Localized Data Assimilation in the Ionosphere-Thermosphere Using a Sampled-Data Unscented Kalman Filter

摘要

We apply the unscented Kalman filter (UKF) to data assimilation based on the vertical one-dimensional global ionosphere-thermosphere model, which models the highly coupled, strongly nonlinear Earth's upper atmosphere. To reduce the computational complexity of UKF, we introduce a localized, sampled-data update scheme with frozen-intersample error covariance, and examine its performance through numerical simulation.