Forecasters that supported the Joint Urban 2003 (JUT) Field Experiment in Oklahoma City required model guidance on winds and other conditions in order to help field-program managers schedule intensive observation periods. The primary model used for this purpose was the MM5-based 4DWX system that has been jointly developed by NCAR and the Army Test and Evaluation Command (ATEC), and deployed for various homeland security applications by the Defense Threat Reduction Agency (DTRA).

The version of the 4DWX modeling system that was used for JUT support employed 5 nested grids with horizontal grid increments of 500 m, 1.5 km, 4.5 km, 13.5 km, and 40.5 km. A model-based data-assimilation system ingested observations that were representative of the model resolution, and the resulting data sets were used to initiate forecasts on a three-hour cycle. The forecasts were 48-h in duration, except for the two inner-most grids where they were 12 h. Even though this model has been used before for operational forecasting in urban areas (e.g., in support of the Salt Lake City Olympics), this is the highest resolution version of the system that has been used to operationally forecast weather on the scale of metropolitan areas. Forecast meteorological variables were used as input to the DOD SCIPUFF transport and dispersion model in order to predict the general properties of the SF6 plumes released in the experiment.

Embedded within the innermost MM5 grid was the DTRA Urban Wind Module (UWM), with a horizontal grid increment of 100 m. Its lateral-boundary conditions were prescribed by the MM5 forecasts. The UWM does not explicitly resolve buildings, but represents their aggregate drag effect on the boundary-layer flow. It also was coupled with SCIPUFF.

This paper describes the modeling system employed and some objective verification statistics, explains how the model products were used by the urban forecasting team, and describes some example products from the model.