By Andy May
This post is inspired by an old post on the CERES cloud data by Willis Eschenbach that I’ve read and re-read a lot, “Estimating Cloud Feedback Using CERES Data.” The reason for my interest is I had trouble understanding it, but it looked fascinating because Willis was comparing CERES measured cloud data to IPCC modeled cloud feedback. I love obscure, back-alley comparisons of models and data. They tend to show model weakness. I learned this as a petrophysical modeler.
Willis wrote the post as a response to a paper by Paulo Ceppi and colleagues on cloud feedback in global climate models (Ceppi, Brient, Zelinka, & Hartmann, 2017). We’ll refer to the paper as Ceppi17. I took the time the last few days to understand Willis’ post and Ceppi’s paper and this is what I figured out, let me know what you think in the comments.
In Ceppi17, N = F + λΔT. N is the energy flux imbalance at the top of the atmosphere, F is a forcing, in W/m2 due to a sudden increase in greenhouse gases. The hypothetical situation used in the paper was a quadrupling of the CO2 instantly, relative to preindustrial conditions. Then they calculated a hypothetical F. “λ” is the cloud feedback and ΔT is the total global temperature change required to regain equilibrium, or an N of zero. Their feedback numbers cannot be duplicated with data due to the implausible scenario. Here are two more versions of the equation for reference.
ΔT = (N-F)/ λ or λ = (N-F)/ ΔT
What is N? N is a force imbalance between incoming (or downward) radiation and outgoing radiation at the top of atmosphere, which we will define as the CERES satellites. N is positive if the downward force is larger (warming) and negative if the outgoing radiation is larger (cooling). The Earth is at equilibrium if the feedback, N and F are zero. Positive feedback (λ) leads to warming and a larger imbalance (N). The higher the feedback, the greater the warming. If the feedback is negative, cooling or slower warming is the result.
“CRE” is the cloud-radiative effect, or the difference between the clear-sky and all sky radiative flux at the satellite (TOA). Clouds reflect incoming shortwave solar radiation (SW), so more SW travels up to the satellite in the presence of clouds, on average the increase is about -45 W/m2. This is a negative number because it means more radiation is leaving Earth, a cooling effect. Clouds also block some outgoing longwave infrared radiation (LW) emitted by Earth’s surface, on average about 27 W/m2, a positive number, since it is energy retained by Earth or less energy reaching the satellite, a warming influence. The difference is -18 W/m2, which means, overall, clouds cool the earth.
One would think that the more clouds, the more rapidly the earth would cool, but it isn’t that simple. Some clouds, especially low level clouds and cumulus clouds tend to reflect more energy during the day than they trap during the night. High level clouds, like cirrus, tend to allow solar SW through and trap a lot of upwelling LW, thus they have a warming effect. So cloud type matters.
Figure 1 is a map of the average TOA (top of the atmosphere) cloud-radiative imbalance (CRE) at the CERES satellites. The blues are a negative energy imbalance or a cooling CRE. The map is an average of CERES monthly data from 2001 through 2019. The CERES variable mapped is “toa_cre_net_mon” or the “Top of The Atmosphere Cloud Radiative Effects Net Flux.” The effect is negative (or cooling) everywhere, except over deserts and the polar land regions. These are areas where the clouds tend to trap surface and lower atmosphere infrared and simultaneously allow solar shortwave radiation through to the surface, this combination creates a positive CRE and strong warming.
The point where the cloud warming and cooling effects meet in the Figure 1 color scale is where the lightest blue meets the light yellow. Right at zero is where the incoming energy equals the outgoing energy, with respect to clouds. Thus, except for the Sahara, the Middle East, Western China, parts of Southeast Asia, Indonesia, Northern Australia, the Southwestern U.S., and Mexico, clouds cool the earth. The darker areas in Figure 1 have more persistent clouds.
Figure 2 has lighter colors for clouds and darker colors for clear skies. Thus the lighter streak, near the equator, in both the Pacific and the Atlantic, is white in Figure 2, opposite of Figure 1. This the Intertropical Convergence Zone (ITCZ) where the trade winds of the Northern Hemisphere and Southern Hemisphere converge. This is where evaporation of seawater is at a maximum. Water vapor is less dense than dry air so it is a zone of rapidly rising humid air and frequent rain and thunderstorms. It is almost always cloudy. The ITCZ follows the Sun’s zenith point and the cooling effect of the clouds in this zone is very high.
The maximum cloud cooling, or the most negative CRE values, are in the small white spots in the middle of the black spots off of southern Peru and in southeastern China, north of Vietnam. These CRE values are very negative (extremely cooling) and off-scale. Figure 1 correlates reasonably well with the cloud fraction in Figure 2, or the lighter colors in Figure 3, with the exception of the polar ice caps.
Figure 3 is the NASA blue marble with ice and clouds shown in a Mercator projection. Notice the similarity with Figure 2, except at the poles.
Figure 4 shows the same data, the CERES EBAF 4.1 variable “toa_cre_net_mon,” as yearly global averages. EBAF means energy balanced and filled. As Norman Loeb and colleagues (NASA Langley Research Center) explain, Earth’s energy imbalance is so tiny, between 0.5 and 1 W/m2, that it is only 0.15% of the total incoming and outgoing radiation. Thus, the number we are looking for is the difference between two large numbers and barely above the uncertainty in the satellite measurements.
The calibration uncertainty in the CERES SW measurement is 1% and 0.75% for the LW. Thus the outgoing LW is only known to ±2 W/m2. There are many other sources of error, and as Loeb, et al. explain, the net imbalance from the standard CERES data products is only ~4.3 W/m2, not much larger than the expected error. Due to the coarse resolution of the CERES instrument, there are a lot of missing grid cells in the one-by-one degree latitude and longitude grid used to make the maps in Figures 1 and 2. To get around these problems Loeb and colleagues use a complex algorithm to fill in missing values and adjust the SW and LW TOA fluxes, within their uncertainty ranges, to remove inconsistencies between the global net TOA energy flux and the heat storage in the earth-atmosphere system (Loeb, et al., 2018).
The CRE, or cloud radiative imbalance value varies a lot from year to year, the average value over the 19 years is -19.1 W/m2, very close to the Ceppi, et al. value of -18 W/ m2 (Ceppi, Brient, Zelinka, & Hartmann, 2017). This suggests that total cloud cover is the main factor, it is shown below in Figure 5. As we would expect, as the cloud fraction goes down, the cooling effect decreases and the CRE becomes less negative. As the cloud fraction goes up the cooling effect increases.
In Ceppi17 a more positive feedback parameter (λ) implies warming. Since they are working with models, they can compute λ by dividing the computed forcing required to counter the original forced energy imbalance, due to clouds, by the resulting temperature change (ΔT). Figure 6 shows Ceppi17’s feedback globally due to clouds.
The units are W/m2/K, where K (Kelvin) is degrees C of warming or cooling due to clouds over the time it takes to reach equilibrium. Figure 6 is cloud feedback and not the same as CRE, but according to Ceppi17, cloud feeback tends to be positive and it suggests that clouds, over the long term, warm Earth and do not cool it. This made Willis question the whole paper. As he points out, Figure 6 is a plot of model output and Figure 1 is data. The data in Figure 1 is massaged and it is close to the edge in terms of uncertainty, but it is data.
Ceppi17 is lucky we cannot derive their feedback parameter from real data, because if we could, I suspect the map would look very different from Figure 6. For example, one of the places where clouds cool the surface the most is offshore of Peru, how does that turn into an area with a positive feedback? The other is southeastern China, OK, we get a little blue there, but nothing like what the actual data shows us. The very cloudy ITCZ is a very hot area in Figure 6, how do you do that?
I agree with Willis, this whole idea that clouds are a net positive (warming) feedback, makes no sense. The worst of it is that nearly every model uses a positive cloud feedback. Cloud feedback is the largest component of the model-computed ECS (the temperature sensitivity due to a doubling of the CO2 concentration) that the IPCC prefers. As we know, clouds cannot be modeled, and must be parameterized (the fancy modeling term for “assumed”). As Steve Koonin’s upcoming book, Unsettled, reports, scientists from the Max Planck Institute tuned their climate model by targeting an ECS of about 3°C by adjusting their cloud feedbacks. He adds “talk about cooking the books.” (Koonin, 2021, p. 93).
Ceppi17 reports that cloud feedback is, “by far, the largest source of intermodel spread in equilibrium climate sensitivity (ECS).” They also point out that cloud feedback is strongly correlated with model derived ECS and supply us with the data, it is plotted in Figure 7.
Oops! Clouds cannot be modeled, models assume their clouds have a warming effect, CERES says clouds have a net cooling effect, a large net cooling effect of -18 W/m2. The models say the entire human influence on climate since the beginning of the industrial era is 2.3 (1.1 to 3.3) W/m2 (IPCC, 2013, p. 661), which puts the cloud impact of -18 W/m2 into perspective. Notice that the variability in Figure 4 is larger than 2.3 W/m2. How much of the ECS from models is due to their assumption that clouds are net warming? How much is due to their assumption that ECS is 3 W/m2? So many questions.
Willis Eschenbach kindly reviewed this post for me and provided valuable input.
Works Cited
Ceppi, P., Brient, F., Zelinka, M., & Hartmann, D. (2017, July). Cloud feedback mechanisms and their representation in global climate models. WIRES Climate Change, 8(4). Retrieved from https://onlinelibrary.wiley.com/doi/full/10.1002/wcc.465
IPCC. (2013). In T. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S. Allen, J. Boschung, . . . P. Midgley, Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge: Cambridge University Press. Retrieved from https://www.ipcc.ch/pdf/assessment-report/ar5/wg1/WG1AR5_SPM_FINAL.pdf
Koonin, S. E. (2021). Unsettled: What Climate Science Tells us, What it doesn’t, and why it matters. Dallas, Texas, USA: BenBella. Retrieved from https://www.amazon.com/dp/B08JQKQGD5/ref=dp-kindle-redirect?_encoding=UTF8&btkr=1
Loeb, N. G., Doelling, D., Wang, H., Su, W., Nguyen, C., Corbett, J., & Liang, L. (2018). Clouds and the Earth’s Radiant Energy System (CERES) Energy Balanced and Filled (EBAF) Top-of-Atmosphere (TOA) Edition-4.0 Data Product. Journal of Climate, 31(2). Retrieved from https://journals.ametsoc.org/view/journals/clim/31/2/jcli-d-17-0208.1.xml
I followed the posting right up until it jumped off into models. I agree, the models make no sense.
I did come up with an idea. I am wondering if it possible to separate the world into 4 time divisions – cloud feedback at noon (including +/- 3 hours), at 6PM, at midnight, and at 6AM. If clouds are really acting as heat transports one should see this represented in the percent of clouds over the Earth at various times of day. I hate this idea of a Global Temperature and a Global Cloud Cover as it seems to just simplify everything into “a miracle occurs here”. It’s like averaging a man and a woman and determining the average person is 48% male and 52% woman.
Just splitting it into daytime and nighttime gives one a much more informed view of cloud radiative effect…I like your idea…I think 6 zones would better….with noon and midnight +/- 2 hours being the most instructive, the others being more obviously intermediate cases than with 4 zones…make little Trenberth type diagram for each…finally add them up….etc.
Higher sea surface temperatures cause a decline in low cloud cover. So in a simple thermodynamic model, it looks like a positive feedback. But in the real world, ENSO and the AMO act as negative feedbacks to changes in the solar wind strength, and the cloud cover changes which they cause, amplify their negative feedback effect.
Such that stronger solar wind states in the 1970’s drove colder ocean phases and increased low cloud cover ,driving global cooling, and post 1995 weaker solar wind states have driven warmer ocean phases and reduced low cloud cover, driving global warming.
https://www.linkedin.com/pulse/association-between-sunspot-cycles-amo-ulric-lyons/
“Some clouds, especially low level clouds and cumulus clouds tend to reflect more energy during the day than they trap during the night. High level clouds, like cirrus, tend to allow solar SW through and trap a lot of upwelling LW, thus they have a warming effect. So cloud type matters.”
Surely at higher latitudes, winter low clouds would cause a net warming? What about the solar near infrared? that is absorbed by any clouds.
That’s not correct, most rainfall is at equator. Rain falls from clouds. SST is high there. SST is a little higher on either side of equator, but because of high cloud cover….
All clouds, including low ones, (actually especially low ones, but that will cause heads to explode) cause warming at night. All clouds reflect sunshine back into space during the day, especially thick ones, which are usually low. High thin clouds let more sunshine through, which is why they are assigned more “warming” effect. You are correct for the far North because there isn’t as much sunlight to reflect into outer space due to cos of zenith angle of incident sunlight, yet good IR transmission between the cloud bottom and the surface.
Yes, but by the same token, there cannot BE much downward radiation from such clouds, because they are not heated well by the sun, well, ’cause, you know, “cos of zenith angle”?
I have to ask: Has anyone actually measured this “down radiation” from clouds? In my universe, insulation explains all the heat under night time clouds, by disrupting convection, and it avoids the pseudophantasmagorigal reasonings thought up by those enforcing the ‘forcing’ paradigm. I climb up a high tower, and my eyes observe the complex stratification of the atmosphere. I read climastrology, and the unsubstantiated formulations and asumptionating makes my eyes water, crying for the total lack of appreciation modellers have for the varied beauty of real life existence.
Those poor buggers really need to get a life!
Higher SST’s do reduce low cloud cover, except for parts of the central tropics, and in the Arctic.
In the blog writing of mine, is a very illustrative figure (number 3), in which way the global temperature and the absolute water amount (TPW) have been developed from 1979 onward. The positive water feedback works well during the ENSO events but strange enough, at the same time the long-term feedback is going in another direction. You will not find these trends in any figures of climate establishment. You can figure out, why.
https://www.climatexam.com/2018
El Nino is a negative feedback, and the water vapour amplifies it. The WV trend is also a negative feedback because the warm AMO phase is a negative feedback to weaker indirect solar forcing, it’s warming since 1995 is also self amplified by the increase in water which it drives. The warm AMO phase reduces low cloud cover and increases surface wind speeds over the ocean, so more water vapour is generated, and the upper OHC also increases.
Interesting that over oceans, clouds always are negative feedback to varying extents
Over land, deserts and other arid lands clouds have positive feedback – however these are areas where clouds are atypical and I suspect most of the warming there is due to slowdown in nighttime cooling. The most positive feedback is over the Sahel (arid but not desert), Greenland and Antarctica. Some searching on “greenland climate” .. ~2 inches precip per year, “Greenland’s climate is … characterized by low humidity“, etc.
So semi-arid land areas, both hot and frigid, are where the strongest positive feedback exists.
I wonder why such large negative feedbacks are indicated in China, NW Pacific and SE Pacific off Peru/Chile coast. Even more so than the ITCG, where I would have expected to see the most negative feedbacks. And near neutral in the big Amazon and Congo rain forests was unexpected – but maybe due to some positive feedback in dry season offsetting negative feedback in wet season?
As a comment, an old flight instructor and professional pilot with over 38000 hours as pilot in command would sometimes relate his student’s observational experiences to those of us sober enough to listen. He had many students, some of whom were really talented aviators. Like sailors, they could live and die by misreading cloud formations and energy content in those wispy fluffy bundles of watery goodness the rest of us see as clouds. Willis has shared this type of observation from the ocean surface. Imagine a vaunted aviator taught by this older pilot. One of the most repeated reminder’s to his students was there are very few old bold pilots.
In the 1960’s with the advent of the turbojet engine many of these aviators were being told that t-cell anvil heads could only very seldom reach 50k ft.(maybe north of the 40th parrallel)
Some of these bright young pilots were relating to their mentor, how after leaving Beale AFB in sunny CA and refueling over the Pacific at 35000ft they would return to their cruising altitude of 60,000ft. Only about 11 miles above Willis but pretty much a world away in terms of perspective. A noticeable curvature at the horizon 600 miles distant a dark blue color to the sky not an almost flat horizon 6-12 miles away in a beautifal light blue sky like for Willis.
But like Willis they looked for clouds.
After several years of crossing the ITCZ numerous times they would become highly indignant when being informed by Air Force meterologists that anvil tops were restricted to at most 65K ft in the ITCZ. They would explain to their mentor that sometimes going up to the top of their aircraft’s operational altitude at 75 to 82k ft (they got pretty vague on actual altimeter readings) they would still be somewhat below the anvil tops. That was usually when the old, bold saw was drug-out and played. When they were so mission driven to attempt to traverse the clouds the amount of available but chaotic energy still in those clouds seemed to convince those younger aviators that their mentor was a very wise OLD pilot, if they got to share a beverage with him after that particular learning moment.
That lack of observational wisdom is a lot of what holds progress back in modelling our atmosphere, because modelling usually is under a dry roof in climate controlled quarters firmly anchores by gravity to terra firma like where most meterologist congregate.
Andy May, Willis Eschenbach, and that old flight instructor are three of my heros that at 2 years old might have stuck their fingers into electric sockets and survived a hazardous experience to learn there is always dangers in exploration and that it can be shocking to make bold assumptions. Thank you all for keeping up the good fight doing such good work and keeping an eye out for clouds.
Excellent point. Field work should be required for a degree in a physical science.
When the choice is between “the data” and a “lying leftist activist pseudo-scientist”… I never chose the leftist.
Clouds block sun and sunshine cannot “go around” the clouds. Heat can go around the clouds.
I question the entire feedback model proposed by the warmistas. Since the geologic record shows that most of the time the earth’s climate has been warmer than current, and the climate did not run away, obviously the feedback model only works over a very short temperature range if it applies at all. Since the fundamental prediction of the warm, humid layer in the troposphere predicted by the proponents of these models has been experimentally shown to not have developed, their entire model has failed its first and easiest prediction.
This strikes me as a very important result. If you run Fig 7 x axis out to -18 w/m2, what value do you get for ECS?
Looks to me like ECS goes negative.
ECS might be negative, in reality. But, positive or negative, it is a small number.
What correlation exists for cloud formation in relation to the level of cosmic ray flux? To that question I now would add formation of clouds at different levels? I could presume cosmic rays trigger far more upper level cloud.
Ewin,
I refer you to this report:
Svensmark, H. (2019). Force Majeure, The Sun’s role in climate change. GWPF. Retrieved from https://www.thegwpf.org/content/uploads/2019/03/SvensmarkSolar2019-1.pdf
They find correlation with low-level clouds.
Cloud cover? Is that it? I remember going to observe a lunar eclipse, the sun was setting on my back and I watched how the moon in front of me pulled clouds towards it, there was a strange warping effect on one of the images, it looked to me like a gravitational effect, but is that a thing? are gravitational forces accepted in science today? When you’re all arguing over a few tenths of a degree in a planetary anomaly or w/m2 do you really understand what’s actually going on?