Louis J. Wicker

The ensemble Kalman filter has become widely used in the geosciences as a data assimilation and analysis method.  Recently work by a number of researchers suggest that it is a viable data assimilation method for Doppler radar in non-hydrostatic convection resolving models.  Data assimilation at these smaller scales is very different than that at large scales.  The problem is primarily a retrieval of the full state variables from only radial velocity and reflectivity.   Verifying analyses is also extremely difficult, since the complete 3D state is never fully observed, even in field programs such as VORTEX2.  Therefore any opportunity to evaluate the EnKF methodology against other types of storm-scale analyses is potentially important.  In May of 2004, a unique high-resolution storm-scale dual-Doppler data set was collected by two mobile C-band Doppler radars during the central Oklahoma 2004 TELEX field program from a tornadic supercell.  Using data from only one radar at a time, a series of EnKF assimilation experiments are performed and the results are compared to a series of dual-Doppler analyses generated from the same radar data from my collaborators, Drs. Biggerstaff and Ziegler, and OU student D. Betten. The results show that at least for this case, the assimilation of single Doppler radar data can reproduce the kinematics of the dual-Doppler wind analyses fairly well, with the added benefits of having analysis estimates of the thermodynamic state and higher temporal sampling of the wind and buoyancy fields between radar volumes.