Earth system science deals with complex systems that pose many significant representation challenges. As depicted in the classic Bretherton diagram of biospheric cycles (Figure 1), modeling the earth system involves numerous interacting components, each of which can be further dissected into sub-components that are studied by specialists in a wide range of disciplines. From this description, problems of both model interoperability and the model simulator interoperability already become evident. Given the complexity of the task and the number of research groups and individuals involved, there exist a wide diversity of modeling approaches, such as models based on differential equations or stochastic methods, that make not only the interoperation of model specifications difficult but the intercomparison of the structure and results of similar models as well, as is evident in the work undertaken by The Global Analysis, Integration and Modeling Task Force (GAIM) (for an example see1). Similarly, in terms of simulator interoperability there is also a wide range programming languages and software in which models are developed, making it difficult to couple a FORTRAN model of thermohaline circulation with an ice sheet model in C++.