Advancing Weather Prediction for Defense Applications
CEE Distinguished Professor Auroop Ganguly is the lead PI for a $1,000,000 grant from the Air Force Weather titled “Weather Ensemble Analytics and Visualization Environment (WEAVE)”. Michael Correll at the NU Roux Institute is the Co-PI. The grant can be credited to Northeastern University’s Institute for Experiential AI (EAI), Roux Institute, and the College of Engineering, and is routed via the Kostas Research Institute. WEAVE focuses on characterizing and visualizing uncertainties from weather model ensembles to support the missions of the US Air Force, the US Army, and other branches of the US Department of Defense.
The Weather Ensemble Analytics and Visualization Environment (WEAVE) project is designed to analyze and visualize weather model ensembles for understanding and characterizing uncertainty in a way that helps military meteorologists, subject matter experts, and stakeholders in ensuring mission readiness through effective decisions. The analysis approaches will include data-driven sciences, such as statistics and Artificial Intelligence, and the visualization approaches will develop methods for uncertainty communication and user-centric analytics. The analysis will inform the visualization, while both will be informed by US Army and DoD stakeholder needs.
Weather forecasting is critical for mission readiness and effectiveness across all branches of the US Department of Defense (DoD), for example, while the Army and Marines depend on forecasts over land for tactical missions, the Air Force and Navy depend on predictions of atmospheric and oceanic conditions for operational missions. The statistics of weather extremes ranging from sub-seasonal, seasonal, interannual, decadal, to multi-decadal scales, are critical for ensuring the resilience of U.S. Army, Air Force, and Naval bases across the globe. Uncertainties in weather predictions, whether resulting from knowledge gaps or intrinsic variability or resolution challenges, which must consider different probability structures of weather variables and weather associated indices of interest to stakeholders, and which may need to be generated based in interfaces between machine-analytics and human-in-the-loop, are important to characterize the ranges of plausible behavior, and hence critical for mission effectiveness across all branches of the US DoD.