Natural disasters such as flood, drought, hurricane, landslide, earthquake, tsunami and volcanic eruption are causing increasing casualties and damage to infrastructure and property. In addition to huge human suffering, the economies of many countries, especially developing ones, are severely affected as disaster strikes the region.

What can we do to mitigate disaster? The most important thing is to provide appropriate and timely information for managers, planners and, last but not least, people affected by disaster. Managers and planners need to know what, where, when, how and what-if disaster happens in order to develop fast warning and evacuation plans and disseminate appropriate information for those (re)acting locally on the ground. To answer these questions of ‘what and how', a decision-support system is necessary, not to say prerequisite. Such systems should be built on GIS platforms, since these have onboard mechanisms to collect, manage, analyse, model and display spatio-temporal data. They are also able to support rapid dissemination of disaster information to the public via the internet.

The development of such decision-support systems is not that easy.

A first challenge concerns real-time data collection over larger regions. Although a disaster may come into being at a certain place, it usually spreads rapidly, causing serious loss over a wider area. Instant and continuous prediction of disaster spread and magnitude is necessary to mitigate casualties and damage.

For example, where I am living in the Mekong Delta in Vietnam, gauge stations would be most appropriate for providing reliable and timely data to warn people affected by disaster and for action taking. Some countries have invested considerably in implementing early-warning systems based on networks of gauge stations. Within such networks warning centres combine resulting data with satellite imagery to indentify location and extent of disaster and inform the public for preparatory action. Disasters do not stop at state or provincial borders; data collection should thus transcend administrative boundaries. Monitoring networks should extend throughout individual and among adjacent countries to arrive at better disaster forecasting.

Once collected, data has to be integrated with other existing geographic data, such as base-maps, to be useful for disaster management. Such data is usually distributed across various national agencies. The challenge is to get rapid access to these multiple data sources, and this makes indispensible well-functioning spatial data infrastructures at both national and regional level.