SPC/NSSL Spring Program 2003

.Members of a forecast team evaluate deterministic model output during the 2003 SPC/NSSL Spring Program. Pictured are (l-r) John Gaynor (USWRP), Dave Sill (Met. Service of Canada), Jim Hansen (MIT), Ken Mylne (UK Met. Office), Greg Mann (White Lake, MI NWSFO), and Harold Brooks (NSSL).

The NSSL and SPC were hosts to another collaborative Spring Program in 2003. This year's program explored two promising applications of numerical models in forecasting severe weather: 1) the use of Short-Range Ensemble Forecast (SREF) prediction systems, and 2) the use of high-resolution deterministic models. As in previous years, forecast/research teams were anchored by SPC forecasters and NSSL/CIMMS researchers. The teams were rounded out with visiting scientists from numerous institutions, including the Environmental Modeling Center (NCEP/EMC), the Forecast Systems Laboratory, the Norman, OK and White Lake, MI NWS Forecast Offices, the University of Arizona, OU, the University of Washington, Iowa State University, MIT, the United Kingdom Meteorological Office, and the Meteorological Service of Canada. In addition, bservers from COMET and USWRP participated.

The SREF systems used in the program included two separate ensembles, one from NCEP and one from NSSL. The NCEP ensemble was an operational system that used an automated "regional breeding" method to perturb initial conditions for individual model runs. It was a multi-model ensemble composed of five Eta-model members, five EtaKF members (the EtaKF is a modified version of the Eta model developed at NSSL), and five members using NCEP's regional spectral model. The NSSL ensemble was a newly developed system based on MM5. It utilized forecaster input to identify regions and parameters of meteorological sensitivity, ingesting this information in an adjoint model to produce a spectrum of initial conditions for the ensemble. With this unique interactive method, forecaster judgment modulated the nature of perturbed initial conditions in the ensemble, rather than automated objective procedures.

In the second area of focus, participants compared mesoscale model forecasts using parameterized convection to cloud resolving forecasts (i.e., without parameterized convection) from the Weather Research and Forecasting (WRF) model. The goal was to provide a preliminary assessment of the forecast value of highresolution models compared to the current generation of operational and experimental forecast models, including the WRF, Eta, EtaKF, RUC, and NCEP's new Nonhydrostatic Mesoscale Model (NMM). The program ran weekdays from April 14- June 6.

-- by Jack Kain and Mike Baldwin

For more information see: http://www.spc.noaa.gov/exper/Spring_2003/