The Temperature measurement Conundrum

Recently, I saw a video by Prof. Bruce Bugbee - founder of Apogee Instruments, saying something to this effect -

If you asked me in highschool - can we measure air temperature accurately? I would have said, ofcourse! Thats a simple measurement involving a thermometer. If you had asked me the same question while I pursued by higher education - I would be a bit less sure. And if you ask me now, as a practicing scientist - I would be not at all sure!

Our experience at APV Parbhani was no different! We wanted to look at the differential air temperature below solar modules as compared to between rows of solar modules. And inside shadenet, as compared to outside shadenet.

Why is it that something that a humble thermometer can measure - so difficult to assess?

Let’s Imagine what happens. Where does air get it’s temperature from - solar radiation. If you switched off the sun - in a few days, all air would go below freezing. Anything the sun’s ray’s touch - gets heated up.

A large part of air is directly heated by the solar radiation, while some part of air gets heated after being in touch with other actors in the scene - the ground, the leaves, a tin roof. So air that is immediately in vicinity of the hot tin roof will be at a higher temperature than beyond. But since leaves are living beings, they cannot get heated to as high temperatures - and so they maintain a small blanket of air around themselves at a somewhat lower temperature. The ground itself is at some temperature higher or lower than air - and so air temperature changes as you move higher from ground. If your eyes could sense these temperatures as colours, you would see a highly colourful image of microclimates around you.

Now - all of the above is in still wind. In the next moment - if there is a breeze, then all air everywhere gets fully or partially mixed and what you have is an average soup of air temperature.

Can you now imagine how difficult it is to even define what should be your temperature of interest for something you want to prove? The metereology guys have defined theirs - for reporting a typical region’s ambient air temperature as something that is 2 m above ground. But that’s arbitrary and may not suit your own purpose.

Now add to this - one more major moving part. If our temperature sensor is exposed to radiation, then the sensor itself will have a solar heat gain and measure a higher temperature than the ambient air. Hence temperature sensors have to be shielded from all radiation in a dark enclosure.

But hey - this dark enclosure cannot have stale air inside, else this air will not be a clear sample of the air outside! So again you are measuring the wrong temperature.

So you put a fan or an aspirator inside - and you have aspirated shields. They work fine - unless there is a morning dew accumulated on the shield. The fan evaporates this mist, makes the air inside cooler by latent heat of evaporation - and gives you a major error in temperature again!

So there you have it. At the minimum - you need to have 2 to 3 independent temperature measuring mechanisms so that you have a reasonable confidence that what is being measured is on the right lines, if not completely accurate. And a healthy suspicion of the measured values, rather than jumping to quick conclusions.