Proc. Natl Acad. Sci. USA 115, 1424–1432 (2018)

Global environmental change, and perhaps climate change in particular, has fundamentally altered the way we look at the future, with past experience becoming an increasingly poor guide to what is to come. A dynamic future combined with multiple, severe and interacting environmental pressures has begun to move many ecologists from undertaking primarily descriptive work to a focus on understanding and anticipating change in order to inform management interventions.

Michael Dietze from Boston University and co-authors argue that there is currently a temporal mismatch between most ecological forecasts, which are predominantly focused on climate change responses over multi-decadal timescales, and environmental decision-making, which tends to operate on near-term — daily to decadal — timescales. The authors argue that a much greater focus on iterative, near-term ecological forecasts is needed. Not only can these forecasts be more easily operationalized into decision-making, but they can also be continually tested and updated with new observational data. Several key areas are identified for development if near-term forecasting opportunities are to be maximized. Improvements are needed in data interoperability, latency and uncertainty quantification and in forecast-specific theory, methods, and cyber-infrastructure. To facilitate these developments, changes in scientific training, culture, and supporting institutions will also be required.