Sunday, 13 June 2010

Measuring the minimum temperature

Some years ago I attended a meeting of the American Statistical Association concerning environmental statistics (at that time a brave new world).  One presenter was expounding on the use of satellite sensing to measure temperature.  A comment from the floor was that "since there can be significant differences in performance of conventional thermometers when the batch of paint (from the same manufacturer) is changed how can you compare satellite and terrestrial observations?"

I think the presenter obfuscated, but the guy sitting next to me wrote quite a long note at this point.  As it was Senator (at that time) Al Gore I guess someone from NOAA got a few tough questions.

At a more earthly level it was rather chilly and foggy here last Friday.  I set off in our car to visit a property about 10km away (in a straight line).  The thermometer in the car gave the external temperature as -0.5C.  On getting to the property - clear of fog - I noticed the temperature had dropped to -3C where it stayed while I did my business.  Driving back home I entered fog again about halfway back.  As I approached the house the temperature  was back to 0C, which shows the insulating effect of moisture rather nicely.  (I have in the past compared the readings on the thermometer with official figures and found it to be quite accurate.)

What actually generated this post was getting up this morning to find the temperature (by a thermometer on the windowsill outside) was -0.5.  I turned up the fire to warm the house.  About 30 minutes later the thermometer read +1.7.  To assess whether this was due to the increasing warmth of the house or a real rise in temperature I moved the thermometer about 10cm out to the edge of the windowsill where it would be less influenced by the indoor temperature,  Within 30 minutes the temperature had reduced to +0.6 and a further 30 minutes later (by which time the sun had risen - although not directly on the windowsill in question) had got to -0.1.

So with thermometers, as with real estate, the key factor would seem to be location, location, location.

No comments: