SnowBallz

I could be wrong here, but doesn’t earth.nullschool just use the raw GFS data to produce wind vectors? I usually see it being used on various news channels whenever there are gale force winds, and they often mislead by presenting it as live data. The only live data I know of comes in from a network of monitoring stations.

Early guidance forewarning of next weeks weather was issued earlier on, and this has been further enhanced with extended outlooks in bulletins. AFAIK (comms not being my area) public agencies and major private infrastructure will begin a daily briefing cycle as of tomorrow. To revisit previous post, latest MR output (MOG) resolving to further mute ‘breakdown’ scenario; with continental blocking signal reinforcing through well into March. Atlantic incursions are expected with southerly deflection; considerable potential for extensive and disruptive snowfall as these air masses interact. Short term, ECM:UKMO evolution favoured and considered likeliest; corresponds well with other products and illustrates fair consistency. Small ramp. Next week = mega. SB

I think BA is simply reinforcing the point that it’s not really worth attaching too much credence to current output, where there is clear inconsistencies being exhibited. It really is more about broader trend and the eventual path, whereas the journey to get there remains fluid and undecided. 12z EC could correspond with its peers or it could reinforce its 0z signal and advertise a more subtle path. What I would strongly disagree with, is NWP assertion of a quick return to mobility (or milder weather) That evolution doesn't correspond to mid range products at all, and can quite confidently be discounted as models reverting to type and to some extent underestimating the degree of in situ blocking. This behaviour has been seen previously so reasonable to view a rapid breakdown with scepticism. What an exciting end to winter! SB

I obviously can't speak for Ian, but I'd suggest he's referring to the point in time where there is currently considerable divergence and the general thought is that this is due in part to the models acting upon SSW influence and increasing entropy values. Not certain, but it's a reasonable position. Naturally, as one progresses through time, you would expect convergence as depth of data influences the direction of travel with attentive entropy moderation. I'd still wager on a colder synoptic developing - I don't think the question is so much about that - it's more the degree and duration which retains uncertainty.

And that is a key point which can be overlooked when SSW’s come into the mix. There isn’t so much a guarantee of colder weather, merely a statistically higher chance and which of course is still sensitive to many influences. To some extent, this explains why SSW events also broadly correlate with mid range NWP output displaying higher variance, or at the very least struggle to demonstrate consistency. As things stand, the house is very much in our favour. I wouldn’t necessarily pay too great attention to when a run defaults to benign or mobility, as the warming effect isn't really showing its tropospheric influence in the 10+ day range to the degree which we know it would. That is where meteorologists have to use more than mere computational output, and also where the biggest focus has been of recent years to tune mathematical modelling within GloSea in order to influence that judgement. Some might view this as pessimistic, but shouldn’t. Uncertainties around an outcome doesn’t make it unlikely to happen, it just retains it as a possibility. But within the context of probability, the likelihood of the SSW not bearing significant influence is the least favoured outcome. That’s a strong position to be in. To use a dice analogy, with an SSW to get cold we maybe only have to hit odd or even numbers - not guaranteed, but fancy our chances; however, without it, we must hit boxcars. If we look at trends, and specifically touching back briefly on my previous post, all evidence still resolves towards continental blocking patterns establishing latter part of the month with evidentially plausible continuation into March. Big, repeating and in some parts amplified signal for that. So, no guarantee but good confidence. The daffodils won’t know what’s hit um! SB

Indeed. Stark correlative response from stratospheric influence. This is an interesting one as GS usually ranges a warming signal circa >20 days, so immediacy is unusual which is influencing CF’s confidence re LR synoptic. What is without question is that there has been some quite remarkable perturbations, and strongly balanced towards a very much colder pattern. If downwelling is as rapid as is being considered, it is plausible to see large scale continental blocking patterns dominate for quite an extended period, and very conceivably into meteorological spring. Should see some interesting runs this week, that’s for sure! SB

Ha! They're booting Hall 1 up (fire was in Hall 2) so feeds should hopefully be back online by this evening, tomorrow morning at the very latest

Which is a very fair assessment. I touched on a few days ago that various products were resolving towards colder outcomes and whilst this broadly remains the case (some moderation in parts, albeit not material) the key uncertainty remains around timing, which I believe is what Ian is also keen to emphasise. In my view, some products appear to accelerate transition towards mobility faster than others, and in a blocking synoptic such as we see it is often wise to take a step back and consider plausibility; is it really likely that there will be sufficient energy within the Atlantic to displace an anchored pattern, or is it more likely that the incumbent synoptic would amplify and deflect (hence reference to advertised battleground scenarios) Very interesting output at present and there really is no strong reason to call any particular outcome. If you're a coldie (as I am) there is plenty reason to maintain interest and see where the models begin to favour with confidence. SB

This. For me, it's a case of micro and macro analysis; I personally find trends are easier (ha!) to identify and analyse when considered within a wider gamut, otherwise I find that you can become lost within the noise and data scatter of intra-run variance - something which I pay little close attention to. There are various techniques which can be adopted and one of those is time benchmarking (comparing 12z to 12z, etc) Lastly, GFS gets a fair bashing on here, but what I would reinforce is that its output is absolutely considered within MO, albeit within the context of holistic NWP output; it isn't the case that it is ignored, more that its output is considered against other more locally sensitive products. WRT such products, without dangling any carrots, there are signs of some confidence around a colder signal - but the timing of this remains less certain. Again - and to reinforce the analytical approach - this is something of an emerging trend, and is probably better described as a moderation of confusion, following a period where no particular outcome held greater confidence. I wouldn't go buying sledges just yet, mind. It is vitally important to see whether such an aforementioned trend consolidates in future runs, or whether it remains rogue and undecided. As a coldie, I am obviously favouring a colder outcome! As you were... SB

2nd this Please keep posting, @DIS1970 I find this a very inclusive forum (I recall how members have supported a couple of other posters over the years, which was nice to see) Quantity does not always equal quality and a pithy comment can be of equal value to that of a thesis! As John says, best wishes with your continuing recovery.

Most of the algorithms I’ve looked over keep it fairly simple: precipitation intensity over time period = rough depth calculation. I’m yet to see any model which introduces soil temperature feedback to calculate an accurate melt rate, and therefore a modifiable depth (assuming further snow was within the forecast) UKV has an experimental field [water_equiv_lapse] which tries to establish this but I wouldn’t say it’s conclusive, reliable or widely used within output. SB

That isn't strictly true, John. Whilst you are correct in respect to how thermodynamics are considered primary drivers in NWP, various other factors and variables need to be introduced to account for regional variation and idiosyncrasies, in order to improve accuracy. The difficulty, is that the layering of variables increases error rate by an order of magnitude, and therefore models require tuning. It has been observed many times over that some NWP models are predisposed to resolving certain patterns or outputs more favorably than others, and there is no real explanation other than how the underling mathematical weighting has been written. Does that constitute a bias? In terms of its output, possibly - but it is entirely intentional, and in any event anomalously recurring patterns are generally synthesized out (ensembles are particularly sensitive to this) SB

I think people are subconsciously managing their own expectations, as well as that of others. Many observers on here have years of experience and have literally analysed thousands of model runs, so it's an informed and reasonable position to take, in my view. The earlier GFS run was, as others have noted, quite exceptional. On the balance of sheer probability, it is more than likely that the outcome will at least be tempered in the 12z's - but the question really is, to what degree. But yes, I agree that there is some background signals which do - to more or less of an extent - support cooler phasing, so the evolution witnessed on the 6z's shouldn't necessarily be discounted out-of-hand. In my view however, and again in that of many others, it is a little bit extreme. Definitely one of the more exciting 12z's coming up, because if it's anywhere near comparable with the 6z's, then eyes shall definitely be raised

Indeed, not so much an outlier but very much a member with little support. What's most evident for me is a couple of things: the level of noise within the signal into the new year, which is clearly dispersed across a wide range of outcomes, and secondly that the mean - whilst acknowledging the noise - does actually trend downwards. I think we'll see a moderate cool down, but nothing notable as such. Sledge remains on stand-by

I have to admit, space weather isn't my forte! Leaves me very Some intriguing signals popping up in the seasonal modelling; I wouldn't say it's decidedly different to previous, but more a meander towards something perhaps a little cooler. Whilst that is no doubt something to observe on the horizon, there is very much more immediacy - and focus on - our current dalliance with the wet stuff.

Welcome back to the madhouse Stuart! It's been a while since we've had a 'seasonal' August, as such; September has recently been our saving grace.

Hi Shotski, I don't know if the '10 days ago' refers to the last post I made, this one: https://forum.netweather.tv/topic/82300-model-output-discussion-16th-jan-12z-onwards-smile-while-you-post/?view=findpost&p=3135716? If you are, you'll note that - whilst I recognised that CFSv2 was advertising its now familar blocking nirvana - I also recognised that, at that point in time, no such signal had any such support from GS5. So, to answer your question: No, GS5 wasn't showing any strong blocked pattern 10 days ago. Broadly speaking, the signal was for average to slightly below temperatures, and perhaps a little drier than average - which is representative of an anticyclonic pattern. In this regard, not a great deal has changed - indeed, if anything, confidence is a bit firmer, as we're 9 days further forward. I'm a proud coldie, and would be happy as larry to see some excellent cold synoptics, but - at this stage - February looks fairly benign. But anyway, I've enough to focus on over the next week or so! SB

I see the omnipresent February signal for blocking prevails in recent CFSv2 output CFSv2 always appears to favour this type of broad synoptic for the latter end of winter, for whatever reason - seldom verifies, I hasten to add. There has been some real Steve Murr eye-candy over the last few days. Alas, none is currently supported by GS5 outputs; advertisement of average to very slightly below, is a broad theme; perhaps drier than average too. Not a great deal of confidence though, so not worthy of any wrist-slitting. ECM is a bit of a turn-up. If that has legs...

The focus within the near-to-mid term is the omnipresent signal for deepening areas of low pressure to swing in; and, unfortunately, it's looking a little 'rinse and repeat' on that particular front. All of the bases are loaded to fuel some powerful systems, and it therefore comes as to surprise to see such strength being advertised in resonance across NWP. As has been previously covered, any significant stratospheric signal has rather diluted itself, therein resulting in a more mobile broad synoptic. An important point to make here: the lack of any meaningful SSW does not preclude the chance for colder periods of weather to develop; indeed, even with SSW, neither is colder weather an absolute guarantee for our locale. Severe weather is most definitely on the way, just not of cold origin.

Just to add: without introducing any hyperbolic nonsense, the interesting signals alluded to remain (and, broadly speaking, have done since I first mentioned back around the time of this post) Gritters: on your marks... SB

Some interesting signals just beginning to appear into December - later on, I hasten to add. A cautious word though: this is well balanced against equally more probable synoptic, and therefore confidence is equally moderated. Near-term remains as stated last week: cyclonic incursions - seasonal, if you will. Apologies if my lack of enthusiasm/bravado for any near-term cold appeared to be 'sitting on the fence' - it wasn't; it was merely a view influenced by data which simply did not resonate to that tune at all. Height anomalies are a frequent topic of discussion, by the way: organisation appears 2-3 weeks delayed, however stratospheric influence could further disrupt. Extrapolate beyond and hence tenuous plausibility for 'interesting signals'. Much to mull over. SB

Hi Ali, Strat f/c is continually fed into discussion, albeit rather cautiously. With regards to OPI, I would say that that falls into the many correlative theories which probably need to be refined a little bit further, before it's used deterministically. That said, the micro correlation appears to present an interesting relationship though. No discernable pattern yet established for the coming season, by the way; that's not to say it'll be as relatively benign (in terms of cold weather, at least) as last year, just nothing of note jumping out. But, it is only early November, and so that's hardly surprising. To that end, I'd suggest that any forecast og unseasonable, near-term cold to be a little optimistic; upstream, whilst anomalously cold air exits the eastern CONUS, the form horse is an accelerated jet stream, feeding cyclogenesis, with a resulting unsettled and stormy pattern closer to our shores. SB

Hey Jason, No, not just your opinion Jason this time; one of the known idiosyncrasies of the GFS model is the bias it lends to over-advertising snow in its precipitations fields. This signal seems to become further amplified when associated with deep cyclogenesis (which appears to develop its own erroneous feedback loop) As Ian often advises, it's wise to avoid placing too much faith in GFS precip fields; usually exaggerated and often atmospherically implausible (as is the case here) As an aside, the major concern really is for the tidal event in the early hours. An extraordinary act of coincidence is seemingly conspiring to deliver extremely elevated levels of danger, and to a region already well tensed. One can only hope warnings are heeded; because this really is no joke.

Hi, Firstly, it's not a daft question at all - it's a perfectly reasonable one to ask. The broad answer is 'yes', NWP does indeed get things wrong; indeed, sometimes very wrong. NWP models - especially those that are stochastic in origin - are highly sensitive and susceptible to variance, the degree to which will depend largely on the integrity of the initialisation data (the 200m odd daily readings that feed into them) If there are ghost areas of data - and this does happen, albeit not often - then you will often see a fairly dramatic consequential effect. To counteract that, many variables are parameterised during assimilation and, where required, smoothed-out through applied algorithms (winsoring, outlier factoring) This essentially 'cleanses' the data, pre-processing; not a perfect option by any means, but nevertheless it calms erroneous signals. You may have heard about the ensemble approach? This is where small degrees of variance is intentionally introduced to the initialisation data, in order to grow forecast depth. You can then analyse that depth to ascertain degrees of confidence in the operational forecast. The ensemble approach has probably been the biggest forward step over the last few years, and that's only really been possible with acceleration of supercomputing power; it takes an incredibly long time to generate forecasts, so you need a LOT of computing power to produce one, let alone an ensemble suite. In terms of model skill (ie: how accurate are they?) there's been a fairly linear growth, as more powerful supercomputing becomes available. In real terms, the 96hrs forecast is - today - as accurate as the 24hr one, 30 years ago; so, a four-fold increase in skill. Further advancement of skill comes from better understanding of atmospheric science and modelling and translating the effects back the models. This is a continual area of research, and where - for example - there's been an up-tick in understanding of the stratospheric~tropospheric relationship. In addition, ever greater data density (so, more observations / sensors) will only ever be of help to models that rely so heavily on data. There's a few graphs here to show how accurate the forecast are; the final one depicting the linear growth in skill, over the last 30 years. Hope that helps? SB

Incidentally, CFSv2 has been rather consistent with its advertising of colder solutions through mid-February and extending into March. This pseudo trend isn't too dissimilar to how it equally advertised colder synoptics in February 2013, albeit a fair bit more moderate. The broad synoptical picture is an omnipresent high pressure cell centred over Scandinavia; consistent in synoptical flavour with stratospheric downwelling propagating through to the troposphere. Whilst GS5 isn't wholly reflective of that which CFSv2 proposes, that isn't to say that its broad synoptical proposal should be dismissed out-of-hand. Variance is high and that isn't unexpected; at mid-range it's not so much about detail as such, but more broader patterning. If consistency can be identified in that context, then confidence can be raised, and detail ascribed once within range of more conventional NWP modelling suites. Boom Boom Boom, let me hear you say way-oh, way-ooh!