An object-oriented characterization of spatial ecosystem information

Abstract

An information system that supports the work of ecosystem modelers should be easy for the modelers to use, should organize both datasets and computational processes in a convenient hierarchical fashion, should be initially populated with information of interest, and should be easily extensible through contributions from its users. We are currently using object-oriented techniques to implement a network-based ecosystem information system (EIS) to meet these objectives. The initial EIS implementation includes a database consisting of key ecosystem datasets and processes “cast” in the object-oriented framework, i.e., as a collection of classes, instances, and methods. In the long term, we rely on EIS users to extend this database by contributing their work to extend the database. Therefore, while a robust and user-friendly implementation of the EIS software is important, the design of the initial database is even more crucial to long term project success. That is, the initial database must present information of sufficient interest to attract the new user, it must illustrate the intuitive concepts and value of the object-oriented paradigm to the new user, and it must serve as a model to interest users in “re-casting” their own work into this paradigm for contribution to the database. The initial EIS database will include datasets and modeling tools from three important ecosystem applications: climate and regional/global scale carbon-exchange modeling, U.S. Fish and Wildlife's biodiversity analysis (a.k.a. “GAP”), and general data visualization techniques for ecosystem modeling. After a brief overview of EIS functionality, the presentation here focuses on the design of this initial collection of ecosystem related classes, instances, and methods.