Guest Post by Willis Eschenbach
It has been pointed out that while many of the global climate models (GCMs) are not all that good at forecasting future climate, they all do quite well at hindcasting the 20th-century global temperature anomaly [edited for clarity – w.]. Curious, that.
So I was interested in a paper from August of this year entitled The energy balance over land and oceans: An assessment based on direct observations and CMIP5 climate models. You’ll have to use SciHub using the DOI to get the full paper.
What they did in the paper is to compare some actual measurements of the energy balance, over both the land and the ocean, with the results of 43 climate models for the same locations. They used the models from the Fifth Climate Model Intercomparison Project (CMIP5).
They compared models to observations regarding a suite of variables such as downwelling sunlight at the surface, reflected sunlight at the top of the atmosphere (TOA), upwelling TOA thermal (longwave) radiation, and a number of others.
Out of all of these, I thought that one of the most important ones would be the downwelling sunlight at the surface. I say that because it is obvious to us—sunny days are warmer than cloudy days. So if we want to understand the temperature, one of the first places to start is the downwelling solar energy at the surface. Downwelling sunlight also is important because we have actual ground-truth observations at a number of sites around the globe, so we can compare the models to reality.
But when I went to look at their results, I was astounded to find that there were large mean (average) errors in surface sunshine (modeled minus observed), with individual models ranging from about 24 W/m2 too much sunshine to 15 W/m2 too little sunshine. Here are the values:
Now, consider a few things about these results:
First, despite the average modeled downwelling sunshine at the surface varying by 40 W/m2 from model to model, all of these models do a workmanlike job of hindcasting past surface temperatures.
Next, the mean error across the models is 7.5 W/m2 … so on average, they assume far too much sunlight is hitting the surface.
Next, this is only one of many radiation values shown in the study … and all of them have large errors.
Next, results at individual locations are often wildly wrong, and …
Finally, we are using these models, with mean errors from -15 W/m2 to +23 W/m2, in a quixotic attempt to diagnose and understand a global radiation imbalance which is claimed to be less than one single solitary watt per square metre (1 W/m2), and to diagnose and understand a claimed trend in TOA downwelling radiation of a third to half of a W/m2 per decade …
I leave it to the reader to consider and discuss the implications of all of that. One thing is obvious. Since they can all hindcast quite well, this means that they must have counteracting errors that are canceling each other out.
And on my planet, getting the right answer for the wrong reasons is … well … scary.
Regards to all on a charmingly chilly fall evening,
w.
PS—As usual, I request that when you comment you quote the exact words you are discussing so we can all understand who and what you are referring to.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Have guest posted here before on unavoidable parameter tuning and the attribution problem it brings into all climate models. Also, AR5 had a lengthly discussion of the cloud problem in Chapter 7, which these results reflect (pun intended). This is a nice post showing the model mess that results.
And per Ross McKittrick, the early CMIP6 sensitivity results are worse (higher, so more discordant to observational energy budget methods). My speculation based on the CMIP5 ‘experimental design’ is the traditional mandatory 30 year temperature hindcast had to incorporate all of the pause as well as half the temperature rise before. That really torques parameter tuning.
Hey, producing a computer model of what has happened is easy. I mean, seriously easy. Even that doesn’t mean it’s right, however, for reasons that anyone familiar with computer models will be all too familiar with. Are variables truly independent, for instance, etc, etc, etc. That is the whole point of models — they help you understand — at least to some extent — what has happened.
However, using models to predict the future isn’t just difficult, it’s impossible, unless of course you’re talking about a truly deterministic system, which is and never will be the case with either weather or climate. If you do use a model to predict the future and it happens to be right the chances are you’ve just been lucky.
There’s a very easy way to prove this, which has nothing to do with climate but everything to do with economics. Why? Because if models could predict the future there’d never be any economic problems again — the models would enable us to avoid them.
If this can’t be done with economics it can’t be done with climate. What’s more it never will be possible because whatever people say the climate ultimately depends on weather and weather is chaotic. It’s exactly the same with economics.
People who believe this rubbish really need to be asked to solve the world’s economic problems. It wouldn’t take long for them to be found out.
Funny! So much in-depth, intelligent discussion about reading tea leaves.
Yes, funny, … in a disturbing sort of way.
Yeh… but the dingbats outnumber intelligent people and are easily winning the debate through weight of numbers and noise.
I drove past a climate extinction gathering in a remote part of Victoria, Australia yesterday. The people living in this and similar regions are not normally regarded as dingbats. Their drug of choice is for chilling not for thrilling.
“global climate models (GCMs)”. Nice one, w.
No this is a physical problem. The cult CAGW/AGW are 100% incorrect.
There is unequivocal evidence from multiple independent lines of reasoning and data that shows humans caused less than 5% of the recent rise in atmospheric C02. Atmospheric CO2 is tracking planetary temperature not human CO2 emissions.
This is an interesting presentation that shows planetary temperature is tracking solar wind bursts which has the same periodicity of past climate change
Interesting that based on past changes we will experience cooling. Cooling is likely the only thing that could stop this madness.
https://youtu.be/l-E5y9piHNU
Supposedly atmospheric co2 is approaching saturation (?) If so, doesn’t the predicted warming require hypothetical feedback forcing? Isn’t it here that sceptics should be turning their attention to?
The Russian INMCM4 model is the one model closest to actual measured temps … it is significantly outside the rest of the ensembles projections.
In the paper Willis posts here the same model also stands out in some of the data noted. I’m not smart enough to interpret exactly how its differences relate to its seeming better accuracy at forecasting, however, Rob Clutz has written some good articles on the INMCM4 and INMCM5 models and their differences.
https://rclutz.wordpress.com/2015/03/24/temperatures-according-to-climate-models/
https://rclutz.wordpress.com/2017/10/02/climate-model-upgraded-inmcm5-under-the-hood/
https://rclutz.wordpress.com/2018/10/22/2018-update-best-climate-model-inmcm5/
Paper on INMCM5 – Volodin 2018:
https://www.earth-syst-dynam.net/9/1235/2018/
Climate model data:
http://www.glisaclimate.org/model-inventory
If a model can hindcast without heuristic massaging, or regular injections of brown matter (“fudging”), then it can forecast… within a limited frame of reference (e.g. conservation of mass, energy, processes). The limits of science are established by incomplete or insufficient characterization and unwieldy processing. We can infer the past and predict the future, but it is philosophy, not science, based on physical myths and assumptions/assertions that may or may not be supported as we reduce the uncertainties, and our hindcast and forecast skills are similarly limited.
People not familiar with multi-variable predictive models may be surprised that hind-casting doesn’t mean forecasting will be accurate, but those of us familiar with such models do not.
Given any set of data and a sufficient number of variables to play with, you can fit the data in hundreds (or thousands) of ways that have NOTHING to do with forecasting. Forecasting requires actual understanding so that one knows which variables are relevant so that the model is actually based on physical realities and not on hunches, wishes, or guesses.
If the system is a complex one (and climate surely is), then it is doubtful one will ever be able to accurately forecast more then some relatively small period of time (maybe 20 years, maybe 30?). So far, they cannot accurately predict next year so they have a lot of room for improvement.
Anyone who thinks we have the ability to predict climate out for 100 years is either naive or just incapable of understanding the problem.
With the increasing number of PV solar installations, I suspect a lot more of that data will be obtainable in the future.
The CO2-driven climate computer models cannot work because they have the time sequence backwards, by falsely assuming that atmospheric CO2 is the primary driver of global temperature.
Minor variations in solar intensity drive tropical sea surface temperatures, which are also modulated by sub-decadal ENSO ocean oscillations and multi-decadal oscillations dominated by the PDO (and AMO).
Then:
* “6. The sequence is Nino34 Area SST warms, seawater evaporates, Tropical atmospheric humidity increases, Tropical atmospheric temperature warms, Global atmospheric temperature warms, atmospheric CO2 increases (Figs.6a and 6b).”
Other drivers such as fossil fuel combustion, deforestation ,etc. may also be drivers of increasing atmospheric CO2, but this is largely irrelevant to climate and hugely net-beneficial, because it greatly increases crop yields.
In summary, the global warming alarmists, including their CO2-driven climate models, could not be more wrong in their fundamental assumptions. “Cart before horse.”
Regards, Allan
* Reference:
CO2, Global Warming, Climate And Energy
by Allan M.R. MacRae, B.A.Sc., M.Eng., June 15, 2019
https://wattsupwiththat.com/2019/06/15/co2-global-warming-climate-and-energy-2/
Excel: https://wattsupwiththat.com/wp-content/uploads/2019/07/Rev_CO2-Global-Warming-Climate-and-Energy-June2019-FINAL.xlsx
Summary from my previous posts:
The current climate hysteria is a well-funded global political campaign, conducted by the wolves to stampede the sheep. Why now? Because the global warming scam will soon come tumbling down, where even the most devoted warmist acolytes will realize they have been duped. How will this happen?
The failed catastrophic very-scary catastrophic global warming (CAGW) hypothesis, which ASSUMES climate is driven primarily by increasing atmospheric CO2 caused by fossil fuel combustion, will be clearly disproved because fossil fuel combustion and atmospheric CO2 will continue to increase, CO2 albeit at a slower rate, while global temperatures cool significantly. This global cooling scenario has already happened from ~1940 to 1977, a period when fossil fuel combustion rapidly accelerated and atmospheric temperature cooled – that observation was sufficient to disprove the global warming fraud many decades ago.
Contrary to global warming propaganda, CO2 is clearly NOT the primary driver of century-scale global climate, the Sun is – the evidence is conclusive and we’ve known this for decades.
_________________________
In June 2015 Dr. Nir Shaviv gave an excellent talk in Calgary – his slides are posted here:
http://friendsofscience.org/assets/documents/Calgary-Solar-Climate_Cp.pdf
Slides 24-29 show the strong relationship between solar activity and global temperature.
Here is Shaviv’s 22 minute talk from 2019 summarizing his views on global warming:
Science Bits, Aug 4, 2019
http://www.sciencebits.com/22-minute-talk-summarizing-my-views-global-warming
At 2:48 in his talk, Shaviv says:
“In all cores where you have a high-enough resolution, you see that the CO2 follows the temperature and not vice-versa. Namely, we know that the CO2 is affected by the temperature, but it doesn’t tell you anything about the opposite relation. In fact, there is no time scale whatsoever where you see CO2 variations cause a large temperature variation.”
At 5:30 Shaviv shows a diagram that shows the close correlation of a proxy of solar activity with a proxy for Earth’s climate. More similar close solar-climate relationships follow.
Shaviv concludes that the sensitivity of climate to increasing atmospheric CO2 is 1.0C to 1.5C/(doubling of CO2), much lower than the assumptions used in the computer climate models cited by the IPCC, which greatly exaggerate future global warming.
At this low level of climate sensitivity, there is NO dangerous human-made global warming or climate change crisis.
__________________________
Willie Soon’s 2019 video reaches similar conclusions – that the Sun is the primary driver of global climate, and not atmospheric CO2.
https://wattsupwiththat.com/2019/09/15/global-warming-fact-or-fiction-featuring-physicists-willie-soon-and-elliott-bloom/
Willie Soon’s best points start at 54:51, where he shows the Sun-Climate relationship and provides his conclusions.
There is a strong correlation between the Daily High Temperatures and the Solar Total Irradiance (54:51 of the video):
… in the USA (55:02),
Canada (55:16),
and Mexico (55:20).
_________________________
http://woodfortrees.org/plot/pmod/offset:-1360/scale:1
Solar Total Irradiance is now close to 1360 W/m2, close to the estimated lows of the very-cold Dalton and Maunder Minimums. Atmospheric temperatures should be cooling in the near future – maybe they already are.
We know that the Sun is at the end Solar Cycle 24 (SC24), the weakest since the Dalton Minimum (circa 1800), and SC25 is also expected to be weak. We also know that both the Dalton Minimum and the Maunder Minimum (circa 1650-1700) were very cold periods that caused great human suffering.
I wrote in an article published 1Sept2002 in the Calgary Herald that stated:
“If [as we believe] solar activity is the main driver of surface temperature rather than CO2, we should begin the next cooling period by 2020 to 2030.”
That prediction was based of the end of the Gleissberg Cycle of ~80-90 years, dated from 1940, the beginning of the previous global cooling period from ~1940 to 1977.
Since about 2013, I have published that global cooling will start by 2020 or earlier. Cooling will start sporadically, in different locations.
Planting of grains in the Great Plains of North America was one month late in both 2018 and 2019. Summer was warm in 2018 and the grain crop was successful. However spring was late and wet in 2019, and much of the huge USA corn crop was never planted due to wet ground; then the summer was cool and winter snow came early, resulting in huge crop failures.
Thousands of record cold temperatures were experienced in North America in October 2019, and temperatures in Britain and parts of northern Europe were also extremely cold.
Recent analysis of the 2019 harvest failure is here:
THE REAL CLIMATE CRISIS IS NOT GLOBAL WARMING, IT IS COOLING, AND IT MAY HAVE ALREADY STARTED
By Allan M.R. MacRae and Joseph D’Aleo, October 27, 2019
https://wattsupwiththat.com/2019/10/27/the-real-climate-crisis-is-not-global-warming-it-is-cooling-and-it-may-have-already-started/
GROWING SEASON CHALLENGES FROM START TO FINISH
By Joseph D’Aleo, CCM, AMS Fellow, Co–‐chief Meteorologist at Weatherbell.com, Nov 18, 2019
https://thsresearch.files.wordpress.com/2019/11/growing-season-challenges-from-start-to-finish.pdf
Bundle up – it’s getting colder out there.
So they just add more sun to explain past warming.
Then observe a lack of sun in the present.
Observe warming.
Models prove CO2 causes warming?
“they must have counteracting errors that are canceling each other out.”
No need to presume such lucky activity in the model black box, just emulate it with Pat Frank’s method, and propagate the uncertainty.
Sorry to be dim. What does “downwelling sunshine” mean? A quickiewiki suggests it may be the proportion of of the sun’s radiation that hits the surface, penetrates the top layers of the sea.
Hywel, the radiation flows in general are characterized as downwelling (headed towards earth) or upwelling (headed towards space). At both the surface and the top of the atmosphere we have both downwelling and upwelling (reflected) sunshine.
w.
The second sentence I agree with completely. The first sentence, without qualification, is not so where I live. In the winter season especially we have warm periods of a couple of days where southerly air flows in advance of an approaching front, and this brings higher temperatures. There may be a good deal of cloudiness in these circumstances especially up against mountain ranges. Clear sunny days in the winter season may be quite cold for several reasons. First, these occur often after the passage of a front and are in polar air, and because overnight radiation produces cold nights the NWS can miss temperature predictions by more than -10F. What I could agree with is sunshine produces higher temperatures caeteris paribus.
I took up a hobby in the past year of purchasing a number of Eppley Laboratory radiometers (PIR, pryheliometers, and PSPs) at auction and putting them back to work. This has caused me to notice local sky conditions as reported by the local AWOS more closely — and these are notoriously wrong. I have also been surprised by the magnitude of solar radiation declines resulting from tenuous high altitude clouds. On clear days, with some widespread wispy cirrus clouds only visible when a less tenuous knot of several such coincide below the sun, there are momentary deviations of
commonly. I am not certain these are even detectable from satellites. I would be not surprised to find we don’t know instantaneous solar irradiance at the ground surface to an accuracy of better than 
. It would take more detail than the climate models possess to get these values right.
A rock climbing friend of mine agrees with me that at the elevations we spend most of our lives (above 7,000 ft) it takes almost no clouds at all to take the warmth out of sunshine.
From:
Remote sensing of earth’s energy budget:
synthesis and review
Shunlin Liang, Dongdong Wang, Tao He & Yunyue Yu (2019)
Kindly sent to me by Mr. Mosher.
After describing the 6 watt spread in surface incident sunshine in various satellite and surface measurements:
“Itisclearthatexistingknowledge on incident solar radiation still has large uncertainties.”
We can’t even measure surface incident sunshine, much less compare it to models based on our faulty information.
Thank you sir. It wur news to oi, and “welling” seems a funny way of describing radiation. Just a word, I suppose.
A transparent ocean absorbs more energy than land. Any land.
“Average land” doesn’t absorb as much energy as land surfaces humans create {UHI effect is an effect because warmer}. Humans can also make greenhouses which warmer than land surface.
It’s my guess, that humans can’t make surface that absorbs as much sunlight as Earth ocean. But a solar pond is something close to an ocean.
Any body of water warms the air more than land surface as there no difference between the water surface temperature and the air above it. Land surface tend commonly get to 60 C, but air temperature doesn’t get this warm. Highest surface temperature are around 70 C and highest air temperatures are about 50 C.
So, the tropical ocean which has average temperature of about 26 C, is the heat engine of the world. And tropical land surface, is not the heat engine of the world.
Now only way I think of tropical land as ever being a heat engine of the world is if the land is at higher elevation. So large percentage of earth surface area being land and it is in the tropics. Anyways with our current configuration of ocean and land, we have about 80% of tropics as ocean area.
I am not sure that higher elevation land in the tropics would work as well as ocean does- and it would depend on the type of land.
Another thing I wonder about is Equatorial bulge, both in regards to higher land elevation and in effect in terms of our present world. But if had a large ocean area at higher elevation, but like land would seem to do, if ocean was higher elevation it would add to the effect of tropical heat engine.
gbaikie, lots of interesting points in there. The ocean does absorb more overall energy (LW + SW) than land. CERES datas says 533 W/m2 for ocean, 447 W/m2 absorbed for land.
Given that ocean area is 72% of the surface, this means that about three-quarters of the absorbed downwelling energy goes into the ocean.
w.
–gbaikie, lots of interesting points in there. The ocean does absorb more overall energy (LW + SW) than land. CERES datas says 533 W/m2 for ocean, 447 W/m2 absorbed for land.–
447 W/m2 absorbed for land in Tropics?
That interesting, it is commonly said tropics {40% of Earth surface} gets more than 1/2 the sunlight which reaches the entire Earth surface. Which is obviously due the the sun spending more time nearer to zenith than outside of the tropics. But never heard of the actual numbers being given.
So according to Energy budgets, the average is about 163 watts per square meter. And of course if not for Earth having oceans the number would much lower than 163 watts per square meter.
Btw this disproves the silly notion that without CO2, Earth would not warm enough to have water vapor- as the tropics gets so much sunlight, it will always have water vapor. One could possibly support idea there might less water vapor, but not the case of having no water vapor.
Or I believe CO2 does cause some warming, and/or I can’t disprove that- but disagree it’s a control knob. And I believe ocean average temperature is actually the control knob. So currently the average is 3.5 C. If it was 2 C, we would without doubt be in glacial period, and if 4 C, one is always in interglacial period. And during our Ice Age the ocean temperature {entire volume average temperature, it has been 1 to 5 C. And if ever gets warmer than 5 C, we leaving our Ice Age, or we still in the Ice Age or Icehouse climate. And will remain in our millions year old Ice Age, for at least 1000 years.
“Given that ocean area is 72% of the surface, this means that about three-quarters of the absorbed downwelling energy goes into the ocean.”
Well it’s said if Ocean was not retaining the heat from the Sun {or heat wasn’t “lost in ocean oceans”} global air temperature would be much higher. Which bad way to “explain it”, I would say they failed to account for the ocean. There models fail because they not including the most important aspect of global temperature, the actual control knob, our oceans.
The big difference between land and water is that land does not store the energy. It loses most of what it gains each day the next night.
Oceans absorb energy primarily in the lower latitudes and releases it at the higher latitudes over an annual cycle. That redistribution of energy drives the global weather cycle.
Another key factor is the view the sun has of earth when it shines brightest:
https://www.google.com.au/maps/@-9.9788079,-164.2336179,2.98z
Almost all water.
Compared to when it shines at its lowest:
https://www.google.com.au/maps/@13.5216001,49.1638295,2.96z
Much greater proportion of land.
So the globe absorbs energy during the austral summer and releases it during the boreal summer.
Could some of these errors be the real world manifestation of the inherent error in the models that Patrick Frank wrote about ?
w. ==> It has been clear to all who bother to do any research of climate models that the modellers have been tuning their models to hindcast — they do this by adjusting various parameters until the model makes the hindcast desired.
Nakamura made this crystal clear in his recent book.
When they say “hindcast” they really mean “an ensemble of hindcasts with the correct mean” — hindcasts have the same chaotic output as forecasts…..results all over the place constrained by parameters hard-coded into the model — the resulting “mean” just being a mean of parameter-allowed outputs.
Yours above is proof of the problem — any value for down-dwelling sunshine can be used as long as other parameters are adjusted to produce “correct” historical outcomes.
And now for something completely different, regarding energy transfer in the atmosphere.
https://blog.friendsofscience.org/wp-content/uploads/2019/08/July-18-2019-Tucson-DDP-Connolly-Connolly-16×9-format.pdf
Slides by Dr. Ronan Connolly and Dr. Michael Connolly
“Balloons in the Air: Understanding Weather and Climate”
The clue to look at this was found at Climate Audit, in a comment from ngard2016 to niclewis.
https://climateaudit.org/2019/10/17/gregory-et-al-2019-unsound-claims-about-bias-in-climate-feedback-and-climate-sensitivity-estimation/
This has big implications for climate models and the greenhouse gas theory.
Andy May and Christopher Monckton of Brenchley have commented on the Connollys’ ideas on WUWT before. But the slides are the best place to start.
Toto
Looks quite interesting.
But I remind my big surprise upon a comparison made 3 years ago, of the full IGRA radiosonde data set (1,500 stations worldwide at that time) with a very small subset of it (RATPAC B, 85 stations).
While IGRA is raw raw data, RATPAC B is highly homogenised and calibrated according to huge work performed at Wien’s University (Austria) by Leopold Haimberger & team, which resulted in the well known RAOBCORE/RICH software package.
I found the data I collected at that time, and show you the difference between homogenised and raw data for the whole Globe.
1. RATPAC B hom vs. UAH 6.0 land at 500 hPa (best fit):
https://drive.google.com/file/d/1a3p4ifwPEXJ8ZetEVZz64E10VrHjtE6p/view
2. RATPAC B hom vs. RATPAC B raw IGRA vs. full IGRA
https://drive.google.com/file/d/1R6UcAkDmFN_fuuEaqEPmKawDqVZmUC17/view
You see how much homogenisation and calibration work would be necessary to obtain from the full IGRA data set
– not only the necessary fit to a sat-based temperature data as provided by RATPAC,
but above all
– a healthy, reliable basis for the credible substantiation of the hypothesis established by Conolly & Conolly.
I hope this work was done at Wyoming’s University.
Regards
J.-P. D.
Ooops! ‘remind’ -> ‘recall’… 3rd language syndrome.
I actually have a method to win at roulette, but it only works at the 1 dollar table.
Bet one dollar on red. If you lose, bet 2 dollars, then if you lose bet 4 dollars. Keep going, doubling your bet if you keep losing.
Eventually you will win. You will have then won 1 dollar.
Keep it up all night and you might win a few hundred dollars.
This does not work for the 5 dollar table, for obvious reasons unless you have really deep pockets.
That’s the Martingale system, won’t win on any table with a maximum stake limit.
Hi Willis, the paper in your first link is from 2014. You might want to look at Wild’s recent paper https://link.springer.com/article/10.1007%2Fs40641-017-0058-x.
Thanks, Geoff, most interesting.
w.
And this recent Wild paper is open access, Estimating Shortwave Clear‐Sky Fluxes From Hourly
Global Radiation Records by Quantile Regression https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2019EA000686
OK, one more by the way, the Wild paper (actually published in 2015) is open access https://link.springer.com/article/10.1007/s00382-014-2430-z