Collaborative Modeling Project Underway at NWS Huntsville

For the past year, the SPoRT modeling team has been collaborating with the Mobile and Houston NWS Offices, incorporating SPoRT datasets such as the near real-time MODIS green vegetation fraction (GVF), Land Information System (LIS) soil moisture, and SPoRT SSTs into their local WRF EMS model.  Using the Model Evaluation Tools (MET) developed by the National Center for Atmospheric Research, objective verification is being conducted to compare model performance with SPoRT datasets versus that of a control configuration without SPoRT data.  SPoRT has developed a set of scripts that interface with MET and the WRF EMS output to generate objective verification statistics.  The improvement in some forecast parameters due to the use of the high-resolution SPoRT datasets have been documented in previous studies.  The collaborative modeling efforts are still ongoing with the Houston and Mobile offices.

Figure 1. 9 km domain from Huntsville WRF EMS. Parameter shown is precipitable water (inches) valid 06Z 23 August 2012. (click for larger image)

The Huntsville NWS office has entered into a new and similar collaborative phase with the SPoRT modeling team.  The Huntsville WRF EMS model is configured with a 9 km outer domain (shown in Figure 1) encompassing much of the southeastern U.S. and a part of the Ohio Valley and Midwest regions.  A 3 km resolution nested domain contains much of the Tennessee Valley region (Figure 2) and covers a slightly larger area than the previous WRF EMS local run at NWS Huntsville.  The model produces “standard” output parameters such as temperature/height/wind at 2m, 925mb, 850mb, 700mb, 500mb, and 250mb out to 36 hours.  However, other output  includes (but is certainly not limited to) composite reflectivity, SBCAPE, SBCIN,  0-3 km SR Helicity, 0-1 km shear vector, and precipitable water.  In addition, forecast soundings are produced for eight geographically diverse points within the forecast area.

Figure 2. 3 km nested domain from Huntsville WRF EMS. Forecast parameters shown are 2 meter temperatures (contoured) and 2 meter dew point temperatures (shaded) valid at 06Z 23 August 2012. (click for larger image)

Data and imagery are currently output to an internal web page for forecasters to use operationally, but model output files in grib-2 format will also make data and imagery available for inclusion in AWIPS II in the near future.  Due to the recent acquisition of a new modeling workstation with enhanced computing capability, NWS Huntsville will be able to run both experimental and control forecasts in real-time, with and without SPoRT datasets, respectively.  This will ensure a clean comparison of verification statistics from model runs made on the same computational platform, all performed in-house.  During this collaborative research period, we will be working with the SPoRT modeling team and using the MET tools to objectively determine model performance utilizing SPoRT datasets.

Figure 3. WRF EMS sounding for Huntsville, AL valid 21Z 23 August 2012. The inclusion of SPoRT datasets are expected to have positive impacts on near-surface heat and moisture fluxes, leading to improvements in the model forecast boundary layer temperature and dew point profiles.

With the inclusion of the SPoRT datasets, we expect to see improvements in the model’s handling of surface heat fluxes and associated improvements in low-level temperature/moisture fields and instability parameters.  In the future, we plan to do some case studies evaluating the experimental and control runs.  Using the SPoRT scripts and MET, this can be done in an objective manner in near real-time.


Thanks to Jon Case and Andrew Molthan of SPoRT for their help with this post.

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s