[3-1] Invited

Linear transforms and averaging for data assimilation in systems with a range of scales

C. Snyder (NCAR)

I will review some work with the ensemble Kalman filter for data assimilation in mesoscale and convective-scale flows. A key issue that arises is the need to simultaneously analyze a range of temporal and spatial scales. Posing idealized statistical models helps to understand the issue better, especially with respect to the challenges it presents for ensemble-based data assimilation. Results from the idealized models also suggest techniques to improve ensemble data assimilation in real systems possessing a range of scales.