NRL SEMINAR
3:00 PM, Tuesday, 8 September 2009
Advanced and High-resolution Data Assimilation Techniques
Andrew Lorenc
MET Office,
Exeter, Devon, United Kingdom

The review intends to shed light on possible strategies for the
convective scales data assimilation (DA) of clouds and
precipitations.

Operational schemes are often based on quasi-linear optimal estimation
theory, and use the variational technique (e.g. 3DVar/4DVar) to cope
with the minimization problem. The review will try to figure out
possibilities and limitations of the seamless extension of such
schemes to higher resolution (Lorenc and Payne, 2007) and cloud
analysis (Errico et al, 2007).

Another method of solving the quasi-linear estimation problem is to
discretize it with ensembles, (through the various forms of Ensemble
Kalman Filters (EnKF).)

Several reviews have already considered the EnKF vs. 4D-Var approaches
(Dance, 2004), so we suggest the review may focus on hybrid strategies
(1DVAR+4DVAR, VAR+EnKF, Lognormal+Gaussian DA systems) in conjunction
with cloud VAR+analysis.

At those scales and for the hydrometeors, non-linearity becomes a
significant issue. It may first appear as a displacement error,
requiring some specific DA schemes, such as Beezley and Mandel (2008).
The review may also give some clues about the notion of fronts and
vertical/horizontal displacement errors in clouds and the way
alternative DA scheme could be designed to solve such problems.


Just as well reallly as I usually rely on online dictionaries these
days and I doubt very much it would be worth carrying my desktop
through US customs. They thayer baying so quaint ayun all.

Errmm.. wots an hydrometeor?