This is a joint activity involving researchers from NOAA and University of Maryland to develop ocean applications of data assimilation. This upgrade will allow the next generation CFS to gain the benefit of a more integrated atmosphere-ocean-land analysis system in which both ocean and atmosphere components use coupled ensemble-based estimates of flux error at the interface. Key personnel include Dr. Steve Penny who is developing the assimilation code and Dr. Gennady Chepurin, who is developing an upgraded ocean GCM. The first task is to upgrade NCEP’s Global Ocean Data Assimilation System (GODAS) from the current 3DVar system implemented in 2003 to the ensemble Local Ensemble Transform Kalman Filter (LETKF). GODAS serves as the ocean component of the integrated atmosphere-ocean analysis system, in turn providing the initial conditions for NCEP’s atmosphere/ocean Climate Forecast System, version 2, (CFSv2). The second task will be to combine the 3DVar and LETKF systems to form a hybrid version of GODAS. We will explore the effectiveness of this hybrid system to represent time-evolving local correlations due, for example to fronts or currents, while at the same time maintaining large-scale correlations. An additional task will be to examine the computational efficiency of the hybrid filter relative to its 3DVar and LETKF alternatives.
We are in the first year of this project.
Link to presentation for 28May2014 telecon: https://drive.google.com/folderview?id=0BwoSvidgOLDMVzgzVnhzUUk5TFU&usp=sharing