Skip to content

There is a contradiction at the heart of climate science. It concerns claims that we understand the climate of the past 50 years or so and can reproduce it with computer models. Then we are told we don’t understand, or have even identified, crucial factors affecting climate change so our future projections will be poor. It is as if the past truly is a different country, and can be explained using climate models that are inadequate for predicting the future.

The past is of course, different. We have data for it, but what we don’t have is a consistent view of it and the future.

It is often said that all aspects of the way that the global climate has changed over the past 150 years (the instrumental period) can be reproduced by the models. We are also told they can identify the point (around 1960 according to the IPCC) at which natural changes in the climate become inadequate to explain the global temperature rise requiring anthropogenic influences (chiefly greenhouse gasses) to be introduced. This is proof that mankind is affecting the climate. The recent post-1980 warming spell cannot be explained any other way.

We are also told that using such models our ability to predict the future course that our climate will take is difficult. One recent paper bemoaned the fact that looking forward a few decades was problematical because “…over the course of the next 10-30 years the magnitude of natural decadal variations may rival that of anthropogenically forced climate change on regional scales…” (Bull. Amer. Meteor Soc, 92, 141 – 156.) See also Judith Curry’s perceptive comments.

Curious then that although natural variations are stronger than man-made ones over several decades we can be so sure that the man-made signal has been detected because natural factors average out and man-made changes are forcing the climate to be ever warmer. However, it depends upon what natural factors you have considered.

Indeed, one of the major themes of climate science over the past few years has been the importance of decadal cycles in climate. Changes due to many factors that occur at decadal cycles are poorly understood and given the fact that our measurements of the climate system only stretch back a few decades in most cases, our observational database is inadequate to recognise them all, let alone quantify them.

It is clear that the oceans – the main reservoir of heat in the climate system whose uptake and release of heat over multiple timescales provide the climate system with its ‘memory’ – are poorly understood. Computer simulations of such factors as the Atlantic Meridional Overturning Circulation and the Pacific Decadal Oscillation, to name but two in the Northern Hemisphere, are poor. Indeed fundamental processes involved in ocean-atmospheric interaction are not understood and new findings contradict current theories.

So, given our poor state of knowledge, who would make projections? Well, projections are useful in that they allow models to be calibrated against real-world data. A recent decadal prediction by the UK Met Office said (in 2007) that half of the years after 2009 would be hotter than 1998. The next few years will be interesting to calibrate that prediction.

My point is that predictions are made in good faith using the best data and models available at the time and they will be changed. But somehow we forget that they are temporary, and will be inevitably superseded.

But one conclusion from using models to explain the past is  downright misleading.

Look at the IPCC’s evidence (only published in 2007) that man-made global warming became discernable in the climate record around 1960. Before 1960 natural factors input into the climate model can explain what happened. After 1960 they cannot.


But hold on. What of our ignorance of decadal climate variations? How can we be so sure that we understand natural climate variations pre-1960?

The fact is that the IPCC’s model only takes into account changes in solar radiation and how those changes are mediated by aerosols in the atmosphere due to volcanoes. Nowhere is there anything about any other climate cycle due to the oceans or the atmosphere, or stratospheric water vapour changes, or what we now understand about how differential solar spectral variations may affect the climate, over periods decadal or otherwise.

Fig 2 is therefore nonsense. Unphysical. An example of after the fact parameter tweaking and curve fitting.

How for instance will the recent thoughts on variations in stratospheric water vapour possibly being responsible for up to a third of the warming between 1980 – 2000 factor into the equation? What difference will it make in the curve fitting and estimates of the 1960 transition date?

It has been reported that influential people have been convinced by the IPCC curve fitting argument. Now we know that it was based on inadequate science. All those who were convinced by it, and use it in arguments, should revaluate their position and await a new statement by the IPCC.

Let us hope that the next IPCC report provides much better evidence for the claim that past climate can be explained using natural factors up to a man-made cut-off point. Another ‘sun and volcanoes’ model will not do.