Several years ago, the Water Environment Research Foundation instituted a research program (the Stormwater Research Challenge) aimed at improving the ability of decision makers to implement solutions “Linking BMP Systems Performance to Receiving Water Protection to Improve BMP Selection and Design”. One need which emerged was found in the models available to solve this kind of problem. It was concluded there was little need to build new tools per se, since the range of available models and options, while still offering opportunities for improvement, basically covers the needs of users. However, a formal canvass of subscribers and stakeholders made it clear that a critical issue remains in that the various available models tend to ‘talk’ to each other only to a limited degree, and the ensuing level of effort of ‘one-of’ or ad hoc data management methods typically required to bridge this gap is substantial. Some initiatives, notably the EU-centric OpenMI effort, tackle this problem by developing standards or methods that provide a means to achieve a very close integration at a core programming level. Other methods also exist. However, it was concluded that there was considerable utility in developing a Framework which would provide a communications hub focused on translating time series results from one model into a form usable by another. This system, also providing the ability to complete basic and enhanced decision support functions (graphical and statistical interpretation, archiving, uncertainty analyses, etc.), enables the user to maintain their investment in ‘best of breed’ tools for various aspects of a modelling program, but also allows them to have those tools interact in a flexible and efficient way. A watershed model might, for example, provide a series to a BMP model, which routes the series and in turn provides it to a receiving water model, and the results of all three can be compared on a common basis despite the fact that the underlying models are quite different. All that is needed is the ability to translate series from one format to another. What makes the Framework a supportable strategic notion is the efficiency gained by implementing an architecture that translates each input/output series via a common internal series, which means that drivers (series format converters) for each model supported by the Framework can be limited to one input/output pair. Once a driver is created, the model in question can interact with any other supported model, subject to some basic common constraints of purpose and definition. The Framework, therefore, is comprised of a suite of supported third party models (that may vary from installation to installation), data stores (e.g. I/O, control, db files), and support tools (e.g. decision and analysis) linked together by “converters” and associated logic structures that translate data files from one format to another in a manner transparent to the user. This paper explores the technical issues in this effort, the current results, and possible avenues for future extension, with an aim to developing a community understanding of how this development may address the common objectives and problems of modellers in this field.