We suggest that Earth observation should be used to monitor ecosystem services in the run-up to implementation of the United Nations Sustainable Development Goals (SDGs; see also A. K. Skidmore et al. Nature 523, 403–405; 2015).

A reliable Earth-observation framework would provide long-term spatial indicators of ecosystem services. It could capture changes in environmental and socio-economic features, for example by comparison with more than 40 years of Landsat satellite data and with information from new sensors such as the Sentinel fleet.

Earth observation could inform management decisions about how to resolve conflicting objectives that arise from the SDGs (see, for example, R. Bosch et al. Nature 523, 526–527; 2015). It would help in evaluating remote effects (teleconnections) for ecosystem-service provision and usage, such as whether biofuel production in one place creates biodiversity loss, pollution or deforestation elsewhere (see J. Liu et al. Science http://doi.org/627; 2015).

We need new forms of data integration and case-study synthesis. Earth-monitoring systems must be developed with input from environmental and social scientists to link up with existing knowledge, for example by relating ecosystem services to biodiversity.