The Mother of all climate models and deglaciation hiccups

In his latest book, The Vanishing Face of Gaia: A Final Warning (Allen Lane, London, 2009, ISBN 978186141850), James Lovelock more or less gives up on the ability of humanity in general, and science and engineering in particular, to fend off looming climatic catastrophe. He reserves his sharpest criticism for what he calls ‘American science’; a fundamentally reductionist approach that is fed into prediction of the future. For Lovelock, the assumption ‘that all we need to know about the climate can come from modelling the physics and chemistry of the air in ever more powerful computers’ has been a disastrous mistake. He is obviously not one for humble retrospection, as his early Gaia writings had at their centre a sort of reductio ad absurdum of that now prevailing genre in Earth system science. Daisyworld, reduced a planet’s life forms to white and black daisies, whose interplay with climatic change was governed by a formula known as a difference equation in the manner of Lotka and Volterra’s work on predator-prey interrelationships. The simplest difference equation is xnext = rx(1-x). Solving such non-linear relationships for minute increments in x led to the unmasking of chaos theory, the first instance being Edward Lorentz’s discovery that the simplest models of climatic turbulence go wonky if you tinker with them: the ‘Butterfly Effect’.

When his Gaia hypothesis drew together all manner of people from New Ageists mathematicians working on complex systems James Lovelock was exposed to friendly criticism and education about non-linearity and chaos. Clearly that revolutionised his world-view, which is fine, albeit a cause of some glumness for him. Far sadder is that he is probably right in criticising climate modelling – now that it has a stranglehold on the entire climate debate and indeed on the ears of the ‘Great and the Good’. A measure of where modelling has led is a simulation of what happened as the Northern Hemisphere emerged from the last glacial maximum, between 22 and 10 ka (Liu, Z. and 13 others 2009. Transient simulation of last deglaciation with a new mechanism for Bølling-Allerød warming. Science, v. 325, p. 310-314). These ~10 millennia saw a return to a see-saw climate that lasted from 60 to 30 Ma as the Earth cooled towards the last glacial epoch, dominated by cooling-warming cycles with a similar pattern of slow cooling-sudden descent into frigidity-thousand year cold spells-sudden warming known as Dansgaard-Oeschger cycles.

The Chinese-US team developed and ran the first synchronously coupled atmosphere-ocean general circulation model to investigate a hiccup in warming of the sea surface one northern ice caps began to melt decisively. It is said to be ‘one of the most epic numerical modelling efforts of the climate community to date’ (Timmermann, A. & Menviel, L. 2009. What drives climate flip-flops? Science, v. 325, p. 273-274). Epic, well yes: one of the world’s largest operational supercomputer (Jaguar at the Oak Ridge National Laboratory, USA) was wrangling for 18 months. Lots of known empirical data for the period were fed in: insolation determined by astronomic effects; changes in greenhouse gases from ice cores; shifts in coastlines and ice-sheet volumes. Tinkering with the model involved varying freshwater influx to high-latitude North Atlantic seawater. The result was crude simulation of what actually happened to sea-surface temperatures at several locations around the North Atlantic, giving some insights into why changes occurred. But climate scientists have long suggested mechanisms for the Dansgaard-Oeschger cycles, Bølling-Allerød warming, and the final frigid paroxysm of the Younger Dryas in much the same framework, the only difference being they didn’t produce numerical models that mimicked reality.

It seems that another 2 to 3 million hours of time on Jaguar are needed to bring the project through to the present. The enormous funding needed to get this kind of number crunching done can only have been on the back of claims that it will help predict future anthropogenic climate shifts. Based on real data, it still didn’t get things right – millennium-long cooling and warmings are not trivial events. There are conflicting kinds of data for changes in the parameters since the start of the Industrial Revolution preceded by 10 ka of relatively stable Holocene conditions. The best that climate forecasting for the next 100 years has been able to do, also using pretty large amounts of CPU time, is a range of straight lines showing increases in global mean surface temperature. Yes, hindsight is wonderful…

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s