How much will global temperature rise if we double CO2 from pre-industrial levels? Based on current behaviour that’s roughly what we are on course to do by the end of the century (see 3 – How much CO2 will there be? And Activists in Disguise and 3.5 – Follow up to “How much CO2 will there be?”).
For this we need a model. But to begin with we can use a much simpler model than a current GCM (global climate model).
The key is to say “all other things remaining equal”. So we double CO2 but assume (in our model) that the vertical temperature structure of the atmosphere doesn’t change, clouds don’t change, water vapor doesn’t change, etc.
This allows us to calculate how the radiation balance gets disturbed. Then we can find the new surface temperature which brings everything back into balance. We don’t need a GCM that attempts to model turbulent flows of the atmosphere and ocean.
It turns out that the global temperature change will be about 1.2ºC from pre-industrial levels.
Well, kind of.
This is with the absolute amount of water vapor staying the same. Water vapor is the strongest “greenhouse” gas in the atmosphere and the amount of water vapor is one key to understanding future climate change.
If you stand next to the ocean in the tropics, barring a strong wind coming from inland, it’s pretty humid. If you stand next to the ocean in the Arctic it’s pretty dry. The reason is that the amount of water vapor that the air can hold depends strongly on temperature. Next to the ocean the air can be close to 100% relative humidity, but 100% relative humidity in the tropics has lots more water vapor than 100% in the Arctic.
So a slightly different simulation has relative humidity staying constant. The result is some amplification from the water vapor. The Earth’s surface gets a little hotter, so there’s more water vapor – which is also a greenhouse gas – so the surface gets hotter still.
In this experiment the global temperature change will be about 2.4ºC from pre-industrial levels. This second experiment is intuitively a better experiment than the first one – at least to get a “finger in the air” kind of result. It doesn’t mean it’s correct, but most people working in climate would expect relative humidity to be more likely to be constant than absolute humidity, when you increase the temperature.
So, our “no feedback” result is 1.2ºC, and our slightly more realistic “some feedback” result is 2.4ºC (currently the global temperature has increased about 0.8ºC from pre-industrial levels).
Both of these results can be obtained without relying on models that have “giant fudge factors” which is what you need to model the atmosphere and ocean “fluid flows” – see 6 – Climate Models, Consensus Myths and Fudge Factors. They only rely on being able to accurately calculate how radiation is absorbed and emitted by the atmosphere – an extremely well understood physics problem. (I can reproduce the results on my home computer using Matlab and the spectroscopic properties of CO2 and water vapor – see Visualizing Atmospheric Radiation).
The real story, of course, is more complicated. However, to understand anthropogenic global warming (AGW) – a better name than “climate change” – it’s useful to know these results (first calculated in the late 1960s by Manabe & Wetherald) – and to understand the difference between simple radiation models and global climate models.
Articles in this Series
Opinions and Perspectives – 1 – The Consensus
Opinions and Perspectives – 2 – There is More than One Proposition in Climate Science
Opinions and Perspectives – 3 – How much CO2 will there be? And Activists in Disguise
Opinions and Perspectives – 3.5 – Follow up to “How much CO2 will there be?”
Opinions and Perspectives – 4 – Climate Models and Contrarian Myths
Opinions and Perspectives – 5 – Climate Models and Consensus Myths
Opinions and Perspectives – 6 – Climate Models, Consensus Myths and Fudge Factors
The Science of Doom 
Question: How much will global temperature rise if we double CO2 from pre-industrial levels?
Honest answer: We don’t know
https://tambonthongchai.com/2018/06/25/a-greenhouse-effect-of-atmospheric-co2/
https://tambonthongchai.com/2018/09/25/a-test-for-ecs-climate-sensitivity-in-observational-data/
https://tambonthongchai.com/2018/11/26/ecsparody/
Eh, we partly know; we have reasonable estimates with error bars.
Very few things in science are ever completely certain. It’s not usually a dichotomy of “we know”, or “we don’t know”. The question is generally how large the error bars are — how well we know.
Water vapor doubles the other feedbacks as well. So add in a little snow/ice melt perhaps some clouds and 3C is very doable in the simplest of models
They only rely on being able to accurately calculate how radiation is absorbed and emitted by the atmosphere – an extremely well understood physics problem. (I can reproduce the results on my home computer using Matlab and the spectroscopic properties of CO2 and water vapor . . ..
SoD, isn’t this a correct characterization only for the case of no radiatively interacting aerosols in the atmosphere?
Dan: The full physics of radiation interacting with matter (radiation transfer) includes the possibility that radiation is scattered and that the distribution of energy in matter may not be determined by a Boltzmann distribution of energy among quantum states.
The second phenomena is important in lasers and LEDs, but not in the atmosphere below about 100 km, where collisions transfer energy much faster than other phenomena. This is called local thermodynamic equilibrium, but doesn’t imply radiation must in equilibrium with matter. (Radiation in equilibrium with matter obeys Planck’s Law.)
SOD is mostly referring to calculating LWR transfer through “clear skies”. Rayleigh scattering by molecules makes the clear sky appear blue, but is only significant for SWR, not the LWR which carries heat to space and thereby cools the planet
Reflection and absorption by larger particles (aerosols) and clouds are important for both SWR and LWR. For calculating LWR radiative cooling to space, we can ignore what happens below clouds and begin with the radiative flux emitted by cloud tops (which have an emissivity near 1).
For more information, see an article I wrote for Wikipedia (and am still trying to improve) mostly based on things I learned first here from SOD. Comments and corrections would be appreciated.
https://en.wikipedia.org/wiki/Schwarzschild%27s_equation_for_radiative_transfer
SoD,
“The key is to say “all other things remaining equal”. So we double CO2 but assume (in our model) that the vertical temperature structure of the atmosphere doesn’t change, clouds don’t change, water vapor doesn’t change, etc.”
This is what is NOT correct, because GHE theory specifically states that the vertical temperature structure will change, i.e. specifically the upper atmosphere (the stratosphere) will cool and the lower atmosphere (the troposphere) will warm.
Warming the atmosphere incrementally according to the prior existing lapse rate warms both the troposphere and the stratosphere, i.e. the GHE theory signature of surface/atmosphere warming is missing. This means the 1.2C is NOT giving you the theoretical intrinsic measure of surface warming (of the 4 W/m^2 from 2xCO2) via the GHE; and this is effectively how the value is being applied to everything (whether the field knows it or not).
This itself doesn’t directly falsify the model results (since the models don’t use that value for anything), but it does falsify every single paper in the literature that is using the 1.2C (or the 1.1C from 3.7 W/m^2) value as the anchor point for sensitivity analysis from observations (i.e. mostly from satellite data).
For new readers, I’ve answered 100s of RW’s 1000s of comments on this blog in the past. It has been a pointless exercise.
Actually I thought RW had reached the select group of people who were banned, must check the records..
You of course don’t have to respond. I’m just pointing out a glaring inconsistency in all of this that apparently no one has ever noticed, and how specious, extremely loose, and unreliable it makes it all.
I chose to accept the validity of the GHE theory — not the 1.2C of supposed intrinsic measure of its effect for 2xCO2. Some pretty basic logic dictates that they can’t both be correct (the GHE and the 1.2C) is my point.
BTW, this part: ““The key is to say “all other things remaining equal”” is fine. It’s the latter part that I’m claiming is in error.
BTW, For Chubbs and SOD and RW, Chubbs was discussing the RSS trends for TLT and I dug into their paper on their new version. It states the following:
” we summarize the long-term global trends in the radiosonde and sampled satellite datasets. The trend in RSS V4.0 is larger than any of the radiosonde datasets, while the trends in UAH V6.0 and RSS V3.3 are less than any of the radiosonde datasets. The older UAH V5.6 dataset shows the best agreement overall.”
Later they also state:
“Given the physical basis for the expected trend ratio of about 6.2% K−1, the results suggest that most, if not all, of these datasets contain residual errors. If we accept our crude estimate of the uncertainty in the vapor trends, then the long-term warming trends since 1988 may be too low in all four TLT datasets, with both UAH datasets being much too low.”
My take on this is that the very indirect water vapor analysis that assumes BTW that GCM’s are right in the tropics about the vertical structure of the atmosphere seems to be contradicted by much more direct evidence from direct measurements by weather balloons. Why would they even bother with the latter line of evidence?
Your concerns regarding the 6.2% value are misplaced. The units are % water vapor per degree of warming, so climate models being off on the amount of warming wouldn’t necessarily skew the value. Also water vapor is concentrated in the lower layers of the atmosphere so water vapor at “hot spot” levels is a small fraction of TPW.
The comparison seems natural to me, since water vapor and temperature are so closely related, though I would have liked to see some error analysis and a discussion of the fact that TPW and TLT may weight different layers differently, which could contribute to the discrepancy.
Note that less warming in the mid-troposphere implies higher ECS, since a steeper lapse rate (greater cooling with height) increases greenhouse warming. Perhaps someone will use model “hot spot” temperature as an emerging constraint.
There is a problem here Chubbs and that is that UAH also did an analysis of water vapor which is far more detailed and it showed that water vapor actually increased in the tropics much more slowly than expected from GCM’s.
You didn’t however respond to my point that RSS shows more warming than 3 radiosonde datasets which agree better with UAH. This makes it unlikely on its face that RSS is still “too cool.” Indirect evidence via water vapor is in my view much weaker than actual thermometer evidence.
I doubt your point about the lack of a hot spot implying higher ECS. Bear in mind that water vapor is far more potent as a GHG than CO2. Lower temperatures in the troposphere imply less water vapor.
dpy –
Need a reference for the uah analysis. The reference below indicates that mid-troposphere water vapor is increasing as expected:
https://www.pnas.org/content/111/32/11636.short
In-any-case OHC, surface temperature and TPW are all increasing rapidly as predicted by models. These are robust measurements. Your need to consider all the data if you want to construct a convincing narrative.
Finally, here is reference which indicates that the time evolution of static stability eventually leads to higher ECS.
https://www.pnas.org/content/114/50/13126
http://www.drroyspencer.com/2015/08/new-evidence-regarding-tropical-water-vapor-feedback-lindzens-iris-effect-and-the-missing-hotspot/
Chubby, Both of your references are basically about using climate models to analyze the troposphere. I don’t give a lot of credence to this type of study given all the issues with the models in the tropics given all the issues discussed by SOD and in the comments.
Sorry Chubbs for the wrong name. Spell checker doesn’t like Chubbs.
dpy –
Likewise I don’t give much credence to TLT-based arguments: 1) Its is a very uncertain measurement that includes a portion of the stratosphere, 2) Per my comment above TLT<models doesn't imply low ECS, and 3) TLT has been used unsuccessfully to argue for low-sensitivity/poor model performance for 25 years.
The goal should be to reconcile TLT with other evidence. The first step is to understand what is going on in the chart below:
http://woodfortrees.org/plot/uah6-land/mean:12/plot/crutem4vgl/last:480/mean:12/offset:-0.3/plot/rss-land/mean:12
Something about altitudes, perhaps.Aboy UAH: “The level nearest to the surface of the Earth is the lower troposphere. The lower troposphere temperature composite include the altitudes of zero to about 12,500 meters, but are most heavily weighted to the altitudes of less than 3000 meters.” RSS is up to 7000 m.
DPY –
I have had a chance to look at the uah blog article. A google search indicates that AMSU – B 83.3 GHz water vapor channel needs to be filtered for precipitation since cold clouds (cirrus) bias the retrieval. The UAH results show such a large drying effect that it looks completely unrealistic, indicating that the precipitation filtering has been skipped. Surely someone else would have noticed such a large drying effect in the tropics.
Notice in SOD’s chart below that there is a large water vapor maximum on the equator. Roy’s climate model matches that pattern well, with a large maximum on the equator and dryness in the subtropics. Meanwhile his satellite observations have a very muted latitude structure that doesn’t match SOD’s chart. This is not surprising since thick cold cirrus will be collocated with the deep equatorial moisture, causing the biggest bias near the equator.
It pays to be a little skeptical of blog science claims.
dpy6629 and Chubbs: As best I can tell, orbital drift has prevented satellites from measuring the temperature (via the microwave flux they receive) at various location at the same time every day. By discussing radiosonde data and water vapor, it seems as if RSS and UAH are not addressing this fundamental problem by introducing fudge factors. Is this an appropriate characterization?
IMO, given all of the technical challenges in radiosondes and homogenization of data, radiosonde trends are much less certain than trends in Ts over land and probably even the ocean. Re-analysis based on all available data (including precipitation) makes far more sensible than giving both groups the opportunity to cherry-pick data that agrees with their biases. Comments?
Frank, I agree that the reanalysis approach is the best way to promote good science because improving the reanalysis should result in improved weather and short-range climate forecasting. I’ve been tracking reanalysis output for surface temperature for several years now and both CFSV2 and ERAI adjusted seem to perform very well (see here). I have not examined the upper air output but I expect it should result in a good symbiosis of all available inputs.
Reanalysis as I understand it is used for initial conditions for weather modeling. I believe it must use radiosonde data however.
Satellite data has problems as you describe. Before their recent version change RSS agreed very well with UAH, which suggests the output was reliable. RSS’s rate of warming increased a lot with the new version. One has to wonder if they were trying to get closer to surface data sets. They are now a lot warmer than hadcrut for example.
Radiosondes are the only direct measurements we have for the troposphere and I would trust them. There are 3 datasets and averaging them is probably not a bad ide.
oz4caster, That’s an interesting blog post. I didn’t realize how much the warming rate varied with lattitude.
Frank:
A couple of comments. First water vapor: The RSS TPW measurement has been operational for a long time and there is good data from different satellites for inter-comparison. There is a published paper describing the methods, data and error at the link below. Note that the measurement only works over the oceans. I covered the UAH analysis above.
Second TLT/TMT: – RSS is in much better agreement with other datasets – both surface temperature and TPW. Agreement with other data is an advantage to me. Most of the difference comes down to NOAA-14, UAH discards it, RSS keeps it.
https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2018EA000363
I’ve been thinking about how CO2 should effect temperature and it seems that wherever there are clouds the CO2 effect is eliminated or at least diminished depending on the thickness of the cloud cover. Also it seems that the CO2 effect should be diminished in places where skies are clear, but the column water vapor is relatively high as can happen mainly at tropical latitudes, but also at subtropical latitudes over the oceans. That leaves mainly deserts and polar regions when cloud cover is absent as the places where the effects of CO2 should be strongest. I’m guessing this is not a large portion of the globe. Is this lack of uniform distribution of impact taken into account when estimating a global average 1.2C impact of CO2 with all else remaining the same?
Clouds are accounted for in the calculation.
See step 4 here:
http://www.realclimate.org/index.php/archives/2007/08/the-co2-problem-in-6-easy-steps/
You can download the paper (Myhre, 1998) here:
https://www.researchgate.net/publication/238498266_New_estimtes_of_raditive_forcing_due_to_well_mixed_greenhouse_gases/download
See table one and two for clear and cloudy radiative forcing for various GHGs.
Your institution that it is higher under clear sky conditions seems correct.
verytallguy, thanks for the info. I will check the links.
verytallguy, in reading your “step 4” link, which mentions biomass burning, I’m thinking it’s likely that biomass burning will increase with increased CO2 because of CO2’s effect at producing more biomass with higher concentrations. Also, they did not directly mention dust or volcanic ash, which are absorbers and reflectors of radiation. Dust could be decreasing as CO2 rises because of greening and shrinking of deserts caused by increased vegetative growth plus increased water vapor from any-cause warming.
As an aside, dust appears likely to have been a very important factor during the glacial maximums in the last million years, as evidenced by huge quantities showing up in polar ice cores as compared to today. During the glacial maximums, cold temperatures reduce atmospheric water vapor in general and in conjunction with much lower CO2 levels and associated decreases in vegetative mass caused massive desertification in the mid latitudes. Also, much stronger temperature gradients between tropics and poles would have driven much stronger winds on average, especially in the mid latitudes, which in turn would help to drive much more frequent, intense, and extensive dust storms than today. The heavy dust loading may have contributed to cooling initially, but ironically might be the primary reason for the end of each glacial maximum because of heavy deposition on sub-polar glaciers contributing to melt with increasing high latitude solar radiation with increasing obliquity. So I guess my point is that atmospheric dust needs to be considered, especially for long-term climate out to thousands or tens of thousands of years, but may play a small role over decades and centuries.
verytallguy, in reading the second link you provided, paper (Myhre, 1998), I did not find any discussion or reference concerning the cloud cover assumptions used to derive the “cloudy sky” radiative forcing. I can imagine that a one size fits all approach to cloud cover effects on radiative forcing might come with a very high uncertainty and a large range depending on actual cloud cover statistics that may vary over time.
Ozcaster,
I think you’ll need to go through the references for the detail you want. The Myhre paper does allude to the differences in cloud handling in the different models used.
oz4caster,
I understand the 3.7 W/m^2 for 2xCO2 is the net increase in the sense that it accounts for cloud and/or water vapor overlap that would be absorbed anyway. The problem is the satellite data I think doesn’t distinguish entirely from cloudy, partly cloudy, and clear sky in the way you might think. They’re mixed together, which means the cloudy sky portion of the data contains areas that are both cloudy and partly cloudy, i.e. it include areas that are effectively clear sky. This is why you will still see some transmittance from the surface into space even through what they’re designating as ‘cloudy sky’. In reality though, anywhere there are clouds, the direct surface transmittance is pretty much zero since clouds are broadband absorbers in the IR.
The real problem with the 1.1C estimate of intrinsic surface warming effect from the 3.7 W/m^2 (from 2xCO2), as I mentioned above, is the clearly missing GHE warming signature in it.
RW wrote
“Yes, but it’s the 1.2C claimed measure of theoretical intrinsic effect via the GHE which isn’t supported by the theory (and which is my whole point). It’s an amount that will restore balance — yes, but it’s arbitrarily used more or less because the field doesn’t have a valid model or a way to quantify the theoretically supportable intrinsic effect from 2xCO2.”
This is unambiguously wrong. Given that the Schwarzschild equation predicts that the enhanced GHG from 2XCO2 will reduce radiative cooling to space by roughly 3.6 W/m2, the question we are faced with is simply: How much does the planet need to warm to emit an additional 3.6 W/m2 and restore a steady state between incoming and outgoing radiation? In other words, what is the planet’s climate feedback parameter in W/m2 emitted per degK of surface warming? (W/m2/K)
If we ASSUME that the planet behaves like a gray body with a temperature of 288K and emissivity of 0.615, it is trivial to use the S-B equation to predict 3.3 W/m2 of increased emission per degK of warming. There is “no theoretical intrinsic effect via the GHE which isn’t supported by the theory” involved. This is rubbish.
To validate the above clearly-stated ASSUMPTION, climate modelers asked the same question of about 20 different climate models: What would happen if the planet were artificial warmed 1 degK everywhere, and unanimously got essentially the same answer: 3.2 W/m2/K.
3.6 W/m2 / 3.2 W/m2/K = 1.1 K the no-feedbacks climate sensitivity.
This is the best answer we have for how much warming we expect from 2XCO2. And it is somewhat irrelevant because it doesn’t account for feedbacks.
RW wrote: “The other main point is that because of this, all of the supposed science involving the sensitivity (and in purported support of the sensitivity) should be taken with a grain of salt. Especially the IPCC’s supported central best estimate of about 3.4K, but even significantly lower results from so-called ‘skeptics’ like Lindze, Spencer, Lewis, etc. are also specious.”
The IPCC’s 70% confidence interval for ECS ranges from 1.5-4.5 K, so they take 3.4 K with a mountain, not a grain, of salt. In fact, AR5 deliberate chose not to provide any “best estimate” for ECS, just a likely range. However, few of the IPCC’s models sample ECS below 2.5 K and none below 2.0 K, so their projections are misleading since aren’t based on their confidence interval.
There are plenty of reasons to be disdainful of AOGCM output, but not the work of those who attempt to illuminate climate sensitivity with observations. like those you mentioned above.
At the link below, I derived the following expression for the planet’s climate feedback parameter (dRi/dT):
dRi/dT = -4eoT^3 – oT^4*(de/dT) – S*(da/dT)
Planck feedback + other LWR feedbacks + SWR feedbacks
https://scienceofdoom.com/2019/01/10/opinions-and-perspectives-7-global-temperature-change-from-doubling-co2/#comment-135346
CERES has been observing 3.5 K of seasonal warming for more than a decade. It tells us that Planck feedback and other LWR feedbacks total -2.2 W/m2/K, meaning that the IPCC gets positive feedback from WV+LR about right, but their LWR cloud feedback is too positive.
There isn’t much seasonal warming in the tropics, so -2.2 W/m2/K may not be appropriate for the tropics. Lindzen looked at the tropics and came up with an LWR feedback of -5.3 W/m2/K. And Mauritsen and Stevens (2015), two supporters of the consensus, obtained an LWR feedback of -4 W/m2/K (confirming Lindzen’s conclusion about negative overall LWR feedback in the tropics). They modified a model to cause storms to cluster as seen in cloud resolving models (a type of IRIS effect) and obtained positive LWR feedback!
Click to access 236-Lindzen-Choi-2011.pdf
https://www.researchgate.net/profile/Thorsten_Mauritsen/publication/275268225_Missing_IRIS_effect_as_a_possible_cause_of_muted_hydrological_change_and_high_climate_sensitivity_in_models
So we no longer need to rely on the IPCCs climate models to have a reasonable idea of what LWR feedback is. We have observations.
The data doesn’t show a linear relationship between Ts and reflected SWR. LC11 found SWR feedback of -1.9+/-2.6 W/m2/K with a three-month lag (R^2 = 25%) and by my eye +3 W/m2/K with zero lag and similar R^2. I personally don’t see any reason to call any of these poorly-linear relationships “SWR feedback”. (For the LWR response, R^2 is about 60%.in the tropics and much higher for seasonal warming.)
There is a simple way to put some sort of limit on SWR feedback. Our planet reflects 100 W/m2 of SWR back to space. If that reflection changes +/-1%/K, that is +/-1 W/m2/K in SWR feedback. Given that the planet was 6K cooler at the LGM, +/-1 W/m2/K is a pretty big change. So I’m personally betting SWR feedback will be somewhere between +1 and -1 W/m2/K. Which puts overall ECS in the vicinity of EBMs.
RW, before you indiscriminately defecate on work you don’t understand, consider asking a narrowly focused question. Some climate scientists actually do real science: They start with observations.
Oz4caster wrote: “I’ve been thinking about how CO2 should effect temperature and it seems that wherever there are clouds the CO2 effect is eliminated or at least diminished depending on the thickness of the cloud cover. Also it seems that the CO2 effect should be diminished in places where skies are clear, but the column water vapor is relatively high as can happen mainly at tropical latitudes, but also at subtropical latitudes over the oceans. That leaves mainly deserts and polar regions when cloud cover is absent as the places where the effects of CO2 should be strongest. I’m guessing this is not a large portion of the globe.”
You can calculate radiative forcing for yourself with Modtran at the link below. A quick look suggest that the higher the cloud tops for thick clouds, the smaller the radiative forcing from 2XCO2, but the forcing isn’t cut in half. You can also adjust humidity. When Myhre (1999) arrived at 3.7 W/m2 for doubled CO2, he combined the radiative forcing calculated for five different regions with a variety of cloud covers. So this figure theoretically has taken into account most of the factors you discuss. Likewise AOGCM produce effective radiative forcing after the atmosphere has adjusted to warming and they average 3.4 W/m2 with a spread of about 1 W/m2.
http://climatemodels.uchicago.edu/modtran/
Oz4caster wrote: “Is this lack of uniform distribution of impact taken into account when estimating a global average 1.2C impact of CO2 with all else remaining the same?”
No. A warming of 1.2 degC for 2X CO2 comes from assuming the planet behaves like a gray body with a temperature of 288.0 K or 289.1 K and an emissivity of 0.615. These gray bodies would emit 239.9 and 243.6 W/m2. So a 1.1 degC rise in temperature of such a gray body would restore radiative balance after a doubling of CO2 – if nothing else changed.
However, the Earth doesn’t emit like a blackbody: Rising absolute humidity with rising temperature reduces radiative cooling to space, changing surface and cloud albedo changes reflection of SWR, changing lapse rate means more warming develops high in the troposphere and more LWR is emitted than for a simple gray body. And the altitude of cloud tops can change, changing their emission to space. All of these changes that develop proportionately to rising surface temperature are called feedbacks (measured in units of W/m2/J).
So 1.1 degC is called the no-feedbacks climate sensitivity.
Frank, thanks for the detailed and informative response. So it would seem that proper handling of a multitude of important different feedbacks is critical for more accurate modeling the land/ocean/atmosphere out to years/decades/centuries/millennia. Because of the complex nature of a variety of interacting feedbacks, I can see how even trying to diagnose what has happened in the past could be very difficult at best and nearly impossible at worst. And thus it may be difficult to properly validate models based on measurements over time.
It might be possible to tune a model to get fairly good performance compared to past measurements and for forecast validation out to a few years for a few parameters, like temperature or precipitation, but the short-term success could be for the wrong reasons. In which case longer-term results would be less likely to be accurate and adjustments would be needed after longer-term validation. I can see how this process of slow gradual improvement could take decades or more because of longer-term validation issues likely to arise.
Second oz4caster’s thanks, Frank. The most interesting sentence in your comment is the one about the spread of effective forcings in models.
“Likewise AOGCM produce effective radiative forcing after the atmosphere has adjusted to warming and they average 3.4 W/m2 with a spread of about 1 W/m2.”
For me this proves conclusively that any skill models have must be due to cancellation of errors.
Frank,
“No. A warming of 1.2 degC for 2X CO2 comes from assuming the planet behaves like a gray body with a temperature of 288.0 K or 289.1 K and an emissivity of 0.615. These gray bodies would emit 239.9 and 243.6 W/m2. So a 1.1 degC rise in temperature of such a gray body would restore radiative balance after a doubling of CO2 – if nothing else changed.”
But this type of surface/atmosphere warming has nothing to do with greenhouse effect warming of the surface, which is specifically how the increased CO2 is supposed to further elevate the surface temperature. Thus, the 1.2C is not giving you a theoretical intrinsic measure of the initial instantaneous imbalance at the TOA’s ability to further elevate the surface temperature.
oz4caster and dpy6629: The radiative forcing one calculates using the Schwarzschild equation for radiative transfer depends on the temperature and composition of the atmosphere through which OLR is passing. To the extent that different models predict different futures – at least regionally – those models will produce modestly different amounts of radiative forcing from 2XCO2.
In theory, models are tuned to match current observed climate and ONLY THEN are experiments run using historical forcing, 1% pa, and 4XCO2 and the optimum model. Isaac Held’s last blog post is a very candid discussion on whether or not models can be and are tuned to agree with the past. IIRC, one of the three models GFDL (his group) submitted for CMIP5 had a very low TCR compared with the other two.
dpy6629: A little more information on model tuning may be useful to you. A decade ago, a group mostly in England experimented with a collection of more than 1000 simplified AOGCMs whose key parameters were chosen at random from within a physically plausible range. The key finds were: 1) High climate sensitivity is far more common and only a few models showed climate sensitivity similar to EBMs. The high climate sensitivity of AOGCMs could be due to chance, not bias! 2) Parameters interact in unpredictable ways, making it likely that models tuned one parameter at a time represent local, not global optima. 3) When evaluated against a panel of 8 observables (temperature, precipitation, albedo), the plausible range for all varied parameters could not be narrowed because part of that range was never consistently associated with poorer performance. One nice thing about this group is that they have a list with mostly valid links to all of the group’s publications and presentation on this subject. Some of them discuss the “philosophy of science” implications of trying to make predictions about a massively important issue under these circumstances. AR4 had a brief discussion which described the IPCC models an “ensemble of opportunity” which doesn’t systematically explore the range of possibilities and whose range of output shouldn’t be used to draw probabilistic conclusions. The IPCC uses their “expert judgment” to characterize the 90% confidence interval from model output as “likely”, not very likely.
https://www.climateprediction.net/publications/?type=29&letter=&theme=39
Stainforth 2005: https://www.climateprediction.net/wp-content/publications/nature_first_results.pdf
Observables: https://www.ethz.ch/content/dam/ethz/special-interest/usys/iac/iac-dam/documents/group/climphys/knutti/publications/sanderson08jc.pdf
RW wrote about the no-feedbacks ECS of 1.1: “But this type of surface/atmosphere warming has nothing to do with greenhouse effect warming of the surface, which is specifically how the increased CO2 is supposed to further elevate the surface temperature. Thus, the 1.2C is not giving you a theoretical intrinsic measure of the initial instantaneous imbalance at the TOA’s ability to further elevate the surface temperature.”
I disagree. The temperature of convective regions of the troposphere (most of the troposphere) is controlled the moist adiabatic lapse rate. So a 1 degK rise in the average surface temperature for a decade is also going to be a 1 degK rise in the average temperature everywhere below the tropopause. The moist adiabatic lapse rate should be lower if absolute humidity rises, but that issue is dealt with by lapse rate feedback and therefore is not included in the no-feedbacks ECS of 1.1 degC.
The only model for the planet that we can use that is simple enough to be used to directly calculate the amount of warming expected from 2XCO2 is a simple graybody model. W = -eoT^4. dW/dT = -4eoT^3 = -3.3 W/m2/K near 288K. AOGCMs make essentially the same prediction when the temperature is uniformly raised everywhere by 1 degK without feedbacks: -3.2 W/m2/K. So the fact that the temperature varies from 190-310 K apparently doesn’t invalidate a simple graybody model.
If one wants to correct a radiative imbalance at the TOA by warming, one simply needs know dW/dT, the climate feedback parameter. And to express that concept in terms of Ts (if you are worried about a changing lapse rate). At the link below, I showed how the concept of dW/dT = -4eoT^3 = -3.3 W/m2/K can be enlarged to include feedbacks.
https://scienceofdoom.com/2019/01/10/opinions-and-perspectives-7-global-temperature-change-from-doubling-co2/#comment-135346
Frank,
“I disagree.”
Then you disagree with established GHE theory, which is that the lapse rate will shift in a specific way, i.e. the upper atmosphere (the stratosphere) will cool and the lower atmosphere (the troposphere) will warm (thus further warming the surface in the process). There are specific physics for why this is case, BTW.
Both solar induced warming of the surface and GHE induced warming of the surface are well established to have significantly different thermodynamic signatures, because different physics are driving them. The signature of solar induced surface warming includes stratospheric warming, which is consistent with the 1.2C being the theoretical intrinsic measure of surface warming for +3.7 W/m^2 of (post albedo) solar power entering the system and NOT that from +3.7 W/m^2 from GHGs via the GHE.
The emissivity relationship you cite is just the reciprocal of the relationship offsetting the input to the system, or it’s the inverse of the system gain, which is about 1.6, i.e. 385/240 = 1.6 and 240/385 = 0.62, where 240 is the post albedo solar power entering the system and the 385 is the radiant power emitted by the surface. The physical meaning of the 1.6 is that it takes about 1.6 W/m^2 of net surface gain to allow 1 W/m^2 to leave the system at the TOA, offsetting each 1 W/m^2 entering the system post albedo from the Sun. It quantifies a linear increase in the aggregate thermodynamic path offsetting watts entering the system from the Sun, and has nothing to do with and/or has no direct connection to GHE induced warming of the surface.
Though of course in the case of the 3.7 W/m^2 for each, both will result in an instantaneous TOA deficit of -3.7 W/m^2 that has to be restored, so you can apply it for the GHGs case as well and it will restore balance, but this is trivially true and has nothing to do with GHE induced warming of the surface.
So Frank, I’m not arguing against basics. I agree the GHE theory is correct (or at least probably correct). I’m inclined to think it is since it was debated and established decades before the science could have been considered to be corrupted by politics or money. Besides, the arguments against it don’t make sense to me and I find that the GHE just intuitively makes sense.
It’s the magnitude of the effect supported by the IPCC and its ensemble of models (about 3.3C per 2xCO2) that I think is way off, by at least a factor of 3 and probably closer to a factor of 10. I highly doubt it’s more than about 0.5C, but one can never be completely sure (of course).
The main point regarding the 1.1C of claimed intrinsic effect is to illustrate how loose and specious it and what it’s used to derive, i.e. the sensitivity, is.
Frank,
What I said here:
“BTW, I don’t really see a fundamental problem with using this for diagnosing net feedback response, but as you know, it’s quite tricky to accurately determine this by looking at it in this way because of the difficultly of determining cause and effect, or forcing vs. feedback, i.e. temperature changes clouds changes.”
Meaning that it is theorized that cloud changes can be both a forcing (causing temperature changes) and a feedback (responding to temperature changes from to some other forcing), and determining this is tricky by looking at TOA fluxes. That is, the direction of causation goes (or can go) both ways, so what are you really measuring at the TOA? Maybe the observed clouds changes are causing an increase in post albedo solar power entering the system, but this is being mis-interpreted as a feedback response to observed surface warming. This is primarily the argument Roy Spencer has been making for years now against the models and papers purporting to support high sensitivity. Lindzen too.
Thanks Frank for the references.
RW wrote about the no-feedbacks climate sensitivity: “Then [Frank] with established GHE theory, which is that the lapse rate will shift in a specific way, i.e. the upper atmosphere (the stratosphere) will cool and the lower atmosphere (the troposphere) will warm (thus further warming the surface in the process). There are specific physics for why this is case, BTW.”
However, the conventional GHE says nothing about what happens to the stratosphere. Nor does lapse rate feedback refer to the stratosphere. Most people think rising GHGs make the atmosphere warmer because they trap heat – which is incorrect. The usual GHE exists only where temperature to falling with increasing altitude and temperature is rising with altitude in the lower stratosphere. See my article on the Schwarzschild equation, particularly the section on the origin of the GHE.
https://en.wikipedia.org/wiki/Schwarzschild%27s_equation_for_radiative_transfer
In order to prove that recent warming is due to rising GHGs rather than the sun, the IPCC correctly emphasizes that a solar forcing will cause the stratosphere to warm while the forcing from GHGs will cause the stratosphere to cool. The warming is due to SWR, which has nothing to GHGs and the LWR they absorb. So warming has little to do with “GHE” and the initial question I (correctly) addressed concerning 1.2 K of warming in response to 2XCO2. To understand cooling, see my article.
Frank,
“However, the conventional GHE says nothing about what happens to the stratosphere. Nor does lapse rate feedback refer to the stratosphere. Most people think rising GHGs make the atmosphere warmer because they trap heat – which is incorrect. The usual GHE exists only where temperature to falling with increasing altitude and temperature is rising with altitude in the lower stratosphere. See my article on the Schwarzschild equation, particularly the section on the origin of the GHE.”
I see where it says that in the article. I understand that the GHE theory does indeed say that the upper atmosphere (which I generally is meant to be the stratosphere) will cool for additional surface warming. While I understand (I think) that temperature increases with height in the lower stratosphere, this has nothing to do with the GHE signature for incremental surface/atmosphere warming compared to that of the signature for solar induced surface/atmosphere warming, where for the GHE induced signature the stratosphere will cool (and for solar it will warm). This is best illustrated with diagrams of each signature, but I don’t know how to post them here. I’ve drawn up Windows Paint files that clearly show the two different signatures.
“In order to prove that recent warming is due to rising GHGs rather than the sun, the IPCC correctly emphasizes that a solar forcing will cause the stratosphere to warm while the forcing from GHGs will cause the stratosphere to cool. The warming is due to SWR, which has nothing to GHGs and the LWR they absorb. So warming has little to do with “GHE” and the initial question I (correctly) addressed concerning 1.2 K of warming in response to 2XCO2. To understand cooling, see my article.”
Yes, but it’s the 1.2C claimed measure of theoretical intrinsic effect via the GHE which isn’t supported by the theory (and which is my whole point). It’s an amount that will restore balance — yes, but it’s arbitrarily used more or less because the field doesn’t have a valid model or a way to quantify the theoretically supportable intrinsic effect from 2xCO2.
The other main point is that because of this, all of the supposed science involving the sensitivity (and in purported support of the sensitivity) should be taken with a grain of salt. Especially the IPCC’s supported central best estimate of about 3.4K, but even significantly lower results from so-called ‘skeptics’ like Lindze, Spencer, Lewis, etc. are also specious.
RW has the ability to hijack complete comment threads with more or less the same confused jumble of words each time.
Anyone interested in RW’s wisdom – search for “RW” in the search bar. Thousands of comments to review.
I don’t believe RW has anything new to add so unfortunately I will be deleting new comments from this esteemed commenter. On the positive side he has a large volume of extant work preserved here for interested parties.
SoD wrote: “So a slightly different simulation has relative humidity staying constant. The result is some amplification from the water vapor. The Earth’s surface gets a little hotter, so there’s more water vapor – which is also a greenhouse gas – so the surface gets hotter still. In this experiment the global temperature change will be about 2.4ºC from pre-industrial levels. ”
A question. How would this globe be? Is it 100% relative humidity all over, or is it just based on the humidity we already have, adjusted for temperatures? I don`t know how the relative humidity gradient is, and how it changes with latitude and altitude over time.
And thank you SoD for the simplification and clarification. It makes it easier to follow.
” Is it 100% relative humidity all over, or is it just based on the humidity we already have, adjusted for temperatures?”
The latter – *relative* humidity is as today.
It looks like relative humidity is going down, perhaps 1% over land since 1981. If this should be a trend it will have some influence. I don`t know how this is calculated in climate models.
https://climate.copernicus.eu/monthly-summaries-precipitation-relative-humidity-and-soil-moisture
I am not familiar with the data and would be very wary of attempting a naive analysis of it.
Generally the IPCC is a good place to go for a first look.
According to IPCC AR5 (Chapter 2):
I think you’ve found the evidence for the “lower levels” bit.
From: Recent changes of relative humidity: regional connections with land and ocean processes. Sergio M. Vicente-Serrano et.al. 2017
Click to access esd-9-915-2018.pdf
“Our results suggest significant RH trends over many regions worldwide and a generally dominant negative trend at the global scale.”
I have a comment in moderation above – but briefly I think you’re looking at surface RH, whereas tropospheric RH is what matters for the GHE.
Have a read of IPCC AR5 2.5.4 and 2.5.5 to see differences in reported trends of the two.
See above discussion chubs and I had of humidity. UAH and RSS have different conclusions. Yet another aspect of climate science where the quantification is poor.
Here is a sample of relative humidity, from WATER VAPOR FEEDBACK AND GLOBAL WARMING, Isaac M. Held and Brian J. Soden, 2000 – an excellent review paper.
Left is actual, right is from a climate model:
Click to enlarge
If you click and enlarge it you can see the values of relative humidity on the contours. Much lower than 100% in the free troposphere (the troposphere above the boundary layer, typically 1km).
My first take is that the model is producing more water vapor aloft than the observations. Could that be due to the model producing a hot spot aloft that the observations don’t contain?
Except over Antartica that is where the model looks badly wrong.
The chart is useful. The model depiction is more than adequate to get water vapor feedback right. After all the model pattern could be completely wrong and still nail water vapor feedback. It is simple physics, warm air holds more water vapor.
How many different climate models have evaluated water vapor feedback? Simple ones discussed in the blog and many generations of 3-D models from the mid-70s to today. All with about the same answer.
“Probability distribution functions of RH from radiosondes and AIRS are similar, suggesting that variability over Antarctica is well reproduced by the satellite. AIRS data are also compared to simulations from the Community Atmosphere Model version 3 (CAM3) and are found to be significantly moister than the model, although the model does not allow supersaturation with respect to ice or liquid water. A climatology from AIRS indicates that it has a repeatable annual cycle over Antarctica. Supersaturation with respect to ice is very common over the continent, particularly in winter, where it might occur almost half the time in the troposphere. This may affect the quantity and isotopic composition of ice over Antarctica.” Repeat: “the model does not allow supersaturation with respect to ice or liquid water”. Why?
From: Gettelman, A., V. P. Walden, L. M. Miloshevich, J. Roth, and B. Halter (2006), Relative humidity over Antarctica from radiosondes, satellites, and a general circulation model.
I have also seen elsewhere that models have problems with the climate of Antarctica. This climate affect a great area of the Southern Hemisphere, so it will also have global implications. An I think it is some systematical model biases when it comes to change of both altitudal temperatures, winds and humidity.
SOD: “This allows us to calculate how the radiation balance gets disturbed. Then we can find the new surface temperature which brings everything back into balance.”
I’m finding it much more useful to add another sentence at this point: “To accomplish this, we need to know how much extra heat the planet radiates (LWR) and reflects (SWR) to space PER DEGK as it warms.” This is the climate feedback parameter (under-discussed at blogs) expressed in W/m2/K.
Suppose the disturbance (forcing) is a 3.6 W/m2 reduction in radiative cooling to space or a 3.6 W/m2 increase in post-albedo TSI. If the planet’s climate feedback parameter is -1, -2, -3, or 4 W/m2/K, it is easy to see that a radiative balance will be restored after a warming of 3.6, 1.8, 1.2 or 0.9 K.
W = -eoT^4. dW/dT = -4oeT^4 = -3.3 W/m2/K. A simple gray body model for the Earth has a climate feedback parameter of -3.3 W/m2/K and will warm slightly more than 1 K in response to ANY +3.6 W/m2 disturbance. There is absolutely no need to express any of this physics in terms of CO2. (F = ma whether the forces comes from gravity, pressure*area, electricity, etc.)
In their CO2-centric world, climate scientists stupidly hide the simplicity and universality of this approach by taking the reciprocal of the climate feedback parameter (K/(W/m2)) and then converting W/m2 into doublings of CO2 (using an uncertain conversion factor). This makes about as much sense as using g (the acceleration produced by gravity at the surface of the Earth) to quantify force. You may laugh and say that force and acceleration have different units (kg-m/s2 and m/s2). However, a forcing (power from W/m2 times surface area of the planet) and temperature (energy) also have different units. The POWER from radiative forcing produces a warming RATE, and that rate approaches zero thanks to a negative climate feedback parameter.
I’ve been struggling for a long time to express the above ideas mathematically. Today, eureka. The planet’s radiative imbalance (Ri) is given by:
Ri = S*(1-a) – eoT^4
dRi/dT = -4eoT^3 – oT^4*(de/dT) – S*(da/dT)
dRi/dT is the climate feedback parameter.
The first term is Planck feedback, the second term is the sun of all of the other LWR feedbacks, and the third term the sum of all of the SWR feedbacks.
Our planet’s emissivity changes with temperature so we must use the chain-rule when differentiating. Emissive changes with changing surface temperature (T not Ts for clarity) because: 1) Absolute humidity may change with temperature and water vapor is a GHG. 2) Changing humidity may reduce the lapse rate may cause more warming higher in the atmosphere than at the surface and more emission per degK of warming that expected from Planck feedback (4oeT^3). 3) Changing altitude/temperature of cloud tops will change their emissivity. Reflection of SWR can change as surface and cloud albedo change with surface temperature, producing SWR feedbacks.
What climate scientists call “feedbacks” is really simple physics: a combination of the S-B equation, and changing emissivity and albedo with temperature.
There is a riddle in climate science: It is about how to explain global heat uptake, of which ocean heat uptake is over 90%.
Activists have their mantra: It is caused by rising greenhouse gas concentrations. And this mantra spills over when scientists present their findings. An important paper; Global reconstruction of historical ocean heat storage and transport. Laure Zanna, Samar Khatiwala, Jonathan M. Gregory, Jonathan Ison, and Patrick Heimbach, 2019
“Since the 19th century, rising greenhouse gas concentrations have caused the ocean to absorb most of the Earth’s excess heat and warm up. Before the 1990s, most ocean temperature measurements were above 700 m and therefore, insufficient for an accurate global estimate of ocean warming. We present a method to reconstruct ocean temperature changes with global, full-depth ocean coverage, revealing warming of 436 ×1021 J since 1871. Our reconstruction, which agrees with other estimates for the well-observed period, demonstrates that the ocean absorbed as much heat during 1921–1946 as during 1990–2015. Since the 1950s, up to one-half of excess heat in the Atlantic Ocean at midlatitudes has come from other regions via circulation-related changes in heat transport.”
It is a great confirmation of the global warming period from 1921 to 1946 (in the northern hemisphere from 1913).” Absorbing as much heat as during 1990-2015″. So this early heat uptake was also caused by a huge increase of GHGs? Short wawe radiation, together with UV, is heating the oceans. It doesn`t convince me of the GHG causality. And I think climate models are far from explaining the global warming from 1921 to 1946.
And climate science has no credibility unless such natural variations are understood. Period.
I doubt that there is enough data to ever fully explain it, and they will have to move on regardless. Period.
And climate science has no credibility unless such natural variations are understood. Period.
This doesn’t track.
If I show you bones from a human that died 100,000 years ago, and we couldn’t determine the cause of death from the available data, does that mean we also can’t determine the cause of death from someone who died yesterday, in a hospital?
Nope. Obviously not.
Why? Because we don’t have enough data to determine the cause of death from the millennia-old bones, but we do have enough data to understand the cause of death from the person who died yesterday.
Sometimes, there simply isn’t enough data to tell what happened. As with past natural climate variations — we lack the data to say what caused them. It doesn’t mean we don’t understand climate.
“Since the 19th century, rising greenhouse gas concentrations have caused the ocean to absorb most of the Earth’s excess heat and warm up.”
This seems like a simplistic causal explanation to me.
But it may be untrue. Where is the evidence?
GHGs may warm ocean surface a little,and backradiation may slow down net IR radiation a little. But it is water circulation which carry away heat. Can it just be heat, and not “exess heat”?
Is it impossible that some of the same ocean mechanisms are present now as it was 80 years ago?
Ah, I think I follow you.
Here’s what I’d say: the ocean is going to take the bulk of the heat that causes the temperature change, whatever the cause. Back-radiation or whatever affects the kinetics of the heat transfer a little, but it doesn’t take that much for heat to get into or out of the oceans.
As for the causes, yeah, we don’t have a lot of great data before 1950 on the specific forcings for the time.
Judith Curry: curryja January 12, 2019: the really interesting OHC stuff is prior to 1950
Steven Mosher: January 13, 2019: yup
Chubbs wrote:
“In-any-case OHC, surface temperature and TPW are all increasing rapidly as predicted by models. ”
In general current models use an antiquated concept (“potential temperature”) which has been replaced with “Conservative Temperature”.
Dr. McDougal points out: “present ocean models contain typical errors of 0.1°C and maximum errors of 1.4°C in their temperature because of the neglect of the nonconservative production of potential temperature” …
I have asked a modeller working on huge models (mainframe types) if they (the developers in her team) have adopted TEOS-10 yet, to which she indicated not yet.
There are still some systemic problems with models in terms of ocean atmosphere interactions (e.g. The Ghost Photons).
I guess one has to know a lot about a particular model to know how it handles the “old software thermodynamics” and the “new software thermodynamics.”
I have a question regarding the forcing from CO2 versus the content of water vapor.
The forcing from methane is dependent of other gasses in a complicated formula.so i wonder if the CO2 forcing really is independent of even the most powerfull GHG, water vapor?
A connected side question: Is the CO2 forcing dependent of the cloud cover?
Please state the assumptions you have to make to answer this rather complicated questions.
I have tried to find answers, but have failed so far. Hope You can help.
Svend,
It’s a complex and technical subject.
There’s some terminology that’s important.
“Forcing” = (in simple terms) change in the radiative balance of the planet as a result of changes in GHGs, like CO2.
“Feedback” = the resulting changes following radiative forcing, for example, the changes in water vapor and clouds and how they affect the climate.
So changes in CO2 are a forcing and the resulting changes in water vapor and clouds are a feedback.
But what you ask – as I read it – might not be what you mean to ask.
Depending on your appetite for learning about the important concepts, you might find these articles helpful:
Wonderland, Radiative Forcing and the Rate of Inflation
Wonderland and Radiative Forcing – Part Two
You might really be asking about whether water vapor overwhelms the CO2 effect?
Visualizing Atmospheric Radiation – Part Four – Water Vapor
– it might be best to start at the beginning of that series to get familiar with the concepts.
Svend: In the past, I have tried to find answers to some of your questions. The online MODTRAN calculator can provide answers many questions. I don’t know enough about the details of the program to be completely sure these are correct answers, but they satisfied my amateur needs.
For example, by changing the water vapor scale from 1.00 to 1.01, I think you are modeling a 1% increase water vapor at all altitudes as a forcing. If you use a temperature offset of 1 K, you can compare constant absolute humidity to constant relative humidity. The difference appears to be water vapor feedback.
For clear skies, OLR is presumably calculated by integrating the Schwarzschild equation from the surface (at surface temperature, ? emissivity) to the TOA. For cloudy skies, OLR is calculated starting from the cloud top (temperature determined by lapse rate?) and integrating to space. MODTRAN offers a variety of cloud options. If you “look down” through clear skies from 2 km, you can compare OLR at 2 km to a hypothetical cloud emitting blackbody radiation upward from that altitude.
http://climatemodels.uchicago.edu/modtran/
I had a follow up to the first answer from SOD, but it somehow disappeared.
My question was rather simple:
The forcing from methane is dependent of other gasses in a complicated formula.so i wonder if the CO2 forcing really is independent of even the most powerfull GHG, water vapor?
How could it be that the forcing from changing CO2 content 5.35ln(C/Co) is independent of the other green house gasses especially H2O.
I find it strange that this very simple forcing formula is completely independent of
other GHG gasses that share the same radiation bands.
If the answer is that it is so, i would like a little explanation of how that could be.
If the answer is that it is dependent i would like a better formula.
I know it is very complicated because water vapor vary a lot with hight, but anyway some approximations would do.
I am not born and raised with english. Concider that if you find my wording a little strange.
There is at least a theoretical connection. More water vapor gives less CO2 forcing, because they are sharing radiation bands. The same with other GHGs. In practice I think this is temperature dependent. Warming will give both more outgassing of CO2 (as a positive feedback), and more water vapor (both positive and negative feedbacks). The goegraphy of warming will also make some difference. So it will be very complicated to place all this into a formula.
Svend: You can answer this question using the online Modtran Radiation Transfer Calculator: For 400 to 800 ppm with a US Standard Atmosphere, the difference in OLR is 3.0 W/m2 with the normal amount of water vapor, 3.2 W/m2 for half the normal amount of water vapor at all altitudes and 2.6 W/m2 for twice the normal amount of water vapor (an impossibility since that would be supersaturation in many places). A 10% increase in water vapor doesn’t cause a measurable increase. So forcing does change with humidity at constant surface temperature, but the change is negligible in practice
In the tropics, where it is both warmer and more humid, doubling reduces OLR by 3.3 W/m2. Warmer surface temperature means there is more OLR that can be absorbed and more humidity that can absorb it. In this case, temperature happens to be more important. Perhaps the opposite is true in polar regions.
http://climatemodels.uchicago.edu/modtran/
Frank. Thanks for answering on what i asked for. Some of the explanations took into account all kinds of side effects, but if you don’t have the first effect clear, how could you incorporate the side effects (feedback).
Take it as a sort of joke, but next time i might ask how the forcing from watervapor depends on the CO2 content.
Svend: IMO, water vapor responds to changing temperature much faster than temperature can respond to water vapor as s forcing. Yes, relative humidity will have some impact on whether the temperature here tonight drops 6 or 7 degC tonight. However, if we think globally and consider how long it takes for any forcing to change the temperature of the atmosphere, surface and mixed layer of the ocean, changes in water vapor are driven by temperature much faster than temperature is changed by water vapor.
Therefore, we can handle the forcing from water vapor in terms of a feedback (W/m2/K) multiplied by a temperature change. It doesn’t need to be done that way, but it is far simpler. Increasing water vapor also reduces the lapse rate, so any surface warming should produced enhanced warming at higher altitudes, causing a greater increase in OLR than expected for a given increase in surface temperature. We could also call changes in the lapse rate another forcing from water vapor, since this does change the radiative balance at the TOA.
It really helps to keep two phenomena separate: forcing and feedback. The planet’s response at the TOA to any change in surface temperature – even one caused by internal variability is a feedback. If we could mix the ocean and lower the average surface temperature everywhere by many degC, we would have a temperature change without any forcing. And the climate feedback parameter (W/m2/K) times that temperature change would tell us how the radiative balance at the TOA would change, including the effect of less water vapor in a colder atmosphere. So it makes sense to include the effects of water vapor as a response to temperature change – a feedback – rather than a driver of temperature change.
This will be correct – of course – only if water vapor is controlled by temperature. If changes in agriculture change water vapor, this approach will be incorrect.
Isn’t water vapor controlled by evaporation, not temperature? The evaporation rate is proportional to wind speed and “undersaturation”, but evaporation can’t continue indefinitely without the vertical convection needed to produce precipitation. And convection doesn’t occur without an unstable lapse rate. As convection warms the upper troposphere, the lapse rate will not remain unstable unless heat is radiated across the TOA just as fast as it is convected upwards. (Modtran can demonstrate that net radiative cooling, OLR-DLR, doesn’t change much with surface temperature.)
Frank and SoD
The radiation at TOA tells about an overall balance, but on the ground where we live it is a different story.
I wonder why i have not heard or seen papers based on the radiation measurements from this network: https://www.esrl.noaa.gov/gmd/grad/surfrad/
And there are other stations around the world.It will be local measurements but we live local and some with more skills and tools than me should dig into these data, so that the surface kan be better connected to TOA.
It is the best available data on radiation at groundlevel.
The very different climate they are placed in could reveal a lot.
Svend,
There are posts on this site that use data from the Surfrad project. For example: https://scienceofdoom.com/2010/07/24/the-amazing-case-of-back-radiation-part-two/
You might also want to look at EBEX 2000, which is mentioned here:
https://scienceofdoom.com/2010/12/23/understanding-atmospheric-radiation-and-the-greenhouse-effect-part-one/
[…] « Opinions and Perspectives – 7 – Global Temperature Change from Doubling CO2 […]
[…] Opinions and Perspectives – 7 – Global Temperature Change from Doubling CO2 […]
There is a recent paper ( https://link.springer.com/article/10.1007%2Fs00382-019-04991-y) in which the authors claim, that the estimations from EBM and obs. are biased low. Not so, Nic Lewis deduced: https://climateaudit.org/2019/10/17/gregory-et-al-2019-unsound-claims-about-bias-in-climate-feedback-and-climate-sensitivity-estimation/
Nic Lewis doesn’t do thermal diffusion into the oceans correctly.
Apparently, neither do AOGCM’s.
I cannot see that Gregory and others are so occupied by “the thermal diffusion into oceans”. It is the currents that matters.
“Our results highlight that the substantial amounts of heat accumulated in the ocean and associated sea-level rise can be influenced by ocean circulation changes and low- to midlatitude air–sea interactions.” From Global reconstruction of historical ocean heat storage and transport. Laure Zanna, Samar Khatiwala, Jonathan M. Gregory, Jonathan Ison, and Patrick Heimbach
So, where does Nic Lewis disagree with this?
Zanna et. al 2019. https://www.pnas.org/content/116/4/1126
And from : The Impact of Recent Forcing and Ocean Heat Uptake Data on Estimates of Climate Sensitivity. Nicholas Lewis, Judith Curry, 2017.
“As in previous energy budget studies, AOGCM simulation-derived estimates of heat uptake are used for the base periods, since OHC was not measured then. The heat uptake values used in LC15, which were derived from simulations by CCSM4 starting in AD 850 (Gregory et al. 2013), scaled by 0.60, were 0.15, 0.10, and 0.20 W m−2 respectively for the 1859–82, 1850–1900, and 1930–50 base periods. The unscaled CCSM4-derived values were consistent with the value derived by Gregory et al. (2002) from a different AOGCM.”
The contradiction of NL is not related to some “thermal diffusion into the oceans” . Gregory et al (2019) claim, that the sensitivity estimations from EBM and obs. are biased low and justify this with some bias in linear (OLS) regressions of the forcing vs. the GMST. But: all the bias comes from the volcano forcing and it’s impact on the GMST. And NL never used some linear regression over some volcano influenced timespans. One can show easily that in the case of one avoids those events by the selection of begin-and end time frames ( as it was done in Otto et all, LC14 and LC18) the by G. (2019) claimed bias is zero. This is the core message of NL’s Posts, also reposted at Climate etc.
“Frank” says:
Of course it explains why Nic Lewis’s analysis is poor. How do you think the ocean acts like a heat sink?
Geoenergymath: “Frank” and “frankclimate” are two different people. “Frank” registered with WordPress as “franktoo” in order to comment at Judy’s and I sometimes forget to log out of that identity here and elsewhere. I find that frankclimate’s comments, including the one above, are usually sound.
Frank(2): some times ago we had both the same nick and this gave some confusion, however we were d’accord in this cases. … IMO you changed to Frank2 and I changed to frankclimate. “We’ll meet again” (with appolgies to Jonny Cash). 🙂
Reblogged this on Climate- Science.press.
Sod, thanks for the extensive information suitable for a range of readers. Can I ask some basic questions (apologies if I’m going over old ground).
My simplistic understanding is that the outer layer of CO2 at TOA is the ‘sphere’ mainly controlling earth’s temperature by downward radiation of IR. If the planet is warming there is less IR radiating into space (one of your models showed 4.3 W/m2 from memory); presumably this also means there is more atmospheric window radiation escaping earth from its warmer surface.
We’ve had almost 1°C warming since the 1970’s so do satellite measurements of earth’s radiation spectra concur with my simplistic ideas?
Also, is there a level of CO2 that would be saturated at some top ‘sphere’ and limit further warming. (This may be too simplistic a question considering complicated feedbacks in the system).
Thanks in advance…
OpenMindedSceptic
It is old ground, but no apology necessary.
Not really.
For a simple explanation (with lots of linked references to other pages) try The “Greenhouse” Effect Explained in Simple Terms.
For getting a clearer picture of atmospheric radiation I recommend Visualizing Atmospheric Radiation.
—
The subject isn’t intuitive for most people. A long time ago, when I started trying to understand atmospheric radiation to understand climate I found myself quite confused. I could see the maths and follow it (once the rust was removed from my brain) but no conceptual picture emerged in my head. Not immediately anyway. When I tried to understand if CO2 was already “saturated”, I also found it very challenging.
There are some simple pictures painted by climate scientists – usually too simplistic – but that’s not to fault them, they are trying to help the average person get some sense of what’s going on. Then there are textbooks and papers – usually much too in depth and detailed for the average person.
Bridging the gap.. in the class of “explaining how more CO2 leads to more surface warming, including proofs for those interested, for someone with some maths or physics background” there isn’t much about. It’s one of the reasons I started this blog.
OpenMindedSceptic said: “If the planet is warming there is less IR radiating into space (one of your models showed 4.3 W/m2 from memory); presumably this also means there is more atmospheric window radiation escaping earth from its warmer surface.”
To clarify cause and effect, it might be better to say: If the planet is absorbing more SWR than it is emitting LWR (due to rising GHGs), then the planet is warming.
OpenMindedSceptic said: “My simplistic understanding is that the outer layer of CO2 at TOA is the ‘sphere’ mainly controlling earth’s temperature by downward radiation of IR.”
This statement is problematic for a number of reasons. Earth’s temperature varies from 190-310K. The temperature at the surface (or any other location) is determined by convection and conduction, as well as radiation. It is always an oversimplification to say that temperature is determined by one particular flux or that a change in temperature is caused by a change in one particular flux. Since only radiation can carry heat to and from space (across the TOA, roughly 70 km above the surface), we can say that the internal energy below the TOA is controlled by the net flux of radiation across the TOA (absorbed SWR minus OLR, not DLR). A reduction in radiative cooling to space with constant absorption of SWR must result in warming somewhere, but this doesn’t tell us where below the TOA warming will be observed, how fast warming will take place, etc.
Some models for the GHE and enhanced GHE use layers or spherical shells. The best of these models is based on a “characteristic emission altitude”, the altitude where the average photon escaping to space is emitted. Rising GHGs will reduce the likelihood that a photon emitted at any altitude escapes to space, thereby raising the characteristic emission altitude. However, higher is colder and less dense, so fewer photons will be emitted from the new characteristic emission level than the old one. That creates a radiative imbalance that will persist until warming raises the temperature at the new characteristic emission altitude. A constant lapse rate will transmit that warming to the surface. Like all simple models, this model may create an initial sense of understanding, but may later slow progress toward a more complete understanding.
The full physics for radiation traveling through an absorbing/emitting atmosphere is given by Schwarzschild’s Equation for Radiative Transfer, which is what climate scientists are using when they say they are performing radiative transfer calculations. (Terms for scattering are also used, especially for visible light). However, Schwarzschild’s differential equation must be numerically integrated along a path (say from the surface to space for OLR), and the results of numerical integration don’t provide many with an intuitive understanding of the GHE. So many rely on simplified models like the characteristic emission level model. When you become dissatisfied with or confused by simple models of the GHE, you need to turn to the complete physics, as discussed here or at Wikipedia.
https://en.wikipedia.org/wiki/Schwarzschild%27s_equation_for_radiative_transfer
https://scienceofdoom.com/2011/02/07/understanding-atmospheric-radiation-and-the-“greenhouse”-effect-–-part-six-the-equations/
OpenMindedSceptic
“We’ve had almost 1°C warming since the 1970’s so do satellite measurements of earth’s radiation spectra concur with my simplistic ideas?”
And the classical Greenhouse theory: ” If the planet is warming there is less IR radiating into space (one of your models showed 4.3 W/m2 from memory); presumably this also means there is more atmospheric window radiation escaping earth from its warmer surface.”
But it looks like the Planet Earth behave in a different way. More planet warming means more TOA LW radiation. And it is a decrease in the shortwave radiation out that warms the atmosphere.
From Decadal Changes of Earth’s Outgoing Longwave Radiation,
Steven Dewitte * and Nicolas Clerbaux, 2018.
“The OLR has been rising since 1985, and correlates well with the rising global temperature. An observational estimate of the derivative of the OLR with respect to temperature of 2.93 +/− 0.3 W/m2K is obtained. The regional patterns of the observed OLR change from 1985–2000 to 2001–2017 show a warming pattern in the Northern Hemisphere in particular in the Arctic, as well as tropical cloudiness changes related to a strengthening of La Niña.”
So the part played by CO2 in this historical changes in long wave and short wave radiation is a bit of mystery for me. How do CO2 make a difference when it comes to lapse rate, water vapor and clouds, if its warming effect is different than we believe.?
Again and again I see this kind of claim from wellknown climate scientists: “Since the 19th century, rising greenhouse gas concentrations have caused the ocean to absorb most of the Earth’s excess heat and warm up.”
If the OLR has been rising, wouldn`t that mean that rising greenhouse gas concentration had a cooling effect on the oceans, especially over the northern hemisphere? An attribution problem of where the ocean heat is coming from?
NK: According to energy balance models, ECS is determined by the forcing from doubled ECS and the climate feedback parameter, lambda.
F + lambda*ECS = 0
ECS = -F_2x/lambda
As best I can tell, lambda can be thought of as the planet’s radiative response to a change in GMST – the increase in emission of LWR plus the increase in reflection of SWR per degC increase in GMST (when warming has eliminated the radiative imbalance at the TOA created by a forcing). OSR = outgoing SWR = reflected SWR.
lambda = dOLR/dTs + dOSR/dTs
F_2x is the change in radiative cooling to space (after the stratosphere adjusts) accompanying an instantaneous doubling of CO2 determined by radiative transfer calculations through today’s troposphere – a change that is independent of temperature.
Some people use effective radiative radiative forcing (ERF) instead of traditional radiative forcing, but the later is simpler in this context.
If correct, Dewitte and Clerbaux tell us dOLR/dTs is -2.93 +/− 0.3 W/m2/K (a negative value for heat loss. If ECS were 3.6 K and F_2x were 3.6, then lambda would be -1.0 W/m2/K and dOSR/dTs would need to be +1.9 W/m2/K. If ECS were 1.8 or 1.2 K and F_2x were 3.6 K, then lambda would be -2.0 or -3.0 W/m2/K respectively. In those cases, dOSR/dTs would need to be +0.9 W/m2/K or -0.1 W/m2/K respectively.
Therefore, even if dOLR/dTs were -2.9 W/m2/K, ECS can take on almost any value – if cloud SWR feedback is positive enough.
In his paper on the IRIS effect, Stevens suggests dOLR/dTs could be as large as -4 W/m2/K in the tropics, but only if fewer clouds let more LWR escape to space, with the consequence that less SWR would be reflected at the same time. Lindzen and Choi claimed dOLR/dTs in the tropics is about -5 W/m2/K. Tsushima and Manabe find that dOLR/dTs is -2.2 W/m2/K in response to seasonal warming. This response is weighted towards the regions of the planet with the largest seasonal change in temperature – ie outside the tropics.
Unfortunately, we don’t know dOSR/dTs, in part because monthly changes in OSR (unlike OLR) don’t have a simple linear relationship with changes in Ts. Some of the response is clearly lagged and (changes sign from positive to negative feedback with longer lags). I expect dOSR observed from space through clear skies due to changes in surface albedo (seasonal snow cover and sea ice) to lag behind dTs by several months, but the same phenomena is observed from cloudy skies. (This could be because (poorly-modeled) marine boundary layer clouds are non-local phenomena produced by descending warm dry air imported from long distances and cold ocean currents and upwelling.)
Let’s suppose dOLR/dTs is -2.9 W/m2/K, dOSR is +1.9 W/m2/K, lambda is -1.0 W/m2/K and ECS is 3.6 K/doubling. After 1.25 K of warming, dOLR is -3.6 W/m2 and the LWR imbalance created by 2XCO2 at the TOA has been reduced to 0 W/m2. However OSR has increased by +1.9*1.25 = 2.4 W/m2. So the imbalance has only been reduced by only 1/3 and it is now observed in the SWR channel, despite initially having been created by GHGs in the OLR channel. IIRC, most abrupt 2X and 4X experiments show OLR returning to balance about halfway to steady state warming, with further warming being driven by the increasing amount of SWR absorbed.
In other words, a value of -2.93 +/− 0.3 W/m2/K for dOLR/dTs is not grossly inconsistent with the behavior of AOGCMs. In my very limited experience, modelers simply focus on ECS and F_2X derived from 4X experiments and don’t spend much time discussing its long and short-wavelength components, dOLR/dTs + dOSR/dTs.
SoD, any cmoment on this moment report on the increase in the lower bound for warming in a 2x co2 world? I thought it was pretty good.
https://www.washingtonpost.com/weather/2020/07/22/climate-sensitivity-co2/#click=https://t.co/Gq30dsrylq
joeyblau,
I can’t see the paper yet. Not via institutional access. Probably need to wait a few weeks. Feel free to remind me in a month – I keep lots of open tabs and then forget about them.
Joey and SOD: The article discussed in the Washington Post article can be found at the link below, but no version of this article (published 7/22) is currently available outside the journal’s paywall. Some of these authors must work for organizations that require public posting of their papers, but that might take some time. This is just a guess, but the paper (with 25 authors from IPCC royalty) appears to be the definitive publication written by the authors of the AR6 chapter on climate sensitivity that will justify the conclusions expected to be reached in AR6. The Journal (Reviews of Geophysics) is an AGU publication that publishes only invited reviews. The have recently published similar articles with huge multi-author teams on sea level and aerosol forcing (same central estimate as AR5, narrower confidence interval).
“An assessment of Earth’s climate sensitivity using multiple lines of evidence”
https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2019RG000678
Abstract: We assess evidence relevant to Earth’s equilibrium climate sensitivity per doubling of atmospheric CO2, characterized by an effective sensitivity S . This evidence includes feedback process understanding, the historical climate record, and the paleoclimate record. An S value lower than 2 K is difficult to reconcile with any of the three lines of evidence. The amount of cooling during the Last Glacial Maximum provides strong evidence against values of S greater than 4.5 K. Other lines of evidence in combination also show that this is relatively unlikely. We use a Bayesian approach to produce a probability density (PDF) for S given all the evidence, including tests of robustness to difficult‐to‐quantify uncertainties and different priors. The 66% range is 2.6‐3.9 K for our Baseline calculation, and remains within 2.3‐4.5 K under the robustness tests; corresponding 5‐95% ranges are 2.3‐4.7 K, bounded by 2.0‐5.7 K (although such high‐confidence ranges should be regarded more cautiously). This indicates a stronger constraint on S than reported in past assessments, by lifting the low end of the range. This narrowing occurs because the three lines of evidence agree and are judged to be largely independent, and because of greater confidence in understanding feedback processes and in combining evidence. We identify promising avenues for further narrowing the range in S , in particular using comprehensive models and process understanding to address limitations in the traditional forcing‐feedback paradigm for interpreting past changes.
Climate scientists have been assessing multiple lines of evidence about climate sensitivity for a long time without being able to narrow the range of climate sensitivity with the historical climate record and the paleoclimate record. For example, the first two authors (Sherwood and Webb) were co-authors of a similar study published in 2016 consistent with AR5’s estimate.
A 2017 paper by three of the above 25 authors updated an earlier IPCC figures showing all published pdfs for ECS and TCR. They reported a range unchanged from AR5.
https://www.research-collection.ethz.ch/bitstream/handle/20.500.11850/197761/knutti17nat_public.pdf?sequence=2
So what’s new? Many AOGCMs being used for AR6 adjusted the parameters of that controlled clouds to better match what has been observed in “cloud-resolving models”. All past AOGCM’s did a fairly lousy job of reproducing the clouds we observe in the real world. For example, they produce far too few marine boundary layer clouds, the clouds that cool the planet most effectively. And the lack of marine boundary layer clouds strongly suggests that compensating errors must exist. So far, I have not seen any claims that the new models reproduce today’s clouds any more accurately than earlier ones. However, the climate sensitivity of many revised AOGCMs is dramatically higher than earlier models. About half of CMIP6 models have an ECS with a mean of about 2.7 +/- 0.5 K and the other half have mean of 4.7 +/- 0.8 K, with almost no models with an ECS near the mean for the entire group (3.7 K). See Figure 1a another publication with large author team likely composed of AR6 authors. They find that the differences in ECS are caused by differences in
https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2019GL085782
Abstract: Equilibrium climate sensitivity, the global surface temperature response to CO2 doubling, has been persistently uncertain. Recent consensus places it likely within 1.5–4.5 K. Global climate models (GCMs), which attempt to represent all relevant physical processes, provide the most direct means of estimating climate sensitivity via CO2 quadrupling experiments. Here we show that the closely related effective climate sensitivity has increased substantially in Coupled Model Intercomparison Project phase 6 (CMIP6), with values spanning 1.8–5.6 K across 27 GCMs and exceeding 4.5 K in 10 of them. This (statistically insignificant) increase is primarily due to stronger positive cloud feedbacks from decreasing extratropical low cloud coverage and albedo. Both of these are tied to the physical representation of clouds which in CMIP6 models lead to weaker responses of extratropical low cloud cover and water content to unforced variations in surface temperature. Establishing the plausibility of these higher sensitivity models is imperative given their implied societal ramifications.
These new high climate sensitivities are grossly inconsistent with the historical record of warming and forcing. As with CMIP5, many have aerosol forcing far more negative than the most likely forcing proposed by aerosol experts, but still within the fat tail of the pdf. The paper below uses global warming since 1970 as a constraint to reject models with a TCR greater than 2.2 K – about half of CMIP6 models.
https://esd.copernicus.org/preprints/esd-2019-86/
The paper you asked about determines ECS from feedback process understanding, the historical climate record, and the paleoclimate record. To the best of my knowledge, “feedback process understanding” comes from AOGCM’s, and cloud feedbacks in half of CMIP6 models are significantly more positive than in CMIP5. There probably haven’t been any major changes in the historical or paleoclimate records, but many people since AR5 have been asserting that natural variability/chaos has caused an anomalous pattern of observed warming since 1970 that was not seen in any of 100 runs of the MPI model. (When models and observations disagree, I’m old-fashioned and reject the models. Climate scientists can’t afford to reject models.)
So this is THE landmark paper that will (rightly or wrongly) likely narrow AR6’s estimate of climate sensitivity. The Science article below I encountered while writing is selling this work as the first progress since Charney in 1979. That article confirms some of my deductions above.
https://www.sciencemag.org/news/2020/07/after-40-years-researchers-finally-see-earths-climate-destiny-more-clearly
Thanks!
The paper (Sherwood 2020) Joeyblau was asking about – which claims a likely range for ECS of 2.6‐3.9 K is now available online. This paper appears to be a preview of what AR6 will assert.
S20: https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2019RG000678
Sherwood (2020) (hereafter S20) argues that we should differentiate between ECS that is relevant to the future and what they call ECS_hist obtained from the past warming using energy balance models. Their conclusions about ECS_hist are not consistent with earlier work including: Otto (2013)/AR5 authors, two papers by Lewis and Curry and many others. To understand how S20 arrives at much higher lower limit for ECS, it helps to compare their calculation with with that of Lewis and Curry (2018), hereafter LC18, for observed warming (dT), the forcing change (dF), the change in heat uptake by the ocean and melting ice (dN), and the forcing for doubled CO2 (F_2x). (Since these calculations start soon after the end of the LIA, I was surprised to learn that dN is less than the 0.8 W/m2 determined by ARGO.) The values for these parameters are shown in S20’s Table 5 next to those of LC18 and others.
Changed estimates for the aerosol indirect effect (aerosol-cloud interactions or aci) are responsible for a big reduction in dF in S20. These new values come from Bellouine (2020), B20, which appears to be a summary of information to be cited by aerosol chapter of AR6. dF is 1.83 W/m2 (90% ci 0.03-2.71 W/m2). In other words, B20 has expanded the possible range of negative forcing from aerosols so much that approximately 1 K of global warming may have been caused by ZERO net anthropogenic forcing at the lower confidence limit! It will be interesting to see if this nonsense survives peer review. Using aci from AR5, dF would have been 2.27 (1.45-2.98) W/m2, and the central estimate for ECS would be about 20% lower (2.5 K). Nevertheless the most likely value for aerosol forcing (the peak of the pdf) in B20 appears to be relatively unchanged at about -0.75 W/m2, AND essentially the same central estimate as used by LC18. Figure 8 of B20 shows that aerosol forcing more negative than -2.0 and -1.6 W/m2 are inconsistent (95% and 85% confidence) with historic warming! Stevens (2015) determined that an aerosol forcing less than -1.0 W/m2 would have produced no net forcing, and therefore no forced warming in the NH through 1950! Nevertheless, Stevens is a co-author of B20. AR5 assessed total aerosol forcing as -0.9 (-1.9 to -0.1) and LC18 revised this to -0.8 (-1.7 to -0.1) for a variety of sensible reasons.
B20: https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2019RG000660
Stevens (2015): https://journals.ametsoc.org/jcli/article/28/12/4794/34961
S20 chose to start where AOGCMs traditionally begin, 1861-1880, and finish with 2007-2018. LC18 computed changes over a period (from 1869-1882 to 2007-2016) carefully selected to match the influences of AMO, ENSO and volcanos, and avoid peak forcing from large volcanos. This change produces differences in dT and dF that may be an improvement over S20. S20’s dT (and therefore ECS) is 29% bigger because of combination of: a) the change in period, b) changing from SST to SAT, and c) using C-W infilled temperature in place of HadCRUT, a 10% increase. (SAT is the air temperature over land and ocean, a quantity we don’t measure over ocean, but is used when AOGCMs calculate GMST. It makes no sense to me for dT to be different when assessed using only SAT rather the traditional combination of SAT for land and SST for ocean or to believe that SAT is more accurate.) Recognizing the superiority of infilled temperature, L&C also used infilled temperature to obtain an ECS_hist of 1.66 K (1.1-2.7 K), but S20’s Table 5 didn’t abstract this calculation. Finally, for reasons I don’t understand, F_2x is 3.8 W/m2 in LC18 and 4.0 W/m2 in S20, causing another 5% increase in ECS_hist. (In my ignorance, I’ve been citing 3.45 W/m2, the mean F_2x of CMIP5 models in AR5.) The end result of all these changes raises ECS_hist from 1.50 K (1.0-2.5 K) in LC18 to 3.11 K (0.9-14.4 K) in S20.
For reasons I don’t fully understand, S20 then applies a 0-20 K prior to a Bayesian analysis of their pdf for ECS_hist, raising it to 4.3 K (2.0-16.1 K). This raises the lower limit for ECS_hist by more than 1 degK! In other parts of the paper, pdfs from paleoclimate data and feedback data are generated and combined with ECS_hist, eventually leading to 90% ci for ECS (not ECS_hist) of 2.3-4.5 K and a 70% “likely” ci of 2.6‐3.9 K.
However, the statistical games aren’t over. pdf’s from other sections of the paper on paleoclimate and feedbacks are used to abstract a revised pdf for total forcing change (dF) with a most likely value of about 2.4 W/m2 (instead of 1.8 W/m2). See Figure 10. (LC18 used nearly the same value, 2.52 W/m2, for a slightly shorter period. If I calculated correctly, using this and other central estimates, S20 would have obtained a central estimate for ECS_hist of 2.3 K.
S20 states that the lower limit for ECS without cooling from aerosols would be 1.2 K. If the non-controversial central estimate for the direct aerosol effect (-0.3 W/m2) were included, the lower limit without an aerosol indirect effect would rise to about 1.4 K. S20’s lower limit for ECS_hist was 0.9 K before applying an (unnecessary) 0-20 K prior. LC20 had a lower limit of 1.05 K. So the pdf for ECS_hist could logically start somewhere between 0.9 K and 1.4 K and rise to a central estimate of 2.3 K, only slightly higher than Otto (2013) and 40% higher than 1.66 K from LC18. This 40% is the result of many small differences, none of which is an obvious improvement over LC18 and some that appear inferior (IMO). So far, we have no good reason to believe an ECS of 1.5 K has been ruled out.
So, what was gained when S20 used an absurd pdf for aerosol forcing and then corrected the resulting absurd pdf for ECS_hist using other constraints? The pdfs for paleoclimate and feedback would have been pulled lower when combined with an ECS_hist centered on 2.3 K. Instead, those other pdf’s were “wasted” correcting an absurd pdf for ECS_hist with an upper limit near 15 K. Possible lower limits from 0.9 to 1.4 K disappeared in the process, allowing climate scientists to claim that they had – after a half-century – finally raised the lower limit for ECS.
(S20 goes on to argue that ECS is higher than ECS_hist due to “pattern effects” and folds this factor in their final range. I haven’t mastered this part of their analysis. This comment is limited to ECS_hist.)
I would love to ask one question. If I load modtran with default presets, it will give me 298.52W/m2 in emissions TOA, with a simulated tropical climate. Now if I double CO2 emissions that will sink to 295.191W/m2. In order to get emissions back up to 298.52W/m2 I have to add a “temperature offset”, which turns out to be 0.76K (try yourself).
If modtran is satisfied with 0.76K, how do you argue 1.1K or 1.2K to 2xCO2?
http://climatemodels.uchicago.edu/modtran/
That’s because you left the water vapor setting at water vapor pressure constant. That means the relative humidity decreases. The standard assumption is that relative humidity is constant with temperature so that water vapor pressure increases with temperature. Water vapor is a greenhouse gas as well. If you set hold relative humidity constant, the required offset is 1.21C.
Well, I am afraid you are only naming the other side of the problem. WITH vapor feedback due to keeping relative humidity constant (thus increasing vapor), the offset is 1.21K, meaning vapor feedback will not double the effect of 2xCO2 either. Rather it is just adding 59%. Without, for only 2xCO2, the effect is just 0.76K, as before.
And that is despite we a very “favorable” preconditions here.
1. The model suggests surface emissivity = 1, which will exaggerate the result.
2. it is true for a “tropical climate”. All other other climates show less sensitivity.
3. That is without clouds. Once you add any cloud scenario the number drops sharply.
And that is despite we have..
Clouds also decrease incoming radiative flux by reflection. There’s still some question about whether clouds in general are a positive or negative feedback in the real world and whether cloud cover remains a constant in a world with rising water vapor pressure. They’re a positive feedback in climate models, but that is a choice, although I think they claim it’s an emergent property.
Btw, whoever said that water vapor would double the effect of CO2? Real question. Your original post asked for how you get to 1.1-1.2C offset from doubling CO2. I told you. Stop moving the goalposts if you want anyone to engage with you in the future.
Actually, if you look at the raw data, the surface emissivity is 0.98, not 1.00. The version of MODTRAN on the web is also several generations old and is nowhere near as sophisticated as a line-by-line radiative transfer model. It’s still more complex than what they use in climate models, though.
No one is “moving the goalposts” except you, so stop playing stupid! The sod article above is pretty clear for anyone who mastered reading. With “some feedback” he means a constant relative humidity btw.
“So, our “no feedback” result is 1.2ºC, and our slightly more realistic “some feedback” result is 2.4ºC”
And I point out to the fact that modtran, with the named “maximising” parameters, has it at 0.76K (instead of 1.2K) for 2xCO2, and at 1.21K (instead of 2.4K) including feedback. Don’t kill the messenger.
One thing I have discovered with Modtran is that the uchicago interface is pretty stupid and will allow any combination of input parameters which can lead to some quite confusing outputs. I found I had to do some extra work work to make sure I was using sensible parameters
Modtran certainly has its flaws. The deviation I name however is not a flaw, I know pretty well where it comes from. Actually it is a necessary consequence of a huge blunder in the whole “GHE” theory (no ever seemed to notice!?), that carries well into ECS estimates. So modtran is largely right, but ECS estimates are blantently wrong.
John Doe: The biggest problem with MODTRAN is the only about 16% of outgoing radiation passes through the clear tropical sky. Roughly 60% of the sky is cloudy and another 60% lies outside the tropics. Clouds and regional variations were taken into account and line-by-line radiative transfer calculations were used by Myhre 1997 and 1998. He found that doubling CO2 from 300 to 600 ppm decreased OLR at the tropopause by 3.7 W/m2 (and would be expected to do the same at the TOA once the stratosphere adjusted (cooled) in response to the smaller flux of radiation entering from below.) With no feedbacks, it would take 1.15 K of warming to increase OLR by 3.7 W/m2 and restore a steady state. If relative humidity remains constant, it takes rough twice as much warming to return to steady state, but lapse rate feedback cause by rising humidity negates about half of the effect water vapor feedback. That gets you to about 1.8 degK needed to restore steady state and about 2.0 degK if you include less reflection of SWR by snow and ice. I’d guess that most knowledgeable skeptics think these are reasonable values. Cloud feedback is the big unknown.
Allegations that there is a huge unspecified blunder in the above summary usually haven’t served scrutiny, but you are welcome to explain.
Today, no one is trying to improve Myhre’s figure of 3.7 W/m2 using regional average temperature data and cloudiness. AOGCM’s do these radiative transfer calculations for about a million grid cells with changing temperature, lapse rate and clouds. Those models come up with a radiative forcing of 2XCO2 ranging from 2.2 to 4.7 W/m2 with a mean of 3.45 W/m2 (IIRC), but these calculations are modeling the atmosphere, not doing calculations based on observations of the atmosphere (as Myhre did). And AOGCM’s don’t do such a great job of reproducing the atmosphere we observe today.
Enjoy..
https://pubpeer.com/publications/728B93557E2BDC0771B38B30C9FA7A#null
John Doe: The work on “Cloud Radiative Forcing” that is being discussed at the link you provided is discussed by our host at this link and the post that followed addressing questions. .
https://scienceofdoom.com/2010/05/30/clouds-and-water-vapor-part-one/
The original 1989 Science paper on this project was and is behind a paywall, so it is difficult to interpret. Above, our host posted information and Figures from a book chapter reviewing this subject that was not behind a paywall. The link he provided no longer works, but the book chapter can be found here.
Click to access The_radiative_forcing_due_to_clouds_and_water_vapor_FCMTheRadiativeForcingDuetoCloudsandWaterVapor.pdf
The critique at your link complains about the difficulty of determining causality, which demonstrates that the author of the critique doesn’t understand the paper. This is a paper about observations of the planet. Before the ERBE mission to observe the planet flew, we knew that clouds reflect incoming sunlight, absorb LWR emitted by the ground before it can escape to space and emit LWR to space. The former phenomena cools the planet, while the latter warms the planet because cloud tops emit less LWR than the warmer surface below. How much less LWR escapes to space when clouds are present depends on the height of those cloud tops because temperature drops with altitude in the troposphere. Before ERBE, no one knew for sure whether globally clouds cooled or warmed the planet. So this is mostly a paper about what we observe rather than causality.
Then the paper provides a much more useful way of quantifying the greenhouse effect (G). Normally the GHE is reported as temperature difference between our planet with and without GHGs, but we can use many different models to calculate what the temperature of our planet would be without GHGs and that causes lots of controversy. However, we know how much thermal radiation our planet would emit without GHGs in the atmosphere (e*o*T^4), and we can quantify G as the reduction in outgoing LWR caused by the atmosphere including clouds And they get into how G varies with temperature when temperature is changed by the seasons (not rising GHGs). There are better observations and simpier math that described this subject in newer papers.
The paywall issue can easily be fixed..
https://www.researchgate.net/publication/6060960_Cloud-Radiative_Forcing_and_Climate_Results_from_The_Earth_Radiation_Budget_Experiment
Nothing in your response and nothing in the paper, the book or the SOD article adresses the problem described. No surprise there, as this is the big “blind spot” of climate science. And of course you do not need to deal with it, there are no obligations.
However, the problem is not going to disappear by ignoring or denying it. And of course it is all about allocating causation. Now the book you linked makes a small but reasonable iteration..
“Clouds reduce the OLR further and enhance the greenhouse effect. This enhancement of the greenhouse effect by clouds can beobtained by letting:Cl=Fc−F”
Sure, you could put it like this. You take the GHGe as given and say clouds enhance it. You could say the same about a doubling of CO2. Going from 400ppm to 800ppm will enhance its GHE by 3.7W/m2 (that is from ~30W/m2 to 33.7W/m2 for CO2 alone).
But is it really true that additional 400ppm ONLY cause a GHE of 3.7, while the first 400ppm cause 30W/m2? Or is it more accurate that both 400ppm parts will cause 30W/m2, of which 26.3W/m2 will be overlapped so that they will not make any difference? The second batch of 400ppm is no less potent as the first batch, it is only due to the overlapping that we see a diminishing effect with more CO2.
And that is exactly the problem with have the cloud radiative effect from the start. Those 30W/m2 of CRE are ONLY what clouds add, not the actual potency they have. The CRE in absence of GHGs is more like 80W/m2.