WRF%204D-Var%20Hans%20Huang,%20MMM/NCAR

1
WRF 4D-VarHans Huang, MMM/NCAR

1. A short introduction to WRF 4D-Var
2. The current status The basic system
3. The structure function single ob exp
4. Results from cold-start experiments
5. Results from cycling experiments
6. Our first radar data assimilation experiments
7. Summary

2
WRF 4D-Var developers

• Xiang-Yu Huang,
• Qingnong Xiao, Dale Barker, Xin Zhang, John
  Michalakes, Wei Huang, John Bray, Zaizhong Ma,
  Tom Henderson, Jimy Dudhia, Xiaoyan Zhang,
  Duk-Jin Won, Yongsheng Chen, Yongrun Guo,
  Juanzhen Sun, Hui-Chuan Lin, Ying-Hwa Kuo

Acknowledgments. The WRF 4D-Var development has
been primarily supported by the Air Force
Weather Agency (AFWA). The Korean Meteorological
Administration (KMA) also funded some 4D-Var
tasks.
3
4D-Var 4-Dimensional Variation data
assimilation
(new)
(initial condition for NWP)
(old forecast)
4
WRF 4D-Var JJo Jb Jc
5
Why 4D-Var?

• Use observations over a time interval, which
  suits most asynoptic data and use tendency
  information from observations.
• Use a forecast model as a constraint, which
  enhances the dynamic balance of the analysis.
• Implicitly use flow-dependent background errors,
  which ensures the analysis quality for fast
  developing weather systems.
• NOT easy to build and maintain!

6
A short 4D-Var review

• The idea Le Dimet and Talagrand (1986) Lewis
  and Derber (1985)
• Implementation examples
• Courtier and Talagrand (1990) a shallow water
  model
• Thepaut and Courtier (1991) a multi-level
  primitive equation model
• Navon, et al. (1992) the NMC global model
• Zupanski M (1993) the Eta model
• Zou, et al. (1995) the MM5 model
• Sun and Crook (1998) a cloud model
• Rabier, et al. (2000) the ECMWF model
• Huang, et al. (2002) the HIRLAM model
• Zupanski M, et al. (2005) the RAMS model
• Ishikawa, et al. (2005) the JMA mesoscale model
• Huang, et al. (2005) the WRF model
• Xu, et al. (2005) NAVDAS-AR
• Gauthier, et al. (2007) MSC
• Operation ECMWF, Meteo France, JMA, UKMO, MSC.
• Pre-operation HIRLAM, NAVDAS-AR

7
Necessary components of 4D-Var

• H observation operator, including the tangent
  linear operator H and the adjoint operator HT.
• M forecast model, including the tangent linear
  model M and adjoint model MT.
• B background error covariance (NN matrix).
• R observation error covariance, which includes
  the representative error (KK matrix).

8
WRF 4D-Var milestones

• 2003 WRF 4D-Var project. ?? FTE
• 2004 WRF SN (simplified nonlinear model). 1.5
  FTE
• Modifications to WRF 3D-Var.
• 2005 TL and AD of WRF dynamics. 1.5 FTE
• WRF TL and AD framework.
• WRF 4D-Var framework.
• 2006 The WRF 4D-Var prototype. 2.5 FTE
• Single ob and real data experiments.
• Parallelization of WRF TL and AD.
• Simple physics TL and AD.
• JcDF
• 2007 The WRF 4D-Var basic system. 2.5 FTE

9
Basic system 3 exes, disk I/O, parallel, full
dyn, simple phys, JcDF
I/O
xb
B
call
WRFBDY
NL(1),,NL(K)
R y1 yK
WRF
BS(0),,BS(N)
TL(1),,TL(K)
call
xn
AD00
10
Single observation experiment
The idea behind single ob tests The solution of
3D-Var should be
Single observation
3D-Var ? 4D-Var H ? HM H ? HM HT ? MTHT The
solution of 4D-Var should be
Single observation, solution at observation time
11
Analysis increments of 500mb q from 3D-Var at
00h and from 4D-Var at 06h due to a 500mb T
observation at 06h


FGAT(3D-Var)
4D-Var
12
500mb q increments at 00,01,02,03,04,05,06h to a
500mb T ob at 06h
13
500mb q difference at 00,01,02,03,04,05,06h from
two nonlinear runs (one from background one
from 4D-Var)
14
500mb q difference at 00,01,02,03,04,05,06h from
two nonlinear runs (one from background one
from FGAT)
15
Real Case Typhoon HaitangExperimental Design
(Cold-Start)

• Domain configuration 91x73x17, 45km
• Observations from Taiwan CWB operational
  database.
• 5 experiments are conducted before Haitangs
  landfall at 0000 UTC 18 July 2005.
• FGS forecast from the background The
  background fields are 6-h WRF forecasts from
  National Center for Environment Prediction (NCEP)
  GFS analysis.
• AVN- forecast from the NCEP AVN analysis
• 3DVAR forecast from WRF-Var3d using FGS as
  background
• FGAT - forecast from WRF-Var3dFGAT using FGS as
  background
• 4DVAR forecast from WRF-Var4d using FGS as
  background

16
Observations used in a 4D-Var experiment
17
Typhoon Haitang 20052005.07.16.00Z
18
Typhoon Haitang 2005
19
A KMA Heavy Rain Case Period 12 UTC 4 May - 00
UTC 7 May, 2006 Assimilation window 6
hours Cycling (6h forecast from previous cycle
as background for analysis) All KMA operational
data Grid 60x54x31 Resolution 30km Domain
size the same as the KMA operational 10km
domain.
20
Observations used in 3D-Var
21
Observations used in 4D-Var
22
Observations Verification
23
Precipitation Verification
24
Observation Verification Precipitation, CSI
25
Observation Verification Precipitation, BIAS
26
First radar data assimilation experiment using
WRF 4D-Var Yong-Run Guo and Juanzhen Sun
27
The OSSE setup
061212Z 061300Z 061312Z
4DVAR time window

• 12/4-km WRFV2.2 run
• Initial and boundary conditions Eta 3-hly
  analysis starting at 2002061212Z
• Domain size 271x241x31 (12-km) and 325x280x31
  (4km)
• The control experiment with the domain2
  (325x280x31)

28
Domain settings
height
landuse
29
Hourly rainfall forecast from the control run
starting from 2002061212Z
30
(No Transcript)
31
(No Transcript)
32
WRF 4D-Var Experiment design

• Physics option as the control run (truth)
• Initial time for Experiments at 2002061301Z (the
  13-h forecast starting from 2002061212Z)
• Domain size 151x118x31, covered the convective
  cells.
• Grid size 4-km, Time step 20 seconds
• First guess from NCEP GFS analysis
• Time window for WRF 4DVar
• 0.25-h, 45 steps
• 0.5-h, 90 steps (?)
• Forecast length 11-h.
• BES interpolated from 12km IHOP BES (Hongli Wang)

33
Control run 4-km domain
118
280
151
WRF 4DVar Exp 4-km Domain
325
34
OSSE Radar data generation

• Gaussian random perturbation (0,1), Xpert, added
  to truth
• The truth is obtained by using iowrf utility to
  extract a box-domain ltx121to 271, y94to211gt from
  the 4-km control run lt325x280gt domain.
• Soichiro Sugimotos code as the reference
• The Radius of Radar OBS 200km
• When Rain water mixing ratio gt 1.e-7, the
  reflectivity truth will be computed as Xtdbz
• It is regarded as clear air when Xodbz lt 5, no
  clear air radar obs considered.
• When the bean angle gt 20o, no radar obs
• The radial velocity data are generated in the
  same way as reflectivity

35
Rain water mixing ratio every 5 minutes from
130100 to 130130
130100 130105 130110 130115
130120 130125 130130
36
OSSE Radar data coverage
10 NEXRAD Radar sites over the Exp. Domain
OSSE Radar data coverage at 0100 UTC 13 June 20
37
Experiment design

• TRUTH ----- Initial condition from TRUTH (13-h
  forecast initialized at 2002061212Z from AWIPS
  3-h analysis) run cutted by ndown, boundary
  condition from NCEP GFS data.
• NODA ----- Both initial condition and boudary
  condition from NCEP GFS data.
• 3DVAR ----- 3DVAR analysis at 2002061301Z used as
  the initial condition, and boundary condition
  from NCEP GFS. Only Radar radial velocity at
  2002061301Z assimilated (total of data points
  65,195).
• 4DVAR ----- 4DVAR analysis at 2002061301Z used as
  initial condition, and boundary condition from
  NCEP GFS. The radar radial velocity at 4 times
  200206130100, 05, 10, and 15, are assimilated
  (total of data points 262,445).

38
INCREMENTS (A-B) B is the NCEP GFS analysis at
2002061301Z
Truth Temperature/wind at lowest h level
(3DVAR-FG) Temperature/wind at lowest h level
39
Hourly precipitation ending at 01-h forecast
40
Hourly precipitation ending at 03-h forecast
41
Hourly precipitation ending at 06-h forecast
TRUTH
42
Summary

1. A short introduction to WRF 4D-Var
2. The current status The basic system
3. The structure function single ob exp
4. A cold-start experiment
5. A cycling experiment
6. First radar data assimilation experiment