in this plot of zonal wind at various pressure levels - There is a marked dip around Jan 22nd which propagates from the lower trop with a minimum at 30mb 51 hours later. This is not to say that the SSW hasn't potentially caused this in the first place, but concluding that these events are the result of SSW is fraught with uncertainty.

How and where are you measuring this reversed tropospheric flow? It may be locally reversed associated with increased meridionality and quasi-stationary blocks but at 60°N for example, the longest continuous reversed mean 500mb zonal flow (1979-2014) in winter was 7 days up to 25/12/00, at 850mb 16 days up to 31/12/00 - there was no technical SSW at that time (close - u1060 down to ~2.8 m/s on 25/12/00). Looking at 30-day mean zonal flow at 850mb, for January it has been negative in 2010, 2001, 1985, 2011 and 2013 - only 1985 followed an SSW (2013 was preceding). In February negative 30-day mean in 1985, 2010, 1980, 1986, 2007, 1994, 1979. Of these 1985 and 2010 followed full SSW, though of course 2010 was already low as shown above. Flow anomalies may be significant, but there isn't wholesale tropospheric reversal at 60°N as seen at 10mb. The predisposition and variability of the troposphere would appear to be more important.

That's seems compelling, but let's see equivalent anomaly correlation charts from any tele expert forecast.

A few from MERRA (1) data - 18/01/03 1 day 09/02/10 3 days 24/03/10 3 days 01/01/85 5 days 22/02/79 8 days 08/12/87 9 days 21/02/89 9 days (initially - after 5 day break another 15 days) 11/02/01 14 days 12/02/18 17 days 26/02/99 21 days 06/01/13 22 days 24/01/09 29 days Edit: as mentioned previously and as tweeted below, this is not a typical split though, it is a displacement / wave 1 event.

Further to the above the ECMWF have created a "stratosphere task force" to improve forecasts - they have an in depth (very!) report from June 2018 with download link here - https://www.ecmwf.int/en/elibrary/18259-report-stratosphere-task-force An obvious challenge to start with is that verification is against observations which may have significant errors. Less obvious is that increasing horizontal resolution degraded some performance -