Skip to content

A primer on real-world temperature measurement

There is a particular design of nuclear plant called a boiling water reactor (BWR).  The name comes from the fact that the steam is generated inside the core rather than in a secondary loop using a steam generator.  Because of this design, BWRs require a tank of water called a suppression pool where the nuclear steam can be quenched in the event of an emergency.

The temperature of this suppression pool has to be carefully monitored to ensure it can absorb all the energy of the nuclear steam.  The suppression pool contains about a half million gallons of water (23 percent smaller than an Olympic pool) and is monitored by a dozen or so instruments, usually resistance thermal detectors (RTDs) or thermocouples (TCs).

The RTDs and TCs are safety-related, nuclear-grade instrumentation, usually with a guaranteed accuracy of about a half degree F.  This is better than the thermostat in your home, worse than a high-end laboratory.  The RTDs and TCs get a regularly scheduled calibration called a surveillance in which the instrument is measured against a standard to ensure it still reads correctly.  The calibration interval is based on the known drift characteristics of the instruments.  The instruments are mounted in guarded enclosures called thermowells to ensure they don’t get damaged.

Once per day, the power plant operators read all the instruments and compute a bulk average temperature for the pool and verify it still meets NRC limits.

Here are some difficulties.  The instruments are good, but not perfect, and normal instrument tolerances can sometimes add all in the same direction to shift the average instead of all the errors canceling out.  Sometimes an instrument goes bad and has to be thrown out.  Sometimes they die from vibration or some other environmental problem.  Sometimes nearby steam leaks influence one or two particular instruments.  Sometimes cold water gets added to the pool, and that also affects instruments where the cold water inlet is located.  Sometimes there are wiring problems on the TCs or power supply problems on the RTDs.  Sometimes HVAC ducts can slightly chill one area.  Sometimes a temperature is measured while the environment is changing but the device hasn’t had time to react.

To moderate all this, the operators usually turn on big pumps to stir the water, mixing it up as thoroughly as possible to even out temperature differences.  Even then, the pumps add heat to the water, so even after they compute what the temperature is, it’s never quite certain what the temperature was.

When you add up all these uncertainties, the gray zone around the bulk average temperature ends up being a degree or so.  Sure, they use electronic calculators, so the digits read out as 93.0123456789 degrees F, but the real world uncertainty amounts to a degree or so, and this is for a tank of water the size of a swimming pool that’s been thoroughly stirred to make it easier to measure.

What are the uncertainties in measuring the temperature of the whole planet?  What, in fact, does the temperature refer to — air, water, dirt?  Some weighted average of the three?  At what depths or heights do we obtain a truly representative measurement of each?  Who gets to decide that?

Further, temperature measurement needs to account for time of day, solar loading on the temperature device as distinct from its surroundings.  It needs to account for the fact that the temperature of any one location is constantly in flux because of the time of day, time of night, and time of year.  It needs to cover the earth where there are no large gaps.  The instruments need to be protected and regularly calibrated (which there’s really good reason to doubt is happening).  They need to be sited away from big city heat islands with no air conditioners near by, no parking lots, no high buildings, nothing that reflects extra solar heating or affects them by cooling, such as shade or a swimming pool (all which we know isn’t reliably happening).

Now, what do you suppose the uncertainties are in any data set from any one day?  I’m thinking the uncertainty is bigger than the measurement errors in the suppression pool at a nuclear plant, which is one degree or so.

And then after all that uncertainty is accounted for, you still need a meaningful method for taking thousands of measurements and computing a global mean temperature.  So how much weight do you give all the instruments across North America where there are many hundreds of weather stations compared to, say, the south eastern Pacific where there are almost none?

It looks to me like global warming is a lot of hot air from people who have insufficient experience with real-world problems involved in temperature measurement.   When people claim the whole earth is warmer by some fraction of degree than it was fifty or a hundred years ago, it indicates, in the judgment of charity, that they truly do not know what they are talking about.

2 Trackbacks/Pingbacks

  1. […] miniscule it’s beyond our power to measure it.  Long-time readers of AVFTA may recall that this is exactly what I said about the warmists’ claims about temperature measurement.  They cannot possibly know what […]

  2. […] the data set is subject to some unpublished degree of instrument uncertainty.  As I mentioned here, I have a modicum of real-world experience with industrial temperature measurement which has made […]

Post a Comment

Your email is never published nor shared. Required fields are marked *