experiments performed July 10th 2024
by Dale Cloudman
Abstract
It is experimentally shown that the thermal radiation from a transparent, colder solid has the capacity to influence a solid warmer than it to become even warmer, under the right circumstances. This dispels the critique of the greenhouse effect that, as heat only flows from hot to cold, the effect is thermodynamically impossible. Even so, significant portions of the theory of the greenhouse effect remain experimentally unproven, signaling caution rather than uncritical acceptance of the theory.
Introduction
A long-running debate about the physical reality of the greenhouse effect centers around whether the thermal radiation from the colder atmosphere can possibly have a warming effect on the warmer Earth’s surface. The argument goes as follows: as heat only flows from hot to cold, the radiation emitted by a colder object cannot possibly cause a warmer object to become warmer. It might have a reduced cooling effect, but under no circumstances can it result in a warming effect.
The most salient features of such debates are that neither side provides experimental evidence in their defense, and that the debates frequently devolve into heated arguments, which is generally an indicator that solid arguments are lacking. As genuine scientific knowledge is acquired via experimentation performed in physical reality, and not theory or “experiments” (sic!) consisting of computer simulations, I sought to settle the debate once and for all with a properly-performed experiment.
I constructed a de Saussure-style hotbox, above which I placed a clear glass plate separated by a good amount of space. I pointed it towards the Sun, and, left to its own devices, the bottom of the box got to around 100ºC, with the glass plate getting to around 30ºC. I then swapped the glass plate with an identical one that was pre-heated to 60-70ºC, and the result was unmistakable: the already much hotter bottom of the box got even hotter as a result. I ruled out any possible conductive or convective effects, concluding it was due to increased thermal radiation coming from the +30-40ºC warmer glass plate. Although this neatly dispels this particular critique of the greenhouse effect, significant obstacles remain before being able to accept that the greenhouse effect behaves as-described.
The Demonstration
![](https://i0.wp.com/wattsupwiththat.com/wp-content/uploads/2024/07/image-95.png?resize=720%2C405&ssl=1)
The apparatus (Figure 1, Figure 2) consists of a hotbox made of styrofoam, with the inside floor being cardboard and the inside walls and floors spray-painted black. Two layers of thin plastic polyethylene film suppress convection with the outside air. Surrounding these is another piece of styrofoam to further prevent heat loss through the sides. Cotton padding between the styrofoam pieces provides further insulation.
On top of the hotbox are two styrofoam walls that support an extra-clear glass plate on top. This is the glass that is swapped with an identical, hotter version during the experiment. A middle mount point allows the optional mounting of another plate in the middle.
Temperatures are measured with Type K thermocouples, labeled and color-coded for ease of reference. Styrofoam pieces immediately above yet not touching them serve as radiation shields. The solar insolation is measured with an Apogee SP-510-SS pyranometer, and the net infrared gain or loss is measured with an Apogee SL-510-SS pyrgeometer.
I pointed the apparatus at the sun and let it heat up on its own. In the meantime, I set aside an identical extra-clear glass plate, on top of which I placed an aluminum plate with a pot of water on top. An immersion heater kept the water temperature at a constant 80-90ºC, which heat diffused downwards to heat the glass plate.
Once the bottom plate exceeded 100ºC, I swapped the cool glass with the hot glass, first by hovering the hot glass above it, then removing the cool glass and placing the hot glass in the same position. This swapping technique provides a dip in measured solar insolation due to both glasses absorbing the sunlight rather than just one. This clearly delineates the swaps, and ensures there is no extra heating effect due to slightly higher insolation that would happen if we first removed one glass and then replaced it with the other glass.
I performed five swaps as above, with Tcoolglass measuring 30-38ºC and Thotglass measuring between 67-75ºC at the start of the swaps. The result was clear and unambiguous in every case: Tbottom shot up rapidly in response. The following two runs are representative (Figure 3):
![](https://i0.wp.com/wattsupwiththat.com/wp-content/uploads/2024/07/image-97.png?resize=720%2C430&ssl=1)
As is evident, before the first swap on the graph, Tbottom was increasing at a certain steady pace. After the swap to the hot glass (delineated by the purple vertical line), the rate of increase rapidly shot up. Once swapped back to the cool glass (delineated by the gray vertical line), Tbottom stabilized around 108ºC. It remained there for ~5 minutes, after which another swap caused it to increase rapidly again.
It is notable that at the exact moment of the swaps, the net infrared radiation Net IR picked up by the pyrgeometer remains negative yet increases. That is, the bottom’s radiative cooling is measurably lessened as a result of this swap.
To rule out differences between the two glass plates, such as the possibility of slightly more solar insolation which could cause a temperature change, I did two swaps with both glass plates at around the same temperature. With these swaps, no difference was observed in the evolution of Tbottom or in the Net IR.
Confirming the Radiative Nature of the Effect
It is important to rule out any convective effects as having caused the increase. The hotter glass will, of course, heat the air around it as well. How do we know that Tbottom didn’t get hotter because the hotter glass heated the air in-between, which then heated the thin plastic films, which then reduced the convective loss of the bottom and caused its temperature to increase?
The evidence is three-fold. First, we can observe that Tlowair did not noticeably increase as a result of the swap. Indeed, it was warmer before the swap than after (Figure 4):
![](https://i0.wp.com/wattsupwiththat.com/wp-content/uploads/2024/07/image-98.png?resize=720%2C432&ssl=1)
Second, by including Tthin1 and Tthin2 in the graph, we can see that Tbottom increased first, and only nearly half a minute later did Tthin1 start to increase, soon after which Tthin2 increased as well (Figure 5). Thus the temperature increase started from bottom to top, not from top to bottom, meaning the observed increase could not be due to the air being heated from the top.
![](https://i0.wp.com/wattsupwiththat.com/wp-content/uploads/2024/07/image-99.png?resize=720%2C428&ssl=1)
![](https://i0.wp.com/wattsupwiththat.com/wp-content/uploads/2024/07/image-100.png?resize=598%2C720&ssl=1)
Third, I repeated the experiments, but with a thin borosilicate plate mounted in the middle of the apparatus (Figure 6). Borosilicate is highly absorbent of infrared radiation. If the effect is radiative, we would expect the borosilicate to absorb any extra thermal radiation from the hotter glass, preventing Tbottom from increasing. This is precisely what happened. During these runs, the Net IR did not change when the hotter glass was swapped in, nor did Tbottom respond to the swaps.
Thus we have to conclude that the increased thermal radiation from the hotter glass is what caused the bottom to increase in temperature, even though the hotter glass itself was much cooler than the bottom.
Analysis
How does this not violate the laws of thermodynamics, wherein heat only flows from hot to cold? The answer is that one must consider all the heat flows in the system. When Tbottom is at a steady temperature around 108ºC, it is because the amount of heat it is gaining from all sources is equal to the amount of heat it is losing to all sinks. At that temperature, the only heat it is gaining is from the sunlight, which is actually the thermal radiation emanating from our nearby star with its surface temperature of 5900K. It is at the same time losing heat conductively with the bottom of the hotbox and the walls, convectively with the air in the box, and radiatively with the plastic wrap layers, the glass, and the sky above. The hotbox itself is also losing heat convectively with the outside air.
When the hotter glass is swapped in, effectively the only change is that now the bottom is losing less heat radiatively to the glass. With all else being equal, it is thus now gaining more heat from the sunlight than it is losing to all other sources, and the result is an increase in temperature. Notably, without the much hotter sun as a heat source, the increase in temperature would not be observed; only a reduced cooling effect would occur.
Conclusion
There are several key differences between the experiment performed here and the theorized radiative greenhouse effect, such that this experiment does not serve as verification of the latter.
- The hotbox is heavily insulated and enclosed to suppress convective heat loss. By contrast, the majority of the heat loss by the Earth’s surface is due to convective loss to the air.
- The temperature ranges are different, with the Earth’s surface being around 15ºC and the air ranging from 15ºC to -55ºC with increasing altitude, contrasted with 100ºC for the black bottom and 30-70ºC for the glass.
- The colder object is a solid plate of glass as opposed to a column of atmospheric air, i.e. a large volume of gas.
- The warmer object is a uniform pitch-black plate, as distinct from the Earth’s surface with its varied terrain, soil, plant, foliage, ice, snow, water, etc.
- The greenhouse effect is due to the air becoming more absorbent of and emissive of thermal radiation at the same temperature, while the experimental result was due to the colder object retaining the same emissive properties yet becoming hotter.
- The observed time duration was minimal and no conclusions can be drawn about the total magnitude of the effect or what effects it may have in the long run.
Indeed, the greenhouse effect theory has a large gap when it comes to the empirical demonstration of the greenhouse effect’s total effect on surface temperatures. It largely relies on a simplified calculation that shows the Earth’s average temperature would be -18ºC without any atmosphere. This calculation even includes the albedo effect of clouds, which would not be present without an atmosphere. Further it ignores all the distinctions listed above, and doesn’t account for the adiabatic lapse rate, which is the main reason why the bottom of the grand canyon is +50ºC warmer on average than the peak of Mount Everest. Such a lapse rate occurs entirely due to non-radiative effects, and therefore would have some effect even if the air were fully transparent to infrared radiation.
- The observed temperature increase here was due to swapping in a hotter glass plate that was externally heated by another heat source (i.e. neither the sun nor the black bottom). This is in contrast with the greenhouse effect, where the atmosphere’s thermal radiation that is theorized to result in a much warmer surface temperature, is initially warmed by the surface itself. This gives the appearance that it presents a situation where an object is able to heat itself up with its own heat – first the atmosphere at the same temperature emits thermal radiation according to its own temperature, which then results in a warmer surface, that then in turn heats the atmosphere further, etc., in a (diminishing) feedback loop.
Although the experiment performed here indicates that the presence of the sun ought to make this possible, and Infrared Halogens appear to exploit an analogous mechanism to reduce energy consumption, an experiment should be done to confirm this is the case. It is notable that past experiments have failed to definitively show a powerful effect: R W Wood’s 1909 experiment showed “scarcely a difference of one degree” between two hotboxes, one with a radiatively-absorbent glass lid and one with a radiatively-transparent rock salt lid, where we would expect the radiatively-absorbent lid to result in a much hotter temperature due its higher radiative emissions as well. More recently, a re-do of Wood’s experiment “performed more carefully” by Pratt actually replicated the result, despite being heavily critical of Wood: a difference of 1.1ºC between the floor of a hotbox with a glass lid vs one with a rock salt lid. Pratt further showed that the glass lid was actually 6.2ºC hotter than the rock salt lid. This further complicates the analysis as convective heat transfer is proportional to temperature difference and indicates the +1.1ºC result is at least in part due to reduction of convective heat loss.
- Finally, even if the basic greenhouse effect were to be conclusively demonstrated, it is well-known that the current levels of CO2 are already saturated with regards to higher concentrations causing higher thermal emissions. The enhanced greenhouse effect is said to occur due to the much colder higher-altitude layers of the atmosphere absorbing and emitting more thermal radiation, thus having a cascading effect on the air immediately below it, which in turn has an effect on the air below, and so on up to the surface. In other words, the hotter absorbing element in the enhanced greenhouse effect is also atmospheric air, and not a solid object. This provides further complications as atmospheric air can freely travel vertically up and down, and hot air tends to expand and rise.
In conclusion, although the thermodynamic possibility of colder objects causing warmer objects to increase in temperature has finally been conclusively demonstrated, a lack of experimental verification of other facets of the greenhouse effect theory warrants caution before accepting its conclusions.
Related