Operational Use of the Rapid Update Cycle

1
RUC Land Surface Model implementation in WRF
Tanya Smirnova, WRFLSM Workshop, 18 June 2003
2
Part 1 Current and Future Initialization
of WRF Land States at FSL
3
Goal for use of WRF in the Rapid Update Cycle

• 2006 - Use WRF model in Rapid Update Cycle (or
  Rapid Refresh) application at NCEP
• First step test WRF model against current RUC
  hydrostatic model using common RUC initial
  conditions
• WRFRUC WRF initialized with RUC-20 initial
  conditions, full-resolution native q/s coordinate
  data, including 3-d hydrometeor, land-sfc data

4
WRFRUC model configuration

• NCAR mass-coordinate dynamic core - v.1.2.1
• 35 vertical sigma-p levels
• Initial conditions including land states for
  WRFRUC
• - native coordinate data from FSL RUC20 cycle
  including assimilation of observations not yet
  used in NCEP operational RUC20 Coupled Data
  Assimilation System (CDAS) available for
  outside users from
• the FSL ftp site in GRIB format
• Lateral boundary conditions from the same FSL
  RUC20 48h forecast
• RUC post-processing adapted to WRF output to
  produce RUC look-alike GRIB output

5
RUC CDAS - four-dimensional system (funded by
GAPP)

• Uses a forward full-physics model
• Cycles surface/soil fields depending on the RUC
  atmospheric forcing
• Cycles 5 hydrometor species cloud, ice, rain,
  snow and graupel. Cloud clearing/building based
  on GOES data
• New compared to RUC operational
• Forecast length (48-hour forecasts with hourly
  outputs)
• Assimilation of
• NEXRAD Radar reflectivity observations
• GPS precipitable water
• Boundary-layer profilers
• Mesonet observations collected at FS
• Main Goal to improve 1-h precipitation forcing
  and the land surface model climate

6
RUC Control
Stage IV Rainfall
24-hour precipitation accumulation ending at 1200
UTC 6 May 2003
RUC CDAS
7
RUC Control
Spatial Correlation fields of 24-h Accumulated
Precipitation ending at 1200 UTC 6 May 2003
(Dongsoo Kim)
RUC CDAS
8
Diurnal cycle of biases from RUC control and RUC
CDAS averaged for the period 1 December 1
March 2003
2-m dew point
Western US
2-m temperature
Western US
9

• Two WRFRUC systems run at FSL in real time
• WRFRUC with 10-km horizontal resolution for the
  TAQ (Temperature and Air Quality) project
• - 48-hour forecasts twice a day (00 and 12
  UTC, runs on JET since June 2002)
• 2. WRFRUC with 20-km horizontal resolution on
• CONUS domain
• - 24-hour forecasts twice a day (00 and 12 UTC,
  runs on JET since February 2003)

http//ruc.fsl.noaa.gov - real-time fields
10
Physics options used in WRFRUC at FSL -
NCEP 5-class microphysics scheme (option 4) -
RRTM longwave radiation (option 1) - Dudhia
shortwave radiation (option 1) -
Mellor-Yamada-Janjic Monin-Obukhov surface
layer (option 2) - RUC land-surface model
(option 3) - Mellor-Yamada-Janjic TKE scheme
(option 2) - Kain-Fritsch (for CONUS) and
Betts-Miller - Janjic (for TAQ) cumulus
parameterization (option 1, 2) as of May 2003
11
Schematic presentation of processes included into
RUC-LSM
6 levels in soil 0, 5, 20, 40, 160, 300
cm State variables - volumetric soil moisture,
soil temperature, snow cover/depth/temperature
- cycled in RUC 1h cycle since 1997.
12

• WRFRUC initialization needed
• Changes to WRF SI (Brent Shaw)
• use of native RUC vertical coordinate rather
  than isobaric levels to provide initial fields
  of atmospheric variables including hydrometeors
  (vapor, cloud, ice, rain, snow, graupel)

The most recent official release of WRF SI
includes all these changes
13

• REAL changes for WRFRUC initialization
• Changes to REAL (Dave Gill)
• accommodate for level structure in RUC soil
  domain
• pass through hydrometeor fields
• Further changes needed to pass through from SI to
  WRF model other land-surface related variables
  such as
• 2 fields for snow temperature
• snow density
• water vapor mixing ratio at surface
• liquid volumetric soil moisture and others

14

• WRFRUC LSM uses
• soil and vegetation parameters, vegetation
  fraction and albedo provided by WRF SI
• cycled soil temperature and moisture from RUC20
• (RUC and WRFRUC use the same LSM, land-use and
  soil classifications, and the same parameter
  tables)
• cycled snow depth and temperature from RUC20
• ice in soil is initialized in WRF
• Atmospheric forcing is provided by WRF. Still
  need from WRF modeling framework
• precipitation type (solid versus liquid)
• option in surface driver for implicit solution of
  energy and moisture budgets

15
Vegetation types both provided by WRF SI (24
USGS classes)
RUC20

• Land-use parameters
• roughness length
• emissivity
• plant coefficient

WRF10
16
Soil types both provided by WRF SI (16
classes)
RUC20
WRF10
Soil parameters look-up table
17
Soil moisture analysis Valid 0000 UTC 17 June 2003
RUC20
WRF10
18
RUC10 Terrain Elevation (dm) TAQ domain
RUC20
19
Surface temperature 0000 UTC, 17 June 2003
RUC20
20
Part 2 Evaluation of LSM performance
21

• RUC LSM participated in
• Project for the Intercomparison of Land-Surface
  Parameterization Schemes (PILPS) - Phase 2d
• Snow Models Intercomparison Project (SNOWMIP)
  Phase 1
• RUC LSM is implemented in
• Operational RUC20 at NCEP
• Real-time RUC20 at FSL (CDAS)
• MM5 chemistry package (Georg Grell) used for
• - air quality predictions
• - regional climate simulations (FSL, Germany,
  Israel)
• WRF model

22
Improved 1-d (PILPS 2d Valdai, Russia) total
runoff and snow water equivalent forecasts with
improved snow and soil physics in MAPS
land-surface model
Total runoff
November 1976 - May 1977
Skin temperature
Snow water equivalent
23
Effects of frozen soil physics on the simulation
of the melting seasons, Valdai, Russia (1966-1983)
Dates when snow ablation starts
Dates when snow pack is all melted
(Smirnova et al., JGR (2000), 105, 4077-4086)
24
SNOWMIP, an intercomparison of snow models first
results P. Etchevers, E. Martin, R.
Brown et al. ISSW meeting,
August 2002
25
Cycled field of snow depth from operational
RUC20 at NCEP Valid at 2100 UTC 8 January 2003
7 January 2003
NESDIS daily snow cover
8 January 2003
8 January 2003
26
12-h forecast Valid 1200 UTC 29 January 2003
RUC
WRF
27
18-h forecast of surface temperature from RUC and
WRF against RUC-20 analysis 1800 UTC 29 January
2003
RUC-10
WRF-10
RUC-20
28
Station verification for TAQ project
Hartford, CT 9 June 2003 0000 UTC 11 June
2003 0000 UTC
http//www.etl.noaa.gov/programs/2002/taq/verifica
tion
29
Station verification for TAQ project
Boston, MA 9 June 2003 0000 UTC 11 June
2003 0000 UTC
http//www.etl.noaa.gov/programs/2002/taq/verifica
tion
30
Station verification for TAQ project
Worcester, MA 9 June 2003 0000 UTC 11 June
2003 0000 UTC
http//www.etl.noaa.gov/programs/2002/taq/verifica
tion
31
12-h surface forecasts verified vs. METAR obs 11
April 11 June 2003 RUC-20 vs. WRFRUC-20 all
METARs in domain
32
12-h winds aloft forecasts verified against
rawinsonde RUC-20 vs. WRFRUC-20
33
WRFRUC-20 (KF cumulus)
21-h forecasts Valid 2100 UTC 10 June 2003
RUC-20 (Grell-Devenyi cumulus)