AI/ML communities across NOAA
Artificial Intelligence (AI) and Machine Learning (ML) efforts are underway across NOAA. As NCAI develops, this will be a space for those Communities of Practice to provide a vehicle for discovery and networking.
Featured NOAA AI Research
Each month the NCAI Newsletter features AI-related NOAA research from our community members. The rotator below highlights research from the current and previous newsletters. Subscribe to the NCAI Newsletter offsite link.
GOBAI: Gridded Ocean Biogeochemistry from Artificial Intelligence
GOBAI-O2 is a gridded data product that provides three-dimensional monthly fields of dissolved oxygen in the global ocean. It was constructed by training machine learning algorithms with observations of oxygen concentration ([O2]) from discrete shipboard measurements and autonomous sensors on biogeochemical Argo floats, then applying those machine learning algorithms to three-dimensional monthly gridded fields of temperature and salinity.
The algorithms used to produce GOBAI-O2 have been validated using real observations and synthetic data from model output, and the data product itself has been compared against the World Ocean Atlas and selected discrete measurements.
Results of these validation and comparison exercises are detailed in Sharp et al. (2023) offsite link.
More Accurate Offshore Wind Profile Projections Using AI/ML in Support of Offshore Wind Energy Development
The NOAA Centers for Environmental Information (NCEI) and NOAA Center for Artificial Intelligence (NCAI), with partial support by NESDIS Office of Common Service (OCS), applied Artificial Intelligence and Machine Learning for significant improvements in projecting sea surface winds (at 10 meters height) into wind turbine height winds up to 200 meters above sea surface with 20 m vertical resolution.
The NOAA/NCEI Blended Seawinds (NBS v2.0) is generated by blending together observations from multiple satellites, and then a Random Forest Regression model is used to extrapolate the blended and gridded sea surface (10 meters) winds to high level wind turbine heights up to 200 meters.
Compared to Lidar observational data, this AI/ML based method significantly outperforms traditional log and power law profile methods (purple and red lines). This AI/ML based model shows best results when compared with independent offshore lidar stations data and NREL’s high resolution profiles.
NOAA tests next-generation wildfire detection and warning tool
Two experimental tools that will speed fire detection and warning got a week-long test run in NOAA’s new Fire Weather Testbed offsite link in June during a series of hands-on simulations with National Weather Service fire weather forecasters, state wildfire managers, researchers, and social scientists.
The first, NOAA’s Next Generation Fire System, or NGFS, uses artificial intelligence to rapidly and autonomously identify fires from observations made by geostationary satellites. By quickly communicating information to forecasters and land managers, it reduces response time when a swift initial attack is most critical.
NGFS, which was developed by NOAA Satellites, uses artificial intelligence to rapidly and autonomously identify fires from observations collected by NOAA’s geostationary satellites offsite link. By quickly communicating information to forecasters and land managers, the new system reduces response time when a swift initial attack is most critical.
The second, the application of the Integrated Warning Team paradigm to wildfire, speeds the exchange of information between meteorologists and land managers and fire agencies to issue fire warnings through NWS channels for rapidly spreading wildfires threatening life and property.
“The tools and concepts we evaluated will turn our meteorologists into scientific first responders in partnership with fire, land and emergency agencies,” said Todd Lindley, the NWS Science and Operations Officer for Norman, Oklahoma, who spearheaded the adaptation of the Integrated Warning Team concept to fire warnings. “These tools will help us provide timely and life-saving warnings of particularly dangerous wildfires.”
Heightened health risks due to wildfire pollutants: Results from an AI modeling study
The Climate Program Office’s Atmospheric Chemistry, Carbon Cycle and Climate (AC4) Program supported new research examining wildfire impacts on air quality and public health in the continental USA from 2000 to 2020. This work is jointly funded by OAR and NESDIS and executed through the NOAA Atmospheric Composition from Space (NACS) team. Supported researchers, Jun Wang from the University of Iowa and Susan Anenberg from George Washington University, collaborated with NOAA scientist Shobha Kondragunta and a team from NASA, and four other academic institutions to examine increasing wildfires and associated health risks in the western U.S. This work, published in The Lancet Planetary Health, contributes to an AC4 initiative to understand long-term trends in atmospheric composition and ultimately help plan for and respond to impacts.
The researchers estimated fine particles (PM2.5) and highly toxic black carbon pollutants using a deep learning model, or a method of artificial intelligence that teaches computers to process data in a way inspired by the human brain. After overall improvements in air quality and a decrease in premature deaths related to PM2.5 and black carbon until 2010, the western U.S. experienced a concerning reversal. Since 2010, there has been a 55% increase in PM2.5, an 86% increase in black carbon, and 670 more premature deaths annually in this region. The study attributes this shift to the escalating frequency and intensity of wildfires. Notably, 100% of populated areas in the USA experienced PM2.5 pollution exceeding guidelines on at least one occasion, with recent wildfires greatly exacerbating exposure risks in western regions. These findings stress the importance of effective wildfire management policies alongside climate mitigation efforts to safeguard air quality and public health.
Researchers Develop Drone-based System to Detect Marine Debris, Expedite Clean Up
NOAA’s National Centers for Coastal Ocean Science (NCCOS), Oregon State University, and their partners are developing a drone-based, machine-learning system to detect and identify marine debris along the coast. In December 2021, the researchers used beaches near Corpus Christi, Texas, to evaluate devices for the system and refine detection methods.
Marine debris, also known as marine litter, is a global problem that threatens the environment, navigation safety, coastal economies, and, potentially, human health. Detecting and identifying debris quickly and accurately is key to cleanup and response actions that can prevent these impacts. Unmanned Aerial Systems, or drones, offer this capability.
Polarized light reflected from human-made objects often differs from natural objects, such as vegetation, soil, and rocks. Installing a polarimetric camera on a drone could improve debris detection from the air. The researchers tested such a camera, both on the ground, and, with the help of the U.S. Coast Guard, from a helicopter.
The helicopter flight allowed the team to mimic the height at which the drone system would be flown and simulate what would happen if the drone used a polarimetric camera. Next, the team trained a machine-learning computer program to find and classify the debris in the imagery collected.
Once fully operational, data collected by the drone-based, machine-learning system will be used to make maps that show where marine debris is concentrated along the coast to guide rapid response and removal efforts.
The researchers will provide NOAA Marine Debris Program staff with training in the use of the new system, along with a standard operating procedures manual. The project is a collaboration among NCCOS, NOAA’s Marine Debris Program, Oregon State University, ORBTL.AI offsite link, and Genwest Systems, Inc.