Winter Model Discussion 18Z 5/2/13 onwards.

[fergieweather]

Not really sure where all this misguided 'GFS is rubbish' twaddle emerged on here (we're discussing computer models, not football teams)... but to return to the focal point of Sunday, which we anticipate could deliver a significant snow event:

At stake here are some marked differences in the rate of E'ward frontal progression into the UK between the 12z UKMO-GM and other operational centre output, notably ECMWF. These appear correlated to the rapidity and scope at which the amplifying upper ridge over the country on Saturday cuts-off. For example, the 12z ECMWF deterministic is markedly slower with this process compared to UKMO-GM, with net result of retarding the frontal progression to a greater extent. This slower solution is favoured by Exeter, who have modified the GM accordingly (away from raw version you'll see on websites).

Clearly this has profound implications in what will prove a high-stakes forecast, should current evolution continue in similar tone. Bear in mind the frontal zone in question is only just now taking shape off E coast of N America... still a way to go on this story.

However, the slower EC (and presently modified GM) solution offers less threat of disruptive snow into London during Sunday, as opposed to raw GM, which did. Hence, this sort of issue - high stakes, as I suggest - remains a clear uncertainty.

The consensus solution follows EC's snow accumulation prognosis (sorry I'm not allowed to share that as actual mapping), which - with frontal leading edge moving into sub-300m WBFL air across much of S England & Midlands/Wales - looks a decidedly bothersome blend. EC offers as much as 10-15cm snow anywhere from approx central Somerset up through Bristol, eastwards to around Reading and north through pretty much all of Wales and West Mids (some peak totals in these areas excess of 20cm). To the SE, similar amounts (to around 10cm) across Wilts, Hants, to W Sussex. Clearly, distribution is uncertain but in broad terms, the combined weight of EPS and deterministic output is indicative of parts of W/S Scotland, eastern Northern Ireland, N England, Wales, West Country north of Taunton, West Midlands and southern-central England being most at risk of disruptive snowfall durIng Sunday and overnight into early Monday. Currently, Exeter suggest 30% chance of some snow into London itself later on Sunday, albeIt any ccumulations here in the capital are thought likely to be pretty small.

Anyway, that's some detailed thoughts for now, ahead of whatever the 00z runs might yield.

Edited February 5, 2013 by fergieweather

[london-snow]

So GFS is rubbish in the short term & completely wrong, but it's quite okay to put up charts from t300+ because they show a GH. You ave to laarf!

Regarding the short term output from the gfs it is showing a lack of consistency in the or just a tad over reliable time frame with each run it has produced it either moves towards the likes of the UKMO & ECM or drops it and goes its merry way.

Longer term and looking at the possibility of a GH all we can say atm that it has been showing this early trend now for the last four outputs. However atm it remains a potential possibility and that's all.

It says something that a so called top model shows no consistency in the reliable but in the longer term shows a trend.

[pages]

Im a little concerned after viewing the 18Z GEFS ensembles.

http://modeles.meteociel.fr/modeles/gens/graphe_ens3.php?x=306&ext=1&y=141&run=18&ville=Londres&runpara=0

We have had alot of scatter around the 9th as I highlighted earlier today but the 18Z ensembles show relatively good agreement for the GFS Op.

http://modeles.meteociel.fr/modeles/gens/graphe_ens3.php?x=306&ext=1&y=141&run=18&ville=Londres&runpara=0

Whilst I hope the UKMO/ECM continue with the same trend these ensembles have placed a seed of doubt in my mind. Lets hope this is just the GFS/GEFS being useless once again.

Dont worry the ECM and nogaps have been consistent for days and have the Bom, CMA and gem backing them up, I place far more faith in those 5 op runs at t120 than 20 ensemble who are wrong to start, with the data being tweaked. I don't know why people look at ensembles surely it more valid looking at the mean of the said op runs as they have the correct(hopefully) starting data.

Ensembles being tweaked t0 must be like the butterfly effect how can the output be trusted when its wrong to start.

[Captain Shortwave]

? Shortwave dropping SE +50% of the ensembles show this and a lot look similar to the ECM/UKMO...

Look at the timing on the gfs runs to the ecm op, it's a good 18 hours earlier and further west, the ecm op drops the shortwave south East into Russia. By then you have pushed more energy into the developing high pressure cell. GFS drops the shortwave whilst the high is weak and disrupts it. ECM holds the shortwave above the high for longer and drops it into a nice position with a cut off high pressure cell.

Of course it might not be just the shortwave gfs is modelling wrong, an increase in strength of the high would prevent the shortwave moving south earlier and push it further east into a more favourable position. That shortwaves trajectory is more complex than I probably meant it to sound like, many factors will affect its movement.

Edited February 5, 2013 by Captain shortwave

[chris55]

Not really sure where all this misguided 'GFS is rubbish' twaddle emerged on here (we're discussing computer models, not football teams)... but to return to the focal point of Sunday, which we anticipate could deliver a significant snow event:

At stake here are some marked differences in the rate of E'ward frontal progression into the UK between the 12z UKMO-GM and other operational centre output, notably ECMWF. These appear correlated to the rapidity and scope at which the amplifying upper ridge over the country on Saturday cuts-off. For example, the 12z ECMWF deterministic is markedly slower with this process compared to UKMO-GM, with net result of retarding the frontal progression to a greater extent. This slower solution is favoured by Exeter, who have modified the GM accordingly (away from raw version you'll see on websites).

Clearly this has profound implications in what will prove a high-stakes forecast, should current evolution continue in similar tone. Bear in mind the frontal zone in question is only just now taking shape off E coast of N America... still a way to go on this story.

However, the slower EC (and presently modified GM) solution offers less threat of disruptive snow into London during Sunday, as opposed to raw GM, which did. Hence, this sort of issue - high stakes, as I suggest - remains a clear uncertainty.

The consensus solution follows EC's snow accumulation prognosis (sorry I'm not allowed to share that as actual mapping), which - with frontal leading edge moving into sub-300m WBFL air across much of S England & Midlands/Wales - looks a decidedly bothersome blend. EC offers as much as 10-15cm snow anywhere from approx central Somerset up through Bristol, eastwards to around Reading and north through pretty much all of Wales and West Mids (some peak totals in these areas excess of 20cm). To the SE, similar amounts (to around 10cm) across Wilts, Hants, to W Sussex. Clearly, distribution is uncertain but in broad terms, the combined weight of EPS and deterministic output is indicative of parts of W/S Scotland, eastern Northern Ireland, N England, Wales, West Country north of Taunton, West Midlands and southern-central England being most at risk of disruptive snowfall durIng Sunday and overnight into early Monday. Currently, Exeter suggest 30% chance of some snow into London itself later on Sunday, albeIt any ccumulations here in the capital are thought likely to be pretty small.

Anyway, that's some detailed thoughts for now, ahead of whatever the 00z runs might yield.

Excellent stuff Ian, many thanks for sharing. Long way to go on this yet but EC has been pretty consistent with this for a few days now.

Edited February 5, 2013 by chris55

[NickR]

Not really sure where all this misguided 'GFS is rubbish' twaddle emerged on here (we're discussing computer models, not football teams)... but to return to the focal point of Sunday, which we anticipate could deliver a significant snow event:

At stake here are some marked differences in the rate of E'ward frontal progression into the UK between the 12z UKMO-GM and other operational centre output, notably ECMWF. These appear correlated to the rapidity and scope at which the amplifying upper ridge over the country on Saturday cuts-off. For example, the 12z ECMWF deterministic is markedly slower with this process compared to UKMO-GM, with net result of retarding the frontal progression to a greater extent. This slower solution is favoured by Exeter, who have modified the GM accordingly (away from raw version you'll see on websites).

Clearly this has profound implications in what will prove a high-stakes forecast, should current evolution continue in similar tone. Bear in mind the frontal zone in question is only just now taking shape off E coast of N America... still a way to go on this story.

However, the slower EC (and presently modified GM) solution offers less threat of disruptive snow into London during Sunday, as opposed to raw GM, which did. Hence, this sort of issue - high stakes, as I suggest - remains a clear uncertainty.

The consensus solution follows EC's snow accumulation prognosis (sorry I'm not allowed to share that as actual mapping), which - with frontal leading edge moving into sub-300m WBFL air across much of S England & Midlands/Wales - looks a decidedly bothersome blend. EC offers as much as 10-15cm snow anywhere from approx central Somerset up through Bristol, eastwards to around Reading and north through pretty much all of Wales and West Mids (some peak totals in these areas excess of 20cm). To the SE, similar amounts (to around 10cm) across Wilts, Hants, to W Sussex. Clearly, distribution is uncertain but in broad terms, the combined weight of EPS and deterministic output is indicative of parts of W/S Scotland, eastern Northern Ireland, N England, Wales, West Country north of Taunton, West Midlands and southern-central England being most at risk of disruptive snowfall durIng Sunday and overnight into early Monday. Currently, Exeter suggest 30% chance of some snow into London itself later on Sunday, albeIt any ccumulations here in the capital are thought likely to be pretty small.

Anyway, that's some detailed thoughts for now, ahead of whatever the 00z runs might yield.

N England meaning the NW and central-UK areas rather than NE? (sorry, it's just that "N England" tends to be used to refer to quite a wide and diverse area)

[Mucka]

Well I think the ECM ensembles are showing the same - they do not seem to have been posted tonight, probably because they show that after the slider a return to less cold conditions seems to be favoured..

I analyse the period from the 13th in detail on the previous page along with the ensembles.

Here they are for London

While it is true scatter appears after 13th it is at least encouraging there is till some tight clustering for cold from there. Obviously we would expect the increased scatter thereafter.

N England meaning the NW and central-UK areas rather than NE? (sorry, it's just that "N England" tends to be used to refer to quite a wide and diverse area)

Where snow is involved this winter, Northern England never means NW England. Even Northwest England doesn't mean my part of NW England.

[NickR]

I analyse the period from the 13th in detail on the previous page along with the ensembles.

Here they are for London

While it is true scatter appears after 13th it is at least encouraging there is till some tight clustering for cold from there. Obviously we would expect the increased scatter thereafter.

Where snow is involved this winter, Northern England never means NW England. Even Northwest England doesn't mean my part of NW England.

Hard to tell when you don't have your location in your sig. BUt parts of NW England have done VERY well for snow this year, including a day or 2 ago!

[london-snow]

Not really sure where all this misguided 'GFS is rubbish' twaddle emerged on here (we're discussing computer models, not football teams)... but to return to the focal point of Sunday, which we anticipate could deliver a significant snow event:

At stake here are some marked differences in the rate of E'ward frontal progression into the UK between the 12z UKMO-GM and other operational centre output, notably ECMWF. These appear correlated to the rapidity and scope at which the amplifying upper ridge over the country on Saturday cuts-off. For example, the 12z ECMWF deterministic is markedly slower with this process compared to UKMO-GM, with net result of retarding the frontal progression to a greater extent. This slower solution is favoured by Exeter, who have modified the GM accordingly (away from raw version you'll see on websites).

Clearly this has profound implications in what will prove a high-stakes forecast, should current evolution continue in similar tone. Bear in mind the frontal zone in question is only just now taking shape off E coast of N America... still a way to go on this story.

However, the slower EC (and presently modified GM) solution offers less threat of disruptive snow into London during Sunday, as opposed to raw GM, which did. Hence, this sort of issue - high stakes, as I suggest - remains a clear uncertainty.

The consensus solution follows EC's snow accumulation prognosis (sorry I'm not allowed to share that as actual mapping), which - with frontal leading edge moving into sub-300m WBFL air across much of S England & Midlands/Wales - looks a decidedly bothersome blend. EC offers as much as 10-15cm snow anywhere from approx central Somerset up through Bristol, eastwards to around Reading and north through pretty much all of Wales and West Mids (some peak totals in these areas excess of 20cm). To the SE, similar amounts (to around 10cm) across Wilts, Hants, to W Sussex. Clearly, distribution is uncertain but in broad terms, the combined weight of EPS and deterministic output is indicative of parts of W/S Scotland, eastern Northern Ireland, N England, Wales, West Country north of Taunton, West Midlands and southern-central England being most at risk of disruptive snowfall durIng Sunday and overnight into early Monday. Currently, Exeter suggest 30% chance of some snow into London itself later on Sunday, albeIt any ccumulations here in the capital are thought likely to be pretty small.

Anyway, that's some detailed thoughts for now, ahead of whatever the 00z runs might yield.

Hi Ian,

The last time i heard that exeter was not favouring the ecm idea past the weekend which has been shown over the past few days now.

Given the consistent pattern the ecm has been showing what is exeter feelings currently beyond the weekend?

Thanks

[Mucka]

Hard to tell when you don't have your location in your sig. BUt parts of NW England have done VERY well for snow this year, including a day or 2 ago!

You calling me a liar?!

The best I have done is brief patchy slushy dusting.

Southwest Manchester

Edited February 5, 2013 by Mucka

[Dunstable Snow]

Not really sure where all this misguided 'GFS is rubbish' twaddle emerged on here (we're discussing computer models, not football teams)... but to return to the focal point of Sunday, which we anticipate could deliver a significant snow event:

At stake here are some marked differences in the rate of E'ward frontal progression into the UK between the 12z UKMO-GM and other operational centre output, notably ECMWF. These appear correlated to the rapidity and scope at which the amplifying upper ridge over the country on Saturday cuts-off. For example, the 12z ECMWF deterministic is markedly slower with this process compared to UKMO-GM, with net result of retarding the frontal progression to a greater extent. This slower solution is favoured by Exeter, who have modified the GM accordingly (away from raw version you'll see on websites).

Clearly this has profound implications in what will prove a high-stakes forecast, should current evolution continue in similar tone. Bear in mind the frontal zone in question is only just now taking shape off E coast of N America... still a way to go on this story.

However, the slower EC (and presently modified GM) solution offers less threat of disruptive snow into London during Sunday, as opposed to raw GM, which did. Hence, this sort of issue - high stakes, as I suggest - remains a clear uncertainty.

The consensus solution follows EC's snow accumulation prognosis (sorry I'm not allowed to share that as actual mapping), which - with frontal leading edge moving into sub-300m WBFL air across much of S England & Midlands/Wales - looks a decidedly bothersome blend. EC offers as much as 10-15cm snow anywhere from approx central Somerset up through Bristol, eastwards to around Reading and north through pretty much all of Wales and West Mids (some peak totals in these areas excess of 20cm). To the SE, similar amounts (to around 10cm) across Wilts, Hants, to W Sussex. Clearly, distribution is uncertain but in broad terms, the combined weight of EPS and deterministic output is indicative of parts of W/S Scotland, eastern Northern Ireland, N England, Wales, West Country north of Taunton, West Midlands and southern-central England being most at risk of disruptive snowfall durIng Sunday and overnight into early Monday. Currently, Exeter suggest 30% chance of some snow into London itself later on Sunday, albeIt any ccumulations here in the capital are thought likely to be pretty small.

Anyway, that's some detailed thoughts for now, ahead of whatever the 00z runs might yield.

Cracking analysis as usual so thank you... really appreciated. I would say that my team fulham are about as effective again Man utd, Chelseaand Man City as the GFS is when calling the antlantic against a cold block from the north East as modelled now until we get into T48 ? Is my analysis fair ?!

[NickR]

You calling me a liar?!

The best I have done is brief patchy slushy dusting.

Southwest Manchester

LOL - not at all... but it has to be said that a LOT of NW Eng has done very well - including many parts of Manchester... they've certainly seem pretty wild snow/wind conditions both with the big front a few weeks back and a few days ago.

[TEITS]

Not really sure where all this misguided 'GFS is rubbish' twaddle emerged on here (we're discussing computer models, not football teams)...

At times the GFS can be rubbish. This was clearly seen in the January cold spell when it took ages for the model to agree with the rest. Remember the system that moved S across the UK? The GFS had this moving NE into Scandi!

When you have followed the GFS for many years you get to learn its biases. One of the main flaws with the GFS is it has a eastward bias whenever blocking is indicated to our NE. Another classic bias of the GFS is to overdeepen LP systems.

What members have to realise is all these models use roughly the same data but what is it that makes the likes of the ECMWF more reliable than say the NOGAPS. The answer is simply down to how they are programmed.

Finally im not sure I agree with the snowfall risk areas you have highlighted. At this stage I feel Ireland/Wales/SW England are more likely to see rain than snow with the risk of snow being further E than you have highlighted. The reason being is I feel the ECM & latest Fax charts are wrong and we shall see a slight shift E in the morning. The GFS on the other hand is far to E with the pattern.

Thanks for sharing the info though. Im going to post my snow risk map in the morning.

Edited February 6, 2013 by THE EYE IN THE SKY

[fergieweather]

Courtesy of a special agreement with UKMO to show select bits of modified output where it helps illustrate and explain their scientific rationale, I can show you just two frames from modified 12z GM, which you can then compare with the raw output. The snowy PPN signal (as x's) needs no further explanation, albeit the very obvious uncertainties at this range need re-stressing.

[Tamara]

We are a long way from having to worry about what might happen after the sliding low from the weekend. I think of this the same way as when we were looking at whether we could even get to a situation where we might get an extension of cold weather from the current northerly - back at the start of last weekend when the present enticements were more than a week away. The ECM at that time had only just started ploughing its lone furrow towards what we see looming (GFS hissy fit disclaimer attached )

So, now, as people are starting to look another week ahead and speculating about what may/may not happen (after *hopefully* a few snowmen have been made up and down the country ) we might want to look to see if there is another repeated pattern to come to give us some more intrigue and another clue. Not yet and I have some doubts we will repeat the same experience of seeing a model pursuing a lone course in the same way that the ECM has been the frosty persistent outlier that has gathered support like a cold rolling snowball in recent days (*GFS erratic child disclaimer once more*) I think the forthcoming 7 to 10 day period at this time is looking much trickier to plot/speculate than the fun period from 2nd/3rd/4th Feb to 9th/10th/11th Feb has been

[Dunstable Snow]

Courtesy of a special agreement with UKMO to show select bits of modified output where it helps illustrate and explain their scientific rationale, I can show you just two frames from modified 12z GM, which you can then compare with the raw output. The snowy PPN signal (as x's) needs no further explanation, albeit the very obvious uncertainties at this range need re-stressing.

Very frustrating for where i am ! half an hour in the car and into a snow fest ! As you say lots of uncertainty and I wouldnt be surpised to see this shifted further east or west dependant upon the strenght of the cold block. Thank you for sharing this with us. You do really add value by sharing with us the latest information that you can from the met office. I am sure you would like to share with us your 'gut feel' but appreciate you could be getting into 'deep watre' for doing so. Cheers

[Mark Neal.]

Thanks Ian! Informative Post on the potential of the weekend and how the Met Office see things currently. Insight like that is what is so great about this forum, top stuff!!

Interesting model watching in coming days, we could be on a cusp of something worthwhile, more snow!

Edited February 6, 2013 by Mark Neal.

[london-snow]

Courtesy of a special agreement with UKMO to show select bits of modified output where it helps illustrate and explain their scientific rationale, I can show you just two frames from modified 12z GM, which you can then compare with the raw output. The snowy PPN signal (as x's) needs no further explanation, albeit the very obvious uncertainties at this range need re-stressing.

Thanks Ian.

One thing is for sure though is as you quite rightly say nothing is certain and stamped yet a slight shift east and the capital could be hit severely.

Any thoughts past the weekend atm or keeping tight lips on things atm?

[fergieweather]

At times the GFS can be rubbish. This was clearly seen in the January cold spell when it took ages for the model to agree with the rest. Remember the system that moved S across the UK? The GFS had this moving NE into Scandi!

When you have followed the GFS for many years you get to learn its biases. One of the main flaws with the GFS is it has a eastward bias whenever blocking is indicated to our NE. Another classic bias of the GFS is to overdeepen LP systems.

What members have to realise is all these models use roughly the same data but what is it that makes the likes of the ECMWF more reliable than say the NOGAPS. The answer is simply down to how they are programmed.

Yeah, I should stress I'm not defending GFS frailties, merely emphasising all models can exhibit known deficiencies. I've lost count over past 4 weeks when NAE was ditched in favour of UKMO-GM; equally of course, GFS suffers from lack of detailed forecaster intervention at UK scale, whereas Exeter folks are daily modifying NAE to overcome known deficiencies (e.g. lack of inland shower penetration) or GM wobbles on shower/ppn intensity through grid-row instability, etc, etc. I dare say if they ran GFS from Exeter with equal degree of hands-on modifications it, too, would prove vastly better.

[Captain Shortwave]

I feel I have perhaps reacted a little over the top over the last run. Grumpy me I think

Just that shortwave brings back bad memories. As it's the gfs it could correct in a lot of other ways which don't involve the shortwave which could affect it's position. It's still very likely to be wrong at this stage. Anyway, time to watch something funny to cheer me up, in fact I had some snowfall tonight (good times)

Thanks Ian for the inside info, as usual very informative and very helpful. Tomorrows model watching shall be fun, like watching England in a cup quarter final, lets hope it doesn't go to penalties and gfs's shortwave is in goal.

[Dunstable Snow]

We are a long way from having to worry about what might happen after the sliding low from the weekend. I think of this the same way as when we were looking at whether we could even get to a situation where we might get an extension of cold weather from the current northerly - back at the start of last weekend when the present enticements were more than a week away. The ECM at that time had only just started ploughing its lone furrow towards what we see looming (GFS hissy fit disclaimer attached )

So, now, as people are starting to look another week ahead and speculating about what may/may not happen (after *hopefully* a few snowmen have been made up and down the country ) we might want to look to see if there is another repeated pattern to come to give us some more intrigue and another clue. Not yet and I have some doubts we will repeat the same experience of seeing a model pursuing a lone course in the same way that the ECM has been the frosty persistent outlier that has gathered support like a cold rolling snowball in recent days (*GFS erratic child disclaimer once more*) I think the forthcoming 7 to 10 day period at this time is looking much trickier to plot/speculate than the fun period from 2nd/3rd/4th Feb to 9th/10th/11th Feb has been

Fully understand where you are coming from Tamara but , for the reasons I stated earlier I think we will descend reasonably quickly into a prolonged cold spell that will see us through until the end of February (not withstandingt the odd milder day) whilst we get there. Having lived in Folkestone as a child in the mid/late 70's and early 80's I k now what a cracking place that can be for cold and snow when the synpotics are right but it does need to deliver over the next three /four weeks for your part of the world. I am confident it will

[london-snow]

So are we saying here the further west the snow line for the weekend the stronger incoming block from the east will be or am i off the mark?

Reason i say that is because those who maybe moaning about eastern & south eastern areas missing the action should not be to despondent if that's the case as our time will come eventually going by the 850s & it being not all that slack of a flow.

[Tamara]

Fully understand where you are coming from Tamara but , for the reasons I stated earlier I think we will descend reasonably quickly into a prolonged cold spell that will see us through until the end of February (not withstandingt the odd milder day) whilst we get there. Having lived in Folkestone as a child in the mid/late 70's and early 80's I k now what a cracking place that can be for cold and snow when the synpotics are right but it does need to deliver over the next three /four weeks for your part of the world. I am confident it will

Yes, it can be a great place to be (hence why the NE'erly early next week would be 'rather' nice to consolidate)

Interesting inside METO info about the latter part of the weekend and snow potential. Best and safest to be on the cold side of the low - but like January just gone by, it is possible we are again going to see another fine line with a cone snow area in the middle, with the dry and coldest weather on the eastern flank which erodes the precipitation quickly as the very dry continental air cuts into the precipitation area, then the heaviest snow area closest just ahead of the frontal zone, and then going further west a cut off zone into a marginal rain/sleet area. The whole boundary from driest cold continental air (and little or no snow) to the modified higher dewpoint airmass might be only a 150 to 200 miles wide (or less?).

Edited February 6, 2013 by Tamara Road

[Panayiotis]

Ye I take that any day! ?

[high ground birmingham]

At times the GFS can be rubbish. This was clearly seen in the January cold spell when it took ages for the model to agree with the rest. Remember the system that moved S across the UK? The GFS had this moving NE into Scandi!

When you have followed the GFS for many years you get to learn its biases. One of the main flaws with the GFS is it has a eastward bias whenever blocking is indicated to our NE. Another classic bias of the GFS is to overdeepen LP systems.

What members have to realise is all these models use roughly the same data but what is it that makes the likes of the ECMWF more reliable than say the NOGAPS. The answer is simply down to how they are programmed.

Finally im not sure I agree with the snowfall risk areas you have highlighted. At this stage I feel Ireland/Wales/SW England are more likely to see rain than snow with the risk of snow being further E than you have highlighted. The reason being is I feel the ECM & latest Fax charts are wrong and we shall see a slight shift E in the morning. The GFS on the other hand is far to E with the pattern.

Thanks for sharing the info though. Im going to post my snow risk map in the morning.

Excellent and very informative post there.

What I would add as a defence to the GFS is that it is second to none when handling northerly airstreams and does tend to predict them from rather out. When it comes to a easterly feed however, it is nonsense.