Search
  • Luke Boxall

Ghost of Winters Past

As I type this blog we currently have a significant difference between the two main weather forecasting models GFS and ECMWF in the outlook by Day 6 - with ECMWF model flaring up a stronger Scandi-high and turning widely cold over Central-West Europe (CWE) whereas GFS keeps low pressure in charge and thus stays around 3-6degC warmer for UK-France-Germany which makes a substantial difference to the level of demand these countries have as temperature on ECMWF dip widely below freezing, but even within the same model - there is strong run-to-run variability occurring.

Images from WXCHARTS - showing GFS OP and ECMWF OP for the same forecast time

Both these models runs then proceed to flare up strong Atlantic low pressures which then push into W Europe on the back-end of the runs in the day 11-15 window despite the risk of on-going surface SSW blockiness persisting and MJO stuck in the blocky phases of 6-7-8 throughout this month.

Polar Cap Forecast on the left from GloSea (MetOffice) and ECMWF shows the risk of continued surface blockiness throughout February associated with the aftermath of the SSW event in Jan printed by METSET and MJO shown on the right from ECMWF model printed by NOAA

One way we can get insight for the future is to look in the past, and when we dig in the past for a similar set-up to the front-end of the forecast with strong high-pressure signal over Norwegian Sea/Iceland/Greenland but with ATL lows - with similar strength features, the main analog is mid-late March 2013 - which is interesting given the on-going surface SSW effects March 2013 had and the MJO being in Phase 7/8. Using 17th March 2013 as an analogue, Forecast for day 5 ECMWF OP from 2013 shown to the left and current ECMWF OP forecast for day 5 shown on the right (Images from Meteociel)



Whilst it's true that in 2013 Europe was already covered more in colder air - so it's not a good analog for the very short-term - the set-up going into Day 5+ looks similar with a very strong high pressure to the North of Europe but then the forecast shifts into more ATL low pressure by day 10 (2013 example Op at day 10 on the left and recent EC OP example on the right)



Whilst it's true that ECMWF Op isn't the most reflective run to use to represent how all models were behaving and also the model was different back in 2013 - I've compared this with a more recent version of ECMWF ENS mean that's been re-run in the past and the result is the same - a "mean reversal" signal appears as the ATL lows gear up and attempt to flush away the warmth.

The teleconnections are similar - the set-up is similar and the forecasts are similar.................

What happened in out-turn?

In 2013 the ATL lows that wanted to push towards western Europe as a southerly flow - ended up further south and weaker as the high pressure over Scandi-Iceland decided to stick around for longer than expected, thus the CWE/UK region ended up staying much colder as shown by the shift south-wards of the coldest air in the forecast-versus-out-turn below - although we actually did end up very windy for UK/DE/Benelux which helped with off-setting demand for around 3-4 days.



Obviously 2013 is just one year and it's difficult to draw too many conclusions from this - certainty not possible to point at this and say it will definitely get colder - but at the very least you can point at this and say the risk of it out-turning colder is there.

How did a recent ECMWF ENS model version deal with the set-up in 2013?


The 3 images - Germany Temps, UK Temps and France Temps show the recent GFS/EC ENS means displayed in Red/Black and Normal is displayed in Green. The various shades of yellow-brown show the 2013 forecasts in the representative time-frame from oldest (light colours) to newest forecasts (dark colours)

Between the 16th-19th March the ENS means hadn't quite got the full extent of the cold in the day 5-10 window flip-flopping somewhat and still showing warmer/mean reversal beyond the short-lived cold snap even as it got closer in forecast lead time - this represents around 1st-4th Feb in our current set-up.


Then suddenly around the 20th March the ENS snapped colder and became stable with a short-lived cold snap but still kept hold of the mean-reversal signal even though again it was closer in time - this represents around 4th-5th Feb in our current set-up.


It wasn't until the "mean-reversal" event was around 3-4days away did the ENS models suddenly flip colder and in the space of a few daily runs dropped by 4-8degC - this represents forecasts beyond 9-11th Feb in our current set-up.


Again, it's hard to draw too many parallels with just one year - but it is interesting that today's models struggled with a similar set-up in the past - suggesting a mean-reversal/Atlantic lows pushing back in but only when it was much nearer - within the 3-4 day did it suddenly flip and was able to see the risk of continued easterly flow/persistent Scandi-high.

Another interesting feature about the set-up into next week is the AO/NAO - the forecast for AO/NAO both dips very negative into the day 6-10 window - that's even on the warmer GFS model - there are only 3 other occasions the AO/NAO in combo dipped this low in more recent times (after 1979) - Nov 1985 and Dec 2009 and Mar 2013 (pops up again - no surprise) - here's the composite of that week (created from NCEP):



This level of atmospheric blocking is associated with dragging in strong cold over most of Europe - adds some weight to the risk of out-turning colder - but can Ghosts of Winters Past help with our current predicament? Only time will tell.



74 views0 comments

The Weather Perspective Ltd