[KaiserJeep]

Australian Dr. David Evans is one of the World's leading experts on climate modeling:

Dr David Evans is an electrical engineer and mathematician, who earned six university degrees over ten years, including a PhD from Stanford University in electrical engineering (digital signal processing): PhD. (E.E), M.S. (E.E.), M.S. (Stats) [at Stanford], B.E. (Hons, University Medal), M.A. (Applied Math), B.Sc.[University of Sydney]. His specialty is in Fourier analysis and signal processing. He trained with Professor Ronald Bracewell late of Stanford University.

David has worked in the climate industry, consulting full-time for the Australian Greenhouse Office from 1999 to 2005, and part-time for the Department of Climate Change from 2008 to 2010. He was the lead modeler analyzing the carbon in Australia’s biosphere for Kyoto accounting purposes, and developed the world-leading carbon accounting model FullCAM that Australia uses in the land use change and forestry sector.

In June of 2014 he published his "Notch-Delay Solar Theory", and in the intervening 18 months he has successfully defended the theory in the peer review process. The very briefest summary would be that Earth surface temperatures track solar activity but with a delay imposed by (are you ready for this?) the sunspot cycle, which although variable has over history averaged eleven and a fraction years, but is currently almost exactly 13 years long. To further his goal of a more accurate climate model, he has invented a new mathematical tool called "Optimal Fourier Transforms", which has been a real godsend to those of his fellow EE's here in Silicon Valley who work on cellphone tech, including both phones and cell tower transceivers. The application of Fourier Transforms to climate modelling has but one justification: accurate results. It seems that there is a 13-year-long "notch filter" in global temperatures!

Here's a link to the model (including formulae, datasets, and spreadsheet): http://sciencespeak.com/climate-nd-solar.html

Those of you who wish to download and use the model are welcome to do so, you will need only a reasonably recent version of Excel.

Dr. Evan's model is unique in one aspect: When used in "hindcasting", his model very closely predicts future temperatures and trends, with an incredible accuracy (+/- 3%) versus the IPCC and GISS climate models (+/- 20%). The dataset includes data from 1749 onwards, which is the period where we have recorded observations of both sunspots and surface temperatures. Indeed (although this is a separate topic) Dr. Evans has in the peer review process poked many holes in the egos of the Paleo-Climate crowd who use "derived temperatures" in their models.

Why have I bothered to make this long-winded explanation? Because when used to forecast current temperatures, two controversial things happen:

1) The Evans model says that the forcing effect of carbon dioxide, although real, is 10% or less than that predicted by the IPCC and GISS models. This appears to be correct, because we are in fact not warming as rapidly as the other models predict.

2) The 13-year sunspot "notch filter" says that the peak global temperature will be reached in 2017, which is 13 years after the TSI (Total Solar Insolation) peak in 2004. Then the model predicts actual solar cooling which will accelerate during the 2020's, until a new "Mini Ice Age" begins around 2030. This is not without consequence, the land will be cooler and wetter than today, and much of Canada and Alaska will again be covered in ice, and sea levels will subside below current levels.

=====> The implications are that we could continue to burn oil and coal until we run out, with noticeable but by no means serious climate impacts, somewhere around a maximum of 15% of the current warming predicted by the IPCC and GISS modelers. For however many decades the new "Mini Ice Age" lasts, we are likely to burn FF's at a rate greater than today, simply for space heating needs that will increase beyond today's.

[Plantagenet]

Thanks KJ

Thats a great post. I read the material on the link page, and I'll download the model and start working through the secondary links.

At first glance I have to say I really liked the way he tied the mechanism for his solar insolation model to changes in the earth's albedo linked to variations in the wavelength pattern in the circum-polar vortex. I've been watching the circum-polar vortex for several years using the realtime NOAA data feeds linked to the "earth.nullschool" website---I set the data stream to show the isobaric wind speeds at 5000 hPa to approximate the jet stream. I got interested in this because the vortex plays such a big role in controlling our weather here in Alaska. Recently its been producing cold winters in the midwest and east coast, and giving us warm winters here in Alaska.

my setup on earth.nullschool.net to see the jet stream in real time

During the 20th century the circum-polar vortex has been changing ---something that wasn't predicted by the GCMs used to model greenhouse warming due to CO2 in the atmosphere.

Its very intriguing that Dr. Evans' model incorporates this real world variation in the circumpolar vortex in his model.

Cheers!

[Lore]

This guy is crackers. Look him up, he's been kicked around for several years now.

[Subjectivist]

Even if Dr. Evens were correct about CO2 having a much lower impact than everyone else calculates the amount we have put into the air would not allow glaciation to resume so suddenly.

[PeterEV]

This graph shows a peak in the Total Solar Irradiance was reached in the 1970's:


http://lasp.colorado.edu/home/sorce/files/2011/09/TIM-TSI-Reconstruction1-1024x788.jpg

[Where does the 2004 figure come from? <--oops, in the current 13 year cycle...]

It's interesting to see the TSI going down with temps going up.

From what Plantagenet is saying, the guy may be on to something. This is what science is about. I'd like to see how this plays out before dismissing his ideas and techniques.

[Lore]

Evens has already recanted much of his research. He's going to come up with some new parameters to replace The Notch and forecast imminent global cooling.

So what does this mean for the notch-delay model? Possibly not much, but there will be a delay (ahem) while I recompute things and update the model and graphs.

The causality of a notch doesn’t support the mandatory nature of the delay that I thought I had established, but there is plenty of other evidence to suggest a delay is needed. There are half a dozen independent findings of a delay around about 11 years in the literature, and a solar influence fits better with a delay (e.g. Lockwood and Froehlich 2007 — they show that solar influence doesn’t fit without a delay). Either the sun has little effect beyond the small TSI variations, and we’re left with CO2-driven climate models that don’t work and a mystery about all the warming before 1900, or there is some solar effect that appears to be delayed from TSI changes. While the notch does not *necessarily* mean a delay, certainly the *possibility* of a delay is strong. So there is sufficient reason to include a delay in the solar model — so the model remains as before, with no change. (Clarification: the form stays the same, the parameters will vary).

Pure gibberish!

[Keith_McClary]

...

What is this "peer review process" ? Blog comments?

I glanced through the OFT paper - he doesn't even define it mathematically. But we are welcome to play around with his spreadsheet.

[Lore]

...

What is this "peer review process" ? Blog comments?

I glanced through the OFT paper - he doesn't even define it mathematically. But we are welcome to play around with his spreadsheet.

Why would you waste your time when the author already says it's a fail.

[Keith_McClary]

...

What is this "peer review process" ? Blog comments?

I glanced through the OFT paper - he doesn't even define it mathematically. But we are welcome to play around with his spreadsheet.

Why would you waste your time when the author already says it's a fail.

That was the notch thing, wasn't it? Has he also recanted OFT?

I didn't spend much time on it, it is mostly textbook FT stuff. Without all that padding it wouldn't be a very impressive "paper" since he doesn't seem to know what his OFT is, exactly.

[Lore]

...

What is this "peer review process" ? Blog comments?

I glanced through the OFT paper - he doesn't even define it mathematically. But we are welcome to play around with his spreadsheet.

Why would you waste your time when the author already says it's a fail.

That was the notch thing, wasn't it? Has he also recanted OFT?

I didn't spend much time on it, it is mostly textbook FT stuff. Without all that padding it wouldn't be a very impressive "paper" since he doesn't seem to know what his OFT is, exactly.

That was pretty well debunked back in 2007. This is about David Evan's Notch and Force X from the sun that delays solar energy from reaching the earth by eleven years, or something like that.

I know, you can stop laughing now.

[dissident]

Innumeracy is a sad thing to behold. The TSI drop, even under the worst Maunder minimum ever, will not exceed 0.6 W/m^2. Meanwhile the CO2 radiative effect, based on first principles physics calculations, will exceed 3.0 W/m^2 by 2100. Also, the little ice age was a regional effect that does not show up in analysis of other locations around the world. And, it was a perturbation of a preindustrial reference state. So this "looming" Maunder-like minimum is not even to be expected to produce the same effect as the one associated with the little ice age. In particular the global ice albedo is different thanks to all the glacier melt and the thermal balance of the upper ocean has changed significantly.

[KaiserJeep]

As far as I know, all of the criticism of Evans and his model derive from how he created it.

Classic climate models are "bottom up" models, where all possible inputs are assessed, measured, and input as data into the model. Then whatever theory is being used is the basis is used to create the algorithms. Data then has to be massaged, tweaked, adjusted, and fitted to the algorithms - and in the process, all relevance to the real world kinda gets lost - no two models ever alter the data the same way, endless debates arise, and laymen publish "scandals" about the AGW cheaters - which is a good word for it. The "hindcasting" test has been where such models display accuracy in the +/- 20% range, over the last ~250 years. The problem has always been determining which inputs are significant and which are not, given the negative feedback mechanisms common in nature.

Evans, by contrast, has no idea what many terms in his model represent. He took the observed temperature data and the other inputs and he derived a formula to fit the data. Now believe me, this Fourier Transform technique works quite well when modelling complex analog circuits, and the technique is considered proven and valid. A brief detour:

Simple analog circuits can be modelled with Algebra and Trigonometry. More complex circuits require complex matrix manipulations and computers are a godsend. But when circuits get so small that quantum effects play a part in performance, and so complex that the matrices grow into multiple dimensions, the models are too complex to use. So EE's developed techniques where we varied all the inputs, measured the impacts on the output waveform(s), and then created a model that had only significant factors and neglects everything else.

Evans applied this technique to climate modelling - and his model simply works better than theirs, by about 6X. The IPCC and GISS (i.e. NASA) modellers are outraged at the approach but also can't explain why the model is so accurate.

Be careful when assessing Evan's work. Anything he published before June 2014 was either in support of another model than his, or in criticism of it. His own model is 18 months old, less a week or so.

[Lore]

The model is worthless and by his own admission he is not a climate modeler. I'm not aware he has revisited this since 2014. He was making up stuff to fit Lockwood and Froehlich 2007. A mysterious undetectable Force X from the sun?

He is suggesting a delay is needed, but can't explain how or why. The subject is dead even from the author's point of view, give it up.

[KaiserJeep]

The model is worthless and by his own admission he is not a climate modeler. I'm not aware he has revisited this since 2014. He was making up stuff to fit Lockwood and Froehlich 2007. A mysterious undetectable Force X from the sun?

He is suggesting a delay is needed, but can't explain how or why. The subject is dead even from the author's point of view, give it up.

What you say is simply not true. The spreadsheet model was last updated online on October 18, 2015. The updates are in the notch filter part of the model, and the result of the changes is a better fit to the actual data - error bands have tightened from +/- 3% to something like +/- 2.2% over 98% of the years of recorded data. Simultaneously the causes of the deviations in the remaining years where the predicted temperatures still deviate from actual records +/- 3% are clarified as anomalies in recorded data (mostly where sunspot counts differ from different observers).

There remains the embarrassing, irritating, and inescapable fact that no other climate model approaches the performance of the Evans Notch Theory model - all display larger and larger error bands and are DECREASING IN ACCURACY the more years pass with the "cooling anomalies" we are observing at present.

Admittedly, Evans has very little theoretical underpinnings for his model, compared to most. The "top down" nature of the mathematical technique used to formulate the model practically guarantees this to be the case, as irritating as it is to "normal" climate scientists.

Something else is lurking under the surface of this discussion. The most complicated and dense integrated circuits that we EE's have used the older and very rigorous "bottom up" models on are on the order of a few hundred components and a few hundred circuit nodes. We had to abandon that approach because the state of the art computers we were using simply took too long to model circuit performance, using matrices of up to 12 virtual dimensions.

Never-the-less, the circuits I am talking about are - at the very least - two orders of magnitude less complex than the natural climate systems in play on our world. The implications are that unless someone creates a "bottom up" climate model that exactly models all significant factors while neglecting all insignificant ones - the "bottom up" approach to modelling of climate is simply beyond the state of the art in computing today.

I say, let the boffins continue to have their scholarly debate, let us apply the Evans-created OFT tool to many more "top down" climate models, and base actual decisions and policies on the results - since these types of models display accuracies far beyond the IPCC and GISS models. But I don't think it's gonna happen, the "scientists" consider mathematicians like Evans as mere technicians - and are so incredibly obtuse that they continue to pursue a modelling premise that can be easily described but simply cannot be calculated - we might be talking about 14 dimensional matrices with dozens of terms per axis - maybe we could make such calculations in 100 years - my guess is more like 500 years. I have tried - and failed - to communicate this problem to other Forum members on at least two occasions. I have hope that some of you can understand the limitations of computer climate models a little better after this thread.

No, it's not my particular area of expertise in EE - I worked in digital circuits which used other forms of mathematics - but as a manager I had to learn enough about spreadsheet models to defend and justify a department budget against the bean counters.

Something my Grandfather taught me (in woodworking): If you can't build that dream design in your head using the tools you have, build something else, and keep thinking about what you want - chances are you can figure out some approach, and often you like what you can build better anyway.

If only those wooden-headed "climate scientists" were as wise as that old (he was then 78) dairy farmer/blacksmith/woodworker/etc.

[Keith_McClary]

Simple analog circuits can be modelled with Algebra and Trigonometry. More complex circuits require complex matrix manipulations and computers are a godsend. But when circuits get so small that quantum effects play a part in performance, and so complex that the matrices grow into multiple dimensions, the models are too complex to use. So EE's developed techniques where we varied all the inputs, measured the impacts on the output waveform(s), and then created a model that had only significant factors and neglects everything else.

Evans applied this technique to climate modelling - and his model simply works better than theirs, by about 6X. The IPCC and GISS (i.e. NASA) modellers are outraged at the approach but also can't explain why the model is so accurate..

You said:

he has invented a new mathematical tool called "Optimal Fourier Transforms"

So he invented this last year and it is already widely used in EE even though it has never been published anywhere?
Or he is using this technique, citing no references, and claiming to have invented it. Can you give any references?

[KaiserJeep]

Evans published the OFT tool in 2009, in some issue of IEEE Abstracts on Computing I have read since I retired at the end of May. I'll try to remember which issue. The "Notch-Delay Solar Theory" and model was published June 2014, some 5 years after the tool was developed and used for circuit modelling.

[PrestonSturges]

This graph shows a peak in the Total Solar Irradiance was reached in the 1970's:


http://lasp.colorado.edu/home/sorce/files/2011/09/TIM-TSI-Reconstruction1-1024x788.jpg

[Where does the 2004 figure come from? <--oops, in the current 13 year cycle...]

It's interesting to see the TSI going down with temps going up.

From what Plantagenet is saying, the guy may be on to something. This is what science is about. I'd like to see how this plays out before dismissing his ideas and techniques.

Indeed, radiance was climbing from the 1930s to the 1970s, which was a cooling period.

[PrestonSturges]

Dr. Evan's model is unique in one aspect: When used in "hindcasting", his model very closely predicts future temperatures and trends, with an incredible accuracy (+/- 3%) versus the IPCC and GISS climate models (+/- 20%).

Yeah I can "predict" the winner of every previous World Series by overfitting a model, which is meaningless.

Post hoc ergo propter hoc, baby.

Then the model predicts actual solar cooling which will accelerate during the 2020's, until a new "Mini Ice Age" begins around 2030. This is not without consequence, the land will be cooler and wetter than today, and much of Canada and Alaska will again be covered in ice, and sea levels will subside below current levels.

That's not a "Mini Ice Age," that's a real Ice Age, so bull shit.

[clif]

Evan's denialist theory needs a reputable equant point or possibly a good deferent, to ground it back in any true scientific based approach to the climate changes we are currently all experiencing.

[davep]

You've posted his rubbish before. He's a known sceptic, and has based his "research" on his idea that all the AGW scientists are on some gravy train. This predates and therefore colours all his (non-peer reviewed) research.

http://www.sourcewatch.org/index.php/David_Evans_%28Australian_skeptic%29