seb

Yes, Ireland has been shown with heavy rain on several different models and runs. It’s more about what happens further East and South.

Just look at that confidence index!! Red from 3 days out. And of course listen closely to what Alex is saying about 4(!) different models showing 4(!) different scenarios by Tuesday.

Appreciate that but at the end of the day the GEFS is still the GFS just with adjustments made for initialisation (from actual observations, which are used for the general GFS runs). That isn't of much help when the actual GFS differs from run to run at 5 days out and when there's nothing even close to consensus comparing to the ECM. Or indeed when historical data is not very useful because of record SSTs and record heat across many areas of the planet.

So did this week, this time last week Also don’t think possibly 28 degrees on Saturday is standard fare, especially after the heatwave we’ve just had

Considering ECM and GFS 00z runs start to diverge significantly by day 5, I would not be comfortable to make any calls on what things will look like post Sunday (in all honesty, even Saturday/Sunday could still end up different to what it looks like at the moment). It was just a few days ago where there was no sign of anything but "normal" autumn weather from yesterday onwards to the end of every run. What seems to be on the cards now for the weekend wasn't shown anywhere but a small minority of individual members in each run. Of course logic and history imply that we are well overdue to switch to westerlies, showers and temps in the mid teens but...

Looked at the last three GFS runs as well as the other models and I’m sticking with what I said yesterday. No point looking at anything post day 5. So much flip flopping! Up to 28 on Saturday seems to be forming up though which is nice

And quite opposite on the GFS lol. Three attempts to push up warm air up north between end of this week and end of run. Might as well look use my crystal ball. If correct we will of course also experience quite some storm in between which will now also flood the west of Scotland (although not all of Ireland any longer)

Anything can happen but I don’t see it and neither do the longe range forecasts. Not a single month from October where 2m temps don’t look to be above or significantly above average. Not just in the UK. ECMWF | Charts CHARTS.ECMWF.INT Same goes for 850hPa temps. ECMWF | Charts CHARTS.ECMWF.INT Then again… models are struggling at 3+ days at the moment. But gut feeling for me is a stormy period thanks to record SSTs followed by benign, boring and mild winter weather.

Well there aren’t any doomsday synoptics and the issues don’t lie, as they usually do, in what we would consider FI. There are differences, at times massive differences, from run to run and model to model even as close as three days. It shouldn’t come as a surprise of course as models use huge amounts of historical data and there simply is no precedent for the SSTs and a number of other parameters.

I think it’s safe to say that we are getting to a rare one day at a time scenario with the models. The differences between the last few GFS runs even as early as midweek are unbelievable. Let’s see what the ECM shows in the next hour. Going by the 00z GFS, Ireland is getting flooded while almost the entire Eastern half of the UK barely sees a drop of rain until at least a week from today. Latest GFS with two attempts of pulling up warm/hot air from the south as far north as Newcastle where a previous run was flooding cold air down south as far as central France! Let’s ignore the differences in wind speeds

Wow what a difference between GFS and ECM. Model chaos

This is a nightmare for the models and I would agree with some other posters that a large part of it is down to the record SSTs. Spot the difference between GFS and ECM for the 18th (feel free to do the same comparison much earlier than that...)

Not sure if overly optimistic. 32.6 was the official high temp yesterday. For a whole swathe of the SE there are no official stations but even here where we are always a degree or two lower due to a limited see breeze we reached an unofficial 32.1. By Sunday the ground will be even more dry, there's another push of hot air from the South. Wouldn't write it off. It's only about 2.5 degrees more than yesterday.

If the latest GFS is anything to go by, of course unlikely to verify as in deepest FI, then you might be on to something... Storm #1 around the 18th Storm #2 around the 23rd Storm #3 looks like the same track a few days later Gusts of 75mph along the South Coast, sustained 10m winds of 50mph and >200m of rain in some areas.

Always having fun. Sitting in the sun with a cocktail, taking a break from work

Indeed. Language matters "After near record or record daytime and most definitely nighttime temps we are likely looking at something more normal from around Tuesday, although still above average temps and below average rain (possibly some exceptions in the far NW) and generally drier and warmer the further SE you are. No washout, nothing cold, no strong winds and most definitely nothing autumnal (away from the far NW where what’s currently shown is normal)." Where appropriate... The first instance is based on the fact that at least three MetO owned sites exceeded the previous September record this morning and of course also the unprecedented duration of the period where a tropical night is occurring at least somewhere in the country during this current heatwave; the second is because there is no chance of what would be considered autumnal weather during the period to the end of the GFS run.

To add to my previous post, let’s throw in the 00z ECM for the same time and the low to the NW of the UK has vanished.

3 subsequent GFS runs plus 1 ECM thrown in, all with very different solutions. Enough said. Midnight Saturday 16 Sep on GFS 12z, 18z and 00z as well as ECM 18z below. Even picking noon on Wednesday 13 Sep gives similarly differing solutions. So as we were. After near record or record daytime and most definitely nighttime temps we are likely looking at something more normal from around Tuesday, although still above average temps and below average rain (possibly some exceptions in the far NW) and generally drier and warmer the further SE you are. No washout, nothing cold, no strong winds and most definitely nothing autumnal (away from the far NW where what’s currently shown is normal).

If the ECM and GFS are correct then there’ll be decent weather all next week for those south and east of Birmingham. Barely a drop of rain, if any. And still above average temps after the cool-down.

And this is why you cannot rely on individual model runs or indeed anything post 10 days (in many setups I’d stick with 7, right now maybe even 5 days). 6z for 18 Sep 6PM vs 12z for 18 Sep 6PM could not be more polar opposite of they tried.

I think the point being made is that August was not ordinary summer. Or at least not what we are used to nowadays. It's was grey, at times too wet and too cool.

So did yesterday’s 00z and as demonstrated by the Mean, both Op and Control are at the lower, at times the lowest end of the scale. Therefore your conclusion is spot on. 7 days+ is FI, add in that there’s no cross-model consensus for this. NB - my previous post stating the entire run stays above average was for a different location where that remains the case.

Do you see the little part with the teal bars at the bottom left? The one with the percentages? Just to demonstrate a situation where there’s NO cross-model agreement:

As a general rule of thumb, the 5 day output is accurate 90% of the time, by 10 days that's down to 80% and after that you get to down to almost zero very quickly. There are of course always exceptions, both ways, but that would need cross-model agreement or at the very least cross-run agreement. Neither usually applies and therefore we safely ignore anything post 10 days apart from broad trends... which at this point is: Following unseasonably hot conditions with near record or indeed record breaking temps, significantly above average for the duration, we will likely get closer to something more normal by mid month but still more likely than not to be above average for temps and below average for rain.

As I just said in my previous post - "What is shown for Tuesday 19th would, but that is of course so deep into FI that other than taking a quick glance at it, it would be foolish to rely on it." You said that the 00z runs shows "[lows] D10 to D16 and persisting" and now you post output for days 14-16 which I have addressed and will do once again - it is so far into deep FI that it's not warranted to take into consideration other than a quick glance. This is the GFS and 9 times out of 10 it shows something post D10 that never materialises, especially when it comes to low pressure.