What is the PoPS Forecasting Platform?
The PoPS model is just one piece of the PoPS Forecasting Platform. Users may download the R or GRASS GIS version of PoPS OR users can use the PoPS Forecasting Platform to run PoPS in the cloud (available for public use soon).
The PoPS Forecasting Platform features an easy to use PoPS Dashboard Interface, the PoPS Database, and the PoPS model running in the cloud. With the PoPS Dashboard Interface, users can run PoPS, compare scenarios, and save data without any programming experience.
How is the PoPS Forecasting Platform structured?
The PoPS Forecasting Platform consists of a series of interconnected open-source packages (available on GitHub) and tools shown below.
End-users may download the R or GRASS GIS packages to run PoPS on their own computer, or they may use the interactive PoPS Dashboard Interface (coming soon).
How did we create the PoPS Forecasting Platform?
We started PoPS in 2018 as a collaboration between the Center for Geospatial Analytics at North Carolina State University and the USDA Animal and Plant Health Inspection Service (APHIS).
PoPS was developed, and continues to be refined, through what we call an “iterative modeling cycle”, a process that we suggest may help guide ecological forecasters in other contexts. Meeting with the stakeholders who will use the forecasts comes first, to discuss their needs and set mutual objectives. After initial data are gathered to feed into the model, four revisionary loops follow:
(1) The Calibration Loop occurs anytime new occurrence data are acquired or new biological information about the pest or pathogen is discovered; new data require that model parameters be re-calibrated, validated, and updated in the database.
(2) The Scenario Modeling Loop involves stakeholders defining a management scenario that they want to test together, using a fully calibrated and validated version of the forecast model, to experiment with strategies that they can then implement on the ground as part of real-world adaptive management; the process repeats as stakeholders compare strategies possible under a realistic monetary budget and decide on the optimal outcome.
(3) The Field Observation and Scientific Feedback Loop is engaged when stakeholders use the forecast to determine management and surveillance priorities; these new surveys and management actions are recorded in the database, triggering another iteration of the Calibration Loop. The Field Observation and Scientific Feedback Loop can also be triggered when new scientific studies reveal information about species characteristics (e.g., environmental tolerances or host preferences); this new information is added to the database and can lead to new insights that change model assumptions; such a change triggers another iteration of the Calibration Loop, and the results of forecasts (and hindcasts) with and without the new knowledge are compared.
(4) The Participatory Feedback Loop consists of iterative back-and-forth discussion with stakeholders, to ensure that research progress matches their needs and vision; stakeholders test and provide input on not only the forecast model but also ways in which they prefer to interact with it.