Get ready for the great MW/GL storm of Dec 19-21

[DaculaWeather]

I think Roger is doing at least as good as the models which right now are crap. Let's see how he does. I don't think he can't do any worse than the models.

[Guest someguy]

I think Roger is doing at least as good as the models which right now are crap. Let's see how he does. I don't think he can't do any worse than the models.

I think that is the dumbest most moronic thing anyone has ever since in the histrry of the entire Human species

at least the Models have been showing an east trend ...

Mr Smith went for Chciago

[DaculaWeather]

All I know is I haven't seen the models do a very good job this winter, and most of the predictions I've heard at the start of the season haven't come true. And while he might be wrong on this one, if he is, he'll fall into the same sinking ship as most everyone else.

[Roger Smith]

At the moment, would just say, some positives and some negatives in this prediction as originally made and discussed since then.

The positives include, system exists, timing not far off, current GFS track runs at angle intersecting postulated track near timing line, snow for Chicago region now in forecast for 20-21 as suggested in discussion earlier than models showed this. Also positive, offshore lows are now trending further east, no coastal snowstorm, no total suppression of timing line one although the stronger offshore system continues to take potential development space away. Comments about upstream generally valid. Setup was broadly similar to LRF across arctic.

The negatives include no Christmas card from someguy, no really major development on this storm (yet), intensity more like a 4 than an 8 on 10-pt scale, if even that. Track error speaks to field error around and east of timing line one mostly. Back at LRF issuance (which was after all mid-Sept) had pictured flow being more zonal but under a high-latitude block.

But with 72h to go now to event time, some scope for either improvement or decline of actual result. There is raw material to work with, a low coming out of the central plains towards the lower Great Lakes and OV. Were this to prove more intense than 00z GFS run, a better result possible.

The Dec 25-26 event crossing the timing line appears to be following a similar path and gets major offshore development. Given the current analysis this is what I would now expect from this system, so possibly the first major snowfall event for some places in the northeast on the 26th-27th.

Would say as another early call, keep an eye on Jan 3-4 as timing for possible east coast major storm.

[dilly84]

Well hope it isn't a trend, but the 0z ECMWF came in a lot weaker than the 0z GFS and the 0Z Nam. The 0z Nam was well north of the GFS, and now the 6z runs are coming in weak like the ECMWF and more north. Hope the 12z runs give a better idea of what is going to happen. I don't know how your model works or whatnot, but the standard models suck. This is the worst I can ever remember them being.

[Feb]

Well hope it isn't a trend, but the 0z ECMWF came in a lot weaker than the 0z GFS and the 0Z Nam. The 0z Nam was well north of the GFS, and now the 6z runs are coming in weak like the ECMWF and more north. Hope the 12z runs give a better idea of what is going to happen. I don't know how your model works or whatnot, but the standard models suck. This is the worst I can ever remember them being.

http://i127.photobucket.com/albums/p151/DarkSoul0703/thmagic8ball.jpg

I believe this is Roger's model

[Ellinwood]

The positives include, system exists, timing not far off, current GFS track runs at angle intersecting postulated track near timing line, snow for Chicago region now in forecast for 20-21 as suggested in discussion earlier than models showed this. Also positive, offshore lows are now trending further east, no coastal snowstorm, no total suppression of timing line one although the stronger offshore system continues to take potential development space away. Comments about upstream generally valid. Setup was broadly similar to LRF across arctic.

This is rather shoddy verification...

To put it into perspective, in the D.C. snowfall contest (first 1+ inch snowfall), 4 people guessed the correct date, and one even hit the accumulation spot on! We did not use a theoretical model. The guesses were made in mid-late November.

How is getting a generally climatological event to occur a good verification?

Could we see any of the actual verification that you're doing? Are there any visuals you can show us?

Well hope it isn't a trend, but the 0z ECMWF came in a lot weaker than the 0z GFS and the 0Z Nam. The 0z Nam was well north of the GFS, and now the 6z runs are coming in weak like the ECMWF and more north. Hope the 12z runs give a better idea of what is going to happen. I don't know how your model works or whatnot, but the standard models suck. This is the worst I can ever remember them being.

And just how long have you been looking at the models? Perhaps I should point you to this: http://www.americanw...post__p__125362

[Hailstoned]

Hey, there was a major storm threat that had models and weenies excited, even if the only thing that came of it was mutual masturbation. So give the guy some credit, though obviously his "model" needs a lot of tweaking, as do those spat out of supercomputers several times a day.

[Ellinwood]

I love all of the "take it easy on him" posts. If I could see some objective analysis and maybe some of the actual output that he's transcribing to us, it would certainly help.

[dilly84]

I love all of the "take it easy on him" posts. If I could see some objective analysis and maybe some of the actual output that he's transcribing to us, it would certainly help.

Actually you're trolling his page. You obviously enjoy it or you wouldn't come to it. And it isn't one storm. He was doing it last year as well.

[Ellinwood]

Actually you're trolling his page. You obviously enjoy it or you wouldn't come to it. And it isn't one storm. He was doing it last year as well.

I'm just making sure he takes his theory through the proper testing/verification procedures in an objective manner. If he has images that he's looking at, why can't WE look at them? It's not going to hurt his research to post some output, especially if he keeps the code all to himself. I would like to see what he's seeing. Who knows, maybe I have some insight for him? Maybe others do?

Yeah the trolling is unfair at this point, but writing out an interpretation of his model is so much longer and convoluted compared to just posting an image. I would like to be able to analyze it myself and see if the conclusions he's drawing are sound.

[dilly84]

I'm just making sure he takes his theory through the proper testing/verification procedures in an objective manner. If he has images that he's looking at, why can't WE look at them? It's not going to hurt his research to post some output, especially if he keeps the code all to himself. I would like to see what he's seeing. Who knows, maybe I have some insight for him? Maybe others do?

Yeah the trolling is unfair at this point, but writing out an interpretation of his model is so much longer and convoluted compared to just posting an image. I would like to be able to analyze it myself and see if the conclusions he's drawing are sound.

Okay, I agree with you where maybe posting an image is concerned. It would help understand what it is he is doing, but based upon reading his posts its obvious he isn't just making something up considering the explanations he gives and, quite frankly, half of it I don't understand. He talks about these timing lines etc. No clue, but the way the models are trending, he's going to be wrong here and then you can gloat over it lol.

[Ellinwood]

Okay, I agree with you where maybe posting an image is concerned. It would help understand what it is he is doing, but based upon reading his posts its obvious he isn't just making something up considering the explanations he gives and, quite frankly, half of it I don't understand. He talks about these timing lines etc. No clue, but the way the models are trending, he's going to be wrong here and then you can gloat over it lol.

At this point I'm not concerned as to how right or wrong his model is. I'm much more concerned about him objectively verifying his forecasts to ensure that he can tweak the model properly after understanding what went right/wrong meteorologically. I'm here to try to keep him from making false verifications. There's nothing to gloat about, especially considering the poor handling of the system by the other models as well.

[Roger Smith]

I'll post some signals derived from the Toronto data when I get a chance. Kind of busy today.

Last winter I posted a thread on the lunar dates associated with the K-U snowstorms. That thread is lost in cyber-space but I saved the table from my posting there. You can see the correlation with lunar dates in these K-U storms and the conclusion drawn in that thread was that new moon events in early February were favoured, however, there are full and new moon signals through the winter season from that data.

The timing line aspect of this theory that people were talking about, is not hard to visualize. Take the NMP rather than the north pole as the centre from which nine timing lines extend poleward. The NMP can be placed at 80N 100W for this purpose (it is actually now at 84N 110W so the grid needs an update). For timing line one, drop this west of Hudson Bay and through northwest Ontario, DLH, ORD, CHS then southeast to a point near 20N 55W. I could go through how this links to a southern hemisphere grid some other time. Now from that timing line, you can find the other eight, by drawing in lines at intervals of 40 deg east (or west) until you have the hemsphere covered. The timing lines used in practice are slightly redrawn to account for topographic forcing effects but this would get you started. There is a concept of meteo-latitude implied in this model since the north pole of the model is ten degrees south of the north rotational pole. In the research model, lines of meteolatitude are drawn in between the poles of the system and the known position of the ITCZ which forms the equator.

Then for research and verification purposes, systems can be tracked as to meteo-latitude and timing "number" which defines where in the timing line system the weather event is located. Since there are nine timing lines, the timing number ranges from 1.00 to 9.99 running eastward -- the value 1.00 would be midway from timing line 1 west to timing line 9. The value 2.00 would be halfway east from timing line 1 to 2. So a point on timing line one has timing number 1.50 in the system. Anyway, probably what's best would be to start a different thread for discussion, so we could examine various systems as they happen, from this research perspective. I haven't looked around the forum enough to know if we have started a research forum like Eastern had. But that's probably the place for this discussion, rather than cluttering up the active weather forum.

[Guest someguy]

Actually you're trolling his page. You obviously enjoy it or you wouldn't come to it. And it isn't one storm. He was doing it last year as well.

and he sucked at it last year as well .

This is a board of science

if you come on here preaching astro meteorology

and are asked for some verifciation that is NOT in anyway TROLLING

in fact it is treating him == roger smoth -- with RESPECT.

[Roger Smith]

Okay, with the event just 12h away, the predicted low is weaker and holding back to the west, actually the 72h GFS had a better solution from verification point of view than 12h real-time models are showing. The position of the low has been drifting back to the west on successive runs and at event time (08z) looks to be around western IL or eastern IA. Clearly the system is weaker than the research model said and is moving WNW-ESE rather than cutting northeast. This would introduce severe errors away from the timing line but for MN, WI and n IL the model forecast is not too bad actually, depending on how much snow actually falls.

I've looked around the hemisphere, checked off the nine predicted lows and found several of them similarly weaker. None are stronger than forecast except the one east of New England which was supposed to be more suppressed. These errors are consistent with the additional strength of the cut-off retrograde PV which was anticipated but not modelled strong enough. The model performance over Europe past 30 days has been good to excellent as can be verified by reading the long-range forecast for UK on Net-weather (plus my call in their forecasting contest for CET 0.8 C which is third lowest of about 100 contestants and at present just above the standing mean which looks to drop to about -0.5 then rise slowly end of month).

Possibly, getting back to the generally weaker storms with this full moon / northern max, the eclipse alignment is a negative, perhaps it demonstrates that the transfer of energy to the geomagnetic field is being shielded, an effect that would not obtain at new moon in eclipse mode. I'm number crunching the part of the data set related to eclipsed Dec-Jan full moons to see if the signal is weaker.

Nevertheless, perhaps the most significant addition I can make is this satellite imagery which clearly shows the rapid formation of a near-perfect clear circle in the cloud structure over the western plains, about 300 miles across, covering much of western Nebraska and adjoining nw KS, ne CO and sw SD.

http://www.weatherof...bimages=1&clf=1

I have linked to an animated satellite image here. You could pause it and check out the evolution. You have to select the animation. If you slow it down, the best view of the clear circle is about 2100-2130z ... streaks of AS/CS start to obscure it at 2200z.

From this satellite perspective, Moon is currently off to the right as time is near sunset in the region and Moon would just be rising in the northeast. Moon's diameter is about 3500 miles (compares to earth's 8,000 miles).

Now, as you know, Moon has same apparent size in sky as the Sun. Sun is about 400 times as far away therefore diameter 400x that of the Moon. If you imagine a cone that contains the Sun, the Moon and a point on the earth's surface, on the average ending at the centre of the earth, but varying by orbital parameters, the size of this cone on the earth's surface will vary from about 100-300 miles. Moon is fairly close to perigee meaning the cone is wider than usual. Possibly this odd cloud feature is the Moon's shadow in the interference pattern, approaching timing line one. Comments on what actually causes this clear circle feature very much appreciated. All the other stuff, whatever ... I feel like I am on to something anyway.

[Ellinwood]

Map? Visual comparison? Correlation co-efficient? I'm only asking because you haven't produced a map or stated that you refuse to give me a map despite my 3 previous requests.

[Roger Smith]

I'll post a map comparing the original forecast track and actual positions when we've passed the end of the 21st. I'm not claiming anything spectacular on this event, but it met one minimum condition -- it exists. That's a start.

By the way, speaking shoddy verification, how about my snow falling through that missing roof?

There's different kinds of shoddy, I guess.

Our roof (at our domed stadium in Vancouver) ripped open in a windstorm a few winters back. I think the design needs work -- same design. This is where they had the opening and closing for the Winter Olympics (speaking of shoddy design), at least they got a better roof on the place.

You can certainly visualize the map, just draw one line from Wyoming to Lake Michigan then ESE, compare that to another line from Kansas to Lake Michigan to Ontario. They intersect over Lake Michigan. Otherwise, bleah.

The system is also late getting to timing positions, meaning, timing number has dropped from about 50 to about 40 in the past three days. The models were fighting this error too, without recognizing it as such. Timing number should continue to drop. This means good things for the east coast, because if the next timing line shifts back west, then it helps intensify storms for the east coast. Last winter, timing number from my analysis shifted rapidly west in late January. Good things started to happen.

[PhineasC]

Congrats on predicting that weather would occur between the 19th and 21st. I was skeptical that weather would occur in the world on those dates, but I am a believer now.

[Roger Smith]

High praise indeed, I am forever in your debt.

[buckeye]

High praise indeed, I am forever in your debt.

not sure what you expect roger. You make a thread entitiled "get ready for the great MW/GL storm of dec 19-21"....and other than some waa snows breaking out as a weak clipper moves well north, you missed it horribly. Claiming a timeline was off or some such excuse, doesn't cut it. When you tout some new scientific method, you need to eventually nail something...even if by luck. Otherwise you rightfully open yourself to criticism.

[Ellinwood]

I'll post a map comparing the original forecast track and actual positions when we've passed the end of the 21st. I'm not claiming anything spectacular on this event, but it met one minimum condition -- it exists. That's a start.

By the way, speaking shoddy verification, how about my snow falling through that missing roof?

There's different kinds of shoddy, I guess.

Our roof (at our domed stadium in Vancouver) ripped open in a windstorm a few winters back. I think the design needs work -- same design. This is where they had the opening and closing for the Winter Olympics (speaking of shoddy design), at least they got a better roof on the place.

You can certainly visualize the map, just draw one line from Wyoming to Lake Michigan then ESE, compare that to another line from Kansas to Lake Michigan to Ontario. They intersect over Lake Michigan. Otherwise, bleah.

The system is also late getting to timing positions, meaning, timing number has dropped from about 50 to about 40 in the past three days. The models were fighting this error too, without recognizing it as such. Timing number should continue to drop. This means good things for the east coast, because if the next timing line shifts back west, then it helps intensify storms for the east coast. Last winter, timing number from my analysis shifted rapidly west in late January. Good things started to happen.

You say the other models had trouble with the timing, not recognizing this "timing number error" you mention. Why didn't your model pick it up? What did you see that indicated this slow-down? Can you show us evidence of this changing timing number other than saying that the models trended later? (that last question is putting words in your mouth, but I assume that that's what you saw that indicated the change in timing)

Congrats on predicting that weather would occur between the 19th and 21st. I was skeptical that weather would occur in the world on those dates, but I am a believer now.

Pretty much this. Like I said before, use the DC snowfall prediction contest as an example. Four of us guessed the correct date of the first 1"+ of snow in DC 20-30 days in advance, but that's just luck. To claim victory on a completely different kind of weather system that's more than 2 days off of your timing is not a good verification at all.

It would be nice if we could get more forecasts so we could do further verification of the model.

[Roger Smith]

Are you saying the low way off the east coast is the low I was predicting? I'm not sure how that's possible to prove or disprove, it crossed the timing line around the 17th and is now closer to timing line two than one. I never thought it was "the" low -- I've been tracking this northern plains low as the verification low. It is not three days late, more like 12-18 hours. And that's because timing lines have shifted west apparently (I'm afraid this is circular reasoning but as the model works from the assumption that events are on timing lines at event time, I have no alternative to it).

The problem is that the low is too weak more than anything else. Weak, but people did mention things like lightning with it, and some locally heavy precip. It is acting like a more energetic system than a clipper even if it looks like a clipper on the pressure charts.

Anyway, could be a while, but I would propose just abandoning this thread as it cannot accomplish anything, and I will post more general research findings in a better place to keep this out of the forecasting forum.

[Ellinwood]

Are you saying the low way off the east coast is the low I was predicting? I'm not sure how that's possible to prove or disprove, it crossed the timing line around the 17th and is now closer to timing line two than one. I never thought it was "the" low -- I've been tracking this northern plains low as the verification low. It is not three days late, more like 12-18 hours. And that's because timing lines have shifted west apparently (I'm afraid this is circular reasoning but as the model works from the assumption that events are on timing lines at event time, I have no alternative to it).

The problem is that the low is too weak more than anything else. Weak, but people did mention things like lightning with it, and some locally heavy precip. It is acting like a more energetic system than a clipper even if it looks like a clipper on the pressure charts.

Anyway, could be a while, but I would propose just abandoning this thread as it cannot accomplish anything, and I will post more general research findings in a better place to keep this out of the forecasting forum.

I'm not talking about that system. Feel free to start a new thread in the research forum, and we can continue forecasting and verification

Long-range signs are good for MW/GL storm Dec 19-21

...

This storm could be expected to form around OK-KS late 18th and start moving northeast on the 19th, reaching Wisconsin on the 20th and central to northeast ON by the 21st. A track somewhat further south is still in play, and could run AR-IN-wNY. Whatever track this storm takes, there would likely be a heavy snowfall of 12-18 inches to the north, bands of freezing rain and rain along and just to the south of the track, and a brief but major warming spike in the warm sector ahead of the low.

...

You were, at the earliest, 36 hours too early.

Here's the 7pm map on the 18th:

We finally start to see the system coming off the mountains at 10am on the 20th:

And it enters the OK/KS area at 1pm on the 20th:

And now it's moving east/southeast, not northeast.

BTW, it takes less than a minute to upload a map. You have them, right? Just post one...

[NEOH]

High praise indeed, I am forever in your debt.

I tried following your posts above regarding the storm... but I was lost... and starting thinking you may be the Riley Martin of the weather world. You had a storm tracking from LA to NY.... there is a weak clipper with WAA snows. Not even close.

Maybe some things are better left unsaid... but calling for the great MW/Great Lakes storm then say you were off a little is ridiculous.