Researchers in France and the United States have found a way to make long-term climate forecasts that not only are valid for more than a year, but are probably better than ever. Presented in the journal Proceedings of the National Academy of Sciences (PNAS), the study shows how climate can be predicted up to 12 months in advance, double the length of time it took previously.
A researcher of the study says climate forecasts are considerably more general than short-term weather forecasts, since they do not predict precise temperatures in particular cities. But Professor Michael Ghil from the Department of Atmospheric and Oceanic Sciences at the University of California, Los Angeles (UCLA) in the United States says they could still influence activities in the agricultural, economic and industrial sectors.
'Certain climate features might be predictable, although not in such detail as the temperature and whether it will rain in Los Angeles on such a day two years from now,' explains Professor Ghil, also from the Environmental Research and Teaching Institute, École Normale Supérieure in France and a senior author of the study. 'These are averages over larger areas and longer time spans.'
The study reveals that long-term forecasts could help scientists make much more advanced predictions of El Niño events, for example. These events are characterised by changes in the temperature of the surface of the tropical eastern Pacific Ocean (warming or cooling known as El Niño and La Niña respectively), and air surface pressure in the tropical western Pacific. These events trigger extreme weather phenomena like flooding and droughts in various regions of the world on an average of every two years.
Keeping natural climate variability separate from human-induced climate change has never been easy. Another challenge for researchers has been to consider natural variability when developing climate models.
For this latest study, Professor Ghil and colleagues evaluated worldwide sea-surface temperatures. In order to give their forecasts a boost, the team created a new algorithm based on fresh insights into the mathematics of how short-term weather interacts with long-term climate. Weather covers a period of days, while climate covers months and longer, according to the researchers.
They used 50 years of climate data and test predictions retrospectively, including climate data from 1950 to 1970, to make 'forecasts' for January 1971, February 1971 and beyond. They also wanted to determine how accurate the predictions were. They succeeded in clinching accurate predictions 16 months beyond what other scientists achieved in half that time.
In a recent development, Professor Ghil and colleagues assessed the macroeconomic impact of extreme events. The results, funded in part by the E2C2 ('Extreme events: causes and consequences') project, which received a EUR 1.5 million grant under the 'New and emerging science and technology' (NEST) Cross-cutting activity of the EU's Sixth Framework Programme (FP6), showed that extreme events pose a bigger problem when catastrophes occur during an economic expansion.
'It is better during a recession,' Professor Ghil says. 'If your roof blows off in a hurricane, it's easier to get somebody to fix your roof when many people are out of work and wages are depressed. This finding is consistent with, and helps explain, reports of the World Bank on the impact of natural catastrophes.'