Last week, ministers from some 60 nations gathered in Brussels to create an integrated Earth observation system, the Global Earth Observation System of Systems (GEOSS). December's tsunami in the Indian Ocean has catapulted GEOSS from relative obscurity to high on the international political agenda. This was clear from the presence of Carlos Gutierrez, the US commerce secretary, on his first overseas visit since being sworn in on 7 February, as well as science ministers from around the planet (see page 789).

The tsunami disaster highlighted the power of Earth observation data, but it has also thrown a harsh spotlight on the patchiness and rudimentary nature of current systems for understanding complex Earth systems and applying that knowledge to agriculture, management of water resources, early-warning systems for natural disasters, and more.

Take ocean currents, which affect climate by shifting large volumes of warm and cold water around the planet. A United Nations body set up in 1991 to observe, model and analyse the world's oceans, the Global Ocean Observing System (GOOS), has been chronically underfunded and has installed barely half of the monitoring instruments envisaged. Similar inadequacies undermine the Global Climate Observing System (GCOS) set up in 1992.

Speak to people working in global networks in almost any area of Earth observation and the message is the same: behind the stunning images and model simulations of planet Earth lies a much more disconnected picture. Countries and agencies tend to pursue their own agendas, resulting in duplication and a lack of sharing, and coverage is disproportionately concentrated in rich countries. Data come in overly diverse formats and units, making them difficult to use. Researchers, particularly those outside the Earth observation community, complain of costs, delays and other obstacles to getting the sorts of data they need.

Today's climate observation system is cobbled together from data from research satellites, weather satellites, atmospheric sounders and whatever ground-based observation stations scientists can get their hands on, rather than being tailored to monitoring and understanding climate change and variability. Research satellites last only a few years and are not replaced immediately, if at all. But for reliable climate-change monitoring over decades, it is essential, for example, to launch a follow-up satellite while its predecessor is still operating, so instruments on both can be cross-calibrated.

GEOSS is key to addressing such shortcomings. But researchers and other user communities should ensure that their needs are heard. Better international coordination promises to make better use of the billions of dollars that are already spent on Earth observation. But there is a limit to the benefits that can be squeezed out of coordinating and networking activities, when support for the basic scientific activity of collecting critical observations is neglected. Ultimately, GEOSS must make the case for, and oversee, an upgrading of systems such as GCOS and GOOS.

An optimistic view is that the political momentum to treat Earth observation more like global ‘big science’ facilities will translate into a better understanding and support of key scientific needs. The decision to house GEOSS within the World Meteorological Organization may bode well in this respect, as this Geneva-based UN agency has a good track record in mounting international operational weather systems.

But as the tsunami fades from memory, there is also a risk that the new-found political awareness will also subside. As GEOSS charts out what exactly it will do, scientists should actively engage with current political will by pressing home compelling arguments as to how better understanding of the spatial and temporal variability of the wide range of Earth-system parameters will result in progress.