In Part One I made the observation:
If the atmosphere has an invariant optical thickness then surely all molecules should be included?
Meaning all ‘radiatively-active’ gases. Then I cited some results from Collins (2006) on the ‘radiative forcing’ for other gases, and added:
..So if total optical thickness from CO2 and water vapor has stayed constant over 60 years then surely total optical thickness must have increased?
In response, Miskolczi supporter Miklos Zagoni said:
Optical thickness was calculated over 60 years for CO2 and water vapor and other 9 IR-active molecular species (O3, N2O, CH4, NO, SO2, NO2, CCl4, F11 and F12), and turned out to be strictly fluctuating around a theoretically predicted equilibrium value
I asked for more details (concentrations of each of these gases over time which were used for the calculations) which weren’t forthcoming.
Later Miskolczi supporter Ken Gregory said:
Only the H2O and CO2 gases were changed. Other minor GHG were held constant.
So, working with this data I thought it would be interesting to see what changes had taken place in optical thickness due to these minor “greenhouse” gases.
I should point out that there are substantial problems identified with Miskolczi’s theory and experimental work and this is a very minor issue – it is more of an interesting aside.
A little while ago I managed to recreate the CO2 transmittance in the atmosphere – as shown in Understanding Atmospheric Radiation and the “Greenhouse” Effect – Part Nine. This was done using the HITRAN database in a MATLAB model I created.
The question about changes in optical thickness over time from other gases was a good motivator to update my MATLAB model to bring in other molecules. It was something I wanted to do anyway.
Note that radiative forcing or (surface emission – OLR) is a much more useful value than total optical thickness (as explained in Part One).
Extracting the HITRAN data proved to be the most tedious and challenging part of the project. It turns out that the “minor gases” like CFC-11 and CFC-12 are stored in a totally different format from gases like CO2, N2O, CH4 etc. These minor gases have a dataset for each temperature and pressure, with different sizes of dataset at various temperature/pressures. Nothing mathematically or conceptually challenging, just very tedious.
Another challenge was working out what concentrations to use for 1948 – the start date that Miskolczi uses. From Collins (2006) it seemed that the main “greenhouse” gases to evaluate were N2O (nitrous oxide), CH4 (methane) plus CFC11 (CCl3F) and CFC12 (CCl2F2). There are other halocarbons to include but time is limited.
Here are the values used:
……………………1948 2008
CO2 311 ppmv 386 ppmv
N2O 289 ppbv 319 ppbv
CH4 1250 ppbv 1775 ppbv
CFC11 0 267 pptv
CFC12 0 535 pptv
The later CO2 value is from 2008 from Miskolczi’s spreadsheet while the other values are from 2005.
ppmv = parts per million by volume, ppbv = parts per billion (109) by volume, pptv = parts per trillion (1012) by volume.
Earlier values of N2O and CH4 are taken from various papers, I can provide citations if anyone is interested – but pre-1980 values are thin on the ground.
In any case, my calculations of total optical thickness are rudimentary and provided as a starting point.
The Model
I used a 5 layer model up to 200mbar, with a surface temperature of 289K. The diffusivity approximation was used to estimate total hemispherical transmittance (see Understanding Atmospheric Radiation and the “Greenhouse” Effect – Part Six – The Equations). The wavenumber step, Δν = 1 cm-1. The calculations were done from 100 cm to 2500 cm (4μm – 100 μm) and the “Planck weighted” transmittance (at 289K) was calculated. This transmittance was converted back to an optical thickness, which is the same approach that Miskolczi uses (see comment).
Water vapor was assumed to be 10g/kg at the surface with a straight line reduction (vs pressure) to zero at 200mbar. Previously I carried out calculations where water vapor was varied from 5g/kg to 15g/kg and the effect on the transmittance change due to other gases was quite small.
Water vapor absorption lines are included from the HITRAN database but the water vapor continuum is not. This is next in my wishlist to include.
Changes in Water Vapor
The model deliberately did not try to follow Miskolczi’s water vapor values. The point of this article is to demonstrate that if (and only if) CO2 optical thickness is canceled out by water vapor changes, then significant increases in optical thickness from other gases impact negatively on his hypothesis.
If his calculations show:
optical thickness (CO2 + water vapor) = constant
then this article demonstrates that:
optical thickness (CO2 + other gases + water vapor) = increasing
Many people might not realize that there are a number of water vapor datasets. The one Miskolczi uses is not the only one. Others show different trends.
Results
Note that water vapor is included, but at unchanged concentration.
- The change in optical thickness, Δτ, for CO2 only changing = 0.0167
- The change in optical thickness, Δτ, for CO2+N2O+CH4+CFC11+CFC12 = 0.0238
The % increase (over CO2) due to the nominated “minor gases” = is 43%.
The total optical thickness is not so important in this analysis. If the number of layers is changed, the total optical thickness changes, but percent changes due to “greenhouse” gas increases are roughly similar.
Conclusion
If (and only if) water vapor has canceled out CO2 increases, then the increase in optical thickness due to these other gases (methane, nitrous oxide plus halocarbons) has destroyed the idea that optical thickness can be considered to be constant.
Of course, my calculations are rudimentary. My model is much less exact than the HARTCODE model used by Miskoczi and it would be interesting to see his results reproduced in full with the correct concentrations of all of the GHGs from 1948 – 2008.
As I commented earlier – this is one of the least important of the criticisms of Ferenc Miskolczi’s papers.
Now I have updated the model I can produce results like these:
Other articles in the series
The Mystery of Tau – Miskolczi – introduction to some of the issues around the calculation of optical thickness of the atmosphere, by Miskolczi, from his 2010 paper in E&E
Part Two – Kirchhoff – why Kirchhoff’s law is wrongly invoked, as the author himself later acknowledged, from his 2007 paper
Part Three – Kinetic Energy – why kinetic energy cannot be equated with flux (radiation in W/m²), and how equation 7 is invented out of thin air (with interesting author comment)
Part Four – a minor digression into another error that seems to have crept into the Aa=Ed relationship
Part Five – Equation Soufflé – explaining why the “theory” in the 2007 paper is a complete dog’s breakfast
References
The HITRAN 2008 molecular spectroscopic database, by L.S. Rothman et al, Journal of Quantitative Spectroscopy & Radiative Transfer (2009)
Radiative forcing by well-mixed greenhouse gases: Estimates from climate models in the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report (AR4), Collins et al, JGR (2006)
The Science of Doom 
10 g/kg water vapor at 289 K is ~90% RH. That’s a bit high don’t you think? Why not use one of the standard humidity profiles like the US 1976 standard atmosphere? I find this link to be a useful reference for atmospheric humidity calculations:
http://www.humidity-calculator.com/index.php
What about tropospheric ozone? Of course, that’s probably not well-mixed.
Good point.
The reason for high humidity was because the relative effect of the minor gases was reduced at higher humidity. So it presented more like a worst case.
I am currently running a set of simulations, (all with surface temperature of 289), of water vapor at 5, 10 & 15g/kg with a 10-layer model for the 3 cases
– 1948
– 2008 with CO2 only increasing
– 2008 with CO2 plus the other gases increasing
Each run takes 2.5 hrs x 9 runs = means results are a little way off.
The spatial variation of ozone makes it tricky. It is something I would like to add.
The other missing part of the model is the Voigt profile to enable accurate calculations into the stratosphere.
For other readers, the Voigt profile is the line shape profile that combines the effect in the lower atmosphere – pressure broadening – with the effect nearer the top of the stratosphere – Doppler broadening.
The Voigt profile is computationally very expensive. I have wasted a bit too much time already looking into some of the faster calculation methods..
Bah, Mis uses TIGR 2. TIGR 2 is well known to be too dry in the tropics. It’s equation souffle with bad data chopped in.
DeWitt Payne said:
“Why not use one of the standard humidity profiles like the US 1976 standard atmosphere?”
The USst-76 atmosphere contains only 1.261 prcm water vapour. This is about half of the true global average.
The USst-76, NOAA and GAT water vapour profiles are shown here:
The NOAA and GAT water vapour contents are 2.618 prcm and 2.610 prcm, respectively. These are very similar.
I asked Miskolczi to prepare HARTCODE runs to test the CO2 sensitivity with the NOAA versus USsts-76 atmospheres. See the results here:
Miskolczi’s graph shows that the change of tau at CO2 doubling is 2.00% using the USst-76 water vapour profile, and is 1.32% using the NOAA water vapour profile. The tau change using the NOAA profile is 1.32/2 = 66% of the tau change using the USST-76 profile.
If you leave out half of the actual water vapour, you calculate too high sensitivity to CO2.
Eli Rabett said:
“TIGR 2 is well known to be too dry in the tropics.”
We are doing global calculations. TIGR2 global average matches the NOAA global average very well. The USst-76 atmosphere is much too dry. This is why Miskolczi write in the caption to Figure 2 of M2010 “Since H2O is the most important greenhouse gas, apparently the USST-76
Atmosphere is not suitable for global radiative balance studies.”
SoD, thanks for the calculation with changing the concentrations of minor GHG. Very interesting.
Nuts, TIGR3 is much better than TIGR2 for tropical water vapor and the tropics are where the action is for the greenhouse. Since increasing greenhouse gases RAISES the tropopause and warms it (e.g. the column water vapor increases) and more water vapor leaks through to the stratosphere you and your friend are playing with a couple of jokers.
Try that
Here are the results from a slightly more comprehensive model with a sensitivity analysis for water vapor.
The model now has 10 layers up to 200mbar.
The wavenumber resolution was 1cm-1 as before. Hemispherical averaging via the diffusivity approximation as described earlier.
Three different values of surface water vapor – 5g/kg; 10 g/kg; 15 g/kg – were used for each of 3 conditions – 1948 values of GHGs; 2008 values of CO2 with 1948 values for “minor” GHGs; 2008 values for CO2 and minor GHGs nominated.
The 9 runs took a total of just under 24 hours on a standard laptop with Matlab.
I think if the water vapor continuum was introduced it might slightly reduce the effect of the halocarbons (CFC11 & CFC12).
SoD,
On your table with 1948 and 2008 values, I think you meant CFC11 and CFC12.
Thanks, I fixed it.
Ken Gregory,
What really determines the sensitivity of CO2 is the height of the tropopause and that’s determined by the surface temperature. In terms of W/m², doubling CO2 changes the forcing in the tropics by ~5 W/m² but only by 2 W/m² at the South Pole. Comparing relative sensitivities by percentage is nugatory. There’s good reason why change in forcing at the tropopause is the proper metric. Forcing tells you how much the surface temperature has to change to restore radiative balance. Percentage change doesn’t.
Humidity is also an exponential function of surface temperature. The reason the planetary average precipitable water is high is because there’s lots of water vapor in the Tropics where the surface temperature is high too. There is no ‘one size fits all’ profile for the planet. That’s another fallacy of the Miskolczi analysis.
Science of Doom,
Before you go any further with your Matlab model, why not get a copy of LBLRTM? It already has all the bells and whistles like the water vapor continuum and its executable is compiled rather than interpreted so it should run lots faster.
Instructions here:
Click to access LBLRTM_MINGW_install_and_compile.pdf
In Miskolczi Part 4 – Emissivity, SoD shows that Su should include the reflected part of the downward emittance Ed. The consequence is that including emissivity will decrease the Ed/Aa ratio. Figure 2 of M2007 shows that including emissivity would increase the Ed/Aa ratio when the reflected part of Ed is not considered.
In comments to Parts 2 & 3, I provided a link to an Excel spreadsheet showing the HARTCODE output data and various parameters considering emissivity, but not the reflected Ed. I have updated this spreadsheet to include the reflected Ed here:
http://members.shaw.ca/sch25/Ken/hartcode_61yearNOAA2.xls
I changed the column headings where emissivity =1 of Ed to Eda. Eda is the downward emittance, and Ed is now the absorbed portion of the Eda.
Columns T to AF show the fluxes and flux ratios with emissivity = 0.967. This is the value given on page 8 of M2010. Su and St includes the reflected portion of Eda.
The graph at cell AK10 shows Ed/Aa vs Su at emissivity = 0.967. The average value is 0.9365.
SoD has argued that the change in the TOA flux, or radiative forcing is more important that the change in optical depth. Since Su can change for reasons not related to greenhouse gases, we should see how the normalized greenhouse factor Gn = (Su – OLR)/Su has changed.
The relative humidity (RH) of the NOAA data from 1948 through 1959 appears to be too high. The relative humidity of the air immediately above the ocean surface would be very close to 1 as the air is in equilibrium with the water. This should stay constant with global warming, so the average RH near the surface should stay constant. The near surface RH for 1960 – 2008 is shown here:
The best fit shows no trend, so I use the data only from 1960.
A graph of the normalized greenhouse factor (Gn) and optical depth from 1960 (with e = 0.967) is shown at cell AT70. Below the graph, I show that the best fit of Gn has increased by 0.19%.
The Gn varies with the log of CO2 content as shown here:
I extrapolate the Gn to double CO2 to calculate the climate sensitivity. With e = 0.967 including reflected Eda, the calculated change of surface temperature at 2xCO2 is 0.26 Celsius. The change surface temperature without considering reflected Eda at 2X CO2 is 0.23 Celsius. You can exclude the reflected Eda in the graphs and calculation by changing cell Z2. Apparently including reflected Eda in the emissivity has very little effect on Gn.
This calculation does not include the changes of the minor GHG. Including the minor GHG would increase the change of Gn, but then the calculated sensitivity would not be 2X CO2.
SOD: Sorry to intrude on a scientific discussion that is way over my head, but,
you say: ‘If his calculations show:
optical thickness (CO2 + water vapor) = constant
then this article demonstrates that:
optical thickness (CO2 + other gases + water vapor) = increasing”.
Wouldn’t this support his point – that it is not the increase in CO2 that is causing GW and that therefore we have to look somewhere else to see what those other gases are and what is causing their increase?
I realize that you prefaced that statement with “if’, but isn’t that the whole point of these articles – to deny that optical thickness (CO2 + water vapor) = constant?
Why look at alternatives, when, what you are trying to do is negate that hypothesis?
DeWitt Payne says:
“In terms of W/m², doubling CO2 changes the forcing in the tropics by ~5 W/m² but only by 2 W/m² at the South Pole.”
Was the graph you linked to generated by a GCM? This is just based on a modelers guess of how water vapour will change in the middle atmosphere. Climate models assume that water vapour will increase in the middle atmosphere with warming, but all the data says it will decrease with warming.
You have to use real data, not climate model output to access climate sensitivity. Climate models generally assume the atmospheric profile follows the moist adiabatic in the tropics resulting in almost constant relative humidity. This is false as shown here:
There is a near infinite supply of greenhouse gases available to the atmosphere in the form of water vapor from the ocean to provide the greenhouse effect, but the atmosphere takes up only a small portion of the water vapour it could hold due to energy balance constraints. The specific humidity in the lower atmosphere is controlled by the saturation limit, as weather turbulence forces parcels of air to 100% relative humidity. This is the limit which causes RH to be near constant near the surface. The relative humidity at the 400 mbar level is only 35%. Air at this level never gets anywhere near the saturation limit, so there is no reason to think that RH stays constant with warming at this level. But the water vapour in limited an energy constraints.
Warming has the a drying effect in the middle atmosphere. Here are three radiosonde profiles as different temperatures taken at the same location.
The darkest line is at the highest surface temperature and highest mixing ratio at the surface. It RH drops to near zero at 8 km.
Satellite measurement of humidity also shows a drying trend at the 300 to 500 mbar layer as shown here:
increasing humidity, and the moist adiabatic profile in the upper atmosphere would lead to higher temperature trends in the tropical troposphere as predicted by GCM, but the actual temperature trends are 1/4 of that predicted as shown here:
Adding CO2 to the atmosphere replaces a large part of the equivalent amount of water vapour to maintain an almost constant greenhouse effect and has only a small effect on global temperatures. Also, Dr. R. Spencer has shown that increasing temperatures causes greater cloud cover and albedo, resulting in a large negative feedback.
Ken, thanks for your comments; you say:
“Warming has the a drying effect in the middle atmosphere”
Could you elaborate on that; is that a reference to Stuart Frank’s work or the recent work on condensation; as williamcg noted on another thread:
“The models do not handle convection and water vapor correctly. The models predict decreasing convection intensity with increasing surface temperature when the exact opposite is true. That’s why they keep looking for that non-existent “hot spot”. They do not handle equivalent potential temperature correctly, probably because they are forcing a constant humidity. They also underestimate vaporization flux by a factor of two or more. Contrary to the models, increasing surface temperature increases convection intensity which in turn increases condensation efficiency. That’s why the upper troposphere dried out during the 1976 – 1998 warming period.”
Ken Gregory,
By data do you mean the NCEP/NCAR Reanalysis ‘data’? You might be surprised to learn that most of that data is generated by model calculations. The calculations are somewhat constrained by actual observations, but you still have to take the results with reservations, as there are few observations over the 70% of the Earth’s surface covered by oceans. If humidity went down with temperature, then the tropics should be the driest place on the planet. But specific humidity doesn’t go down with temperature. τ isn’t constant.
Cohenite says:
““Warming has the a drying effect in the middle atmosphere”
Could you elaborate on that; is that a reference to Stuart Frank’s work or the recent work on condensation; as williamcg noted on another thread:”
I don’t know who williamcg is, but the drying effect is described is some detail in the first article of the “Climate Models” section of my website here:
http://www.friendsofscience.org/index.php?id=222
The article by William Grey and Barry Schwartz shows that warming creates greater evaporation and precipitation, increased albedo due to increased cloud cover, “and extra return mass flow subsidence associated with extra IR energy being emitted to space.” They say “Our observations indicate that upper-level moisture actually goes down as precipitation rates increase.” and “Saturated air from the upper tropospheric outflow of Cb clouds which sinks to levels only 100 mb below it has its RH greatly reduced by values as much as 60 to over 90 percent (Table 2).”
The second article by William Kininmonth says the GCM underestimates evaporation and precipitation by 2/3. “Empirical studies suggest that the vertical energy
transport within deep convection is inadequately specified;
the updraught mass flow must be significantly augmented
by saturated downdraughts if the required vertical mass and energy exchanges are to be achieved.”
The three links to graphs in my previous post shows the upper troposphere drying with warming.
DeWitt Payne says:
“If humidity went down with temperature, then the tropics should be the driest place on the planet.”
Warming at a specific location, or globally, reduced humidity at the upper troposphere, while increasing humidity near the surface. This does not imply that the upper troposphere over a warm region should be drier than over a cool region. I agree that humidity over the tropics is higher than elsewhere.
karl sniderman:
No, Miskolczi’s hypothesis is that total global optical thickness (tau) of the atmosphere is invariant with time.
He “demonstrates” this by calculating tau for CO2 and water vapor.
I demonstrate that other gases have significantly increased tau therefore demonstrating that even with Miskolczi’s calculations – real tau is not invariant.
This is a minor detour.
The bigger questions experimentally are why clouds weren’t included seeing as they are very optically thick (see Part One) and which is the right water vapor dataset to use (not yet covered). Along with whether total optical thickness is the best measure of the “greenhouse” effect.
The theoretical aspects to his paper are fatally flawed as demonstrated in Part Two, Part Three and Part Five
SoD: OK, so you say: “The theoretical aspects to his paper are fatally flawed”.
Now, keep in mind that I am completely naive about the science here.
Other people have brought up that Miskolczi’s hypothesis destroys the argument that CO2 is the cause of AGW. I want to be able to argue back that you have conclusively rebutted his theory.
Is that “kosher”, valid, justified? Or do I have to wait for your analysis to go through peer review?
And, by the way, did his paper go through peer review, or is this blog part of the peer review?
karl sniderman:
These are all good questions. Now for a somewhat length reply..
First of all, the only way for you to know whether my analysis is correct is to try to understand it.
I realize that it is a very technical subject and so making your own assessment might be very difficult.
The problem is, most people who claim that “Miskolczi’s hypothesis destroys the argument that CO2 is the cause of AGW” are probably equally “naive about the science”.
That doesn’t mean everyone. But some papers get large numbers of supporters because the conclusion has a happy ending.
This is probably also the case for papers supporting a more “consensus” point of view.
There is no “final arbiter” of science and you have to make a judgement based on attempting to understand the arguments put forward, how the other side responds, and so on.
It is also justified to say “I don’t know”.
If two people debate in Chinese then I can’t say who “won” (if anyone did), because I can’t speak Chinese.
Perhaps I can suggest one approach. For those claiming his theory is correct, ask if they can explain how kinetic energy can be equated with flux (emission of radiation). When you get no response you will know that they never worked through any of the equations and probably don’t actually understand it at all. (By the way, I have asked the author this question and haven’t got a response – see Part Three).
Or – ask how a theory which assumes no convection and a “grey atmosphere” (=no change in absorption properties with wavelength) can be used to explain something significant about the real atmosphere. I expect that 99% of the time, or 100% of the time you will get a confused response, or no response.
Secondly, you ask about peer-review. I am not submitting anything to a journal. A little about why a paper getting peer reviewed doesn’t guarantee its infallibility in New Theory Proves AGW Wrong!.
The fact that something is in a journal is no guarantee of anything. Perhaps the editors don’t understand the subject.
The author himself has admitted that two of the steps in his “theory” (in the 2007 paper) are not from theory but are “experimental results”. “My experiments support my experiments”. And one of these is clearly not correct anyway – see Part Four.
Many more steps can easily (and have been) criticized.
Not much to write a paper on. Who would publish this?
Herein we review Miskolczi 2007.
1. Kirchhoff’s law cannot be invoked for an atmosphere not in thermal equilibrium.
2. Equation 7 is invented (is not derived from any theoretical principles).
3. Kinetic energy cannot be equated with flux as kinetic energy is proportional to temperature while flux is proportional to the 4th power of temperature.
4. A semi-gray model with no convection cannot be used to demonstrate anything useful about the atmosphere. (The derivation is wrong anyway).
The end.
What journal would I submit it to?
You are too modest SoD; that is an adequate abstract sans the whimsical “The end”.
The fleshout has already been done by you in previous posts; go for it.
SoD: Thank you.
Ken Gregory,
Humidity in the upper troposphere is of minor importance compared to the lower troposphere. The scale height for water vapor is 2 km. That means that 86% of the water vapor in the atmosphere is below 4 km altitude and 95% of the water vapor is below 6 km, which is at ~500 mbar pressure so well short of the tropopause. Downwelling IR and absorption of upwelling IR is going to be affected most by humidity in the first kilometer or so. You’re not going to get drying there.
As SOD says, however, it is the GH effect that is important.
Drying in the upper troposphere has the effect of lowering the level at which emissions from water vapor escape to space. And generally within the troposphere this means at a higher temperature, and thus, more energetically.
The question remains open about exactly what the humidity trends have been.
The sonde data do indicate a large decrease in humidity, but of course some, if not more than all of that trend is from changes in humidity sensors with a tendency toward faster response ( reduced 1/e times ). Since humidity generally decreases with height this induces a spurious signal. Still, the trends indicate a continuing decreasing trend, even through the 2000s when presumably humidity sensors have converged.
Satellite approximations lack vertical resolution and have some contradictory data from satellite to satellite.
But it is the change in OLR that is the important measure, not the total precipitable water.
Ken,
Ken said:
I am William C. Gilbert, the owner of the paper here:
http://www.friendsofscience.org/assets/documents/Gilbert-Thermodyn%20surf%20temp%20&%20water%20vapour.pdf
You linked to Figure 8 of this paper in your 10:32 pm post which shows RH radiosonde plots for three different surface humidity levels.
My comments to Cohenite on the other thread were based on my own work plus the Kininmonth paper you linked to and the as yet unpublished paper by Noor van Andel CO2 and Climate. In the van Andel paper, this is covered in detail on pages 14 – 15 where he also says:
“The main error in the climate models is that they suppose heating and moistening, and thus higher θe, of the upper troposphere by CO2, in contradiction with radiosonde and satellite measurements. This assumed heating & moistening leads the model to assume an increase of θe at this height, which makes deep convection decrease as a result of increasing SST, very unphysical as we have seen here above”.
The Gray and Schwartz paper you linked to adds another dimension to the science since he also measured OLR and empirically demonstrated the negative feedback nature of the lowering upper tropospheric water vapor concentration. This should no longer be contested, but unfortunately it still is.
As to these articles by SOD concerning the Miskolczi papers, they have been very interesting and technically well done. Unfortunately he has missed the bigger point of the whole exercise. Miskolczi has empirically established some very key radiative relationships of the atmospheric system. I have heard nothing to disprove the empirical evidence. Miskolczi has then done a good job of trying to explain the empirical results based on theory. While SOD and others may quibble about some details of Miskolczi’s theory, they have not been able to explain the empirical results with their textbook radiative theories, either – not even close. They try to falsify the theory but they can’t falsify the empirical data. This just leads to falsification of their own theories.
SOD is under the assumption that the measured radiative properties of the atmosphere can be explained using only radiative processes. This is his major error. The radiative properties of the atmosphere can only be explained via both radiative and non-radiative processes considered together. All of the independent variables of this very complex system interact with one another and concepts such as “forcings”, where certain variables are held constant, cannot work – especially when you have completely left out the most important non-radiative variables in the analysis altogether. I believe Miskolczi understands this, SOD does not.
Bill Gilbert
Sorry, I messed up the links. I’ll try again.
Here is my E&E paper:
http://www.friendsofscience.org/assets/documents/Gilbert-Thermodyn%20surf%20temp%20&%20water%20vapour.pdf
Here is van Andel’s paper:
http://xa.yimg.com/kq/groups/4401572/269546772/name/CO2_and_climate_v8.pdf
Bill Gilbert
The empirical radiative flux relationships of Miskolczi have not been shown to be applicable at all optical depths. In fact, we know that they cannot be applicable over a wide range of optical depths, equation (7) for example. Let’s look at the equations found on page 13 of the EGU presentation:
Aa ≅ Ed is a reasonable approximation, but by itself proves nothing.
Su = OLR/f is wrong, as it is a conclusion of an incorrect solution of the the Schwarzschild-Milne equation.
Su = 2Eu is only true if Su = OLR/f and even then only approximately true as Eu/Su goes to zero as τ → 0 and Eu/Su → 0 as τ → ∞
The same applies to Su = 3OLR/2
The conclusion that using these relationships to prove that τ is constant is backwards. The actual conclusion is that the empirical relationships are only true at τ ≅ 1.8. They do not constrain the value of τ in the real atmosphere.
Also πBg = Su + K, not Su.
The variable that maximizes entropy production is probably K. But you cannot calculate K from Schwarzschild-Milne alone.
williamcg:
Miskolczi has presented a paper which has the appearance of theoretical proof of experimental results.
Yet it is not a theoretical proof of experimental results. If I have demonstrated that successfully then I achieved my first objective.
If Ferenc Miskolczi would like to revise his claims to simply that water vapor has declined against model predictions that will be a very interesting discussion. Unfortunately, it still isn’t clear exactly what is claimed to be “experimental work” and what is “theory”.
On experimental work, so far I have demonstrated that two of his specific claims are not true.
First, optical thickness is not constant:
a) clouds are ignored yet overwhelm any calculations of optical thickness from clear skies.
b) as demonstrated in this article, minor gases also have increased the optical thickness of the atmosphere yet were neglected when Miskolczi concluded experimentally that optical thickness was constant from 1948 -2008 (actually he concluded “unchanged over the time period”, because it’s definitely not constant in his graph).
On a) the author has made no comment. We can all understand why.
On b) – written since the author departed – he has yet to comment and perhaps this might take some time to review.
Second, Ed≠Aa:
a) Ken Gregory has helpfully confirmed that with a real world surface emissivity Ed is nothing like Aa (instead of Ed=Aa, Ed=0.94 Aa). Perhaps this is irrelevant, or perhaps the original error is propagated through the experimental calculations which are all incorrectly calculated.
And as a fairly easy-to spot-error, how many other experimental values suffer similar problems? Note that almost all of his values are derived (calculated) from experimental results.
As yet I have made no comment on the long term water vapor trends.
My comment at this stage would be that if Miskolczi’s work is simply an experimental claim in declining water vapor then at the minimum there should be some discussion of why the particular data set was chosen against others. What results follow if other datasets are used? and so on.
As you will see when I post an article about long term water vapor trends most of the writers of papers on this subject devote a lot of time to comparing different datasets and evaluating one against the other.
As always, it is easy to make claims.
Now is your opportunity to explain specifically what in any of these six articles is flawed.
williamcg:
And as I reread your comment I wonder if you understand what is really in Miskolczi’s papers:
Miskolczi’s values are calculated via radiative relationships. Some of which have been demonstrated to be flawed.
What empirical data in Miskolczi’s papers is there to falsify?
Tau is calculated
OLR is calculated
St is calculated
Su is calculated
Ed is calculated
Aa is calculated..
SOD:
Calculated from what? ……..Empirical data!
Bill
What empirical data is it that we are trying to falsify? (Your claim)
What empirical data has Miskolczi presented that we are “trying to explain with our textbook radiative relationships” and yet “are not even close“?
Claiming is one thing.
Demonstrating is another.
Even being specific would be a start.
I made a model of the atmosphere and I analysedMiskolczi’s variables
Click to access IRabsW27102011.pdf
JWR,
Why do you think the GHE or K&T’s diagram violates the 2nd Law? I’ve seen this claimed in many places, but I don’t get the objection because it’s not about heat going from the cold to warm through a conduction process.
Do you actually think that an emitted photon cannot travel from the colder atmosphere toward the warmer surface?
williamcg,
The problem is that Miskolczi’s ’empirical’ data proves nothing by itself. It’s only when he elevates approximate relationships to identities and then plugs them into his simple radiative transfer theory that we get ridiculous conclusions like τ is a constant equal to 1.86756 so increasing CO2 will have no significant effect on the average temperature because water vapor will magically go away as CO2 increases. Without that conclusion, nobody would have paid more than the slightest attention to the papers.
For example, Equation (12) in M2010.
(3 + 2exp(-τA))/5 = 2/(1+τA+exp(-τA))
What’s empirical about that equation? And please don’t refer me to Figure 7 in the same paper for the right hand side of the equation. The agreement with the TIGR2 data is not very good. Or Figure 8 which purports to prove equation (10) Ed/Eu =5/3, which is required for the left hand side of equation (12).
DeWitt and SoD, your considered opinion about Miskolczi applies only to M2007 and M2010? Or do you intend that it applies to M2004 as well:
Click to access IDOJARAS_vol108_No4_01.pdf
Absolutely. M2004 uses the same flawed equations as M2007.
I can play the empirical relationship game too.
Suppose we assume that Bg = OLR*(2+τ)/2 as would be true for the classic solution of the gray atmosphere model. Then let Su = Bg – K. If I use the Su, τ and OLR calculated from NCEP Reanalysis in Ken Gregory’s spreadsheet, I can calculate a value of K for each year to make the data fit. It turns out that K ≅ 101 W/m² with a standard deviation of ~1 W/m². TF&K09 has K = 97 W/m² so that looks pretty good. If I plug the average value in, there is no fit. But if I look at K vs OLR, there’s a linear relationship with K = 0.7885*OLR – 101.21 with R² = 0.53. So K increases with OLR, but not as fast. That seems reasonable too. I suspect that if I threw in more variables like τ, the fit would be even better.
Yes, clouds are better greenhouse forcers than any of the gasses.
But if you’re looking for trends, quality data doesn’t seem to be available.
I recall two of the satellite assessments coming up with different signs in the trends, especially for the high level clouds which are the most important to longwave forcing.
Because there is directional scattering from clouds, dependent on type, orientation, and microphysical properties, clouds and the albedo they change, remain elusive measurements.
I am on board with the assessment that Tau is not the same as greenhouse effect (what energy actually leaves earth ) which is the important measure.
CO2 is well mixed in the homosphere, but imagine it wasn’t. Imagine all the CO2 in the atmosphere was magically constrained to the layer right at the tropopause. The average CO2 emissive temperature would decrease.
Now imagine all the CO2 was magically constrained to the lowest 100 meters. The emissive temperature would increase, and so the energy lost to space would increase.
In reality, CO2 will remain well mixed, but H2O will remain constrained ( dropping off dramatically from the surface upward, and also being subject to transport horizontally and vertically. )
Similarly, total cloudiness could remain constant but the proportion of high clouds could change, which would alter longwave out.
Here’s another interesting tidbit from the Ken Gregory spreadsheet: albedo. If we assume that radiative steady state exists, then OLR = Fo = (1 – α)*TSI where OLR is Outgoing Longwave Radiation, Fo is total solar radiation absorbed, α is the albedo, the fraction of solar radiation reflected and not absorbed and TSI is the solar radiative flux at the top of the atmosphere. Taking TSI = 341.3 W/m² from TF&K09 and OLR by year from 1948-2008 from Ken Gregory’s spreadsheet, one finds that α = 0.249 +/- 0.005.
Now there’s a number neither fish nor fowl. Global albedo is ~0.3 and clear sky albedo is going to be close to surface reflectivity or 0.125 in the absence of significant atmospheric aerosols other than clouds. That means that calculated OLR is either too high by 17 W/m² if it’s actually supposed to be global OLR or it’s too low by about 40 W/m² if it’s supposed to be clear sky. In other words, it’s wrong. If OLR is wrong, everything else is wrong too.
Another tidbit:
If you plot OLR vs time for the 61 year period in the Ken Gregory spreadsheet, you get a linear increase with a slope of 0.0482 W/m²/year (R² = 0.9)or an increase of 2.9 W/m² over the 60 year period. That’s a direct result of the decrease in albedo of over 3.4% (-0.000141/year) over the same time period. I don’t think so.
Ken Gregory,
I get 1.44 prcm for the US76 clear sky profile in Archer MODTRAN, but that’s not important. What’s important is that only 40% of the sky is clear. I get 2.28 for the cumulus cloud, base 0.66 km top 2.7 km with the US 1976 standard atmosphere. That averages to 1.94, which still isn’t 2.5, but is a lot closer without the unrealistically high surface relative humidity required by a clear sky with a surface temperature of 289 K to get total precipitable water of 2.6 cm (90+%), not to mention the intermediate value of OLR that’s produced.
You can’t feed some global average profile into a radiative transfer program and expect it to produce a result equivalent to a partly cloudy planet. You also can’t feed in radiosonde profiles taken in cloudy conditions and assume the humidity measured in the cloud layer is good enough to get the correct results. Water droplets have vastly different radiative properties than water vapor.
And we return once again to a familiar theme: Miskolczi’s analysis ignores clouds.
DeWitt, in M2004, they specifically look at radiative emittances in the context of the atmospheric water column. They also do a specific comparison between clear-sky and all-sky OLR based on ERBE measurements.
There is no EQN 7 equivalent in M2004. Are you really saying there is no worth in M2004?
cohenite:
On my side I haven’t reviewed M2004 again. I did read through it when I was working through M2007 & M2010.
If you have a specific question I might be able to summon up the enthusiasm to consider it. I have invested a lot of time in the first two papers already so don’t really have the energy for a general review of this one.
Water vapor trends await. Lots of papers..
SoD, not so much questions as a couple of observations, so just consider but don’t deviate from the water which is obviously the key. Much of the criticism levied against M is his focus on clear-sky conditions. In 2004, figure 4 he compares his findings with ERBE such that:
“The total OLR curve from the TIGR data set is between the clear-sky and all-sky OLR curves from the ERBE, and the shape of the curves are very similar.”
I had been rereading some of the old discussions and came across this from Steve Short who is probably as disenchanted with M as you and DeWitt:
“The K&T97 and TF&K08 reviews both imply that LW IR emitted by clouds is ~30 W/m^2. This means that both the CERES and ERBE period best estimates of all sky S_T should be about 31 W/m^2 respectively and that the clear sky (cloud free) estimates of S_T again for both these periods should not exceed about 61±10 W/m^2 i.e there is about a less than one chance in 40 (2.5%) (assuming binomial distribution) of a clear sky S_T exceeding 81 W/m^2. [It also means that both the CERES and ERBE period best estimates of S_U should be about 395 W/m^2. Thus the best estimate of clear sky mean tau should be about 1.87+0.18,-0.15 but also indicating a mean global all sky tau is hardly likely to be as low as 1.87]”
Steve concludes:
“Miskolczi HARTCODE measurements of clear sky S_T are thus probably consistent with mainstream science values.”
So, I guess there are 3 observations:
1 Are clear sky findings of no value given Figure 4 from M2004
2 Is M’s steady state tau of any value if just for clear-sky conditions
3 M2004’s finding “that on a global scale, the far infrared contribution to the clear-sky normalized greenhouse factor is
significantly increasing toward the polar regions” would seem to be of at least some modest interest given the bell-whether status of the poles in the AGW lexicon.
cohenite,
That’s the problem in a nutshell. The total OLR curve from the TIGR data set is not representative of the planet as a whole. The albedo calculation proves it. Yet he goes on to write two more papers as if it were representative and proves his (fatally flawed) theory. It isn’t, and it doesn’t.
As William F. Buckley once said about Ayn Rand: Where Miskolczi is correct, he’s not original. Where he’s original, he’s wrong.
cohenite,
I disagree with Steve Short’s analysis of clear sky S_T. Only in the Tropics is clear sky S_T as low as 61 W/m². Everywhere else, it’s higher. In fact, K&T97 and TF&K09 use something close to US76 for clear sky emission and get S_T ≅ 100 W/m². I think that’s too high as US76 does have very low humidity. But the 30 W/m² for S_T from cloud tops is a more interesting number. Cloud tops are colder than the surface, but they’re also well above the surface. Most absorption of IR is from water vapor and well over half the water vapor resides in the first 2 km above the surface.
For example:
MODTRAN tropical atmosphere with low cumulus cloud cover, base 0.67 km and top 2.7 km. The surface temperature is 299.7 but the cloud top temperature is 280.5 K. Transmittance up from the cloud top is 0.3636 for the range 100-1500 cm-1, which makes S_T = 121.5 W/m². Higher clouds will be colder and emit less, but the transmittance will be greater too. The only way I can come up with something like 30 W/m2 directly to space from cloud tops is to subtract the difference in flux between the surface and the cloud top due to the temperature difference, (121.5 – (448.29-351.01)) = 24 W/m² for the Tropics. Maybe that’s it. That still makes global S_T = 70 W/m². Which makes global transmittance = 70/396 = 0.1768 for a value of τ = 1.73.
DeWitt, this issue of the emissions from the tops of clouds was discussed by Steve in the thread I referred to; rather than continue to post slabs of the conversation the link is to David Stockwell’s site and the relevant comments begin 21 from the bottom:
http://landshape.org/enm/the-value-of-tau/#disqus_thread
I was about to bomb you with a copy/paste rant, but you are actually posting data. Good for you.
In this otherwise entertaining & informative thread, there is an issue which I have not seen addressed. My understanding is that SoD (with many others) treats the atmosphere as a radiatively-driven assemblage, where Miskolczi treats it as a radiatively-coupled system. While I find M2007 obscure in parts, falsifying M’s detail doesn’t eliminate his general model. It wouldn’t be the first time that a scientist got something (approx) right for the wrong reasons.
Is anyone else taking this coupled-system approach? If not, I have a few ideas for comment.
Toth (2010) shows that the Virial Theorem can apply to an atmosphere, provided the potential energy (PE) is gravitational and associated only with the parallel (vertical) degree of freedom (DF). He restricts his finding to diatomic gases, but I suggest it extends to any mixture of diatomic and rigid linear molecules at a temperature too low to excite vibrational modes (so N2, O2, CO2, N2O). Note that KEz = PE/2, so each of horizontal-translation and rotation modes is associated with PE as each has 2 DF. He associates PE with troposphere depth.
For simplicity, assume a single GHG with a single saturated stop-band. On a radiance-wavenumber graph, pencil in three BB curves – middle, the bare Earth (controlled by the Solar constant and Earth’s albedo); upper, the “radiatively-forced” Earth; lower, the TOA. Mark the stop-band. For steady-state, the area between middle & upper outside the stop-band = area between middle & lower in the band. Aa & Ed differ considerably. The ratio of in-band OLR to Ed should be a fairly smooth function of their temperature ratio. The coupled system is “pumped” only by in-band Su-Ed. If this is large, the system PE rises, driving both of Ed & OLR up and reducing in-band Su-Ed; ie a negative feedback loop. The system thus has a credible mechanism to maintain itself.
Actual Su-Ed around this blog appears ~10-20 W/m2. The total non-LWR input (SW to atmosphere, latent + sensible heat) is 169 (K&T1997 Fig 7). This would reduce but not invalidate the feedback, and offset the original steady-state condition, allowing temp at TOA to be higher than bare-Earth. Ed approaches Su. Note that in a coupled system, point of input is immaterial.
Adding water (bent molecule, condensable) is beyond my expertise.
Can anyone falsify this so far? If not, has anyone examined actual TOA/BOA temperature ratios over their typical ranges, for comparison with OLR/Ed ratio (both in- & out-of-bands)? Does M2007 imply a prediction of the environmental lapse rate of -6½K/km?
btw, can I suggest that the basic model Earth is a water planet? This would place evaporation centrally in surface processes, where I think it should be.
The Virial Theorem does not imply a lapse rate at all. It is true for any lapse rate or combination of lapse rates. What would change would be the pressure as a function of altitude. An isothermal atmosphere, for example, would have a higher altitude for a given pressure less than the surface pressure for the same surface temperature than for an atmosphere where the temperature declines with altitude.
Miskolczi is still wrong.
RipVanWinkel,
I don’t understand your terminology.
Heat transfer between the atmosphere and the surface is by convection and radiation. Heat transfer between the surface/atmosphere and space is by radiation.
So in broad brush, the atmosphere is in radiative-convective equilibrium with the surface. (More specifically, it is never in equilibrium).
Miskolczi’s model treats the atmosphere as one where convection doesn’t exist. An atmosphere in radiative equilibrium.
His model is based on detail. Your statement just means you haven’t understood his model.
To demonstrate this I challenge you to describe his model, its premise(s), and how it could be falsified.
His model is “empirical evidence” disguised as a theoretical paper.
The empirical evidence is not clear. It ignores the effect of clouds on optical thickness for example. This is important because clouds cover 62% of the sky and have a very high optical thickness. It ignores the effect of other GHGs.
Apologies for any inadvertent obscurity. I had in mind something simple that the general reader might take from this blog even if they don’t follow the detail.
From Toth:
> The virial theorem applies to the troposphere (given care with the DF.s);
> The tropospheric system’s energy is expressed in its height (or depth).
If there is a simple argument requiring that increased CO2 gives higher tropospheric energy = deeper troposphere (with riders that neither effective TOA temperature nor lapse-rate can decrease), surface AGW is proven. SoD’s detail then addresses the “how much?” following question.
Is this do-able?
If this works, it should apply whatever happens to cloud effects.
Pace, DeWitt Payne; if the virial theorem applies to the troposphere, it’s a fairly safe bet that other constraints in the system dictate its average lapse-rate.