Kevin Kilty
There is great disagreement and discussion about how heat once planted on the Earth’s surface redistributes itself and makes its way back to space. Not that it makes much difference to settling our debates, but I see value in what observations tell us.
SURFRAD stations are highly instrumented platforms capable of revealing anything one would wish to know about insolation, surface albedo, upward long wave infrared (LWIR) flux, and yes, even the thing that a few people say has never been demonstrated, downward LWIR flux commonly known as the greenhouse effect.
There are only eight such sites in the continental U.S. and I have chosen two of them, representing the end members of summertime climate in the lower 48, for detailed review. In conducting this review I add balloon soundings from nearby stations and calculations from the U of Chicago version of MODTRAN.
The suite of instruments at each SURFRAD installation include paired upward/downward oriented pyrometers and pyrgeometers. These instruments integrate radiant intensity data over a hemisphere of view and over the wavelength range of 4 to 40 micrometers (um). There is also a suite of meteorological instruments. People can reach this data through the Global Monitoring Laboratory website.
Desert Rock, Nevada
Figure 1 shows the Desert Rock site. It is a desert region of sparse vegetation and loose soils common to over half the Western U.S. Figure 2 shows solar and LWIR data over a solar day organized from SURFRAD data. The principal features of interest are: a peak in downward solar flux of just over 1,000W/m2 and upward solar flux that is a constant 20% of downward (surface albedo). The integrated net solar flux for the day is 24 million J/m2.
Figure 1. The default radiation plots from SURFRAD stations are synchronized to Coordinated Universal Time (UTC), so these plots are spliced to provide a clearer view of daily evolution.
Of greater interest is the LWIR record that Figure 2 illustrates. I will first note that the green curve is that which some people claim has never been measured – the greenhouse downward LWIR flux nearly all of which originates in the atmosphere.
Figure 2. Red – downward solar flux; Blue upward solar flux; Amber upward LWIR; Green downward LWIR.
The amber colored curve is upward LWIR flux. The earliest hour just prior to dawn exhibits an average value of 465 W/m2. Since the 10m of intervening atmosphere between the surface and the pyrgeometer contributes less than 1W/m2 from my MODTRAN calculations, 464 of that must represent the black body contribution from the surface. This suggests a surface temperature of 30C (86F). The meteorological station records an average of 31C 2m air temperature during this interval.
At its peak of the day, however, upward LWIR reaches an average value of 670W/m2. The implied ground temperature is now 59C, far above the 2m air temperature of 39C. Intense solar radiation produces a superadiabatic layer near the ground surface which is the source of the ubiquitous dust-devils and thermals similar to those which I described at Albuquerque in this essay and its addendum.
The day ended in all respects near in temperatures and fluxes where it began. Thus, we can conclude that energy storage during this day was close to zero. The cooling due to LWIR is positive upward at a net of about 13.6 million J/m2. In other words, on this day the near surface Earth was cooled 57% by LWIR and 43% by other means – thermals or horizontal transport but no significant latent heat probably because the day was cloudless and RH remained low.
One final interesting observation involves the downward LWIR flux before dawn. The hourly averaged value is 383W/m2. I thought it would be interesting to compare this value to what MODTRAN would calculate from its midlatitude summer model, with an appropriate adjustment of surface temperature and surface humidity.
Figure 3. A comparison of balloon soundings and a MODTRAN model atmosphere.
This turned out to be 325W/m2; the 58W/m2 difference between the two seemed a large departure. I wondered about it.
To investigate this I obtained balloon soundings from Las Vegas, Nevada which is the closest I could find to Desert Rock. Figure 3 shows a comparison of the MODTRAN midlatitude summertime model atmosphere temperature structure offset by 30C, to the morning and midday soundings. A notable observation is that a 1,000m thick layer of the atmosphere near the ground surface cools overnight through radiation. This is so common throughout dry and elevated portions of the Earth that it is only notable about one time.
Figure 3 shows the MODTRAN model atmosphere has a more or less constant temperature gradient as the orange curve indicates. The morning sounding (6 am Las Vegas time on the 26th) shows a structure that if the orange curve was offset to the surface temperature of the sounding would leave a thick segment of air aloft with a temperature above the model atmosphere. I did an approximate calculation using MODTRAN and found about 33W/m2 of adjustment. There is little doubt in my mind that the entire 58W/m2 is explainable by atmosphere temperature structure.
Goodwin Creek, Mississippi
This SURFRAD station is located near the Waterways Experiment station at Vicksburg. Figure 4 shows the location which is lush with surface grasses and nearby woods. The tower with the pyrometers is apparent as is the meteorological station.
Figure 4.
Figure 5 shows my reconstructed solar day data. I find it a remarkable example of how clouds modify insolation. The downward solar flux is greatly depressed while clouds, especially thunderheads, pass over. However, clouds do more than just shade the Earth. The peak insolation is nearly 1250W/m2 showing that clouds can redirect insolation to places other than the places they shade. The net full day insolation here is 8.6 million J/m2.
The LWIR flux curves are remarkable also. Note how close they are to one another. They even intersect at times. This is a consequence of the extremely wet atmosphere near the surface which ranges in relative humidity from 80% to 100% from the surface to 700mb at 3.2km above the surface. Net LWIR, while being directed upward, does not supply nearly the fraction of cooling that it does at Desert Rock. Net LWIR is only 2.1 million J/m2. Thus, LWIR provides only about 24% of the cooling needed to offset insolation. Assuming that at the end of day we return near enough to conditions at the start of the day, bulk heat transport must provide 76% of the cooling.[1]
The upward directed LWIR in the darkness before dawn (449W/m2) is exactly what the 26C air temperature at the time would predict. At the peak of upward LWIR in the 1800 UTC hour, the implied surface temperature is just 6C warmer than air, but this is enough to run thermals as the observations at Albuquerque and its addendum showed.
Figure 5. Radiation plot from Goodwin Creek, Mississippi for the day beginning at about 6 am local time on July 25 and ending at roughly 6 am on the 26th. The default radiation plots from SURFRAD stations are synchronized to Coordinated Universal Time (UTC), so these plots are spliced to provide a clearer view of daily evolution.
Figure 6 is a Skew T plot of the balloon sounding from Jackson Mississippi at 7 am local time on the 25th. The extremely humid atmosphere from surface to 700mb is apparent. The atmosphere at this time consists of layers of humid air interspaced with dry air. A similar structure is observed from balloon soundings across the plains and Mississippi valley this same day.
What a person can see with Skew T plots at west coast stations from a few days prior is that dry air throughout the troposphere is a boundary condition to the continental U.S. established by flow over the Pacific. Modifications occur each successive day as air traverses the country. Exactly how the atmosphere obtains the structure of Figure 6 is too complicated to unravel with the limited data at hand. Suffice it to say that radiation and bulk heat transfer work hand in hand to produce it.
Figure 6. Skew T diagram of the balloon sounding from Jackson, Mississippi at 7 am local time on the 25th of July, 2026. Note the extremely moist atmosphere below 700mb.
Appendix A
People like to quote Richard Feynman. His most famous quotation may be his definition of science being the primacy of experiment over theory. Yet, I keep in mind another quotation I find more valuable. Feynman stated it in reference to the classical theory of paramagnetism being just plainly wrong, but it applies to all of science. Just eliminate the word “classical” in this.
“If you start a [classical] argument in a certain place and don’t go far enough, you can get any answer you want.”[2]
I am also tempted to invoke Fredric Bastiat’s broken window fallacy to prove this idea is universal, but maybe that’s whipping a dead horse. We all know about politicians getting any answer they want.
There is a common refrain in discussions here that bulk heat transport dominates at the Earth’s surface, takes the job away from radiation so to speak, and only hands off the job back to radiation very high in the atmosphere. I think it’s important to dispel this notion because from data and other considerations, it’s difficult to view the job of transporting heat from surface to space as anything other than a shared responsibility of radiation and bulk transport.
The commenter known as “Frank from NOVA” urged me to look at an online document, which I located and downloaded from Andy May’s website, by Tom Shula and Markus Ott. Its explanations of the interaction between radiation and matter are 90% correct, but just as Feynman warns, they aren’t taken far enough for a valid conclusion.
Let’s begin with their Figure 2.
All that this diagram shows us is that a CO2 molecule may absorb a 15um photon then release a photon (spontaneous or stimulated) to return to its ground state. According to the authors this is a slow process. So far, so good.
Proceed now to their Figure 5.
Their Figure 5 shows the process of de-excitation of a CO2 molecule without emission of radiation. The molecule of CO2 in an excited state of vibration has collided with another molecule and converted the energy of its excited state into altered kinetic energy of that other molecule. So far, so good, also.
Now they embark on a discussion about how often the process of their Figure 5 occurs with respect to the process of their Figure 2. They find the vast majority of CO2 molecules in an excited state will proceed via Figure 5 back to a ground state. Also, so far, so good.
Yet here is what they state as the main take-away of their discussion to this point.
“That means that in air under ambient conditions the thermalization rate is by a factor of about 50,000 higher than the emission rate of the excited state. As a result of this rough calculation, we can say that at atmospheric pressure and ambient temperature the thermalization of the excited state of the CO2 bending vibration is by four to five orders of magnitude faster than the spontaneous emission from this excited state. Even if an excited state somehow escapes the thermalization and emits a 15-micrometer photon, this photon will be absorbed after traveling a few meters through the atmosphere by another CO2 molecule. It is highly probable (~50,000:1) that the excited state resulting from this lucky escape will then be thermalized. Under these conditions a radiative transport of energy through the lower layers of the atmosphere by a cascade of emission and absorption events is not possible.”
First, keep the words “not possible” in mind. I’ll return to this. Then let me simply state that the rapidity measured by (50,000:1) is really a red herring in this context. All that rate tells us is that exchanges of energy within the volume of gas occur in tremendous numbers per unit time. But what seems underappreciated is that through such exchanges there is a process of establishing thermal equilibrium. Molecules within a sea of photons and subject to a flux of photons from a source like the Earth’s surface, or out of a neighboring volume of gas, is going on at a very healthy rate. This rapidity is what produces and maintains a state of Local Thermodynamic Equilibrium (LTE). It’s what returns a state of momentary disequilibrium back to equilibrium.
Now, refer to their Figure 8.
Figure 8 shows us how the alleged impossibility of the release of a 15um photon down in the lower atmosphere actually becomes possible throughout the atmosphere. The de-excitation process is actually two-way. Through collisions the CO2 molecule may be raised to an excited state and release a photon.
This is where Feynman’s dictum about not taking an argument to a conclusion becomes essential. The processes of their Figures 2, 5, and 8, all take place within each volume of gas down at the Earth’s surface, and each one high in the atmosphere, and each one every place in between. It is an example of the principle of detailed balance in operation.
The authors even calculate what the results of detailed balance and thermal equilibrium imply – a Boltzmann distribution. They correctly figure a stationary ratio of molecules in the excited state to those in the ground ground state at any time as a function of local temperature – 1.3% at -53C, 3% at 15C, 4% at 30C.[3]
What this implies is that at Earth’s surface, where a mole of atmosphere is contained in around 25 liters of air, there are 6×1023 molecules of air, 2.4×1020 molecules of CO2, and around 9×1018 of those CO2 molecules in an excited state that will emit a photon in the next half-second. There is a vigorous diffusion of photons, and other energy conversions, all serving to keep the system as a whole near LTE. Moreover, since the process of their Figure 8 occurs at a boundary of the gas volume, 15um radiation will be detected by a radiometer of suitable design. A pyrgeometer is just such a beast of suitable design.
Citations:
1-Balloon soundings taken over the full day at Jackson show that the atmosphere’s structure has changed. No doubt some of this change is due to the heat conveyed away from the Earth’s surface but I haven’t enough information to analyze how.
2-Feynman’s Lectures on Physics, Vol. II, 34-6, First Paragraph.
3-We might include the Lectures on Physics, Vol. II, 35-8, midpage starting with “Now statistical mechanics tells us …”
