Guest Post by Willis Eschenbach
Impelled by my restless curiosity, I’ve returned to the TAO buoy dataset to investigate a claim by Dr. Ramanathan of a “super-greenhouse” effect. The TAO buoys are a number of moored buoys located across the Pacific. The TAO data is available here.
Figure 1. Locations of all of the sites of the TAO buoys, stretching from above Australia on the left, across the Pacific to off of South/Central America on the right.The buoys collect information on some 17 different variables. The graphic is from the data selection page linked to above. Solid blue squares show buoys which record the currently chosen variable (in this case SST). Empty blue squares show buoys which do not measure the current variable.
I am using the sites on the Equator itself because they have the widest variety of data, including rainfall, air temperature, sea surface temperature, pressure, winds, etc.
Now, here’s the statement by Dr. Ramanathan that I wanted to investigate:
The greenhouse effect in regions of convection operates as per classical ideas, that is, as the SST increases, the atmosphere traps the excess longwave energy emitted by the surface and reradiates it locally back to the ocean surface. The important departure from the classical picture is that the net (up minus down) fluxes at the surface and at the top-of-the atmosphere decrease with an increase in SST; that is, the surface and the surface-troposphere column lose the ability to radiate the excess energy to space. The cause of this super greenhouse effect at the surface is the rapid increase in the lower-troposphere humidity with SST; that of the column is due to a combination of increase in humidity in the entire column and increase in the lapse rate within the lower troposphere. The increase in the vertical distribution of humidity far exceeds that which can be attributed to the temperature dependence of saturation vapor pressure; that is, the tropospheric relative humidity is larger in convective regions.
The “convective regions” are the warmer tropical regions where convective thunderstorms are a frequent occurrence. And his claim is kind of logical, since evaporation is in part a function of temperature, with increasing temperature leading to increasing evaporation.
However, my own experience of living in the tropics led me to suspect that contrary to Ramanathan’s claim, the relative humidity (RH) would in fact be lower in the convective areas, and lower during the times of day when there are the most thunderstorms. I thought this for two reasons.
The first is my own experience of a couple of decades of working in these tropical regions. My observations are that before the afternoon thunderstorms come rolling in, the air is often “sticky” with moisture. After the thunderstorms, on the other hand, the air feels dryer. Anecdotal, I know, but I tend to trust my own experience over theory …
The other reason is that although there is a lot of moisture moving around during the thunderstorm regime, it’s mostly concentrated under and inside the thunderstorms, and that moist air is moving rapidly upwards to have the water wrung out of it by the thunderstorm. But in the much larger area in between the thunderstorms, you have dry descending air. This is air from which the water has been stripped by the thunderstorm through a combination of condensation and freezing.
And as a result, my expectation was opposite to that of Ramanatan—I expected that the more convection, the lower the relative humidity.
So, off to the data, with a few digressions along the way around and back. First, let’s look at sea surface temperatures. This is all two-minute data, that is to say the sea surface temperature (actually one metre below the surface) is recorded every two minutes.
Figure 1. The daily average variations in sea surface temperature at eight equatorial Pacific TAO buoys.
Now, I’ve colored the data from light blue (coldest) to red (warmest). Note that this is also in order by location—the further west you go along the Equator in the Pacific, the warmer are the ocean temperatures. Note that the water temperatures rise evenly and fairly rapidly from early morning to a peak at about three pm. Then over the next sixteen hours or so, the ocean gradually cools down again.
There is kind of a subtle oddity in the daily variations. This is that the warmer the ocean overall, the less daily variation there is in the sea surface temperatures. To illustrate this, Figure 2 shows those same daily ocean temperature cycles as anomalies around their respective averages.
Figure 2. The daily variations in sea surface temperature at eight equatorial Pacific TAO buoys, expressed as anomalies about their respective means. Red shows the warmest buoys, light blue shows the coolest buoys.
Curious. The sea surface temperature in the warmer part of the Pacific don’t vary as much on a daily basis as the temperatures in the cooler part.
As might be imagined, a similar situation holds with the air temperatures. The further west you go, the warmer the air temperatures you’ll find. Figure 3 shows the air temperatures at the same buoys shown in Figures 1 & 2.
Figure 3. The daily variations in air temperature at eight equatorial Pacific TAO buoys.
As with the sea, the temperatures increase with the distance west. However, the changes in the air temperatures are more complex, because of the emergent atmospheric phenomena of cumulus clouds and then thunderstorm clouds. This becomes visible when we look at the air temperature anomalies.
Figure 4. The daily variations in air temperature at eight equatorial Pacific TAO buoys, expressed as anomalies about their respective means. Red shows the warmest buoys, light blue shows the coolest buoys.
Figure 4 is perhaps the strongest evidence of the existence of a cloud-based temperature regulation system that I’ve found so far. Let me see if I can explain why. Here’s a graphic showing the situation at dawn …
Figure 5. The general situation in the tropical convection areas in the early morning.
As you can see, at this time of day clouds are uncommon. As a result, Figure 4 shows that the temperature rises very rapidly for a couple of hours after six AM. However, as the day warms up, at some point a threshold of emergence is passed and the first thermal cumulus clouds start to form, resulting in a change of atmospheric state. Within an hour or so, in place of clear skies, there will be a fully developed cumulus field covering the entire surface.
Figure 6. The general situation in the tropical convection areas in the late morning, with a fully developed cumulus state.
In the colder areas, the cumulus do not form as early or as strongly, so they don’t have as large an effect. But as you can see in Figure 4, in the warmer areas there are so many clouds that the temperature actually drops for three hours, from about nine o’clock to about noon. And as Figure 4 shows, the further west you go, the warmer it gets, and the stronger the cumulus cloud effect gets.
However, even in a fully developed cumulus state, there is not continuous cloud cover. The cumulus clouds can be thought of as flags, each one marking an area where there is an upwelling column of air. However, in between the upwelling air columns and their respective clouds, perforce there must be larger areas of slowly downwelling air. And these areas don’t have clouds. As a result, although the temperature rise is reduced or reversed from nine AM until noon, the sun still gets stronger over that time, and at some point around noon the cumulus shield is not enough to stop further temperature rise.
In the afternoon, with the continuing temperature rise, a new threshold is passed and we get another change of state. This one involves the formation of thunderstorms. These astounding emergent entities pipe air vertically at very high speeds, removing heat from the surface and converting it to mechanical motion. They also cool the surface in a number of other ways.
Figure 7. The general situation in the tropical convection areas in the afternoon to night, with a fully developed cumulonimbus (thunderstorm) state.
There is an oddity, which is that when the thunderstorms develop, the albedo goes down. This is because the vertical motion is so fast in the thunderstorms that they have a proportionately much larger surrounding cloud-free area of dry descending air on all sides of them.
Now, I said at the outset that there would be “a few digressions along the way around and back” to Ramanathan. So with those as the digressions along the way around, let me come back to the topic by saying that these large areas of descending dry air are the reason that I thought that Ramanathan was wrong. Remember that I’d disagreed with Ramanathan’s claim, viz:
The cause of this super greenhouse effect at the surface is the rapid increase in the lower-troposphere humidity with SST; … the tropospheric relative humidity is larger in convective regions.
And what do the TAO buoys say about the relative humidity (RH)? Well, here are the daily cycles in RH for the same eight TAO buoys …
Figure 8. The daily variations in relative humidity (RH) at eight equatorial Pacific TAO buoys, expressed as anomalies about their respective means. Red shows the warmest buoys, light blue shows the coolest buoys.
Now, the colors of the buoys are the same. Coldest is light blue, warmest is red. But instead of the RH increasing with sea surface temperature (SST) to engender a “super greenhouse effect”, the reverse is true. As the ocean temperature rises, the relative humidity falls.
How about during the course of the day? My hypothesis regarding emergent phenomena says that the relative humidity should be lowest during thunderstorm time in the afternoon. The next figure shows the RH all of the buoys once again as anomalies, so we can compare their daily variations.
Figure 9. Daily variations in relative humidity (RH), shown as anomalies about their respective means.
Comparing this to the SST, we see that contrary to what Ramanathan claimed, when the SST is largest, the relative humidity is the lowest.
Now, all these findings shown in Figs. 8 & 9 are curious, because Ramanathan clearly believes that relative humidity is invariant under changes in the climate. He says elsewhere (emphasis mine):
A simple explanation for the water-vapor feedback among the early studies of climate sensitivity was the fact that the relative humidity of the atmosphere is invariant to climate change. As Earth warmed, the saturation vapor pressure (es) would increase exponentially with temperature according to the Clausius–Clapeyron relation, and the elevated (es) would (if relative humidity remains the same) enhance the water-vapor concentration, further amplifying the greenhouse effect. Although it is well known that atmospheric circulation plays a big role, a satisfactory answer as to why the relative humidity in the atmosphere is conserved is still elusive.
But according to Figure 8, the relative humidity in the convective zones of the Pacific varies inversely with sea surface temperature. And this is true both for long-term average sea surface temperature, as well as for the daily average temperature variation.
I can’t say that I have any great conclusions from all of this. However, it does appear that the modelers’ claim of strong water vapor feedback rests on the idea that relative humidity stays constant in the face of warming. If these TAO data findings are correct, and if relative humidity more generally is not constant with respect to temperature, it would seem that this would greatly reduce the amount of purported water vapor feedback …
In any case, it would seem to falsify the idea of a “super greenhouse effect” that is driven by relative humidity as Dr. Ramanathan claimed.
Always more to consider, always more to learn.
My best wishes to all,
w.
PS—If you disagree with someone, please have the courtesy to quote the exact words you disagree with. In that manner, we can all understand exactly what you are disputing.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
I can’t help but love the term ‘super-greenhouse effect”. The hype never stops. Although super-greenhouse is even a little subdued. That must mean mega-greenhouse or greenhousenado were already taken.
Hi logos.
That is the beauty and the whole fun with it.
The hype is actually “killing” AGW.
More hype, more of “supers” and “mega” with the greenhouse effect, bigger the missing heat, bigger the problem with the AGW hypothesis, bigger its problem with reality, more, far more ridiculous the claim that greenhouse effect is or will be a climate changer.
cheers
You have to watch out for those assumptions. That warmer air holds more water is only true for water in a jar. Otherwise, dynamics can change it, as Willis demonstrates. The assumption in GCMs of constant relative humidity is a moronic simplifying assumption of the sort that a modeler might use to get started. It is not a sound basis for building world-changing models. There is lamentably poor data on actual humidity over time.
It’s true for industrial size wet cooling towers and those house top evaporative coolers. Not exactly closed jars.
Craig, Judith Curry and I had a go around on this which drove me to work through the NCAR CAM3 model documentation before finishing my humidity post for her in 2012. The models do not embed an explicit Clausius-Clapeyron UTrh lapse rate. Well, for sure not CAM3, and per AR4 not the others either. But when run, they all behave as though they did. AR4 WG1 Black box 8.1. AR4’s attempt to show this model result accorded with observation is a massive selection bias cherry pick, which was the real subject of my post and that illustration in my ebook The Arts of Truth.. My hypothesis for how the models actually misfire on this is below. If that hypothesis is correct, then the flaw is inherent and unfixable.
These tropical thunderstorms are not only regulating air temperature,they are regulating the concentration of CO2 being pumped out their tops. The regulating processes are evaporation/condensation, freezing/thawing, absorption/extraction, and the vertical changes in their rates. Any detectable “greenhouse effect” from CO2 or water vapor will be lost in the variability of these processes.
Willis
An interesting study.
The real problem with the super greenhouse effect is that there appears to be no effective mechanism whereby LWIR can heat the oceans.
About 60% of all DWLWIR is fully absorbed within just 3 microns. Almost none finds its way past 10 microns.
The only proferred mechanisms mixing the very top of the oceans is wind and waves, and ocean over turning, but these are slow mechanical processes.
Willis, what is the rate (or your best guestimate) of the rate at which wind and waves mixes the very top of the ocean when say local conditions are BF1 or less, BF2 or less, BF 3 or less.
Willis, in the Doldrums how does the action of the wind and waves effectively mix the DWLWIR being absorbed in the top few microns of the ocean?
Willis, you mention ocean overturning, and note that this is a diurnal phenomena. What is the rate (or your best guestimate of the rate) at which ocean overtuning effectively mixes the very top of the ocean (and hence the DWLWIR being absorbed within the first few microns) say between 00:00 to 06:00 hrs, between 06:00 to 12:00 hrs, between 12:00 to 18:00 hrs, and between 18:00 to 24:00hrs?
Can either and/or both of these mechanical processes realistically mix the very top of the ocean (and hence the LWIR absorbed within the top few microns) and sequester the energy absorbed in the top few microns to depth (where volume will disipate that energy) at a rate faster than the energy from LWIR being absorbed in the top few microns would otherwise drive evaporation?
If those slow mechanical mechanisms cannot sequester the energy being absorbed in the top few microns at a rate faster than that energy would otherwise drive evapporation, all that is happening is that DWLWIR goes to drive evaporation and not to heat the ocean
Richard V. How do you think seaweed and plankton and such grow several meters below the surface? What is the ‘heat energy’ in such wavelengths?
UV and visible light get deeper than IR.
Solar.
It would appear that the oceans are only heated by Solar, and this provides all the energy not only to heat the oceans but also for plankton and other life forms that dependent upon its energy..
Fortunately for us the absorption characteristics of Solar in water is very different to that of LWIR (about 50 to 60% of DWLWIR is fully absorbed within the top few microns and about 90% within the top 10 microns).
Instaed of being almost fully absorbed in the top 10 microns, almost no Solar is absorbed in the top 10 microns of the ocean. Instead, it is absorbed at depth predominantly between 50 cm and 5 metres (about 50% is absorbed within the first 1 metre of the ocean and a further 30% within the next 9 metres), but some will make its way to great depths (depending upon purity) of around 100m. This allows Solar to heat gradually the oceans, rather than boiling them off from top down as would be the case if Solar was absorbed in the same manner as LWIR is absorbed.
See for Solar:
See for LWIR:
But when considering the absorption of DWLWIR, bear in mind that DWLWIR is omni-directional, ie., some of it is intercepting the ocean at a grazing angle of 10 degrees, or less, some at 20 degrees, or less, some at 30 degrees or less etc such that very little is intercepting the oceans perpendicularly; hence the reason why a greater proportion of DWLWIR is fully absorbed in 3 microns than would be the case if it was compiled exclussively of LWIR intercepting at the perpendicular (the plot detailed above is vertical penetration).
“It would appear that the oceans are only heated by Solar”
This is because the radiation from the Sun is a transfer of energy to the ocean AS HEAT [Q]. And an input of heat is what … heats.
A postulated – but ultimately imaginary, purely mathematical – flux of energy from a cooler atmosphere to a warmer ocean surface, like the DWLWIR, is not a transfer of energy as heat. The heat flux (the transfer of energy as heat) between the ocean surface and the atmosphere above it is … UP. The atmosphere is the surface’s COLD reservoir. The surface is the atmosphere’s HOT reservoir. Whence the direction of the heat flow.
So please stop it with this nonsense idea that DWLWIR might heat the ocean or might provoke evaporation (which thermodynamically is an equivalent process, requiring a ‘net’ energy transfer). It couldn’t. Only at the times and in the places where the atmosphere is in fact warmer than the ocean surface. Only THEN this whole discussion about how far down the IR can penetrate becomes relevant …
Great analysis. Lends strong observational support to Lindzens adaptive infrared iris hypothesis (BAMS 2001).
AR4 argued strongly for constant Upper Troposphere relative Humidity (UTrH) because that is what the CMIP3 models produced. Their observational evidence was one satellite/ model comparison. Gross selection bias evidenced by dismissing two papers that showed UTrH declined, and ignoring two others reaching that conclusion. Story is in the climate chapter of last book.
Since 2007, a number of papers and studies have come out showing this decline. Specific humidity increases, but not enough for constant UTrH. Documented in essay Humidity ia still Wet. So CMIP3/CMIP5 overstate water vapor feedback; they model a tropical troposphere hotspot that does not observationally exist. Reason is that neither CMIP3 nor CMIP5 can simulate tropical convection cells (Tstorms) due to grid scale computational limitations. See essay Models all the way Down. So have to be parameterized.
CMIP5 parameters were selected to best hindcast 1975-2005 as part of the formal experimental protocol. A period of warming partly from natural variation when model attribution was GHG. So now the pause and the divergence which falsifies the models, and so their high sensitivity, and so CAGW.
“CMIP5 parameters were selected to best hindcast 1975-2005 as part of the formal experimental protocol. A period of warming partly from natural variation when model attribution was GHG. So now the pause and the divergence which falsifies the models, and so their high sensitivity, and so CAGW.”
No. they were not.
Hi Rud.
I am not well informed in this Upper Troposphere relative Humidity or the relative Humidity and its actual play in GCMs.
But the way you have put it in your comment it has sparked some interest and I would like to explore it a bit further especially in the angle you have put it. Especially in regard to the missing predicted tropical troposphere hotspot.
Selecting from your comment above:
“AR4 argued strongly for constant Upper Troposphere relative Humidity (UTrH) because that is what the CMIP3 models produced. Their observational evidence was one satellite/ model comparison.
So CMIP3/CMIP5 overstate water vapor feedback; they model a tropical troposphere hotspot that does not observationally exist.”
——-
Do actually the CMIP3/CMIP5 keep the UTrH as constant (forced to)…. and in the same time are allowed to fluctuate the relative humidity for the Lower Troposphere?
If that the case, then there could be an interesting explanation for the artificial hot spot, especially if the SST-relative Humidity relation is as shown by Willis.
Looking forward for a reply. Thanks.
cheers
See my response upthread Craig Loehle. UTrH is not forced, for sure not in Cam3. It is an emergent property. The error has to be in the parameterization. Tropical convection cells (T storms ) can be adequately resolved in weather forecast models with grid scales on the order of <4km (ECMWF is 1.5 km, initialized from a 30 km regional grid and observed regional boundary initial conditions). See essay Models all the way Down for an illustration using an Arizona thunderstorm system.
The finest GCM resolution in CMIP5 is 120km (1.1degree); typical is 2.5 degree which is about 280 km near the equator. GCMs cannot do finer grid scales because of computational constraints. Halving the gird size quadruples,the number of cells, and nearly quadruples,the number of time steps needed. You run out of MIPS. So tropical convection features arenecessarily parameterized to some presumed average for the cell.
Each model/group has different parameterizations. Cam3 is particularly understandable because the design deliberately uncoupled the dynamical core (physics) from the parameterizations. CAM3 parameterized deep convection (tech manual 4.1), condensate and precipitation (tech manual 4.5), dry adiabat adjustment (4.6), cloud fraction (4.7), cloud nature (4.8), and OLR greenhouse effect(4.9). The water vapor parameterization is 4.9.2 starting page 121. Standard absorption/emission stuff for typical frequency ranges. Usubnw is pressure weighted precipitable water (specific humidity) and Tsubp is the absorber (water vapor) weighted path temperature. Relative humidity is a dependent parameter calculated from these, used only to modify aborption/ emission line broadening and line strength. All this for large atmospheric grid cells coupled to a slab ocean model.
Further clarifications. 1. Each model grid cell and time step,is calculated identically; that is the dynamical core part of NCAR CAM3. But each grid cell lat lon and altitude, has to start with unique initial conditions and responds to changes in those when the dynamic core throws its estimate of a time step change in the cell against the cell’s parameters. Tropical convection at the ITCZ is intense; it is very scarce over the Sahara desert.
2. In the previous reply, I went back to the archived documentation to be precise. It is NCAR/TN-464+STR, June 2004. Free on line.
3. The CMIP5 experimental design is decribed by Taylor, Stouffer, and Meehl, originally 2009 with a corrected version 2011. Available free on line. Table 1 summarizes decadal prediction experiments. 1.2 is 30 year hindcasts using RCP4.5 from 2005 (and 1980 and 1960) with an ensemble size of 3 for each start date. It is evident bt comparison that parameterization was chosen to best match 1975-2005. Initial conditions were supposed to be ‘in some way’ oberved within 4 months prior to 1/1/75. The parameter choices (whatever each group did) amount to the model ‘tuning’ Mosher apparently thinks did not happen. Well, CAM3 was not parameterized by immaculate conception.
Whiten, final short comment. It is the specific humidity in the upper troposphere that is significant for the CO2 greenhouse effect. See essay Sensitivy Uncertainty. Lower down rH does appear pretty much to follow the C/C equation. Main way any humidity can get that high is cumulonimbus. Most frequently in the tropics Willis analyses here. I speculate that the reason UTrH is not preserved is that with warming, there is more evap, more WV near surface, and consequently for whatever reasons disporportionately more humidity lowering precipitation from Tstorms. CMIP3 and 5 models underestimate precipitation, by up to half.
Rud Istvan
February 18, 2015 at 1:50 pm
Hi again.
Thank for the reply.
Sorry for bothering you again.
I actually was not meaning that UTrH is forced on the GCMs when I asked about it.
What I was meaning was if the considered UTrH as a constant was forced instead of being a variable allowed to fluctuate in accordance to the variation of the other parameters related to!?
In another way, you say UTrH (Upper Troposphere relative Humidity), and I say LTrH (Lower Troposphere relative Humidity), does this LTrH make any sense, and if it does is it treated same way by the models with no any strong argument for it to be a constant.
That what I am interested to know, if this has any substance.
I am asking you because clearly I do not have your experience and expertise in this particular matter, and it may take a lot of time to get there for me.
And because is very interesting to me.
For a long time I have been convinced that the GCMs are run under an exaggerated CS so to speak.
But to be fair and honest I been scratching my head for as long too , because even in that case there should not have being so much projected warming if the replication in to the GCMs of atmospheric functioning through all the mechanisms and couplings known and considered was a good and accurate enough replication.
Now thanks to you and Willi too, I can see a possibility that may end that head scratching, so to speak..:)
Please if you can take it up to consideration again and try to clarify this for me, will very much appreciated.
cheers
Sure. Surface UTrH is observed to follow C/C over the oceans. Less over land; simple lack of sufficient water. So the question really becomes the rH lapse rate with altitude and temperature. No model I am aware of forces that rate to be equal to C/C by some explicit equation or constraint. (Caveat: personal detailed knowledge is CAM3). But AR4 clearly says that is the general model result. Discussed multiple places; so important got AR4WG1 special note 8.1.
It is obvious that lapse rates are not locally constant. The lapse rate inside a T storm is much lower, and in the descending exterior dry air column much higher. So a cell average is influenced both by the number and power of these convection cells.
It is now quite certain that models overstate upper troposphere specific humidity. This is knowable directly in three ways. First radiosonde observations corrected for known instrument dry biases. Many instruments, lots of corrections. Second, via microwave satellite sensors. Third, by the newest GPS enabled method. See Arts of Truth for examples. It is knowable indirectly by the modeled tropical troposphere hotspot that does not exist observationally. See Christy’s APS testimony in essay Humidity is still Wet. A direct consequence of overstated water vaper feedback.
Thanks again really appreciated.
Ok, now if I have got this right, hopefully…..I am very aware that I have being driving blindly thus far only by been guided from you.
But for what is worth that what I have got thus far:
In the case of a tropospheric rH variation is required only a slight “overloading” of either UT or the LT in regard to each other, meaning that the rH of one varies slightly more than the other (to a significant point) for a condition of a troposphere heat accumulation to arise, either sourced from a UT or LT heat accumulation, which ever be the case…………while that will be manifesting as a hotspot in modeled tropical troposphere, which actually does not exist,………. or and as an increased warming (artificial) in the projections of the GCMs.
Seems like all needed is the overstated upper troposphere’s specific humidity, which may be enough overstated by the models, for these given condition to arise (the slight rH variation “overloading” of LT in regard to UT).
Gosh hopefully that is something that makes sense, otherwise I will be scratching my head for some time longer….:-)
thanks very much…appreciated.
cheers
Sounds like Ramanathan is enamored by the postulated “super greenhouse effect” for the earth when in several billion yrs the oceans completely evaporate and reside in the atmosphere.
Prb’ly needs to get back to the more relevant present….
Great work as always, Willis! It would be interesting to see what the plots of dewpoint (or water vapor mixing ratio, or water vapor pressure) look like, as these measure humidity in more absolute terms (whereas RH is highly inversely correlated with temperature.)
+1…beat me to it.
Trane has an interactive psychrometric program/graph on their commercial site. Easy to see how moist air parameters/moisture content/energy content, etc. are related.
http://www.trane.com/commercial/north-america/us/en/products-systems/design-and-analysis-tools/calculators-charts.html
Willis, I really enjoy reading your posts. Here, you study RH and draw conclusions based on it.
“…But instead of the RH increasing with sea surface temperature (SST) to engender a “super greenhouse effect”, the reverse is true. As the ocean temperature rises, the relative humidity falls.”
I have never been comfortable with RH because it’s derived from two variables: Air temperature and Dew Point (water content). In your figure 9, the lion’s share of the RH fluctuations are driven by temperature masking the fluctuations in water content (DP) resulting from the competing influences of ocean surface evaporation and downwelling of drier air aloft. Most of the time it’s easier for me to see relationships by viewing dew point rather than RH because the fluctuations of temperature are removed.
Nevertheless, yours was a great post. Thanks for all your hard work.
RH is the ratio between the current grains of moisture per lb dry air and grains per lb of saturated air at constant dry bulb temperature. This change also requires an enthalpy increase from 22.75 Btu/lb to 30.06 Btu/lb. Dew point isn’t involved. Dry bulb and dew point only tell you what the RH is.
Air (CO2?) moves energy at 0.24 Btu/lb-F
Water moves energy at 1.0 Btu/lb-F
Evaporation/condensation of water vapor moves energy at about 1,000 Btu/lb.
Guess which one is the 500 lb gorilla?
It is interesting to parse some of Ramanathan’s statement
Of course this can go on only so long because it represents a constant increase in surface temperature. One of two things occurs. 1)At some point some process limits the temperature increase, or, 2) the posited feedback does not actually occur in the first place.
This is an odd one. More properly one should say that an “assumption” not explanation for the water vapor feedback is the idea of constant RH, but there is no physical basis for such a conservation principle, is there? Usually the justification is the clausius-clapeyron relationship, but this pertains to equilibrium, and where exactly does equilibrium hold over large regions or for long periods of time?
Of course, Ramanathan might reply that while the relationship that Willis illustrates here is true, and is also true in the enhanced greenhouse world, but all of the curves are found at slightly elevated values.
I have heard climate scientists and meteorologists alike state that convection raises the lapse rate and makes the atmosphere less stable, but surely the atmosphere is more or less inherently stable and convection and precipitation are responses to disequilibrium disturbances that return the system toward a more stable state. Also I have trouble understanding why the lapse rate necessarily increases. Within a convective storm the lapse rate must decrease because of the latent heat released (i.e. look at the difference between the dry adiabat and a wet adiabat). The final state of the atmosphere after pronounced convection and the release of latent heat might be to decrease the lapse rate. Anyone have thoughts on this, or data?
“Of course this can go on only so long because it represents a constant increase in surface temperature. One of two things occurs. 1)At some point some process limits the temperature increase, or, 2) the posited feedback does not actually occur in the first place.”
That’s not precisely true. If x is the input and there’s positive feedback f, then the output y is given by y = (x + fy) A -> y = Ax / (1 – Af). This wouldn’t blow up if Af < 1.
You are correct here. I just saw his statement as being ambiguous, as it didn’t mention feedback (i did), but rather the trapping and return of “the” increased emittance.
Region-wide/Ocean basin-wide pressure patterns are essential components of the smaller scale convective areas, and the outcome of the generated water vapor. Will it get transported over a long distance to fall far from its source water area, or will the water precipitate out more locally?
Of course to review, large high pressure areas are subsiding stable air masses with clear skies, and warming surface temps which allow sw sunlight through to the water column. Convection is suppressed due to inherent stability feedbacks. Large low pressure areas support uplift, convective formation, heat dissipation above the tropopause. Tropical cycles tend to move around the periphery of large area high pressures. The predominate area-wide pressure patterns jostle around (hundred of km shifts), and seasonally move north and south with the Hadley cells shifts and jet stream kinks.
Examples: The Ridiculous Resilient Ridge of high pressure forms in the NE north Pacific Ocean, the surface temps and SST warms as SW heating of the water column is favored. When it abates or shifts westward, a cloudier, convective water column cooling patterns moves in its place.
Over in the North Atlantic Late summer, the clockwise Bermuda High keeps the highly unstable tropical disturbances rolling off the west African super-heated Saharan desert generally moving westward with the trade winds. Tropical storm development needs to operate within a larger stable air mass if they are to strengthen, since horizontal shearing away of convection cells that manage to punch through the tropopause are the biggest inhibitors to further growth.
Understanding humidity at the surface is just one variable in a complex system.
With knowledge that temperature is a measure of a kinetic distribution, then the notion of ‘heat trapping’ in the lower troposphere must, by definition, increase the mean kinetic energy beyond that calculable without its incorporation. There must be a significant discrepancy if this effect exists and modifies the lower tropospheric temperature profile.
Taking values of temperature, pressure and height of the tropospheric/tropopause boundary from
http://www.researchgate.net/profile/John_Austin5/publication/228926705_Long-term_evolution_of_the_cold_point_tropical_tropopause_Simulation_results_and_attribution_analysis/links/0f31752e7b2aac62d9000000.pdf
T= 192.5K
P=93mb
Altitude 17km
We can calculate the potential temperature (surface) for a diatomic mix of nitrogen and oxygen from the isentropic flow equation by substitution of T(trop min) at the tropopause boundary pressure.
Isentropic flow equation
T(1)/T(2)= (P(1)/P(2))^(gamma-1/gamma)
Where gamma for a diatomic is 1.4 as used for engineering purposes, ((7/2)/(5/2)).
This gives an isentropic equilibrium temperature for the surface pressure of ~381K.
This differs from measured surface temperatures by around deltaT= 76K higher, for measured T(surface)~305K (32degC)
This is an energy gap of some 76 times the specific heat capacity of dry air difference. Cp for dry air is 1,005J/kg
Therefore the energy gap is 76,380J per kilogram of dry air.
How much liquid water can we vaporise with 76,380J/kg,
Evaporation heat of water is 2501J/g
Therefore 76,380/2501= 30.5g per kg of previously dry air.
32K and specific humidity of 30.5g/kg are not unreasonable figures for tropical humidity conditions calculated from the conditions of the upper troposphere and ignoring back radiative enhancement.
There is NO EVIDENCE of enhanced surface energy due to radiative heat ‘trapping’ within the lower tropical troposphere. Nor anywhere else!
As usual, very interesting, Mr. W.
As luck (?) would have it, I happened to be reading “Water Vapor Feedback and Global Warming,” by Isaac Held and Brian Soden, Annu. Rev. Energy Enfiron. 2000. 25:441-75, which has some bearing on the assumptions you are questioning with this analysis.
Some relevant bits:
“The first results of sensitivity of such a climate model to an increase in CO2 were presented in 1975 by Manabe & Weathereld with an atmosphere-only model over an idealized surface with no heat capacity, no seasonal cycle, and with fixed cloud cover. The equilibrium sensitivity of global mean surface temperature obtained was app. 3 K for a doubling of CO2. The model produced only small changes in relative humidity throughout the troposphere and thereby provided the first support from such a model for the use of the fixed-relative humidity assumption in estimates of the strength of the water vapor feedback.” p. 453.
“In particular, all comprehensive climate models of which we are aware produce increases in water vapor concentration that are comparable to those predicted by fixing the relative humidity.” p. 454.
Their validations, like the comment above, compare annual or monthly means in observations to model output averaged over comparable periods:
“Following Raval & Ramanthan, in Figure 2(a) we use ERBE observations to plot the annual mean clear sky greenhouse effect … over the oceans … . A simple inspection of these figures reveals several important features regarding the processes that control the atmospheric greenhouse effect.”
“The magnitude of greenhouse trapping is largest over the tropics and decreases steadily as one approaches the poles. Moreover, the distribution of the clear-sky greenhouse effect closely resembles the vertically-integrated atmospheric water vapor (Figure 2(b); see color insert)[annual means]. The thermodynamic regulation of this column-integrated water vapor is evident when comparing this distribution with that of surface temperature (Figure 2c; see color insert) [annual means]. Warmer [sea] surface temperatures are associated with higher water vapor concentrations, which in turn are associated with a larger greenhouse effect. Regressing G(clear) versus Ts over the global oceans, one finds a relationship that is strikingly similar to that obtained from radiative computations assuming clear sky, fixed lapse rate, and fixed relative humidity.” p. 449-450.
Willis is looking at hour by hour data, which reveals important phenomena that invalidate the assumptions of fixed relative humidity. These phenomena are obscured by the use of monthly or annual means. This is the final interesting bit I will quote:
“If the value of [water vapor feedback] were larger than unity, the result would be a runaway greenhouse. The outgoing infrared flux would decrease with increasing temperatures. It is, of course, self-evident that the Earth is not in a runaway configuration. But it is sobering to realize that it is only after detailed computations with a realistic model of radiative transfer that we obtain the estimate [water vapor feedback] = app. 0.4 (for fixed relative humidity). There is no simple physical argument of which we are aware from which one could have concluded before hand that [water vapor feedback] was less than unity. The value of [water vapor feedback] does increase as the climate warms if the relative humidity is fixed. On this basis, one might expect runaway conditions to develop eventually if the climate warms sufficiently. Although it is difficult to be quantitative, primarily because of uncertainties in cloud prediction, it is clear that this point is only achieved for temperatures that are far warmer than any relevant for the global warming debate.” p. 449.
So, have they fooled themselves into believing the assumption of fixed relative humidity is correct?
“But it is sobering to realize that it is only after detailed computations with a realistic model of radiative transfer that we obtain the estimate [water vapor feedback] = app. 0.4 (for fixed relative humidity). There is no simple physical argument of which we are aware from which one could have concluded before hand that [water vapor feedback] was less than unity”
Actually there is a simple physical argument to that effect which I have set out previously.
In so far as water vapour is a greenhouse gas it can radiate energy directly to space from within the atmosphere.
In doing so it reduces the total amount of energy (kinetic and potential) held within the atmosphere.
The result is that less energy is returned to the surface in adiabatic descent trhan is removed from the surface in adiabatic ascent.
That reduction in energy returning to the surface is evidenced by the water vapour feedback being less than unity.
However, it does’t cool the surface because at the same time those water vapour molecules are also warming the surface to an equal extent with the net thermal effect at the surface being zero.
The thermal effect of radiative molecules in an atmosphere is always offset by an equal and opposite thermal effect from changes in the balance between adiabatic ascent and adiabatic descent.
In the more complex CMIP5 archive it is about 0.5, cloud feedback is about 0,15 and the total is about 0.65-0.67. All the Bode 1/(1-f) feedback equation. Correct values seem be on order of cloud ~0, WV ~0.25-0.3. Those give an ECS on order of 1.7-1.8. Derivations in Arts of Truth and essays in Blowing Smoke and elsewhere in comments. When two parameters (of 5) are better estimated in Moncktons irreducibly simple model (see my comments on that thread), the equation produces ~1.75. Remarkable agreement.
I concur Willis – today the air is cool and dry in Hervey Bay – the very distant outskirts of an approaching cat 2 cyclone.
http://www.frasercoastchronicle.com.au/news/state-alert-possible-cyclone-looms-bundaberg/2547906/
Willis, thank you for another good essay.
The recurrent claim of unchanging relative humidity is a puzzle.
Willis,
Excellent analysis but there is a flaw. Pardon me if others already touched on this point.
The flaw is that you consider relative humidity rather than absolute humidity.
You wrote:
“But instead of the RH increasing with sea surface temperature (SST) to engender a “super greenhouse effect”, the reverse is true. As the ocean temperature rises, the relative humidity falls.”
If we heat a parcel of air containing water vapor, the “relative” humidity will fall even if total vapor content is unchanged. This is because relative humidity expresses the amount of vapor in the air relative to the maximum amount that could exist at that temperature. Air at 50°F and 100% relative humidity as the same total water vapor content as air at 100°F and 18% relative humidity.
It’s often said that hotter air can hold more humidity but this is not a precisely correct statement because air doesn’t “hold” vapor. It’s just that more water is able to exist in vapor from when the water molecules are hotter. As ocean temperature rises, total water vapor (absolute humidity or vapor pressure) rises slightly, due to increased evaporation, but relative humidity falls dramatically because more vapor can exist when the air is warmed.
Relative humidity is always highest at night, when it’s cool, and lowest in the afternoon, when it’s warm— unless a weather system brings drier air to the area. Other things being equally, total vapor content over a windless ocean would be lower at night (cool) and higher in the day (hot).
It would be interesting to see a rework of your analysis using absolute humidity rather than relative humidity.
Relative uses pounds, absolute uses cubic feet. It’s the pounds that carry the energy.
Before you ask about analyzing absolute humidity instead of relative humidity, let me ask you to verify my calc’s for generating relative humidity from 2 meter air temperature, wet bulb temperature, and local air pressure.
This info is for a site much colder than the tropics (Barrow AK) but I do want to verify my conversion from the raw numbers above is right.
This might be a mess, we’ll see. First, of all I’m steeped in English units. Second, your table doesn’t seem to include wet bulb. Doesn’t matter. This is from Trane’s program which is produced by HandsDown software.
(H – Dry bulb * 0.24 Btu/lb-F)/lb water vapor
These are my results. Don’t see your RH, WB, or W for comparison.
[Set to text format for the table. (html code (pre) and (/pre) – in angled brackets for WordPress .mod]
nickreality65
The column labeled “Dewtemp” is their wet bulb temperature in degrees C.
Wind speed (for each of these sample lines at least) was all 0.0 knots. But wind speed doesn’t change dewpoint and 2 meter air temperature – which was also in degrees C.
Pressure is in millibars in the last column (Yeah – metric again.)
What is 2Mtemp and what are the units?
2MTemp is the dry bulb air temperature at 2 meters above ground, degrees C.
I looked again through emails for the original text file, but have only the resultant Excel file.
The other column, what I improperly said above was “wet bulb temperature” earlier, is DewPoint temperature, also in deg C.
RH equations are:
Spreadsheet-ready equations for each unknown in terms of the two knowns:
RH: =100*(EXP((17.625*TD)/(243.04+TD))/EXP((17.625*T)/(243.04+T)))
TD: =243.04*(LN(RH/100)+((17.625*T)/(243.04+T)))/(17.625-LN(RH/100)-((17.625*T)/(243.04+T)))
T: =243.04*(((17.625*TD)/(243.04+TD))-LN(RH/100))/(17.625+LN(RH/100)-((17.625*TD)/(243.04+TD)))
(• replace “T”, “TD”, and “RH” with your actual cell references)
(• T and TD inputs/outputs to the equations are in Celcius)
Ref: http://andrew.rsmas.miami.edu/bmcnoldy/Humidity.html
“but I tend to trust my own experience over theory …”
Well, there’s your problem …
And I see your fish is back on your diagrams. Nice to see it again, even if it is not essential for realising the Tao of Humidity.
Matches my observations in the subtropics and tropics. Gets humid and hot, then the vertical flow reaches the precipitation point and the rains come, wringing water out of the air as precipitation. The dry and cooled air then descends and everyone enjoys the cooler less humid result. Seen in even more spectacular form when a hurricane happens. Leaves much cooler water behind it and with millions of liters of water precipitated back to the surface.
Once water is hot enough to get the evaporate, rise, precipitate cycle going, added heat does not result in more temperature increase, but shows up as more mass flow of water vapor; thus the flatter temperature curves. As a heat pipe engineer and I’m sure they can give the engineering details of it.
Nicely done W.
It’s as simple as this: Without convection you will get a hot and humid (‘sticky’) layer of air close to the surface. The heat transferred from the surface to the lowermost layer of the atmosphere (by conduction, radiation and evaporation) simply cannot effectively escape to higher regions, and so it piles up (after all, the heat is still coming in from the Sun). Convection on, however, and both the accumulated energy and the accumulated water is swiftly brought up and away. Cooling and drying.
Excellent post, thank you Willis. I enjoy the clear logic and curiosity of your work.
While what you have done here raises hundreds of good questions – most of which I am ill equipped to help with – the question that really interests me is why are these people so bad at science?
I think there are 2 reasons:
1. They are not prepared in the sciences to do science and so examination of their work always seems to deconstruct their work.
2. They lack scientific curiosity, which you demonstrate here.
When they happen upon a graph or a number that meets their need, they think they’re done. Not interested in looking further. They are not curious which is fundamental characteristic of a real scientist who now has bunches of new questions to consider.
Thanks Willis,
You seem to have answered a question I had.
I have been looking at the meteorological data at an Australian Bureau of Meteorolgy site called Giles in central Australia. The dew point is highest in the early morning and deceases during the day as the temperature rises. Dew point is a measure of the water content of the air.
So I wondered; where does all the moisture go? Well the answer seem to be that it goes up (with the thermals, which go over 10.000ft) and is replaced by drier air coming down from above.
Willis,
I think you’re close to showing compelling evidence of your posited equatorial cloud-based thermostat but you’re missing some fundamental physics.
“If these TAO data findings are correct, and if relative humidity more generally is not constant with respect to temperature, it would seem that this would greatly reduce the amount of purported water vapor feedback …”
Specifically, “if relative humidity … is not constant with respect to temperature.”
If water vapor is constant, “relative” humidity will NOT be constant with respect to temperature because it is a relative value—relative to the amount of vapor that could exist at the now elevated temperature. If vapor content is constant and temperature is increased, relative humidity will decrease.
Also, the intertropical convergence winds carries moist air from the subtropical seas to the band of equatorial convection clouds. It’s not gaps in the equatorial clouds that represent the down-welling air. It’s air over the oceans north and south of the equator that converge at the equator to supply the moisture for the equatorial clouds. These areas are oceanic deserts where very little rain falls but much moisture evaporates.
Willis,
Correction: It is unquestionably true that equatorial convective storms have a damping effect on the earth’s temperature. They reflect sunlight and shade the surface below so that the surface is cooler than it would otherwise be. Still, your analysis errs in using relative humidity because relative humidity is not a measure of total water vapor content and it is total vapor content that controls the local greenhouse effect.
RH is a measure of total water vapor. It’s the ratio between pounds of actual water vapor and pounds of water vapor at saturation.
Thomas February 18, 2015 at 10:36 pm
Actually, Thomas, your analysis errs in not noticing that I was REPLYING TO RAMANATHAN’S CLAIMS about relative humidity. Absolute humidity is a valid question, but it’s not what Ramanathan was discussing and as a result, it’s not what I discussed.
w.
Willis, you are right about the relative humidity thing. Your analysis actually fails because you were looking at surface level humidity and Ramanthan was looking at total column humidity.
You quoted from the abstract of Ramanthan’s paper “Physics of Greenhouse Effect and Convection in Warm Oceans” (1994) wherein he wrote that, “tropospheric relative humidity is larger in convective regions.” Ramanthan also wrote that the increased greenhouse effect was caused by an, “increase in humidity in the entire column.”
These statement are clearly correct. Convection causes an increase in total column relative humidity because water vapor is lifted far above the sea surface. Without that lifting there would be no cloud formation so it’s obvious that it is actually happening.
Your charts of TAO data show changes in relative humidity near the surface. This does not address Ramanthan’s statements about total column, tropospheric humidity (whether relative or absolute) because you’re looking only the humidity only near the sea surface.
However, it’s interesting to note that Ramanthan also says that his “radiation model calculations do not include the effects of clouds.” He used a model to study the greenhouse effect of increased water vapor but ignored the fact that the water vapor will cause cloud formation, which will cause surface cooling because clouds reflect sunlight.
Ramanthan’s whole exercises is rather obvious—(a)water vapor is a strong greenhouse gas, (b) convection of moist tropical air increases total column vapor, so (c) the greenhouse effect is larger. I don’t think we benefit much from a detailed model-study of this effect. Any first-year metrological student would know the answer.
Furthermore, calling it a “super greenhouse effect” is a gross exaggeration. There is nothing “super” about the effect, unless you ignore clouds.
Ramanthan’s model experiments certainly do not provide evidence to support the hypothesis that CO2-induced warming will produce a significant positive feedback. As you have so elegantly explained, the warmer the ocean gets, the more clouds form, so the ocean gets cooler. Ramanthan ignored the clouds so his paper can’t be a refutation of your theory.
As for the effect of tropical storms on surface relative absolute humidity, it’s somewhat counter intuitive but the near-surface humidity is often lower when it rains. When rain falls it drags air with it and this air has a low water vapor content because most of the water is now in the rain droplets. These downdrafts warm due to compression and through mixing with warmer near-surface air. This means water vapor content near the surface when rain is falling is often lower.
I looked at weatherunderground.com and found a random day in Miami that had a thunderstorm (August 20, 2014).
Here’s what I found:
Just before rain began (13:30)
32.8°C with 57% RH = 0.018 kg of water vapor per kg of air.
When rain was falling (14:00)
21.1°C with 77% RH = 0.016 kg of water vapor per kg of air.
The absolute humidity at the surface decreased by 20% when rain was falling.
“The column labeled “Dewtemp” is their wet bulb temperature in degrees C.”
Dew point and wet bulb are NOT interchangeable. Dew point says how much vapor is in the air, wet bulb says how much more vapor the air can absorb. Wet bulb is a measure of the moist air’s heat content. With dew point heat content decreases as it condenses out reducing the sensible heat, i.e. dry bulb.
Willis, I always enjoy reading your musings and follow-up analysis. Especially since you follow the data wherever it leads. I have commented before on the water-cycle “thermostat” concept that you are exploring further in depth in this blog, and might as well throw the same thoughts into the comments on this one as well. Some of my past observations to add to yours:
1) IR photos from space looking down on the oceans during very large storms (hurricanes) at night always show areas of light and dark “radiation”. IMHO this has to point to where IR radiation is given off from the storm. You do in fact see striations in these images, areas where active “radiation” is occurring and areas where no (or less) radiation is occurring (see for example: http://www.space.com/18236-frankenstorm-hurricane-sandy-satellite-photos.html)
2) Also IMHO, the evaporation of water (liquid to gas), sublimation (solid to gas) as well as the transport of water vapor upwards all have to be processes that are energy losing processes (that is it takes energy from the system to make each of these processes occur). The energy that is “lost” (ie that is transferred to the water molecules) from the “system” is radiative. During the day (while radiation is actually coming into the system) some of the radiative energy is absorbed by water vapor and some is also reflected back to space. The radiation that enters the top of the atmosphere and actually makes it to the ocean is absorbed, and “lost” to excited (liberated) water vapor molecules. Therefore, when condensation (or melting) occurs in a cloud the converse must also be true as well – radiative energy must be released. During the night, no net radiation is coming into the top of the atmosphere. Therefore, these IR pictures of a huge storm at night must show us WHERE radiative energy loss to outer space from the cloud is actually occuring. The picture tells the story of where the cloud is “glowing” and giving off extra energy in the infrared. (You’ll also notice in these pictures that although there is a similar amount of CO2 over the uncloudy water regions, you don’t see anywhere near as much IR radiation as you do coming from a cloud – in fact you see black, none, nada, zip.)
3) Also IMHO, when you look at a nice thunderstorm forming on the horizon, what you see visually shows evidence of where you should also see the most IR radiation. You nearly always see a distinct band where the bottoms of the clouds begin. You also see a distinct band where the top of thunder heads form the “anvil”. Both planes of demarcation give evidence that this is the elevation where the water cycle is dramatically changing phase (crashing?). IMHO it should be at those phase change locations where the most release of stored energy back to IR occurs.
4) In his excellent presentation “Water, Energy, and Life: Fresh Views From the Water’s Edge”
( https://www.youtube.com/watch?v=XVBEwn6iWOo ), Dr. Gerald Pollack introduced us to the concept of water droplets in a cloud actually self forming a liquid crystal phase all around the outer surface of each tiny condensing water droplet. In his presentation, he provided evidence that the energy engine that created this phenomenon was infrared energy. He also noted the appearance of a charge separation within each condensing droplet that does at least two amazing things. First, it enables clouds to self organize and not “disperse” or diffuse away. Second it stores even more energy (ie sucks IR radiative energy out of the surrounding milieu and into condensed cloud droplets). So this suggests that “inside” clouds water droplets that are forming liquid crystal films around there outer surface, must be IR sinks. The most dramatic event in a cloud that releases this accumulated energy is a lightning strike (radiative across the visible spectrum – extremely energetic radiative release).
5) Finally, its a shame that the buoys you got the data from for this analysis don’t also provide:
a) a straight-up-looking visible and IR spectrum of the intensity of light hitting the buoy to get an instantaneous picture of what radiative transfer is occurring back down from under clouds vs back down from when rain is occurring,
b) a simultaneously measure of CO2, methane etc. to give us a picture of what the other (inconsequential) green house gas concentrations are doing. I suspect that every day after a good rainstorm the CO2 dramatically drops (is scrubbed from) the atmosphere. I also suspect that the change is much greater that the slow average annual drift that has warmists all in a tizzy.
So, have you ever had any thoughts about adding “liquid crystal” thermodynamics to your overall thermostat concept? I suspect it is NOT a minor effect but rather may be nearly as large as latent heat transfer, and specific heat/convective transport. Secondly, in your analysis of the data, can you distinguish a difference between when a buoy goes through a day where it is in the path of a storm – experiences the before and after effects of a good rainstorm – versus when it goes through a day and no rain falls on it? What do those data “pictures” show regarding thermostasis?
Respectfully, DonV
“The energy that is “lost” (ie that is transferred to the water molecules) from the “system” is radiative.”
I disagree. As water evaporates it influences the air/liquid sensible temperatures through conduction not radiation. Consider what is happening in one of those enormous utility wet cooling towers.
Wow, did I halt another thread?
Willis,
I do not understand the x-axis on your figure 4. Perhaps that should be in whole integers, not tenths.
Willis, never mind. I see the diurnal temperature variation is less than 1° C, most remarkable.