Air Conditioning Nairobi, Refrigerating The Planet

Guest Post by Willis Eschenbach

I’ve mentioned before that a thunderstorm functions as a natural refrigeration system. I’d like to explain in a bit more detail what I mean by that. However, let me start by explaining my credentials as regards my knowledge of refrigeration.

The simplest explanation of my refrigeration credentials is that I have none at all. As with many trades I’ve pursued, I have no training in refrigeration. But the challenge was simple. When I was 37, a good friend of mine and I had taken the job of installing a blast freezer system in a 60′ (18m) steel sailboat in Fiji called the Askoy. I was sure we could do it … despite the fact that at that point in my life, neither of us had ever taken apart a refrigerator, or could even explain how a refrigerator worked.

How the heck do I get things coldFigure 1. The Challenge SOURCE

But we had two months before the job started, and one rule of thumb has never failed me—Do Your Homework

I was laughing about this with my friend this afternoon. We’ve been partners in various oceanic ventures and adventures over the last forty years. He reminded me that I’d bought my refrigeration gauges and my freon sniffer at the local flea market, I’d forgotten that detail. He was to do the metalwork, the piping and the soldering and such, while I had to design the system and charge it and get it working. We discussed our ignorance at the time, and he said “I never had any doubt that you’d do the refrigeration part.” I laughed and said “I never had any doubt that you’d do the metalwork part.”

I learned refrigeration the old-fashioned way. I taught myself.

WARNING—this post is a 50/50 mixture of science and autobiography, call it autosciography. If that makes your brain explode, DO NOT continue reading.

I went to a technical bookstore in San Francisco and bought a college refrigeration textbook, and a refrigeration technicians textbook. I started with the college refrigeration text, just like I was in college again. I read every word of every chapter, and then I answered all of the questions at the end of each chapter. I went back on the ones I missed until I understood them as well. At the end of the first month I could knowledgeably discuss superheat and the difference between the kinds of Freons and how different types of refrigeration systems worked and what the units called “tons” measure in refrigeration (the cooling power equivalent to the melting of 2,000 pounds of ice starting at 0°C in 24 hours).

Then once I understood the theory backwards and forwards, I got out my refrigeration gauges and my sniffer, and I found some old refrigeration systems and I started working through the refrigeration technicians manual. By the end of the second month I could test and charge and repair a system, fine tune the setup, discharge the system and recapture the freon, tear it down and build it again, whatever you wanted. I was ready to go.

So that’s how I learned about refrigeration, and my friend did the same regarding the metalworking and silver soldering and all the rest of the knowledge he needed. And after we finished the installation of the blast freezer, I subsequently made good money at various times diagnosing and repairing marine refrigeration systems. I’ll return to the question of my credentials and the lack thereof in a bit. But first, for those who like me couldn’t explain how a refrigeration system works, here’s my explanation.

A refrigerator cools things in exactly the same way that sweating cools your body—by evaporation. Of course instead of using water like your body does, a refrigerator uses freon, or one of the modern refrigerants. But the principle is exactly the same regardless of the “working fluid”. You use an evaporating liquid to remove the heat from whatever you want to cool down.

Now, if the working fluid is actually boiling, you get the maximum evaporation. So for a particular refrigeration application, you might pick a liquid (one of the various Freons in the old days, now other liquids) that boils at say ten degrees below freezing.

Of course, your body uses up the water that cools us when we sweat. We don’t try to recapture that water, it condenses somewhere else.

But we don’t want to waste valuable freon. We’d rather condense it back to a liquid. One way to do that, of course, is to pipe the vapor to some cold place, where it naturally condenses back into a liquid. In The Inventions of Daedalus, the eponymous author propounds another of his crack-brained but plausible schemes, this one for air conditioning Nairobi by running the vapor up to the top of Kilimanjaro, where it would condense and run back down by gravity. Here’s my sketch of his plan:

nairobi airconditioningFigure 2. Daedalus’s plan for air-conditioning Nairobi (click to expand). The working fluid boils at say 5°C (41°F). The liquid return pipe is insulated so the fluid doesn’t boil on the return path to the evaporator.

The working fluid boils in the evaporator on the lower left. The evaporator is like a car radiator, and a fan blows through it, and the resulting cool air is used to air condition Nairobi.

The vapor then moves up to the top of Kilimanjaro, where a condenser (also looks like a car radiator) has icy natural winds blowing through it to condense the liquid. This liquid then flows by gravity down an insulated tube and back to Nairobi to start the cycle again. And like most of Daedalus’s inventions, there’s no reason you couldn’t actually build that.

Now, that’s the basic principle underlying how a refrigeration system works. A liquid turns to a vapor, absorbing heat in the process. This is called “latent heat”, because it doesn’t increase the temperature.

That vapor, containing the latent heat, is piped away from the object you want to refrigerate. Then somewhere else, it’s condensed back to a liquid, and in the process releasing the latent heat as sensible heat. Finally, the liquid is returned to the original location, to repeat the cycle.

Note the importance of the two phase changes in the process—evaporation (picking up latent heat) and condensation (releasing that latent heat elsewhere as sensible heat). Those two processes, evaporation and condensation, are the central part of the whole process of refrigeration. It’s just a very efficient way to move heat from A to B.

Now, consider Figure 3, which shows what a tropical thunderstorm is doing, and how it functions as a refrigeration system.

Down at the surface, the water is evaporating and refrigerating the surface. The thunderstorm forms over a local hot spot. The evaporation cools the surface, and the energy is transferred to the air as latent heat. The hot, moisture-laden air moves upwards.

Up above the “lifting condensation level”, the elevation of the base of the clouds where condensation begins, the water condenses into larger and larger droplets. The latent heat is released back into the air as sensible heat. The water then falls as rain, to complete the cycle.

thunderstorm refrigeratorFigure 3. Natural refrigeration system. Just as in a domestic refrigerator, a working fluid (in this case water) is evaporated to remove heat from the surface, the area to be refrigerated. After rising up into the thunderstorm, the water is condensed, releasing the latent heat as sensible heat.

As you can see, this is the same system that Daedalus proposes to air condition Nairobi. It uses the same principle as your home refrigerator. Evaporation cools what you want cooled, and somewhere else, you condense the working fluid and get rid of the heat.

Now, let me start by making one thing crystal clear.


Instead, it is a natural refrigeration system, capable of cooling the surface well below its starting temperature. Treating it mathematically as a feedback is a huge mistake. It is nothing of the sort. It is a threshold-based emergent phenomenon which actively refrigerates the surface.

[UPDATE: In the comments, people have been confused by this question of feedbacks, obviously I was not clear enough. When I say it is not a feedback, I mean it is not a simple linear feedback of the only kind considered by the IPCC. Instead, it is a control system which utilizes feedbacks of a host of kinds to maintain a constant temperature. -w.]

Not only that, but it selectively refrigerates the hot spots, forming just where it is needed. As a result, it is very difficult to represent by averages. This is especially true because its response time is minutes to hours, not days. The hot spot doesn’t really have time to get going before it is refrigerated into submission.

It gets better, much better. You see, up until now, I’ve just described the parts of the system that correspond exactly to manmade refrigeration systems. Let me point to some very clever wrinkles that thunderstorms use to increase their refrigeration capacities and to cool the surface more efficiently and more widely.

• Wind

The thunderstorm generates wind at its base, and evaporation is proportional to wind speed. If the wind underneath the storm cloud increases from say 5 knots to 20 knots, or say from 2 m/sec to 8 m/sec, evaporation goes up by a factor of 20 / 5 = four-fold. In other words, the self-generated wind alone multiplies the strength of the refrigeration by about four.

In addition, the wind increases the evaporative area by blowing water into the air as spray and fine droplets. These have a large surface area and evaporate rapidly. This also increases the strength of the refrigeration.

• Dual fuel

Thunderstorms run on both temperature and moisture. Moist air is lighter than dry air. The four-fold increase in evaporation yields a proportional increase in the vertical speed of the air moving through the thunderstorm, because it is much lighter. It also keeps the thunderstorm from dying out if the temperature drops, because once the wind starts, the moist air is light enough to keep the thunderstorm going to well below the temperature required for initiation.

• Direct surface refrigeration by cold working fluid.

In most manmade refrigerators, evaporation is the only mechanism for cooling the objects to be refrigerated. The working fluid is not used directly to cool down what is being refrigerated. Instead, it’s brought to the evaporator in an insulated tube and immediately evaporated to carry away the heat.

But in addition to the evaporation, a thunderstorm also delivers large quantities of chilled water directly to the surface. This is a separate and distinct refrigeration mechanism, one not generally utilized in manmade refrigerators.

• Refrigeration via entrained wind.

You’d expect that the rain would warm as it fell through the warmer lower atmosphere, and to some extent it does. But it also entrains the air around it as it is falling, carrying it along. This sets up a vertical entrained wind that falls right along with the rain. That wind is constantly cooled as it falls by the evaporation of the rain that it is mixed in with. And since the rain and the chilled air fall together as a package from aloft, they both arrive at the surface much cooler than the surroundings. Often when standing out on the apartment deck in the Solomon Islands at night, the first sign of the approach of a thunderstorm would be the arrival of the cool entrained wind.

The entrained wind falls vertically with the rain, but unlike the rain it’s not absorbed by the surface. So it blows out cold air horizontally in all directions from the base of the rainfall. This blast of cool air is quite distinct. It smells of the upper atmosphere where it originated, and it is very refreshing on a hot night. It is also a separate and distinct refrigeration mechanism.

• Re-use of heat of condensation.

This one is sometimes done in manmade installations. In the thunderstorm, the heat is used to drive and sustain the building of the “tower”, the tall vertical part of the cumulonimbus cloud. This in turn increases the speed of the upward flow through the core of the thunderstorm, and allows for the possibility of another phase change.

• Additional phase change

It would certainly be possible for humans to design a system using a second phase change in the working fluid. Right now, our refrigeration systems utilize the phase change from gas to liquid and back again. But there’s another possibility, to go from gas to liquid to solid and back again.

The advantage is that you can move more heat that way. Instead of just the heat from one phase change, you could move the heat from two phase changes as latent heat.

So why don’t humans utilize both phase changes for extra efficiency? Well, we haven’t figured out an easy way to get the solid working fluid from wherever it was frozen, back to the evaporator to start the story over. I mean, we could freeze the Freon after it’s condensed into a liquid … but then how do we move solid Freon back to the evaporator to continue the cycle? With wheelbarrows?

Nature doesn’t mind these small problems, however. Nature continues to cool the water past the point where it condenses, and all the way to where it freezes … and then it uses gravity to return the solid working fluid back to the surface as ice. I can only bow my head in awe, what a clever setup. At the surface the ice will first melt (cooling the surface) and then warm up to the local temperature (further cooling the surface) and then evaporate to continue the cycle.

• Inter-storm coupling.

When the need for surface refrigeration gets high (anomalously warm surface temperatures), a new emergent pattern appears. The thunderstorms start to align themselves in long rows, called “squall lines”. These in turn have long canyons of descending air between them. This is a type of Rayleigh-Bénard circulation that greatly increases the throughput, and thus the refrigeration capacity, of the mass of thunderstorms.


• At all times and all around the planet, thunderstorms are constantly refrigerating tropical hot spots to prevent the globe from overheating. This constant refrigeration is what controls the surface temperature of the planet, not CO2. If this refrigeration system failed even for a week, we’d fry.

• The thunderstorm refrigeration system utilizes the same familiar principles of manmade refrigeration—evaporation removes the heat from what you want to refrigerate, and condensation gets rid of the heat somewhere else.

• In addition, the thunderstorm refrigeration system utilizes some unfamiliar processes, all of which combine to greatly increase the refrigeration capacity of a given thunderstorm.

• The refrigeration is selective, responding to local temperature—the hot spots get refrigerated until they confess, and the cold spots get nothing.

• The current generation of climate models deal with feedbacks. This is nothing of the sort. It is an emergent mobile self-sustaining refrigeration system, not a feedback of any kind. It needs to be analyzed as such, and it is very difficult to do so by means of parameters or averages.

• The system responds to temperature. It is not driven by the forcing, nor does it respond to areas of high forcing. Instead, it actively responds to surface temperatures. The formation of a local hot spot is quickly followed by the formation of a corresponding refrigeration system to cool the hot spot down.

• The system is extremely sensitive to the formation of local hot spots. It puts a refrigeration unit right to work on the problem. On the other hand, it is indifferent as to the cause of the hot spot. It chills them all the same.

• In addition, the surface temperature of the system is relatively insensitive to the number of hot spots—you just get more or less refrigerators to match the number of hot spots, and that keeps the temperatures within bounds. And this in turn means that the surface temperature of the system is relatively insensitive to the forcing.

• As a result, the system doesn’t care about CO2, or about small variations in the sun, or about the effect of volcanoes. The threshold for refrigerator formation is based on surface temperature, not on CO2. If there are more hot spots, the system simply makes more refrigerators, whether the hot spots are from CO2 or from a clearing of the aerosols or from a 5% increase in sun strength over a billion years.

• In such a system, the idea of “climate sensitivity” doesn’t go anywhere or mean anything. The system is relatively insensitive to the forcing, not sensitive. The system responds to hot spots by building refrigerators … and as a result the surface temperature is maintained despite variations in the forcing. The problem is not that the relationship is non-linear. In a thermostatically governed system such as the climate there often may be no relation all between forcing and temperature.

• This is a relatively simplified (but very accurate) explanation only one of a host of interlocking emergent phenomena that maintain the surface temperature within ± half a degree in a hundred years. Yes, there are lots of details I’ve left out, and manmade refrigeration systems have other valves, bells and whistles … if you’re interested lets discuss them, but please don’t bust me for leaving them out.


That’s what I wanted to say about refrigerators and thunderstorms. When you are analyzing our climate, which contains powerful emergent refrigeration systems like thunderstorms, you can’t analyze them as a feedback. It’s very difficult to parameterize them. You have to get out your refrigeration tables and analyze them as what they are, huge natural refrigeration units, and very efficient ones at that.  My takeaway message is this:

The surface temperature of our amazing planet is set and maintained by the constant refrigeration of the surface hot spots as they form, not by the forcing, whether from CO2 or anything else. 


To close, earlier I said I’d return to the question of my credentials for talking about refrigeration. Well, before we went to Fiji my friend and I researched the available marine refrigeration systems. We went and talked to the people freezing the product and saw what they used. Harlow, the owner of the boat, wanted to be able to purchase and process what they call “crayfish”, the tropical ocean lobster. To do that, you want to flash-freeze them with a wind colder than 40° below zero or so (either -40°C or -40°F, they’re the same). They need to be snap-frozen, an ordinary freezer won’t do it.

So my friend and I located a killer packaged refrigeration unit, all assembled, self-contained, ready to go. We could bolt it in and go processing in a couple weeks, because that was the dream. We’d finish the freezer and go processing crayfish around the tropical South Pacific … what’s not to like? So we went down to Fiji. Harlow was going to buy the packaged unit and bring it down.

But Harlow decided he knew better. So he shows up in Fiji with a refrigeration compressor, and an evaporator, and a condenser, and some fittings and valves and pipe, and tells us he wants it built from scratch. Oh, and he wants it water-cooled, not air-cooled. Oh, and not driven electrically, but run off a “lay-shaft”, a separate shaft driven by the main engine which drives other machinery in turn.

Ooooooh kaaaaay … we can do that, Harlow, but it’s gonna take a while.

So for our very first refrigeration project, my friend and I got to design and build an entire marine engine-driven commercial-type blast freezer system with a water-cooled condenser … from scratch, from the individual pieces. Might as well set the bar high, I figure …

So I got out my texts and tables and designed it up, and we got started, nothing else to do. We built the lay-shaft, and installed the refrigeration piping from the engine room up forward to the freezer room, belt-drove the lay-shaft off the main engine, and then got the water pump and the refrigeration compressor to run off the lay-shaft, and laid out and cut and soldered and tested all of the piping, and that’s only a tiny fraction of all of the tasks … in a foreign country, with not a whole lot of refrigeration parts available, and no instruction manual.

Like we warned Harlow, it took a while, just about six months to do it, with my gorgeous ex-fiancee serving as the ship’s cook and nurse and general hard worker. But finally, after scraping Suva dry of various refrigeration parts and pieces, one fine day I charged up the system, and we gave it the first test … and the wind off of that blast freezer was at -50°F (-46C) just like we planned. Indeed, the blast freezer worked like a champ. As long as the main engine was on it could be clutched in or out to run it, and the lay-shaft could also be driven by the auxiliary engine to keep the freezer hold and the seafood frozen if the main engine died. It was a sweet rig.

So naturally, Harlow decided that we should have a party to celebrate, and we were all up for the plan. We’d been anchored right offshore from the Royal Suva Yacht Club the whole time, so we invited everyone.

There was a trimaran owned by a friend at the Yacht Club at the time, so we tied her up alongside the Askoy. That was for the big wide stable dance floor, we hung speakers from the rigging on each side. My friend Doc Lowry used his 28′ open skiff as the shore boat, to bring out loads and loads of people from the Yacht Club dock. He spent most of the night moving folks from the Club to the party and back again … then out to the party again …

As each person arrived, we took them a couple at a time down into the blast freezer. You entered through a hatch in the deck, and down below it was cold, cold, cold. I had put a number of bottles of vodka into a basket right in front of the blast of the freezer. I stuck a thermometer through the cap in one of them, it registered minus forty degrees … I still have a picture of it around here somewhere. [UPDATE—found the picture. You can see the red indication at minus forty]

askoy freezer party

So as each person came down into the freezer, we’d prop them up in front of the wintry blast. The wind was strong, and blowing at minus fifty degrees F. Most of these folks were Fijians, who had never seen any place that was as cold as plus 50°F (10°C), much less minus 50°F. They started shivering as soon as they got inside, they’d never in their lives felt a wind at minus fifty. So I or their other host of the moment would pour each person a reasonable glass of pure vodka at minus forty degrees.

Vodka at that temperature hardly has any taste, and the folks were in a hurry to get out of that damned freezing cold, so they’d drink it down straight like it was water … then we’d take them back out into the balmy tropical night. They’d get about ten steps across the Askoy deck, maybe somebody would pull them onto the dance floor, maybe not, but in either case, a few steps later, the combination of the initial freezing cold wind, the vodka, and the subsequent heat would make their knees wobbly and their eyeballs jiggle, and the Askoy Freezer Party got just that much merrier. The Yacht Club Bar eventually closed, and the bar guys and the Club office ladies joined the party. People kept arriving. Big Jenny showed up at three AM and shouted “Am I late?”

“No, party just starting, girl!” We took her into the freezer and gave her a double shot, and indeed the party restarted when she came back out.

I’ve not been to too many parties like that one. Rumor had it that it resulted in one marriage and a couple of divorces. Both a wallet and a set of eyeglasses committed suicide by jumping into the Suva Harbour sometime during the night.The amount of debris on the deck and the dance floor was overwhelming. And me, I proved it was a magical party. By the end of the night I was so drunk that shortly before dawn I went to sleep on a nice, soft pile of rope I’d discovered up front near the bow of the boat. It was covered by some gunny sacks, and I nestled in and got comfortable and was gone.

When I woke up, though, I sadly concluded that some ungrateful bastard must’ve replaced the rope while I was sleeping, because I found I was nestled on the usual pile of Askoy anchor chain … in its usual spot up in the bow … covered up as always with dirty gunny sacks …

Sleeping on chain, I found out, makes a man say very bad words upon awakening. If you are given an alternative mattress choice, say a bed of nails, or a small barnyard stall with two chickens and a rabid goat, I’d advise taking it. Plus it seemed that the entire Fijian mosquito tribe had taken advantage of the party to do some in-flight refueling. My body was royally whupped in the morning, big anchor chain marks and dents in my hips and side, covered in mosquito bites.

But I didn’t care a bit. The freezer was done, the icy blast off the evaporator was at minus fifty, the vodka had been at minus forty all night long, the party was a success, and my gorgeous ex-fiancée and I had danced away the night.

And that’s all there is to my refrigeration credentials.

Regards to everyone,


0 0 vote
Article Rating
Newest Most Voted
Inline Feedbacks
View all comments
Lance Wallace
March 11, 2013 8:06 pm

I guess the ex-fiancee found a nice warm comfortable bed somewhere and left you to your gunny sacks and anchor chain. Some kind of lesson here about practical-minded females and romantic but goofy males.

Leonard Weinstein
March 11, 2013 8:24 pm

I enjoyed most of your story (as I enjoy your other ones). There is one issue. The air temperature is colder at higher altitudes (on average) due to the lapse rate. However, when the air circulation carries this air down, it warms up due to adiabatic compression at lower altitude, and does not supply colder air to the surface. That is not part of the cooling effect. The cooling from evaporation, dropping of colder drops (which do not heat due to adiabatic compression), and the fact that wind increases surface evaporation and cools your body (removing sweat faster) are the total effect.

March 11, 2013 8:33 pm


March 11, 2013 8:33 pm

I think they give Pulitizers for stuff like newspaper stories. Why not an internet award? That’s actually an interesting idea. If fact I predict [without benefit of models] there will be a Pulitizer for something on the internet within the next ten years.

March 11, 2013 8:34 pm

Having recently sold and retired from the HVAC business, I can now understand refrigeration.
BTW, I’m sure you’ll be the one to leave the 1,000,000th reply. LOL
Thanks w.

March 11, 2013 8:35 pm

Funny you mention refrigeration. I had to replace some R-12 in my car’s AC a while back. The stuff is illegal to make in the US, but not to buy. But you have to have an EPA certificate to be able to buy it. So I went online and took some Mickey Mouse course, and paid $15 and took the test (this was probably about 10 years ago), and now I have an EPA refrigeration mechanic certificate thingy good forever. Never did get around to the AC project.

March 11, 2013 8:37 pm

Marvellous yet again, Willis. Since reading your writings about clouds, I find I look at them differently. I’m learning quite a lot in here, and having a blinking good read while I’m at it. Thank you so much.
Cheers. 🙂

March 11, 2013 8:39 pm

Willis, did you ever have the opportunity to use a sling psychrometer while at sea? I had one years ago and unfortunately sold it. I’d use it to describe to my children why evaporation cools. Kinetic energy is directly proportional to absolute temperature – so as water molecules leave the surface of skin, kinetic energy decreases, and so does temperature.

March 11, 2013 8:45 pm

In such a system, the idea of “climate sensitivity” doesn’t go anywhere or mean anything. The system is relatively insensitive to the forcing, not sensitive.

Insensitivity IS a sensitivity. It is a sensitivity of zero (in the case of complete insensitivity, or complete transport away of any hotspot hotter than the thermostatically controlled level). This sensitivity of zero also corresponds, by definition, to a particular feedback (a feedback of negative one, or completely offsetting). Thus it seems quixotic for Willis to deny that the thermostatic mechanism he describes can be viewed in terms of sensitivity and feedback. It has strong implications for sensitivity and feedback.
There may be good reasons to dislike the forcing-feedback straightjacket that the IPCC has imposed (pursuant to its stated and highly-biased objective of analyzing the human impacts on climate), but the fact is that this formulation HAS been imposed and is the language of 95% of climate science. For Willis to eschew this language when his own formulations actually have strong implications for feedbacks and sensitivities is to muffle his own voice.
Noting the flaws in the forcing-feedback framework does not require denying the implications of one’s own ideas for analyses done within this framework.

Ian Cooper
March 11, 2013 9:00 pm

Willis, would it be fair to say that if tropical thunderstorms can act as refrigeration units at their scale then a full blown tropical cyclone could have a similar effect but on a larger scale? I ask this because I have evidence of such a thing happening here in New Zealand in 1968. I was graphing T-Max data for the austral summer of 1967-68, a brilliant summer that came to a sudden holt with the arrival of TC Giselle, the infamous storm that took out the inter-island ferry Wahine at the mouth of Wellington Harbour on April 10th that year. The ferocious winds that lashed the Cook Strait area that terrible morning had caught everyone by surprise. Forecasters had thought that in typical fashion the storm would lose a lot of its strength once it crossed land. This was pre-satellite and dopler radar days so the unexpected increase in ferocity was unseen in the night.
The main thing that showed up in the T-Max data for my home town some 90 miles north of Wellington was the staggering drop of around 8 degrees C that ensued directly after the passing of Giselle. The memorable heat of the that lingering late summer-early autumn was gone and never returned. A massive step change had occurred in a way that I haven’t seen in subsequent transits of tropical cyclone remnants (they are hardly ever still officially T.C’s when they make landfall in N.Z.) over New Zealand. A few have had a short term affect before temperatures have rebounded, but Giselle was agame changer for sure.

March 11, 2013 9:08 pm

Willis. After all these years, you have dashed my belief that refrigeration was a product of black magic. If you yearn to visit Gualala, and eventually everyone does, drop in on a second or fourth Tuesday. I and the rest of the Lions would love to hear some of your tales.

March 11, 2013 9:11 pm

It’s certainly true that hurricanes can cool the surface. I made some animations here of high resolution daily SST in the N Atlantic with hurricane tracks superimposed. Although the resolution sounds good, hurricanes can move a long way in a day. Still, some do leave very marked cooling tracks.
Of course, they have to cool something – they are heat engines and transferring heat from warm to cold is where their energy comes from.

Mike Wryley
March 11, 2013 9:39 pm

Since you mentioned silver soldering and clorofluorocarbons in nearly the same breath, and because many who tread this site may not be as well read, crafty and lucky as you but never the less inspired to follow in your sandal tracks, I feel it might be worth mentioning that open flames and Freon are a deadly combination.

ed mister jones
March 11, 2013 9:45 pm

I understand T-Storms, as I fly around them. I understand the transport of heat from the surface and near surface hot spot. I understand the cooling effect. But the Heat does wind up somewhere, and if that somewhere doesn’t pass the heat somewhere else yet again, then the first somewhere is going to get warmer (as happens when you run an ordinary refrigerator in an unventilated room). . . . So the next piece might I suggest, should deal with the “and then . . .” aspect of the voyage of the heat from the Hot Spot to it’s next destination(s)?
Do we know the effect of the dissimilar temperature water on the Ocean’s vertical circulations as you discussed previously? Can it be quantified? (rhetorical).
So many moving parts and variables, the Idea that ‘they’ have enough knowledge to model it and predict outcomes Decades into the future is laughable.

March 11, 2013 10:03 pm

I tried freezing vodka with a big block of dry ice once, same effect on a warm night – people don’t realise how drunk they are. But not nearly as much fun as a blast freezer!

March 11, 2013 10:51 pm

Isn’t “Reversing Climate Change” the same as “Manufacturing Climate Change”, thus returning the Climate to a previous state? And what would that state of Climate Be?

March 12, 2013 12:19 am

Couple of things, I have always looked at a refrigerator as a woking model of planet earths heating and cooling, with the sun as a compressor(heat of compression) and thunderstorms as condensor, #2 what was the reliability of an open shaft refrigeration compressor in salt water environment, great post as always Willis!

March 12, 2013 12:23 am

Great post again Willis,
Askøy is an island off the western coast of Norway, close to the city of Bergen. I am sure the vessel hails from there originally.
How it ended up in Fiji is probably a story in its own right…

March 12, 2013 12:38 am

This, IMHO, is one of the most important posts of yours to date with regards to the technical aspects. Like you I am self taught in refrigeration. Bought the books and read every word and made it a point to understand every aspect inside out. Engineering in college only gives one the basic learning tools to work with.
The other engineers reading this post are certainly enjoying it. For all, understand the detailed principles of the refrigeration cycle to the extent that you know your own name. You should then know equally well the gas/liquid/solid properties of a gas (that could be stated better but the point is made). With this knowledge in hand/mind a much better understanding of the so called climate science issues can be evaluated.
Now the real kicker is the absorption refrigeration cycle. That one is like the one in your camper where heat (propane) is burned to cool your refrigerator. That one takes a couple of reads to fully understand. Please note the use of the term “fully”.
Again Willis, this was an excellent post that should help the majority of readers here. There are others who will nit pick on the proper use of certain terms or conditions but that will not take away from what you have written. Any nit picking can be covered in the comments/discussion as you noted in your post. Good work!

Greg Goodman
March 12, 2013 12:51 am

Another excellent, informative and entertaining articel Willis.
For the most part I would say accurate and clearly explained. However, there one point I think you’re going a bit off the rails.
“Now, let me start by making one thing crystal clear.
Instead, it is a natural refrigeration system, capable of cooling the surface well below its starting temperature. Treating it mathematically as a feedback is a huge mistake. It is nothing of the sort. It is a threshold-based emergent phenomenon which actively refrigerates the surface.”
Well, despite saying it isn’t you provide a very detailed explanation of why it IS a feedback. A strong negative feedback. Now if this strong negative feedback produces what you call an emergent phenomenon, fine. The two are contradictory , you’re just giving a name to the mechanism.
You repeatedly say how any hotspots provoke a storm that cools them. I can’t imagine a simpler explanation of what a negative feedback does.
The amplifying effect of the updraught causing increasing surface wind and spray evaporation is a clear positive feedback. So the negative feedback process itself has a positive feedback.
That makes it a very STRONG NEGATIVE FEEDBACK.
In fact, it’s strong enough to create localised emergent phenomena. 😉
The overall effect on tropical climate that you point out is also exactly that of a strong negative feedback: the output variable (temperature) varies only over very small range in response to potentially large changes in the input drivers.
This could be phrased as it having a very low sensitivity to input forcings.
This thread on Climate etc. discusses one paper that is finally recognising the “sensitivity” varies regionally.
Be patient , scientific orthodoxy has the turning circle of massive oil tanker with a full load !
The next question after the tropics is what is the sensitivity at the poles where a lot of the heat pumped out by the mechanism you describe here ends up.
I hope to post on that soon.
One clue is at the end of this graph.
Arctic ice cover has reduced notably in response to warmer Atlantic waters, yet the rate of change in ice cover is still swinging around the same value close to zero (longer term average over this data, which only covers post 1980 warming period, is negative.).
That would suggest that the new geography with more exposed water also shows a negative feedback to temperature. Not the positive feedback and supposed tipping points we always hear about.
If the warmer N. hemisphere conditions were simply melting the ice with no negative (or a positive) feedback then the rate of change would be matching the still warm conditions. The divergence at the end suggests something else is happening.
More detail on that soon.

March 12, 2013 12:53 am

One more comment that I would like to make that is OT but has some relevance (in a weird sort of way) for those who understand the gas laws fully, especially engineering and physics. For an equally mental challange go to the and archive to the era of the Deepwater Horizon spill and follow the discussions paying close attention to Dave Summers who tags/comments as Heading Out. Don’t do this unless you have a few hundred hours to spare or are curious in a technical sort of way.

Greg Goodman
March 12, 2013 1:02 am

Willis: “The main problem is that the “forcing-feedback” paradigm doesn’t contain a governor, so yes, I do have to deny that their framework is sufficient to encompass the climate.”
It will be interesting to see you details of “governor”. You are correct that a simple linear feedback is inadequate. That does not mean you need to abandon the whole idea of feedbacks. I suspect you governor may be a non linear negative feedback as I outlined above: a negative feedback that whose magnitude itself has a positive feedback. I’ll have to wait to see you details but I think you, I and Alex are saying the same thing.

March 12, 2013 1:05 am

Sometimes I think you are a clone of me. I learned refer systems on cars and later turned that to cooling yachts and work boats. Crazy world, eh. Anyway – your analogy goes all to hell because none of this works withing energy entering the system. Don’t care if it is in Tristan d’Cunha, PNG, or in the core of my Kelvinator, nothing does anything without energy going in. I’d accept any refinement that ends up with an imbalance of energy along the transfer path.
You start your cycle with a hot spot but that hot spot didn’t arrive from nowhere. To be complete you need to show how that spot got hot. It was put there and that putting put the rest in motion. Here’s the point. This is all – repeat that – ALL energy positive. Yes, some places may be cooled by energy movement, but this is, and always will be an energy positive operation. The only way the excess energy can leave the system is by radiation to the great voids of infinite space.
Any energy that does not immediately go there remains in the local (Earth) system but displaced from the point of origin. So you can cool stuff here but at the expense of heating or sustaining warmth there. No free lunch.
But that isn’t the end of it – that heat is moved by T-storms somewhere useful if you are a mad greeny crazed with passion of containing global warming. Those T-storms move that heat from the surface to very high altitudes possible faster than it can have gotten there by radiation, ping ponging between CO2 molecules (convection), and physical abuse (Conduction). This is good because it puts that heat very close to the blackness of space with little in the way of CO2 density to impede. The extinction range of CO2 at 30,000′ is far greater than it is in a laboratory where they haul in great wealth to misunderstand all this. The consequence is that surface energy which radiates only from GHG’s by the way, are radiating madly into space. It is worth knowing that the GHG’s radiate like a point – in all directions, but at altitude more directions point at stars than our Earth and so more leave the system with fewer intervening hops among GHG molecules. And this radiated heat will include in large part that heat that started the process in the first place.
So you are ultimately correct but not perfectly accurate, perhaps. I do agree with you that T-storms on average will move net positive energy out of the Earth system by convoluted means powered by the sun.
As you know we cool yachts and workboats by piping cabin heat into the ocean. You wouldn’t be, but lay people would be amazed what a refrigeration mechanic can yard in during a warm Acapulco summer. Axe me how I know.
I can see a fat carbon tax opportunity coming around the corner when Hansen puts a bug in Gore’s ear about this.

March 12, 2013 1:22 am

@Willis: you are a rainmaker!
On refrigerator without compressor see:

March 12, 2013 1:52 am

The bit about thunderstorms reminds of travelling through Greece when I was 14.
Every day about 5pm, at the end of a hot hot day, the clouds would form and POW, down came the rain. Pelting down for about 15 minutes. Then the rain had kind of had enough, all stopped, and things were cooler and of course far more humid.
This happened pretty much every day. I’m pretty sure it was Greek summer time.
But my point is, that cycle of evaporation, clouds, rain, cooling, was very very obvious. In the places where the are thunderstorms every day, or pretty much every day, then anyone reading what you wrote would be instantly familiar.

March 12, 2013 2:30 am

Your science credentials might come across a little better if you included a phase diagram and maybe some simple explanation as to how much energy would be required to move the refrigerant to the mountain top instead of just using the compression energy on site. Otherwise stick to the autobiography …

March 12, 2013 2:55 am

Superb yarn and exposition on thunderstorms! You’d have got on well with Frank Ludlam – he was so keen on weather he taught himself a lot – he was appointed as a lecturer without even a bachelor’s degree in 1951, and was a professor of meteorology by 1965 – at a time and place where being a professor was a great achievement. I was a postgrad student in his department the year before he died in 1977. I feel sure he would have had no truck with the casual alarmism which has poisoned his beloved subject. I remember him smiling at us in one of his all-too-rare lectures when we demanded guidance us as to what we should do about the threat of a new glaciation – the media scare of that time, aided and abetted by such as Stephen Schneider. He paused, smiled gently at us, and said if we really must do something, we might advise our children to build their houses 50m (I can’t recall the exact figure, but it was small) closer to the equator than our own. Now that helped us get the scare into perspective. His posthumously published book, ‘Clouds and Storms’, can still be obtained on the second-hand market ( It is a beauty.

March 12, 2013 3:14 am

Just a bit worried about having people drink -40° vodka. From living in a cold climate in my experience is that vodka left outside in -35°C will freeze to a slur, so yours must be really strong. AND downing -40° liquid may cause permanent frost damage to your throat

March 12, 2013 3:40 am

following your response to Leonard Weistein, I can assure you that the air descending along with the rain fron a cumulonimbus cloud is colder then the surrounding surface air, but not because its origin is aloft, at colder temperature.
High in the troposphere, expecially in presence of a convective tower, potential temperature is much higher than that at the surface. Every parcel of air descending from aloft would gain adiabatic heat and have a much higher temperature when it gets to the surface.
Anyway, the downdraft is colder.
Because rain passes trough drier, non saturated air before it gets to the surface.
Some of that water will evaporate again (in some very dry climate even 100% evaporates), cooling the downdraft much more than the heat gained from adiabatic compression.
Also, melting of snow/ice, higher up along the downdraft, will have a significant colling effect.
In conclusion, again, the phase change of water is what adds to the cooling of the Earth surface.

Joe Public
March 12, 2013 3:50 am

@ ed mister jones on March 11, 2013 at 9:45 pm
“But the Heat does wind up somewhere, and if that somewhere doesn’t pass the heat somewhere else yet again, then the first somewhere is going to get warmer (as happens when you run an ordinary refrigerator in an unventilated room). . . . ”
When you leave the refrigerator door open in your kitchen, it warms the room. Simply because you’re using an external power source to move the refrigerant around the system.

March 12, 2013 4:10 am

Having installed refrigeration plants on fishing vessels I was alerted to the effect of leaking refrigerants on internal combustion engines(,d.ZGU&cad=rja), so never did the lay shaft solution for refrigeration. Did lots of other lay shaft solutions in marine work boats, great fun and greatly satisfying to see designs making someone a living. (Also worked on Saturn C5, the ‘interstaging’ ring?.)
Absolutely love your writings. I find them technically very accurate and illuminating, and your life stories most entertaining.
As you point out moist air is lighter than dry air and as the rising air expands against gravity the work produced drives the weather, the giant gas turbine in the sky. (Not a mention of CO2).

March 12, 2013 4:16 am

Anyone who doubts should spend a summer on the Gulf Coast. The subtropical weather pattern is dominated, not by continental weather patterns, but by a 30% chance of “popcorn showers” every afternoon. (On time lapse images the storms pop up from nothing like popcorn.) Rarely does the thermometer reach 100F. Hurricanes have the same effect on a grand scale.

William Truesdell
March 12, 2013 4:43 am

“WARNING—this post is a 50/50 mixture of science and autobiography, call it autosciography. If that makes your brain explode, DO NOT push the button marked “Continue reading →”.
I appreciate your warnings but the main issue most had with you was not just the lack of science in your Vietnam War post, but the politics and the standard comments from the anti-Vietnam War playbook.
I have had no issues since then, so your warnings (not as mature as they could be- “makes your brain explode”) are appreciated just in case you want to pontificate politically again.
Also like it that you are now including gradations in the content (50/50).
I do read most of your posts and appreciate your exceptional insight into the problems of weather/climate. Keep it up.
W.M. Truesdell

March 12, 2013 4:47 am

Just a note on the Einsteinian refrigerator, he and Leo Szilard invented it while thinking about magnetorhydrodynamics – Einstein was both a theoretician and and experimentalist when he was young. It worked brilliantly, but the cores in the pump screamed like a banshee from magnetically induced pulses – it got quickly turned off.

March 12, 2013 5:11 am

I think you would have got on well with Frank Ludlam. He taught himself a lot of meteorology. He was appointed as a lecturer without even a bachelor’s degree and became a professor a few years later – at time and place when getting to be a professor was a very big deal indeed. I was a postgrad student in his department the year before he died all too young in 1977. I like to think that he would have scorned the casual alarmism which has poisoned so much of his beloved subject in the decades since. I remember at the end of one of his all-too-rare lectures, we students pestered him for guidance about what to do in response to the then current scare in the media about the threat of an imminent glaciation – a media position aided and abetted by such as Stephen Schneider. He paused, smiled gently at us, and said if you really must do something, then you might consider advising your children to build their houses 50m (I can’t be sure of the number he used, but I do recall it was small) closer to the equator than your own. Even that, he hinted, might be a gross over-reaction given the level of knowledge on the topic. His posthumously published book ‘Clouds and Storms’ is still obtainable on the second-hand market, and is a beauty.

John Shade
March 12, 2013 5:16 am

I think you would have got on with Frank Ludlam. He taught himself a lot of meteorology. He was appointed as a lecturer without even a bachelor’s degree and became a professor a few years later – at time and place when getting to be a professor was a very big deal indeed. I was a postgrad student in his department the year before he died all too young in 1977. I like to think that he would have scorned the casual alarmism which has poisoned so much of his beloved subject in the decades since. I remember at the end of one of his all-too-rare lectures, we students pestered him for guidance about what to do in response to the then current scare in the media about the threat of an imminent glaciation – a media position aided and abetted by such as Stephen Schneider. He paused, smiled gently at us, and said if you really must do something, then you might consider advising your children to build their houses 50m (I can’t be sure of the number he used, but I do recall it was small) closer to the equator than your own. Even that, he hinted, might be a gross over-reaction given the level of knowledge on the topic. His posthumously published book ‘Clouds and Storms’ is still obtainable on the second-hand market, and is a beauty.
[Note to moderator/ I have tried twice to get this in via WordPress log-in, but each time I failed to get a system acknowledgement. My apologies if they, and this, are in some queue for you somewhere. I hope you will delete the least well-written of the three if so, and if you let one through at all]

Duke C.
March 12, 2013 5:25 am

Here in the Mojave Desert, we’ve been using a variant of this principle for as long as I can remember: Swamp Coolers. Nothing more than an enclosed circulation fan that pulls hot, dry desert air over a series of pads saturated with water. The cool moist air moves through the interior where it absorbs heat, then exits through an open window on the far side of the house. Biggest advantage is no compressor or exotic refrigerants. Only disadvantage is when the late summer monsoons blow in, which raise the relative outdoor humidity.
These units are efficient enough to keep the interior ~30F cooler than the outside temp, even when it’s 110-115F.

March 12, 2013 5:26 am

Sensible heat is the heat difference you measure with a thermometer. The phase changes in the refrigerant are called Latent Heat of Vaporization and Latent Heat of Condensation. The atmosphere is an ‘open refrigerant’ as there is no ‘piping’ separating the working fluid during these changes. This same system is at work in the ocean with high temp/pressure gases at sea floor vents, which condense at the sea floor and evaporate in the water column or at the surface. Massive amounts of air move veritcally within thunderstorms, creating vertical wind shear. The Daedalus cycle would never work as the thermal gradient can never overcome the liquid and gas line friction losses. Man has designed systems using the Latent Heat of Solidification, called “Heat of Ice” systems which use a smaller system working 24 hrs a day and storing needed daytime peak load as ice. There are also phase change systems which “store” thermal energy using salts. Check the for the thermal capacity of water, it is double that of any other listed substance. We are blessed to have a planet covered with a life giving refrigerant.

Mike M
March 12, 2013 6:25 am

I also think of it as a feedback mechanism working in a localized fashion. In the tropics, the time thunderstorms occur is on average between ~3PM to ~5PM after the sun has heated up the surface – not at 9AM, not at 8PM. I think it explains why the tropical areas exhibit the least ‘global warming’, they have the water to enable negative feedback as well as positive feedback. The hotter it gets just after noon – the bigger the thunder storms. At night the water vapor helps hold the heat in. Desert conditions = no feed backs either way = wild temperature swings.
And, as DP above noted, I’ve been screaming for years about how the latent heat is going up to altitude totally unimpeded by GHG’s along the way. The simple measurement of how much heat is re-radiated up there is precipitation. Every drop of liquid water that comes down precisely equals a drop’s worth of latent heat that went up as a vapor; the only variability in the net outbound radiation is how high it went before it condensed.

Josh C
March 12, 2013 6:34 am

Governor system, or a ‘Regulator’ as I understand it – I think the process is only really screwed up by the fact we have large land masses that interfere with the normal system. Lacking the ability to rapidly generate clouds due to insufficient available water means the land mass (and related land covering) is the real culprit for disorganizing the system from being a regulated constant max temperature no matter what the input.
Thanks for the insight – and the stories. Having been self taught on quite a few subjects myself I enjoy hearing it in action.

March 12, 2013 6:37 am

As you’ve point out in other posts and imply here, Bejan’s Constructal Theory describes how this and other natural systems tend toward the most efficient ways for moving energy and mass through the flow channels. Empirical evidence has piled up in so many cases that it’s very hard to argue, unless you’re blind to it, that human-caused alterations of the environment would be globally catastrophic. Local disaster, yes. It’s plenty easy to make a mess like Chernobyl or the Dust Bowl. But even then the natural systems get to work repairing and restoring flows to optimum levels. There’s an inherent characteristic of systems to revert to the mean. Equally fascinating is the capacity to temporarily diverge from the mean a bit and provide some variability and thus more “surface area” (a/k/a “niche space” for occupation.

Gary Pearse
March 12, 2013 6:41 am

Willis, you are indeed a “can do” man and an inspiration for us comparative wafflers. You could have easily gone right into the refrigeration business (or maybe any business with your confident attitude and smarts). I’m an engineer (and geologist and economist – didn’t know what I was going to end up doing) and here is my refrigeration story. I worked for the Geological Survey of Nigeria (GSN) in the mid-60s and was sent off to the very dry and hot northwest province of Sokoto. When I went to “stores” to get my gear, I noticed a few small (2cubic foot?) kerosene refrigerators – you know the magical ones that use fire to freeze the icebox? I learned that, in the past, some geologists had used them in the field. I decided to take one after trimming the wick and firing it up to be sure it was working. Here is an explanation of how it works.
Set up in my camp – two tents and refrigerator in a shaded area- myself and the GSN driver, I would go out for a day’s mapping, returning to the luxury of cooled foodstuffs and a tiny ice cube tray. It was in the middle of Ramadan (no food or drink while the sun is up for the faithful) and news of my magical refrigerator spread widely, finally reaching the Emir of the region. He sent an envoy to ask me if I would kindly give him some ‘kankale’ for his drinking water for after the sunset, to which I agreed.( pron.”kunkelly” it is the Hausa word for hail to which they attribute magical and medicinal values (scorpion bites, etc) and which they gather during a storm and place in straw lined covered pits (a village of people can probably gather a ton or so from a good storm). I also put my thermos into the fridge to have it cool for the job. The envoy would return in the evening for the kankale. Wouldn’t you know it, that I returned a couple of hours before sunset to find black smoke coming out of the fridge chimney! Yoikes, I checked the ice cubes and found little platelets of ice floating in the melt water. I had to trim the wick and get it burning properly and I had also to fan the condenser vigorously for an hour – I didn’t want to disappoint! At the appointed hour, the envoy arrived on foot with a small gourd to receive the ice. I quickly put the not quite totally refrozen cubes into the thermos, got driver, envoy and myself into the Landrover and raced off to the Emir’s main village where the Emir was waiting with an enamel cup of water in a crowd of villagers there to witness the event. I removed the thermos cork and poured with a musical tinkle the woefully thin wafers of ice into his cup. Everyone cheered as he downed the water which at least had been partially. After that I didn’t want for firewood, bananas, the odd chicken (live), etc.

Mark Bofill
March 12, 2013 6:48 am

Greg Goodman says:
March 12, 2013 at 12:51 am

You repeatedly say how any hotspots provoke a storm that cools them. I can’t imagine a simpler explanation of what a negative feedback does.
I don’t disagree with you strictly speaking, but I took Willis’s point to be that tropical thunderstorms shouldn’t be evaluated as a ‘traditional’ negative feedback because of the localization, quick response time, and lack of dependence on forcing of the phenomenon; basically the whole ‘governor’ mechanism thing.

March 12, 2013 6:57 am

Seen in the classified ads in our local (U.K.) newspaper:
‘For sale. Refrigerator. Not working; otherwise good condition.’
Make a good bookcase, then…

Matt Skaggs
March 12, 2013 7:01 am

I don’t remember a “governor” being a fundamemtal building block in control systems theory, so I for one am very interested in hearing how it is different from a feedback.

March 12, 2013 7:06 am

Man, you can sure spin a yarn! Enjoyed this immensely! As to the people not ‘getting’ the “Not a feedback” bit — the Chicken Littles of the world have all warped the definition of the term to mean solely positive feedbacks — as always, their motto remains “If the data does not fit, you must omit.”

March 12, 2013 7:11 am

Also the fact the icy top of a thunderhead spreads out as the “anvil-top” spreads out that air in a manner that effectively radiates the heat into space.
I always suspected that the model’s “tropical hot spot” (which shows no sign of actually existing) was created because some agenda-driven modeler knew, down deep in his soul, that tropical thunderstorms governed the tropical heat, and therefore he wanted to create a hot “lid” that would prevent those thunderstorms from occurring, because otherwise Global Warming would fail to happen within his model.
(Sorry, but those modelers activate my suspicious side.)

Greg Goodman
March 12, 2013 7:12 am

First time I’ve heard mainstream recognition of neg feedback in Arctic:
The smaller relative decline in winter volume highlights an interesting “negative feedback”.
“Thin ice grows more quickly than thick ice in the winter. Ice acts as an insulator – the thinner the ice, the more heat can be lost to the atmosphere and the faster the water beneath the ice can freeze,” Dr Giles told BBC News.

John Shade
March 12, 2013 7:19 am

The ’50m’ is my comment above denotes 50 metres (and not miles). I have only just spotted the ambiguity.

John Shade
March 12, 2013 7:21 am

Oh dear, my comment appears 3 times. If the moderator is watching could you please delete the second two?

Mark Bofill
March 12, 2013 7:33 am

Matt Skaggs says:
March 12, 2013 at 7:01 am
I don’t remember a “governor” being a fundamemtal building block in control systems theory, so I for one am very interested in hearing how it is different from a feedback.
I’ll admit that the more I think about it the hazier I am on what Willis is getting at as well. I think the idea is that the traditional notion of feedbacks oversimplifies the situation too much, but I’m looking forward to hearing him talk about it in more detail as well.

March 12, 2013 7:41 am

Per Wiki….”Daedalus is a fictional inventor created by David E H Jones for his column in ‘New Science’ and ‘Guardian’….prima facia science evidence of fiction.
“Heat of Ice” thermal storage systems around since the 1990’s are described here…
Propane absorption refrigerators were a common item in the 1950, made by Seval, later owned by Arkansas Gas. Slow cooling but known to operate for 50 yrs without service. York Air Conditioning Company made absorption chillers for areas with low power availibity. I did drafting on such a system for the Bank of Monterrey (Mexico) in the sixties.
The average culumlus cloud weighs 800 tons and is kept aloft by INTERNAL vertical wind shear. As an errant student pilot i flew solo through a series of these clouds, where alternating bands of raindrops pounded the top, then the bottom of the wing. These direct observations of internal cloud conditions are in “Science Goes Over Under, Inside Out”. The droplet size increased during the decent. The vertical wind shear made the pitot tube and static port instruments useless. Life experience trumps classroom hypothesis everytime.

March 12, 2013 8:03 am

Another good story willis. I am old, but still capable of learning, thank you.

Mike M
March 12, 2013 8:08 am

Matt Skaggs says: I don’t remember a “governor” being a fundamemtal building block in control systems theory, so I for one am very interested in hearing how it is different from a feedback.

There’s appears to be no problem with your memory because governors do not exist in theory, they exist in reality. A real control system is a device based on control theory – not control theory itself. A ‘governor’ is a control device exploiting a negative feedback, (e.g. – a throttle governor used in an IC engine), a thermostat is another, etc.

John Stover
March 12, 2013 8:16 am

Willis, thank you for a very succinct explanation of refrigerators and thunderstorms. Has added greatly to my own understanding.
In my glider flying days in North Yorkshire (RAF Dishforth) I became intimately aware of the power of the atmospheric heat engine you described. Early in the summer mornings there is almost no wind as the sun hasn’t yet heated the ground enough to prime the heat engine. The sky is a boring blue. By about four hours after sunup things start popping and you see the little puffs of cumulus appearing. A little later the puffs become more numerous until the clouds join and begin to form stately drifts with their bases lined up parallel to the wind and all at the same altitude, where the dew point and the air temperature coincide. In summertime Yorkshire that was 3,000-5,000 feet. Since each of those clouds would be linked to the ground by a rising column of air conditions were then suitable for a glider pilot to scratch around looking for lift.
At that point it was “pilots man your planes” and we would start winch launching gliders. Three thousand feet of steel cable attached to a WW2 surplus barrage balloon winch spun up gets a glider airborne in about 75 feet. Yank back on the stick to maintain a 45 degree climb and you are at a thousand feet in about a minute. The cable ring slides off the nose hook and you are away.
People who haven’t flown in the rising air column from ground to cloud can’t truly imagine how powerful it is. The most important instrument in a glider is your variometer (vertical speed indicator in American.) It shows if the air around you is rising (good) or sinking (bad). Sometimes the needle would peg at 10kts up and you could climb to cloud base in only minutes. If it is strong enough, you don’t even have to bank to stay in the column. It just carries you straight up like an elevator.
By 1300 or so the clouds have lined themselves up in streets and you can work your way cross country by soaring to a cloud base and then jump the gap by flying across to the next street and gain back the height you lost at the next cloud.
Alas, all good things come to an end and at some point the connection from the ground to the cloud is broken, the cloud begins to disintegrate, and your lift collapses. The variometer tells the sad story of sink and you have to scramble to get back to your field before you run out of altitude.
If the lift has been very strong and it is getting later in the afternoon you can see the clouds getting dark and angry looking on the bottom. They start morphing into a cumulonimbus pattern and it is definitely time to go home because the more energetic the rising air, the more powerful the downdrafts of sink outside the column of lift. If the cloud base begins to dish upward you could find your aircraft being sucked into the into an incipient thunderstorm. Gliders have long wings and are very susceptible to torsional forces from asymmetric lift/sink which can tear them off. Also never-to-exceed speeds are usually only about 90 knots so airframe damage is why glider pilots usually wear parachutes.

March 12, 2013 8:22 am

Nice post, Willis…But I think this paragraph is where the problem lies in terms of interpreting this as somehow getting around the standard picture of AGW:

• Once inside the cloud, the air is moved vertically through the center of the cloud to the upper troposphere. During the passage is shielded from interaction with the lower troposphere. It dodges the lower region where the air is dense, and moves rapidly to the rarified upper atmosphere. There, where both CO2 and water vapor are low, the air is free to radiate to space. Rather than radiating the energy from the surface through several recaptures and re-emissions from water vapor and other GHG gases, the barrier is breached mechanically by physically and very rapidly moving the air from the surface to up near the troposphere … don’t move the radiation … move the air.

But…the lapse rate feedback is already included in the climate models. I don’t really see how you have introduced anything new here. You might argue that the lapse rate feedback is being underestimated, but that would mean that the upper troposphere would have to be warming even faster relative to the surface than the models predict. As you know, if anything the data shows it to be the other way around; this is likely attributable more to data issues than model issues, but I don’t think there is much room for saying that the models are underestimating the so-called “hot spot” in the tropical troposphere!

Mickey Reno
March 12, 2013 8:32 am

Great story, Willis. I’ve never slept on the anchor chain, but a nice pile of nylon lines, properly arranged, can make a lovely bed or recliner.
Do we have any idea of how much infrared is transmitted into space by a thunderhead?

Chuck Nolan
March 12, 2013 9:37 am

Willis your diving stories reminded me of this:
Ready for the open ocean dive the instructor asks his blonde student “Why do Scuba divers always fall backwards off their boats?”
To which the blonde replies “If they fell forward, they’d still be in the boat.”

March 12, 2013 9:43 am

Often when standing out on the apartment deck in the Solomon Islands at night, the first sign of the approach of a thunderstorm would be the arrival of the cool entrained wind.

This I have experienced many times. It is the sign to rush indoors and shutter your windows before the great downpour. The air is fresh, just like in the open sea. Sometimes, the first drops of rain come in a light wave; then it seems as if the rain is subduing, then kaboom, the heavens unleash.

March 12, 2013 9:46 am

Also the sudden drop in temperature when you get this vertical downdraft is quite amazing.

Gary Pearse
March 12, 2013 9:57 am

Willis, a possible way to move the solid from the melt/evap phase changes back to the condenser is to have the refrigerant in a closed container that could mechanically be raised (or lowered or horizontal action or rotated on a shaft away from freezer to condenser/compressor) to the freezer box to melt the ice and evaporate the liquid and then head back to where it can be compressed and cooled. A double acting piston might be more efficient in the rotating alternative to have the expanding gas in the cooling end assisting with compression in the condensing end. Engineering might have some challenges, though.

bobby b
March 12, 2013 10:24 am

Excellent instance of pure teaching. Concepts that have fuzzily evaded me for my 55 years now make sense. Thank you.
Two questions:
– “Moist air is lighter than dry air.” – This seems counterintuitive to me. True when temps are equal?
– “The system responds to temperature.” The response affects temperature. Isn’t this feedback?
Again, wonderful post, to my ultimate benefit. Thank you.

March 12, 2013 10:28 am

The miniscus is getting thicker…:) I made a comment about water steaming(evaporating) at a lower temperature than boiling point(100 C) and was refered by another commenter to this informative link:
Where i learned about phase changes and sensible/latent heat of water and i still think that that vapor is steam and water steams(vaporizes/evaporates/phase changes) and cools the air and surfaces it has contact with.That’s what clouds are made of,and it sure isn’t 100 C anywhere on the surface and if it was there would be a cloud forming(like above a cooling tower).and i know that it has to do with dew point and relative humidity and that just makes the cloud visible or not.(contrails?,fogaggedon?thermal lows?)Water is sure a magical compound and we’re lucky we have a lot of it to work with.
Thanks for the interesting articles and comments

G P Hanner
March 12, 2013 10:32 am

I have one experience at detecting the conversion of latent heat into sensible heat as it is done by Mother Nature.
We were picking our way through the tops of a middling strong Tropical Depression out over the South China Sea, about half way between Danang and Manila. We were at FL410 (about 41,000′) and picking our way through the CBs (cumulonimbus) that ran up another 5 to 10 thousand feet above us. We were in clear air, but everything below us was solid cloud cover right down to the surface.
At some point our air conditioning system, which worked by taking ambient air and both dehumidifying and warming it to something much more comfortable than the -50 degree C it was at FL410, started pumping lots of very warm and humid air into the cabin of that KC-135A. First, ambient air at -50 C is not usually very humid; second the warmth of the inflowing air was considerably more than it had been a minute or two before. A minute or so after the warm humid air started flowing, there was a loud boom. One of the J-57-59 engines swallowed a big chunk of ice that had built up on the engine inlet, overwhelming the engine anti-ice system. Fortunately, the J-57 engine was a very rugged engine and it took the ice with no discernible damage.
And then the warm humid air stopped flowing and everything was back to normal.

March 12, 2013 11:00 am

Willis Eschenbach says:

No, this stuff is not in climate models, Joel, because it’s all sub-grid. No, there’s no independent mobile refrigerators in climate models. No, there’s no way in a climate model for air to go from the surface to the top of the troposphere without interacting with the lower troposphere all along the way.

I agree that it is sub-grid. But, it is not correct to say that sub-grid processes are not described at all. They are just described in some average way. Now, that may present problems in some cases, but I don’t think you have demonstrated it does so here.
In particular, your notion of air going from the surface to the top of the troposphere and somehow bypassing things has an empirically-testable consequence: Since, at the end of the day, the only way that energy gets transferred out into space is via radiation and radiation occurs according to the temperature, then what you describe has to manifest itself as higher temperatures in the top of the troposphere in order to radiate the energy away.
I don’t really see how you can easily get around this fundamental fact. (The only conceivable way that I can envision is something like the following: Different parts of the atmosphere with different concentrations of water vapor will have different optical thicknesses and hence different amounts that a certain volume radiates. So, I imagine it might be possible to argue that it is not that the upper troposphere is overall warming faster, but that the wetter regions of it are or something of that sort. If this is your hypothesis then it ought to be testable…and one would also need to understand more about what climate models currently predict for the variation in temperature and in humidity in the tropical troposphere.)

Bob Maginnis
March 12, 2013 11:53 am

I’ve been reading your excellent stories, have had similar experience refrigerating fishboats, anchored my sailboat off Makena for a few days when it was a wild beach, etc. One issue with the Nairobi passive reefer system is that the sea level evaporator needs a low enough pressure to allow evaporation, but a 10,000 foot high column (on what we usually call the suction side) will have, even for a vapor, a gravity pressure that would be to high to allow evaporation, unless you used a working fluid much less dense than the usual ‘freons.’ Maybe CO2 with a high evaporator pressure. BTW, During previous interglacials, the World temps are said to have been 3 C. warmer than today, and I presume thunderstorms existed back then, so however well thunderstorms regulate temps, we can still have 3. C. warming, especially with CO2 GHG AGW.

Greg Goodman
March 12, 2013 12:04 pm

These devices range from Watt’s steam engine governor to the sophisticated microprocessor controllers found in consumer items…
Indeed Watt’s governor is possibly the origin of the term in the engineering sense. I remember seeing them on stationary engines at country fairs when I was a kid, and staring at them until I sussed how it worked.
Centrifugal force separates the bob weights which have a central pivot. This raises an arm giving mechanical displacement roughly “proportional” ( which depends on ) the rpm of the motor. That is used to control a valve in the case of a steam engine of fuel input in the case of an ICE that tends to increase/reduce the motors power. This means that as a load (plough , thresher etc) varies the engine keeps running at roughly the same speed.
The linear displacement is not strictly proportional to rmp and neither is the engine response to fuel/steam input. But linearity is not the aim and it keeps the motor running at pretty much the same rmp from idle to near full load.
In engineering terms: a negative feedback.
Now, as has already been noted, there is a strong positve feedback affecting the amplitude of the negative feedback in the T-storm , so the feedback will be very strong and decidedly non linear.
That does not prevent it from being modelled in the usual engineering terms of feedbacks if it can be adequately described.
I seem to recall an article Willis did about a year ago on ARGO float data then showed surface sea temperatures had a ceiling of about 38 C. They just hit a barrier and went no higher.
However, the negative swings seemed fairly sine wave like IIRC. It did not appear that the mechanism prevented negative swings, so I’m not sure how will the ‘governor’ idea fits.
But Willis has a talent for clear, simple explanation so it will be interesting to see his account of thunderstorms as a governor of tropical climate.

March 12, 2013 12:17 pm

I suspect (though stand to be corrected) that what Willis is getting at when he asserts this is “not a feedback, but a governor” is that a governor is a very specific form of negative feedback.
A simple negative feedback will tend to reduce (but not necessarily cancel out) the effect of a change in input to a system. A governor will tend to cancel out changes in input in order to maintain a system at a constant state. In that simple sense it’s equivalent to a negative feedback of 1.
But it will also adjust its response to EXACTLY (on average) counteract any other changes in the system that might try to alter the base state – including changes in any other feedbacks that are operating or even, to some extent, fundamental changes in the system itself. If the input varies, a governor will cancel it. If a positive feedback elsewhere in the system increases, the governor will cancel it. If a negative feedback elsewhere fails , a governor will take up its job as well.
If you have a two cylinder engine governed to a given RPM you can even disable one cylinder (a pretty fundamental system change) and the governor will STILL maintain the set output RPM provided the remaining cylinder is capable of supplying enough power.
In that snese, even though governors rely on negative feedback to work, Willis is entirely justified in drawing a distinction between a governor and a feedback.

Mark Bofill
March 12, 2013 12:29 pm

Are we overcomplicating matters in pondering Willis claim that this is not a feedback? Again, maybe in a pure theoretical sense there’s an argument to be had. In practice though, with respect to climate scientology, don’t we essentially end up just adjusting our notion of climate sensitivity to CO2 based on feedbacks? If the answer to this question is ‘yes’, then I think Willis has a point in making the distinction he’s making. This regulation system doesn’t have a darn thing to do with CO2 – it shouldn’t be looked at as a feedback if feedbacks in practice end up having some sort of implied logarithmic relationship with atmospheric CO2.
Or maybe we’re not overcomplicating matters. I keep putting off posting because I’ve got this nagging sense that there’s more to it than this. I can’t identify or articulate the significance concisely to myself. Still missing stuff. Maybe it’ll come to me later. Oh well, break times over for me, back to work.

March 12, 2013 12:31 pm

Bob Maginnis says:
March 12, 2013 at 11:53 am
[…]] BTW, During previous interglacials, the World temps are said to have been 3 C. warmer than today, and I presume thunderstorms existed back then, so however well thunderstorms regulate temps, we can still have 3. C. warming, especially with CO2 GHG AGW.
If Willis is right that thunderstorms act as a governor then increased CO2 won’t have any overall effect – it will be just another change to the system that the governor cancels.
As for quibbling over a 3 degree difference – that’s all of 1% or so. Even if we allow the whole 12 degree range from ice age to previous interglacial it’s only about a 3% variation. Anything that can hold a system as complicated as the earth’s climate stable within 3% over millions of years against all the changes in inputs (from multiple sources), outputs and changes in internal dynamics is doing a pretty good job.
In fact, it’s such a good job that it almost HAS to be a governed system!

March 12, 2013 12:44 pm

bobby b says @ arch 12, 2013 at 10:24 am:

– “Moist air is lighter than dry air.” – This seems counterintuitive to me. True when temps are equal?

I was wondering about that and I think I have it figured out. Air is made up mostly of nitrogen and oxygen molecules. Both molecules are diatomic with the weight of two atoms. Moist air gets diluted with lighter water molecules with the weight of one oxygen atom added to two protons, the hydrogen atoms.

March 12, 2013 12:46 pm

Willis Eschenbach says:

I don’t understand this. Higher temperatures than what? Than the identical situation without thunderstorms? How do you plan to measure that? If you mean higher temperatures than the surrounding air, the inside of the thunderstorm tower has to be warmer than the surrounding air or it wouldn’t convect upwards.

Higher than are predicted in the climate models that you feel are not properly representing this process.

I’ve said that when air travels from the surface to the the upper troposphere inside of a vertical cloud tower, basically it doesn’t interact with the lower troposphere. That is to say, it doesn’t exchange energy, either latent heat, sensible heat, or radiated energy, with the troposphere.
You claim that this is false … but physically, how is it false?

I am not saying that it is false. I am saying that you have presented no evidence that this is not properly accounted for in climate models. In other words, I have argued that if your picture is not correctly accounted for in the climate models, it ought to have a consequence and that consequence is that the temperatures in the upper troposphere would have to warm in such a way that the Earth can radiate the extra energy away (that it is entering the system, for example, as a result of rising greenhouse gases) without the average surface temperature rising. So, the question is whether there is any empirical evidence to support (and/or any theoretical reason to believe) that the warming would occur in this manner.
What you are essentially proposing is a negative lapse rate feedback that is so strong that it allows the Earth to radiate more energy back out into space by just warming at altitude and not at the surface. That is in essence what your “governor” is.

March 12, 2013 1:28 pm

The Inventions of Daedalus: A Compendium of Plausible Schemes

Amazon price New from Used from
Hardcover — $475.81 $49.98
Paperback — $139.38 $10.24

Pricey! I found one copy in our library system and put a hold on it. 🙂

March 12, 2013 1:57 pm

Add me to the list of those who think you are confusing the issue by saying the mechanism you discuss “is not a feedback”. I say this as someone whose academic concentration in graduate school was feedback control systems, and who has made my living for the past 25 years designing these systems.
A “governor” is an engineered negative feedback control system, usually mechanical in nature (modern electronic feedback control systems are almost never referred to as “governors”). But even the links you cite show a governor as being a type of feedback.
I believe the distinction you wish to make is this: the mechanism you discuss is not a mathematically linear proportional feedback, and analyses that model it as such are not going to come close to capturing the overall behavior. That I can totally get behind.
Most engineered feedback control systems, in addition to having a proportional feedback term – if you double the error from the set point, you get twice the restoring control action – have an “integral” feedback term. This term continually increases the restoring control action until the error is eliminated.
In a cruise control system, to use an example you have employed, if the controller employed only a linear proportional feedback control term, you could get “steady state” errors in response to different external conditions. Let’s say you had it set for 60 mph and it was holding that on a flat road with no wind. If you then hit a 10 mph headwind, or a 1% upslope, that slowed you down, this proportional term would increase the engine output, but not enough to keep you at 60 mph, but it might keep you at 59 mph.
Of course, a real cruise control system will increase the engine output even further, enough to get you back to 60 mph. This is almost always done through some kind of integral action.
In engineered feedback control systems, there is usually at least an attempt to make the control law as mathematically linear as possible, and to design the physical system under control to be as mathematically linear as possible, because this makes the analysis and verification much easier. Of course, there is no particular reason a natural system has to be anywhere near mathematically linear, and I think you have identified a strongly non-linear system here.
(Interestingly, while writing this note, I was interrupted by someone asking how a new input circuit of ours could work at a much wider voltage range than our older circuit. The key, which the guy, being accustomed to thinking in linear terms, had a little trouble getting, was that we had replaced a “linear” resistor, whose current is directly proportional to the voltage applied, with a “non-linear” constant current diode, whose current is constant regardless of the applied voltage – within reason of course).
It seems to me that you are proposing that what is happening in a thunderstorm-refrigerator employs either integral feedback action, non-linear feedback action, or both. But I think you will be better off considering these as specific forms of feedback.

March 12, 2013 2:01 pm

Willis Eschenbach says:
March 12, 2013 at 1:05 pm
Joe, that’s a point I’ve repeated over and over, with little effect, so I’m very glad to see someone with an understanding of just how amazing that feat is … particularly in a system governed by things as ephemeral as clouds and wind …
Trouble is, Willis, I’m not a climate scientist either. So what to you and me is obvious doesn’t actually count.
Of course, on that basis, when your mum told you not to run in front of traffic you should have ignored her unless she could provide a model of the forces involved in high speed collisions 😉

March 12, 2013 2:07 pm

Your explanation in Refrigeration via entrained wind is nonsense. It completely ignores the concept of “potential temperature”. Paolo M. provides a very good explanation.
Your comment above

However, the temperatures inside a mature thunderstorm may be quite different from those outside, so two parcels of air (one from inside and one from outside) would not necessarily end up at the same temperature when brought to the surface.

is correct .. except that the air inside a thunderhead is much warmer than the air outside it. MUCH! There are 3 lapse rates to consider, the Dry Adiabatic Lapse Rate (DALR) 9.8 K/km, the Environmental Lapse Rate (ELR) 6.5 K/km, and the Saturated Adiabatic Lapse Rate (SALR) variable 3.9 K/km – 7.2 K/km. As moist air rises and water condenses, the air remains significantly warmer than the air outside the rising column because the heat of condensation keeps it warm.
The true effect of potential temperature is seen when moist air passes over mountains (orographic lifting). When the air is pushed up one side of a mountain, it is cooled (via adiabatic expansion) and rain occurs. When the remainder of the air goes down the leeward side, it is heated (via adiabatic compression) to a much higher temperature than when it contained water vapor. If there had been no rain, the clouds formed as the air cooled would have simply evaporated and the temperature on both sides of the mountain would have remained the same.

March 12, 2013 2:18 pm

Dear Mr Willis Eschenbach
I loved this chapter . Again.
I remember your telling about the blast freezer for the crayfish, and the party, and your waking up on top of a coil of chain, in a fragment of that first story of yours ” It’s not about me” ( at least, it was the first of your autobiographical stories that I read here ). And I loved that story so much that I asked your permission and translated it to Spanish, and Plazamoyua put it in his blog. And I have the original in English among my Favourites list, and read it again and again.
This time, the part of the blast freezer is even better, because you explain how it works. And how tropical storms work.
I love it when you tell an adventure, ( or a tall tale ), and you explain why things in the story work the way they do.
So, Please, please, please: Will you publish all your writings, the autobiographic and the technical ones together, and with photographs and maps, and charts and all, in book form ?
I know I ask for the same thing every time I write, and I should not insist so much, but I am not getting any younger, and patience is not my strong suit…
Love from your old Spanish admirer !!!

Mark Fraser
March 12, 2013 2:56 pm

Willis, a control system typically comprises a forcing function of some kind, a (usually) negative feedback loop to counter the forcing function according to some formula, and an output which, hopefully, remains somewhat constant over a wide range of input (forcing functions) we can call this a “preset”. A “error signal” is created when the desired output differs by some amount from the preset. So much for simplistic terms.
In a typical temperature control loop, let’s use a fridge, when the measured temp goes above the preset by some arbitrary amount, said error signal is used to start the compressor, which cools the fridge until the temp is (probably) just under the preset value.
A “proportional” control system in a fridge would provide a variable amplitude of cooling in proportion to the error signal (degrees above setpoint), so as to restore the desired temp as quickly as possible, unlike the simple thermostatic control above. A typical governor acts like this, increasing the correcting signal in proportion to input (engine rpm, for example).
In proportional systems delay becomes important, both in sensing a temperature change, and in responding thereto. Who knows what the delays are in the earth’s energy balance proportional control system? Who knows the relationship between energy input changes and the balancing forces applied (more/less albedo, elevations of clouds, winds, ocean currents, etc.)? Not me, but rest assured, it isn’t linear and it isn’t instantaneous and it isn’t predictable at the moment.

March 12, 2013 3:12 pm

[replying to Willis under another name at Anthony’s request]
Per Wiki….”Daedalus is a fictional inventor created by David E H Jones”….then Wiki goes on to describe one of the Jones “inventions”….a locomotive powered by grass clipping grown between the rails. Sorry Willis, but that would qualify as FICTION. The Kilimanjaro A/C would suffer insurmountable friction line losses and would therefore qualify as FICTION. The noice-cancelling patent is credited to Amar Bose, U S Patent #4,455,675 on 6/19/84 with no mention of Mr Jones.
The tons of water vapor in clouds defy gravity, not by air flow outside the thundercloud, but from the vertical shear winds INSIDE the cloud. I am available for confidntial peer review of your work prior to posting, as we are all students that can all benefit from dialogue. My previous post name is my website address, which is VERY popular and i sign all my articles with my real name. I only use the prefix “Joseph” and the suffix “PE” to distinguish my articles on the web….you may call me Joe. I am a great admirer of your adventures and your story telling.

Matt Skaggs
March 12, 2013 3:19 pm

“However, with all due respect, in what I quoted you correctly describe a governor as a “control system”, to distinguish it from other things. And I think you would agree that a feedback in general is not a control system, or they’d call it a control system. And I think you would also agree that the refrigerators I describe above indeed ARE a control system … and ergo, they are not a feedback.”
“Control systems,” a branch of engineering, is a shorthand notation method for showing the basic functions of machinery. In control system diagrams, you use things like dashpots to represent damping, and a spring can represent a simple feedback mechanism. The climate can be easily diagramed using control system theory, and in doing so thunderstorms forming in the tropics would need to be represented by one of these basic items. The item used to represent the mechanism you describe above would be categorically considered a feedback-type device.
Saying something is a control system and therefore not a feedback is, to borrow a phrase, not even wrong.

March 12, 2013 3:45 pm

Willis, Sorry I was not clear. I guess my problem is with

This sets up a vertical entrained wind that falls right along with the rain. Since it comes from well aloft, it is much cooler than the surface air.

If “Since it” refers to the “vertical entrained wind”, then we disagree. But if it refers to “the rain”, then I misunderstood what you wrote.
In general, the atmosphere cools 6.5 K/km when measured by thermometers carried by balloons, but heats 9.8 K/km in a downdraft. As a result air at 20,000 feet (6.1 km) would be about 40 K cooler than the surface, but would produce a temperature 60 K warmer if there was a downdraft. The fact that it feels cool is 100% because the rain cools it, not because of the temperature it started at.The relatively small temperature variations within the clouds are not enough to explain the fact that the air feels cool.
Potential temperature = actual temperature + 9.8 K/km * altitude (in km)

Mark Bofill
March 12, 2013 3:46 pm

Matt Skaggs says:
March 12, 2013 at 3:19 pm

Saying something is a control system and therefore not a feedback is, to borrow a phrase, not even wrong.
Right! Willis you darn fool, what were you thinking? Every idiot knows you represent thunderstorms with spring symbols when diagramming the climate!
Matt, did you wake up on the wrong side of the bed or something?

March 12, 2013 4:21 pm

Joe says:

Anything that can hold a system as complicated as the earth’s climate stable within 3% over millions of years against all the changes in inputs (from multiple sources), outputs and changes in internal dynamics is doing a pretty good job.

Well, we know what the main thing holding it that stable is. It is the fact that as the temperature increases, the amount of radiation that is emitted increases very rapidly, as T^4 power.
The question is whether the inputs have really changed by amounts large enough that we should be surprised that the temperature hasn’t changed that much. When attempts are made to, for example, measure the change in inputs between the last glacial maximum and now and the resulting change in temperature, then the central estimate for what the climate sensitivity necessary for this is not that small. In fact, it is roughly in line with the IPCC estimates of the climate sensitivity (although there is admittedly a fair bit of uncertainty in the estimate).

Greg Goodman
March 12, 2013 4:21 pm

I thought shouting was forbidden here on WUWT. Take a deep breath and try to avoid getting draw into pointless arguments.

Greg Goodman
March 12, 2013 4:33 pm

w: James Watt did not control his steam engine by using a feedback. He controlled it by using a governor.
As I explained in detail above Watt’s governor uses feedback to perform its function. So a governor is not “a feedback” but is a control system based on feedback.
If climate models include the right feedbacks they may, one day, end up showing a behaviour that could be described as a governor.
You seem to be getting a but irritable today. Chill out and see if you can learn something from some of the posts without getting too concerned about those that are wrong. Blogs are full or wrong.

March 12, 2013 4:35 pm

I can’t seem to follow the arguments above.

James Watt did not control his steam engine by using a feedback. He controlled it by using a governor.

I was taught that a governor is a mechanical device the provides negative feedback. Are you claiming that that is not correct?
Likewise, from

The term feedback is used to refer to a situation in which two (or more) dynamical systems are connected together such that each system influences the other and their dynamics are thus strongly coupled.

Therefore, if the Sun heats the Earth, which causes thunderstorms to occur, and those then cool the Earth, then thunderstorms provide negative feedback.
Like I say, I think I am missing something.

March 12, 2013 4:51 pm

That PS about electrical energy and weather dynamics in the above comments was interesting(PE+KE=?)potential energy plus kinetic energy.By the lifting of water molecules and letting them fall back to earth.At least that’s what i was thinking when you mentioned that.Maybe that’s where lightning comes from?
Thanks for all of the interesting articles and comments

Greg Goodman
March 12, 2013 5:08 pm

Robert Clemenzi says: I was taught that a governor is a mechanical device the provides negative feedback. Are you claiming that that is not correct?
I would say that a governor is a mechanical device that uses feedback to govern (control) a system. But all this is getting to be knit-picking semantics.
I don’t think it makes much sense to talk about a governor while saying feedbacks are not involved.
Storms: initial feedback to temperature rise : increased evaporation (negative). Water vapour plus air is lighter and rises. Draws in more air to replace it which causes more evaporation, wind, which causes spray droplets and more evap. ie a positive f/b that increases the magnitude of the negative f/b.
Since this +ve does not turn into an explosive chain reaction there must be some further neg. f/b that limits it. This will be viscous drag in both air and sea water that ultimately limits wind speed and wave height.
So self-amplifying neg f/b all wrapped up in neg. f/b envelop.
Now knowing how horribly complex fluid dynamics gets when you try to write it down as an equation, I don’t know what the chances of anyone correctly modelling one thunderstorm would be, let alone the whole tropical climate system.
Maybe by the latter half of the century maths and processing power will be sufficient to do it if society has not fallen apart for some other reason by then.
One thing is for sure , we will have stopped worrying about climate change long before we manage to model it from first principals. A fools errand IMHO.

Steve Fitzpatrick
March 12, 2013 5:11 pm

“The threshold for refrigerator formation is based on surface temperature, not on CO2.”
The threshold has nothing to do with CO2 or other GHG’s; it is a temperature difference (surface temperature compared to the temperature of the air above), not strictly a surface temperature which sets the threshold for convection. If the air above is slightly warmer, then the surface has to be slightly warmer for convective transport to get going. The primary effect of GHG’s (absent any potential feedbacks) is to raise the average temperature of the entire troposphere by restricting radiative loss to space from well above the surface (mainly in the upper troposphere). The convective processes of thunderstorms (of course) act as you say, cooling any area which warms too much… but ‘too much’ is always relative to the temperature of the air aloft, not relative to some absolute temperature. Warmer air aloft means a warmer (on average) surface temperature.

March 12, 2013 5:19 pm

How could anybody possibly write any computer program that dealt with a rotating sphere with a tilted axis traversing a elliptical orbit, which is possibly not even reproducible due to orbits of other spheres gravity, having a variable period over it’s surfaces over time, subject to varying energy inputs, especially unknown magnetic effects, moderated by chaotic variable cloud induced albedo changes, including energy “trapped” by conversion to unknown quantities of mass through photosynthesis and including unmeasurable mechanical work through wind action ?
The shear complexity is stupendous and will likely never be solvable.
“settled science” – BS.

Greg Goodman
March 12, 2013 5:23 pm

Here, Willis, this is up your street. From you endless travels, are you aware of anything similar to what the Aussies call Morning Glory waves anywhere else in the world?

Greg Goodman
March 12, 2013 5:29 pm

Steve, convection is caused by temp. relative to surrounding air , not air aloft. Hot air goes up the chimney because the air in the room is denser, not because of what is happening on the roof.

March 12, 2013 5:38 pm

I thought climate science tells us that –
1. The major constituents of air are transparent to IR. Some even claim they do not radiate.
2. Most of the IR escaping to space is due to GHGs.
If these are true then climate science is a no-brainer.
99% of air cannot be heated by IR and cannot be cooled by IR – only GHGs can.
Therefore it is difficult to see why increasing the concentration of the only things that can radiate to space will not increase the cooling.
There are gaps in radiation theory.
I always find the visible radiation from “Cyalume” sticks interesting – you know the ones used in emergency where you break the glass inner tube and the chemicals mix emitting green light – a significant amount of light for quite a long time.
This occurs without generating any significant heat yet green visible light is associated with very high temperatures according to Wien’s displacement law.
Can someone give some explanation about this ?

March 12, 2013 5:45 pm

Robert Clemenzi
“I was taught that a governor is a mechanical device the provides negative feedback. “
I think this is degenerating into semantics, I’m an electronics technician, (not an engineer) and know as much about feedback, positive or negative, as any other techie.
To me a governor on a steam engine is NOT controlling the speed by “feedback” it simply increases or decreases the the flow of steam. In other words controls a valve.
What does it feed back?
As I said, it boils down to semantics.

March 12, 2013 5:46 pm

joeldshore asks about energy transferred up through a cloud.

Since, at the end of the day, the only way that energy gets transferred out into space is via radiation and radiation occurs according to the temperature, then what you describe has to manifest itself as higher temperatures in the top of the troposphere in order to radiate the energy away.

Temperature is not the ONLY thing that affects radiation. For instance, water vapor radiates in only a part of the spectrum and cloud drops use all the spectrum. Therefore, cold cloud tops radiate more energy than water vapor at the same temperature. When you include the increased albedo from the clouds, the total energy to space by cold clouds could be more than expected if there were no clouds at all.
When forming, cloud tops tend to be warmer than the surrounding air (due to the heat of condensation). As they dissipate, they will be colder (because the droplets are evaporating due to the low humidity).

March 12, 2013 5:58 pm

“What does it feed back?”
Sorry I should have said “what feed back ?”

Steve Fitzpatrick
March 12, 2013 5:58 pm

Geg Goodman,
Convection is always due to differences in density, of course. But please consider what would happen if the if the air outside the top of the chimney was less dense than the air in the chimney. Would convection up the chimney continue under those circumstances? Warming of the upper troposphere pretty much has to translate into some warming at the surface; how much surface warming takes place for a specified warming in the upper troposphere is a subject of legitimate debate. The satellite data suggest that the warming aloft is less than warming near the surfaces (the TMT trend is consistently less than the TLT trend), which indicates that the GCM’s have serious problems with tropospheric transport, and maybe some other factors as well, since they consistently predict greater warming aloft than near the surface. That doesn’t mean warming aloft (and near the surface!) has not happened; it clearly has. (

old construction worker
March 12, 2013 6:29 pm

As I have said many times- We live in a big swamp cooler.

March 12, 2013 6:44 pm

I don’t know which of the two following quotes was the best:
“Sleeping on chain, I found out, makes a man say very bad words upon awakening. If you are given an alternative mattress choice, say a bed of nails, or a small barnyard stall with two chickens and a rabid goat, I’d advise taking it. Plus it seemed that the entire Fijian mosquito tribe had taken advantage of the party to do some in-flight refueling. My body was royally whupped in the morning, big anchor chain marks and dents in my hips and side, covered in mosquito bites.”
Or, “WARNING—this post is a 50/50 mixture of science and autobiography, call it autosciography. If that makes your brain explode, DO NOT push the button marked “Continue reading →”.
You have wonderful talent in explaining complex systems. I wish I were able to force upon this nation your installation as “Primary Science Adviser” for certain and varying fields. This explanation of TStorms and refrigeration gave me no insight, but it DID give me an appreciation for the way this type of subject can be presented.
… and thanks… I will save this and hopefully be presenting it to some folks I know.

March 12, 2013 7:13 pm

Steve Fitzpatrick says:

The satellite data suggest that the warming aloft is less than warming near the surfaces (the TMT trend is consistently less than the TLT trend)

The reason that the TMT trend is smaller than the TLT trend is because those channels actually measure over a fairly broad range of altitudes and, in particular, the TMT is measured from the T_2 channel which has a significant tail into the stratosphere ( ) where, over the past 35 years, the temperatures have cooled at a steeper rate than they have warmed in the troposphere.

March 12, 2013 7:29 pm

Robert Clemenzi says:

Temperature is not the ONLY thing that affects radiation. For instance, water vapor radiates in only a part of the spectrum and cloud drops use all the spectrum. Therefore, cold cloud tops radiate more energy than water vapor at the same temperature. When you include the increased albedo from the clouds, the total energy to space by cold clouds could be more than expected if there were no clouds at all.

The first part of what you said here is right, but you forget something that makes the second part of what you say incorrect: By Kirchhoff’s Laws ( ), an object that is a better emitter at a certain wavelength is also a better absorber.
So, no, at the end of the day, colder clouds don’t radiate more energy to space than if there are no clouds. What you have at the end of the day is that the amount of emission you see from the earth at a certain frequency is a (weighted) average of the temperature (really, T^4) over the range of heights where the radiation emitted can successfully escape to space without being absorbed again. Hence, in places with high cloud tops, the emission comes mainly from those high…and hence cold…cloud tops and is less. In places with no clouds and at frequencies where the greenhouse gases (water vapor, CO2, …) are not good absorbers, you have the emission coming mainly from the Earth’s surface…where it is warmer and hence the emission is greater. This is how an IR satellite shows the clouds…The clear regions with greater emission show up as blacker and the cloudy regions, particularly those with high clouds, where the temperatures are coldest have lesser emission and show up as white.

March 12, 2013 7:34 pm

As an addendum to my last post, here is emission from the Earth as seen by satellites, with blackbody emission curves for various temperatures also shown: At any particular wavelength, the temperature of the blackbody curve at that radiance represents the weighted average temperature of the region of the atmosphere from which the radiation that successfully escaped to space was emitted.

March 12, 2013 8:34 pm

Very good article, Willis. It will have links from your “Thermostat Hypothesis” pages in ARVAL.

March 12, 2013 9:07 pm

Another wonderful story.
However, for me your theory, Willis, raises a question. Based on my experience living elsewhere I agree with your theory that thunderstorms function as a refrigeration system. However, here in Toronto, it seems that most of the time the temperature and humidity are increased following a summer thunderstorm. On a hot, sticky summer day it is a supremely disappointing phenomenon. Would anyone here know why that would be the case? I suppose it could just be due to the movement of weather fronts.

March 12, 2013 9:43 pm

Re. “feed back”.
Feel as I’m teaching my grandma how to suck eggs in this company of self confessed engineers but here goes.
Take two electronic examples both achieving the same outcome, and tell me which is using Feedback?
We have an automatic level control, monitoring the output of the last stage of an amplifier by taking a small amount of the output signal, rectifying it, and if it’s over a set level feeds this control signal back to the appropriate stage and reduces the gain, the level goes down, same in the opposite direction.
Second set-up we have a rheostat (potentiometer) directly in line with the output signal. Hypothetically we rig up some sort of crude mechanism that turns the rheostat forward or back as required thus controlling the output.
Would you say that the second set-up is using feedback?

March 12, 2013 10:25 pm

Willis, I’m not quite sure where you want to go with this. I think we are both agreed that this phenomenon cannot be explained by a simple linear feedback term. But I’m not sure that “control system” or “governor” is the best way to name the alternative.
The notebook computer I am writing this on has a heat pipe to cool the processor. When I turn on the computer, this heat pipe does not provide much cooling and the processor heats up quickly. The temperature rise is rapid until the processor gets hot enough to start boiling off the water in the heat pipe, which rapidly carries away heat to the external “universe” where the water is condensed and flows back to the processor to complete the cycle. Even though this employs only one of the cooling mechanisms of a thunderstorm, it is a pretty good analogy for the cooling mechanism that thunderstorms provide.
For me, the key in both cases is that at a certain temperature threshold, a new and much more potent cooling mechanisim kicks in, strong enough to effectively cap the temperature rise of the system in question. Looking at both regimes, this is a strongly non-linear feedback effect. I would explain it in terms of it providing a saturation effect. That is, after a certain point, additional energetic inputs do not cause any additional increase in temperature.
(I would also state the essential lack of CO2 effect somewhat differently. I would say that even if additional CO2 inhibited radiative loss in the mornings, causing more retained heat and a faster temperature rise, all this would do is slightly bring forward the time of day at which the secondary loss mechanism — the thunderstorm — kicked in, capping the temperature.)
I am inclined to believe that you are onto something significant with your thunderstorm hypothesis, that the formation of these storms puts a pretty hard upper limit on tropical temperatures.
But is it appropriate to call it a “control system” or “governor”? I don’t really think so. For most people, these terms imply deliberate engineering, and response to some sort of set point, and these things don’t exist here. This may be just a semantic quibble, but semantics can be important in getting a point across properly. To me, control systems employ feedback to make the state of a system match a desired state. I think you will always get a lot of pushback from people who at least intuitively consider a control system to imply design.
So anyway, I would explain your argument as having identified a strongly non-linear increasing negative feedback that creates a saturation capping the possible temperature that can be achieved. I am certainly open to your arguments as to why it should be considered a control system or a governor.

Frank de Jong
March 13, 2013 5:54 am

Dear Willis,
I do like your theories and I see a lot that makes sense. However, if the thermostats automatically appear at certain temperatures and humidity, how come there is so much temperature difference across the globe? Comparing two equally humid areas with different temperatures (say, England and some tropical rain forest), something prevents the thermostats from completely controlling the higher temperature area to go down to the lower temperature area’s temperature.
Is it that the thermostats only start working above some critical temperature? In that case, the colder places (England in the example) could still experience significant warming up to this critical temperature.
Or is it that the thermostats can only do a limited amount of cooling? In that case, both places could still experience significant warming, despite the thermostats.
Maybe I missed something, but I would say the amount of thermo-regulation of the thermostats is rather limited, which would undermine your conclusions 1, 8, 9 and 10.

Gary Pearse
March 13, 2013 6:56 am

Willis Eschenbach says:
March 13, 2013 at 12:16 am
joeldshore says:
March 12, 2013 at 12:46 pm
“Like others, you seem totally unclear about the difference between a simple linear feedback (whether positive or negative) and a governor.”
I seem to be having a problem with this, too. The JW type governor for example “detects” an unwanted amount of output and operates to reduce it (steam pressure). It is impotent, however, to deal with a reducing input of energy into the system which causes the system to slow, and at the extreme, stop (I suppose you could have a governor that would react to stoke up the fire in this case if you wanted to stretch the idea). Now your “governor” is a very fine metaphor and I like it very much,but, like all metaphors it has a limitation and defending it to the death, I believe, mars it. In the climate case, you have an input that automatically sets up conditions to mitigate itself – both ways – too hot, it builds a refrigerator, too cold, the refrigerator goes away and lets things warm up. Now I know, unlike the steam example, that the sun is a more or less constantly stoked fuel so the engine won’t quit and this makes the metaphor a good one, but it is also the definition, of a negative feedback (positive feedbacks – like the series electrical motor, I believe are not as important in climate, precisely because your overriding negative feedback “governor” keeps us in the zone more or less in the zone (within 3% as you say. Sheesh, semantics is a waking nightmare.
Having gotten that safely out of the way(?), I’m surprised that you haven’t applied the governor metaphor to the thunderstorm’s big sister – the hurricane – in the kind of detail it has been applied to the thunderstorm. It cleans up the big hot spots in essentially the same way.

Greg Goodman
March 13, 2013 7:20 am

Curt says:I am inclined to believe that you are onto something significant with your thunderstorm hypothesis, that the formation of these storms puts a pretty hard upper limit on tropical temperatures.
I think that was what was shown in Willis’ look at ARGO data last year.
What that data does not seem to show is a similar regulation of negative swings, so I’m not sure how far the idea of a governor can be applied.
During cold periods less storms can allow greater solar energy to hit the ocean but once the sky is clear it can help no more. The control mechanism hits the end stop. Maybe another process limits the negative excursions.
If tropics clamp the upper limit maybe poles control the cold end. Increasing ice cover area prevents most upward IR, evaporational losses. A couple of km of ice provides pretty good insulation and the only part of the planet not covered is the topics where the heat comes in.
Cryosat2 data is leading to the first recognition of a negative feedback in the Arctic, rather than the alarmists favourite “tipping point” positive feedback.
Maybe that’s why we see a bistable system, swinging between glaciation and interglacials.
As Willis has pointed out, tropical storms can provide very fast ( on an hourly scale ) response to input variations. Changes in polar ice cover take several years to build up a significant change. Though, as Alley found in Greenland core, even major swings like de-glaciation can happen in just a few decades.
Bistable systems usually indicate some kind of positive feedback at work, constrained by a more powerful negative feedback to keep the whole thing long term stable.
A combination of negative and positive feedback provided hysteresis, the latching to one extreme or the other. Positive feedbacks can also explain emergent phenomena mushrooming out of what seems like insignificant variations.

Steve Fitzpatrick
March 13, 2013 7:54 am

“The reason that the TMT trend is smaller than the TLT trend is because those channels actually measure over a fairly broad range of altitudes and, in particular, the TMT is measured from the T_2 channel which has a significant tail into the stratosphere”
There is some overlap; it is in the range of 10%-15% contamination of TMT with the lower stratosphere trend. But even taking that into account (adding 15% of the inverse of lower stratospheric trend to the TMT trend, or about +0.045C per decade) the adjusted TMT trend (~0.124C per decade) is still below that of the TLT (0.131C per decade), while the models predict warming in the mid troposphere that is >20% higher than in the lower troposphere. The models do appear unable to match the best available tropospheric profile data, which indicates they do not properly handle tropospheric heat transport. The discrepancy of the TMT with the model projected surface trend (rather than TLT) is even greater.

March 13, 2013 1:36 pm

Greg Goodman says: “Bistable systems usually indicate some kind of positive feedback at work, constrained by a more powerful negative feedback to keep the whole thing long term stable.”
I tend to agree. I am certainly not the first engineer to look at the Vostok core data and see the action of Schmitt trigger. This is an externally digital circuit whose analog innards use positive feedback around an op-amp to drive the output quickly from low to high or back. There is significant hysteresis on the input to prevent a noise spike from causing an output transition.
During a transition, the internal positive feedback dominates the effect. Once the transition starts, the output goes very quickly to the opposite “rail”. But as the high or low voltage “rail” is hit, this effect is saturated and the negative feedback that stabilizes the power supply is now the dominant effect.
In the glacial/interglacial transitions, the ice-snow/albedo feedback certainly could be a prominent positive feedback. Possibly, changes in water vapor and CO2 concentration and the resultant IR trapping changes could be significant as well.
But as the world warms coming out of a glacial period, the strength of the ice-snow/albedo positive feedback weakens. Certainly, the effect is much stronger when glaciers are at 45N latitude than at 75N. I’ve often thought that the transition stops (repeatedly!) at the point where it does – it could go further, of course – because this feedback weakens so much that it is reduced to insignificance.
At the tropical end, a mechanism such as Willis’ “thunderstorm hypothesis” could put a clamp on the tropical temperatures as well when the world gets to about present conditions.
An interesting source of hysteresis in these transitions comes from the height of continental glaciers. What falls as snow on the top of these glaciers at about a kilometer of elevation would turn to rain by the time it hit the ground underneath if the glacier weren’t there. I have seen calculations that concluded that if you could remove the Greenland glacier, it would not regrow under present conditions because it would melt each summer before it could build up the height to stay frozen.

March 15, 2013 1:24 pm

Wow, that is pretty dang cool (pun intended). Complex, too, you really know your stuff!

March 17, 2013 9:56 am

David says: March 12, 2013 at 6:57 am
“Seen in the classified ads in our local (U.K.) newspaper:
‘For sale. Refrigerator. Not working; otherwise good condition.’”
That description fits climate models.

March 17, 2013 9:58 am

As for “feedback”, it sounds to me as though the question is what a “feedback” does or is good for.
(I try to distinguish a “feedback” as a data/signal, from a “feedback mechanism” (which a governor system is), Willis distinguishes “feedback” from “controller” (March 12, 2013 at 2:17 pm, he’s thinking of a governor.
perhaps that’s not necessary if we are explicit as to causal mechanisms.
A classic governor uses the feedback signal to regulate output to a value set by the design. The classic mechanical governor mechanism controls speed, by using a measure of that speed to facilitate its control. The output value is “fed back” to the governor which adds fuel or friction to regulate speed. The feedback is applied in the negative direction.
A refrigerator uses a thermostat mechanism to control temperature to a set point. The output (internal temperature) is used by the thermostat to know when to turn the cooling device on or off.
It seems to me that in the climate context the term has been used to mean that a change in temperature results in a change in some quantity that tends to counteract the change. For example, a change in water vapour may result in more clouds which block some incoming heat or radiate it back to space.
That “feedback” in the climate might limit variation of temperature from an average, not govern it to the small range of variation we might like. What controls that average? (Ice ages being an extreme example of variation – but for some reason climate warmed again. Likely a complex control method with wide variation.)
If we are unlucky there will be some major disruption. An extreme example of a major disruption is a diesel engine exposed to a natural gas leak – that’s much additional fuel, so a means is needed to prevent runaway to destruction. (Overspeed shutdown is used near natural gas production wells.) Earth’s climate is affected by variations in external inputs – orbital variations change the heat influx (infrared), variations in other solar emissions may cause changes (spawn high altitude clouds, for example). In the past many millenia or more the variations have not been too extreme for human life to survive. (Well, it even survived ice ages because the ice did not cover the entire earth.)
The extra aerodynamic and friction drag on a vehicle whose speed increases from going downhill will work against the acceleration caused by gravity, but is not enough to prevent runaway to a dangerous speed. So you might say there is a “feedback mechanism” (higher speed causes resistance to increasing speed, but not enough to control). OTOH, a loaded parachute reaches terminal velocity that is survivable, especially if you flare just above the ground. 🙂
Of course some things are not governed closely – the loping idle of a race-tuned car engine is an example, engine speed varies quite a bit but variations are limited in magnitude. (Lessee, have to think through the carbureted engine – engine rotation creates vacuum which pulls fuel through metered orifices in the carburetor. That’s not a governed system, as variation in load is compensated for by the driver, it’s a calibrated system. Well, the cruise control is a governor that maintains vehicle speed despite changes in wind and hills?)
So, …
– let’s used feedback as a signal, negative feedback mechanism as one that counteracts, and governor as able to control (yah, I know, we have to keep in mind the IPCC’s imprecise lingo when dealing with their theories)
– recognize that inputs vary
– recognize that output may vary more than desired, due to imprecise governing or large changes in inputs (fuel – such as sunlight, load, etc.), but the system is still “governed” because output is within usable range
– some systems are stable without governing, but likely more susceptible to instability in the face of external changes

%d bloggers like this: