Emergency management personnel at federal, state, and local levels can benefit from the increased use of simulation and modeling for emergency preparedness, including planning, exercises and training. The Department of Homeland Security (DHS) is chartered to identify and address modeling opportunities for national preparedness. Dozens of modeling efforts relevant for emergency response have been identified, but knowledge of existing resources and the expertise needed to provision a simulation with data, execute it, and synthesize results are not uniformly available to all. To meet these needs, the DHS Science and Technology Directorate has spearheaded the Integrated Modeling, Mapping, and Simulation (IMMS) program to create a framework that brings relevant tools to the emergency response community.

IMMS defines a software framework that brings together distributed codes using metadata, heuristic domain knowledge and a uniform interface, to provide integration capability and automated execution. A fundamental goal is to connect users such as training and exercise planners with modeling resources. A fundamental challenge is to bridge the gap in expertise and technical skills between these two communities, a gap which hinders model discovery, provisioning, execution, and interpretation.

We present a platform-neutral, distributed computing framework that connects users (logged in as clients) with models (located on servers). Similar frameworks have been proposed or prototyped, including FAIT for a small set of infrastructure models, iCAV and Palanterra for GIS modeling, and DIAS for discrete event simulations. IMMS contributes two distinct innovations that bring together users and model providers: a discovery process based on web ontology taxonomy trees, and an abstraction for modeling emergency planning scenarios. The discovery process allows subject matter experts to contribute metadata, while enabling less sophisticated users to find relevant models. The abstraction uses "simulation templates" to group models according to function, and then combine functionalities to address hazard and threat scenarios.

The paper will describe the framework architecture, the innovative discovery and simulation template components, and an initial reference implementation to be used in FEMA National Level Exercises (NLE 2010 and NLE 2011).