A Problem With “Average Temperature”

A new paper by David Lobell and his colleagues at Lawrence Livermore illustrates one of the problems with the whole “average temperature” notion ensconced in our discussions of climate change. The researchers compared outputs from 12 climate models, looking at the differences between daytime highs and overnight lows. The lows went up a little bit more than the highs but they were similar. The difference shows up in the variability. Clouds, as we know, are one area of uncertainty in climate models, and clouds play a big role in the summer daytime high. But they do not play as much of a role in overnight lows. As a result, the modelers found much greater variability in the daytime highs than in the overnight lows:

These results highlight the importance of considering separately projections for Tmax and Tmin when assessing climate change impacts, even in cases where average projected changes are similar. In addition, impacts that are most sensitive to summertime Tmin or wintertime Tmax may be more predictable than suggested by analyses using only projections of daily average temperatures.


  1. John-

    I would not describe this as a problem with average temperatures, but rather a recognition that it’s not just the average, but also the higher order moments of the distribution that matter. This is something that (I think) is well known in the climate community.

    Global average temperatures is adopted as a metric not because it’s particularly meaningful, but for a host of practical reasons.


  2. Andrew –

    Thanks. I guess the “problem” I’m referring to here is the way the knowledge of the climate community is communicated to the interested public(s). In terms of impacts, winter temperatures and summer temperatures matter in different ways and for different reasons. Ditto day vs. night temperatures, both overall and seasonally. Boiling all that down to a single number, as is almost universally done in communicating to the public(s) about climate change, is therefore a problem. Note that the “problem” I’m referring to here may be more of a problem in my community (the communicators) than yours.

    I am guilty of perpetuating this problem, though I will note that my efforts to get beyond the single number problem, to differentiate between day and night temperatures, seems to fall into the “more information than a newspaper reader wansts” problem space. (Note to self: do not use phrase “diurnal temperature range” again in a newspaper story.) Not so much the winter-summer problem. People seem to get that one.

  3. An old discussion on sci.environment comes back to me. The discussion was about night-time radiative cooling and several people noted that even in the desert ice could form as the radiative exchange to cold space took hold and how this did not happen on cloudy nights. There was a guy from coastal Mississippi who did not get it and insisted that it was an old wives story. After some time I finally figured out that his air was at 100% relative humidity. I think they have to control for a few more things.

Comments are closed.