Synthetic Weather Analogs for DoD Exercises, Training, and Simulation

Weather has a profound, cross-domain effect on Department of Defense (DoD) operations. Besides severe weather events such as large hail, high winds, or heavy rains that can halt operations and damage assets, even simple weather details such as wind speed and direction impact plans. To maintain effective operations, personnel must be trained to handle a variety of weather events. Currently, training event planners cannot generate typical weather that is realistic and consistent, not to mention rarer events. To fill this gap, the Air Force needs tailored synthetic weather data that can be disseminated across domains to support live, virtual, and/or constructive training exercises. These training exercises prepare personnel as part of the National Defense Strategy (NDS), supporting efforts such as the Joint All?Domain Command and Control (JADC2), Advanced Battle Management System (ABMS), and Agile Combat Employment (ACE). However, existing synthetic weather generation capabilities do not provide realistic weather conditions, leading to poor preparedness. Thus, for effective training, the U.S. Air Force (USAF) needs a system that generates realistic synthetic weather based on user specifications. Creare is developing the Realistic Environmental Impacts (using Generative Neural Networks) Map (REIGNNMap) system. REIGNNMap gives users an intuitive graphical interface to specify desired weather conditions and then generates similar, realistic weather scenarios. It is trained on location?specific archival weather data to generate an ensemble of realistic weather starting from the user specifications. By learning the statistics of underlying weather patterns, REIGNNMap extends the variation of weather conditions beyond what’s captured in the weather archive while maintaining realism. REIGNNMap enables commanders to realize their intent for training exercises, ensuring DoD personnel are prepared for any weather event.