Benoît Rittaud is a maths lecturer and blogger at “Changement Climatique”. This article, from November 11th 2014 is one of a series analysing a recent series of reports by Meteo France containing climate predictions for France (Metropolitain and Overseas). The original article is at http://www.skyfall.fr/?p=1366
The Climate Models Are Capable of Anything (That’s how you recognise a Climate Model.)
by Benoît Rittaud
A quick glance at this August 2014 report on the Climate in France in the 21st century prepared for the Ministry of Ecology and Sustainable Development is good for a laugh.
See for example on page 17 this Table of Doom, the terrifying pith and moment of which has been reproduced just about everywhere in the media.
This table summarises the mean temperature anomalies for the periods 2020-2050 and 2071-2100 in comparison to the reference period 1976-2005 according to the climate models WRF and Aladin-Climat for each of the three scenarios RCP2.6, RCP4.5 and RCP8.5. The columns C25 and C75 correspond to the 25th and 75th centiles of an ensemble of regional models. WRF and Aladin-Climat are the two models considered in detail in the report. RCP means “representative concentration pathways.” The important thing to note is that the higher the number after the letters RCP, the greater the emissions of CO2 envisaged and therefore, according to the IPCC, the higher the resulting temperature. I see no other way that a model can be compatible with the findings of the IPCC.
However, in the table above, this necessary characteristic is flouted in no less than six places.
Whichever column you look at, you’ll find at least one case in which higher emissions lead to to less warming. The WRF model is the clear winner of the prize for the most bizarre results. For the period 2021-2050, which is practically tomorrow in terms of climate, in two out of four cases WRF predicts less warming for France in a scenario with a far higher level of CO2. And the differences are not negligible: from +1°C to +0.6°C in the summer, and from 1.3°C to 0.8°C inwinter. An English-speaker might be tempted to rename the WRF model, replacing the R by a T…
When I was a young university student, a physics lecturer had the bright idea of introducing us to the notion of the criterion of relevance. The idea is that when you’ve finished a calculation, that doesn’t mean your work is over. You still have to cast a critical eye on the result and the way in which it was obtained, in order to spot any possible error.
The fact that not one of the six climate scientists who worked on this report considered applying such a simple and obvious criterion of relevance as the one detailed above tells us much unfortunately about certain practices in climatology: worship of the Great Computer, devotion to “statistics”, and a complete lack of a critical spirit.