SnowBallz

I honestly can't stress enough how it's wise (and better for your health) to stop being so reactive to the outputs of each of the models. You have to accept that there's going to be large scale cross-model variance; one extreme to the other, and often in the same day. Viewing this thread the last 48hrs has been hilarious; it's such a dramatic soap opera! What will the 18z show? Who knows, but how about we view it in the context of how much variance we've all seen across all models, over the last 48hrs, ie: try not to over-react to it. Obviously that's easier said than done (for both extremes) but I'm just trying to explain how these up's and down's are all very likely to happen, until such time as certain variables are more defined. Nick, superb commentary again - first class stuff

I think what the suites are perfectly illustrating - in glorious technicolour - is the current level of material uncertainty; on the one hand you have outputs from GFS predicting what would be a quite epic cold period, but which is tempered by the relatively more mundane ECMWF. To a degree, it's from one extreme to another, and all in the space of, what, 3 hours or so?This is why I'm not an advocate of hanging off the back of each run, of each model; I've seen them make utter fools of quite a few on here, far too often. But trends can definitely be observed from them. I'd say current trending from the MRF models is increasingly towards colder outputs, and this ties-in closely to LRF signals which are favourable to this solution. To that end, the outputs do not come as much of a surprise to me, and neither - it has to be said - is the material variance, as outlined above. Can I also add, that I think we've had some quite fantastic commentary on this thread over the last 10 days or so, and many of you are a credit to the science! Obviously each of us has a different preference, but I think it's nice that we all agree on one thing: we're passionate and interested in the weather, and may it continue to be as interesting and thought-provoking as possible! As I said a few days ago; we've waited a while for cold to arrive, but it is gonna be worth it ... For humour, I've renamed GFS to: Getting Fairly Stressed Anyway, bedtime

With regards to whether specific runs tend to trend more milder than others, it is a belief which is commonly repeated. I don't personally do any re or cross-analysis to validate that, so nothing really empirical to go against. That said, if you accept that the different runs must have slightly different base input variables, then I would imagine it plausible that runs are differentiated by consistent subtle +/- variable characteristics. You then let the computers crunch the maths, and then human intervention is to assess whether the subsequent output is plausible against upstream patterns which you 'know' (to all intents and purposes) IS going to have a material effect. Therein, it's often suggested that the 6z is very mild and the 18z is at the other end of the scale. Implicit within that, it's not improbable to promote that those two runs are the boundary thresholds, and that the runs in-between are probably a blend of the two. From a scientific perspective, that then gives you four different outputs covering a good range of variable tweaking. I doubt there is any great emphasis based on any particular run; I honestly believe each output is taken on its merit and assessed accordingly. But, as I said yesterday, mathematical models are not perfect and they have to be written with rules that largely discount improbable scenarios. If they didn't, the spaghetti range of outluts would simply be so wide that it would be unusable. Personally I'm really getting into the mathematics which falls behind LRF modelling; similar dimensions, but even wider parameters! Very interesting though as you deal in different overarching variables (GP ~ teleconnective enthusiasts would know what I mean) I did say about 10 days ago that I though the first signs for a pattern change would begin to appear in runs towards the end of the first week in January and that is definitely what is subtly starting to show. This is purely down to resolution and pure mathematical reasoning. I fully expect models to deliver more consistent runs with even more pronounced pattern change. It has been a wait, but I believe it's going to be worth it...

Having read quite a few inputs, it's clear to me that there's two types of model-watcher: i) those who feverishly hang on every single output of the GFS, ECM, UKMO etc, and chop and change their 'forecast' like - no pun intended - the weather. ii) those who tend to avoid the 'chop and change' method, and place their reliance more on recurring or consistent long-range trends which imply signals. If you like drama in your life, then - at the moment - chose option number 1; because there's plenty there to keep you occupied. Those of us who are firmly in the number 2 camp are happier with outputs, as there is an encouraging level of consistency. I also think its worth bearing in mind that mathematical models are not that fantastic at applying credence to signals which go against their pre-determined bias. At the moment, the implied background is so far upstream, that - actually - the models are kinda doing the right thing, by discounting any major pattern change. After all, if mathematical models were that reactive, then they'd be all over the shop, all of the time. Computers are wonderful things, but they can't do everything. In the short term, certainly they can be used as tools - but, in the longer time frame, I think it wise to revert back to old fashioned human intervention; old skool forecasting, if you will. It IS all interesting of course, but I think those in the number 1 camp should realise that, I they chose to follow every whim on the models, then are likely to often be very disappointed. Personally, I don't and wouldnt really encourage it.

PL, why do defensive? He hasn't - in my reading - been "critical" of you; he's merely asked that you reference to the evidence which suggests or corroborates that which you've attested to, ie: the colding trend.As this is an evidence-based science, I think that's a quite reasonable request.

An 'expert' who tears up their forecast, isn't really deserving of the term 'expert', in my humble opinion. I honestly don't understand that bit. Personally, when I 'forecast' something, I stick to it. I don't change it as it turns out I'm wrong, because - if that's acceptable - then what's the point of 'forecasting' anything; because anyone can do that! Those false prophets whom scrap their 'forecasts' would do better to re-analyse their original data, and see where it was that lead them to make false assumptions and thereafter, predictions. Scientifically, this would be a more productive, constructive and progressive approach to the science which we're dealing with here. As for the models of late... I think the problem is that, although a pattern change has been mooted amongst discussions, I don't think that it's a change which is likely to filter through into model algorithms over night. It simply doesn't work like that. So, to those eager to see such rapid change, I would urge caution and patience. For, otherwise, you're falsely amplifying an inference which, for example, GP proposes. GP supposes a change, but you're ignoring the context and caution which he's wrapped around that conceivable change. That's wrong, and that - I believe anyway - it why there is a lot of glum faces, as each revolution of the models are output. Currently, anyway. I personally think the various models are only going to start showing notable changes in runs which take place in the 2nd week of January. Why? Simple: the upstream pattern - which is theorised - will then be within a more mathematically sound range, and the calculations which are run on that dataset will have a sounder footing. Personally, I tend to avoid GFS as its verification evidence proves it is not a reliable model. That said, others are more than free to follow it's daily runs if they so wish, but - again - I would urge caution when, especially, looking into 'FI' on it. Last winterr, I found that BOM had a good handle on upstream signals, and you must always weight both ECMWF and UKMO, as their concentration, patterning and resolution is preferable over the GFS - that's just plain common sense really. CFS is quite fun too. I'm currently writing a paper on its performance as a case study - great fun Happy Model Watching gang

Met Office Provisional Model Upgrade Timetable 2010 • January 2010 Global model to 25kmUpgrade Global ensemble to N216 (~60km) / L70Upgrade regional ensemble to 18km/L70Replace the stochastic kinetic energy backscatter (SKEB1) and stochastic convective vorticity (SCV) schemes with the SKEB2 scheme (MOGREPS-G).Revised screen T diagnosticNew soil ancillaries and van Genuchten soil hydraulics • Spring 2010 Physics upgrade including new PC2 cloud scheme

NAE

+144... Hmmm, pretty potent

Which is the same as the one I posted. You need to make sure you don't cache images, as that hinders your browser from displaying the most recent data. Those uncached would've seen the updated 12z, whereas you saw the out-dated 0z.

Not a bad blast from the ECM I'd say...

Gorden Bennett...

I made errors? News to me. That's never happened before; you do realise that's a world first? It's incredibly important not to make errors, because that makes you look incredibly foolish. I mean, can you imagine if I made them on a daily basis? I might as well not bother getting out of bed. Speaking of which, what the hell am I doing looking at model at 12.35? Nite nite TETIS

I second the motion to ban ~> It's annoying habitual overuse by one particular member, whilst we had the 'SE Cold Discussion' threads, had me googling for local gun-hire shops. Rest assured, my AK-47 wouldn't miss him like the snow clouds were off the NE Kent coastline. With any luck, Darwin will be right, and evolution will phase him out. I don't mind so much for ~> which usually means a nice pretty picture to click on. However, if it's used mid-waffle, then I just glaze over it and look for some more interesting hyperlinks :lol: Oh, and I like both of these ones: :lol: & :wub:. Don't ever get rid of those or I'd be lost!

Their servers interpret data which is fed through to them by the Met Office. These servers then deliver 'forecasts' across all of their media, ensuring consistency in their output And yeah, the Met Office DO have their own forecast modelling system. They consider the output of all the available models - but (understandably) trend towards their own one. That's not to say that theirs is always accurate: because it's been proven that it isn't - but, on the long-term output, it more than likely performs considerably better - for the UK anyway - than the others which are out there, which I doubt have such a focus on the British Isles.

Yup, cold evolution very much back on the cards by the looks of things - ECM painting quite a potent one at that. Remains to be seen whether the signal propagates across all models, and remains consistent. Considering the proposed mild scenario was fairly disparate, then you could argue that there is already a degree of underlying relative consistency.

Look at it like people who buy a lottery ticket, then scream blue murder when their six numbers haven't come up: "eh?? how can this be; I bought a ticket, therefore I should win! x 1,000,000" If there were only better appreciation of odds and likeliness, then there wouldn't have to be such public self-flagellation. Considering that, these people pretend to interpret science, statistics and data - then it baffles me why cold, base probability is seemingly ignored. Indeed, some clown on that thread - I can't remember who as I'd fallen off my chair with laughter - asserted that his staunch belief in an Easterly lay within a special forecasting technique. Wow, sounds interesting: very mystical, tell us more! Turns out, it was a flock of pigeons which he hadn't seen for a while. Who needs data nodes when all you'd need ask is Bill Oddy, huh? I was hoping he'd return from his bird watching and say, "haha, I was only joking" but no - it appears he was actually being deadly serious. Worrying I haven't bothered looking at the model thread this morning, but I can guess the amount of hot air resonating from it would rival any Atlantic flamethrower. I guess now that the models are kinda boring for the fantasists, then that thread will also go back to its 'usual' one thread a day update - because that's what always happens. I take it 'weather' goes on holiday when there isn't snow around - because, unless it's apocalyptic - they're seemingly not interested. This is where I empathise with the folk down at the Met Office who are not afforded breaks from forecasting: they have to do it 24 hours a day, 365 years a year - whatever the weather. Now that's science. Would like to see a bit more snow this winter, but - with the days getting longer now - am also quite looking forward to perhaps an early spring, and some nice, wild weather. I was driving around a lot last week and what I hate about the cold is what it's done to all the roads: they are damaged Unfortunately I also live on a hill (a very nice hill, but still - a hill) and people drive/slide down it when it's icy and use the cars either side as crash barriers. Cost me £300 to have some idiots dent taken out from the last toboggan run, in early January. I also dread to think how much I'll have spent on my heating also this winter, so am quite looking forward to days when I don't have to have it going full blast. These are all the knock-on effects of cold weather, so - in a ways - there's a lot to be said of the idiom "be careful what you wish for." Anyway, Arsenal v Man Utd today and I hope Arsenal get battered harder than the GFS just has SB's Mood of the Day: Effervescent :wub: xx

I'm genuinely disappointed that both the UKMO, and then subsequently the ECM didn't 'fall into line' with what the GFS was suggesting. I'm not surprised though. But then this is where I think it's healthy that people should maybe take a step-back, and assess how the models have fared - but more importantly, how much faith can you afford in a signal which is generated by that same debunked model, for the foreseeable. I'd love for the cold lovers to have their day out at the fair and a congratulatory orgasm - but how likely is it all? Perhaps it's understandable considering that GFS is the only real long-term model (that I know of anyway) so - to a degree - people are kinda devoid of choice, and if you don't have choice - then you don't have standards, or at least a comparative. It's tempting to go onto the thread and 'haha!' a few people, but I didn't intend on the other model thread to mutate into what it eventually did, and I genuinely regret that. In any case, it'd only be hypocritical of me, as - even though the models are veering away from an Easterly - it's still only 'one run' from both UKMO and ECM (and now a subsequent GFS run) so it's not enough of a robust correction, to completely discount it. Well, I don't think so anyway - but what do I know. Worryingly, I read how someone has spotted - wait for it - an Easterly deep in FI. I sigh as I type that, but hey - as long as they are having fun, then I can't - and I don't - wish to spoil any of it for them. I agree, it'd be a very boring forum if we didn't fantasise somewhat, some of the times. More worryingly though, is that it's 11.30pm...on a Saturday...and I'm sober as a Judge :lol: See, this is what happens when Mr Flu comes and knocks on your door: I get cranky, and faaaaar more dangerous than any I.E.D. I wish I had some alcohol right now Oh and (not that any of you are the slightest bit interested...) my team - the Mighty Spurs - conspired to throw away yet another two points in our quest to win the league finish in the Top 4 And to top it all off, Keano might be leaving Hmm. Mind you, at least the best looking guy on the forum said ello to me today, so that aint so bad - I can sleep with a smile on my face now :lol: Jesus I'm boring

Stop trying to make a martyr of yourself. You were the one who chose to have a go at me - it wasn't me.

EXACTLY! I have to admit, I do laugh when I read some on the model thread asking for belief or some other rubbish - as if sitting there with their fingers crossed is going to change the weather :blink: I wish it did, but unfortunately we can't control it at all! When I see super computer models not being able to make head nor tail of it, that's when I wonder why us humble little humans torture ourselves so much over it. Yes there are patterns, but they are incredibly general - and that's all. Hate posting in that thread anyway, 'coz my posts just either get moved or deleted. They seem to hate reading anything which sails against their faux tide - and I stand out like a sore thumb. So I just lurk and that's it. On that thread, am now being asked to quantify why I don't think Doomsday is going to happen. How hard is it to explain that these models have an appalling record, and thus their output (at mega long-range) shouldn't be trusted? I don't mind checking out charts/models/whatever out to +72, but anything beyond +92 seems like you might as well be playing a game of Russian Roulette :o lol

Whoops, sorry I meant TITIS or whatever his moniker is. Everyone uses acronyms on here, can get very confusing at times.

"Signals". Signals change all the time - surely I don't have to tell you this? The "signal" for today can (and does) vary widly to what the "signal" might be suggesting in 48, even 24hours time. All the NAO and AO numbers mean is that it is more favourable for the Jet to sink South and colder air to follow with it. But that's it - it does not mean de facto cold spell (which many seem to think) The Atlantic is weak? Says who, says you? Says anyone who wants their cold forecast to verify, that's who. They said the Atlantic was weak in early January, and that - because of this "fact" - then the blocking would remain and "reload after reload" would ensue. Oh, but what happened? The Atlantic kicked it out of the way. I guess it wasn't so weak as the fantasists convinced themselves into believing it was. How can you seriously reference possibly the one and only time when something waaaay out in "FI" has turned out to verify, as good cause to dismiss the thousands of other times when it's been spectacularly wrong? A success ratio of 1:1000 isn't all that great y'know. Go to the toilet often enough, and eventually you'll have a poo. I've already voiced my unimpressed view of 2c mean ensemble data in another thread, so shant repeat myself. Oh I know it's going to be cold - but hey, it's winter - it's supposed to be. But I guess this cold - our normal, quite dull BI cold - just isn't exciting enough for some people. Unless I start seeing some of these fantasy charts verifying at +72hr range, then I'm just not gonna bother worrying about them. I mean, isn't weather about what happens outside - not on a computer model anyway? I thought it was, maybe I was wrong though.

Why "well done," when none of this has verified yet? If we've learnt anything, is it not that the chances are that by tomorrow, it'll be a completely different picture? I don't get this whole need to congratulate, when there's nothing to actually congratulate. Obviously it'd be nice if we had an Easterly - because that's a bit more exciting than our usual run-of-the-mill Atlantic rubbish. But, until we experience otherwise, then I think it's healthy if all model watchers remember what's the dominant feature to us. It just sounds like people are so desperate for something exciting to be within the models, yet why look for the extraordinary when, 9 times out of 10, it never verifies? That's gotta be incredibly depressing. Oh and, BFTP, I said of the last cold spell that it was incredibly stupid for people to suggest that the "UK would shutdown." Which it is, because it didn't. It's called exaggeration, which - ironically enough - is exactly what some of the models tend to do when they find something odd; it gets amplified out of all proportion. However, are these ridiculous, stupid charts ever dismissed? No, instead people weirdly save them to their hard drives. I mean, WTF??!! And no BFTP, a 2c mean doesn't signal - to me - anything extraordinary at all. What it suggests - to me - is that there is quite obviously rather large discrepancy between peak and trough values, across the runs. It's cold - I'll grant you that - but it does not signal 'End of Days' scenarios which I read some try to continually push. But we shall see, won't we - that's the beauty of the weather. However - as I said - I fully expect 8th Feb (or anywhere either side of it) to pass without any note whatsoever, and yet - as sure as night follows day - they'll be charts posted up dated 22Feb10, promising - yup you guessed it - Easterlies. As I've always maintained, I'd love to be proven wrong - but I just feel that a healthy dose of realism is needed here. Let's make an informed, educated critique of the models, not slavishly follow whichever path they meander down. If something moronic is being suggested, then let us proclaim that it's moronic - not try and build a case for it, only for it to be blown out of the water, as resolution tidies it up. Because, for me, isn't good meteorology as much about disproving a scenario, as it is building a case for the unimaginable? I see a lot of the latter, but when someone tends to favour the former, they are dismissed as mere pessimists. Well, they could easily counter by accusing their 'adversaries' of being fantasists, couldn't they?

Nothing spectacular about them at all. Am I surpised? Nope.