Better Than The Models ?

[Roger J Smith]

Long-time model watchers are familiar with a phenomenon on the GFS known as the "dartboard low" which sometimes appears on the output at about day 6 through day 12 and then gradually fills to less spectacular central pressures, often with a significant deviation in track as well. I remember back in October 2005 the remnants of hurricane Wilma were at one point coming in at day 6 or so as a 950 mb low but eventually it arrived more like day 7 as a 975 mb low.

There have been a few other examples, so at this point, I am not getting very excited about the 945 mb low on the current day 10-11 output. The reason why the GFS is developing lows to this radical extent would seem to be a signal from an extreme deepening of the eastern Canada PV, note on these same charts 500 mb temps of -50 C. I haven't looked at the thickness charts yet but I would imagine they are showing extremely low values near Baffin Island and western Greenland.

I'm sure people will follow this storm evolution very closely but two things about it at this point, it is coming from that WSW direction that I mentioned for major storm development, and there will be issues about the increased randomness of prediction outside the boundary of a stated forecast, but if it came off the way it is now depicted, people would certainly want to know more about the reasoning you apply to severe storm timing. Like I said early on in this discussion, the science is not so well developed that anyone is currently holding long-range forecasters to nearest day precision on 90-day or even 30-day outlooks, and we need to go one step at a time rather than expecting this "eureka" moment -- I think the atmosphere is such a complex system that understanding it on this sort of time scale is going to come about very gradually (as in fact it already is doing). There is also a pretty strong chance of this GFS storm being a partial or total "figment" so I would keep watching for a few more days before getting too excited about it.

As to why it might come three days late (overlooking the track difference), from my research perspective, that could be a response to timing line structure, in other words, the right concept of storm development but applied to one timing line downstream from where it actually takes place (since the time separation averages 3 days). Here again, my research tends to support a number of other energy peaks beside new and full moon, so if the postulated energy peak was the 3 Feb new moon, that could trigger the storm on timing line 2 (eastern N America) and send it on to timing line 3 for the next energy peak, which I have timed as 6-7 Feb. There is a complete list of energy peaks in our LRF in the winter discussion thread (somewhere around end of Nov not in the first post of the thread).

The smart money on this GFS forecast would say less intense low somewhat further north, but in December the GFS kept trying to move lows too far north with an average 5-day error of about 7 deg of latitude (which was often taking lows into Italy rather than Germany).

I've done today's stat validation for the GFS and am waiting for the ECM to finish the post.

[MurcieBoy]

Many thanks indeed for the thorough explanation RJS. I must one day read more about your approach - your concept re energies at Full/New Moons etc is beyond me, but sounds truly fascinating nonetheless.

[LomondSnowstorm]

Hmm -

[Roger J Smith]

Day 7 validation on the 3 Feb maps

_____________________________________

GFS continues to show rather weak systems across the northern half of the grid, but today has weakened the high pressure centre it has been advertising near the Channel, to values below 1030 mbs. Otherwise it continues similar to its previous trends and shows a correlation of 0.84 with the previous day. The continued weakening trend of pressure gradient across Scandinavia has dropped the MB correlation to -0.22, while the correlation with RJS continues about the same as day 8, at 0.56 now.

ECM has sharpened the trough near 0-5E but other features remain similar, giving a correlation of 0.64 from the previous day's output, and 0.81 with the GFS for this run. The correlation with MB has jumped up considerably today, to 0.48, but a sector analysis shows this is all in the region south of 55N, so there is little change in pattern similarity across the northern half of the grid. The RJS correlation with the ECM has fallen to 0.17 due to the phase difference in the low (RJS has this west of Ireland, ECM east of the UK).

We are now apparently entering a period of model divergence, there are no increases in model correlation as a general trend across the northern half of the grid, only slowly divergent solutions. This seems like a typical response to uncertainty in breaking down a block that global models now have as an initial condition. The 7d UKMO and GEM models (from extrapolation) have nothing very different to say on this evolution.

Thread followers will probably be more interested now in day 10-11 output on GFS where there are signs of major storm development. See my previous post for a discussion of that. Finally, ECM day 9 is also showing intense storm development southwest of Iceland. Presume this will be heading for Shetland region on day 10 but have to head out and post this now. Same comments would apply here, but interesting to see the GFS-ECM consensus on strong development at this time scale for regions just north of Scotland. This may develop into something similar to the Jan 1990 windstorm scenario.

[OldMetMan]

I have been studying the model output closely today and there are 3 factors that are niggling me, and this usually means that they are relevant! I have already made reference to them previously but I will expand.

1. NOGAPS today continues with the trend from yesterday of veering the jet from 1st February to a more NW or WNW direction. The flow is shown to become much more zonal by then but the point is that systems will come at us from the NW. All along I have felt the models were handling the movement of the train of lows coming out of the US the wrong way. I base this on the synoptic development this month so far and the fact that I have more often seen HP retrogress when in its current position that progress E or SE. The models are keeping the low procession on a very similar track heading NE and then E for an extended period and that just does not look right to me.

MB, the synoptic development depicted in your forecast is for low pressure to be coming from the NW. From where we are right now, in order for this to happen, we need to see the jet back in direction off the US eastern seaboard and for the main Atlantic HP to take over in the W Atlantic, gradually moving NE to mid-Atlantic. It's been trying to do this all week:

This would all allow the upper long wave pattern to amplify a bit more, it's a bit too flat at the moment. This leads me to the next point:

2. The GFS shows some quite intense short-wave features moving NE to the NW of Scotland and ending up in Scandinavia around 2nd Feb.. These are the kind of depressions that could give the strength of wind in the Storm forecast, but their movement is too far NE. An amplification in the upper wave pattern would turn the trajectory more E and SE, which is what is hinted at in the NOGAPS. There's plenty of energy there in this strengthening jet, it's just going the wrong way!

3. We still have a strong jet way, way south over N Africa which is pretty unusual. So there is also an associated upper trough with this bringing unsettled to stormy conditions in the Mediterranean. In the past, I have seen an upper wave development take place that brings a deep trough S to "join up" as it were, with the more southerly jet. For this to happen it would need a significant low development or developments W of Scandinavia, down the North Sea and into W Europe. The upper pattern then becomes - deep trough over Canada and US; upper ridge rebuilding from over and to the W of the UK to be positioned W to mid-Atlantic and a deep and very long wave trough going as far S as N Africa. That is simplified of course, but that is the general idea.

I must stress at this point that the above is based solely on my knowledge and experience of Atlantic and European surface and upper air patterns gained over many years' of poring over synoptic charts. So, I will quite understand if any of the more technically minded posters shoot me down in flames!

The interesting time will be when the developing, very strong US/Canada jet heads out into the Atlantic within the next 3 or 4 days. In the long term GFS, as Roger as has already stated, shouldn't be taken very seriously. I want to see what the UKMO says for 120hrs later today as I shall be curious to see whether it, too decides to start turning the depression paths in a more E direction.

I have just checked the ECM 12z run and this IS interesting, and a complete departure from previous runs:

That is some gradient off the NE of Scotland!

Edited January 27, 2011 by OldMetMan

[MurcieBoy]

Wow! Thanks very much indeed for that analysis OldMetMan. This is exciting!

According to the Storm forecast, the Arctic winds will change direction by 1 Feb to from Westerly to North Westerly.

The ECM 12z chart for 3 Feb you provided was highlighted by LS too. For me, that chart shows my "methods" classic expectation around the UK - just a wee bit further south for that time (nb: I appreciate the Storm forecast is different for 3 Feb)

[carnoustie]

Wow! Thanks very much indeed for that analysis OldMetMan. This is exciting!

According to the Storm forecast, the Arctic winds will change direction by 1 Feb to from Westerly to North Westerly.

The ECM 12z chart for 3 Feb you provided was highlighted by LS too. For me, that chart shows my "methods" classic expectation around the UK - just a wee bit further south for that time (nb: I appreciate the Storm forecast is different for 3 Feb)

Great thread Murcia Boy. Hope your forecast verifies. It is quite unbelievable some of the nonsense you have had to put up with.

And thank you OldMetMan. I enjoy reading your expert analysis - absolute quality. Yourself, RJSmith and Bluearmy (Nick L) are far and away the most knowledgeable posters on this and other UK weather forums (TWO and UKWW).

Edited January 27, 2011 by carnoustie

[weatherguy]

Good read OldMetMan, thanks

Regardless of the end result, it's amazing to think that amongst all the "boring" high pressure we've had for so long, the dates MB has predicted a storm for so long ago, looks to be showing up at all (regardless of exact location/strength). I know it's not a spectacular forecast to say there will be a storm on x date, but it does show a bit of credibility for it to happen out of sheer probability from a long way away.

Here's an interesting chart anyway...

[OldMetMan]

Wow! Thanks very much indeed for that analysis OldMetMan. This is exciting!

According to the Storm forecast, the Arctic winds will change direction by 1 Feb to from Westerly to North Westerly.

The ECM 12z chart for 3 Feb you provided was highlighted by LS too. For me, that chart shows my "methods" classic expectation around the UK - just a wee bit further south for that time (nb: I appreciate the Storm forecast is different for 3 Feb)

You're welcome MB. I hadn't noticed the reference LS made, sorry. But it's a remarkable turnaround, it will be interesting to see if any of the other models pick up on it.

[MurcieBoy]

But it's a remarkable turnaround, it will be interesting to see if any of the other models pick up on it.

Indeed!

[OldMetMan]

Great thread Murcia Boy. Hope your forecast verifies. It is quite unbelievable some of the nonsense you have had to put up with.

And thank you OldMetMan. I enjoy reading your expert analysis - absolute quality. Yourself, RJSmith and Bluearmy (Nick L) are far and away the most knowledgeable posters on this and other UK weather forums (TWO and UKWW).

Thank you carnoustie, you are most kind.

Good read OldMetMan, thanks

Many thanks weatherguy.

[BLAST FROM THE PAST]

I have been studying the model output closely today and there are 3 factors that are niggling me, and this usually means that they are relevant! I have already made reference to them previously but I will expand.

1. NOGAPS today continues with the trend from yesterday of veering the jet from 1st February to a more NW or WNW direction. The flow is shown to become much more zonal by then but the point is that systems will come at us from the NW. All along I have felt the models were handling the movement of the train of lows coming out of the US the wrong way. I base this on the synoptic development this month so far and the fact that I have more often seen HP retrogress when in its current position that progress E or SE. The models are keeping the low procession on a very similar track heading NE and then E for an extended period and that just does not look right to me.

MB, the synoptic development depicted in your forecast is for low pressure to be coming from the NW. From where we are right now, in order for this to happen, we need to see the jet back in direction off the US eastern seaboard and for the main Atlantic HP to take over in the W Atlantic, gradually moving NE to mid-Atlantic. It's been trying to do this all week:

This would all allow the upper long wave pattern to amplify a bit more, it's a bit too flat at the moment. This leads me to the next point:

2. The GFS shows some quite intense short-wave features moving NE to the NW of Scotland and ending up in Scandinavia around 2nd Feb.. These are the kind of depressions that could give the strength of wind in the Storm forecast, but their movement is too far NE. An amplification in the upper wave pattern would turn the trajectory more E and SE, which is what is hinted at in the NOGAPS. There's plenty of energy there in this strengthening jet, it's just going the wrong way!

3. We still have a strong jet way, way south over N Africa which is pretty unusual. So there is also an associated upper trough with this bringing unsettled to stormy conditions in the Mediterranean. In the past, I have seen an upper wave development take place that brings a deep trough S to "join up" as it were, with the more southerly jet. For this to happen it would need a significant low development or developments W of Scandinavia, down the North Sea and into W Europe. The upper pattern then becomes - deep trough over Canada and US; upper ridge rebuilding from over and to the W of the UK to be positioned W to mid-Atlantic and a deep and very long wave trough going as far S as N Africa. That is simplified of course, but that is the general idea.

I must stress at this point that the above is based solely on my knowledge and experience of Atlantic and European surface and upper air patterns gained over many years' of poring over synoptic charts. So, I will quite understand if any of the more technically minded posters shoot me down in flames!

The interesting time will be when the developing, very strong US/Canada jet heads out into the Atlantic within the next 3 or 4 days. In the long term GFS, as Roger as has already stated, shouldn't be taken very seriously. I want to see what the UKMO says for 120hrs later today as I shall be curious to see whether it, too decides to start turning the depression paths in a more E direction.

I have just checked the ECM 12z run and this IS interesting, and a complete departure from previous runs:

That is some gradient off the NE of Scotland!

An excellent post this and one worth reading and taking note of.

Very interesting BOM showing the potential storm on a more southerly track but I suspect that we will see more chopping and changing as we get nearer the time as there potentially is a lot to go on. The increased amplification in the Atlantic is a development worth watching as I too think HP being pushed E or SE looks wrong. Anyway D Day approaches and this is what we are here for.

BFTP

Edited January 28, 2011 by BLAST FROM THE PAST

[nking]

Today’s posts (27/1) have been a pleasure to read. It’s so nice and refreshing to see a balanced and unbiased critique of the situation.

[MurcieBoy]

Here's the comparison GFS 18z and ECM 12z with the 3 Feb 2011 Storm forecast chart:

[Paul Sherman]

Hi Nick

Any chance of doing 1 of those for all the days as well or is that too much work. Would be good to see the 28th until 5th February so we can look back and see how close individual days are.

Many Thanks

Paul S

[OldMetMan]

An excellent post this and one worth reading and taking note of.

Very interesting BOM showing the potential storm on a more southerly track but I suspect that we will see more chopping and changing as we get nearer the time as there potentially is a lot to go on. The increased amplification in the Atlantic is a development worth watching as I too think LP being pushed E or SE looks wrong. Anyway D Day approaches and this is what we are here for.

BFTP

Thanks BFTP, just had a look at the UKMO 144hrs, it seems to be following the NOGAPS trend, shifting more towards a NW flow. The GFS is rather similar to its earlier run so we'll see what tomorrow looks like!

[Coast]

Met Office coastal wind forecast for today here:

http://www.metoffice.gov.uk/weather/marine/inshore_forecast.html#All

Met Office high seas forecast and storm warnings here:

http://www.metoffice.gov.uk/weather/marine/highseas_forecast.html#All~All

Shipping forecast and gale warnings here:

http://www.metoffice.gov.uk/weather/marine/shipping_forecast.html#All~All

Beaufort scale explained with equivalents in MPH etc here:

http://en.wikipedia.org/wiki/Beaufort_scale

Latest marine observations here:

http://www.metoffice.gov.uk/weather/marine/observations/

Live data form the K7 buoy (60.701N, 4.500W) here:

http://www.ndbc.noaa.gov/radial_search.php?lat1=60.701N&lon1=4.500W&uom=E&dist=250

Others here:

http://www.ndbc.noaa.gov/maps/United_Kingdom.shtml

Forecast chart for 12.00 hrs:

Met Office view of 0000 UTC surface analysis

High pressure is centred over Scotland, covering the UK. High pressure centres also lie in the Atlantic and over Denmark. Low pressure centres to the west of Norway and east of Greenland have associated fronts lying over northern Scandinavia and Iceland. Low pressure also lies to the west of Spain and in the Mediterranean.

http://www.metoffice.gov.uk/weather/uk/surface_pressure.html

[MurcieBoy]

Ok Paul S, here are the comparison charts for 1-5 Feb at 28 Jan 0z:

1 Feb

2 Feb

3 Feb

4 Feb

5 Feb

28 January 2011 Forecast v Actual

Here is the latest comparison chart (with detailed 28 Jan 6z METO FAX chart):

[Paul Sherman]

Many Thanks for doing that Nick

28th Looks good pressure slighly lower but was around the 1040mb Mark just a few days ago, the Isobars further North also good running from the west into Scandinavia.

Paul S

[MurcieBoy]

Many Thanks for doing that Nick

28th Looks good pressure slighly lower but was around the 1040mb Mark just a few days ago, the Isobars further North also good running from the west into Scandinavia.

Paul S

My pleasure Paul, it was a good idea!

Edited January 28, 2011 by MurcieBoy

[Coast]

OK, a few midday current analysis charts coming out so you can compare the first milestone in MB's forecast. I am aware that Nick has his chart drawn for 14.00 hrs, but many sources only churn out at 6 hours intervals. I'll try and grab something nearer 14.00 if I can:

Weather symbol legends

Red (upper-left): temperature in deg C

Green (lower-left): dewpoint in deg C

Blue (upper-right): abbreviated mean sea-level pressure in hPa (=mb). To get the pressure value, just put a 10 or a 9 in front of the displayed value and divide by 10: only one of these choices will give a realistic value.

Magenta (center-left): weather symbol

Blue (center): cloud cover in octa

Blue (barb): wind speed in knots and wind direction

Orange: pressure tendency in the last 3 hours in tenth of hPa

Purple (left-upper-left): Wind gust in knots

Upper-center and lower-center (few stations): cloud type symbols

Black lines: pressure (hPa=mb) lines each 4 hPa analyzed with Barnes algorithm from surface observations and first guess model data (digital filter)

Red dashed-lines: geopotential height at 500 hPa each 6 dam analyzed with Barnes algorithm from upper-air observations and first guess model data

[Coast]

MetO midday fax not out yet, here is a a current one for pressure and wind but is quite tight around the UK boarders:

[Coast]

MetO is out and I have tried to superimpose Nick's chart onto it. Please note, I have not altered the position or any of the fronts/pressure lines on the attached. I have altered the tint of the original to differentiate it from the black and white of the MetO fax.

I have also resized it to fit the slightly different MetO map as the original from MB was 'flat' and the MetO one is more representative of the curvature of the earth, it still fits the coastlines pretty well. I hope you can make out the two distinct charts. Below this are the originals:

It seems for this one we don't have the high winds (80mph) in Scotland or the low pressure of 965mb at 70N, 20E (off the top of Finland) although there is an area of 985mb very close by. The front across Scandinavia through the middle of Europe down into Southern France is actually two fronts, but both are broadly in line with Nick's synopsis albeit separate. The 1040mb high off Ireland into South West England is not the same shape, but it is roughly in the same area as the 1028mb pressure reading taken around the same time. Generally the isobars seem to have the same spacing over Scotland and are in a similar orientation. Nick's high over Spain is in the same area as the current low of 1005 mb.

[Coast]

Next milestone is Sunday. I might not be around for it, so if someone could stick up the 12Z MetO fax that would be jubbly!

Here's MB's next phase of the forecast:

[Great Plum]

http://www.wetterzentrale.de/pics/Rtavn1681.png

Hmmm...