Posted 4 December 2010
"Average temperatures or temperature ranges are often used as a simple proxy for climate.....Records kept over the years give us historical figures to make
comparisons between average temperatures then and now. This sounds
simple, but the very concept of an average temperature has no simple
definition. First, we have to realise that temperature is what is known
as an intensive property of matter. This simply means that it does not
depend on the nature or size of the material for which it is measured." - Newsletter of the Scientific Alliance
3rd December 2010
Can we really measure the climate?
Average temperatures or temperature ranges are often used as a simple proxy for climate. In combination with some description of rainfall, they encapsulate the essentials: in the Mediterranean it is typically hot and dry in summer and cooler and wetter in winter, and a continental climate is hot and dry in summer and cold with snow in winter, for example. But quantifying climate more precisely is fraught with difficulty.
Records kept over the years give us historical figures to make comparisons between average temperatures then and now. This sounds simple, but the very concept of an average temperature has no simple definition. First, we have to realise that temperature is what is known as an intensive property of matter. This simply means that it does not depend on the nature or size of the material for which it is measured.
So, for example, air and a body of water may have the same measured temperature at a particular moment, but their behaviour is very different. Air has a low thermal capacity (it take little heat to change its temperature), while water has a high thermal capacity and its temperature changes relatively slowly. In the present long cold spell in western Europe, ponds and lakes need a period of consistently sub-zero temperatures before they begin to freeze. Equally, as air temperatures rise, the ice may take many days to melt. A given volume of water has a very different thermal energy content than the same volume of air. This can be easily quantified and, in contrast to temperature, is an extensive property.
When trying to average temperatures, the first obvious rule is that the measurements must all be of the same material: you cannot average air and water temperatures, for example, and get a meaningful answer. This in itself is pretty obvious and, in discussing climate change, air and water temperatures are considered separately. However, the difficulties with averaging do not stop there.
Even if temperatures are measured under carefully controlled conditions as expected for official records, they will fluctuate quite rapidly depending on wind direction and strength, cloud cover, time of day etc. The convention is to measure a maximum and minimum shade temperature each day. These readings can then be used to provide average maxima and minima per month or year, or combined to give an overall 'average temperature'. And the figures for individual stations can themselves be combined to give national, regional and global averages.
These figures tell us something, of course, but the desire to quantify also obscures the detail. Say, for example, that place X has an average maximum temperature of +15°C and an average minimum of +5° and place Y registers +25° and -5°. Both have an overall average of +10°, but the actual climate experienced would be quite different. In a similar way, measured air temperatures in the shade bear little relationship to the apparent temperature in the sun. Although the measured shade air temperature might be the same whether or not the sun is shining, the effect on the Earth's surface of the sunlight is significant and, once the ground has been warmed, it will release its heat at night to keep the air somewhat warmer, at least temporarily.
Simple averaging can be deceptive in other ways as well. Depending on the weather conditions or time of year, either the maximum or minimum temperature might be more typical of the day as a whole, yet both are implicitly given equal weight. Nevertheless, it is arguable that such issues are not important when comparing time series of measured temperatures. For example, the Central England Temperature record (CET) is the longest continual record available, with monthly means being recorded from 1659 and daily means logged from 1722. Looking at this it is easy to see the recorded range and note that temperatures do indeed appear to have been higher in the latter part of the 20th Century, although they have dipped again since 2000. It is the changes which are significant rather than the absolute values, provided that all measurements are strictly comparable.
This, of course, introduces yet another concern. The same instruments would not have been used in the 17th Century as 300 years later and, with the best will in the world, it is difficult to guarantee that no artefacts have been introduced. Equally, it is hardly conceivable that the surroundings of the measuring stations will be unchanged over this period (although hopefully none of the weather stations is now in an urban area, on tarmac or near heat sources as some have been found to be in other countries).
A final problem to bring up with averages is that, to avoid giving a misleading picture, data should be taken from stations spread evenly over the Earth's surface. This is certainly not the case. In particular, there are large areas of the Arctic and Antarctic with no data being collected. The same is true for the open oceans, where collecting surface water temperatures reliably is enough of a challenge, without trying to measure air temperatures.
What we are left with then is an incomplete record of imperfect data, from which conclusions about climate change are drawn. This is the basis of the 'global warming' message. But actually the concept of global average temperature is again a little misleading, since the summary of the IPCC Fourth Assessment Report shows that the warming pattern is regional rather than global. Warming over the 20th Century was recorded on all continents apart from Antarctica, but was considerably greater in the northern than the southern hemisphere. Given the greater proportion of ocean in the south, this is not surprising.
But global averages are still the main measure and this is the time of year when preliminary conclusions are drawn about the current year, as the annual meeting of the UN Convention on Climate Change takes place. So far, the message being put out by the World Meteorological Organization is that 2010 is likely to be among the warmest three on record. Based on the temperature record, this is doubtless correct, but how meaningful is this?
The WMO points towards record high temperatures in Russia, China and Greenland to support its case. Meanwhile anyone mentioning record lows and pointing out that new records are set nearly every day somewhere in the world is told that this means nothing. In practical terms, life has to go on and adapt to whatever climatic conditions turn out to be. Measuring temperatures remains a useful thing to do, but we must be careful not to read too much into the average figures. And we should never forget that, whatever the temperature is, we still have only a hazy idea about what controls it.
The Scientific Alliance
St John's Innovation Centre,
Cambridge CB4 0WS
Tel: +44 1223 421242