Eager: harnessing the power of AI edge computing for near real-time flux tower analysis

The National Science Foundation Division of Biological Infrastructure awarded West Virginia University Research Corporation $294,580 on September 1, 2026, under the Biological Sciences program (CFDA 47.074) to develop artificial intelligence and edge computing technologies for near real-time analysis of ecosystem carbon, water, and energy fluxes using data from the National Ecological Observatory Network (NEON).

The project will develop AI-based approaches for computing and quality-controlling ecosystem fluxes on distributed edge devices, validate results against NEON's standard processing pipeline, and deploy FluxMapper, an AI-enabled framework that spatially attributes ecosystem fluxes and produces dynamic maps using thin-edge computing. The recipient will provide open access to these data products through interactive visualizations and a large language model interface enabling novices and experts to query and analyze flux data. The technologies will enable scientists to detect short-lived ecological events and access observations in near real time rather than with months or years of delay, improving the quality and timeliness of environmental observations and broadening access to ecosystem information for researchers, students, and other stakeholders.

Performance occurs in Morgantown, West Virginia, with a period of performance from September 1, 2026, through August 31, 2028. The project integrates advanced AI methods with NSF's SAGE GRANDE edge-computing cyberinfrastructure to establish a framework for near real-time eddy covariance flux processing and analysis.

Abstract

This project will develop new technologies that enable scientists to observe how various ecosystems exchange carbon, water, and energy with the atmosphere in near real time. These exchanges play a critical role in regulating ecosystem health, water resources, and other crucial services, yet current processing methods often delay access to these data by months or years. By combining artificial intelligence, edge computing, and the national ecological observatory network (neon), this project will create an open, scalable system that rapidly processes and visualizes ecosystem fluxes as they occur. The resulting tools will help researchers detect important short-lived ecological events, improve the quality and timeliness of environmental observations, and make complex scientific data more accessible through intuitive visualizations and AI-assisted data exploration. The project will advance the nation's scientific cyberinfrastructure, support environmental research and education, improve operational data quality, and broaden access to ecosystem information for researchers, students, and other stakeholders. This eager project will integrate advanced AI methods with NSF's sage grande edge-computing cyberinfrastructure to establish a framework for near real-time eddy covariance flux processing and analysis. The project will; (1) develop and evaluate AI-based approaches for computing and quality-controlling ecosystem carbon, water, and energy fluxes on distributed edge devices, with validation against neon's standard processing pipeline; (2) deploy fluxmapper, an AI-enabled framework that spatially attributes ecosystem fluxes and produces dynamic maps of ecosystem fluxes using thin-edge computing; and (3) provide open access to these data products through interactive visualizations and a large language model interface that enables both novices and experts to query and analyze flux data. Together, these innovations will establish a scalable cyberinfrastructure for real-time ecological intelligence, enabling more rapid detection of ecological change, improved data quality assurance, and new investigations into the spatial and temporal drivers of ecosystem processes, particularly during extreme events and other transient periods of biological and hydrological activity. This project aligns with the NSF priorities of supporting research in epscor jurisdictions and artificial intelligence (AI). This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the foundation's intellectual merit and broader impacts review criteria.- subawards are not planned for this award.