2.8. DATA MANAGEMENT

The long-term flask data management strategy adopted in 1992 (see Peterson and Rosson [1993] for a detailed description) maintains a sampling history for over 85,000 ambient air samples and 23,000 special samples, and grows at a rate of ~7000 samples per year. Analysis and processing information is maintained for a suite of measurements that includes CO, CH, CO, H, SF, NO, and the stable isotopes of CO. The present database management strategy has several important features. While individual measurement projects retain a degree of independence, they are inextricably linked by way of the database, making multiple­species comparisons reliable and straightforward. Because the database employs UNIX text files as a storage format, database access does not require specialized software; files can be readily viewed using a variety of operating system commands and off­the­shelf software tools. Software written for the statistical and graphical analysis of flask measurement data can be species independent because the format is identical for different species, providing powerful tools to a greater number of flask programs.

Database maintenance includes regular integrity checks designed to verify consistency throughout the database and to ensure the adequacy of security precautions. Although the current strategy is working well, it is anticipated that the approach will continue to evolve to meet the growing needs as both the sampling network expands and new projects are added. Recently, the need to log procedural changes or equipment modifications into a central repository was recognized so that complete histories of analytical systems or projects can be readily reconstructed. In response, software tools are being developed that will maintain a meta­database, e.g., a database containing information about the data, that will complement our existing strategy.

[back] [contents] [next]