Understanding complexity and globality - the example of climate change

Few scientists today doubt the reality of global warming. And according to the Second Assessment report (1995) of the Intergovernmental Panel on Climate Change (IPCC)1 there is mounting evidence that human activity is responsible for part of the increase. This is mostly due to a steep rise this century of greenhouse gases, like carbon dioxide, in the atmosphere as a result of the motor car, burning fossil fuels, industrial activities and destruction of tropical forest. The gases trap the sun's heat, just like the glass in a greenhouse. But what is proving more difficult is to predict just how much temperature is likely to rise by the end of the next century. This is important to describe impact 'scenarios' and to guide decision-makers in the actions they need to take. Current IPCC estimates put the increase at between 1.5° C and 3.5° C by 2100, if, as predicted, current practices double carbon dioxide emissions.

To make long-term climate predictions more accurate, scientists need data, complex climate forecasting models and fast computers to run them. But it has only been in the last fifteen years that these resources have started to be available. And this is partly due to a massive mobilisation of experts world-wide in a series of joint climate observation projects.

Ocean-atmosphere interactions control climate
Just as a shepherd looks at the evening sky to predict if it will rain the next day, climate forecasters must study the oceans - or rather the ocean-atmosphere interactions - in order to look months, even decades ahead. The oceans transport vast quantities of heat around the globe. Water at the tropics and equator warmed by the sun moves towards the poles where it cools, sinks and returns. Deep underwater, slow-moving 'thermohaline' currents distribute heat around the globe, taking up to 1000 years to complete a cycle.

At the same time, the oceans and atmosphere interact, with winds pushing surface water into currents, while the great inertia of the oceans influences longer-term atmospheric patterns. The best-known example of this is El Niño, the erratically recurring reversal of winds and surface currents in the tropical Pacific that not only has dramatic local effects in Peru or Australia - but by 'teleconnection' via atmospheric jet-streams, can cause floods in Kenya and droughts in South Africa.

Predicting weather: the example of El Niño
While El Niño is not evidence of climate warming, its increased frequency and intensity might be. The two strongest El Niño events on record were in 1982/83 and 1997/98. And prediction of El Niño would be a powerful test of the complex computer models needed to look at climate patterns more than a few days or weeks ahead. Peruvian fishermen have known the disruptive effects of the Christmas El Niño on anchovy populations for hundreds of years. But they did not know when it was coming until it was too late. And when the 1982/83 El Niño caused an estimated $13 billion of damage world-wide and over 2000 deaths, the international science community found their governments keen to back research. If they could see an El Niño coming six months or a year ahead, farmers could plant crops resistant to drought or to high rainfall, while governments could prepare for severe weather, food and water shortages.

Scientists faced some major obstacles - especially a lack of data and an incomplete understanding of what drives world climate. The main problem is one of scale. We can currently model weather disturbances where the elements are fairly large, like land masses and ocean hundreds of kilometres across. But to get down to a resolution of even 50 km needs computing power we do not yet have. Meanwhile, scientists know that even minute cloud particles can influence weather. But to incorporate these in models will need computers over 1000 times more powerful than today's supercomputers. With so many variables, the mathematical models can only be approximations, averaging ocean and atmospheric values over areas about 300 km across. But to obtain even this resolution needs a dense array of observations.

Getting the data
In January 1985, under the auspices of ICSU, WMO and UNESCO's Intergovernmental Oceanographic Commission (IOC), an 18-nation Tropical Ocean and Global Atmosphere Program (TOGA) started a decade of observations. At the heart of the project was a planned array of about 70 buoys - the Tropical Atmosphere Ocean Array (TAO) - straddling the equatorial Pacific. These now send real-time measures of sea surface temperature, surface wind and salinity via satellites, to research centres that use the data in climate models to make forecasts. The TAO data are supplemented by expendable drifting buoys dropped from aeroplanes, by radar altimeter measurements of sea surface level from the Franco-American Topex/Poseidon satellite, and by sea surface temperature measurements made from satellite infra-red sensors.

Testing the models
Using the continually updated data from Topex/Poseidon and the TAO moored buoys, scientists were able to predict El Niño in 1997/98. Early signs came in March 1997 when the TAO buoys detected an anomalous sub-surface cooling in the western tropical Pacific and a surface warming to the east. It was then possible to chart the build-up of the anomaly on a day-by-day basis. The US National Oceanic and Atmospheric Administration (NOAA) and the Topex/Poseidon websites published updated images and forecasts almost daily, while the US National Climate Centre issued El Niño warnings.

When the predicted floods in Chile started to happen in August-October, while forest fires burned out of control in Indonesia, it became obvious that the investment in research had paid off. Using the Internet warnings passed on by government offices in some vulnerable tropical countries, farmers and fishermen were better prepared and many potential catastrophes were offset. The success of the TOGA experiment in the tropics has given new impetus to a planned global program of observations. In response to Agenda 21 a joint initiative of IOC, WMO, UNEP and ICSU has set up a Global Ocean Observing System (GOOS) - a unified, co-operative ocean monitoring network implemented by nations. With parallel global systems also set up to observe the atmosphere and the land, scientists will begin at last to have the data they need to monitor and forecast the impacts of climate change in the next century.

1 a panel of 2000 experts co-sponsored by the United Nations Environment Programme and the World Meteorological Organization