Spycher, G.; Cushing, J. B.; Henshaw, D. L.; Stafford, S. G.; Nadkarni, N. 1996. Solving problems for validation, federation, and migration of ecological databases. In: Global networks for environmental information: Proceedings of Eco-Informa '96; Lake Buena Vista, FL. Ann Arbor, MI: Environmental Research Institute of Michigan (ERIM): 11: 695-700.
The H.J. Andrews Long-Term Ecological Research site hassupported production and access of ecological data for over 20 years.We have developed a flexible mechanism to generate validation codebased on a standard set of metadata. This approach reduces the timerequired for data production and permits the maintenance of multipledatabases. New challenges are migration of existing databases toinclude new data formats and federation of existing databases forgeneral access. The more constrained environment of the Wind RiverCanopy Research Facility could be used to examine new distributiontechnologies and incorporate spatial data into the existing structures.Shared research sites oiler incentives to individual researchers to useprescribed protocols and tools. and provide a test bed for solutions tomigration and federation problems.
The H.J. Andrews Long-Term Ecological Research (LTER) site. together withresearch groups of the USFS Forestry Sciences Laboratory in Corvallis, has supportedan organized effort to maintain and store ecological data in a usable form for overtwenty years (Stafford et al., 1986). The Forest Science Data Bank (FSDB) at OSU(Stafford et al.. 1984,1988) stores over 200 long-term and opportunistic databases fromdiverse scientific disciplines for access by database owners and secondary users. TheWind River Canopy Crane Research Facility (WRCCRF) now plans to use FSDBservices including its database validation system, and in return offers an opportunity fortesting migration paths and federation of FSDB databases.
This paper first briefly describes our understanding of problems of validating andstoring ecological data based on the experience of the FSDB. and a mechanism devel-oped there to automatically perform validation checks based on standard metadata andspecific database rules. We note both some data validation problems that remain and
intrinsic limitations with the database technology we have chosen. We then proposesonic ways to overcome these problems and suggest how database research involvingecological data for the WRCCRF at The Evergreen State College (TESC) and OregonGraduate Institute (OGI) might address these problems.