Skip to content

It is often interesting to see the reaction that some papers have. The recent paper by Huber and Knutti in Nature Geoscience that deals with the attribution of observed recent global warming to its cause has been hailed by some to be almost iron-clad proof that humans are responsible for practically all of the warming observed in the past 110 years. Contrast this with another recent paper suggesting that climate sensitivity to increasing CO2 might be less than was thought, that has been described as not the last word on the matter.

Huber and Knutti deal with simulations and projections. Often with climate simulations the proof of their effectiveness is not so much in their ability to hindcast effectively, but rather to make forecasts accurately. I feel that running a climate simulation using data already obtained is one thing, using it to predict what new data will show is another. Often hindcasts tend to be an exercise in what was once called in another context ‘saving the appearances.’ It’s all too easy to fit parameters to past curves. The more parameters one has to modify the more careful one should be.

Using a computer model (the Bern 2.5D intermediate complexity thermohaline ocean model) Huber and Knutti ran thousands of simulations using different parameter variations, keeping those that fitted the observations. They considered estimates of the global increase in heat since 1850, and used it as a limit on the contributions of estimated radiative forcings. They then determined how much of the increase went into the oceans and how much is lost to outgoing radiation. Considering timeframes starting in 1850 and 1950 they conclude that for both starting periods anthropogenic greenhouse gas emissions are the dominant cause of global warming.

The study has been widely misunderstood in the media, who have said that it determines that three quarters of the temperature rise since 1950 is due to mankind, with 25 per cent due to nature. Some have reported that 75 % of the warming seen since the start of the 20th century is due to mankind. This is not strictly what the paper says. It actually concludes that less than one per cent of the warming since 1950 could be due to non-anthropogenic factors, such as ocean cycles, volcanoes and solar influences.

This is not what the IPCC concludes which says that most of the warming seen in the past 50 years is due to humans. Before that it was due to the sun mediated by volcanoes. Like the IPCC the authors suggest that aerosols have reduced the warming that would have taken place by about half.

However, not all the climate forcings and feedbacks are well enough known to be well represented in the climate models. Also, we do not understand the full effect, beyond that of simply blocking sunlight, which aerosols might have.

In addition, the temperature standstill seen in the past decade seems incompatible with their predictions. The ocean temperatures are already below the Bern 2.5D climate model predictions. Simulations of future temperature trends using this climate model performed around the turn of the century have had a tendency to overestimate the warming. Indeed, the authors remark on the near constant ocean temperature seen over the past 5 years is not well simulated by the model, and that the cause of this standstill is unclear.

Looking at the graph presented in the paper. The temperature should have already been a record, which it has not. Click on figures to enlarge.


So whilst this interesting work is not the last word in identifying the human signature in climate change it does make predictions for the future. The researchers predict a sharp rise in annual average global temperatures starting almost immediately. Looking at their curves the rise should be somewhat larger than the ENSO and La Nina variations we have seen in the past decade, so it should quick and obvious if it takes place or not.