Generating Physically Consistent, Integrated Environmental Databases from Distributed Data and Modeling Resources

The Environmental Scenario Generator (ESG) provides environmental data that portrays the state of the natural environment in a physically consistent manner, across and within the space, atmosphere, ocean, and/or terrain domains, with resolution appropriate for the intended application. To accomplish this goal, the ESG must provide capabilities to: 1) capture and interpret customer requirements 2) identify a set of resources to meet those requirements 3) produce an environmental database from those resources

The ESG has been under development since 1998, and currently exists as a distributed software application providing data mining and database production services. The data mining functionality allows the customer to identify specific environmental events within reference data resources that appear to provide the conditions desired by the customer. The ESG data mining architecture is presented in Presentation 4.22, "The Environmental Scenario Generator" (Kihn et al., 2001). This presentation addresses the architecture developed to support the production of environmental data bases from distributed data and modeling resources.

The critical area of capturing and interpreting customer requirements has not yet been developed in ESG, and will be the focus of continuing research and development. This capability will allow the ESG to select a set of data and modeling resources with a consistent level of fidelity that will result in a physically consistent, integrated environmental database that meets the customer requirements. The ESG currently uses a Scenario Generation Template (SGT) to capture the data and models to be employed, their inter-dependencies, and the customer requirements for the structure of the final database. The SGT conveniently separates the problem of requirements interpretation from database production, while providing a useful static record that can be developed, reviewed, approved, and subsequently repeatedly used to control database production.

The focus of this presentation is on the Java architecture that has been constructed around the use of an SGT to control the production of environmental databases. The SGT is created and stored as an eXtensible Markup Language (XML) document, and subsequently parsed and used to define the structure of, and relationships between, the Java business objects required to produce the desired environmental database. The architecture consists of a common data model for exchanging data, a process-controller that tracks the sequence of data orders and model runs that must be executed, and business objects representing the final scenario database, a Master Environmental Library (MEL) data order, and a remote model run. The common data model defines a five-dimensional (two spatial dimensions, a vertical dimension, time, and parameter) collection of two-dimensional spatial grids (each one representing a single parameter valid at one vertical level and one instance of time). The XML structure of the SGT, and the mechanism by which instances of the common data model are created from it, will be presented. The subsequent process by which environmental databases are produced, in accordance with the SGT, will also be presented.