In pure science, as in art, little is urgent. Gravitational waves were discovered — a triumph for curiosity-driven science — thanks to physicists’ patience and imaginative power. That they had waited decades is irrelevant. Alas, not all science has the luxury of timelessness.

Urgent science touches on issues that rank high on the social agenda. Theorists have classified fields such as climatology and global-change research as post-normal science, in which socio-economic stakes are high and decisions are pressing. That is the case with the agenda of Future Earth, an international sustainability-research platform set up in 2012 to tackle complex social challenges, from climate change to finance.

The scheme replaces a number of narrower programmes, including the international Geosphere-Biosphere and Human Dimensions programmes, which — to the regret of many — are now all closed.

Sustainability researchers will need to follow a multidisciplinary— nay, transdisciplinary — approach that goes beyond what many scientists have been used to. Future Earth’s ‘co-design’ intends natural and social scientists to plan and carry out research with outside experts. Whether that will win over academic researchers, stakeholders and, crucially, funders remains to be seen. To convince sceptics, the scheme needs to provide a successful example of how it will work in practice.

Preservation of the natural commons, such as atmosphere, water, land and oceans, for future generations is vital and a cause to which any responsible scientist will happily subscribe. But combining the conventional scientific methods of the natural and social sciences with knowledge from various other sources — land owners and planners, insurance companies, conservation groups, emergency organizations and political decision-makers — poses conceptual and organizational challenges.

A cross-community Future Earth workshop on adaptation and responses to extreme climate events, held last month in Berlin, offered a taste of such challenges (see go.nature.com/6utfmi) and might serve as a test for the design of research networks on sustainability issues. Under time pressure, participants had to draft a research strategy to address the drivers and implications of extreme events, and make it fit with Future Earth’s conceptual framework — a tough issue. The workshop asked scientists from different academic cultures to do that work, which produced semantic confusion and the odd unhelpful generalization.

But the workshop was not in vain. Many participants (a healthy number of whom were from developing nations) said that they revelled in being pushed out of their comfort zones. They produced several meaty research questions, including some genuinely new ideas for how the social and natural sciences could interact. For example, when do climatic and socio-economic factors combine to amplify the impacts of climate extremes and induce cascading harm? Are there ‘tipping points’ at which social or natural systems might fail to recover from shocks? And how might science-based adaptation work in data-scarce regions?

The ideas found an audience. Representatives of funding agencies at the workshop cautiously indicated that the proposals stand a good chance of getting funded by the Belmont Forum, a worldwide group of 21 major funders of global environmental-change research.

But governments and grant-giving agencies have not yet firmly committed to funding Future Earth as a whole. The reluctance comes from uncertainty over what the scheme might be able to deliver. The closure of successful programmes in favour of something fashionable but conceptually unproven has earned Future Earth sceptical glances. But then, it was launched in response to complaints that previous programmes were not sufficiently linked and that the knowledge they produced was scarcely picked up in practice. There is no lack of studies, for example, on how extreme heat, rain and wind affect farmers, city dwellers and coastlines in many parts of the world. But the results are almost useless if they never make it beyond the pages of academic journals.

Future Earth will need to make sure that scientific evidence comes to the desks of decision-makers, no matter what they might then make of it. But the programme should also avoid retreading familiar ground. The mountain of data from previous programmes, including countless climate-change studies, remains relevant — even if the information hasn’t yet been put to constructive use.

Future sustainability research, no matter how interdisciplinary, should build on that heritage and focus on finding and closing knowledge gaps. In doing so, scientists involved in Future Earth can provide an invaluable service to society. And researchers in niche disciplines — palaeo­climatology or behavioural science, say — who work to fill those gaps will get a welcome chance to put their work into a broader context.

Future Earth might also become a showcase for linking natural and social sciences — a real necessity given that human activity is altering the planet at worrying speed. But sustainability research must not become tied in the straitjacket of conceptualism and utilitarianism. Scientists are not merely service providers. As in any other field of science, sustainability research must remain at its core a curiosity-driven affair.