OOS 15-7 - Operationalizing ecological forecasts: Two case studies in moving from research to production

Tuesday, August 8, 2017: 3:40 PM
Portland Blrm 256, Oregon Convention Center
Emily Read1, Megan Hines1, Heather Schreppel2, Hilary Stockdon3, Jordan I. Walker1 and Jake F. Weltzin4, (1)Office of Water Information, USGS, (2)Cherokee Nation Technologies, (3)St. Petersburg Coastal & Marine Science Center, USGS, (4)National Coordinating Office, USA National Phenology Network, Tucson, AZ

Ecological forecasts often originate as scientific research products, such as peer-reviewed publications and static datasets. But, to fully realize the societal value of ecological forecasts, they should be operationalized: scientifically vetted, reliably delivered to users in accessible formats and relevant timeframes, and used for decision making. In most cases, the process of operationalizing ecological forecasts will require sustained investment in infrastructure, information technology (IT), software engineering, web design and stakeholder communication. Here, we describe the transition of two forecasting systems from applied research towards operations: the USA National Phenology Network (USA-NPN; https://www.usanpn.org/) and the Coastal Change Hazards Portal (CCH Portal, https://marine.usgs.gov/coastalchangehazardsportal/).

The USA-NPN provides national short-term (6 day) forecasts of accumulated temperature (Accumulated Growing Degree Days) and an associated plant leaf-out threshold model (Extended Spring Indices) that drive phenological, or cyclical, events at local scales. The CCH Portal provides real-time forecasts of the probability of coastal change - beach erosion, dune overwash and coastal inundation – during storm landfall on the Gulf or Atlantic Coasts. Both systems’ forecasts are updated in near-real time at frequencies appropriate for forecasted processes (daily for USA-NPN products and twice daily for CCH Portal landfalling-storm forecasts).


The USA-NPN and CCH Portal have matured along the research-to-operations continuum, and continue to evolve. In both cases, communications with key stakeholders, investment in infrastructure, IT, software engineering, and web design began several years prior to the forecasting system’s release.

Lesson learned:

  • Co-production of systems by scientists, software engineers, and stakeholders is critical; an adaptive management framework can be used to improve tools over time.

  • Sustained support for operations and maintenance by professional software developers is necessary.

  • Web design should be intentional and evidence-based so that users get easy access to information products.

  • Forecasted output should be provided in a variety of formats (e.g., web interface, web services, and other modes of communication) and include ample documentation and training materials for all anticipated user types.