Assessment of Kalman Smoothing for TIEGCM-Driven EMPIRE Outputs
Trevin
Cox
Illinois Tech
Abstract text
This project investigates the application of a Kalman smoother to TIEGCM-driven EMPIRE outputs using Weimer and AMIE forcing. The original goal was to reduce high-latitude error and improve agreement with AMIE reference data. Initial results show that smoothing decreases the estimated covariance, but the resulting state estimates move farther from AMIE rather than closer. These results suggest that the current covariance formulation may be overconfident, and that additional error tuning, such as latitude-dependent covariance structure, is needed before smoothing can improve physical accuracy.
Student in poster competition
Poster category
DATA - Data Assimilation, Data Analytics, Methods and Management
Poster number
11