Better Than The Models ?

[Roger J Smith]

I had mentioned that the two forecasts (MB and RJS) when compared by pressure points are negatively correlated but I estimated that from experience with correlation, the actual correlation is -0.38. What makes them not totally opposite would be the slight overlap of high pressure in the south-central part of the grid and a vaguely similar tendency to lower pressure in the northeast. Anyway, on to the meat of this thread ...

Starting today, I can record the pressure grids for the 12z GFS and ECM maps. So as not to overload the site, I will do this on my own computer system until the final day when reality replaces prediction. However, if there are any questions about the validation post-facto, I can place these grids on-line. An excel file will be created to record this validation permanently and that will be posted after 3 Feb.

The summary of the grids will be posted as follows:

10d GFS (issued 24 Jan) -- 1036 mb ridge approx eastern Ireland to southern Germany, with extended ridge northwest through Iceland to strong Greenland high. North of this ridge, broad WNW flow of moderate intensity with a 975 mb low in the vicinity of Novaya Zemlya. This flow originates from NNW flow between Greenland and Svalbard. Very weak frontal trough bisects the ridge from central Norway through Iceland then southwest.

from correlation performed here, this map has +0.15 correlation with MB 3 Feb and +0.53 correlation with RJS 3 Feb.

10d ECM (issued 24 Jan) -- compared to GFS the ECM has a similar pattern shifted 10-15 deg east. The ridge (which peaks at about 1032 mbs) runs from Svalbard south-southeast into Denmark and then into Poland. The UK and Ireland are under a broad moderate southeast flow while northern Scandinavia into northwest Russia are in a moderate northwest flow. A weak low is located southwest of Valentia, in a trough that extends to the western Atlantic.

from correlation performed here, this ECM map has -0.37 correlation with the GFS map, -0.79 correlation with MB 3 Feb and +0.15 correlation with RJS 3 Feb.

Day 10 summary -- The two experimental forecasts are correlated with the GFS, in particular the RJS map. MB is highly opposite to the ECM while the RJS comparison is a weak positive with that. The two models themselves are moderately "opposite" in correlation.

One other statistic of interest, the average pressure over the grid:

MB ... 984 mbs

RJS .. 1016 mbs

GFS 10 1023 mbs

ECM 10 1018 mbs

As I mentioned before, the MB tendency to lower pressure will not be as big a concern (or feature) in correlation as it may be in implied wind speeds.

I think the negative correlation of (MB-RJS) equalling more or less the negative correlation of (ECM-GFS) is both ironic and interesting. Seems that the current state of play is that RJS and GFS are in one camp, ECM and MB represent opposite case outliers from that two-model consensus. At the same time, in the sector west of Ireland there is high agreement RJS-ECM. So this starts to look like a traditional model discussion and perhaps both the RJS and the MB models have just about as much right to entry into that field for a 10-day forecast as your more expensive models.

Expect a similar report each day as we approach 3 February.

It can be noted that the 16d GFS now shows a very strong low southwest of Iceland approaching that country with a central pressure below 940 mbs and 80 knot winds depicted. This would develop from a weak low south of Newfoundland at day 14. The arrival in Iceland of this low would occur on 10 February. Otherwise, there are no strong winds implied on either model run. The GFS is of course rather infamous for phantom storms that come and go in "FI" so we'll see if this storm becomes a feature of model runs closer to the time.

[SMU]

I will throw my hat into the ring-

7 years ive been here, every year we get one of these threads & every year sooner or later people realise that its a load of nonsence-

It was Robert' the other year & before that it was Ken Ring' lol- IIRC

Just to put it into context you have data fed into the ECMWF or the American models has essentially come from the most advanced scientific minds of this era & yet with the energy & flux that goes into the atmosphere the models can just about cope at day 4/5-

To have joe public come along & pronounce Armageddon @ X time interval is not forecastable -period!, to specific & certainly way OTT in terms of expectation V reality- 160 MPH gradient winds- hum... a 100 year event -

*IF* there had been presentation around expected 500 MB anomalies in certain locations across a period of time I would have been half interested-

As it is this will be my one & only post in the thread-

The you tube forecast has been down the pub with the 18z-

S

Edited January 24, 2011 by Steve Murr

[jethro]

Who is who then? I'm a taciturn, strikingly handsome man of action and adventure, so I'll claim the "The Good"...

I admit I'm bad.

So that leaves........Coast.............. anyone seen Coast?

[OldMetMan]

Well, just for the record, and for what it's worth, this GFS chart is interesting to say the least:

http://www.netweather.tv/gfsimages2/gfs.20110124/18/177/hgt500-1000.png

Granted, it is the 18Z run, it is still a long way out, but this particular solution has appeared in yesterday's run also. I mentioned earlier that, given the predicted synoptic pattern around 29th/30th, strong cyclogenesis south of Iceland is entirely possible. The GFS seems to think so too.

But wouldn't it be something if it verified?

[AtlanticFlamethrower]

Not a bad 18z GFS for MB. That's the new closest yet.

[osmposm]

Today's (22 Jan) Met Office long range forecast is just a description of the UK climate. Anyone could do this.

UK Outlook for Sunday 6 Feb 2011 to Sunday 20 Feb 2011:

This period looks to start mainly dry across southern parts of the UK, with cloudier, wetter and milder conditions in the north and west. Overnight frost and fog patches could continue for many southern parts of England and Wales. Meanwhile, spells of rain or drizzle, perhaps heavy at times with strong winds, are more likely further north across Scotland and Northern Ireland, with hill snow also possible here at times. Updated: 1145 on Sat 22 Jan 2011

Here's my forecast for April 20th 2011.

Temperatures about normal for the time of year, but with periods of warm spring sunshine at times in the midlands, east and south-east England. Generally dry, but some light showers are possible, especially in the north and west. Mostly light winds.

Well, it's a description of one sort of UK climate, certainly; but the North being forecast to be generally milder than the South, it's not exactly the only - or even the most usual - one for early/mid February. And granted it was talking of the situation two to four weeks ahead, it's not surprising it was pretty loosely worded for the latter part of the period. We all know how unreliable models and predictions are at even ten days!

As for your general-purpose forecast for three months hence....yup, not bad. Could be that....or not. It'll be fun to see how it pans out! :winky:

Edited January 25, 2011 by osmposm

[Coast]

I admit I'm bad.

So that leaves........Coast.............. anyone seen Coast?

:blush:

I guess that leaves me as the ugly one then!

Not a bad 18z GFS for MB. That's the new closest yet.

Tried to align and match up the coastlines from these two, hope everyone can see the two distinct images and can compare:

Personally, I can't draw any conclusion at this stage but the period under scrutiny is only three days away now.

[MurcieBoy]

We know that +180 is a long way out (and the models do tend to move about a lot), but (for what it's worth) here's my comparison of the latest GFS for 12:00 on 1 Feb 2011.

Edited January 25, 2011 by MurcieBoy

[OldMetMan]

We know that +180 is a long way out (and the models do tend to move about a lot), but (for what it's worth) here's my comparison of the latest GFS for 12:00 on 1 Feb 2011.

The models have been very ambivalent, as I mentioned yesterday, it all depends on whether the deep low forecast to develop near Nova Scotia within 3 -4 days deepens enough, and turns to the NW, so as to allow the downstream trough to amplify and thus allow a rapid low pressure development to occur as per your forecast. It's this area that the models don't seem to have a handle on yet.

Purely from an historical point of view of synoptic development, your scenario is the more likely as I see it, given the present pattern. Blocking highs such as we have seen in recent days more often retrogress.

[MurcieBoy]

The models have been very ambivalent, as I mentioned yesterday, it all depends on whether the deep low forecast to develop near Nova Scotia within 3 -4 days deepens enough, and turns to the NW, so as to allow the downstream trough to amplify and thus allow a rapid low pressure development to occur as per your forecast. It's this area that the models don't seem to have a handle on yet.

Purely from an historical point of view of synoptic development, your scenario is the more likely as I see it, given the present pattern. Blocking highs such as we have seen in recent days more often retrogress.

Your posts have been extremely useful to my further analysis/evaluation of the likely outcome.

Edited January 25, 2011 by MurcieBoy

[OldMetMan]

Your posts have been extremely useful to my further analysis/evaluation of the likely outcome.

Thanks MB, interestingly the latest GFS shows a curious departure from previous runs:

http://www.netweather.tv/gfsimages2/gfs.20110125/06/204/hgt500-1000.png

Such a development is capable of VERY tight gradients if it comes off.

I think the models will start to consolidate a trend from this point on. The UKMO is the one to watch, as I consider it to be the best at picking up rapid low development, I think the October 1987 storm did much to improve their forecasting parameters in this area!

[Roger J Smith]

Continuing the formal validation, day 9 GFS

The 12z GFS at day 9 (25 Jan) has developed a somewhat stronger circulation pattern but remains similar to the previous day. A 1042 mb high sits near the Channel, with a ridge NNW into the Greenland Sea. A 984 mb low is located over northern Finland and Karelia. There are no significant features to the west of the high other than a broad southwest flow that terminates west of Svalbard. The day to day correlation from day 10 GFS is 0.87. Both experimental forecasts have dropped in correlation with this map relative to their scores on day 10, but remain positively correlated, RJS at 0.36, MB at 0.05.

The 12z ECM, which now shows high pressure ridged east-west near 50N through the grid, and a broad WSW flow to the north terminating with weak lows near 70N, has changed considerably over 24h and is correlated at -0.19 with its predecessor, while jumping to a correlation of 0.65 with the GFS. It is also highly correlated with the RJS at 0.71. The correlation with MB while negative at -0.15 has improved a similar amount.

At day 9 then, the model "ensemble" has shifted so that GFS, ECM and RJS are broadly similar and all inter-correlated, while the MB remains an outlier from all but by no means opposite in sign from any of them, so an alternate solution that would not require a total pattern change from the two models in flux to achieve higher correlation nearer the time.

The RJS feature of a separate low moving towards Ireland is not supported on either 9d model run and remains the main reason for a lower than 1.0 correlation. The MB feature of deep low pressure in northern-central Sweden is mildly supported on the GFS and mildly discouraged on the ECM. All four model solutions have a highest pressure near 50N but the position seems to move west in the order of RJS-ECM-GFS-MB.

The feature of a very strong low near Iceland by 9-10 Feb has been replaced by more of a moderate zonal flow with unspectacular pressure or wind patterns on the longer range GFS. The 10d ECM hints at a similar trend if that model went further out.

[MurcieBoy]

Here's a graphical representation to supplement RJS's work:

[Harsh Climate]

Here's a graphical representation to supplement RJS's work:

If those charts come off you will end up with a forecasting accuracy of about 0% (Of your big storm forecast.)

[bluearmy]

could be very stormy at some point between 6th and 10th feb if the spreads on ecm and naefs are anything like accurate. not the same set up as forecast but if we get a big winter storm first half feb, what kind of statistical support would that have - kevin ???

[MurcieBoy]

If those charts come off you will end up with a forecasting accuracy of about 0% (Of your big storm forecast.)

Maybe, but what are the chances GFS is right at +216? Here's their Xmas 2010 effort:

[Kentish Kiwi]

Seems to me that there needs to be a pretty huge swing in the models in a relatively short space of time to see anything close to what MB predicted...

[Harsh Climate]

Maybe, but what are the chances GFS is right at +216? Here's their Xmas 2010 effort:

To be honest these charts arent massively different.. Certainly not compared to the monstrous changes needed for your forecast to verify. All we see here is the high presure over the UK being slightly differently situated, what you need is a monster storm to develope out of nowhere in the areas you have mentioned! (And not just a common deep area of low pressure that the charts have hinted at..)

[LomondSnowstorm]

If the control run verified - we'd have to give MB at least a little credit? http://91.121.84.31/modeles/gens/run/gens-0-1-276.png?12

[MurcieBoy]

To be honest these charts arent massively different..

The GFS were back on form the next day at +192:

[mountain shadow]

Seeing as its Burns Night I would have to say MB's forecast was on " a shoogly peg the necht".

[OldMetMan]

To be honest these charts arent massively different.. Certainly not compared to the monstrous changes needed for your forecast to verify. All we see here is the high presure over the UK being slightly differently situated, what you need is a monster storm to develope out of nowhere in the areas you have mentioned! (And not just a common deep area of low pressure that the charts have hinted at..)

At least 3 GFS runs over today and yesterday have shown striking similarities to MB's forecast. I do not see why "monstrous changes" are needed given the type of flow predicted for early Feb.

If you take a look at the synoptic pattern for 31st January 1953 storm you will see that the "monster storm" which produced so much devastation did, in fact, develop out of nowhere. A weakish low pressure area was off the west coast of Scotland and rapidly intensified as it moved E and then SE.

I mentioned earlier that, as far as predicting any intense cyclogenesis is concerned, the only model I would trust would be the UKMO, so by tomorrow or Thursday we should have a better idea of what, realistically, could happen in the first days of February.

I get the feeling here that some posters really want MB's forecast to fail because should it prove close to the mark, it would seriously undermine conventional forecasting methods. We are looking at a radical departure from accepted ideas on how atmospheric systems come about, that's kind of heretical to some folk unfortunately.

Edited January 25, 2011 by OldMetMan

[AtlanticFlamethrower]

18z again shows MB's pattern for Feb 3rd which would be impressive if that verifies but the detail of the forecast - namely it being windy - doesn't look close to being accurate at this stage.

[La Bise]

I get the feeling here that some posters really want MB's forecast to fail because should it prove close to the mark, it would seriously undermine conventional forecasting methods. We are looking at a radical departure from accepted ideas on how atmospheric systems come about, that's kind of heretical to some folk unfortunately

It would undermine the very understanding of our universe, rewrite history, most probably make a mockery of a number of fundamental laws of physic (wild guess here...) as MB makes it quite clear, in his initials posts on here that his method is one that has been passed on by an ancient civilisation that had means of communications far surpassing ours, an intimate knowledge of nature far surpassing ours and that in order to gather that knowledge you need to attain a kind of higher sphere of consciousness, possibly through some kind of time travel method (assuming when he talks about 4thd dimension, it is the intended popular convention about time being the 4th dimension).

MB would be, on Netweather and a few other selected websites, basically overthrow the very fundament of our civilisation. We're not talking about some guy who think some factors not yet recognised by the dreaded drones of meteorological academia do have an influence on our weather, we're talking about someone on the brink of offering a first proof towards something of cosmic grandeur...

Or he is just a very naughty boy...

[OldMetMan]

I get the feeling here that some posters really want MB's forecast to fail because should it prove close to the mark, it would seriously undermine conventional forecasting methods. We are looking at a radical departure from accepted ideas on how atmospheric systems come about, that's kind of heretical to some folk unfortunately

It would undermine the very understanding of our universe, rewrite history, most probably make a mockery of a number of fundamental laws of physic (wild guess here...) as MB makes it quite clear, in his initials posts on here that his method is one that has been passed on by an ancient civilisation that had means of communications far surpassing ours, an intimate knowledge of nature far surpassing ours and that in order to gather that knowledge you need to attain a kind of higher sphere of consciousness, possibly through some kind of time travel method (assuming when he talks about 4thd dimension, it is the intended popular convention about time being the 4th dimension).

MB would be, on Netweather and a few other selected websites, basically overthrow the very fundament of our civilisation. We're not talking about some guy who think some factors not yet recognised by the dreaded drones of meteorological academia do have an influence on our weather, we're talking about someone on the brink of offering a first proof towards something of cosmic grandeur...

Or he is just a very naughty boy...

Do I detect a note of sarcasm here?

I do not think you actually read his initial posts (or his posts on other forums) or you wouldn't be so rude about his methods. Besides, what you have just posted only proves my point that MB is speaking heresy as far as some are concerned.

What cannot be understood is ridiculed, it's been happening for centuries.