The Seasonal Forecast Thread

[saintkiptanui]

I think Rogers forecast is looking good

[Bobd29]

RJS LRF (don't u just luv abbrevations!!!)

Roger's methodology is hardly mainstream and it seems human nature to be reluctant to accept anything 'unconventional' at face value. What cannot be denied however is the time and research invested by Roger over many years to back up any forecast with incredibly detailed data, whether conventional or not the approach is certainly scientific. To date the LRF of 3 months ago is almost spot on with a validation that beats anyone else brave enough to have made a prediction.

I for one always read a RJS comment with real interest, I know it will be factually well researched. Roger has backed his February forecast with his CET prediction of -1.1 against nearly all others. So hats off to you Roger, your dedication to your science is outstandng, please continue to enlighten/entertain us.

Bobd29...

[Terminal Moraine]

RJS LRF (don't u just luv abbrevations!!!)

Roger's methodology is hardly mainstream and it seems human nature to be reluctant to accept anything 'unconventional' at face value. What cannot be denied however is the time and research invested by Roger over many years to back up any forecast with incredibly detailed data, whether conventional or not the approach is certainly scientific. To date the LRF of 3 months ago is almost spot on with a validation that beats anyone else brave enough to have made a prediction.

I for one always read a RJS comment with real interest, I know it will be factually well researched. Roger has backed his February forecast with his CET prediction of -1.1 against nearly all others. So hats off to you Roger, your dedication to your science is outstandng, please continue to enlighten/entertain us.

Bobd29...

Agree with that; always keep an open mind is my motto.

If a theory seems a bit 'out there' it doesn't automatically mean it's wrong, rubbish or even flawed.

If it's proven wrong repeatedly then perhaps it's time to dismiss it but RJS has as good a record with his long range forecasts as anyone else I've read ( better than a good many ) so I read his thoughts and forecasts with interest.

Let's not forget that even luminaries like Copernicus were thought to be 'out there' at one time.

[Roger J Smith]

If this isn't non-random, it is pretty damn close.

Edited February 3, 2013 by Roger J Smith

[Rollo]

Roger Smith is very quiet considering his LRF is looking pretty good now,well done Roger! I trust all is well back in Vancouver.

[Methuselah]

Aye Fred: the use of analogues certainly has a bright future?

[BLAST FROM THE PAST]

Aye Fred: the use of analogues certainly has a bright future?

Analogues though Pete don't give you timings, that's the extra bit

BFTP

[Roger J Smith]

The whole foundation of this cold February idea was that the research model showed a pleasing combination of weak progressive warm signals and strong retrograde index. If that combination can't bring cold weather, nothing can. The retrograde index is what we're currently discussing in the thread that I promised (in the science forum) and at the rate I am finishing charts and posting new material, it will overlap most of this month to complete the discussion.

I've come to the belief that our atmosphere is really just the interface between the planet's surface and the surrounding magnetic field environment as modulated by lunar interference patterns. People always say that the Sun should dominate and obviously it does just in the fact that it provides all of the heat radiation for the system, but my theory is about variation in the Sun's output and how we can predict and model that for long-range forecasting. This variation is small in absolute terms but apparently small variations can lead to the scale of atmospheric variation that we observe. Now it is obvious to me and others that long-term changes in solar output connected to weaker solar activity at certain intervals (like the Maunder, Dalton and the less prominent case from about 1875 to 1915) will play a significant role in lowering the general base-line for any other concept of temperature variation. I don't think we have entered into such a strong solar-weakening period yet, that the model needs to be recalibrated down for that baseline temperature. I am not observing any continual bias towards colder actuals than model predictions. Perhaps this is because the AGW signal is just about cancelling out the weak-solar signal since 2008 or so.

One large theoretical problem for my work to overcome is that it is necessarily at the mercy of quasi-random wanderings of the grid that is imposed by earth's magnetic field. Although there has been considerable change in position of the NMP since the 17th century (it has performed about half a counter-clockwise rotation now taking it from Victoria Island then through the northern Canadian mainland in the early 19th century, through the heart of the Canadian arctic islands in the 20th century and now out onto the polar ice pack towards 85N 130W) that series of changes has probably changed the grid mostly over North America and less so for Europe. You can see a stronger warming effect in the late 19th century in the Toronto data compared to CET data as the implied "meteo-latitude" gained 5-10 deg north over North America but probably less than 2 deg over Europe. Also the timing structure can be pulled east or west by these variations. On the very long time scale, even if this theory were to be doing very well at year to year scale forecasting, it might not be very easy to determine forecasts hundreds of years ahead because there would also need to be accurate prediction of where the grid would be located.

Anyway, thanks for dropping in Rollo, I saw your comment first over in the model thread and somebody apparently moved it over to here, oddly enough I was typing out a post when you posted about not seeing me and so that appeared right after yours which is why I didn't mention your post there, hadn't seen it while typing.

[syed2878]

@blast from the past. Just wanted to say an excelent post on the moddel thread yesterday afternoon.

[sandstorm2]

Think it is time the plaudits were handed out to RJS for his LRF and all his work regarding the same. The amount of data that he has analysed is nothing short of phenomenal.

You have another convert here RJS! also thanks to Blast from the Past for his input also.

Sandstorm

[hillbilly]

i dont think anyone can predict the weather in this country for more than a few days when the low pressure systems are coming from the atlantic if we have high pressure maybe a week or so for someone to say we are going to have a cold february i guess its going to happen sometime.i am saying we will have a hot and sunny july because we havnt had one since 2006,when was our last cold feb by the way?

[Methuselah]

I'll be the first to congratulate RJS and BFTP, should February's CET be anywhere near -1.1C...You can't get fairer than that...

[Tim Bland]

I have a feeling that as feb goes on we will see undercuting lows and the possiblity of further snow. Im not sure we can tap into enough cold air though for the whole of the UK to see continued snow cover like mid Jan. I expect instead that high ground in the north will see quite a bit whilst Midlands south will see transient snow / rain mix.

[hillbilly]

to get -1.1 for the month we will have to have savage night frosts for 3 weeks or very low maxima and night frosts for the entire month,december 2010 only made -0.7 cet although coldest cet in 100 years for dec,1981 beat it here making -1,only beaten by 1986 at -1.1 as the coldest month for50 years

[Methuselah]

So long as we keep an open-mind though, hillbilly...

[hillbilly]

So long as we keep an open-mind though, hillbilly...

would love to see those kind of temps but its the 7th now i guess you never knowit is northerlies now

[Roger J Smith]

RP, I feel your post above is in the wrong thread, if you want to criticize monthly CET forecasts do that in the monthly thread, if you want to criticize seasonal forecasts, then use the data in the seasonal forecasts (it says quite clearly 2.0 or lower in earlier posts in this thread for my February forecast). I think regular readers of this thread know that my expectation was always for a cold or very cold February but I understand that snow cover and feedback are difficult second-order variables to model well in advance and I was merely trying to draw attention to a cold signal for the month in question when making and posting the LRF in October or whenever it was.

If the mean comes in closer to 2 than -1, that may be a setback for my CET monthly contest hopes but it won't really have much if any impact on the research program in the LRF domain because going as low as 2.0 was already about as cold as I would normally go with a seasonal time-scale forecast.

[VillagePlank]

Nov: Fore 8.26C, Actl 6.8C, Err: +1.46C

Dec: Fore 4.40C, Actl 4.8C, Err: -0.40C

Jan: Fore 2.74C, Actl 3.5C, Err: -0.76C

My Feb forecast is 3.68C

Edited February 8, 2013 by Boar Wrinklestorm

[SnowObsessor90]

So it looks like yet another forecast could well be on it's way to going down the pan.

I guess all it does is reaffirm the point that we're still a million miles away from making a breakthrough in long range forecasting and that's no disrespect to those who do try and give a real go year on year with exception of the likes of Madden of course.

[Methuselah]

RP, I feel your post above is in the wrong thread, if you want to criticize monthly CET forecasts do that in the monthly thread, if you want to criticize seasonal forecasts, then use the data in the seasonal forecasts (it says quite clearly 2.0 or lower in earlier posts in this thread for my February forecast). I think regular readers of this thread know that my expectation was always for a cold or very cold February but I understand that snow cover and feedback are difficult second-order variables to model well in advance and I was merely trying to draw attention to a cold signal for the month in question when making and posting the LRF in October or whenever it was.

If the mean comes in closer to 2 than -1, that may be a setback for my CET monthly contest hopes but it won't really have much if any impact on the research program in the LRF domain because going as low as 2.0 was already about as cold as I would normally go with a seasonal time-scale forecast.

What criticism?

[TEITS]

I still remain unconvinced of our ability to make accurate LRF. Im not picking on anyone in particular but one of the main problems with internet LRF is they all seem biased. What I mean by this for example is many of us love the cold weather & extremes and because of this we seek signals that point towards cold and possibly ignore those that hint at mild. Take for example another well known forecaster who I won't mention and does not post on this forum is always predicting mega blizzards with 200mph winds and a few twisters thrown in at the same time. These forecasts from this person are always OTT but they catch the eye and make great headlines!

This is one of the reasons why I don't make any LRF. Firstly I don't have the scientific knowledge and secondly I don't think I could resist allowing my bias to affect my judgement and as a result I would be predicting endless E,lys. I prefer to stick with the models and look at the Met O 6-30 day forecasts.

Many should follow the example of John Holmes. I know JH doesn't produce LRF but does post his thoughts on what the next few weeks may be like. Now in all the years of being on this forum I haven't known John to post anything but an honest interpretation of the charts with no bias whatsoever. I also like the methods John uses.

So basically if someone produces a LRF that is less dramatic and more in tune with our climate im more likely to read it. If anyone is hinting a record breaking blizzards or hurricanes then I won't bother.

[Methuselah]

Agree with you there, Dave. I do hope, however, that the next generation of supercomputers will considerably extend forecasting abilities...

Snowballz will, no doubt, be far more conversant with all that stuff than me, though...

[UV-RAY]

With winter drawing to an end, we can now start to evaluate LRF for this season, tbh none of them have been spectacular and if anything it just goes to show how difficult it is in making a LRF for the UK. Hats off to all those who try, but IMO we are still a long way off in making an accurate call for any season,perhaps we will never be able to correctly call a season for the UK, but I still enjoy reading them all .

[Gibby]

You can see my thoughts and predictions for Spring 2013 on my website. I will post them here when I have more time.

Without blowing my own trumpet I thought I done reasonably well from my winter prediction giving myself about 75% for accuracy and trend predictions. I did slightly underestimate the wetness of the first two months of Winter though.

Edited February 22, 2013 by Gibby

[Roger J Smith]

Just a brief report on the output of the research model for spring (March to May) 2013 and a peek ahead at summer 2013.

The temperature trends for the spring months are generally near normal with fairly large variations indicated. The model output is currently showing below normal so I have confidence in the starting point; this below normal trend reverses sharply about 7-10 March and it may be equally above normal for a while mid-month, declining to somewhat colder than average near the end. April shows a series of ups and downs on either side of normal, ending with a warm spell that may last a few days into May. The rest of May looks rather backward with the next indication of significant warmth being early to mid June.

I would expect this pattern to bring generally near normal rainfalls as it appears rather blocked at times and mobile at others.

The peek ahead to summer as promised is encouraging; July has the same above normal output that was shown last year in August, where a +1 to +2 anomaly was verified. Therefore unless my research uncovers a lot of new variables that I wish to incorporate, the outlook would be for a warm to hot month of July. Although the August output was near normal, if that followed a warmer than average July there could be feedback to keep anomalies slightly positive, so with June showing about +0.5, the overall forecast for summer is that there may be a warmer pattern than we've seen since 2005 or 2006 in general.

I will leave the winter post mortems to others except to say that the forecast has probably more strengths than weaknesses and maintains my optimism that the research method is worth pursuing. The stated CET values of 4.8, 5.0, 2.0 appear likely to produce a mean error of about 1 C deg and that is not too bad a result, I feel.