The next BMB data assimilation system
Chaolin Zhang, Yingchun Wang, Shuiyong Fan, Jiqin
Zhong (BMB/IUM)
Xiang-Yu Huang, Ying-Hwa Kuo, Yongrun Guo (NCAR/MMM)
1. Introduction
The NCAR/MMM has developed a
three-dimensional variational data assimilation (3DVAR) system based on the
Weather Research and Forecasting (WRF) model. The WRF 3DVAR system has been
implemented successfully for both operational use and for research. It has been
applied to a wide range of data assimilation research problems, including the
assimilation of Doppler radar data, the ground-based GPS water vapor data, and
satellite measurements.
The WRF 3DVAR system will be
the next operational system at BMB. A collaboration project between BMB and
UCAR started on
2. Domains
The current operational data
assimilation system at BMB runs on 3 domains using two-way nesting. These
domains will be used for the next operational system for continuation. In
addition, a 1 km domain will be added.
The resolution of the 4
selected domains will be 27 km, 9 km, 3 km and 1 km. The domains will have
151×151, 142×184, 172×199, 211×211 (NS×WE) horizontal grid points,
respectively. The domain centers will be all at 116ºE, 40ºN. These domains are
shown in Figure 1.
Two-way nesting strategy will
be used.
3. Levels
In the vertical direction, 38
levels will be used for all selected domains. The full s levels will have the following s values:
1.00,0.995,0.99,0.985,0.98,0.97,0.96,0.95,0.94,0.930,0.92,0.910,0.9,
0.88,0.86,0.83,0.8,0.77,0.74,0.71,0.68,0.64,0.6,0.56,0.52,0.48,0.44,
0.4,0.36,0.32,0.28,0.24,0.2,0.16,0.12,0.08,0.04,0.
The model variables are
placed on full s levels or half s levels, as shown in
Figure 2.
Figure 2. Vertical levels.
Dotted lines denote the full-s levels
(total:38), dash lines denote the half-s levels
(total:37), top level is at 50 hPa.
4. Forecast model
4.1 Dynamics
The model uses an Eulerian
solver based on a flux formulation to solve the fully compressible
non-hydrostatic equations. A third order Runge-Kutta time integration, third to
fifth order advection operators, and split-explicit acoustic/gravity wave are
used.
4.2 Physics
A basic physics packages will
be used for all domains, including the YSU (YonSei University Scheme) PBL
scheme 5-layer soil model(not LSM), RRTM (Rapid Radiative Transfer Model)
longwave radiation, Dudhia shortwave radiation, and the WSM-6 (WRF
Single-Moment 6-class) microphysics scheme, second-order horizontal diffusion,
and vertical velocity damping. KF (Kain-Fritsch) Cumulus parameterizationwill
be used for the 27 and 9 km domains.
4.3 Lateral boundaries
For all domains, the analysis
and forecasts from the CMA/NMC global data assimilation system will be used as
lateral boundaries.
5. Data assimilation
5.1 Observations
In addition to the conventional data from GTS, local
observations made and managed by BMB will be assimilated. Enhancements in WRF
3DVAR system to incorporate AWS data and GPS ZTD/PW will be made, and
experiments to assess the impact from the local observations will be carried
out in the BMB-UCAR project. Other local observations may also be considered in
a later stage of the project.
5.2 Cycles
The resolution of AWS and
GPS-met networks of
For the 27 and 9 km domains,
6 hour cycles will be used. For the 3 km domain, 3 hour cycles will be used.
Experiments will be carried out using the 3 km domain and 1 hour cycles.
In order to decide whether a
regular restart of cycles is needed, data assimilation experiments will be
carried out.
6. Remarks
Although the next BMB data
assimilation system will be based on WRF, MM5 may play an important role at the
beginning of the project as it has been the operational forecast model at BMB
for several years.
Two months (one summer month
and one winter month) MM5 forecast data have been archived and will be used to
compute background error statistics for the 27, 9 and 3 km domains.
A similar data set, generated
by WRF, will also be made available during the BMB-UCAR project.