The first five years of the RCPs
Aug 11, 2015
Bishop Hill in Climate: Models

Further to yesterday's post on butterflies and RCPs, I wondered just how things were panning out for the RCPs since they were issued five years ago. I wasn't really expecting very much from this analysis since five years is not very long, but it turned out that there is more of a difference than might be expected.

The RCP data is for the mid-year carbon dioxide concentration and it turns out that the June figure from Mauna Loa has just tipped the 400ppm mark. RCP8.5 predicted that 2015 would be the first year in which the 400ppm mark was breached at the mid-year point, so at first glance we are indeed on the RCP8.5 pathway.

However, it turns out that at the point the RCPs were issued Mauna Loa was already several PPM ahead of the RCPs, so it is necessary to put everything on the same baseline to get a meaningful comparison. Here's the result - a graph which shows the growth in CO2 concentrations for each time series since 2010.

As you can see, RCP8.5 carbon dioxide concentration is already more than 1ppm more than reality. That's not a great deal in absolute terms but is quite a lot given we have only had five years for any divergence to show itself. We actually seem to be tracking RCP4.5 most closely. (Slightly counterintuitively, RCP6 produces the lowest carbon dioxide concentrations for the first few decades before accelerating in the second half of the century.)

Of course it's really a bit early to draw strong conclusions from such an analysis. The big question is over what happens in the middle of the century, when RCP8.5 and RCP6 start to curve upwards and RCP4.5 heads the other way. But, as Matt Ridley and Blair King have so ably described, a brief look at the assumptions behind RCP8.5 suggests that it is more a rather implausible worst case scenario than anything else. Certainly any suggestion that it represents "business as usual" is likely to mislead the lay reader.

Article originally appeared on (
See website for complete article licensing information.