Atmospheric profiling by GPS Radio Occultation

GPS RO is a global sounding technique providing atmospheric profiles useful for numerical weather prediction and climate studies. The system sounds the Earth’s atmosphere evaluating the delay affecting the passing radio signal. Such a delay is processed to retrieve profiles of refractivity at global scale, and subsequently temperature, pressure and humidity. In the lower atmosphere the classical retrieval schemes need of additional information, e.g. of the knowledge of temperature from independent observations at each GPS occultation. This book proposes a method based on neural networks to retrieve atmospheric profiles from RO refractivity. Once trained, the neural networks provide a suited transfer function able to obtain atmospheric profiles overcoming the requirement for additional information at each GPS occultation. The analysis should be useful to shed some light on GPS RO data at different stages and to develop smart and simple retrieval algorithms handling tricky data sets.