WUWT readers may recall this story from November 3rd NOAA deletes an “inconvenient” kids science web*page where NOAA took down a web page called “It’s a gas, man” that talked about a tabletop science demonstration that kids could do themselves to “prove” that CO2 retains more heat. Problem was, the experiment as presented then was flawed, and when it received some attention from skeptical websites, NOAA recognized the flaw and took it down, replacing it later with an updated page.

Fast forward past Climategate to this past Thursday Dec 17th, and we find that the BBC decides to try essentially the same experiment on live TV for an impressed and non questioning audience.

Click to play the video at the BBC website*
Only one problem, the BBC presenters botched the experiment. Fortunately we can show why, because WUWT reader* Professor Kevin Kilty of the University of Wyoming, who took an interest in recreating this experiment with students in his physics class well before the BBC did their experiment, has conclusively demonstrated its scientific shortcomings in an experiment log he sent me on December 20th showing results of a November 23rd experiment run.

What got me connecting what Professor Kilty had done to the BBC live TV experiment was a comment from WUWT reader Bryan C of the UK. Here’s an excerpt:*

Dear Anthony

Here’s something I found shocking and that you don’t see every day: the British government’s former chief scientific adviser Professor Sir David King flagrantly lying on national television to boost the dubious idea that some foreign agency (the Russian secret service?) was behind Climategate.


This was in the context of BBC 2’s Newsnight staging a peculiar experiment, with a politically-correct black female “space scientist” heating two bottles – one containing “air” (last time I looked, that included carbon dioxide anyway) and one containing “atmospheric air with a greater concentration of carbon dioxide” (they didn’t say how much they were adding, of course, but I’d bet it was substantially more than 0.000388%!). Surprise, surprise — the latter bottle grew hotter… Of course it did. A greater amount of carbon dioxide will be warmer when heat is applied. This is not a surprise! The proportions are key, of course, as you know.

Newsnight itself characterised the effort right at the start as a “very unscientific experiment” — so why do it at all?! In fact the “science” as presented was misleading and selective to the point of deception.
Indeed when you watch the BBC video, it is clear that there’s no sort of control of any kind, the thermocouples were placed haphazardly at different angles into the bottles, and there’s likely alignment differences between the lights illuminating the bottles. It seems so from my viewing of the video.

Professor Kilty also viewed the BBC video and writes:

You can see that the two bottles start at temperatures of 32+ C. Perhaps the house is this warm, we don’t keep ours this warm, but more likely they have run the experiment and know pretty well in advance how it will turn out. I tried to see from the size of the spot on the bottle if one or other is obviously closer to the lamp–I can’t– but what really matters is the thermocouple, of course. The NOAA description in “its a gas, man” looks like the epitome of careful research in comparison.

This is just kid science. The BBC did their best. Not as good as the ten-year old of a couple of weeks ago, though. It is funny that the journalist sells this as “proof” of global warming early in the sequence.
Here is what a properly conducted experiment looks like, as performed under professor Kilty’s supervision by students at his lab at the University of Wyoming.


Date: December 20, 2009.

Figure 1. Two separate set-ups running at the same time. While it looks like our lab is bathed in mood- lighting this is an illusion. The extremely bright filaments fooled my automatic camera. The room was brightly lit. The nearest set-up uses Moll-type thermopiles, while the distant setup is more like the NOAA description, except with thermocouples replacing lab thermometers.*
Are there endless silly or meaningless experiments and demonstrations that one can do with carbon dioxide (CO2)? We’ve seen a few on WUWT recently.1 On Tuesday November 3, 2009,WUWT exposed one endorsed by a major scientific organization under the headline NOAA deletes an inconvenient kids science web page.

Indeed, all reference to this page appears now gone at NOAA. But, thanks to the efforts of WUWT, and the help of the way-back machine,2 selected physics students in three of my courses at LCCC got to try the experiment as someone at NOAA designed it. As it turns out, this experiment is silly for what it attempted to show, but it provides darned good lessons about scientific experiments.

The first group of physics students to get a crack at greenhouse warming in a two liter bottle were from my Physics 1050 course – physics without math. They set the experiment up as closely to the NOAA specifications as possible and made Runs 1 and 2 as I describe below. The algebra based physics course got a stab at it next, then the calculus-based physics class had their try. These classes modified the experiment to get a better picture of what was going on. They performed Runs 3 and 4, respectively.

1. Procedure

The NOAA web-page suggested doing the experiment according to the following recipe.
(1) Partially fill both bottles with water. In fact, we filled each with the same amount of water – about two inches worth.
(2) Add the seltzer tablets to one of the bottles. We delayed this step until we had the apparatus assembled.
(3) Suspend the thermometers inside the bottles in such a way that you can measure the temperature of the air and seal the tops with molding clay. We thought there was little reason for sealing the top completely, so we used a cork stopper with hole large enough to allow gas generated in the bottle to pass out around the thermometer.
(4) Place the lamp at equal distance between each bottle. This is the tricky step in this seemingly simple experiment.
(5) After an hour, measure the temperature of the water in each bottle. We thought the word “water” was a mistake here because there was no instruction to make the amount of water in each bottle equal, nor any reason the water would be of interest when the thermometers were suspended in air. Accordingly we monitored the temperature of the air to equilibrium at least, which was less than an hour.
Despite the simplicity of the procedures, we encountered plenty of experiment design issues. These included:

1) the typical lab thermometers have fiducial marks at one-degree interval and so temperature can be read to a resolution of about 0.5oC at best,3

2) the marks are actually not of uniform size,

3) it is really difficult to get a label completely off a two-liter soda bottle, and so there is a readily available shield or
reflector to confound one’s results. Finally, there is that deceptively simple step 4; Place the lamp at equal distance between each bottle.

Figure 2. Thermocouple in a two-liter bottle. Note that the thermocouples are not perfectly vertical, nor are they likely to be perfectly centered. The near thermocouple points away from the lamp and residue from the label shields the thermocouple.*

Although a person can purchase clear light bulbs that allow one to see precisely where the filament is, and what geometry it has, there is almost no way to decide what is the exact center of radiation. After all 95% of the radiation leaving the lamp is infrared and invisible. From outside the lamp does radiation appear to come from the filament? Or does the bulb envelope appear as the source? Moreover, even if a person can decide where is the center of radiation, there are a host of other ways to get the set-up wrong. Figures 2 and 3 show some. Students rarely noticed if the thermometer was centered and vertical or if it stayed that way during the course of the experiment – and as one might expect to happen sometimes, thermometers in the CO2-filled bottle tipped toward the lamp, as Figure 3 shows, while those in the control bottle tipped away like Figure 2.

Figure 3. A thermocouple in a two-liter bottle. Note that this thermocouple points toward the lamp, and has a reflector from the residue of the label torn from the bottle.*
2. Results

The table below summarizes our research of November 23, 2009. The first experimental run, using ordinary lab thermometers, appeared to detect an increased temperature rise in the CO2-filled bottle. However, students failed to appreciate at this point that repeating this experiment, no matter how exactly, could arrive at a different outcome.

Indeed, Run 2, using six thermocouples read to a temperature resolution of only 1oC indicated no average difference in temperature rise, but showed greatest temperature change in some bottles without CO2.

Run 3, using thermocouples read to better resolution of 0.1oC, showed the greater average temperature rise to occur in the non-CO2 bottles. In this case students swapped thermocouples among bottles to make certain no variation was the result of mis-manufacturing of these sensors. We concluded from these results that sufficient replications of properly randomized runs would likely show no detectable difference at temperature resolution typical of equipment in K-14 science labs.

Run 4 made use of Moll-type thermopiles. These devices capture a very broad spectrum of radiation, from far IR through visible, and conveys it to a highly absorptive collector at the base of a conical reflector. A series connection of 17 type-K thermocouples indicates the temperature rise of the absorber. These thermopiles have a sensitivity of 0.28mV/?W; a voltage that good quality bench multimeters can read easily. Figure 4 shows one of these devices.

Figure 4. A Moll-type thermopile. Picture from Cenco on-line catalog.*
In these runs we organized a moll-type thermopile to look at the lamp through our plastic bottles. When the potential of the thermopile became stable we then dropped two selzer tablets in the bottle and monitored the decline in potential until it became stable again. In this manner we managed to avoid all confounding influences except variations in one plastic bottle to another, and possibly extremely small variations in aim of the thermopile. The average decline was 0.095mV .
This translates into a typical decline of 0.34 ?W of radiation power entering the conical collector.

3. Discussion

The presence of CO2 in a plastic bottle reduced radiation collected by a thermopile looking through that bottle. But what radiation is reduced, and what causes the reduction? We are pretty sure that visible light isn’t reduced as there is no visible difference between bottles with CO2 and those without. Thus, the difference is likely in the infrared (IR) part of the spectrum. CO2, as we have heard interminably for the past 25 years, absorbs certain bands of IR radiation, most notably in the IR near 2, 3 and 4 micrometers wavelength, and in longwave bands between 13 to 17 micrometers wavelength. At thermal equilibrium CO2 will radiate in these same wavelength bands as much power as it absorbs. The radiated radiation does not travel in the same direction as the absorbed radiation was traveling, however. It is radiated uniformly in all directions. In the case of our experiment this leads to a small decrease in power reaching the Moll-type thermopile.

Applying this to the case of a simple Earth atmosphere, containing nothing but CO2 and having no weather, leads one to conclude that longwave radiation leaving the top of Earth’s atmosphere will decline in magnitude slightly. This decrease in longwave power traveling away from the surface forces the Earth’s surface temperature to rise slightly in order to maintain its thermal equilibrium. This is the “greenhouse effect” in its pure form.

Table 1. Various runs of our experiment. Thermometers run showed the expected enhanced ?T of the CO2- filled bottle. First run with thermocouples, though, showed no average difference, but was fraught with con- founding influences. Temperatures were displayed at the whole number resolution because of the digital readout. Run 3 thermocouples read with a digital display having 0.1oC resolution and showed the largest effect in bottle with no CO2. Thermopiles were read with a bench DMM having 10 ?V resolution.*
4. Conclusions

When this experiment is set-up according to the prescription on the NOAA webpage it is quite possible to get a difference of temperature of 1 oC between or among thermometers even if none of them contain any CO2. A properly randomized experiment will likely result in no discernible difference among thermometer readings irrespective of CO2 in bottle or not. The issue is one of not enough magnitude of effect to resolve on typical lab thermometers.

An instrument as sensitive as a Moll-type thermopile can detect a small difference in radiation passing through bottles filled with CO2 as compared to an identical bottle not filled. The amount of IR power re- directed by a two-liter, CO2-filled bottle appears to be about 100?W/m2.

The most important result of this experiment is how it shows students so many issues of experiment design. First, there is the issue of how difficult temperature measurements are to make accurately. Students are quite surprised at this. They are equally surprised that seemingly identical temperature sensors will not measure identically. Second, there is also the difficulty of proving conclusively that A causes B when the experiment includes confounding factors. This is an important lesson about the value of skepticism in climate change research, observations, and publicity. If X, Y, and Z cause B just as readily as does A, then what allows one to claim A causes B?


1See for example: http://wattsupwiththat.com, 2009/11/18/, Climate Craziness of the week.

2The way-back machine still has a copy of this web-page at:
http://web.archive.org/web/20060129154229/http://www.srh.noaa.gov/srh/jetstream/atmos/ll gas.htm

3Actually it is possible to tell that the liquid in the thermometer is above half
way, but below the next fiducial mark. Thus, I suggested students could resolve
the least significant digit as .0, .2, .5, .8, respectively.

A complete report on this experiment from Professor Kilty in PDF form is available here


Back to the BBC video, Bryan C points out some problems with statements by Professor King, who joined the group after the CO2 bottle experiment was performed. Here is his comment, continued.

Professor King adroitly avoided key questions. Anyone there with any knowledge of the science could have taken him apart. The BBC clearly wasn’t interested in finding anyone equipped with the facts who could have countered the orthodoxy. In contrast, we had an ignoramus who expressed scepticism at the beginning saying he was now completely convinced. Others taking part who maintained their scepticism unfortunately didn’t have the facts at their fingertips to back up their positions.

Professor King’s assertions about Climategate (from 6:20) were particularly shocking. He conceded that the behaviour shown was unacceptable, but no conclusions were then drawn by him — the program simply moved on! But I was most stunned by his obfuscatory introduction of the conspiracy theory about “agencies” which went unchallenged, and involved a direct fabrication about mobile phone conversations.

“Remember that these emails go back to 1998 and they’ve been accumulating them and just released them in the week before Copenhagen…

“Let me also make this allegation for the first time in public. It’s an extraordinarily sophisticated piece of work to hack into all of these emails and mobile phone conversations, right? What agencies have got the sophistication to manage that? I leave you to think about that.”

Of course, the most likely scenario is not of an outside hacker but a whistleblower inside the CRU who pulled them together and released them. The suggestion of “an extraordinarily sophisticated piece of work” doesn’t really hold up if you’re just referring to emails, but introducing the idea of monitoring mobile phone conversations (a complete lie as far as I’m aware) serves to boost the conspiracy theory and muddy the waters. And this man was Britain’s most senior scientist?

I hope you can draw people’s attention to this deception!

Regards Bryan C
Clearly there has never been any mention of “mobile phone conversations” in any known discussion about the Climategate incident. This appears to be a complete fabrication by Professor King. It is troubling that the BBC has not corrected this.

All in all, this was not a well thought out or well researched video presentation by the BBC, and in my opinion it does a disservice to the citizens that pay taxes through television licenses to support the BBC.

UK readers are encouraged to make the issues and independent experimental results known to the BBC and to media monitors there.