Rjs And Bftp Winter 2010/11 Lrf

[Yorkshiresnows]

Totally agree with some of the above supportive comments.

For the past couple of years your forecasts (Roger and Blast) have been very good.

Of course, there is now a real test with notable divergance from G.P's forecast (who also has a superb pedigree in this area of long range forecasting), and Joe B (now stating that winter is pretty much over for Europe ....... so it will be very interesting to see how the last third of January and February pan out.

If the forecast is even close to being as suggested (very cold), then folks are going to have to look seriously at solar and lunar factors to run alongside the more traditional teleconnections. As a cold lover I'm crossing my fingers

Good Luck

Y.S

[BLAST FROM THE PAST]

Yes its approaching the time where we go off on a different direction from GP who remains steadfast thus far...good to see. And Joe B but he seems to have abandoned ship and changed his call already. His cold to be centred more S and E in Europe has gone bust as he expected a typical Nina response in his outlook with pattern matching.

It at least gets a little juicy now....and I'm away for a few days to Norfolk for wife's nan's funeral so will be offline until Friday....which will be a good day to see what the models have for show for the following week's developments for final third of Jan....a crucial weekend of model watching possibly? :smiliz19:

BFTP

Edited January 10, 2011 by BLAST FROM THE PAST

[Mr Frost]

I see nothing more than wet, windy, average winter weather for the rest of this month. I do hope you are right though.

[BLAST FROM THE PAST]

I see nothing more than wet, windy, average winter weather for the rest of this month. I do hope you are right though.

That's if you can trust the models wholeheartedly....I don't :smiliz19:

BFTP

[Roger J Smith]

Just a few notes here in response to some earlier posts.

Yes, the mild spell has arrived a few days earlier than implied in our forecast, and the "near-record warmth" as mentioned is likely to hit a few days before the two peaks in energy level mentioned (17th, 20th). This may not look like a very large error at the end of the winter, if we then see a downward slide in temperatures later in January and a cold February. However, there was no expectation (from me at least) of 100% accuracy so if we can stay on the right side of normal temperature more than 70% of the time it will actually represent an improvement in the methods being used. I should leave comparisons to others but I really don't think any other research program in LRF is ahead of that standard today.

Kevin raised the question about absolute accuracy that he thinks an astronomy-based forecast should have. I agree that the potential exists for that. The map that I published on Net-weather for Dec 31, 2005 demonstrated the potential to get an accurate map at very long range (over 30 days in that case). However, what Kevin was suggesting is really premature for a number of reasons, and therefore at this point in time does not constitute a valid criticism of the method ("it isn't perfect, therefore it must be flawed" seems to be what he's saying).

Let's review the reasons why a forecast based on this approach cannot be 100% accurate at present. First of all, a large element of the temperature trend forecasting is index-based from past analogues. The theory is quite easy to visualize -- we attempt to model the solar system magnetic field in sectors, apply that system to the global atmospheric pattern, and use conventional meteorology to deduce a forecast. One can see right away how that is not going to produce a deadly-accurate day by day forecast, but let me run a second point by you first -- the day to day pressure variations are postulated to be lunar interference in gravitational wave patterns. There again, analogues are used to derive approximate forecasts.

Okay, with those two concepts in mind, the problems are numerous for getting complete accuracy. One, the various components of the model never repeat in combination, so we are always assessing different signals derived from the data independently, and trying to visualize how these will combine or interact. That creates uncertainty. Also, the derived index values could be as accurate as you like, but the grid into which the signals are being received is slowly changing over time (the shifting magnetic field of the earth). So we are always using the signals from past cases that were in a different grid to the present day. That factor is assessed but obviously adds uncertainty.

The lunar event analogues can be precisely timed but the analogues are subject to the same set of problems as per the above, a shifting grid -- a perfect analogue storm from say 1953 might now track several degrees further north in the shifting grid over that period of time. I think these examples give some indication of how approximate the system becomes at long time intervals. When I'm applying the research to short-range forecasts I can get some very precise results because the initial condition is so precisely defined. When trying to model events 60-120 days ahead, there is bound to be some uncertatinty even if correlations are high in the index values.

A final point adding uncertainty is simply the nature of magnetic field sectors, these could be expected to ebb and flow within defined boundaries, it's not quite as straight-forward as astronomical orbit prediction might have one believe. We are trying to predict the three-dimensional position in space of a field sector from past analogues and initial time locations. The field sector could then pulse, oscillate, rather like a bowl of jelly rather than moving through very precisely defined locations in orbit like a planet in orbit around the Sun.

I honestly don't know if more precision than this current forecast is possible. The broad general concept, very cold, then very mild briefly, then cold again, would obviously be more accurate if you got the starting dates and end dates of the mild interval exactly right, but using this method, I am not optimistic that we could ever be more precise than plus or minus three days on such sector changes. For the lunar events, the error is more in terms of track than timing, and I think further accuracy can be gained by better understanding how to predict track (the timing is already quite accurate).

I hope the reader can see that it wouldn't matter how complex a computer one had to throw at this puzzle, if the cause and effect has a certain amount of non-linear non-rigid "give and take" due to the porous nature of the field sectors, then we might just have to live with a range of uncertainty and take the wait and see approach to exact timing of sector changes. What's more important for planning is the general menu of events. If authorities know in advance that there will be a lot of cold and snowy weather, then they can plan in general terms and take the actual events on a case by case basis. Anyone expecting this to be a mild winter would have been caught short on winter supplies. And also, let's say somebody read our forecast in November and saw the mention of a milder spell in mid to late January, and said "okay, sounds like I may not get much outdoor work done in December or early January, but I might be able to make up ground later in January," then they are going to be well prepared for this set of changes and a few days difference in timing won't matter. But if they tried to cut it too fine and went off to the south of Spain planning to return on 15 January to resume work, then they will have missed perhaps a week, so that sort of indicates the concept of taking the forecast as a broad general guide rather than a calendar for future events. We should perhaps not bother with the daily details except that I find it helps my research to consider them and study the errors involved.

[Catacol]

Roger

I doubt too many people would demand precise daily accuracy for a lrf to be worthwhile. So far your forecast has the patterns spot on, although the cold in the first third of Janaury was not as bad as expected with the trough obstructed somewhat. I am impressed purely by the general pattern accuracy, and would agree that for industry/business this would be an excellent asset for useage - and if the cold or warm or whatever arrives a day or two early/late then so bet it, but the knowledge that the overall pattern was accurate would be sufficient.

It will be interesting to see if you can score a near 100% on this in the second half of winter - lots of suggestions around that high lat blocking and/or easterlies will not be on the cards - getting that pattern correct is the key.

I am new on Netweather this year: I note you spoke of posts as far back as 2005. Over the last several years how would you rate your accuracy in getting the overall patterns right? Do you keep data like that?

[Roger J Smith]

For my part, I joined NW in 2005 as part of my efforts to extend my research from North America (where I've been at this most of my adult life, since 1980) to western Europe. I would describe the long-range forecasting that I've done as improving with experience, and partly as the consequence of building up a CET data base that I can study in the same way that I approached my North American research. Have been collaborating with BFTP for about three years now as we found that we were looking at similar theoretical paradigms, and forecasts have improved since that began.

We have not done other seasons very intensively, sometimes I have posted a brief outlook. In general, I believe that the forecasts are staying on the right side of normal when you break down the seasons by weeks or thirds of months, at least 65% of the time and possibly more like 70%. This is encouraging. We both think it could improve further.

It's a rather fragile process, given that neither of us are full-time employed in the field, although as I'm now retired from outside work, I suppose I am full-time committed to it. At my age, there's no telling how many more years I may have to advance this research. I find that it is straining both my mind and my home computer network to the limits, just to continue making progress rather than maintaining a status quo. There are an awful lot of computer files to maintain, study and frankly, remember for future reference. I have identified over 100 factors in my theoretical model (other researchers doing this kind of forecasting have only the Moon in their modelling which must be a luxury). The sheer complexity of the global atmosphere and my limited reources have conspired to limit my research to specific regions. I have hardly scratched the surface of the eastern hemisphere, or the southern hemisphere, although the theory can make predictions for these regions -- I don't have data bases nor do I have time to study the daily weather there in much detail. My day is pretty much overloaded now just keeping up with two separate internet "interest areas" on about a dozen forums (I am equally interested in politics, as it happens, but that's the other side of my Jekyll and Jekyll personality).

Was Jekyll the good or the bad one? Anyway. We have a ton of melting snow that also requires my immediate attention before somebody sues my sorry ass. Better get dressed and deal with it.

[Weather-history]

Kevin raised the question about absolute accuracy that he thinks an astronomy-based forecast should have. I agree that the potential exists for that. The map that I published on Net-weather for Dec 31, 2005 demonstrated the potential to get an accurate map at very long range (over 30 days in that case). However, what Kevin was suggesting is really premature for a number of reasons, and therefore at this point in time does not constitute a valid criticism of the method ("it isn't perfect, therefore it must be flawed" seems to be what he's saying).

What I'm saying Roger is that there are far too many random elements for me. You got El Nino, La Nina, SSTs, solar activity, volcanic influences etc and to me they have a far bigger influence than some small planet ie Mercury, retrograding, which I'm afraid I am totally unconvinced with. For instance, how would your forecasting technique cope if we had a Tambora type eruption ie an Earth based event? For me, the biggest players that affects Earth's weather and climate are the Sun and the Earth itself. All others are very minor players unless they hit the Earth ie cometary or asteroid bodies.

Edited January 13, 2011 by Mr_Data

[Roger J Smith]

The points above regarding volcanic dust veils are entirely irrelevant, this would be the same problem for any method of long-range forecasting, but also would not just come out of the blue, there being a lag time between major volcanic eruptions and impacts on climate at the distance of the U.K., so I don't see how that's a valid objection to our method in particular.

I'm sure this will be obvious to most, but of course we have factored in the role of the Sun in our forecasts, otherwise, we would be predicting monthly averages of -273 degrees, or if we got the main point but missed the axial tilt, each month the same. Some of the other objections have been discussed at length before, the bottom line being, these teleconnection aspects of climate are part of the forecasting challenge as well as being signals, and there's no reason to suppose that our methodology does not incorporate them indirectly as being predicted elements themselves. Anyway, I can tell you that most traditional La Nina based forecasts for this winter have proven very shaky indeed for North America, so there's a lot more going on than just sea surface temperature anomalies. I am in the school of thought that says that SST anomalies are driven by pattern and not vice versa.

Otherwise, most of the stuff in the post is a rehash of earlier objections that I've dealt with numerous times. All I am willing to repeat at this point is that interactions are not postulated to be gravitational in certain cases, therefore the lecture about gravitation is also irrelevant, since we don't claim the processes are gravitational. The processes are occurring in magnetic fields and processes within these are scaled much differently. Some of the processes I am studying fit best to the metric of

M^(1/6) / R^ (1/6)

which is to say, the sixth root of mass divided by the sixth root of distance. This metric tends to reduce the differential aspects of gravitation by a huge amount.

Anyway, this thread is about our forecast, there are other places to discuss the theory behind it. Basically, I need to renew my efforts to get published so that I can end this tedious situation.

Edited January 14, 2011 by Roger J Smith

[lorenzo]

Roger i wouldnt if i were in your learned position - refer to things as tedious.

For many of us, myself included albeit claiming some weak esoteric grasp of anomalies here and there - the posts you create are dynamic, they are forthright they personify.

I se contradicting views on here as counter equations it is almost like the matrix vs merrovingian or whatever his name is.. ( sic)

It is life..

I knoww I know..

[BLAST FROM THE PAST]

Re the sun being a major player, for me now it is an integral part of my imput into the forecast. There are many many cycles but one of the recent obvious one that played a big part in this forecast from my side was the perturbation cycle [we are in a period of la Nina domination and a -ve PDO and is a 36 yr solar cycle]. This changed in Feb 2007 and some of its impacts are a well correlated prevalent -ve NAO pattern [something we for sure have been witnessing], a pushing equatorward of the jetstream thus enhancing northern blocking [something which we have been witnessing]. Now with this effect then one would then anticipate a moving of position of any peak energy signal? The fact that we are seeing the quietest sun for a century also pushed me towards readily going for a consecutive cold winter without hesitation.

From this I could bring up an overall pattern and as I have said for specifics Roger's research is more advanced and able to forecast that much better than I.

Incidentally the ECM 12z is looking quite juicy for a reload of very wintry weather. I know I am a little more 'excitable' than Roger :smiliz19: but I don't expect 100% hit rate but I am now of the thought that we can very much get the pattern of the winter pretty close.

BFTP

[The PIT]

Ah well Jan forecast is bust not quiet as bad as my attempt but bust nether-less.

Agree with Blast though about the sun being a major player and I think the next few years will confirm this.

Edited January 15, 2011 by The PIT

[❄️❄️❄️❄️❄️❄️❄️❄️❄️]

Is it still on track to be a cold Feb?

[Roger J Smith]

I don't consider the forecast a bust, cold to start, near-record warm mid-month, looks like turning colder at some point especially near the end to set up a colder February. Where's the "bust" in that? January CET 3.0 still looks about right if the last few model runs are right.

Edited January 16, 2011 by Roger J Smith

[ZONE 51]

I Have to say both of you your forecast is amazing, clever and incredible in the way its put together, and as i can see RJS your re-edited post (i did read that earlier) but just after reading that i went to your forecast page and re-read through again and its passed so far! i don't see any thing un-forecast so far, whats written has happened! and whats building is! so great stuff, relax, and enjoy your(both of yours) next predicted coldspell!

[geoffw]

it is very obvious february will be nowhere as cold as jan. i very much expect fred and rogers forecast for feb to go KA-BOOM and BUST big time!

[sundog]

it is very obvious february will be nowhere as cold as jan. i very much expect fred and rogers forecast for feb to go KA-BOOM and BUST big time!

Well i dont think its obvious at all what Feb holds for us i would not be surprised if it is colder then jan. I mean how can it be obvious what feb will be like when the start of it is still 2 weeks away???

[BLAST FROM THE PAST]

Ah well Jan forecast is bust not quiet as bad as my attempt but bust nether-less.

Agree with Blast though about the sun being a major player and I think the next few years will confirm this.

The only thing so far is that it didn't start as cold as anticipated in my own view, it is only the 16th and calling it bust is a tad premature at present. Lets see whn we get to 1st Feb?

Geoff you are entitled to your opinion but what evidence or methodology are you using to call us wrong re Feb this early? What is your reasoning behind such a post?

Ta

BFTP

[Tim Bland]

Fred, I recall you saying you where going to do a more in-depth forecast for Feb at some point?

[geoffw]

come on ... give us the details fred and then we can judge your accuracy!

[Optimus Prime]

come on ... give us the details fred and then we can judge your accuracy!

The way you just posted I'm surprised he'll even bother!

Does look as if colder weather will return. I look forward to seeing how it stands come March 1st. Probably nowhere near as bad as some are making it out to be!

[The PIT]

I don't consider the forecast a bust, cold to start, near-record warm mid-month, looks like turning colder at some point especially near the end to set up a colder February. Where's the "bust" in that? January CET 3.0 still looks about right if the last few model runs are right.

Now that is good spin I'll sign you up for Tony Blair's press officer.

As for Feb not mentioned it it could end up right.

As for my cock up here it is

January will start with another northerly outbreak which will be fairly short lived. One or two Snow showers around but nothing major. After a brief quiet cold spell another Northerly outbreak with possible frontal snow working south. The second week shows once again the familiar southerly jet will be the main theme so cold is the likely outcome. Overall I think a cold month ahead with some snow around. So cold with dry frosty spells.

There's misses and there's misses.

I tend to monthly forecasts here http://www.sheffieldweather.co.uk/Last%20Months%20Report.htm I use CFS raw and GFS and ECM for the first two weeks. GFS and CFS were still showing a fairly blocked scheme while ECM was hinting at a longer milder spell. Still I got November and December pretty well spot on.

[Weather-history]

it is very obvious february will be nowhere as cold as jan. i very much expect fred and rogers forecast for feb to go KA-BOOM and BUST big time!

Not even sure how January will end up so I can't see how you can say it is very obvious that February will be nowhere near as cold as January when we are not even sure of the January benchmark.

I don't consider the forecast a bust, cold to start, near-record warm mid-month, looks like turning colder at some point especially near the end to set up a colder February. Where's the "bust" in that? January CET 3.0 still looks about right if the last few model runs are right.

Temperatures should be only part of the validation process, Roger. Synoptics validation is a big player as well and if you end up with a temperature prediction similiar to the final outcome but under different synoptics than predicted then the forecast IMO is not a good one. For instance I could predict a CET of 1.6 with frequent northerly outbreaks with Scandi trough and Greenland high and it turns out 1.6C but because an anticyclone sat over the UK throughout the month ala December 1933.

[The PIT]

Not even sure how January will end up so I can't see how you can say it is very obvious that February will be nowhere near as cold as January when we are not even sure of the January benchmark.

Well at the moment it's not looking to great for any real cold although ECM is having the odd half hearted attempt on the 12 oz runs deep in FI. Temps should drop off this week for a while but the question then how cloudy will this high be. Clear nights the best scenario while the worst is cloudy day and night so suppressed day time temps but above average night time temps.

[Guest mycroft]

The only thing so far is that it didn't start as cold as anticipated in my own view, it is only the 16th and calling it bust is a tad premature at present. Lets see whn we get to 1st Feb?

Geoff you are entitled to your opinion but what evidence or methodology are you using to call us wrong re Feb this early? What is your reasoning behind such a post?

Ta

BFTP

Must be the Time And Relative Dimension In Space method

T A R D I S

Edited January 16, 2011 by mycroft