Thank you Santa - that's just the ticket: 1987 eye candy under the tree, and deserved, as I've been very good this year...

Here goes the roller coaster ride of model watching as we get nearer to mid Jan... But I love the ride.

great timing too...like 1987 it was the perfect time for a real coldie memory - just need double the length and double the snow depth...not sure Cantebury dwellers would agree with that, but 2 ft will do me!!! Listening still Santa - and I have been very good !! LOL

Ops at the bottom in FI....Control and mean stay above average....few cold ens in far far away land...nothing trending yet.....give it a couple of days see where this looks like heading....

Ops at the bottom in FI....Control and mean stay above average....few cold ens in far far away land...nothing trending yet.....give it a couple of days to do where this looks like heading....

depends on your definition of trending, more members going colder is a trend compared to previous runs. Whether it comes to fruition is a different question

 

 

depends on your definition of trending, more members going colder is a trend compared to previous runs. Whether it comes to fruition is a different question

--------------------------------------------------------------------------------------

 

agreed 50% showing a colder trend....but need to see it sustained ....hopefully it's the start....

---

Home : Mid Leicestershire ...135m ASL

 

We have this debate every year - and there can’t be a debate about data depletion (roughly 20K datapoints compared to 35K on normal days). The depreciation on accuracy however is debatable - I’m not convinced it has much effect on the 5-7day range given it’s only around 80% on the verification stats anyway.

This has always been my point; the effect on accuracy, not denial of reduced data input. Even the stats historically available on those links show barely enough difference to poo-poo or as some posters do, discount the output altogether.

 

This has always been my point; the effect on accuracy, not denial of reduced data input. Even the stats historically available on those links show barely enough difference to poo-poo or as some posters do, discount the output altogether.

 

 

I think the reduction in data is significant: over 40% less actual data means the opening approximation of the state of the atmosphere is going to be more suspect than usual. Now bearing in mind that the ensemble suites are designed to test the sensitivity of the model output to small errors in the starting position, and given how we see quite substantial changes in the evolution across those ensemble sets, it seems obvious to me that we are adding a significant layer of uncertainty to an already uncertain position.

GEFS fast improving , some decent charts starting to show their hand 

When you see so many of the purtabations showing similar patterns, I start to get a tad more interested. The latest GFS is snowing several similar mid Atlantic blocking scenarios, with screaming northerlies, with embedded polar lows. Remember that one that dumped a shed load all down the west side from Birmingham to Bristol, must have been the nineties or naughties? I was expecting the scenario shown on the 12z op to have gone for a day or two, but it seems to have multiplied from the control to many more versions. Whether this is data mehem or an emerging trend, the eye candy is most welcome.

When you see so many of the purtabations showing similar patterns, I start to get a tad more interested. The latest GFS is snowing several similar mid Atlantic blocking scenarios, with screaming northerlies, with embedded polar lows. Remember that one that dumped a shed load all down the west side from Birmingham to Bristol, must have been the nineties or naughties? I was expecting the scenario shown on the 12z op to have gone for a day or two, but it seems to have multiplied from the control to many more versions. Whether this is data mehem or an emerging trend, the eye candy is most welcome.

Just a minor point: AFAIK the models still don’t pick up true polar lows until at quite short range because they’re mesoscale features that develop in highly unstable Arctic flows.

ECM 12z ensemble suite for London - 2m temperature:

It’s interesting that even by Boxing Day two distinct clusters are developing, one giving slightly above average temperatures (10-11c) and the other below average (6c). The op starts in the milder group but drops by 28th into the cooler group.

Reasonably consistent with the 00z beyond Day 10 with a clear cooling trend and some cold or very cold options appearing and a near-absence if mild options.

When you see so many of the purtabations showing similar patterns, I start to get a tad more interested. The latest GFS is snowing several similar mid Atlantic blocking scenarios, with screaming northerlies, with embedded polar lows. Remember that one that dumped a shed load all down the west side from Birmingham to Bristol, must have been the nineties or naughties? I was expecting the scenario shown on the 12z op to have gone for a day or two, but it seems to have multiplied from the control to many more versions. Whether this is data mehem or an emerging trend, the eye candy is most welcome.

I think you are wishing a bit too much judging from your Freudian slip 

 

I think you are wishing a bit too much judging from your Freudian slip 

GFS operational 00Z is a snoozefest until the end of the run but then suddenly shows a major northerly incursion from T+360. 

Don't get excited,  'he only does it to annoy because he knows it teases'

 

I think the reduction in data is significant: over 40% less actual data means the opening approximation of the state of the atmosphere is going to be more suspect than usual. Now bearing in mind that the ensemble suites are designed to test the sensitivity of the model output to small errors in the starting position, and given how we see quite substantial changes in the evolution across those ensemble sets, it seems obvious to me that we are adding a significant layer of uncertainty to an already uncertain position.

only 40% less data doesn’t equate to the minute difference in accuracy score. This is why I think the whole thing is not as credible. 

A ~10% drop isn't minute.

As I say, it is of course your right to ignore what's been shown to you (and backed up with evidence these past few years). If you want to convince people, though, you need to do your own research to back up your views. That's how science works!

You'd best get a move on if you want to start this year, though, as the data loss will kick off in around 24 hours!

GEFS certainly not as good as last nights