Reduced-Order Modeling of Stratospheric Winds and Its Application in High-Altitude Balloon Trajectory Simulations

The knowledge of weather conditions at the stratosphere is important for the planning and execution of high-altitude balloon flights, which require an accurate modeling of weather data over a period of time. Various methods based on statistical analysis, artificial neural networks, and cluster analysis have been employed to model the temporal variation of weather parameters. In the present study, a proper orthogonal decomposition (POD) method has been used to study the spatial as well as temporal variations of wind data in Singapore. The use of POD facilitates a compact representation of the weather dataset and aids in faster computation of wind profiles for use in balloon trajectory simulation. Further, the results reveal the existence of the quasi-biennial oscillation phenomenon, which is characteristic of equatorial easterly–westerly winds. This phenomenon enables the development of a Fourier prediction model, which can be used in real-time balloon trajectory simulations. The Fourier model is observed to be sensitive to wind velocity fluctuations, especially in the vicinity of alternating wind directions. However, it provides a reasonable projection of balloon trajectory, which can be used in preliminary planning and testing of high-altitude flights. Thus, a prior knowledge of wind profiles based on POD or a Fourier model aids in balloon station keeping. A simple case of altitude-controlled balloon flight is presented, and the results highlight the advantages of the present method in balloon station keeping.