This article is reproduced with the permission of New Scientist for exclusive use by Nova users.

Global warming: the new battle
13 September 2003
From New Scientist Print Edition.
Jenny Hogan

Enlarge Figure 1

Stop debating whether global warming is happening. The priority now is to start preparing for its consequences, be they droughts and heatwaves or floods and harsh winters. That is the message from meteorologists, who say adapting to climate change will be one of the world's major challenges for the 21st century.

It marks an astonishing shift in how experts view global warming. Confident that they have all but convinced the world that warming is taking place, climate scientists are concentrating on predicting its local effects. The problem, they say, is that so far we have no way of knowing what those effects will be.

Pushing the idea of adapting to climate change is not popular with all climate scientists, and some worry that it could detract from protecting the environment. But Mike Hulme, director of the Tyndall Centre for Climate Change Research in Norwich, UK, is one of those who argue that decisions made now will affect us for decades to come as temperatures rise. Changes to the climate affect all areas of human activity, he points out, from what clothes people are likely to buy next summer to long-term plans for transport, flood defences and agriculture (see Timeline).

Until now, the benchmark of climate change has been the Earth's average temperature. Scientist have geared their models towards forecasting this simple measure, and used the results to shock policy makers into action. The UN's Intergovernmental Panel on Climate Change collects predictions from centres around the world, and in its 2001 report concluded that human activity is likely to cause the average global temperature to rise by between 1.4 °C and 5.8 °C by 2100.

Politicians responded with efforts to curb greenhouse gas emissions and promote renewable energy. And with refinements to the global picture appearing almost every week, climate scientists feel there is now an urgent need to move on. "People are fighting less about whether there is climate change," says Jonathan Gregory of the University of Reading in the UK. "It is now more sensible to think about adapting to it."

When it comes to coping with the changes, global temperatures are essentially irrelevant. "All people are interested in is the regional and local details," says Rowan Sutton of the Centre for Global and Atmospheric Modelling at the University of Reading, UK - and climate science is completely unprepared for this challenge, he warns.

At the Royal Meteorological Society's annual meeting in Norwich last week, Sutton presented his audience with a topical British example. During this summer's heatwave, rail companies had to cancel trains, fearing tracks would buckle as the temperature soared to almost 38 °C. "Should Network Rail invest in upgrades to avoid future heat-induced speed restrictions?" he asked. "Clearly there will be a need for climate information there. What are you going to tell them?"

There was silence. Answering this question would require climate models that can predict peak temperatures in the region of south-east England where the problem occurred. This is not something current models are capable of. Even when answers can be given, the uncertainties are often huge (see "Fuzzy forecasts").

The need for local data is urgent. According to the European Commission, droughts have already caused crop yields to drop across southern Europe (New Scientist, 23 August, p 5). But without predictions of exactly what will happen where, farmers can't prepare for the future.

The floods that devastated central Europe in August 2002 are another example. The industries left to clean up after the river Elbe flooded were eager to blame global warming (New Scientist, 26 April, p 24). Were they right and should we prepare for more of the same? Research published this week that re-examines historical records of the area concludes that summer floods around the river have not increased, while winter floods have become less common (Nature, vol 425, p 166). This fits climate models, which suggest warmer surface temperatures stop the river from freezing in winter, making it less likely to become blocked by breaking ice.

Add social factors to the mix, and the picture changes again. Edmund Penning-Rowsell, head of the Flood Hazard Research Centre at Middlesex University and his colleagues reported the results of a study of flood risk in the UK at the annual meeting of the Royal Geographical Society in London last week. "Our aim was to come up with a national picture for 100 years' time," says Penning-Rowsell. The researchers included factors such as population density, land use and flood defences and found that even though global warming seems to be reducing the risk of severe floods, the risk facing Londoners and people living along the coast of south-east England looks set to rise significantly over the next century. "If we build poor houses in dumb places, the flood risk increases," says Ilan Kelman, part of the research team. The worst-case scenario predicts that an extra million people in the UK will be at high risk by the 2080s.

Detailed local information about climate change will be invaluable for studies like these. But getting it right won't be easy. Because local climate is heavily dependent on what is happening elsewhere, the whole planet has to be included just to look at one particular place.

Running the existing global models at higher resolution would take more computing power than we have. Even the Earth Simulator in Japan, the most powerful supercomputer ever built, struggles to reach a resolution of 10 kilometres. What's more, as the scale gets smaller, "you have to completely re-evaluate the science," says Alan Thorpe, who chaired the Norwich conference. Local models will need to include features such as clouds and mountain ranges, and make different compromises from global models, embedding a high-resolution local model within a lower-resolution global one, say.

For climate scientists, there is a huge task ahead. So far, says Sutton, hardly anyone is working on solving these problems.

From issue 2412 of New Scientist magazine, 13 September 2003, page 6

Fuzzy forecasts One thing climate modellers don't tend to let on is how confident they are in their results. "We just give the forecast, with no idea of how likely or unlikely that outcome is," says Mat Collins from the UK's Hadley Centre for Climate Prediction and Research in Bracknell, Berkshire. But that won't be good enough to base decisions on, he warned colleagues at the Royal Meteorological Society meeting in Norwich last week. Putting error bars on climate forecasts is one of next challenges for the field, says Collins. "There seems to be a body of agreement that this is a real priority." Collins and his team presented the first results from a project that aims to quantify the uncertainty in their climate model. They tweaked various settings, such as the parameters describing clouds or convection, then checked to see whether the new versions of the model matched today's global climate. The researchers took the 53 that did, and ran them forward to predict how temperatures will change if levels of carbon dioxide in the atmosphere double. The different versions predicted temperature increases of anywhere from 1.8 °C up to 8.3 °C - a far wider spread than the Intergovernmental Panel for Climate Change's prediction of 1.5 to 4.5 °C, which is based on results from all the 10 or so climate models run around the world. In the short term at least, says Collins, we will have to accept that climate prediction will become less certain. The drawback of this approach is that running several versions of a model eats up computer time that could be used to make more precise forecasts. But a unique project to be launched this week could help climate modellers do a little more. Climateprediction.net will use idle time on people's computers to calculate how the climate is going to change. Each volunteer runs a slightly different model, and the results are sent back for comparison. Anyone can download the software at www.climateprediction.net and watch the weather patterns on screen as they take part in what the organisers hope will be the largest distributed computing project in the world.

For the latest from New Scientiist visit www.newscientist.com