Guest Post by Willis Eschenbach
I have theorized that the reflective nature of the tropical clouds, in particular those of the inter-tropical convergence zone (ITCZ) just above the equator, functions as the “throttle” on the global climate engine. We’re all familiar with what a throttle does, because the gas pedal on your car controls the throttle. The throttle on any heat engine controls the running conditions by limiting (throttling) the amount of incoming energy.
Similarly, in the climate heat engine, the throttle is the tropical albedo (reflectivity). The tropical albedo controls how much incoming solar energy is rejected back to space at the hot end of the heat engine. In other words, the albedo throttles the incoming energy to control the entire system.
I have further said that the tropical albedo is a threshold-based and extremely non-linear function of the temperature. So I thought I’d use the CERES satellite data to take a look at how strong this climate throttle is in watts per square metre (W/m2), and exactly where the throttle is located. If such a throttle exists, one of its characteristic features would be that the amount of solar energy reflected must increase with increasing temperature. Figure 1 shows the results of that analysis.
Figure 1. Average change in reflected solar from a 1° increase in surface temperature. Red areas show greater reflection with increasing temperature. The change in reflected energy is calculated on a per-gridcell basis as the change in albedo per 1° temperature increase for that gridcell, times the average solar radiation for that gridcell. Gray line shows zero change in albedo with temperature. Dotted lines show the tropics (23.45°N/S) and the Arctic/Antarctic circles (66.55°N/S).
Clearly, then, such a throttle mechanism exists. It is also where we would expect to find it, located near the Equator where the maximum energy is entering the system. On average, the throttle operates in the areas enclosed by the gray line. I was surprised by the strength of the mechanism, however. There are large areas (red) where a one degree C warming in temperature increases the solar reflection by 10 W/m2 or more. Obviously, this thermostatically controlled throttle would be a factor in explaining the observations of a hard upper open ocean temperature of about 30°C.
The throttle mechanism is operating over much of the tropical oceans and even some parts of the tropical land. It is strongest in the ITCZ, which runs below the Equator in the Indian Ocean and over Africa, and above the Equator in the Pacific and Atlantic.
Next, it is worth noting that overall the effect of temperature on solar reflections is about zero (global area-weighted average is -1.5 W/m2 per degree, which is smaller than the uncertainty in the data). In addition, large areas of both the land and the ocean in the extra-tropics are quite similar, in that they are all just slightly negative (light orange). This is another indication that we have a thermoregulatory system at work. Since over much of the planetary surface the albedo is relatively insensitive to changes in temperature, small changes in temperature in the tropics can have a large effect on the amount of energy that is entering the system. Figure 2 shows the relationship (land only) between absolute temperature in °C, and the change in reflected energy per degree of warming.
Figure 2. Change in reflected solar (W/m2 per °C) versus absolute surface temperature (°C) over the land. Note that where the annual temperature averages below freezing (0°C), there is little variation in surface reflection with temperature. From freezing to about 20°C, the amount reflected is generally dropping as temperatures increase. Above about 20°C, there are two kinds of responses—sizeable increases or sizeable decreases in reflected solar with temperature.
Next, over the oceans the areas near the poles show the reverse of the behavior in the tropics. While the tropical albedo changes cool the tropics, near the poles as the surface warms, the albedo and the reflected sunshine decreases with increasing temperatures.
Figure 3. Change in reflected solar (W/m2 per °C) versus absolute surface temperature (°C) over the ocean, annual averages. Where the annual temperature averages near freezing, there is strong negative variation in surface reflection with temperature. From freezing to about 20°C, the variation is stable and slightly negative. Above about 20°C, there are two kinds of responses—sizeable increases or sizeable decreases in reflected solar with temperature, up to the hard limit at 30°C
What this means is that in addition to limiting overall energy input to the entire system, the temperature-related albedo-mediated changes in reflected sunlight tend to make the tropics cooler, and the poles warmer, than they would be otherwise. Clearly this would tend to limit the overall temperature swings of the planet.
Finally, the use of monthly averages obscures an important point, which is that the changes in tropical albedo occur on the time scale of minutes, not months. And on a daily scale, there is no overall 10 W/m2 per degree of temperature change. Instead, up to a certain time of day there are no clouds, and the full energy of the sun is entering the system. During that time, there is basically no change in tropical albedo with increasing temperature.
Then, on average around 11 am, within a half hour or so the albedo takes a huge jump as the cumulus clouds emerge and form a fully-developed cumulus regime. This makes a step change in the albedo, and can even drive the temperature down despite increasing solar forcing, as I showed here, here, here, here, and here
From this we see that the thermal regulation of tropical albedo is occurring via changes in the time of the daily onset and the strength of the cumulus/cumulonimbus regime. The hotter the surface on that day, the earlier the cumulus and cumulonimbus clouds will form, and the more of them there will be. This reduces the amount of energy entering the system by hundreds of watts per square metre. And on the other hand, during cooler days, cumulus form later in the day, cumulonimbus may not form at all, and there are fewer clouds. This increases the energy entering the system by hundreds of W/m2.
I bring this up to emphasize that the system is not applying an average throttle of e.g. 10 W/m2 over the average area where the throttle operates.
Instead, it is applying a much larger throttle, of a couple hundred watts/square metre, but it is only applying the throttle as and where it is needed in order to cool down local hot-spots, or to warm up local cold spots. As a result, the averages are misleading.
The final reason that it is important to understand that the albedo changes are HOURLY changes, not monthly average changes, is that what rules the system are instantaneous conditions controlling cloud emergence, not average conditions. Clouds do not form based on how much forcing there is, whether the forcing is from solar or CO2 or volcanoes. They form only when the temperatures are high enough.
And this means that things won’t change much if the forcing changes … because the cloud emergence thresholds are temperature-based, and not forcing-based.
I hold that this immediate response is the main reason that it is so hard to find e.g. a solar signal in the temperature record—because the thermoregulation is temperature based, not forcing based, and thus operates regardless of changes in forcing.
This is also the reason that volcanoes make so little difference in the global temperature—because the system responds immediately to cooling temperatures by reducing albedo, opening the thermostatically controlled-throttle to allow the entry of hundreds of extra W/m2 to counteract the drop in temperature.
There is plenty more to mine from the CERES dataset, and although I’ve mined some of it, I still haven’t done lots of things with it—an analysis of the efficiency of the climate heat engine, for example. However, I think this clear demonstration of the existence of a temperature-regulated throttle controlling the amount of energy entering the climate system is important enough to merit a post on its own.
Best regards to all on a sunny December day,
w.
Duster,
That places clouds into the “micro” scale was far as weather is concerned, effectively would force modelers into the position of attempting to derive climate from weather, rather than the other way around.
I should have thought that if we (OK Willis) were to take a long term data average of this weather he would arrive at a real world data based climate model input for the thermoregulatory effect (TRE). The climatic input from this effect should straightforwardly slot into GCM models which I believe Willis has previously shown to amount to a one line equation which can be executed itteratively on an abbacus by a team of low cost chimpanzees.
Perhaps Willis could repeat his calculations using the chimpanzee team on his simplified but empirically accurate version of the GCM models with the TRE included? Then by comparing the new output with real world temperature data we can see just how important the TRE is in climate modelling.
If it transpires that Willis is able to show that his now modified GCMs are more accurate than those used by the IPCC to justify their fat salaries we might reasonably replace the IPCC with the chimpanzee team and organise a “spot the difference” competition.
I live near Singapore, so it is interesting to note the transition to a negative reflective effect by temperature rise in this area, and generally around any equatorial landmass. Here we tend to not see the precise timing of cloud build up around noon, we can have predominantly cloudy days for weeks on end.
I suspect this may primarily because of the airflow changes around the land masses.
A most beautiful rendition of wind currents on the earth:
http://earth.nullschool.net/#current/wind/isobaric/1000hPa/orthographic=110.77,2.94,256
It reveals the huge effect the landmasses have on airflows around the earth.
Double click to zoom in, shift and double click to zoom out, drag to rotate, click to show wind speed direction at any point.
I find the comment Any speculation by way of heat engine analogies would have to be for open systems, not closed systems, such as the earth’s climate system.. The theoretical Carnot Cycle describes a closed cycle heat engine, with an externally heated hot end and an externally cooled cold end – quite a good analogy for the Earth described as a heat engine, I would have thought.
Re the above, Sorry a bit off topic; But interesting.
Re temperatures at the north pole vs the south pole.
While not discounting ocean currents, perhaps it is as much a function of the airflow disrupting effects of all the land mass surrounding the north pole as it is that the Arctic region is a sea rather than a land mass.
North pole view of airflows:
http://earth.nullschool.net/#current/wind/isobaric/1000hPa/orthographic=100.45,90.13,256
South pole view of airflows:
http://earth.nullschool.net/#current/wind/isobaric/1000hPa/orthographic=97.64,-83.19,256
Perhaps what is so amazing is the stability of the Earth’s climate over the past 500 million years with the occasional near extinction event obviously excluded.
There obviously has to be some kind of self-regulation mechanism for our planet’s temperature and climate.
After all:
1. We have been steadily moving further away from the Sun.
2. The Sun has been steadily increasing its energy output. Also, we can measure the change in energy output over the known 11 year cycle, but we do not know if there are more much longer cycles.
3. The continents move around the face of our planet having a long term impact on climate.
4. Our planet has an elliptical orbit.
Has what Willis describes been accurately modelled? My five bucks says it: i) either has not been considered, ii) is too complex to model, or iii) produces inconvenient results of the no future funding variety, i.e. No CAGW.
Well, that’s a model for a thermostat, which is the default setting within one geological era.
You also need a model of how to over-ride the thermostat, creating either very warm periods or ice ages with temperatures ten degrees or more hotter or colder than currently.
A rampaging out of control fire removing the cloud cover over what is currently equatorial forest?
A freak set of storms covering an abnormally large NH area with long-term snow cover??
A meteorite??
More subtle effects??
Anyone come up with any practical models for that part of the equation??
rtj1211 says:
December 29, 2013 at 1:43 am
Well, that’s a model for a thermostat, which is the default setting within one geological era.
You also need a model of how to over-ride the thermostat, creating either very warm periods or ice ages with temperatures ten degrees or more hotter or colder than currently.
That would be Svensmark’s theory – Galactic Cosmic Ray modulation of the tendency of water vapour to form clouds. All else being equal, when GCR levels are high, Willis’ cumulus clouds would form a few minutes earlier, resulting in a cooler Earth.
Sorry, and Milankovitch cycles of course 🙂
The very strong negative albedo effect at high latitudes tallies nicely with the Milankovich curves. The Milankovich parameter that correlates with the glacial-interglacial cycles is the amount of sunlight at high latitudes.
The tropical thermostatic throttle mechanism explains why there has been almost no warming in the lower latitudes which is a paradox (an observation that directly contradicts the predictions of the IPCC general circulation models (GCM) and that indicates there are one or more fundamental errors in the GCMs.)
http://bobtisdale.files.wordpress.com/2013/11/figure-72.png
It is truly strange (surreal, twilight zone weird ) that billions upon billions of dollars have been spent to study ‘climate’ change and five IPCC reports have been published and there is no comment that the observed latitude warming pattern on the planet disproves the extreme/catastrophic AGW theory. If there is no CAGW problem we do not need to spend trillions of dollars on green scams that do not work even if there was a EAGW problem which there is not.
The warmists and media have screamed from the roof tops the fact that the planet has warmed. They have completely hidden the fact that lower latitudes have not warmed which disproves catastrophic AGW. They have hidden the failure of CAGW theory by focusing on the polar warming and calling the polar warming amplification with no comment that the polar warming is not predicted by the GCM.
Curiously the polar warming has abruptly and suddenly reversed. There was a 50% increase in summer sea ice in the Arctic and there is now two sigma record sea ice in the Antarctic for every month of the year. Atmospheric CO2 has not changed. Something must have changed to cause the sudden cooling of both poles which also indicates that something else besides the increase in atmospheric CO2 caused the warming. (Hint the sun.) Another curious observation is the sudden inhibiting of the La Niña / El Niño cycle. Changes in observations require a physical explanation.
Latitudinal Warming Paradox
As CO2 is more or less evenly distributed in the atmosphere the potential for CO2 warming is the same for all latitudes. The actual theoretical warming (if theory worked in reality which it does not which indicates there are one or more fundamental errors in the models) due to CO2 is linearly dependent on the amount of long wave radiation at the latitude in question before the increase in CO2. As most amount of long wave radiation that is emitted to space is in the tropics the most amount of warming due to the CO2 increase should have occurred in the tropics. That is not what is observed as shown in Bob Tisdale graph. The following is a peer reviewed paper that supports the above assertions.
http://arxiv.org/ftp/arxiv/papers/0809/0809.0581.pdf
“These effects do not have the signature associated with CO2 climate forcing. (William: This observation indicates something is fundamental incorrect with the IPCC models, likely negative feedback in the tropics due to increased or decreased planetary cloud cover to resist forcing). However, the data show a small underlying positive trend that is consistent with CO2 climate forcing with no-feedback. (William: This indicates a significant portion of the 20th century warming has due to something rather than CO2 forcing.)”
“These conclusions are contrary to the IPCC [2007] statement: “[M]ost of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations.”
Willis, I would guess that , like most people doing the sensitivity thing, you are doing OLS regression of dRad against dT, with T on the x-axis.
Since OLS _requires_ negligible error in x variable, doing this on two variables with significant uncertainty is not really giving the correct regression slope. This is known in some circles as regression dilution. The regressed slope will (almost always) be lesser in magnitude that the real ratio.
How much less depends upon the spread of the data. First check is invert the axes and do the regression the other way around.
It’s quite easy to have a factor of 2 difference with broadly scattered data !
It could be that your figure of 10 W/K is significantly underestimating the effect.
This is one of the biggest blunders in climate science in attempting to assess climate sensitivity in both models and obs data. There is a tradition of doing the plot (and hence the regression) this way around and since climate sensitivity (CS) is derived from the inverse of the slope everyone using this kind of method _over-estimates_ CS.
Spencer, Trenberth, Dessler …. just about everyone is doing it. Notable exceptions being Lindzen & Choi, who get much lower sensitivities !
I really need to find time to write this up in detail. Simple OLS regression on scatter plots is wrong but just about everyone is either ignorant of that or thinks it doesn’t “matter”. Well it often does.
It would be interesting to see how much your estimations change if you invert the regression. I would imagine that would be pretty easy to flip variables in your R code lm() function.
In support of:
Eric Worrall says:
December 29, 2013 at 1:59 am
rtj1211 says:
December 29, 2013 at 1:43 am
Well, that’s a model for a thermostat, which is the default setting within one geological era.
You also need a model of how to over-ride the thermostat, creating either very warm periods or ice ages with temperatures ten degrees or more hotter or colder than currently.
That would be Svensmark’s theory – Galactic Cosmic Ray modulation of the tendency of water vapour to form clouds. All else being equal, when GCR levels are high, Willis’ cumulus clouds would form a few minutes earlier, resulting in a cooler Earth.
William:
As the solar system revolves about the galaxy it bobs in and out of the galaxy arms which results in a 3 to 5 fold increase in galactic cosmic rays (GCR, mostly high speed protons) for the period of time when the solar system passes through the galaxy’s arms.
http://ruby.fgcu.edu/courses/twimberley/EnviroPhilo/Phanerozoic.pdf
Celestial driver of Phanerozoic climate?
http://www.phys.huji.ac.il/~shaviv/articles/long-ice.pdf
The Spiral Structure of the Milky Way, Cosmic Rays, and Ice Age Epochs on Earth
Willis, what you are describing with the timing of a major switch of state, say between cloudy and clear skies, is analogous to digital pulse width modulation (PMW). It is the duty cycle that controls the effect, rather than amplitude. The strong positive feedback of the ’emergent phenomenon’ makes it snap on and off like a switch.
Small scale variability in SST means that on a regional scale there can be a continuous response to variations in input.
In doing a linear regression you are attempting to model the response as linear. (I’m not necessarily against that idea but I thought you were, maybe try regressing T^2 against rad if you are still of that opinion. Again, you’d want to plot and regress the other way around. ).
Willis: “I hold that this immediate response is the main reason that it is so hard to find e.g. a solar signal in the temperature record—because the thermoregulation is temperature based, not forcing based, and thus operates regardless of changes in forcing.”
If you are correct, the reason it’s hard to find a solar signal is that there won’t be one !
Actually, there will be one but outside the tropics.
“This is also the reason that volcanoes make so little difference in the global temperature—because the system responds immediately to cooling temperatures by reducing albedo, opening the thermostatically controlled-throttle to allow the entry of hundreds of extra W/m2 to counteract the drop in temperature.”
http://climategrog.wordpress.com/?attachment_id=278
http://climategrog.wordpress.com/?attachment_id=310
Since you mention volcanoes
http://climategrog.wordpress.com/?attachment_id=750
Stratosphere is nice since it’s much less noisy than down below.
Note the definitive 0.5K drop in TLS after each event.
Milankovitch cycles are at best unhelpful insofar as it is already possible to make planetary comparisons with other planets and draw proper conclusions based on a climate spectrum between zero degrees (Equatorial climate) and 90 degrees (Polar climate).
A planet with zero inclination maintains the same surface weather conditions across its orbital period,something like Jupiter with its 3 degree inclination hence storms like the Great Red Storm have no real orbital input to dissipate the storm whereas at the other extreme,Uranus has a Polar climate with its 82 degree inclination and extreme swings in surface conditions across its orbital period.
The old ‘no tilt/no seasons’ has to go in order to research planetary climate properly by introducing the climate spectrum between o Degrees and 90 degrees as a working principles.Each planet has a global temperature budget and it is the degree of inclination that determines how the surface spends that budget over the course of its orbit,if is has an Equatorial climate with zero degree inclination then the heat to col differential across latitudes with be more or less steady whereas a 90 degree inclination then the expenditure will be extreme and more extensive across all latitudes much like moving the Arctic circle close to the Equator.
Milankovitch cycles operate off a false assumption because the heat budget of the Earth is the same regardless of inclination so an increase or a decrease in inclination doesn’t affect the level of solar energy a location experiences across an annual cycle,only the way it experiences that radiation. If people assume a polar climate means a ‘colder’ climate then they are mistaken,it only means how the surface responds to inclination across latitudes.
Gerald Kelleher
Milankovitch cycles operate off a false assumption because the heat budget of the Earth is the same regardless of inclination so an increase or a decrease in inclination doesn’t affect the level of solar energy a location experiences across an annual cycle,only the way it experiences that radiation. If people assume a polar climate means a ‘colder’ climate then they are mistaken,it only means how the surface responds to inclination across latitudes.
Wrong, the connection between Milankovitch cycles and ice ages appears to be the level of insolation received by the Northern Hemisphere.
From https://en.wikipedia.org/wiki/Milankovitch_cycles
Changes near the north polar area, about 65 degrees North, are considered important due to the great amount of land. Land masses respond to temperature change more quickly than oceans, which have a higher effective heat capacity, because of the mixing of surface and deep water and the fact that the specific heat of solids is generally lower than that of water.
If earth reacts in a way its surface temperature keeps constant, then a simple question what is this temperature? You can answer me the temperature that is actually observed…but it is weird.
So even it you caught a process in the complexity of climate…you still don’t understand the climate because nobody can tell you,given the sun radiation, given the astronomical situation, given the physical composition of earth the temperature at the surface will be….
whatever will be the calculation to try to know the temperature they always will start from the current state of earth climate and we don’t know if we are in a equilibrium state or not.
To make it shorter , you may have found a process to keep a changing temperature constant…
but so nice and so rare to see some real data …
This is so straightforward it is beautiful. I have no way of knowing if you are right but it really is a beautiful idea!
Surely “Climate Science” is aware of this albedo = f(Temp) relationship. But, from what I gather from Spencer’s shall I say “input”, Willis’ time of day hypothesis is not recognized.
With the hundreds of watts-m^-1 involved in Greg’s “switching duty cycles”, the amount of time the switch is “on” is not trivial. Doesn’t seem like it would be that hard to model this “large signal/large response” phenomenon… and get a model with some skill in reflecting reality (past and future).
Willis, I remember well your “spot the volcanos” charts. Greg might be onto something higher up.
Re: land, sea and clouds over them. I was lying on the beach at Giardini, Sicily having climbed up Mt Etna to look into the bubbling cone. It was impressive but so were the clouds.
During the bright sunny days there were white clouds drifting from over the land straight over the sea, the path square to the shoreline. As the clouds passed the shoreline going East they disappeared in a period of a few minutes. They literally evaporated. Cloud after cloud met its death at the shoreline.
I conclude, after reading your piece today that the temperature over land was higher than over the (still lovely, but cloudless) Mediterranean Sea. You can probably extend your thesis to include land where there is enough moisture available.
Another great post Willis. Thanks much.
I first got interested in the climate in the early 70s when so many “experts” were predicting another ice age real soon. There was a fair amount of hysteria and would have been even more if everyone was not thinking we would all die in a nuclear war before the ice froze us to death. (plus I live in Florida — hard to get excited about ice here)
I have come to believe, based on decades of information and observing predictions, that the climate of this planet is very, very complex and that it has many “controls” or “thermostats”. During the life of the planet, if one can believe the paleo-climatologists, the climate has remained remarkably stable within certain large cycles. Earth has been an “ice ball” at times with much of the northern hemisphere under ice and it has also been approximately the present temperature. (or even a bit warmer at times in the past)
The real unanswered question is what causes an interglacial period of warmer global average temperature lasting thousands of years and what causes a return to the glacial periods within the present ice age. Other than this oscillation between glacial and interglacial, all other questions pale in comparison.
Alarmists holler that an average warming of 2 degrees would kill most of us with great disruptions, and yet some of my family left the north and came here to central Florida where it is perhaps 20 degrees warmer than their home area is. They seemed to handle the rapid increase in temperature (20 degrees in a day!) just fine. It is freezing temperatures that kill people.
In conclusion, I would love to know what causes the onset of glacial periods, and I would love to see the socialists who want to use CO2 to scare people into submission to shut up. (hoping for 1 out of 2 before I die)
Willis, have you ever thought of getting this kind of article published?
A little reworking and it would be more or less there.
One thing’s for certain, given the non-linearity of any such negative feedback function, hell will freeze over before we’ll see it incorporated properly into a GCM, the modelers can’t seem to even get the coefficients of the linear stuff correct.
http://imageshack.us/a/img850/8112/wm4.gif
http://imageshack.us/a/img716/5372/ypj.gif
http://imageshack.us/a/img850/876/f0z.gif
Credit: JRA-25 Atlas.
@ur momisugly dp & markx : Thanks for links.
Annual Cycle
• Concise overview of heat engines = p.433 [pdf p.10] here:
Sidorenkov, N.S. (2005). Physics of the Earth’s rotation instabilities. Astronomical and Astrophysical Transactions 24(5), 425-439.
• Elaboration on heat engines = section 8.7 (begins on p.175 [pdf p.189]) here:
Sidorenkov, N.S. (2009). The Interaction Between Earth’s Rotation and Geophysical Processes. Wiley.