2012-03-03 19:13:19Prediction: New Surface Temperature Record in 2013
dana1981
Dana Nuccitelli
dana1981@yahoo...
71.137.110.129

Looks quite likely that 2013 will break the record.  In fact there's actually a decent chance that 2012 will break it.  So I put together a post crunching the numbers.

Prediction: New Surface Temperature Record in 2013

2012-03-03 19:19:33
Alex C

coultera@umich...
67.149.101.148

>>>multiplying the MEI value by approximately 0.075 provides an estimate of its influence on global surface temperatures, according to FR11.  Changes in TSI or sunspot number have a 1 month lag, and multiplying a change in TSI by roughly 0.07

Only for the multivariable regression FR11 did.  Those coefficients are not valid w/o taking into account aerosol forcing.

2012-03-03 19:20:37
Alex C

coultera@umich...
67.149.101.148

Actually let me modify that: the model could be improved using aerosol forcing.  The coefficients as they are now are not optimal.

2012-03-03 19:25:03
Alex C

coultera@umich...
67.149.101.148

I don't understand the 2005/2006 change in the prediction.  If both TSI and ENSO dropped, why did the predicted temperature rise so much?

2012-03-03 19:28:54
MarkR
Mark Richardson
m.t.richardson2@gmail...
134.225.187.225

Perhaps you could estimate a margin of error based on the standard deviation of the difference between the residuals and the observations so far based on Foster & Rahmstorf?

How well does the model do with long term and big El Nino's/La Nina's? I can't remember, it might be worth a glance at the past to see if there's any 'memory' to particularly long/strong ENSO features, since we've just gone through a double dip.

Also, have you looked at the sort of shape of El Nino that would be required to give that average value you looked at?

 

 

I think a big, clear summary sentence or two at the end would be good:

If the Sun warms up and if ENSO averages slightly above average, and if there are no major volcanoes or other changes in aerosol release then the F&R model expects a temperature in 2013 of...

Remember how the 'skeptic' liars have misrepresented Hansen & the IPCC in the past by playing word games to ignore the actual predictions.

2012-03-03 23:15:04
Riccardo

riccardoreitano@tiscali...
2.33.129.107

It would be good to show it graphically using tamino's corrected data over the full period, maybe adding the uncertainty using the variance of the residuals. I'd also make clearer that it's just an exercise and not a real forecast.

2012-03-04 03:50:22
dana1981
Dana Nuccitelli
dana1981@yahoo...
71.137.110.129
Alex - I should explain better, I rather cheated. I'm only 'predicting' changes from one year to the next, so for example I applied the 2005-2006 changes to the 2005 measured, not predicted temperature. I'm just trying to figure out how accurately I can reproduce 1-year changes. If I was trying to reproduce long-term data, then I'd apply the method to the previous predicted temperature.
2012-03-04 06:22:58
logicman

logicman_alf@yahoo.co...
109.150.152.138

I should explain better, I rather cheated.

Is there a statistical method for sticking error bars on that?  ;-)

 

My own 'method' is this: I consider that we are probably going to see a full El Nino this year, I consider it likely that 2012 will produce the lowest ever Arctic albedo, and I conclude that you are probhably correct in your forcast.

2012-03-04 07:35:02
Rob Painting
Rob
paintingskeri@vodafone.co...
118.93.197.214

I agree with MarkR - you should stongly emphasize that your prediction is contingent upon El Nino developing, the solar cycle behaving as expected, and no major volcanic eruptions. We are in 'no analogue' times - there could be a few surprises, so we need guard against that.

Having said that, the data indicate the next two years (and the next half-decade) are going to be hot, hot, hot!.  

2012-03-04 08:56:01
Kevin C

cowtan@ysbl.york.ac...
94.9.112.106

The error bound would be determined by both the uncertainties in predicting ENSO etc, and the uncertainties in the weightings of the exogenous factors in F&R. The latter are fairly significant.

The aerosol forcing is a big unknown (both because it is poorly measured, and the scale is unknown), which could easily be accounted for by the uncertainty inn the weightings in F&R. And there are other unaccounted for factors like stratospheric drying.

I think you're probably right (and have changed my position on the size of the solar term substantially over the past 6 months). Personally, I wouldn't want to gamble SkS' credibility on it, bit that's not my call. OTOH, it might encourage some of the deniers to roll out their own predictions. Which could be interesting.

2012-03-04 10:22:22
dana1981
Dana Nuccitelli
dana1981@yahoo...
71.137.110.129

Oh yeah you guys reminded me.  I meant to talk a bit about aerosols, and also make it clear that this is my prediction, not representative of SkS as a whole (though as John has come to learn, Dana is always right!).

I think barring a volcanic eruption, 2013 will break the record.  Even if my estimate is a full tenth of a degree high it will still break the record, and if I'm right, it will break it by a lot.  Unless the ENSO models are totally wrong and we go into yet another La Niña, and even then it would have to be a reasonably strong one.

The interesting thing will be whether 2012 can break the record, which will depend on how this El Niño shapes up.  If it does, we will probably have 2 consecutive years breaking the record in 2012 and 2013.  But I wouldn't put my money on 2012 breaking the record since January was so cold.  Maybe a statistical tie with 2005 and 2010, again, if El Niño shapes up as expected.

Speaking of which, I'll put together an ENSO graph as Mark suggested, to show what sort of MEI changes my numbers are based on.

2012-03-05 13:09:31
Agnostic

mikepope_9@hotmail...
118.208.167.232

“For example, the Multivariate ENSO index has a 2 to 4 month” … I suggest adding (MEI) after the word “index”

Mention of China’s reduction of particulate emissions is a good point.  It is an on-going government policy aimed primarily at reducing respiratory health problems experienced in major conurbations – but it will certainly have an upward impact on average global surface temperature.  I think the Indian Government has adopted similar policy but have no idea of its implemewntation.

2012-03-05 16:34:11
dana1981
Dana Nuccitelli
dana1981@yahoo...
71.137.110.129

Okay, I added a bit about aerosols and other factors not included in this analysis, a bunch of caveats, the fact that it's my prediction and not that of SkS, and added an MEI plot in.

2012-03-06 09:45:32
skywatcher

andycasely@hotmail...
122.107.164.176

Dana, thanks for doing this!  It's nice to see numbers on the vague thoughts I was having the other day, also interesting (and somewhat troubling) that according to both your estimation and Arthur Smith's, the next few years are forecast to be rather warmer than I would have guessed.  We could actually see the record broken by a year that is more or less ENSO-neutral, and I don't suppose that would be expected to happen very often.

Thumbs up from me.

2012-03-06 10:59:48
dana1981
Dana Nuccitelli
dana1981@yahoo...
64.129.227.4

I was really surprised the numbers came out so high for 2012.  I saw Arthur's post before I ran the numbers, and told him I was skeptical that 2012 could break the record.  Then I got the same answer!

We'll see if they're borne out, given how cool January was.  It all depends on that El Nino.  It would go a long way to shut the deniers up if both 2012 and 2013 break the record.  I'm still skeptical about 2012, but I'd put money on 2013.

2012-03-06 17:40:48
Rob Painting
Rob
paintingskeri@vodafone.co...
203.173.247.60

It would go a long way to shut the deniers up if both 2012 and 2013 break the record

Sadly no, they will just parrot "It's El Nino!!!!"

2012-03-06 17:51:01
dana1981
Dana Nuccitelli
dana1981@yahoo...
71.137.110.129
Heh, not just El Nino, but also a climate shift!
2012-03-06 17:52:38
Rob Painting
Rob
paintingskeri@vodafone.co...
118.92.61.227

Bugger!, not the climate shift!

2012-03-06 18:09:35
Andy S

skucea@telus...
209.121.15.232
This cries out for a Monte Carlo simulation, using probabilistic distributions of the input variables (ENSO, sun, underlying warming trend, aerosols, etc) based on past variation and trend, along with random chances of another Pinatubo or Agung. it would be easy to work out probabilities of any given year beating a certain threshold and to show how those probabilities would increase with time, as well as PDFs for every year. Climate modeling lite.
2012-03-07 11:05:27
michael sweet

sweetdreamfiji@hotmail...
72.91.99.84

It might be interesting to update this post monthly to check how it is panning out.  As ENSO updates are made and monthly temperature data is released you can see how the projection updates.

2012-03-07 12:05:58
dana1981
Dana Nuccitelli
dana1981@yahoo...
71.137.110.129

Well, maybe not monthly, but perhaps every few months.

2012-03-09 16:32:09loonie idea
jyyh
Otto Lehikoinen
otanle@hotmail...
193.64.23.101

Do the 2012 prediction now, save it with full modification details and dates and publish on 1/1/2013. Then do a follow-up post when the actual measurements have been accounted and summarized. This would shorten the interval between the prediction and the actual measurements, something that has been called for by the so called skeptics... (even though it's impossible to do normally). Some of them might actually remember the prediction if it was published a week before the actual measurements come in ;-). (well of course one might link to the prediction published earlier, but it might still be good to publish it w/ the full details of modifications in case someone like McI tries to say its fake.)

Also I'd prefer if the prediction included a variant with a VEI5 volcano going off in the NH (there's already been 20 years since the last one, VEI5s have a frequency of once in 13 years (OK the Pinatubo eruption occurred the same year so the next might be delayed.)) I'd beleive a VEI5 in Iceland would create a 0.5 deg dent in the NH (say 0.25 global) so that woud keep the temperature ~level if there's no large el Nino at the same time.

http://www.volcano.si.edu/world/largeeruptions.cfm

2012-03-09 20:38:21
Riccardo

riccardoreitano@tiscali...
2.33.129.107

The link to ENSO prediction is broken.

You need to specify the surface temperature series you used in the tables.

2012-03-10 03:04:19
dana1981
Dana Nuccitelli
dana1981@yahoo...
64.129.227.4

Yeah I'll post this next week and update with the latest ENSO weekly prediction first (and upload it to SkS so we don't lose the link).  Will also note I'm using GISTEMP.

I'd prefer to just publish the whole thing now, logicman.  I've got a caveat in there about a volcanic eruption nullifying the prediction.  It's too hard to predict either timing or size of eruptions.