The Magnificent Climate Heat Engine

Guest Post by Willis Eschenbach

I’ve been reflecting over the last few days about how the climate system of the earth functions as a giant natural heat engine. A “heat engine”, whether natural or man-made, is a mechanism that converts heat into mechanical energy of some kind. In the case of the climate system, the heat of the sun is converted into the mechanical energy of the ocean and the atmosphere. The seawater and atmosphere are what are called the “working fluids” of the heat engine. The movement of the air and the seawater transports an almost unimaginably large amount of heat from the tropics to the poles. Now, none of the above are new ideas, or are original with me. I simply got to wondering about what the CERES data could show regarding the poleward transport of that energy by the climate heat engine. Figure 1 gives that result:

Figure 1. Exports of energy from the tropics, in W/m2, averaged over the exporting area. The figures show the net of the energy entering and leaving the TOA above each 1°x1° gridcell. It is calculated from the CERES data as solar minus upwelling radiation (longwave + shortwave). Of course, if more energy is constantly entering a TOA gridcell than is leaving it, that energy must be being exported horizontally. The average amount exported from between the two light blue bands is 44 W/m2 (amount exported / exporting area).

We can see some interesting aspects of the climate heat engine in this graph.

First, like all heat engines, the climate heat engine doesn’t work off of a temperature. It works off of a temperature difference. A heat engine needs both a hot end and a cold end. After the working fluid is heated at the hot end, and the engine has extracted work from incoming energy, the remaining heat must be rejected from the working fluid. To do this, the working fluid must be moved to some location where the temperature is lower than at the hot end of the engine. 

As a result, there is a constant flow of energy across the blue line. In part this is because at the poles, so little energy is coming from the sun. Over Antarctica and the Arctic ocean, the sun is only providing about a quarter of the radiated longwave energy, only about 40 W/m2, with the remainder being energy exported from the tropics. The energy is transported by the two working fluids, seawater and air. In total, the CERES data shows that there is a constant energy flux across those blue lines of about six petawatts (6e+15 watts) flowing northwards, and six petawatts flowing southwards for a total of twelve petawatts. And how much energy is twelve petawatts when it’s at home?

Well … at present all of humanity consumes about fifteen terawatts (15e+12) on a global average basis. This means that the amount of energy constantly flowing from the equator to the poles is about eight-hundred times the total energy utilized by humans … as I said, it’s an almost unimaginable amount of energy. Not only that, but that 12 petawatts is only 10% of the 120 petawatts of solar energy that is constantly being absorbed by the climate system.

Next, over the land, the area which is importing energy is much closer to the equator than over the sea. I assume this is because of the huge heat capacity of the ocean, and its consequent ability to transport the heat further polewards.

Next, overall the ocean is receiving more energy than it radiates, so it is exporting energy … and the land is radiating more than it receives, so it is getting energy from the ocean. In part, this is because of the difference in solar heating. Figure 2, which looks much like Figure 1, shows the net amount of solar radiation absorbed by the climate system. I do love investigating this stuff, there’s so much to learn. For example, I was unaware that the land, on average, receives about 40 W/m2 less energy from the sun than does the ocean, as is shown in Figure 2. 

(Daedalus, of course, would not let this opportunity pass without pointing out that this means we could easily control the planet’s temperature by the simple expedient of increasing the amount of land. For each square metre of land added, we get 40 W/m2 less absorbed energy over that square metre, which is about ten doublings of CO2. And the amount would be perhaps double that in tropical waters. So Daedalus calculates that if we make land by filling in shallow tropical oceans equal to say a mere 5% of the planet, it would avoid an amount of downwelling radiation equal to a doubling of CO2. The best part of Daedalus’s plan is his slogan, “We have to pave the planet to save the planet”  … but I digress).

Figure 2. Net solar energy entering the climate system, in watts per square metre (W/m2). Annual averages.

You can see the wide range in the amount of sunlight hitting the earth, from a low of 48 W/m2 at the poles to a high of 365 W/m2 in parts of the tropics.

Now, I bring up these two Figures to highlight the concept of the climate system as a huge natural heat engine. As with all heat engines, energy enters at the hot end, in this case the tropics. It is converted into mechanical motion of seawater and air, which transports the excess heat to the poles where it is radiated to space.

Now, the way that we control the output of a heat engine is by using something called a “throttle”. A throttle controls the amount of energy entering a heat engine. A throttle is what is controlled by the gas pedal in a car. As the name suggests, a throttle restricts the energy entering the system. As a result, the throttle controls the operating parameters (temperature, work produced, etc.) of the heat engine.

So the question naturally arises … in the climate heat engine, what functions as the throttle? The answer, of course, is the clouds. They restrict the amount of energy entering the system. And where is the most advantageous place to throttle the heat engine shown in Figure 2? Well, you have to do it at the hot end where the energy enters the system. And you’d want to do it near the equator, where you can choke off the most energy.

In practice, a large amount of this throttling occurs at the Inter-Tropical Convergence Zone (ITCZ). As the name suggests, this is where the two separately circulating hemispheric air masses interact. On average this is north of the equator in the Pacific and Atlantic, and south of the equator in the Indian Ocean. The ITCZ is revealed most clearly by Figure 3, which shows how much sunlight the planet is reflecting.

Figure 3. Total reflected solar radiation. Areas of low reflection are shown in red, because the low reflection leads to increased solar heating. The average ITCZ can be seen as the yellow/green areas just above the Equator in the Atlantic and Pacific, and just below the Equator in the Indian Ocean. 

In Figure 3, we can see how the ITCZ clouds are throttling the incoming solar energy. Were it not for the clouds, the tropical oceans in that area would reflect less than 80 W/m2 (as we see in the red areas outlined above and below the ITCZ) and the oceans would be much warmer. By throttling the incoming sunshine, areas near the Equator end up much cooler than they would be otherwise.

Now … all of the above has been done with averages. But the clouds don’t form based on average conditions. They form based only and solely on current conditions. And the nature of the tropical clouds is that generally, the clouds don’t form in the mornings, when the sea surface is cool from its nocturnal overturning.

Instead, the clouds form after the ocean has warmed up to some critical temperature. Once it passes that point, and generally over a period of less than an hour, a fully-developed cumulus cloud layer emerges. The emergence is threshold based. The important thing to note about this process is that the critical threshold at which the clouds form is based on temperature and the physics of air, wind and water. The threshold is not based on CO2. It is not a function of instantaneous forcing.  The threshold is based on temperature and pressure and the physics of the immediate situation.

This means that the tropical clouds emerge earlier when the morning is warmer than usual. And when the morning is cooler, the cumulus emerge later or not at all. So if on average there is a bit more forcing, from solar cycles or changes in CO2 or excess water vapor in the air, the clouds form earlier, and the excess forcing is neatly counteracted.

Now, if my hypothesis is correct, then we should be able to find evidence for this dependence of the tropical clouds on the temperature. If the situation is in fact as I’ve stated above, where the tropical clouds act as a throttle because they increase when the temperatures go up, then evidence would be found in the correlation of surface temperature with albedo. Figure 4 shows that relationship.

Figure 4. Correlation of surface temperature and albedo, calculated on a 1°x1° gridcell basis. Blue and green areas are where albedo and temperature are negatively correlated. Red and orange show positive correlation, where increasing albedo is associated with increasing temperature.

Over the extratropical land, because of the association of ice and snow (high albedo) and low temperatures, the correlation between temperature and albedo is negative. However, remember that little of the suns energy is going there.

In the tropics where the majority of energy enters the system, on the other hand, warmer surface temperatures lead to more clouds, so the correlation is positive, and strongly positive in some areas.

Now, consider what happens when increasing clouds cause a reduction in temperature, and increasing temperatures cause an increase in clouds. At some point, the two lines will cross, and the temperature will oscillate around that set point. When the surface is cooler than that temperature, clouds will form later, and there will be less clouds, sun will pour in uninterrupted, and the surface will warm up.

And when the surface is warmer than that temperature, clouds will form earlier, there will be more clouds, and higher albedo, and more reflection, and the surface will cool down.

Net result? A very effective thermostat. This thermostat works in conjunction with other longer-term thermostatic phenomena to maintain the amazing thermal stability of the planet. People agonize about a change of six-tenths of a degree last century … but consider the following:

•  The climate system is only running at about 70% throttle.

•  The average temperature of the system is ~ 286K.

•  The throttle of the climate system is controlled by nothing more solid than clouds, which are changing constantly.

•  The global average surface temperature is maintained at a level significantly warmer than what would be predicted for a planet without an atmosphere containing water vapor, CO2, and other greenhouse gases.

Despite all of that, over the previous century the total variation in temperature was ≈ ± 0.3K. This is a variation of less than a tenth of one percent.

For a system as large, complex, ephemeral, and possibly unstable as the climate, I see this as clear evidence for the existence of a thermostatic system of some sort controlling the temperature. Perhaps the system doesn’t work as I have posited above … but it is clear to me that there must be some kind of system keeping the temperature variations within a tenth of a percent over a century.

Regards to all,

w.

PS—The instability of a modeled climate system without some thermostatic mechanism is well illustrated by the thousands of runs of the ClimatePredictionNet climate model:

Note how many of the runs end up in unrealistically high or low temperatures, due to the lack of any thermostatic control mechanisms.