There is only one thing that is certain about the climate: it changes. But working out just how much it might change in the future is a fiendishly difficult business.

Our knowledge of many important climate processes is incomplete. Descriptions of cloud formation, for example, are approximate at best.

If we get more of one type of cloud, the climate could cool because sunlight will be reflected back into space. On the other hand, more of another type of cloud could raise temperatures by trapping heat trying to escape from the planet.

And clouds are but one part of a highly interconnected system. Ice cover, ocean currents, volcanic aerosols, soil moisture, vegetation - all have an influence on our daily weather and long-term climate. Some feed back on the system to accelerate cooling; others work the other way.

Ups and downs

The media are generally very bad at conveying these complexities. Climate has been presented as a sort of constant. The impression has been created that the Earth has always been a temperate world.

This is far from the case. The Earth has been much hotter and much colder than it is now. The current atmosphere that blankets the planet helps to keep the globally averaged surface temperature up to about 15 degrees Celsius.

But there is geological evidence that this figure has been down to about 7 deg C and possibly as high as 27 deg C. We are currently in a not-so-cold phase of an ice age - The Quaternary Ice Age.

Sea level, too, has changed constantly. It has varied from its present position by hundreds of metres. About 20 thousand years ago, during a colder phase in this ice age, sea level was up to 130 metres lower than at present.

About 100 million years ago, sea levels were about 300 metres higher than at present because of changes in the volume of ocean basins due to movement by the Earth's techtonic plates.

'Noise' in the system

That the world is now warming up is not in itself a surprise. If there is a surprise, it is in the extent and pace of that warming.

A single event - a hurricane, a flood, a drought - cannot and should not be attributed to an enhanced greenhouse effect.

In climatological terms, events assume significance only when their frequency and/or severity over a considerable period of time are clearly different from previous averages.

The issue is whether there is an enhanced "greenhouse signal" in that long-term data. The mainstream of science believes there is.