Jump to content
Problems logging in? ×
Snow?
Local
Radar
Cold?

seb

Members
  • Posts

    609
  • Joined

  • Last visited

  • Days Won

    1

seb last won the day on August 8 2022

seb had the most liked content!

1 Follower

Profile Information

  • Gender
    Male
  • Location
    Singleton, Kent

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

seb's Achievements

Community Regular

Community Regular (8/14)

  • 1000 reactions received Rare
  • 500 items posted
  • 30 days in a row
  • Dedicated
  • Week One Done

Recent Badges

1.2k

Reputation

  1. Someone shared a document with me in the summer but I can't find it now. It explained that the SST data used for the daily model runs is normalised data which comes from SST forecast rather than raw data. The current way of normalising the SST data, if I remember correctly, outputs values that are too low compared to actuals. Don't quote me on it; maybe someone here has more details. Having said that, just looking at the temp maps it would appear that both ECM and GFS simply do what they've always done and add a couple degrees for a very narrow strip along the immediate coast. To me that doesn't look like doing much with the fact the the SSTs around the UK are currently several degrees above where they should be.
  2. Of course we can return to this at some point on Monday I've explained in a few of my subsequent posts why I don't believe the ECM Op raw figures for Monday morning (or indeed Sunday) - GFS sits comfortably several agrees above; more than half of the ECM members are above; high SSTs are not taken into consideration (along with several other factors) and the ECM more often than not puts nighttime minima several degrees lower than actuals, especially in autumn.
  3. Well quite. And the elephant in the room that so many don't want to see is that this isn't just shown on the most recent GFS run but also the previous one... and the one before that. And the ECM. That's what I mean I say that there's wide confidence levels and that this is when you switch to probabilities of what is more likely on balance. The mean is in sync with the LT average or above, apart from a very short period starting Saturday. The scatter is off the scale starting from Tuesday next week.
  4. Of course it is a different matter but that is exactly what doesn't seem to be understood. I don't know how much clearer I can explain it - there is currently nothing to indicate that there will be widespread out of the ordinary cold 2m temperatures. There may be two mornings where some areas will experience temps which are a little below average for this time of the year but even that is far from guaranteed. People seem to forget that we are an island which is 300 miles wide at its widest point. SSTs are still higher than they should be and that will affect 2m temps and is not accounted for by models.
  5. For 1 or 2 days which is exactly what you are showing; and which is in line with MetO, and with what I (and others) have said. It's also not "much colder than average". London October average night time temps are 8 degrees and average daytime temps are 14. Even on Sunday which so far looks to be the "peak" of the "colder" air, the low for London is 5 degrees and the high is 14 degrees. Add to that the ECM more often than not underestimates night time temps and it could well be 24 hours with bang on average temps. As for frost - a few areas north of Manchester on Sunday morning and the same area (roughly) again on Monday morning (unless you want to take the ECM for Monday morning at face value where some patches show up across the central parts of country... but this is of course not supported elsewhere).
  6. You are free to disagree with my assessment, just like anyone else. This is however a discussion so I don't see the need for your labelling of posts as misleading. If you wish to nail down your forecast despite everything shown, then that is your prerogative. The generally accepted definition of "wide confidence level" is higher uncertainty. Where uncertainty is high, as it arguably currently is x-model at a much earlier point than what we are used to (and as it has been for months), it is foolish to to state that x will happen on Monday, which is what some members have done and which is what I referred to in my post. Whenever such a scenario presents itself then forecasters apply terms such a "likely", "probably" or something like "more likely ... than"; they often also add "confidence is low" or "confidence is lower than usual". At this moment in time, it is almost certain that temperatures will drop to more seasonal values after Friday, albeit even at this range it is not clear whether they will be below October average for a time or not (this is more likely in central and northern parts). Any output for the period post Monday currently suffers from increasing uncertainty, between models and also from run to run as demonstrated by the 00z ECM vs it's 12z run; this uncertainty increases more than usual the further into the modelled period we go. This uncertainty is perfectly reflected in the latest MetO update for 15-24 Oct from this morning which states that temps (nationwide) will be a little below average at the start of that period, rising to little above average. The extended outlook (through to 7 Nov), although from yesterday afternoon, states that temps are generally expected to be a little above average with frost less likely than expected for the time of year. While you may disagree with that assessment, I happen to fully agree (and I'm not someone to hold back with criticism of the MetO). As it stands, temps will drop but the likelihood of anything cold is low. In fact I wouldn't even be surprised if save for some areas in Scotland and possibly the very far north of England, we will not see any frost even on the "coldest" nights in the next 7-10 days.
  7. My eyes are wide open and they see wide confidence levels, average or above average temps and when looking at ensembles and comparing models a whole lot uncertainty across the board. Whoever wants to be foolish enough to make a call on what the weather will look like exactly from Sunday onwards, be my guest. I’ll stick to cooler, somewhat more unsettled but unlikely to be cold or below average. Again, probabilities, confidence levels and many other factors matter. Personal preference does not.
  8. Oh yes, all FI and more so this year. But looking at that scatter probabilities look to be above average for now. Too early to make any calls.
  9. Cooler yes but nothing to suggest something cold and also not for long. Blended output for my location literally has 12 days of temps between 12 and 17 degrees. That’s far from cold and still mainly above October average.
  10. Which really is what model output, taken together, is showing. Probabilities matter in forecasting which is why I really never understood why some here still take individual runs as gospel… regardless of scatter in the ensembles and constant chopping and changing of the op runs.
  11. Yes there will definitely be chopping and changing re the areas that get decent rain and the amount generally is likely over-egged at this point.
  12. Yes. Although still decent and above average temps southeast of a line Cardiff - Birmingham and not much in the way of rain until next weekend. So let’s see where we end up. There would have been times where 17 degrees with sunny spells on 13 October would have been considered amazing weather.
  13. Yes potentially a lot of rain for large parts of the UK and surprisingly good x-model agreement. Obviously a lot of variables at the moment that may change the outcome but this scenario has a much higher chance of actually happening than the apparent cold wave some got excited about
  14. @ANYWEATHER as you disagree with my post, would you point to any output which contradicts what I’ve said? Just so I can make a final decision on whether to take your input seriously going forward or if better to disregard (which would be a shame).
×
×
  • Create New...