5-D Weather Forecasting

The emergence of extreme-endurance aerospace platforms, with flight endurance greater than 90 days, places an unprecedented demand on atmospheric forecasting. These platforms are highly sensitive to wind and weather conditions and mission planners are required to operate them within tight requirements. Due to their slow airspeeds these platforms cannot adapt quickly to errors in the weather forecast, making them highly dependent on the accuracy of long-term weather forecasts for both mission utility and safety. Errors in weather forecasts can cause these platforms to be positioned poorly and unable to station keep underneath clouds or thunderstorms. Although it is effectively impossible to accurately predict weather out to 90 days, the possibility exists to create a forecasting system that can provide a spectrum of reasonable weather scenarios that have a high probability of capturing the true weather environment far into the future within a band of uncertainty. The Skydweller team proposes to build AI-driven five-dimensional (5D) weather forecasting tools that model weather across latitude, longitude, altitude, time, and uncertainty. The fifth dimension of uncertainty is a critical innovation for extreme-endurance platforms. This capability will allow mission planners to run flight simulations against all possible weather scenarios far into the future, within a forecasted range of uncertainty, and then adapt the flight plan of the air vehicle to maximize the probability of mission success and flight safety.