DWFE Experiment Design
Design of DWFE
The WRF Developmental Testbed Center is conducting a high resolution
NWP forecast
experiment during the winter season:
- Period: December 2004-March 2005
-
Horizontal resolution: 5km
-
Vertical resolution: 38 levels
-
Domain Size: CONUS
-
Emphasis: Eastern US
-
Forecast period: 48 hours
-
Cycle: 00 UTC
-
Dynamic Cores: ARW and NMM
-
Parameterizations:
- ARW NMM
- Microphysics WSM5 Ferrier
- Radiation RRTM/Dudhia MYJ
- Plantery Boundary Layer YSU Eta
- Land Surface Model Noah Noah
- Cumulus none none
For more information on WRF, go to http://wrf-model.org.
Initial and boundary conditions for these forecasts are based
on Eta 212 grids.
In addition, a high-resolution land-surface data assimilation system (HRLDAS)
is being used
to specify the initial land surface fields for the ARW forecasts (insert link
to
HRLDAS document).
Objectives
- Evaluate the value of high-resolution NWP forecasts over
a large domain during the winter to determine whether operational
forecasters find the high-resolution precipitation forecasts
useful,
- Identify whether other small-scale flow features resolved
by the high-resolution grids have forecast value.
- Evaluate the
relative value of high-resolution deterministic forecasts
to the lower resolution ensemble forecast techniques,
- Evaluate current
NWP models and determine where further model research is
required
Evaluation of the DWFE forecasts will involve both
subjective feedback from NWS forecasters(link to forecaster
feedback form) and the general NWP community (link to general
feedback form), as well as objective evaluation through the
computation of verification statistics.
Verification statistics will be computed using NCEP's verification package,
as well as FSL's RTVS. Verification results for DWFE can be viewed at the
verification website. The objectives and design of the DTC Winter Forecast Experiment
involved close consultation between operational forecasters,
researchers and the DTC. This close interaction between the
DTC and NWS will continue during the assessment stage of this experiment.
Why use 5-km horizontal resolution?
Results from the summer experiment suggest that convective systems are very
well represented for grid spacings on the order of 5 km. DWFE forecastswill
also be run without parameterized convective processes to test the performanceof
WRF with explicit convection during the wintertime regime.
Why use the CONUS domain?
The larger the domain, the longer the value of a high-resolution forecast is
retained. For the CONUS, the East is the main beneficiary of this large domain,
particularly in the winter, due to the rapid west to east movement of weather
systems during this season. Although the focus is on the East Coast, all National
Weather Service (NWS) regions in the continental US have expressed interest
in participating in the experiment.
Computer Resources
- FSL (NMM core) and NCAR (ARW core) supercomputers
- Forecasts
made available to NWS forecasters through AWIPS, FX-NET and
the web
- The research community, as well as the private sector
will have access to the forecasts through the web
|
