Storm in a test tube

Natural disasters claim half a million lives every decade. Julian Hunt and Brian Lee explain how science saves lives.

ABBC TV presenter covering Hurricane Floyd's arrival on the United States east coast in September asked the spokeswoman from the National Hurricane Center in Miami, "Aren't these hurricanes totally unpredictable?" "No," she replied, "as a matter of fact, this hurricane has been forecast pretty accurately."

She might have added that the error for the location of such a hurricane 24 hours ahead is now 140km, while in 1992 it was more likely to be 210km. She did not add either that the UK Meteorological Office forecasts compiled in conjunction with the City University of Hong Kong for the Miami area are often the most accurate.

This is one example of the advances in science, technology and engineering in international disasters that have been achieved over the past decade. This era has been designated by the United Nations as International Decade for Natural Disaster Reduction (IDNDR).

Although this period has seen an increased government focus on the problems of climate change, there has also been more international public awareness of the threat from natural disasters. The human suffering involved is estimated at 500,000 lives lost per decade, and the worldwide economic costs are estimated at $1billion per day.

Many of these issues will be discussed at an international meeting, "Dealing with natural disasters - achievements and new challenges in science, technology and engineering", to be held at the Royal Society in London from October -29.

This is being organised jointly by the Royal Society and the Royal Academy of Engineering as a finale to UK work on IDNDR.

Research into natural disasters has typically been conducted by specialists, many of whom have a narrow focus to their work. Nevertheless, the international bodies bringing together expert knowledge have made great progress in drawing groups together and in establishing general conclusions, listing information about different types of disaster and developing general guidelines and techniques for better risk assessment and forecasting.

There has also been extensive research by social scientists on the social, economic and political effects of disasters, and of response at national, regional and local community levels. The Department for International Development's funding of the UK Flagship Programme on Forecasts and Warnings insisted that the social science aspects be an integral part of the programme.

We have a new understanding of how warnings should be issued and how communities such as those in cyclone-prone Mauritius and volcano-prone Montserrat act on them. One finding is that continued warnings will be heeded only if they are known to be accurate.

The ultimate interdisciplinary problem facing us today is how to reduce the impact of natural disasters on the growing urban areas in which more than half the world lives. Wealthy societies can effectively prevent the major impacts of natural disaster on their infrastructure by massive investment, but for disaster-prone developing countries the first goal is simply to ensure that people survive.

The need now is for adequate risk analysison major infrastructure elements so that these are protected against the onslaught of natural hazards. Intermediate technology solutions often strengthen traditionally designed housing in many developing nations and are also essential for the survival of infrastructure on which a struggling population may be totally dependent in the days following a natural disaster.

Finally, are there economic and administrative solutions to the complex problems of natural disaster reduction?

The debate is just beginning but hopefully the world will take heart from the successes of the IDNDR. Already it is evident that the UN intends to continue this work through the establishment of a new inter-agency task force.

Natural disasters claim half a million lives every decade. Julian Hunt and Brian Lee explain how science saves lives Julian Hunt is in the department of space and climate, physics and geological sciences at University College London. Brian Lee is in the department of civil engineering at the University of Portsmouth. Both are members of the UK national coordination committee for the International Decade for Natural Disaster Reduction.

HOW EXPERTS HAVE TAMED NATURAL EVENTS

Volcanoes Because volcanoes exemplify the specific nature of a disaster phenomenon, it was decided that during the IDNDR there would be an intensive study of a certain number of "decade volcanoes", whose list includes Vesuvius, Mount Pinatubo in the Philippines, and Soufriere Hills in Montserrat.

At Mount Pinatubo in 1991, an array of measuring instruments was able to detect seismic tremors and give warning of the main eruption.

Effective international cooperation between volcanological, meteorological and civil aviation organisations ensures that information about volcanic eruption is passed to designated meteorological centres to calculate the movement of volcanic dust to warn jet aircraft of dangers.

Earthquakes Intrigued by the claims of a few scientists that earthquakes can be predicted, a special meeting was held at the Royal Society in 1995. The sceptics were not convinced - and this is why the risk of earthquakes continues to be assessed by the laws of probability.

The great conceptual breakthrough in understanding why earthquakes occur in relatively narrow bands on the earth's surface came in the 1960s with the realisation that these bands are where the earth's tectonic plates grind against one another.

Today, minuscule movements of these tectonic plates can be seen from satellites using interference measurements.

Hurricanes and storms Tropical cyclones, with their secondary effects of heavy rainfall and flooding, cause more deaths and economic loss than any other weather-related natural disaster.

A three-day forecast is now as accurate as a one-day forecast was 20 years ago.

The main qualitative change over the past decade has been the improvement in the monthly to seasonal forecasts of meteorological extremes. Over such periods variations in ocean temperature strongly affect both atmospheric extremes.

Studies in the United States demonstrate how improved measurements of ocean temperature from satellite data and improved modelling of ocean currents show why some hurricanes, such as Mitch in 1998, rapidly increase in strength by suddenly moving over warm water.

Work at University College London supports such investigations and is providing greater reliability in forecasts.

Flooding In Europe, inland floods and tidal surges produce the greatest economic loss from natural disasters.

These events are determined by levels of precipitation and wind strength.

Theoretical and computational modelling could help predict flows, surges and wave actions that affect water levels.

Big improvements have been made in predicting coastal surges, particularly where wind and tide combine.