Forecasting Challenges provide an opportunity to create a community of practice that builds capacity for ecological forecasting by leveraging targeted datasets.
EFI has cyberinfrastructure set up to support stand-alone Forecasting Challenges that allow people to submit forecasts that are checked for alignment with the metadata standards, scored, cataloged, and visualized on a dashboard. There is also shared code available for individuals who want to synthesize across forecasts that have been submitted to those stand-alone Challenges. Reach out to info@ecoforecast.org if you are considering setting up your own Challenge.
See these resources to learn about the workflow and applications of the Forecasting Challenge developed for NEON data targets, the cyberinfrastructure that can be deployed for forecasting challenge using other data targets, and resources to use the challenges in the classroom.
- Thomas, R.Q., C. Boettiger, C.C. Carey, M.C. Dietze, L.R. Johnson, M.A. Kenney, J.S. McLachlan, J.A. Peters, E.R. Sokol, J.F. Weltzin, A. Willson, W.M. Woelmer, and Challenge Contributors (2023) The NEON Ecological Forecasting Challenge. Frontiers in Ecology and the Environment 21 (3): 112–13. https://doi.org/10.1002/fee.2616.
- Thomas, R.Q., and C. Boettiger. 2026. Cyberinfrastructure to Support Ecological Forecasting Challenges. Ecosphere 17(6): e70682. https://doi.org/10.1002/ecs2.70682
- Best practices and resources for using the NEON Ecological Forecasting Challenges in the Classroom
EFI Supported Forecasting Challenges
- EFI NEON Ecological Forecast Challenge:
This project was funded by an NSF Research Coordination Network grant. The Challenge revolves around five theme areas that span aquatic and terrestrial systems, and population, community, and ecosystem processes across a broad range of ecoregions that use National Ecological Observatory Network (NEON) data products collected at NEON sites. The target variables for each theme are:- Aquatic Ecosystems – Daily forecasts of dissolved oxygen, water temperature, and chlorophyll-a from lakes and river/streams NEON sites
- Terrestrial Carbon and Water Fluxes – Half hour and daily forecasts of net ecosystem exchange (NEE) of CO2 and evapotranspiration (LE) at terrestrialNEON sites
- Tick Populations – Weekly forecasts for Amblyomma americanum nymphal tick abundance at 9 NEON sites
- Beetle Communities – Weekly forecasts for beetle abundance and species richness at NEON terrestrial sites
- Phenology – Daily forecasts of plant greenness (gcc) and redness (rcc) at deciduous broadleaf forests, evergreen needleleaf, evergreen broadleaf (tropical), grasslands,tundra, agriculture, and shrubland NEON sites as measured by Phenocams
- See the NEON Forecast Challenge Dashboard here.
- Doñana Biodiversity Forecasting Challenge:
This is the first European EFI Chapter-supported forecasting challenge. It focuses on data streams for Doñana National Park in Spain and is part of the Doñana Digital Twin project, funded by the European Union and the Open Science Clusters’ Action for Research & Society (OSCARS).
The Challenge provides detailed instructions on how to submit forecasts for- Abundance of four shrub species at 18 long-term monitoring plots OR across the entire Doñana area
- Kilometer Abundance Index (KAI) for rabbits
- Abundance of two dung beetles at 6 long-term monitoring plots OR across the entire Doñana area
- EFI-USGS River Chlorophyll Forecasting Challenge:
Co-hosted by the Ecological Forecasting Initiative and U.S. Geological Survey (USGS) Proxies Project, this challenge provides a unique opportunity to forecast data from the USGS. By participating, you’ll sharpen your forecasting skills and contribute to vital research aimed at addressing pressing environmental concerns, such as harmful algal blooms (HABs) and water quality. - Virginia Ecoforecast Reservoir Analysis (VERA) Forecasting Challenge:
This project is hosted by the Virginia Tech Center for Ecosystem Forecasting and is part of the NSF-funded Virginia Reservoirs LTREB program. The Challenge includes 11 focal freshwater variables with over 11 years of historical data from two reservoirs. Target variables include physical (n=3), chemical (n=6), and biological (n=2). There are many additional datasets collected in near-real time at the reservoirs that can also be used for forecasting even though they are not the focal target variables evaluated in the VERA Challenge.
- Wildfire recovery forecast using MODIS leaf area index (LAI):
The Cyberinfrastructure and Statistical Methods working group developed this spatial forecast challenge and rolled it out at the EFI2025 Conference. You can find more information about the project here: https://github.com/eco4cast/modis-lai-forecast/tree/main.
A tutorial for the spatial challenge is available at: https://github.com/eco4cast/modis-lai-forecast/tree/main/tutorials.
And the rendered html is available here.