This paper addresses the problem of providing software support for simulating complex physical systems that require multiple physical models, potentially at multiple scales and resolutions and implemented using different programming languages and distinct parallel programming paradigms. The individual models must be coupled to allow them to exchange information either at boundaries where the models align in physical space or in areas where they overlap in space. Employing multiple physical models presents several difficult challenges, both in modeling the physics correctly and in efficiently coupling multiple simulation codes for a complete physical system. As a solution we describe InterComm, a framework that addresses three main parts of the problem: (1) providing comprehensive support for specifying at runtime what data is to be transferred between models, (2) flexibly specifying and efficiently determining when the data should be moved, and (3) effectively deploying multiple simulation codes in a high performance distributed computing environment (the Grid).