Dr Roy Spencer has done an interesting experiment using two containers with water above which one has a silver foil which partly obscures the container from the cold sky.
This shows that water exposed to a higher radiant temperature is warmer. However, reading the comments I found some people arguing that this didn’t prove IR could heat water.
The controversy
For those who do not know, the controversy arises because IR is absorbed in the top few um of the water which evaporates off and/or because some people don’t believe in IR. This is all part of the argument that “back radiation” either doesn’t exist or cannot heat the sea. I usually keep away from the arguments, because back radiation exists. But this concept comes from greenhouse warming models which I describe as “noddy science”. They are just bad physics, not because of back radiation but for other reasons. So I have a lot of sympathy for those who don’t like these “backradiation” models of how greenhouse gases work and try to find fault with them.
Also, I don’t think the argument about heating water is completely daft – but I wish those who argue above water heating spent more time doing practical experiments and less time arguing the theory.
The Story
As I drank my whisky last night I thought “This is easy to prove”
I filled a tray ~1cm deep with cold water (at 18.6C using IR thermometer – which is possibly a few degrees higher than I expect, but perhaps it was sitting in a pipe). I put it under the grill which I then turned on as per this exhaustively researched experimental set up:
I set the timer to 5 minutes and then after five recorded the surface temperature without moving it – so when still irradiated. The IR thermometer read 63C. I then removed it – which inevitably stirred it, the IR thermometer read around 50C, I then gave it a good shiggle to ensure it was mixed up and it still read around 50C.
The 63C was at a low angle with IR still present and probably raised by reflected IR – but it might be interesting to repeat it with deeper water to see if there was noticeable surface warming. But the key measurement of bulk temperature had definitely increased by a good 30C.
For interest I measured the various surfaces inside the grill and they varied from 30 to 60C. The 30C was at the far end which had much less radiant heat, but was in much the same air temperatures. So, this is the worst case scenario of “air heating”. Air heating cannot explain the increase in the water.
The 60C was the surface closest to the tray – so it was being heated, but only to a very similar temperature as the water.This shows that IR was heating both the side surface and water in a similar way. And clearly there was no significant temperature gradient that could explain the rise in temperature of the tray from the surrounding surface.
I then touched the grill base and confirmed it was much cooler than the heated water (so no conductive transfer from the base – probably the reverse).
And whilst I did not measure it directly, the hotter air leaving the grill remained pretty cool and certainly not high enough to heat the water this quickly. (It would be like trying to make toast on a central heating radiator!)
Heating was not from visible!
Finally, to avoid some comments on visible heating, the grill was so dark I had to use a very long expose (~10seconds). The element was actually much darker than the picture suggests.
This tells me that the intensity of the visible was less than if I had left the tray exposed to the kitchen light.
Having worked with LED lights, I can be very confident the light level was much lower than a 20ma LED. This means the available visible heating was much less than 20mw and probably less than 1mw. So the visible was around 1millionth that of the IR and if anyone even suggests the heating was visible red, …
Conclusion
IR certainly heats water from above: and if the effect is just at the surface, then thermal conductivity rapidly transfers that heat into the bulk of the water.
It might be worth trying this with deeper water. But I can probably calculate it quicker (see: Caterpillar III) than the ~20min experiment would take with deeper water ~5cm deep.




