- JoNova - https://www.joannenova.com.au -

Has CO2 warmed the planet at all in the last 50 years? It’s harder to tell than you think.

Joint Post: Jo Nova and Tony Cox

Even most skeptics agree that the world has been warming during the last 50 years, but there is apparently no significant underlying warming trend in 46 out of 47 years of data. Something decidedly unusual happened to the world in 1977 and we don’t know for sure what it was. The world got warmer, and the change “stuck”. But there were no extra emissions of CO2 in that year, so there is no reason to pin this to CO2.

It’s difficult to believe we are not sure – but the last 50 years of warming trend depends on that single stepwise leap in 1977. Look at the graph below. Does it show one strong underlying warming trend, or is it really a trend so insignificant that it wouldn’t exist if there was not a step change that artificially bolstered it?

A series of two flat lines can appear to be a continuous warming trend if a linear trend line is fitted because it ignores the step change.  McKitrick and Voselgang

This step effect was first noted by David Stockwell in 2009

The continuous warming appears to be obvious in the records of the lower atmosphere when we draw a linear trend line.

McKitrick and Voselgang (click to enlarge)

But look again at the data, allowing for the step-change in 1977.

McKitrick and Voselgang (click to enlarge)

The problem with step changes is that two flat lines with a rising step change will artificially produce a rising linear trend that is meaningless. McKitrick and Voselgang[i] account for the step change and show that significant warming trends in the low to mid troposphere only occur from 1958-2005 if we include the step-change. If the step change is removed, the underlying trend is not significant. In other words, 46 out of the 47 years of atmospheric data do not produce a meaningful rising trend.

If CO2 is having an impact on our atmosphere it is impossible to say with any certainty what that effect has been.

 


[i] McKitrick, R. and Vogelsang, T. J. (2011), Multivariate trend comparisons  between autocorrelated climate series with general trend regressors, Department of Economics, University of Guelph. [Discussion paper PDF]

8.4 out of 10 based on 87 ratings