Formalized Linkage of Atmospheric Dispersion Models to Transport in Other Media

To conduct comprehensive environmental assessments, an integration of atmospheric dispersion models with groundwater, surface water, and soil dispersion models in a single system is needed. Standalone atmospheric dispersion models are widely used in regulatory applications to assess potential human and environmental impacts.  Although these models often use data from other models and produce data for models, those connections are typically accomplished by a manual transfer of data.   This paper describes the creation and application of an integrated modeling system that formalizes and automates these types of input and output data transfers.

 

 As background, linkage methodologies used in our integrated modeling systems has evolved over the years.  The efforts reported here started in the mid-1980s.  The first system, Multimedia Environmental Pollution Assessment System (MEPAS), was a hard-wired software system with linkages of release, environmental transport and dispersion, exposure, and risk models.  This system was developed for the U.S. Department of Energy (DOE), who used that system to do a nation-wide survey of the potential health impact from environmental releases its 16 major facilities (DOE 1988).  The second linkage methodology replaced the code specific linkages with standard flux files defined for each media to media transfer.  One aspect of this effort was specifically to create the capability of implementing legacy standalone dispersion models into the integrated assessment system.  The standard flux file model linkage methodology is implemented in the Framework for Risk Analysis Multimedia Environmental Systems (FRAMES) V1.  This system is currently being used in hazardous material assessment systems including ARAMS (maintained by the US Corp. of Engineers) and as the platform for a multi-pathway radiation exposure assessment code, GENII.

 Based on the experience with FRAMES V1, the need for a more formalized system for defining the parameter transfers between models was identified. The greatest challenge encountered in previous model integration efforts was assuring the correctness and appropriateness of information being transferred between media.   A new more generic linkage methodology was developed based on the concept of using dictionaries to define the contents of databases to be transferred information between models (Whelan et. al 1994).  The linkage of a model, such as atmospheric dispersion code, requires definition of what information the model consumes (i.e. input data) and what information the model produces (i.e. output data).  A dictionary, or set of dictionaries, defines the parameters needed for model linkage operation.  The set of specifications for each parameter to be transferred – including parameter attributes such as name, dimension, units, type, limits, relationship to other parameters, etc. The dictionary-based approach has proved to be very effective in avoiding the sources of model linkage errors that had been encountered in earlier systems.  The dictionary-based system also has proved to be much more flexible -- allowing the generic linkage of many types of models.

 The implementation of legacy codes is supported in FRAMES V2.  This capability allows applications to use previously-developed standalone air dispersion models in a much broader context.  Software tools are described that expedite the formal implementation of such legacy atmospheric dispersion models with minimal changes in the original software coding. 

 The dictionary-based linkage of datasets and models implemented in FRAMES V2 provides a “plug and play” capability for doing compressive assessments.  Examples are provided of 1) the input and output dictionary files for two air dispersion models and 2) the linkage of these two air dispersion codes in multiple-media environmental assessment computations.  A multiple-media simulation linkage example shows the input source term being generated by a “Facility” model and the air dispersion model outputs (air and soil concentrations) linked to a sequence of “Receptor”, “Exposure”, and “Risk” models.  The result is a single source to air-pathway risk simulation involving five different models.