Guest Post by Willis Eschenbach
I discussed the role of tropical albedo in regulating the temperature in two previous posts entitled Albedic Meanderings and An Inherently Stable System. This post builds on that foundation. I said in the latter post that I would discuss the diurnal changes in tropical cloud albedo. For this I use a marvelous dataset called the TAO dataset. It is measurements from a number of moored buoys in the tropical Pacific.
Figure 1. Locations of all of the TAO buoys ever in operation. Background shows the sea surface temperature.
Sadly, despite the billions spent on “global warming”, the TAO buoys don’t have funds for maintenance. As a result, the records from some have ceased entirely. But I digress … the great thing about the TAO buoy records is that they are either hourly, or every ten minutes, or even every two minutes in some cases. This lets us accurately reconstruct the daily cycles.
To refresh your memory, my hypothesis is that variations in the timing and strength of the emergence of tropical cumulus and tropical thunderstorms act to regulate both the amount of incoming energy and the tropical surface temperature. I say that whenever there is a hot day or a hot area, we get earlier and more dense cumulus and thunderstorms. The cumulus clouds act solely by reflecting the sunlight. Thunderstorms, on the other hand, cool the surface in dozens of ways. This prevents the surface temperature from overheating.
So with that hypothesis in mind, let me start by looking at the daily air temperature cycles. Because of availability of data, I’ve used data from a string of buoys along the Equator. The buoys I used stretch from 95°W (buoy just to the left of the “E” in “Equator”) to 165°E (on the Equator northeast of Australia). Conveniently, the average temperature increases steadily along the line. Figure 2 shows the daily variations in surface air temperature for those Equatorial buoys:
Figure 2. Average daily air temperatures measured at ten minute intervals at eight different buoys. Colors represent temperatures.
Using just locations along the Equator gives me a peculiar advantage. All of the locations receive exactly, precisely, the same amount of top-of-atmosphere solar energy every single day. This means that the differences between them can’t be from different solar forcing. It eliminates a variable from the equation.
Now, there is an oddity about these records, which no doubt you’ve noticed. The temperature doesn’t warm steadily during the day. Let me show you what I mean. Here’s a chart I made a while back showing temperatures at Santa Rosa, California, the met station nearest to where I live.
Figure 3. Hourly temperatures averaged over a year in Santa Rosa, CA. About 20 miles (30 km) from the ocean. The photo shows wine grape trellises.
As you might expect in a generally marine climate that usually doesn’t get much in the way of afternoon clouds or thunderstorms, the graph is simple. As the solar energy increases the earth warms. It continues to warm until around 2:00 and then starts to drop. It cools rapidly at first, then more slowly towards early morning.
However, that’s not the pattern we saw in Figure 2. Instead of a steady straight rise from dawn to noon, there is a bend or a “dip” in the rate of temperature rise. This can be seen more clearly when we look at the same records shown in Figure 2 as anomalies (variations about their individual averages). Figure 4 shows the same data as in Figure 2, but with each individual average subtracted from its respective record.
Figure 4. Same data as in Figure 2, but expressed as anomalies about the individual means (averages). Colors indicate buoy average temperature as shown in Figure 2.
Here we see a most interesting progression. The cyan (light blue) colored trace of 95°W, the coolest buoy, shows only a slight bend in temperatures from 6 am to the afternoon peak. It’s nearly straight. But as we look at warmer and warmer buoy locations, the bend becomes more and more pronounced. In the warmest five locations, there is an actual “dip”, a reduction in temperature as the day progresses.
In addition, the peak temperature anomalies start decreasing with warmer temperatures. Since there is identical solar input to all of the buoys, this must reflect some local phenomenon.
To me, the “dip” in the morning records is the clear sign of the phenomenon I described in my last post—the emergence of the cumulus clouds starting in mid-to-late morning. Through variations in their emergence time, as soon as a certain temperature threshold is surpassed these clouds “throttle” the incoming solar energy by reflecting some of it back to space. This cloud throttling effect is so strong and comes on so suddenly that in the warmer locations, the temperature actually drops despite the continually increasing morning sunshine.
However, in no case is the throttling effect of the morning albedo change sufficient to overcome the full strength of the tropical sun. This is because there is no way for these cumulus to cover the entire sky—there needs to be clear descending air around each cumulus cloud to maintain circulation. As a result, there is only so much the cumulus reflections can do … and so past noon the day continues to warm. The later reduction of the peak afternoon temperature values is due not to increased albedo but to the emergence of afternoon thunderstorms. These “chop the top” off of the temperatures, imposing a high temperature limit and preventing further surface temperature rise.
Having seen that, let me move on to another way that we can see the effect of the morning-time cloud albedo. Note that the clouds that create the reflective albedo which helps regulate the tropical temperature only emerge in response to the surpassing of a temperature threshold. Once that threshold is passed and the increased cloud albedo has come into existence, it acts to reduce the high temperatures by cutting way back on the incoming solar energy.
Given the nature of the regulation, which depends on reflecting the sun’s rays, we can make the following predictions.
• The regulation of the temperature will be stronger in the day than in the night. No sun, no reflection …
• The regulation of the temperature will be greater in the morning than the afternoon. This is because the early morning is often clear and the late morning is cloudy, whereas there are generally clouds throughout the afternoon. As a result, controlling the onset time of the cloud formation will provide powerful regulation, and generally that happens in the morning.
• The regulation of the temperature will be greater up at the warm end of the scale than down at the cool end. This is because the emergent phenomena act to reduce peak temperatures.
With those predictions in mind, I cast around for some way to visualize the effects of the thermal regulation due to clouds and thunderstorms. Figure 5 shows my solution. It is the record of the hourly air temperature from the TAO buoy on the Equator at 165 East. This is the warmest of the buoys in the graphs above (red line in those graphs).
Figure 5. Boxplots of the hourly air temperatures at 0N165E. There are 59,429 observations, or about 2,500 for each hour of the day.
A “boxplot” gives various information about the distribution of the data, including outliers. The green boxes show the range that contains half of the data (the “interquartile range” or IQR). The heavy black line is the median of the data, which is the point with half the data above it and half below. The dotted “whiskers” show a distance from each green box of 1.5 times the IQR for that data. Black crosses show “outliers”, which are data points that are further from the boxes than the extent of the whiskers.
An examination of Figure 5 shows that the predictions of the distributions are borne out by the data. First, daytime regulation, from 6 AM to 6 PM (18:00 hours), is much stronger than night-time regulation. Daytime temperature regulation is so strong that there is not one single outlier on the warm side from dawn until noon, and only one (or in one instance two) outliers in each hour from noon to sunset. In fact, daytime regulation is so strong that there are many night-time temperatures that are greater than the record noon-time temperature … go figure.
Second, the regulation is stronger in the morning than the afternoon. The variations in the timing of the albedo changes are able to oppose the sun successfully until about noon (see Figure 4). After that, the continued solar input starts driving the temperature higher, and the regulation is not as certain.
Third, it is clear from the number and distribution of the outliers above and below the row of boxes that there is extensive downward pressure on any warm temperatures. This shows the cloud/thunderstorm control system is pushing back at the hot spots, cooling them down. Nor does this downward pressure only exist on the warmest temperatures. A close examination of the location of the median line shows that the median is in the middle of the green box from midnight to dawn. But during the day, the median is high up in the green box, showing that downwards pressure from the regulatory mechanisms extends well down into the body of the data.
My conclusion is that this downward pressure is the combination of cumulus clouds throttling back solar input in the morning, and thunderstorms and squall lines moving heat from the surface to up near the tropopause in the afternoon. It is this regulation of each day’s maximum tropical temperature via a host of inter-related mechanisms that keeps the earth from overheating on a daily basis.
And as I mentioned in my previous post, my insight was that if there are mechanisms that reliably keep the earth from overheating for a single day, they would keep the earth from overheating for a million years …
I may return to these topics in a future post, I’ve only scratched the surface of the TAO data.
My best wishes to each of you,
w.
My Customary Request: If you disagree with someone, please quote the exact words you disagree with. That way, everyone can understand your objection.
Data and Code: I’ve been wrestling this for too long, I’m burnt. I’ll post up the code when I get time if someone wants it. This code a dog’s breakfast, no order, functions used before they’re defined, sections of dead code exploring blind alleys. The data, on the other hand, is from the TAO website.
What we see here is a depiction of the average diurnal cycle not of albedo, but of surface temperature based upon TAO buoy data. The former would require satellite measurements of radiance. Meanwhile, the general characteristics of the latter and the underlying physical mechanisms have been well-known since at least WWII (see Hamilton & Archbold, 1945), and the particular features of TAO data have been studied in depth by various authors for over a decade. Sadler and Schroll (1997) present a reasonable analytic model for the diurnal cycle. What’s new in Willis’ work is the terminology of “emergent phenomena” and the splendid graphics.
George E. Smith:
Avoidance of aliasing is why WMO calls for 4 daily observations of temperature at periodic standard times. The question of using Tmax and Tmin to obtain the mid-range temperature (often mistakenly called the daily mean) is quite separate from that of aliasing, because the extremal times are NOT periodically spaced.
Well samples of continuous functions, don’t have to be periodically spaced; they merely have to occur at least once in any half cycle of the highest frequency component of the band limited continuous function. And they need to occur at least once in ever full cycle of that frequency in order for an extracted “average” to not have aliasing errors.
Willis’s diurnal Temperature plots clearly go beyond a second harmonic frequency component limit, so even four samples per day, is not above reproach.
I’m not a fan of combining different data taken at different times and different places, and trying to somehow claim that you are adding information to just what the raw data records.
When they can synchronize their instrumentation so that all spatial data samples are taken simultaneously, at all spatial locations, then they might start to have some idea of what is going on.
Physical thermal processes, such as “heat” transport or propagation, depend on what Temperature difference there is at a specific moment in that region, so measuring the Temperature of two different places at two different times tells you nothing about what thermal processes will happen in that region. You are just recording anecdotal events, that are unrelated to each other.
g
Aliasing is an artifact EXCLUSIVELY of periodically sampled continuous-time signals. Shannon’s Sampling Theorem and band-limited interpolation depend critically upon the STRICT periodicity of the sampling. Granted, spectrum analysis of actual (not just averages of diurnal data) much-more-finely-sampled records typically show significant spectral content only at the fundamental (24hrs) and the second and fourth harmonics. While the WMO standard is indeed not beyond all reproach, the fourth-harmonic content nevertheless pales in relation to the lower-frequency content. It serves the practical function of providing SYNOPTIC (simultaneous) data, as well as the best basis available for determining the true daily mean temperature, rather than the daily mid-range value.
Unfortunately, the mid-range temperature is all we have historically throughout the English-speaking world when it comes to century-long records. I’ve never been a fan of a variety of ad hoc methods to bring the mid-range value into closer conformity with the true daily mean, upon which the monthly “average” temperature is based. Your stated criteria for avoiding “aliasing” in NON-periodically sampled series are patently off the mark in every respect.
“””””…..
1sky1
June 10, 2015 at 3:29 pm
Aliasing is an artifact EXCLUSIVELY of periodically sampled continuous-time signals. …..”””””
Well that would be useful to know for those people who put weather stations, to sample weather parameters, at totally random locations on the planet.
Obviously their strategy, being anything but periodic in space (rather than time) must be quite immune to aliasing noise,
This is wonderful new knowledge that you give us, that aliasing noise occurs only in time sampled data systems; well according to Shannon’s Sampling Theorem, you say that is.
Also you might point out to Tektronix, that they shouldn’t be selling sampling oscilloscopes, that use random sampling, to achieve results that can’t otherwise be obtained without difficulty.
I believe they have been doing that since I was working there in the early 1960s.
I’ll have to point out the error to some of my old colleagues when we next cross paths.
In any case, your new information explains why Dr. Mann, was able to succeed with his sampling of just a single Charlie Brown Christmas tree on the Yamal peninsula.
My library text books on sampled data systems, seem to assert that sampling works with ANY continuous function of any variable, or sometimes combination of variables.
So what is the point of spatially sampling weather data, if “Shannon’s sampling theorem” is not applicable to such sampling.
Sampling, and reconstruction from such samples, is a purely mathematical construct, and works even with no linkage at all to any physical system, so if I replace (t) for time, with (s) for space, or (lambda) for wavelength the mathematics derives exactly the same result. so sampling is not “exclusively” restricted to continuous time functions.
But evidently, people who are doing that are unaware of your new information on the subject.
g
Well, George, the very fact that you think that discrete-time sampling produces “aliasing noise” reveals a fundamental miscomprehension of what aliasing really means. As the name implies, it refers to the mathematical artifact of high-frequency sinusoids above the Nyquist frequency (N = 1/(2 delta t)) appearing as the frequency alias of a baseband frequency due to under-sampling with a FIXED sampling rate delta t. It is not the same as signal distortion or noise! For any baseband frequency f, the potential alias frequencies are specified by 2kN +/- f, for all integer k. You’ll find this explicated in any competent introductory text on discrete time systems (e.g., Freeman, 1965).
No doubt, t need not be the time variable; it can be any variable which supports a continuous process. Weather stations are, of course, NOT distributed on any spatially periodic grid, but their measurements, aside from Tmax and Tmin thermometers, are are all sampled PERIODICALLY in time, according to the WMO standard. BTW, the lack of a periodic station grid is what makes BEST’s resort to global kriging wholly inadequate, because there is no spatially-periodic data for estimating the spatial covariance function or its spatial stationarity, aka homogeneity.
Hope this helps.
Not convinced Willis. Take the 165 E bouy, do a temperature plot for days when there is little or no cloud and you still get the rapid rise from 6.00 am to about 8.00 am then the dip around midday followed by a slight afternoon rise. As the buoy also measures SW radiation averaged over 2 minutes it is easy to see when a day has little cloud.
Not sure what is going on but the diurnal temperature difference isn’t much 2m above the ocean.
I do think your hypothesis has merit over larger areas of warm ocean forming convective cloud over longer time periods.
Can’t see your objection. What happens one cm above the surface Temperature wise, is as little concern as what happens 2m above the surface.
It is the very surface itself (couple of molecule layers) that determines the thermal (BB like) radiant emittance, and also the evaporation with its latent heat content, and also the conduction from the surface to the atmosphere.
The atmosphere is such a poor thermal spectrum radiator, compared to the ocean surface or the land, that the Temperature fluctuations of the lower atmosphere, are only of interest to the local thermal (heat related) processes.
The surface drives the atmosphere; the atmosphere is not driving the surface.
g
“To refresh your memory, my hypothesis….”
So Lindzen’s work is now your hypothesis? Just because you don’t read the scholarly works doesn’t mean you are the first to come up with something.
From Lindzen’s MIT page, “His research involves studies of the role of the tropics in mid-latitude weather and global heat transport, the moisture budget and its role in global change,…..”
So go read his research. Like this for instance: D. Stevens, R.S. Lindzen and L. Shapiro (1977). A new model of tropical waves incorporating momentum mixing by cumulus convection. Dyn. Atmos. and Oc., 1, 365-425. [pdf]
Or maybe this, R.S. Lindzen and S. Nigam (1987). On the role of sea surface temperature gradients in forcing low level winds and convergence in the tropics. J. Atmos. Sci., 44, 2418-2436. [pdf]
No, Lindzen’s work is nothing like my hypothesis, and the fact that you can’t seem to tell the difference disqualifies you from trying to judge either one. Lindzen is a brilliant man in my opinion … but his “iris hypothesis” doesn’t have a lot to do with my hypothesis other than that it is another part of the complex system that keeps earth’s temperatures within fairly narrow limits.
w.
So… You haven’t read all his papers on the tropics… Thought not.
Dinostratus, you are 100% correct. I haven’t read each and every one of Lindzen’s papers on the tropics, just as I suspect you haven’t read each and every one of my posts on the tropics … so what? There is no man alive who could possibly read every scientific paper on the climate. So sue me …
If you had any actual objections to my scientific statements, you’d have raised them. My work is either correct or incorrect, no matter what I have or haven’t read.
Instead of raising scientific objections, first you falsely accuse me of stealing Lindzen’s work [ when a) his is nothing like mine and b) you can’t give a single example of what I’m supposed to have stolen ], and now you want to criticize my reading list … my reading list? Get a grip!
Please come back when you actually have something scientific to contribute to the discussion. You see, Dino, I have a rule of thumb that covers people like you:
And as a result, when some random internet popup like yourself decides to start heaving muck at the walls as a way to demonstrate his infantile frustration, I just bow out until they come to their senses and return to the science. It’s simpler that way.
Regards,
w.
“I suspect you haven’t read each and every one of my posts on the tropics … so what?”
Because you could LEARN from it. That’s what. You don’t have to reinvent the wheel. Original research takes so much longer than study. If you want to see farther, stand on the shoulders of giants. Having the ability to recognize that while you are intelligent, you are not as intelligent as everyone who has come before you COMBINED, is the beginning of adding to the world instead of duplicating what already exists.
Dinostratus June 10, 2015 at 10:08 am
Dinostratus, I’ve come into science in a curious manner. I have no formal training in science beyond a year of introductory college physics and one of introductory college chemistry. That’s it. Period. That’s my formal scientific education.
Everything else that I know, every scrap of it, I’ve had to learn from the works of others, the work that you accuse me of ignoring. Over the years I’ve laboriously taught myself enough to have a peer-reviewed “Communications Arising” of mine published in Nature magazine, along with peer-reviewed articles in several other scientific journals. So I’d say I’ve done quite well considering where I started.
Now, one thing about educating myself is that in addition to a huge mountain of past studies and papers on climate science, there is also an unending river of publications coming out of the journals on a daily basis. In all of this, it’s easy to get lost. So I have to pick and choose, admittedly based on my own rather eclectic method, just what is worth spending my precious and limited time reading.
So I read things like Bejan’s work on the application of Constructal Law to climate … have you read it? I read Ramanathan’s hypothesis about the “supergreenhouse effect”. I read Dick Lindzen’s work on the Iris Hypothesis … but there are plenty of things that I haven’t read, even by those three men. However, that doesn’t mean that I don’t want to “stand on the shoulders of giants”, for goodness sake, how do you think I got here?
So I’m sorry, Dino. I’m not objecting to learning from the past, I do that all the time, whenever I can. I’ve had to do so to get where I am.
I’m objecting to you thinking that I should follow your reading list in the matter, and that if I don’t I’m neglecting the giants of the past. I’m sorry, but I’ll follow my own reading list … and if you haven’t read Bejan or if you haven’t heard of the Constructal Law, then let me recommend that part of my reading list to you.
The only difference is, I won’t diss you if you decide to read something else …
w.
“So I read things like Bejan’s work on the application of Constructal Law to climate … have you read it?”
Hell no. Anti-entropy theories are like rear end holes. Next time my sink is stopped up I’ll yell at it, “For a finite-size system to persist in time (to live), it must evolve in such a way that it provides easier access to the imposed currents that flow through it.” We’ll see what good it does.
Okay fine. I’ll give you a gold star for self directed curiosity (really, I do) but let’s be serious. You are reinventing the wheel. Also, a little humility would go a long way with me.
I just looked up this Bejan guy. Not the anti entropy stuff but just the regular what’s he published stuff. Never heard of him before. Odd as he works in an area that is similar to mine. Moreover he has some of the same academic lineage I have. Weird.
After scanning some of his stuff, I’d recommend moving on. He seems to have recast theory thoroughly covered by BSL (the case of Le=1 stuff) as his own. He currently publishes in at least some journals that don’t have a strong reputation.
Dinostratus June 12, 2015 at 9:56 pm Edit
Well, since “this Bejan guy” was listed by ISI as one of the top 100 cited engineers in the world in all engineering disciplines, and since he has published over 600 peer-reviewed papers and 28 books, and since he has not one but two physical constants named after him (the “Bejan numbers” in both thermodynamics and fluid flow), and since he occupies the J. A. Jones Distinguished Professor of Mechanical Engineering chair at MIT, I fear that your ignorance of his work says much, much more about you than about him … but sure, go ahead and tell us he’s a non-entity.
After all, that’s why we have random anonymous internet popups, isn’t it, to give us the benefit of their lack of experience in these questions?
w.
It’s not the “…Distinguished Professor of Mechanical Engineering chair at MIT”. It’s the “…Jones Distinguished Professor of Mechanical Engineering chair at Duke”. Duke’s not even the best school in NC when it comes to heat transfer. NC State is Duke’s better.
Combining known non-dimensional numbers in the case of Le=1 isn’t new science. It’s redundant and confusing.
GoreSat just reached station. When they have the Earth-facing camera working, maybe we’ll get to watch the clouds repeat the albedo pattern from its “noon” viewpoint.
http://www.nasa.gov/feature/goddard/nation-s-first-operational-satellite-in-deep-space-reaches-final-orbit
Personal observation: In Florida, as seasons progressed from winter to summer, at first there were no thunderstorms (only occasional frontal storms from cold northern air). Then late afternoon storms near leaving work time. In mid seasons, it would rain conveniently between lunch and going home. Then, at the hottest of summer, rain at lunch was an issue.
My desk had a partial window view, and “Is it raining?” was the daily question at lunch and leaving… I ended up with 3 umbrellas so one was always near. Home, work, car. Eventually I got skill at predicting and only used one of them…
Thanks, Chiefio. That’s exactly the temperature pattern I’m pointing to. The warmer the day, the earlier the clouds and thunderstorms form.
w.
https://medium.com/the-physics-arxiv-blog/cause-and-effect-the-revolutionary-new-statistical-test-that-can-tease-them-apart-ed84a988e
Mosh, that’s absolutely fascinating, thanks immensely. If I understand their brilliant insight, it’s like this, correct me if I’m wrong.
We have a cause C with noise NC.
We have an effect E with noise NE.
Their insight was, if we look carefully, under certain conditions we can partition the noise in the effect, NE, into two parts.
One part is the noise in E.
But the other part is the original noise from C, NC, transformed by whatever process it is that connects the cause C with the effect E. So if you find that in one but not the other, you’ve separated cause and effect.
Brilliant.
However, as clever as their method is and as useful as it is likely to be, I fear that for the most part it won’t be too useful in climate studies. This is because in climate we don’t often see simple cause and effect.
Instead, the usual situation is not just what I call a “chain of effects”, a sequence of events wherein one event leads to the next with no obvious “first cause”. No, it’s worse than that.
The usual situation is what I call a “circular chain of effects”. As an example, consider the relationship between the daily temperature and the albedo in the tropics. Clearly, changes in temperature are followed closely by changes in albedo … but just as clearly, changes in albedo are followed by changes in temperature. That’s a circular chain of effects.
And this feedback in the circular chain of effects will wreak havoc on their lovely method. Consider my previous example. Any noise in the temperature will cause corresponding transformed noise in the albedo … which then will feed that transformed temperature noise plus some more purely albedo noise back into the temperature.
They do use climate variables in their example dataset, but they are only where a “circular chain of effects” is not possible. Here’s their list:
Obviously, no matter how much it rains it won’t change the longitude … but for most interesting climate relationships there is no such simple one-way causality.
Their description of their pairs data does contain the following curious claim:
The “Haigh 2007” turns out to be a book called “The Sun and Earth’s Climate” … sorry, not buying that claimed cause and effect. I wonder what their method said about it … so much to investigate, so little time.
In any case, Steven, my great appreciation for pointing that out. Do you know if there is R code available? I found the location of their CauseEffectPairs data in their paper, it’s here.
Regards,
w.
There is also one problem with their notion.
In the real physical world, that we inhabit, cause and effect are simultaneous.
That is true down to the shortest time intervals that we are able to observe, and that is some where in or beyond the atto-second region.
I don’t see how one observes “noise” or noisiness, in two different things that happen simultaneously.
g
Meanwhile, SOI increases.
Recent (preliminary) Southern Oscillation Index (SOI) values.
https://www.longpaddock.qld.gov.au/seasonalclimateoutlook/southernoscillationindex/30daysoivalues/
The ITCZ snuffed my heat wave. ITCZ birthed a hurricane, which became a TS. TS turned into TD as it came up from Baja. Then just a barely organized moisture blob. Blob got caught in mid latitude circulation and fed a cut off low. The subsequent clouds and rain took us from a heat wave Monday, to an overcast barely warm day yesterday to a rainy coolish day today.
Good essay. Thanks again.
“All of the locations receive exactly, precisely, the same amount of top-of-atmosphere solar energy every single day.”
I’m not sure the equator is all that special. From one day to another, every latitude receives almost exactly the same amount of TOA insolation. The orientation of the earth with respect to the sun doesn’t change that much in the span of a day. In that respect, the equator is no different from any other latitude. It is around the days of the equinoxes that the equatorial insolation is most steady, since the equator is in the plane of the ecliptic.
It isn’t that special, MfK. I picked those because they were on the same line of latitude.
w.
w.
I will continue posting with my thread with updates until i hear from someone.
lo
Willis wrote: “My conclusion is that this downward pressure is the combination of cumulus clouds throttling back solar input in the morning, and thunderstorms and squall lines moving heat from the surface to up near the tropopause in the afternoon. It is this regulation of each day’s maximum tropical temperature via a host of inter-related mechanisms that keeps the earth from overheating on a daily basis.”
The TAO buoys provide hourly surface air temperature, but you haven’t cited any direct information linking changes in surface temperature to changes in clouds or thunderstorms. You are inferring what “must be happening” in the sky above the surface based on your personal experiences in the tropics and your “thermostat” hypothesis. Only real climate scientists are allowed to claim that their inability to conceive of an alternative explanation for some data provides strong evidence their preconceived notions are correct. Since neither of us are real climate scientists, let me point out a phenomenon that could account to some of your observations.
Most of the time, the “skin layer” of the ocean (the top 10 um) is colder than the water immediately below because all OLR and evaporation comes from this thin layer and DLR isn’t enough to compensate. Most SWR passes through the skin and is absorbed in the top 1 m. For most of the 24-hour day, the skin layer is getting heat from the water below by conduction and convection. As SWR weakens, the surface water gets cold and dense enough to sink and initiate convection. Around mid-day, however, enough SWR may be absorbed by the skin layer that it becomes warmer than the water below and upward conduction and convection cease. The air temperature measured by the TAO buoys probably closely follows the skin temperature of the ocean.
In Figure 4, the rapid fall in temperature around 17:00-18:00 followed by a slower cooling through at least midnight could represent cooling of the skin layer from 17:00-18:00 followed by the onset of convection and a dramatic increase in the heat capacity of the ocean in rapid equilibrium with the air above. In this case, convection should cease sometime in the morning.
In Figure 4, the rapid rise in temperature around 6:00 could be due to the fact that incoming SWR is entering the skin layer of the ocean at a very shallow angle, providing a long path length through the skin layer (or the top of the ocean in equilibrium with the air). However, the same phenomena should happen near sunset.
Perhaps the TAO buoys are collecting other data that will allow one to distinguish between changes in heat transport near the surface of the ocean, clouds, thunderstorms, and alternative hypotheses.
(Infrared thermometers measure the average temperature of roughly the top 10 um of the ocean, the microwave detectors used in satellites measure the average temperature of the top 1 mm and thermometers the bulk of the ocean below this.)
True in part, because I’m laying this story out in sections. You’ll just have to wait. However, I will note that my first graphic shows the timing of the emergence of the clouds in the Pacific, which coincides exactly with the timing of the changes I showed in the TAO buoy data.
False in root and false in branch. The fact that I haven’t written up all of the studies I’ve done, or presented all of the evidence I have, doesn’t mean I haven’t done them or don’t have them.
As to the idea that it is the ocean temperatures that are making the changes, I do love how people jump on stuff without doing the actual work to find out what’s happening. The variations in sea temperatures are almost totally disconnected from the air temperatures ,,, which you’d know if you gotten the data yourself and actually taken a look.
w.
Willis wrote: “The fact that I haven’t written up all of the studies I’ve done, or presented all of the evidence I have, doesn’t mean I haven’t done them or don’t have them.”
Thank you for the reply. My apologies for not guessing that you have undisclosed direct evidence that changes in clouds cover and the development of thunderstorms limit surface warming. I should have stock to the facts and not speculate about your thinking.
I’m very interested in any data that might explain what phenomena produced Figure 4 (or often limits SSTs to a maximum of about 30 degC). I wrote above: “Perhaps the TAO buoys are collecting other data that will allow one to distinguish between changes in heat transport near the surface of the ocean, clouds, thunderstorms, and alternative hypotheses.”
You can find more information about daily changes that occur at and near the surface of the ocean here:
http://ghrsst-pp.metoffice.com/pages/sst_definitions/
http://www.researchgate.net/profile/Akiyoshi_Wada/publication/225649603_Diurnal_sea_surface_temperature_variation_and_its_impact_on_the_atmosphere_and_ocean_A_review/links/0a85e539f0d1e1c144000000.pdf
Willis wrote: As to the idea that it is the ocean temperatures that are making the changes, I do love how people jump on stuff without doing the actual work to find out what’s happening. The variations in sea temperatures are almost totally disconnected from the air temperatures…”
Ouch! When I first commented, it seemed obvious that SST and SAT were tightly linked in the middle of the ocean. I tried (and mostly failed) to find some useful information about the relationship between SST and SAT. In terms of monthly averages, the difference is usually less than 0.5 degC in the Indian Ocean. The difference is larger (1 degC) near land, when the prevailing wind comes from land.
http://www.academia.edu/5322744/Relationship_between_sea_surface_temperature_and_surface_air_temperature_over_Arabian_Sea_Bay_of_Bengal_and_Indian_Ocean
The problem is that the annual range of SSTs in tropical oceans is only a few degC. So the correlation between SST and SAT in the Indian Ocean is usually, but not always, greater than 0.5 and sometimes greater than 0.75.
Even worse, I need information about the hourly, not monthly, relationship between SST and SAT. Since all of the heat exchange between the surface and the atmosphere occurs from the skin layer, the most relevant SST will be that of the “skin layer” (measured by IR thermometer). I believe satellite sensors record SST for the top 1 mm of the ocean (the depth from which microwaves escape) and traditional thermometers record the bulk temperature at various depths below the surface (where temperature varies less between day and night). The difference between these different measures of SST (up to 2 degC) is bigger than the daily warming of the air above the surface! See the graph at http://ghrsst-pp.metoffice.com/pages/sst_definitions/.
The rate of heat and humidity transfer from the surface of the ocean to the air depends on wind speed as well as the temperature difference between the air and surface and the relative humidity of the air. With turbulent mixing in the boundary layer, you are correct to disdain speculation in the absence of data. Unfortunately, ignorance of the diurnal changes at the surface of the ocean (where heat is exchanged) might lead one to miss any relationship that does exist.
Willis wrote: “And as I mentioned in my previous post, my insight was that if there are mechanisms that reliably keep the earth from overheating for a single day, they would keep the earth from overheating for a million years …”
That’s a bit of a stretch. Figure 4 suggests that the same mechanisms are influencing the daily cycle of temperature no matter what the average local temperature is: At 95W, where the average SST is 23 degC, the average daily warming is 0.8 degC. At 165E, where the average SST is 28.5 degC, the average daily warming is about 0.7 degC. IF global warming ever raised SST at 95W to 28.5 degC, I’m not going to feel any better if the average daily warming is limited by clouds and thunderstorms to 0.7 degC.
Your interesting post showing the SSTs are rarely above 30 degC suggested that some mechanism limits warming above that threshold. The development of hurricanes apparently requires a threshold SST of 26.5 degC. The data in this post PROVES that there is NO threshold SST where daily warming is significantly limited. Perhaps you will find a limit to daily warming around 30 degC. That would be exciting.
Areas of rising air that produce clouds require other areas of descending air that almost certainly will be sunny – even in the warmest regions of the tropics. The relative proportion of cloudy and clear sky is determined by the relative rate air rises and falls. Perhaps that proportion changes with warming. Perhaps convective cumulus clouds or thunderstorms are different above 23 degC topical oceans than they are above 28 degC tropical oceans. That difference could limit warming in the tropics.
Thanks Willis. Here at 12 N in South India on the coastline I see the same thing: the hotter, the more and faster clouds develop. Hot meaning above 36 C. Thunder storms come from the strongest and longest heating and when they drop their load they cool the ground by at least 5 C or more.