12z Model Comparisons Results

[Thundery wintry showers]

12 December

Missed yesterday, due to being at Leeds University graduation when I received my MRes with Merit award.

A large 964mb low to the north of Scandinavia, 985mb low north of Scotland and a very strong westerly flow over the British Isles. Weather consisted mostly of sun and showers.

GFS was unavailable at T+144. T+120 and T+96 were disappointing with too strong a HP over Europe, giving strong WSW winds in Scotland, and an anticyclonic regime in the south; only 6/10.

UKMO also overdid the Euro High but to a much lesser extent than the GFS. T+144 and T+96 were particularly close to being spot on, while T+120 had the high covering the extreme south. I give this an 8/10, as everything else was accurate.

ECMWF also scores 8/10. T+144 was unavailable, but T+120 and T+96 had the Euro High only slightly too far north- similar to UKMO.

NOGAPS had the strong westerly pattern right, but seemed uncertain of the positioning of small secondary lows to the north, in particular T+120 had Britain under a weak ridge of high pressure; only 6/10.

JMA was poor (only 4/10) as the Euro High was too far north and east at all three timeframes. T+144 had a slack SW flow over Britain, T+120 had a moderate SW flow and T+96 had a strong SW flow. Low pressure was always centred over Iceland.

GEM was quite poor at T+144 with a cyclonic W/NW flow and probably wintry showers for Scotland, but T+120 and T+96 were reasonably good, so this run scores 7/10.

Surprising that GFS and JMA overdid the Euro High the most; that's a trait I more commonly associate with UKMO and to a lesser extent ECMWF, but the Euro models handled it best this time.

[johnholmes]

being at Leeds University graduation when I received my MRes with Merit award.

congrats Ian and again many thanks for all the work in doing this.

John

[tooby7]

12 December

"Missed yesterday, due to being at Leeds University graduation when I received my MRes with Merit award".

Well done that man.

Edited December 13, 2006 by tooby7

[Thundery wintry showers]

Many thanks for the congratulations above.

Now, unlucky for some:

13 December

A classic Bartlett situation with 1030mb Euro High and strong south-westerlies over Britain, 985mb low off NW Scotland.

GFS was quite impressive this time, scoring 8/10. The Euro High was slightly overdone at T+144 and T+120 but we still had a strong SW flow over Britain, and T+96 was pretty much perfect.

UKMO was about as good as GFS. T+144 wasn't as good, with the SW winds only just coming into Britain, and the lows too far west, but T+120 and T+96 were excellent- better than on GFS, so I give it 8/10 as well.

ECMWF was poor at T+144 with a westerly over Britain, and everything too far west by a long way. T+120 and T+96 were very good, although T+96 had an overdone low to the NW and winds even stronger than actually happened. I only give this 7/10, although it was still a reasonable output.

NOGAPS had the Euro High overdone at T+144 and T+120, and at T+120 the high was over southern England in an anticyclonic westerly regime. T+96 was unavailable. Overall just 5/10.

JMA was reasonably good- another 8/10- with some niggly problems at T+144 and T+120 in terms of low positioning, but Britain always in a SW flow, and then a very good T+96.

It's fair to say that GEM was the worst of the outputs. T+144 and T+96 did have a SW'ly over Britain but with the Euro High somewhat overdone, and T+120 had us in a showery NW'ly with the Euro High pushed out of the way. A 4/10.

Overall, UKMO, GFS and JMA were about as good as each other- some impressive outputs there- with ECMWF not too bad either. NOGAPS and GEM were poor though.

[Thundery wintry showers]

14 December

965mb low north of Scotland. It was another very mild and windy day, 1030mb Euro High.

Another good GFS output, worthy of an 8/10. At T+120 the Euro High was slightly overdone resulting in a flow from a south-westerly point, but on the whole, it was a very good output, T+144 and T+96 both accurate.

UKMO had a SW flow over Britain at T+144 and T+120 with the Euro High overdone, but was accurate at T+96. It thus scores 7/10; not quite up to the GFS's standard.

ECMWF scores 8/10, being about as good as the GFS. It had Britain in too much of a SW flow at T+144, but was accurate at T+120 and T+96.

I don't have the NOGAPS T+120, but it scores only 5/10 on the evidence of T+144 and T+96. T+144 had a high to our east and SSW winds over Britain, while T+96 committed the error of having too much of a SW flow, especially over the south- none of the main three models committed this error at T+96.

JMA had too much of a SW flow at T+144 and T+120 but was very good at T+96- similar to UKMO, so scores 7/10.

GEM was poor again- showery NW flow at T+144, pretty good at T+120, but at T+96 had a strong low off W Scotland resulting in a SW flow. A 5/10.

NOGAPS and GEM had problems, but the other models handled it pretty well again, with GFS and ECMWF being marginally better than UKMO and JMA.

[Thundery wintry showers]

15 December

Another westerly dominated chart, 980mb W Scandinavia Low, 1000mb Icelandic Low, 1030mb Euro High, 1020mb shallow low to the SW.

I don't have the T+96 outputs- that was the day I was away!

GFS had another good 8/10 run; at T+144 it didn't have the 1020mb low to our SW but the general WSW-wind pattern, and positioning of lows, were good, and at T+120 it was very accurate.

UKMO had us in a broad WSW flow at T+144 and T+120, with the maritime air well away to the west, and no secondary 1020mb low over the SW. Thus, it only scores 7/10.

ECMWF was good at T+144 but had too strong an Icelandic Low, so winds in the Atlantic were more SW'ly. T+120 was very accurate. Overall about as good as GFS, so scores 8/10.

NOGAPS was available only at T+120 so no mark, but it was poor, with a very SW'ly airflow and a low off western Scotland.

JMA overdid the Euro High at both timeframes, with pressure at 1030mb over south-east England, and a secondary low over Scotland, rather than the SW. Thus it scores only 5/10.

GEM is another 5/10; it had Britain in a SW pattern with a low off western Scotland at both timeframes, T+120 rather better than T+144, at least the polar maritime airmass was close enough at T+120.

Overall, another good performance by GFS, UKMO and ECMWF, with the GFS and ECMWF just edging out UKMO. The other three models, though, were rather disappointing.

[Thundery wintry showers]

1 January 2007

The model comparisons return for the New Year, following a long period of no updates.

A very strong WNW flow over Britain, with a 960mb low off NW Scandinavia, and a 988mb Icelandic Low, cutting off the coldest air and rendering the westerly flow a moderately mild one.

I only have the T+96 outputs for New Year's Day, but here they are:

GFS had too flat a westerly flow, failing to pick out the wave feature that affected the southern half of Britain, but otherwise it was accurate, an 8/10.

UKMO was a 9/10- it had an accurate westerly flow and also the trough over the south.

ECMWF also scores 9/10- again evidence of the trough over the south, and a strong westerly flow.

NOGAPS was reasonable- 8/10- but didn't pick out the trough over the south.

JMA wasn't so good- it had a westerly regime over us but the trough over the south was shown as being like a secondary low, with only slack winds. Only a 6/10.

GEM scores 7/10- it overdid the Icelandic Low, causing more of a WSW flow to approach the British Isles.

Overall a good set of T+96 outputs, as one would probably expect, but JMA and GEM weren't so good. Tomorrow the T+96 and T+120 outputs will be available for the analysis of 2 January.

[Thundery wintry showers]

2 January 2007

975mb low off NW Scandinavia, a NW flow over Britain but not an especially cold one, with 1040mb high over Spain.

GFS was dodgy at T+120 with a secondary feature in a WNW flow, and the high over Spain underdone and too far south- overall a more unsettled theme than actually happened. T+96 was very good, so it scores 7/10.

UKMO delayed the passage of all systems at T+120- it was reasonably accurate, but everything was slightly too far west. T+96 was accurate. Slightly closer than GFS at T+120, so scores 8/10.

ECMWF had everything a bit too far east at T+120- too progressive, mild SW'ly winds coming into the west, but was very accurate at T+96. Overall about level with the UKMO, so another 8/10.

NOGAPS was poor, at T+120 the high over Spain was way too far north, affecting southern England, with a flat westerly flow over Britain. At T+96 everything was a good 300-400 miles too far east- far more so than on ECMWF T+120- with SW winds for most. Only a 5/10.

JMA was poor at T+120 with a low east of Iceland giving SW winds for Britain, but it was good at T+96, though the high to the S was slightly overdone. I'd put this behind GFS but ahead of NOGAPS; 6/10.

GEM had everything too far east at both timeframes, with SW winds for most; only 5/10.

Overall, good performances by the three most respected models; disappointing from the other three.

[Thundery wintry showers]

3 January

A very familiar picture by now; 972mb low south of Iceland, 1032mb Euro High, and a strong SW flow over the British Isles. T+144 and T+120 are available.

GFS was reasonable, but slightly out at both timeframes. T+144 had Britain in a WSW'ly but with winds lighter than actually happened, and high pressure over Europe too far north. T+120 had Britain in too much of a westerly flow, but otherwise good. I give it 7/10.

UKMO was pretty good at T+144, though with the Icelandic Low too far west, and more of a SSW flow. T+120 was also good, but Icelandic low was slightly too far east, giving westerlies. A 7/10 overall.

ECMWF was also good but not brilliant; both timeframes had clearer showery weather spreading into the west and more of a cyclonic westerly flow over Britain, associated with the Icelandic Low being too far east, overall another 7/10.

NOGAPS was poorer; it was good at T+144 with the SW flow over Britain- indeed possibly the most accurate of the T+144s- but at T+120, a showery westerly flow was indicated, the low a long way too far east, overall 6/10.

JMA was way out, with a low over Scandinavia and east of Iceland at T+144, giving westerlies, showery in the north, warm sector in the south. T+120 went for a showery westerly with the Icelandic Low a long way too far east. Only 4/10.

GEM had Britain in a NW'ly at T+144, and a high coming in from the SW. T+120 was better, with a showery westerly moving in, and SWs for eastern areas, but everything way too far east, also a 4/10.

Disappointing JMA and GEM outputs. The pressure chart was quite standard, but nonetheless, no models got it absolutely right. NOGAPS was closest at T+144, but threw it away with a poor T+120 output.

[Thundery wintry showers]

4 January 2007

I wasn't available on 30 and 31 December, and thus only T+144 is available for this day of north-westerly winds, 975mb Scandinavian Low and 1032mb high over Spain.

GFS was poor, with Atlantic lows way too far east, and SW winds over Britain, only a 3/10.

UKMO was no better, again lows way too far east, and SW winds, 3/10.

ECMWF didn't have the lows quite as far east, with more of a WSW flow covering Britain, and highest pressure wasn't displaced as far east, but still a poor output, 4/10.

NOGAPS was quite accurate- slightly too flat a westerly flow, more of a WNW, and the lows west of Iceland were over Iceland, but not bad at all overall, 7/10.

JMA is another 3/10- lows way too far east, high displaced into Europe, SW winds.

GEM likewise- 3/10, SW winds over Britain.

It's hard to conclude much from just a T+144 output, but NOGAPS was the only model not to get it completely wrong.

[Thundery wintry showers]

5 January 2007

An elongated Icelandic Low at 970mb; unfortunately only T+96 is available (T+96 and T+120 will be available tomorrow) and 1032mb high over Spain.

GFS and ECMWF were both spot on at handling this; a 9/10 for GFS and 9/10 for ECMWF, only very minor differences with what actually happened. UKMO didn't have the 'wave' in the westerly flow, and so scores only 8/10, but still very good.

NOGAPS scores 7/10- lowest pressure too far west, and a slacker WSW flow than actually happened, but still reasonably good.

JMA scores only 5/10- only a slack WSW flow and the Icelandic Low well in excess of 500 miles too far west.

GEM scores 7/10- a reasonable run but the Icelandic Low was about 300 miles too far east, so polar maritime air was closer to affecting Britain.

Once again, very good outputs by GFS, UKMO and ECMWF, and less good outputs by NOGAPS, JMA and GEM.

[J10]

9th December

Does anyone have a copy of the model comparisons files that I used to have? When my computer's system registry packed in, Windows XP decided that Drive H (where I keep my files) was faulty, due to a system registry error, and forced me to run Chkdisk on startup to "repair" the files. The result was that it corrupted many of the files by embedding system registry data into them. I've lost both the model comparisons.doc and .xls files to this.

model_comparisons.xls

Updated to 1st November. A Word document will soon be arriving in John's email inbox.

This is the last backup I can find for the excel version,I don't know if this helps atall.

[Thundery wintry showers]

Thanks for reminding me- yes, it does help, I've been able to download it, and have only lost the data for November and December now. Since there are large gaps in my coverage over that period, I may well be able to catch that up in the near future.

[Thundery wintry showers]

6 January

We now say goodbye to 2006 properly; both the T+96 and T+120 outputs are from 2007. 2007 remains mild, wet, and windy.

970mb lows to the west and east of Iceland, a WSW flow over britain, and 1032mb high SW of Spain, weak 1012mb low over southern England. It's getting quite repetitive seeing the same synoptic pattern in charge over and over again! B)

GFS scores an 8/10 as it had the WSW pattern correctly over the British Isles at both timeframes, though the low over southern England was incorrectly placed- about 200 miles too far east at T+120, and slightly too far to the NW at T+96.

UKMO didn't have the southern England low at all at T+120, giving Britain in a WSW flow, while T+96 had the trough, but like GFS, a bit too far NW. Still, an 8/10 as the general pattern was spot on.

ECMWF also scores 8/10- indeed it was almost identical to UKMO.

NOGAPS was also quite good, but failed to get the low over the south, and at T+120 in particular, everything was too far east, with colder mP air getting towards Ireland. Overall a 7/10.

JMA was dodgier- a SSW flow at T+120, with pressure too high over Europe, and at T+96, the lows were too far east, with mP air starting to come into the west. Just a 6/10.

GEM is only a 4/10. Polar maritime air covered the British Isles on a westerly flow, and at T+96 a large 970mb low was placed off western Scotland.

Again, little to choose between GFS, UKMO and ECMWF, the Euro models holding their own against GFS during this zonal spell. They're out on their own, with the NOGAPS, JMA and GEM performing less well.

[Thundery wintry showers]

After two days of internet trouble, the summaries return, with 7 and 9 January under the microscope today (8 January's chart was unavailable)

7 January

This is getting repetitive to say the very least- no wonder even West Is Best is getting fed up with the constant zonality. 972mb low north of Scotland, 976mb Icelandic Low, 996mb low off western Ireland, 1032mb high to the south with a ridge into Europe, and SW winds.

GFS did quite well for 7 January. At T+144 there was too flat a westerly flow, with sunshine and showers the likely result, but T+120 and T+96 both had a strong low to our west (slightly overdone and too far west) and SW winds over Britain, most things were accurate. I rate this at 8/10, the T+120 and T+96 outputs were very close to the mark and T+144 wasn't bad.

UKMO at T+144 was similar to GFS; a straight westerly flow. T+120 was good- similar to GFS- but T+96 was dodgy again, with the low having passed away to the east, and a westerly flow. Only a 6/10.

ECMWF with 9/10 did best; it had the pattern correct at all three timeframes, the only quibble being that at T+120 and T+96 the low was slightly overdeepened and 100-200 miles too far west, similar to GFS. Because of the better T+144 it beats the GFS.

NOGAPS was unavailable at T+144. The T+120 and T+96 outputs were good, the low about 200-300 miles too far west but giving a SW flow, overall an 8/10.

JMA was less good; everything was displaced too far east at T+144 and T+120, giving westerlies. T+96 was accurate, with the low west of Ireland accurately positioned, but only a 6/10 because of the relatively poor T+144 and T+120.

GEM was accurate at T+144 and also T+96, but it was poor at T+120, giving a showery westerly and everything displaced too far east. Overall, a 7/10.

A reasonable set of outputs on the whole.

[Thundery wintry showers]

9 January

9 January had a 960mb low off Shetland, 996mb secondary low off western Ireland, and westerly gales- the chart does look different to the constant SW'ly winds of the past several charts.

GFS only scores 5/10. The T+144 and T+120 outputs had the main area of low pressure west of Scotland, giving south-westerly gales over Britain; only at T+96 was it reasonably accurate.

UKMO was surprisingly good at T+144; low slightly underdone, but WSW winds shown over Britain and low correctly placed, it even had a 995mb secondary low west of Ireland. Only T+120 was even slightly dodgy, with the low of the right intensity and correctly placed, but more of a WNW flow over Britain with the secondary low west of Ireland too far north. T+96 was pretty much perfect. A good 9/10.

ECMWF scores only 5/10; T+144 and T+120 had the low over western Scotland giving WSW gales. Although those outputs were closer than those of the GFS, the T+96 was worse than GFS, with Britain still in a WSW flow although the 960mb low was at least east of the meridian in that output.

NOGAPS was unavailable at T+96. At T+144 and T+120 it was extremely poor with the low off western Scotland giving southerly gales, changing to SW'lys in the west. Only a 3/10.

JMA was also unavailable at T+96, but in any case, like NOGAPS, it was poor at T+144 and T+120 with the low well to the west of Scotland, and SSW winds over Britain; only 3/10.

GEM was decent at T+144 with the low slightly underdone and over N Scotland giving westerlies, so only slightly too far west. However, T+120 was poor with the low west of Scotland giving SW winds; T+96 had the low correctly placed but underdone giving moderate westerlies. A score of 5/10.

Today's outputs were not so good, with the exception of UKMO which had it nailed on.

[Thundery wintry showers]

10 January

960mb low SW of Iceland, NW winds over Britain, 995mb low off NE Britain, and 1032mb Azores High with a ridge into Europe.

GFS was only good enough for a 6/10 as it had too flat a westerly flow at all three timeframes, most notably T+144 when there was no discernable ridge of high pressure and low pressure was too far south; T+120 and T+96 were better though.

UKMO is good enough for 7/10 as it was closer than GFS, but it still had too flat a westerly flow, with the ridge of high pressure underdone especially at T+144.

This time it was ECMWF's turn to have it nailed on at all three timeframes; only minor niggles like orientiation of the NW flow, and overall superb; 9/10.

NOGAPS was unavailable at T+120, but at T+144 it had Britain in a cyclonic westerly flow, and at T+96, the ridge of HP was underdone, so only a 5/10.

JMA has only T+144 available so I can't give it a mark. At T+144 it had Britain in north-westerlies, but the cold air was vastly overdone, projecting snow showers for the northern half of Britain.

GEM was poor at T+144- low pressure over Scotland giving W winds. T+120 had a fairly long-draw northerly with snow showers likely for the north, while T+96 was quite accurate but with more of a NNW orientation to the flow. Only a 4/10.

Overall, an excellent run by the ECMWF.

[J10]

08 January Actual Pressure Chart (From Weatheronline Expert Charts Section)

[Thundery wintry showers]

8 January is therefore to be updated:

Elongated 980mb low to the north of Britain, 975mb centre south of Iceland, 1030mb high over W Spain with ridge into Europe.

GFS only scores 5/10 on the evidence of this output. A deep low N of Scotland at T+144 and T+120, giving very strong to gale force westerlies. The positioning of the lows was completely wrong. T+96 was better but had Britain in westerlies and the Icelandic Low too far west.

UKMO had the Icelandic Low too far south at T+144 giving SSW winds. T+120 and T+96 were more accurate but had too much of a SW-NE orientation to the flow; overall 7/10.

ECMWF was almost identical to GFS and scores only 5/10; T+144 and T+120 had a deep low north of Scotland.

NOGAPS was pretty good at T+144, with the elongated low complex well modelled. T+120 was poor, with a large low off NE Scotland giving WNW winds, then T+96 had a very slack westerly flow, not really resembling the actual output. Overall just a 5/10 despite the good T+144.

JMA had a long draw of SW'ly winds at T+144, a large low to the NE with WNW winds at T+120, but was reasonably accurate at T+96; overall another 5/10.

GEM was decent at T+144; WSW winds, elongated low complex, but at T+120 and T+96 it had us in a showery westerly with no secondary low development; overall a 6/10.

Not such great outputs this time.

[Thundery wintry showers]

11 January

The relentlessly zonal weather continued. 960mb low to the north of Scotland, severe gale force westerlies, 1036mb high to the south covering Spain- classic zonality.

GFS had the low too far east with more of a N-S orientation in the flow at T+144; as a result the flow projection was NW'ly, and conditions were projected to be cool zonal with snow in the north. T+120 and T+96 were reasonably accurate though, particularly T+96, so overall a 7/10.

UKMO scores 8/10, as although it wasn't perfect and underdid the low to our north, it was good at all three timeframes- even the traditionally dodgy T+144.

ECMWF scores only 7/10. It had the low to our north underdone, and at T+120 it was about 300 miles too far north, giving a slacker westerly flow and high pressure closer to the south. However, the output at least resembled the correct outcome at all three timeframes.

NOGAPS was poor, with the low underdone at T+120 and T+96 (T+144 unavailable) and a cyclonic moderate westerly flow, of polar origin. Only 5/10.

JMA was available at only T+96 so I couldn't give it a mark. Just as well, as it was very poor, having Britain in a warm sector SW'ly and the low a long way too far west.

GEM was poor at T+144 with the low too far west and SW winds. T+120 was better, and T+96 was pretty good, though the orientation of the flow was too cyclonic. I give this a 6/10.

Overall, the UKMO just shaved it as the most accurate model. I think at the moment, there seems to be little to choose between GFS, ECMWF and UKMO.

[Thundery wintry showers]

12 January

A relatively limited set of outputs today because I was changing ISP on 8 January, thus only T+144 and T+120 are available. 972mb low north of Scotland, westerly gales (yet another very windy day) and 1032mb high over Spain and France, in "the Bartlett position".

GFS was okay at T+144, the 972mb low about 300 miles too far east giving WNW winds. T+120 was worse, with the low way too far east, secondary lows piling in from the Atlantic and westerlies. The westerlies were correct, but there were serious misjudgements on the positioning of the lows, so only a 6/10.

UKMO was consistent at both T+144 and T+120, underestimating the speed of approach of the low and locating it off NW Scotland (about 300 miles too far west), giving WSW winds. The T+120 output was closer to being correct than that of the GFS, so the UKMO scores 7/10.

The ECMWF also had the low slightly too far west at both timeframes, but less so than the UKMO- indeed at T+120 there was very little difference from the actual chart- so a score of 8/10.

NOGAPS was pretty good as well, the low generally too far west and slightly underdone, like on the UKMO, and WSW winds over Britain, 7/10.

JMA available only at T+120, when it was pretty poor- a slack WSW flow and no deep low to the north.

GEM was very poor at T+144- a low to the NE, with northerlies spreading south though not especially cold. T+120 was better, with the low about 300 miles too far west and WSW winds- a 5/10.

Overall, it seems that the Euro models were best at handling this output, ECMWF just shading UKMO, while NOGAPS also deserves a honourable mention.

[Thundery wintry showers]

13 January

992mb low over north-west Scotland with strong south-westerly winds; another very windy day, with Mr Bartlett firmly in charge over Europe, a centre of 1028mb over Spain.

All models bar GFS are unavailable at T+120, but thanks to the new N-W Extra archives of GFS runs dating back to May 2006, GFS is also available at T+120.

GFS was quite remarkable in that it was very accurate at T+144- almost nothing to choose between the GFS T+144 and the actual output- but both T+120 and T+96 were poor, with the 992mb low over NW Scotland not picked up, and a moderate westerly flow over Britain. Overall just a 6/10.

UKMO was decidedly dubious at T+144 with a WNW flow, and a rather cold airstream- possibly cold enough to support wintry showers at low levels in the north, and no low off NW Scotland. T+96 was no better than that of the GFS- no low off NW Scotland, westerly winds indicated over Britain. I only rate this at 4/10.

ECMWF had no 992mb low off NW Scotland at T+144, and a westerly flow over Britain. At T+96 there were signs of it picking up the low, but it was somewhat underdone and too far south, with a moderate WSW flow covering Britain. T+96 was therefore better than that of GFS, but T+144 was worse, so the ECMWF scores 6/10 as well.

NOGAPS was almost identical to ECMWF- straight WNW flow at T+144, but started to pick up the 992mb low at T+96 but somewhat underdone; overall 6/10.

JMA had a deep low over N Scotland at T+144 and cold NW winds coming in; it was even worse at T+96 with the low off eastern Scotland giving NW winds over the whole of Britain. Only a 3/10.

GEM had a straight westerly flow at both timeframes; at T+96 also a shallow low over northern England. It scores a 5/10.

Surprising that the models did quite poorly considering the conclusions drawn from Stratos Ferric's excellent analysis thread; I'm guessing that the models must have timed and positioned the track of the low pressure incorrectly, while the southward extent of the -5C 850hPa line was quite accurately modelled. GFS, ECMWF and NOGAPS provided the best stabs at the correct outcome; GFS T+144 was by far the best output, but ECMWF and NOGAPS were closest at T+96.

[Thundery wintry showers]

14 January

All models bar GFS are unavailable at T+144 due to internet problems on 8 January. The synoptic chart shows a 1032mb high just off southern England, an elongated 980-985mb low off northern Scotland, and a low at 988mb south of iceland; the Atlanticathon continued.

GFS was good at T+144; the high slightly too far south with less of a pronounced ridge, but westerlies correctly shown over Britain. T+120 was poor though, with a 1000mb low over NE Scotland and vigorous westerlies and HP well to the south. T+96 was accurate, though high again slightly underdone. I give this a 6/10; it can't really get a high mark with the wobble at T+120.

UKMO scores 8/10. At T+120 the high was slightly too far north, with a strong ridge over southern England, then it was spot on at T+96.

ECMWF scores 7/10; the T+120 output was dodgy with south-westerlies already coming into Britain and the high displaced east, but T+96 was very accurate, more so than GFS and UKMO.

NOGAPS scores only 6/10; the high to our south was displaced to the east giving south-westerlies at both timeframes.

I can't rate JMA as it was available at only T+120; the WZ charts didn't update properly onto the site (Entire Grid Undefined). It had the high displaced to the east giving WSW winds.

GEM scores an 8/10; it was very good at both timeframes, high marginally too far south at T+120.

I sometimes slate the UKMO model for overdoing high pressure to the S and SE, but on this occasion, it was entirely correct to go with a high in that location. GEM was also surprisingly good.

[Thundery wintry showers]

15 January

A day of strong south-westerly winds across Britain, with 1032mb Euro High and 988mb low off NW Scotland. Predictably most areas had a cloudy mild day as a result.

GFS was an okay run that deserves a 7/10. At T+144 it was okay with WSW winds and a large high over Europe but had a secondary low in the southern Atlantic that didn't materialise. T+120 had the Euro High too far SE, then T+96 was very good; low off NW Scotland slightly too far north.

UKMO had the Euro High way too far north-west at T+144 with the high over eastern England giving southerlies elsewhere. T+120 was better but the high was still too dominant, then T+96 was reasonably accurate with the low off NW Scotland well positioned; a bit too far north as on GFS. Overall a 6/10; I don't think it was generally as good as GFS, committing the common UKMO error of overdoing high pressure over Europe.

ECMWF had too straight a westerly flow at T+144, with a weak low to the NW. T+120 was similar to GFS with the Euro High too far SE and a strong WSW flow, and T+96 was accurate. Overall very little to choose between this and the GFS in terms of accuracy, so a 7/10.

NOGAPS was poor at T+144 with low pressure in charge off W Scotland and a cyclonic SW flow, but was very good at T+120. Then at T+96 the Euro High was overdone and the low pressure significantly too far north. Overall, though, at least it had the SW'lys in charge; a 6/10 overall.

JMA was unavailable at T+120. At both T+144 and T+96 the Euro High was overdone, at T+144 covering south-east England with lows well to the north, and at T+96 too far northeast giving southerly winds; overall 5/10.

GEM was okay at T+144, lows too far east really with a colder WNW'ly moving towards western areas but a SW'ly over Britain with a correctly positioned Euro High. T+120 overdid the Euro High slightly, but T+96 was reasonable- again, low off NW Scotland slightly too far north. A 7/10.

ECMWF, GFS and GEM were okay, although on the whole the models struggled with the Euro High, tending to overdo it especially in the case of JMA; UKMO was particularly strong in overdoing it at T+144.

[Thundery wintry showers]

16 January

A ridge of high pressure over most of Britain, but with a shallow low off SW England, and 976mb low fast approaching western Scotland. Southern England lay to the south of the polar front where it was cloudy and mild, while in the north it was sunny and cold with, for many, the first air frost of the year- and even some snow for the Central Lowlands of Scotland later on.

I thought GFS had handled this one well, looking at the outputs- and indeed it had. The ridge of high pressure was generally slightly underdone at all three timeframes (most notably T+144) with a stronger high to the SE than actually occurred, but slack westerlies changing to southerlies were shown, and it was shown as being cold in the north, mild in the south. 8/10.

UKMO had us in a Bartlett at T+144, mild SW winds for all. T+120 was better but still had the colder, clearer air too far north and too strong a high to the SE, only the far north of Scotland was cold and clear. T+96 was quite accurate. Overall a 5/10; it overdid the Euro High.

ECMWF was good at T+144, but at T+120 it had the low off SW England too far east, residing over East Anglia instead, and T+96 had too pronounced a ridge of high pressure with the large Atlantic low too far west. Overall a good set of outputs, but slightly less good than GFS, so 7/10.

NOGAPS was okay at T+144 and T+120 but had the ridge too far west, too strong a low to our north, and thus, too high a chance of wintry showers. T+96 was again erroneous in having lowest pressure too far west; overall 6/10.

JMA available only at T+120 and T+96. At both timeframes the low in the Atlantic was too far east, especially T+96 where most of Britain was under southerlies; overall just 5/10.

GEM was good at T+120, but at T+144 and T+96 the Atlantic low was too far east, giving southerlies for all; 5/10.

Overall GFS handled this one best, but ECMWF was pretty good too.