Stay cool about short-term climate forecasts

An ultra-sensible comment by Jason Palmer at the New Scientist (Environment): What are we to make of headlines like this: "Next decade 'may see no warming'"? That, and others like it, have been inspired by a paper in Nature this week that has taken a new look at climate change model data. Skeptics are loving it – but what does it really mean?

Noel Keenlyside of the Leibniz Institute of Marine Sciences in Kiel, Germany, and colleagues used sea surface temperatures recorded over the last 50 years to constrain other variables in an existing climate model. They validated the approach by "hindcasting", that is, by checking that the model predicted the right conditions over recent history. Then to the forecasting: they reckon that over much of Europe and North America, the coming decade will see cooling, while Pacific regions remain the same temperature.

If it's true, it's a fleeting reprieve, not a commuted sentence; the paper is peppered with the word "temporarily". In about 2020 - and here's the take-home message - the model predicts the temperature curve will begin to rise again, in line with prior predictions like those of the IPCC. Any slowdown in cooling would be a decade-long ripple on an upward curve centuries and millennia long….

This graph of oil prices came from the US Department of Energy (via Wikimedia Commons). Its presence in this article is somewhat mysterious. Maybe it just here to suggest the mysteries of climate modeling.