“I need to hear an explanation of why carbon emissions have been going up over the last decade and temperatures haven’t been going up.” — Senator Steve Fielding

Steve Fielding is confused about something that puzzles a lot of people — why doesn’t the temperature trend follow the essentially monotonic increase in atmospheric CO2?

To some extent the answer is found in his graphs.

Over any time scale you choose, there are multiple influences on climate, and these influences in turn vary on different time scales. For example, the sunspot cycle (which affects the sun’s solar output) varies across about 11 years. The southern oscillation (ie. El Nino/La Nina cycles) — the dynamics of which are not particularly well understood or predicted — has a quasi-decadal cycle (three to seven years).

And this is just one of a number of such regional oscillations internal to the climate system that have global consequences for climate. Other events such as volcanic eruptions also have transient effects.

Given that the climate state over a given period is the result of the combination of these effects, you are bound to see a fair bit of noise in the time series. Steve Fielding’s HadCRU data shows this noise, and it should be clear that over shorter periods – say, of much less than a decade — it would be dangerous to draw conclusions about trends in climate.

For example, if you fitted a linear trend through, say, the 1992 to 1998 segment of the series, you would conclude that the world was warming at a terrifyingly rapid rate.

But 1992 was a cool year, largely because of the effects of the previous year’s Pinatubo eruption, and 1998 was a very hot year, largely because of a record el Nino.

Thus to expect CO2 and temperature to change in perfect lock step is to imply that climate must be influenced only by CO2 or otherwise not at all. The messier truth — as with most complex phenomena — is that there are multiple influences.

These are borne out by the sawtooth pattern of change seen in the data. So it is the underlying trend, and that only, that we believe to be driven by CO2. In light of this, it is neither surprising nor unprecedented that trends in temperature and CO2 don’t match over shorter periods.

The RealClimate site did a good job of illustrating this point (see here):

The blue lines are trends for successive eight year segments of the record. It makes clear that you can find just about any trend in the data if you look in the right place (or time, more accurately) — it also makes clear that even if you insist on doing so, there is no cooling trend in recent years.

Full disclosure, however: the above graph uses the NASA global surface temperature analysis, which is slightly different to, but broadly consistent with and equally well regarded as the Hadley/Climate Research Unit (HadCRU) data used by Fielding.

The complete NASA GISS record is below, and shows slightly different results for the most recent years.

Seen in the context of the longer term variability in the record, in neither case does the “so-called” cooling of recent years appear unprecedented — it is entirely consistent with the variability seen throughout.