At present, the majority of regional numerical weather prediction models are configured through the use of an input file in which the user sets several variables to locate the grid structure, define input and output data streams and other parameters that control the behavior of the numerics and physical parameterizations. Depending on the design of the software there may also be several preprocessing steps for converting the input datasets into a format usable by the model. Once the model is configured and the input data streams are defined the model is executed to generate forecast or simulation analysis files. The users then utilize an analysis program to generate graphical output to examine the model forecast. When used in an operational mode over a same region, computer scripts are used to manipulate the input file, execute the model and generate graphical output which is often disseminated through a web page.

Operational use of a numerical model over a same region, once the scripts are properly set up, is a relatively straight forward process. However, configuring a numerical model to conduct simulations over any specified area on demand is a more involved process since a user will have to configure model grids over an area that he/she may not be familiar with and also assemble the necessary data streams specific to that area. This is not a serious drawback when the models are used in a research setting. However, for emergency management applications that require customized numerical simulations to be initiated and analyzed with urgency, a Graphical User Interface (GUI) that allows rapid configuration and analysis of numerical model simulations is desirable. A Java based GUI is being developed which will allow users to rapidly develop model simulations. The system is designed with multiple levels of interactivity; one with low levels of interactivity assumes a non-technical user while the one with a higher level of interactivity is provided for a technical user. The system is designed to remove the complexities associated with assembling the input data streams, as well as to easily locate the model grid domain and allow rapid analysis of the model output.