Guest Post by Willis Eschenbach
I got to thinking about how I could gain more understanding of the daily air temperature cycles in the tropics. I decided to look at what happens when the early morning (midnight to 5:00 AM) of a given day is cooler than usual, versus what happens when the early morning is warmer than usual. So what was I expecting to find?
Well, my hypothesis is that due to the emergence of clouds and thunderstorms, when the morning is cooler than usual, there will be less clouds and thunderstorms. As a result the day will tend to warm up, and by the following midnight it will end up warmer than where it started. And when the morning is warmer than usual, increased clouds and thunderstorms will cool the day down, and by the following midnight it will end up cooler than when it started. In other words, the emergent thermoregulatory phenomena will cause the temperature to tend to revert to some mean, not over months or years, but on a daily basis.
Now, this is the third post in a series discussing the effects of albedo and thunderstorms on the tropical temperatures. In order they were Albedic Meanderings, An Inherently Stable System, and The Daily Albedo Cycle. This post will make more sense if you’ve read those three first and seen the Figures.
So to investigate warm and cold days what I did was to take the air temperature data from some sixty-seven TAO buoys. I sorted them by average temperature, and I started to look at them. Figure 1 shows the temperature data from one of the coolest TAO buoys, where the mean temperature is 24°C. I split the data into “warm” and “cool” days, based on the average early morning temperature from midnight to 5 AM, and then took an hourly average of the warm and cool datasets individually.
Figure 1.Cool TAO buoy, averages of the days with warmer early mornings (Midnight-5AM) and the days with cooler early mornings. Straight lines connect the temperature at midnight at the start of the day with the midnight temperature 24 hours later. “Mean” is the mean temperature of all days. “Recovery” is how much the following midnight averages have moved towards the mean compared to the opening midnight averages. “Recovery Percentage” is the same as “Recovery”, expressed as a percentage of the distance from the beginning temperature to the mean.“Warm Recovery” is how much the warm temperatures have moved towards the mean, and “Cool Recovery” is how much the cool temperatures have moved to the mean. Horizontal black line shows the mean (average) temperature of all midnights. Red and blue straight lines connect the starting and ending midnight temperatures.
My hypothesis says that the temperatures should move towards the mean. That is to say, the temperatures at midnight of the end of the day (hour twenty-four in Figure 1) should be closer to each other than the temperatures at midnight at the start of the day (hour zero in Figure 1). So I have measured the difference between the opening distance (warm-to-cool temperature difference at opening midnight), and the closing distance (warm-to-cool temperature difference at closing midnight ). This I have called the “recovery” in Figure 1. This movement towards the mean is reported both in °C and as a percentage of the opening warm-to-cool difference. I’ve also noted how much the ending midnight temperatures of the warm and cool days separately have moved towards the mean midnight temperature.
However, there’s not a lot happening in Figure 1. The temperatures are barely moving towards the mean. When the day starts out cold it seems that it stays cold, and when it starts out warm, it stays warm. There is very little difference over the 25 hour period shown (0-24). Looking at other buoys I found that at the coolest end of the TAO buoy locations, there is little indication of my hypothesized thermoregulatory mechanisms. None of the TAO buoys in the cooler locations show any significant thermoregulated recovery to the mean.
But then I looked at the records from a TAO buoy at one of the warmest locations, where the mean temperature is over 28°C. There, the situation is totally different.
Figure 2. Warm TAO buoy, averages of the days with warmer early mornings (Midnight-5AM) and the days with cooler early mornings. Straight lines connect midnight at the start of the day with midnight 24 hours later. “Mean” is the mean temperature of all days. “Recovery” is how much the following midnight averages have moved towards the mean. “Warm Recovery” is how much the warm temperatures have moved towards the mean, and the same for “Cool Recovery”.
Now, this is quite different. At the warm end of the TAO buoy locations, the warm days end up cooler, and the cool days end up warmer, exactly as my hypothesis predicts.
One of the most interesting things about Figure 2 is how rapidly the restorative forces are able to move the temperature back towards the mean. In only one day, on average the temperature at midnight moves sixty percent of the way back to the mean midnight temperature … that’s a very rapid and rigid temperature control compared to what is happening in the cooler TAO buoy locations.
To close out this part, here’s a typical record from an intermediate temperature TAO buoy, with average temperatures of 27°C:
Figure 3. Intermediate TAO buoy, averages of the days with warmer early mornings (Midnight-5AM) and the days with cooler early mornings. Straight lines connect midnight at the start of the day with midnight 24 hours later. “Mean” is the mean temperature of all days. “Recovery” is how much the following midnight averages have moved towards the mean. “Warm Recovery” is how much the warm temperatures have moved towards the mean, and the same for “Cool Recovery”.
As you can see, the recovery towards the mean in this medium-temperature TAO buoy is somewhere in between the coolest and warmest buoys. In a single day the midnight temperature moves about a quarter of the way back to the mean.
One oddity that I noted was that although in absolute terms (°C) the recovery was different between the cold and warm days, in percentage terms (for the buoys shown above at least) the recovery is about the same.
This led me to ask, what is the overall dependence of the restorative thermoregulatory forces on the temperature? To see this, I took a scatterplot. Since I wanted to also see if the warm/cold recovery percentages were different, I used a scatterplot of the warm recovery percentages and the cool recovery percentages separately as a function of temperature. Figure 4 shows how the recovery percentage is related to temperature. I have again used the average temperature from midnight to 5 AM as the dividing factor for warm and cool days.
Figure 4. Scatterplot, daily thermoregulatory response to warmer (red) and cooler cooler (blue) days versus annual mean temperature. “Recovery Percentage” is how much closer to the mean the temperature of the midnight at the end of the day is, compared to midnight at the start of the day. If it moved all the way back to the mean it would be 100%.
First, it’s clear that the strength of the thermoregulatory response is a function of temperature. There is almost no thermoregulation at the low end of the temperature scale, while at the high end the midnight temperature moves halfway back to the mean or more in the course of a single day.
Next, it’s kind of hard to see the red and the blue because there is so little difference between them. I’ve printed them transparent so when they overlap they make purple … but in no case is there any significant difference between the warm and cold recoveries when expressed as percentages. This is despite the fact that often they are different in absolute terms (°C), as is shown in Figure 5 below. I have no explanation of why this should be so. Always more puzzles …
Figure 5. Scatterplot, absolute daily thermoregulatory response to warmer (red) and cooler (blue) days versus annual mean temperature in degrees C. “Recovery Amount” is how much closer to the mean the midnight temperature at the end of the day is, in degrees C, compared to the midnight temperature at the start of the day.
Here, we see that the thermal regulatory mechanisms at the upper end of the ocean temperature range can warm or cool a single day by a third to half of a degree C, midnight to midnight …
CONCLUSIONS: Well, I can say that this result is certainly consistent with my hypothesis that there are emergent thermoregulatory mechanisms that warm up the cool days and cool down the warm days in the wet tropics.
Now, scientists have previously proposed temperature mechanisms which they thought were involved in holding the temperature down in the Pacific Warm Pool (PWP), where we find the warmest of the TAO buoys. Sea temperatures in that area are the warmest in the open ocean … but despite that, the sea temperatures rarely exceed 30°C. Ramanathan proposed a “Super-greenhouse” effect to explain this temperature limit, and Lindzen proposed an “Iris Effect”, in order to explain the strong downward pressure on the temperature in the PWP. And those proposed mechanisms may indeed exist, they are not in opposition to my hypothesis.
What I have not seen mentioned previously, however, is that in addition to there being the strong downward pressure on the temperature of the warm days in the PWP noted by previous researchers, there is also a strong upward pressure on the cool days in the PWP … and as far as I know, mine is the only one of those three hypotheses that predicts such an effect.
However, it’s a big world out there, and I certainly could have either missed or misinterpreted previous art …
Finally, my hypothesis is that the temperatures displayed above are regulated by the emergence of cumulus, thunderstorms, and organized squall lines. HOWEVER, this analysis can say little about whether my hypothesis is the actual reason for the remarkably strong daily recovery towards the mean of warm tropical ocean temperatures. All it can say is that such a powerful thermoregulative effect certainly exists, that it operates on both the cool and the warm days, and it is consistent with my hypothesis.
It does not provide evidence that the mechanism is cloud-based. That’s hard to establish with the TAO buoys because they don’t contain information on the cloud coverage. But I think there’s a way to do it, which will be the subject of an upcoming post.
w.
You May Have Heard This Before: If you disagree with someone’s words, please have the courtesy to quote the exact words you disagree with. That way we can all understand your objection.
“Now, this is the third post in a series …”
Fourth, right?
You plot the “mean temperature at midnight”, but I am more interested in the mean temperature each day. Currently we get this calculated, globally, through just two reference points, minimum and maximum temperatures. How stable is the temperature at the equator, I wonder? If we measured the temperature every 15 minutes to get 96 data points each day and took the mean of this, how stable or variant would the temperature be each day? Would cloud cover have any impact at all? Would humidity? Bearing in mind the assertion that the Greenhouse Effect is supposed to add 33C of heat to the surface and that water vapour is supposed to be by far the biggest part of that increase, do the changes in its concentration day to day, bare this out?
There is so much basic testing, in Climate science in general, that should have been done, that hasn’t been done. So many assumptions made without testing the basics.
The equator is a good place to do a lot of this. Equal amounts of day and night, with only the distance from the sun throughout the year to consider (at first, though solar activity can be added in later).
As interesting as I find these investigations into cloud cover and weather patterns, it seems to be jumping ahead of the process.
Perhaps you could also check the mean temperature for the entire day on your two separate starting scenarios and comparing that (if there aren’t 96 equal intervals of temperature checks each day go with however many there are (hopefully there will be more than two!!)) Checking for the mean temperature the following midnight, leaves me feeling I am looking at an incomplete picture.
Hi Willis
Do you have enough data to see if there are any changes in the outcome if the data is sliced by day of year (or week, or month) ?
I would be interested to see whether the recovery percentage varies with the seasons.
Unfortunately, the data has lots and lots of gaps, I wouldn’t really trust sub-setting it …
w.
Well, just model the missing data!
/postnormal
“Strong wind shear—a large change in wind speed and/or direction going up through the atmosphere—can destroy a hurricane or even prevent one from forming. The top map shows the difference from average (1981-2010) in the strength of the vertical wind shear during August-October 2013, measured between 200 and 850 millibars of pressure (roughly 4,500 feet and 38,000 feet altitude). The vertical wind shear across much of the western half of the main development region and the Gulf of Mexico was stronger than normal.”
?itok=80TABy5a
http://www.sat24.com/image2.ashx?region=am&time=false&index=7
I may be missing something:
Figure 1 has both end of the day, beginning of the day, data the same (midnight avg is midnight avg, says so in the text) Trendline is flat.
Figure 2 midnight avg end, is not same a midnight avg beginning, so you get a trend line.
Why are the 2 midnight avgs not the same? The hypothesis is seemingly proved by using a different “midnight avg” calculation.
BTW – TAO? as per Juan Slayton June 16, 2015 at 8:01 pm
And if you think about it, Tropical Atmosphere Ocean doesn’t make a whole lot of sense for a buoy system that is presumably in the water and not in the atmosphere.
Thanks.
Thanks, Rocky. The time shown at zero is the midnight at the start of the day. The time at 24 is NOT the repeat of the time at zero. It is the midnight at the end of the day.
w.
We live in a water world, most of it is covered in water, we live on a blue planet, not a green one. Without water, and the amount that is on earth and its unique properties the world would not be a place where life could flourish. I live most of my life in a cold climate, and when I learned in high school that water freeze on the surface because it get lighter at 4C I realized if not for that simple property, earth would be a giant snow ball, since without it lakes and oceans would freeze from the bottom up, and once froze not likely to ever thaw out, but since water gets lighter a 4 C and ice forms on the surface and ice itself is and insulator lake and oceans do not often if ever freeze solid. Water keeps things from getting to hot or to cold, add on the heat conveyor, that the thunder storms present water even more wondrous. Water is the control knob of climate, not CO2, to bad the morons of the world have not figured that out. Even worst some of the same moron are trying to set limits on CO2 production thinking their fools errand might make a difference, I only have contempt for such morons. Oh as a disclaimer those who are born with limited intelligence please do not take offense, my derogatory comments about morons are only for the self made morons.
Reconsidering my earlier comments about regression to the mean, I realise that your “recovery” is a temp diff over 24h, ie it is dT/dt. That maybe tested against a null hypothesis of a gaussian distribution.
It has to be fairly symmetric since what you’re doing is defined relative to the mean. However, abs vs proportional change is probably informative.
If it was gaussian distributed data they would be equally distributed either side of the mean in degree units.
Proportional change is what you would get from a negative feedback. NB I’m not talking about the non-linear processes inside the emergent phenomena, but the resulting daily averaged effect that you are considering.
I think that what you have shown is further evidence that this is feedback not regression to the mean.
Figure 4 may give some indication as to whether it is a linear or non-linear -ve f/b. It seems fairly flat <26 deg C above that there is upward curvature indicating non-linear ( stronger than linear ) feedback, with a very strong rate of change at the top end.
I'm inclined to see a very nice power law curve at the bottom of the envelop, and a more linear rise in the rest but that could be just in the eye of the beholder. It may be interesting to see whether that is due to some geographical variation. Are there two groups of data behaving somewhat differently here?
PS. the slope of a log plot will give you the power law. ( log % vs temp ).
It seems to me that there is two cut off points on the graphs and three sets of data in figure 5?
Looking at figure 4&5 there is cut of points in behaviour at 26 degree c. And one in figure 5 at 27.5 degree c.
The three sets of data in figure 5: There seems one set of results the (base line) looking like a, dare I say it hockey stick!!! Another one I would call offset 26 to 27.5 degrees raised above the base line. And one above 27.5 which seems reactive.
Could the difference be in the position of the buoys with say a under currents or prevailing winds or another factor? I am curious as to what’s going on. Any one tell me the answers please.
Thanks.
Willis, This is way out of the domain of anything I know even the slightest bit amount, so just ignore it if it is mindless. But it seems to me that you assume that all temperature changes seen by the TAO buoys are local and are the result of insolation, clouds, convection, precipitation,etc at or near the buoy. But I should think that at any given time, some of the buoys will be outside the intertropical convergence zone and in the domain of the trade winds with air from elsewhere constantly replacing the local air. I would also think that even very modest changes in temperature caused by air mass transport will affect your arithmetic.
Read the following:
http://www.pmel.noaa.gov/tao/proj_over/sensors.shtml
http://www.pmel.noaa.gov/tao/proj_over/mooring.shtml
The variable ‘control’ at the higher t locations suggests something from control theory. You need two things for control: forward response with gain (energy), and neg feedback for stability. Most systems rely on constant gain. In this case, gain is a variable because energy in system is increasing before onset of violent convection. So it’s a control system that can’t work until gain passes some threshold
Willis,
I wonder if you’re not looking at something caused by how you split the data. Statistically, most of the colder days must get warmer, while most of the warmer days must get colder. I guess this is what other posters are calling regression to the mean.
I did a quick test in R to see what completely random data would look like. It turned out that the random data also shows the two trends, so you’d probably need to add something to take out any trend effects caused by your choice of warm and cold. Then again, maybe I’m missing something, I see your code.
Here’s my code, should be possible to run it as is (sorry, I’m new to R)
# Generate random data
meas = runif( 1000, 22, 27 )
# Moving average to get something resembling a random walk
f10 <- rep(1/10, 10)
mavg <- filter( meas, f10, sides=1)
# Split in "days" of 10 hours
mavg <- t(matrix( mavg, 10, 100 ))
# Split in hot and cold, find average
num_cold <- 0
num_hot <- 0
cold <- vector( length = 10 )
hot <- vector( length = 10 )
for( day in (2:99) )
{
if( mavg[day,1] < (22+27)/2 ) {
cold[ 1:10 ] <- cold[ 1:10 ] + mavg[day,]
num_cold <- num_cold + 1
}
else {
hot[ 1:10 ] <- hot[ 1:10 ] + mavg[day,]
num_hot <- num_hot + 1
}
}
hot[ 1:10 ] <- hot[ 1:10 ] / num_hot
cold[ 1:10 ] <- cold[ 1:10 ] / num_cold
plot( 1:10, hot, col = 1, ylim=c(22,27),xlim=c(1,10),type="l" )
lines( 1:10, cold, col = 2 )
> Then again, maybe I’m missing something, I see your code.
Err, I *didn’t* see your code.
Frank says : “# Moving average to get something resembling a random walk”
You’re heading in the right direction.
Don’t use running mean ( it’s crap ) , besides what you need to do to get a random walk is integrate. Random temps would be random walk.
What is needed is to fit the parameters of your random model to the data, not just choose any random series. See my comments above. Willis’ 24h dT is rate of change so that could be modelled as AR1 ( ARIMA or some other may be better )
It looks like when scaled to the data the 24h regression would be negligible but this needs to be tested and would firm up the result.
oops, more proof reading reqd. Temps could be modelled as AR1, 24h dT should be gaussian random series, as a basic test.
Thanks Mike.
The point I was trying to make was to show that even a random set of values, when split into “hot” and “cold” intervals will always show two trends towards to average. The running average was just to get something quickly that goes up and down somewhat smoothly in the 22-27 degree range. I don’t think it matters much to the point I was trying to make. I don’t want to model anything.
F
OK, it’s good illustration of the effect.
Output looks like this (for one particular realisation):
http://i59.tinypic.com/156wmjl.png
Frank de Jong June 17, 2015 at 5:01 am
Don’t know how many times I have to say this, but I guess it’s at least one more …
IF this were regression to the mean we’d see it at both low and high temperatures. But at lower temperatures, we don’t see this at all. No significant reversion to the mean occurs at lower temperatures.
THEREFORE … I’m sure you can fill in the rest.
Nor would we really expect to see any sizable “regression to the mean”. Remember, we’re looking on a daily basis. A summer’s day doesn’t revert to the annual mean in the course of a day.
As to the code, well … you’re more than welcome to it, it is here, but I’m not sure it will do too much good. Far from being user-friendly, It’s kinda user-aggressive …
w.
Line 14 of that file reads ‘source(“Willis Functions.R”)’. Where may I find that file?
Not Sure, it’s here.
Willis Functions.R
w.
Thank you, sir. I took the liberty of uploading your code here: https://gitlab.com/notsure/willisclimate/tree/master
I’ve given you credit, of course:
https://gitlab.com/notsure/willisclimate
I can’t get it to work at the moment. I’ll make commits as I make hopefully positive changes.
Keep us posted.
w.
I’m stalled again. I commented out problematic code starting at line 143, and added some documentation (Notes.rmd). Now the call to “initialize()” at line 17 is failing. Is there another missing file?
https://gitlab.com/notsure/willisclimate/commit/dac1efaac79dbc594caf41a8cb07eed9996a407b
I commented out that problematic code and got a little further. I’ve factored out your path-handling utils into a separate file. It looks like that missing “initialize()” function downloads the TAO data to a folder
called “~/New TAO Folder Highres/”.
https://gitlab.com/notsure/willisclimate/commit/b1e572b789d07fda6e8be592f60b50c300dfe6ef
Not Sure, the “initialize” did something at one time, but not now. Now all it does is remind me to run all the various functions, which are all down near the end of the file.
w.
Thanks. I’ve had to table that for now while I work on downloading all the TAO data and putting it into the directory structure the code expects. It looks like you download the various different data sets into numbered sub-directories, is this correct? This is very much still in progress:
https://gitlab.com/notsure/willisclimate/commit/8be9d4cfde67450038f1f2f04338a798cb1025df
The directory structure looks like this
Head Folder: New TAO Folder Hires
Folders inside: “SST data”, “AIRT data”, “RH data”, etc.
The folders and files aren’t numbered, they have names. I just use the numbers to traverse them.
w.
Frank. One key error you have is that your rand series is the full range of the TOA data , not the range for one site.
If you look at Willis’ fig 1 etc the range of data are much narrower. You need to scale your integrated random walk to fit the data. Ultimately this needs to be done by matching the statistical model to the actual data, but your idea is a good first hit.
I would like to see the rate of change in SST on calm cloudless nights. SST closely correlates with dew point at the surface and the rate of change in dewpoint is a measure of the energy being radiated to space.
Wish I could make a constructive comment, but you can see what the scatter plots say as well as I can. The only way you could do better is by adding dimensions using more data, for example if the buoys were equipped with a simple full-spectrum (all the way down into the LWIR so that it includes “back radiation”, even better a split that records separately SWV and LWIR split at the midpoint between incoming solar and downwelling LWIR) photocell that gave you “top of buoy” insolation as a function of time. This is the thing that would really help you, because it would let you connect what is actually happening with incoming radiation as a function of time/temperature and vice versa. Humidity/precip as well, as I’m guessing getting actually rained on might affect temperatures, or else the peak in temperature might correspond to achieving saturated humidity near the surface.
In the meantime, ascribing causes is going to be basically impossible, because all of the causes discussed are in hidden dimensions and because the regression effect you record is a fraction of a degree and hence is associated with some comparatively weak (on average) phenomenon.
To some extent you may just be seeing the emergence of a seasonal split (it would be interesting to correlate e.g. location of the cool buoys with their interaction with e.g. a jet stream, or the dates over which they return cool behavior vs warm behavior in a scatter plot as your “split” may be purely seasonal). The warm buoys might regress to a mean because there is only one stable mean with little seasonal behavior. The cool buoys clearly have a variable mean with a temporal persistence greater than a day, which suggests seasonal or at least weather cycle persistence.
I’ve read elsewhere about the apparent clipping of open ocean temperatures around 30C. The one thing that is immediately apparent is that that energy doesn’t stop coming down at the TOA just because SSTs reach 30C. It must therefore be a critical point of some sort for a mechanism that either blocks incoming radiation, increases sensible loss via e.g. latent heat, or increases sensible loss via some sort of active transport (or all of these together). It also has to be one that has a rather sharp onset because your scatter plot is clearly the flipping of some sort of climate “switch” (or switches) in higher dimensions. It is reminiscent of the clipping of temperature in water at the boiling point until you add enough energy to break latent heat bonds and start a boil. The energy isn’t going into temperature any more, it is going into something else (but obviously not Hansen’s “boiling oceans”:-) or going somewhere else.
I like going somewhere else, but where? Does this temperature correspond to one that triggers advective rolls in the salinity downward transport cycle, so that any extra heat just speeds up the transport? That could even be the missing mechanism for the missing heat — an entire equatorial band that simply speeds up downward transport of high density warmed water directly proportional to any increase in net downward flux to maintain a constant surface temperature of 30C. More heat equals faster, not warmer.
This sort of thing wouldn’t function alone. It would have to regulate the rate of surface evaporation to match, cope with the extra water vapor produced and increase sensible heat transport laterally in that form poleward, and might actually alter albedo several thousand miles away, several weeks later to further implement your feedback mechanism. That is, the mechanism doesn’t need to be local to affect global average temperatures — increasing water vapor production in the tropics and transporting it poleward could push temperatures down (on average) in the temperature zone three weeks later by increasing the cloud albedo there by a tiny amount.
This is why climate models suck. One would love to be able to build a model for this. But how can one build a model for this? It’s not even a hypothesis. One is postulating an alteration of a probability connected to a purely turbulent phenomenon in the coupled Earth-Ocean system, one involving density not of water but of highly saline ocean water in a way that depends on the local details of its density/depth/thermal/saline profile. The turbulent rolls involved are going to be completely invisible on a scale of 100km square cells. They might have a characteristic length scale of 50 meters! The entire phenomenon I described above might involve them growing to 51 meters when the surface is receiving more energy than it loses to the atmosphere! There might be a whole spectrum of these rolls that varies with surface wind speed and direction and past history, and the entire regulatory mechanism could be a tiny shift in the spectrum.
The only way to build it into climate models is by fiat. Assert that it exists, make up some semi-empirical function that affect tropical 100 km square cells, implement it in a model, and then run the model. At which point voila! The model will produce what you built in as that function! But does this reflect reality? It is literally Impossible with caps, boldface, and italics all intended, to tell. Climate models produce an insensible shotgun blast of future climates. To even detect the phenomenon might require real time detectors in an array the same order of magnitude as the advective rolls themselves place in the open ocean, and one would expect that the existence/presence of the detector might be more than enough to disrupt the phenomenon so that they are literally not observable except with extremely sophisticated techniques, perhaps a doppler sonar, perhaps collecting signals from a few million tiny (oversized bobber sized, too small to affect the evolution of the rolls) research buoys that ride the local currents and return some sort of picture of the local flow of energy in the ocean where temperatures appear to be clipped.
We are decades away from having the data needed to solve mini-problems like this, and centuries away from having enough compute power to solve the larger problem (the coupled Navier-Stokes equations for the Earth-Ocean-Atmosphere-Sun system) that dictates the evolution of the climate at adequate spatiotemporal resolution to be able to actually model the emergence of emergent negative feedback phenomena.
This is where GCMs utterly fail. 100km square cells basically mean that they can only develop self-regulatory nonlinear phenomena — emergent dissipative structures — at that scale. 5 minute timesteps and enormous amounts of compute time mean that they cannot with any plausibility function well enough to capture (predict) the known, named long period emergent phenomena, the decadal oscillations, that clearly have an enormous impact on global climate. Yet much shorter length scales are almost certainly critical (literally) in the heat transport they are trying to model. Thunderstorms have length scales dictated by their height — a cell with an anvil that reaches the top of the troposphere has a length scale order of 10 km, and there can be a lot of variability on scales down to a km laterally by kms vertically. Thunderstorms represent a huge local transport of heat latent and otherwise straight up through the greenhouse gas barrier.
The ocean no doubt has many similar transport mechanisms that exist on length scales from meters (size of surface waves) to a few hundred meters, pretty much clipped at 700 m if not before at the thermocline. But within the first 700 meters there can be entire encapsulated rivers transporting heat coherently and spinning it off in turbulent chaotic featherings to mix with surrounding water at different temperatures. This process is almost certainly fractal down to very small scales indeed — one can see the featherings in any surface temperature representation of e.g. the gulf stream, but only at a horrendously poor resolution, and of course a lot of what happens is beneath the surface, where the thickening of a layer by a factor of 2 might not affect the surface temperatures at all but which might represent a doubling of the energy being transported at constant temperature.
One hates to plead ignorance as an excuse, but I do so plead on behalf of the human species. We are cosmically ignorant about the climate. We are decades away from enlightenment, decades that are not well-served by a research climate rife with confirmation bias and contaminated by a “religious” desire to save the world from an unconfirmable hypothesis “predicted” only by GCMs that were built for the sole purpose of predicting it.
AGW is a plausible hypothesis that is directly supported by evidence that increased CO_2 is largely anthropogenic and a single, comparatively simple theory of radiative heat transport that, as a mean field theory, predicts a logarithmic warming that is reasonably well-supported by data with a thoroughly unalarming empirical climate sensitivity. Catastrophic AGW is (so far) purely a pipe-dream of a few individual people, amplified beyond all reason and outside of all actual evidence into a trillion-dollar public trough from which the very energy companies that are supposed to be “the bad guys” gobble down billion dollar chunks, money that if otherwise invested might have (by now) brought about an end to global poverty (instead of its perpetuation), a stable and healthy national economy (instead of a fragile energy economy that at this point would collapse if the public trough were to suddenly be emptied), and hey, who knows, World Peace.
Me, I’d be content if we simply invested it in making energy cheap. Energy is the fundamental scarce resource. Given enough, cheap enough, energy, all of the world’s problems can be solved. With energy we can (if necessary) desalinate the oceans and make deserts bloom, we can make fertilizers and run tractors and combines and grow food to feed a hungry world, we can build sewage treatment systems and bring the miracle of flushable toilets and the consequent reduction in misery and disease to the third of the world’s population that uses the nearest open field or the side of a handy street as a bathroom. We could provide energy to heat food without burning charcoal or dried dung and break a cycle of dung-based contamination of food and respiratory illnesses that claim several million lives a year in the third world, most of them children.
Personally, I could care less about where the energy comes from as long as it is cheap. Cheap solar is just fine. Cheap nuclear is peachy. Cheap coal is lovely, as long as coal remains cheap. Cheap fusion is my dream — that is the technology that would allow us to run civilization for so long that we would no longer be recognizably human long before it runs out (if then — by then we can probably mine Jupiter or its moons and never run out in the lifetime of the Sun). Sure, we might die off in the meantime, or be wiped out by an asteroid or gamma ray burst or war or an engineered super-virus built by a crazy person, but at least we wouldn’t do any of the above because of a lack of the fundamental resource.
At heart, all poverty is energy poverty. The units of energy are the units of work, and work, one way or another, is wealth.
rgb
This is what got me interested in Climate.
But I have one nit
The problem is that the data is infilled based on the same hypothesis, and none of the temp series are based on actual measurement alone, so you have GCM’s being compared to temp series that are augmented by the very same hypothesis they are used to confirm.
Turtles all the way down, turtles, inside of turtles, inside yet more turtles….
“The problem is that the data is infilled based on the same hypothesis, and none of the temp series are based on actual measurement alone, so you have GCM’s being compared to temp series that are augmented by the very same hypothesis they are used to confirm.”
wrong
[for a man who professes that we all need to show our work (raw data, code, procedure, etc.) you sure have a lot of one-word responses without providing the slightest portion of what you preach -Anthony]
So you infill based on lat and elevation, and a fudge factor for weather from stations up to 1,200km away, you do exactly the same thing for 1880 as you do for 1980, and you do it before you adjust the hell out of the 1880 data right?
Which makes the value of the TCS more uncertain, but probably doesn’t change its sign or make it particularly likely to be zero. Personally I think it is somewhere between 0.5 and 2 C, most likely around or a bit over 1 C, and one reason for the size of this range is that it is very difficult to know how much of the warming post 1850 in HadCRUT4 is an artifact of the sort you describe or lost in what should be much larger error bars than HadCRUT acknowledges in the 19th century. Almost certainly more than 0.1 C, possibly (IMO) as much as 0.3 C. That would knock my own “best fit” TCS of 1.8 C down by almost a factor of 2, which would put it — perhaps unsurprisingly — at the low end of the range of the central theoretical estimates for TCS, which run from 0.9 C to around 1.5 C. The low end would simply suggest that feedbacks, if any, are on average slightly negative, not positive.
But even 0.1 C would knock it down pretty easily to the high end of the no-feedback range, around 1.5 C. This is what at least one climate model that uses the newly measured lack of any significant climate response to aerosols obtains — TCS around 1.45 C. I don’t think it would surprise anyone that even if every single adjustment made to HadCRUT has been made with the very best of will, the simple fact that nearly all of them have had the effect of cooling the past and/or warming the present has perhaps overshot the mark by 0.1 C. Not that there isn’t an acknowledged uncertainty in HadCRUT4 greater than that anyway, nearly 3x that much (and still not enough) on the early end of things.
The fundamental problem is this. What the heck, accept HadCRUT4 at face value. Fit the predicted log increase in temperature as a function of CO_2 concentration historically with no lag and no additional assumptions. It is a simple two parameter nonlinear least squares fit, a child could do it with R (and I have). One of the two parameters is irrelevant as it just matches the floating scale of the “anomaly” to the equally floating scale of the log. The money parameter is then only the no-lag equilibrium sensitivity.
When one does this, one gets a really excellent fit with a log factor that implies ECS equals 1.8 C, a bit over 1/2 of the IPCC’s central estimate. One also gets a fit that is good enough that it leaves little room for “uncommitted warming”, although of course long relaxation times wouldn’t necessarily show up in this sort of stationary fit as long as everything is smooth enough.
Given this, what is the motivation for spending a fortune on GCMs that obviously don’t work as well as the essentially one parameter fit? How much of the climate’s observed variation over the last 165 years remains “unexplained” by this fit? The answer is:
Not much.
rgb
I’m going with 0.8C +/- 0.4C (per doubling), From a little above Co2 alone (1.1C) to about half that.
I would even buy that during the day max temps are a bit higher*, but it’s completely lost at night.
* though the more I look at surface data the harder it is for me to believe what I just wrote.
rgbatduke wrote: “one gets a really excellent fit with a log factor that implies ECS equals 1.8 C”
Unless I misunderstand what you wrote, that gives a Transient Climate Response of 1.8 C. Smack dab in the middle of the IPCC range of 1.0 to 2.5 C. To get ECS, you need to account for net heat transfer into the ocean. But your response is high since it does not account for other greenhouse gases. Careful analyses of this type (such as those by Nic Lewis) do give ECS around 1.8 K.
micro6500: I’m going with 0.8C +/- 0.4C (per doubling), From a little above Co2 alone (1.1C) to about half that.
Looking only at surface energy fluxes, I calculated a surface climate sensitivity to a doubling of CO2 of about 0.3C-0.9C. I put the calculations up at RealClimate and ClimateEtc. A part can be downloaded from my web page at ResearchGate.
“At heart, all poverty is energy poverty. The units of energy are the units of work, and work, one way or another, is wealth.”
Thumbs up.
rgbatduke: The only way you could do better is by adding dimensions using more data, for example if the buoys were equipped with a simple full-spectrum (all the way down into the LWIR so that it includes “back radiation”, even better a split that records separately SWV and LWIR split at the midpoint between incoming solar and downwelling LWIR) photocell that gave you “top of buoy” insolation as a function of time. This is the thing that would really help you, because it would let you connect what is actually happening with incoming radiation as a function of time/temperature and vice versa. Humidity/precip as well, as I’m guessing getting actually rained on might affect temperatures, or else the peak in temperature might correspond to achieving saturated humidity near the surface.
Rainfall and humidity are available, but downwelling lwir is generally not, except on a few buoys for a few years.
Clearly the tropics buffer the planet.
One need not look further than the satellite surface temperature record, which is flat as a board since 1979.
The electromagnetic spectrum goes like this from short wave high energy to long wave low energy:
Gamma rays, X-rays, UV, visible, SWIR, LWIR, microwave, radio wave.
Step 1: UV, visible, SWIR from the sun heats the earth.
Step 2: The earth heats and per S-B emits LWIR.
Step 3: GHGs absorb LWIR.
Step 4: GHGs emit an equal amount of down welling LWIR that over heats the earth.
Step 5: I don’t think so.
Per S-B and Einstein’s award winning photoelectric equation balance GHGs cannot emit LWIR. Objects/molecules absorb some of the incident energy through heating, vibration, oscillation, etc. and emit at a lower energy, longer wave.
The down welling back radiation from GHGs cannot be LWIR, but microwaves, good for heating water, not much else.
And the upwelling and down welling fluxes shown as equal on so many of the popular global heat balances cannot be correct.
Really Nick?
Greenhouse gases emit radiation at the same wavelengths they absorb radiation. This is Kirchoff’s Law.
Besides, down-welling radiation is measured; it is not guessed at. Spectral plots show that down- radiation is in the long wave infrared. It is not the result of guessing, assuming or modelling – it is measured.
And yet actual surface temps don’t match the models.
Kirchoff’s law applies to electric circuits. What does that have to do with GHGs?
Sure they do. Just not General Circulation Models, not ill-tuned weather models run out to a few thousand times their useful time span. They match the simple model of CO_2 driven average warming extremely well:
http://www.phy.duke.edu/~rgb/Toft-CO2-vs-MME.jpg
Better, on average, than all of the climate models (literally) put together.
Seriously, the physics for the GHE is pretty sound, and as you can see, explains almost all (all except for a 0.1 amplitude, 67 year period sinusoid for which I have no explanation) of the observed warming of the past 165 years (I’m only displaying the last century plus, but it works over all of HadCRUT4 back to 1850 well within reason).
rgb
Oh, Kirchoff’s radiation law! Which applies to an ideal black body in an isolated system at thermodynamic equilibrium. Just exactly like the atmosphere!
NNNNNNOOOOOOTTTTTTT!
https://en.wikipedia.org/wiki/Kirchhoff%27s_law_of_thermal_radiation
Gosh, I guess Kirchoff had more than one law in more than one context. Kind of like Newton, and Gauss, and a few other early physicists and mathematicians.
Aside from Kirchoff’s Law, molecules absorb radiation according to its radiatively coupled quantum structure, and emit radiation according to its radiatively coupled quantum structure. There’s a bit of slop (from broadening and multi-molecule events), but on average the slop works out to be nearly zero.
The funny thing is that people have a problem with this even after MikeB points out — quite correctly — that the downwelling spectrum is not the result of guessing, assuming, modeling, or applying a theoretical law because it is MEASURED! Let me see if I can isolate that word for you again:
MEASURED!
Do you understand what that means? It means that denying its reality is literally as insane as denying that the sun is lighting up the earth, or denying that my weight is slightly in excess of 1000 Newtons. It isn’t badly measured, or measured with bleeding edge equipment. Measuring it is cheap and easy and several methods for measuring it have been around for a very long time. You could measure it yourself if you invested a comparatively small amount in a piece of equipment to do the measurement with.
rgb
I can only repeat what I said yesterday to ren under similar circumstances. As a physics professor and with the greatest of good will, you should never, ever post again on the subject of radiation or physics in general, certainly not unless you are going to take the time to learn some of both.
If you want to be educated — instead of just posting a short diatribe that proves that you don’t have the foggiest idea of what you are talking about — you could read Ira Glickstein’s nice article on the greenhouse effect on WUWT (I can find the link if you are interested and can’t manage Google). Or better yet, you can buy a copy of Grant Petty’s book “A first course in atmospheric radiation”, which isn’t horribly expensive. You will without question struggle with the physics if you are so confused as to state that downwelling radiation is “microwaves” and not “LWIR” and that there is some difference in what they do, or that CO_2 can interact with one and not the other, but if you get a textbook on introductory physics (I have a free one online if you can’t afford one of the paper ones) and perhaps continue on to learn a bit of modern physics and a smattering of more advanced electrodynamics and quantum theory you can manage it.
In the meantime, please — if you post as “fact” exactly the kind of articles that give skeptics a bad name, things that are so obviously erroneous that (logical fallacy that it is) makes it easy for warmists to assert that all skeptics are equally erroneous, you do science no service.
rgb
APS IPCC Workshop 1/8/14
Dr Koonin’s opening remarks
“While not all or even most of the APS membership are experienced in climate, it’s important to realize that physicists do bring a body of knowledge and set of skills that are directly relevant to assessing the physical basis for climate science. Radiation transfer, including the underlying atomic and molecular processes, fluid dynamics, phase transitions, all the underpinnings of climate science are smack in the middle of physics.”
In a similar vein:
In order to earn, emphasis on “earn”, my BSME I had to demonstrate competence in chemistry, physics, heat transfer, thermodynamics, statistics, calculus, algebra, etc. Got the picture? And engineering is more that models, it actually has to work.
The notion that these “climatologists” have some kind of special knowledge or scientific insight the rest of us haven’t got is just snake oil BS. We all recognize a used car salesman when we hear one.
All of this esoteric academic chatter about LWIR up/down is all beside the point. These next three points are all that matter.
According to IPCC AR5 industrialized mankind’s share of the increase in atmospheric CO2 between 1750 and 2011 is somewhere between 10% and 200%, i.e. IPCC hasn’t got a clue. IPCC “adjusted” the assumptions, estimates and wags until they got the desired result. Ta dah! It’s all about man!
At 2 W/m^2 CO2’s contribution to the global heat balance is insignificant compared to the heat handling power of the oceans and clouds. CO2s nothing but a bee fart in a hurricane.
The hiatus/pause/lull (IPPC acknowledges as fact) makes it pretty clear that IPCC’s GCM’s are useless trash.
rgb you are hand picking data to make the case that somehow global temperatures are agreeing with the models which is so not true. In addition the correlation between CO2 and temperature does not exist.
http://wattsupwiththat.com/2012/04/11/does-co2-correlate-with-temperature-history-a-look-at-multiple-timescales-in-the-context-of-the-shakun-et-al-paper/
rgb your data or what you are trying to present does not reconcile with this data or for that matter most other data.
http://hockeyschtick.blogspot.com/2014/09/new-paper-finds-natural-ocean.html
Here is what correlates to all of the temperature changes to the highest degree.
Also if one plots radiosonde and satellite temperature data versus CO2 /model temperature projections one can easily see how wrong they both are.
In addition if one goes back to the Holocene Optimum some 8000 years ago the overall temperature trend has been in a gradual down trend from then — present punctuated by periods of warmth due to solar variability being superimposed upon the slow gradual moving Milankovitch cycles which were more favorable for warmth 8000 years ago as opposed to now.
.
Further refinement of the temperature trends from the Holocene Optimum to present fit very well if the PDO,AMO and ENSO phases are put into the mix and for even a further refinement can be brought about if volcanic activity during this period is accounted for. The correlation being very high as evidenced by the data I sent over.
Co2 having nothing to do with anything. A trace gas with a trace increase which probably has a negative feedback with water vapor as evidenced by STILL no tropical hot spot. This trace gas is not going to run the climatic system of the earth. In addition all data shows CO2 concentrations are in response to the climate /temperature which is why it always follows the temperature never leads it.
That is not to say there is not a GHG effect, but that GHG effect is a result of the climate not the cause
of it.
One last note is the weakening earth magnetic field is being trivialized in all of this which is a mistake. This field acts to enhance solar effects primary and secondary effects.
It ceases to amaze how a non energy source in trace amounts (co2)can be thought to have more influence upon the climate then the sun which is not only the source of energy for the climate system to begin with, but what drives the climate system.
Yet the scam has legs thus far ,despite every major atmospheric process it has predicted to be verified as false but I think time is running short.
I’m doing no such thing. In the graph above, I’m fitting all of HadCRUT4, an accepted temperature series, using the raw data they publish on their website. If that’s “hand picking data” to you, all I can say is that we have very different definitions of hand picking.
In addition, I have to say I’m a bit flabbergasted that, when I publish a graph above that shows a truly excellent — according to R — relationship between the unpicked data (HadCRUT4) and the standard log model for CO_2 driven warming, you would assert that the correlation I’ve plotted “does not exist”. That puzzles me. You might assert that the correlation is accidental. You might assert that the data I’m fitting is corrupt (several people have, and I don’t argue with them but as I clearly state, I’m taking HadCRUT4 at face value and suggest that the buyer beware, especially with regards to error bars and the possibility of biased overcorrection). But to assert that something I plot doesn’t exist seems to me to be some sort of state of pretty serious denial.
I plot the precise functions I use to fit HadCRUT4. You too can download its data, punch the function I obtained into R, and either fit it yourself or verify that the fit I plot really does match the data as well as I portray, but to claim “it does not” when clearly it does is silly.
You go on to plot various other timeseries for CO_2 and temperature, most of them geological low resolution even higher error data, and assert that this somehow “proves” that the greenhouse effect doesn’t exist or that CO_2 concentration has no effect on temperature. I would take issue with the former, just as I would take issue with the assertion that the good fit I plot above “proves” that it does exist, and the latter is almost certainly not true, although there is a great deal of uncertainty as to the magnitude of the “average” effect it has in a complex system that can behave quite perversely with respect to any single “cause”.
As I go to some pains to point out, the plot above is certainly not good evidence against the assertion that CO_2 causes a logarithmic warming of the planet, on average. The fit permits one to estimate an ECS, and the ECS is very much in line with theoretical expectations. I should emphasize again that the fit also does not imply that we are on the brink of any sort of catastrophe — quite the contrary, it suggests that what the majority of surveyed climate scientists in at least the Georgetown survey apparently believe is correct — that the planet is warming, the anthropogenic greenhouse effect is real, and that future warming caused by it will probably not be catastrophic. I think this view is shared by our host, by Christopher Monckton, by Dick Lindzen, by myself, and by a whole lot of other folks who have gone to the trouble to work through a paper or two describing the details of the greenhouse effect, or read through e.g. Grant Petty’s book.
On to Nick:
Nick, I try not to judge somebody’s scientific knowledge on the basis of their degrees or lack thereof but on what they say and the arguments they make. My comments above were very specifically aimed at some absolute nonsense about downwelling LWIR not really being LWIR but being “microwaves” followed by the assertion that microwaves are somehow different. As I’m sure you know given your own understanding of E&M, this is pure piffle, nonsense, balderdash. The downwelling radiation is indeed (partly) LWIR, its spectrograph can be and has been made at many sites and portrayed on this website and elsewhere, and for that matter, there is no substantive difference between LWIR and “microwaves” except for their wavelengths — both are either absorbed or reflected by anything they fall on depending on that something’s particular properties.
I fail to understand why “academic chatter” about LWIR is beside any point. If downwelling LWIR exists — and I think you truly have to be a bit crazy to assert that it doesn’t, especially when you can feel it and measure it and do a full spectral decomposition of its intensity, all of which suggest reality one would think — then that fraction of it that hits the surface of the Earth or Ocean is part of its overall energy budget. True, it is order of 1% of the total energy budget. OTOH, the greenhouse warming we are talking about is order of 1% of the total absolute temperature. It isn’t unreasonable.
As for your other points. I personally respectfully disagree with the assertion that humans haven’t caused all or most of the CO_2 increase. Note well that I did not arrive at this position casually — you can look back a few years at comments I’ve made in the WUWT archives (if you can find them) and for a rather long time I was unsure myself because it was clear that multiple models all of which predict rising CO_2 “could” have been correct, but over the last few years I’ve learned more about the data and its consistency with the various alternatives, and at this point I’m fairly certain that humans have contributed all or most of the current CO_2 increase — around 100 ppm over the last 100 to 150 years. Ferdinand Englebeen, a frequent WUWT contributor who is if anything a climate skeptic or lukewarmist himself AFAIK, has built a lovely website where he exhaustively tallies the evidence and multiple consistent lines of support and IMO his conclusions are very difficult to challenge.
Regarding 2 W/m^2, recall that this is 2 W/m^2 all the time, very nearly everywhere on the planet. It is interesting that you compare this, unfavorably, to the awesome power of hurricanes. Consider a year of downwelling radiation. A year being around 3 x 10^7 seconds, that’s 60 million joules per square meter — just about enough energy to shoot a kilogram sized mass to infinity from the surface of the planet for every square meter of the planet — times 510 x 10^6 x 10^6 = 5 x 10^14 square meters. That’s 3×10^22 Joules, which is, actually, a lot of joules, around 1% of the total solar energy budget.
An average hurricane releases (depending on how you estimate it) just about half of the 10^15 watts of global power downwelling from those 2 lousy watts. If there were over 700 additional “hurricane days” a year, the earth would break even compared to its energy budget without the additional forcing of those 2 watts. That would roughly triple the number of hurricanes. I don’t think you can count on hurricanes to help much with balancing it.
If you were trying to argue instead that the 2 watts is irrelevantly small compared to e.g. the heat capacity of the ocean-earth system, I’d have to say that you are getting your units mixed up. Power is the rate at which the planet receives energy. Its temperature is related to its heat capacity and energy/enthalpy content, but then you can’t use “the planet” because temperature and heat storage are local, not global averages. As I pointed out above and will point out again, 2 watts is a bit less than 1% of the Earth’s current average solar power budget and adding it to the open system can perfectly reasonably be expected to raise the absolute temperature of the Earth-Ocean system by an amount on the order of 1%, although of course it could be half this, or even a tenth of this, or (less likely) more than this (always possible in a nonlinear system). If we assume the average temperature is around 300 K, 1% is 3 C (same degree size) and half of this is — gasp — the ballpark of the climate sensitivity expected of CO_2.
But this isn’t entirely fair, as the actual computation of a response is a bit more work than that. Again, I’m not asserting that I can solve the climate problem in my head, only pointing out that the assertion that increasing CO_2 will cause warming is more than “plausible”, more than just “reasonable”, it is probable. The issue isn’t whether the world will warm, it is “how much”.
Finally, we are in complete agreement that the IPCC’s GCMs are complete trash, as they don’t even do as well as my trivial one-effective-parameter model above head to head on the same data. True, I’m just fitting the data, but the fit is completely consistent with the theoretical prediction if one really assumes that all things remain equal in the climate but the total downwelling radiation increases by 2 measely watts.
The main thing to remember, however, is that while the IPCC’s computers cannot solve the problem of the climate, neither you nor I can solve it in our heads, either! The IPCC’s predictions of doom and gloom could be true, or could be false. The question is one of probabilities based on observation and evidence. I would say that it is probably true that humans are largely responsible for the last 100 ppm of CO_2 in our atmosphere. I would say that it is almost certainly true that this CO_2 has been enormously beneficial to the planet so far, so much so that if we soberly assessed it we would probably have chosen to raise planetary CO_2 to at least 400 ppm, maybe even (as we learn more) 500 to 600 ppm. I would say that one of the many benefits of the CO_2 so far is its probable contribution (however large or small it might be) to the warming that lifted us out of the Little Ice Age. The real problem we face isn’t that the IPCC is right or wrong to worry — it is that we don’t know enough to predict the results of continuing to increase CO_2 in the atmosphere.
This puts humanity on the horns of a dilemma. There are obvious, enormous, overwhelming benefits to all of mankind right now from generating electricity burning coal. Without coal generated electricity, the world would enter a depression the likes of which it has never seen, a loss of around a century of steady rise in global average standard of living, a return to abject poverty in all of the energy starved economies of the world. That is the immediate risk of catastrophe that we are playing with right now, not eighty years in a murky future. There are other risks — huge political power shifts, risks to personal and political freedom. Yet there is an undeniable risk of a climate catastrophe — we certainly don’t know enough to be certain one won’t occur if we push planetary CO_2 over 850 ppm up towards 1000 ppm.
I just don’t like both sides exaggerating the certainty of their knowledge that a catastrophe will definitely occur or that it will definitely not occur, that the greenhouse effect is real and catastrophic, real and harmless, or not real at all when nobody can solve the climate problem in their heads or using the world’s best computers. The real political debate is best served by Feynman’s brutal utter honesty about what we do not know, and how little we can reliably predict, not about what might happen in some imagined worst-case scenario that might actually be rather improbable. We needn’t turn this into the modern version of Pascal’s Wager in the Green Church of the Climate Apolcalypse, which is the way it is currently presented.
rgb
Just like to say this is one of the best and most informative threads for a long time. Willis, as usual, is posting really well reasoned and thought-provoking stuff, nicely presented: no surprise there. But also the quality of the responding comments is brilliant. Something from so many disciplines. Much sensible caution and counter-argument. Much right on the money, some perhaps not so relevant, but all really constructive stuff. Wonderful.
Many thanks, folks.
mothcatcher, I second that.
one of the most interesting article series with a very valid set of hypothesis and supported by data sinca a long time. digging this out does not show a lot of other scientists that try to explain these mechanisms.
thanks a lot for this!
Excellent, as usual!
I can’t help feeling that plotting this same temperature recovery data, from pole to pole as opposed to equatorially, that the seasonal changes due length of day and sun angle would some how reflect variation in polar sea ice formation/melting from year to year. What could influence the recovery from year to year? The ENSO and other events don’t always jibe with sea ice levels.
As mentioned in previous articles in this series, 30C seems to be the magic temperature at which things really start to happen. This just appears to confirm that.
If the data were available, it might be interesting to look at the spread (N-S) over which this effect occurs. I suspect that as temperatures N-S increase towards 30C the band of clouds expands, and as temperatures N-S decrease, the band narrows.
So a global increase in temperature would see the normal equator thunderstorm pattern spread out towards the tropics.
Philip,
“As mentioned in previous articles in this series, 30C seems to be the magic temperature at which things really start to happen.”
It never actually gets to 30C, it looks like things start to happen at 27 +/1 C. That happens to be the critical T for forming hurricanes. So the mechanism might involve triggering some sort of strong convection.
“So a global increase in temperature would see the normal equator thunderstorm pattern spread out towards the tropics.”
That would be in agreement with what is in the paleo record. Tropical T’s don’t change much as the planet gets warmer of colder. To a large extent, what changes is how much of the planet experiences tropical T’s.
The addition of water vapor to air (making the air humid) reduces the density of the air, which may at first appear counter-intuitive. This occurs because the molar mass of water (18 g/mol) is less than the molar mass of dry air (around 29 g/mol). For any gas, at a given temperature and pressure, the number of molecules present is constant for a particular volume (see Avogadro’s Law). So when water molecules (water vapor) are added to a given volume of air, the dry air molecules must decrease by the same number, to keep the pressure or temperature from increasing. Hence the mass per unit volume of the gas (its density) decreases.
The density of humid air may be calculated as a mixture of ideal gases. In this case, the partial pressure of water vapor is known as the vapor pressure. Using this method, error in the density calculation is less than 0.2% in the range of −10 °C to 50 °C. The density of humid air is found by:
http://en.wikipedia.org/wiki/Density_of_air
Those popular greenhouse and blanket analogies are both totally bogus because they are woefully incomplete.
As JoNova (maybe Curry) observed in some thread, the popular GHE is exclusively about radiation, LWIR, in, out, trapped, etc. Without water vapor a greenhouse is an oven. Water absorbs heat when it evaporates and releases heat when it condenses. And a little bit can carry mega-KJ of energy without changing the surrounding temperature. Water vapor is the only GHG that matters because it runs the entire show.
The blanket analogy also ignores the power of water vapor. When you blanket your house with more insulation it doesn’t get hotter inside because the furnace thermostat cuts back the firing rate just as oceans and clouds moderate the atmosphere. Chop wood on a cold day while wearing a heavy “blanket” coat. What happens to the trapped heat? You sweat! And cool off! Your own personal water vapor thermostat. Powerful stuff, that H2O.
Good work Willis Eschenbach, please do more.
So, the planet is self regulating and clouds are important for weather. I’m glad that those lectures I had decades ago were correct all the time. Now, if only Environmental Terrorists, erm, I mean global warming enthusiasts could go back to school.
To summerize in my words “The hottest part of the pacific has a negative cloud based feedback that tends to keep temperatures constant.”