Assimilation of Cloud and Precipitation data
Research at the Met Office
Andrew.Lorenc
UK Met Office
Abstract:
The Met Office have been assimilating cloud and precipitation data into their
operational mesoscale NWP system for ~10years, using a "nudging" approach. With
careful attention to bias correction and quality control of the observations, and tuning
of the assimilation, the data have been shown to have a positive impact on rainfall
forecasts. However it has been difficult to progress beyond a moderate impact, or to
fit the data closely, so development is now focussing on variational methods.
1D-Var methods have been developed for deriving the height and amount of cloud
from IR sounders, and for retrieval of cloud liquid water from microwave sounders.
The operational 3D-Var is being extended to use cloud observations through a
diagnostic relationship with relative humidity, and (resolved) precipitation through a
relationship with balanced vertical motion. A 4D-Var scheme is being developed for
implementation in the first instance for synoptic scale NWP. It is being extended to
allow the direct assimilation of precipitation and cloud observations.
The Met Office has a research programme aimed at developing the Uinified Model
system to give improved very short period precipitation forecasts for 5km*5km boxes
(running on ~2km grid). We are currently following a pragmatic development
programme, using as much as possible of the existing model and assimilation.
However convective scale NWP brings with it both a requirement and an opportunity
for basic research into various problems associated with modelling and data
assimilation with partially resolved convection.