Statistical Database Management for Ecosystem-Effects Analysis

摘要

A team of scientists in the Environmental Sciences Division (ESD) at Oak Ridge National Laboratory has developed a multi-year study to estimate the environmental effects of a synthetic oil spill into lentic freshwater ecosystems. The toxicity of synthetic fuels was evaluated at three levels of increasing complexity; single species bioassays, laboratory microcosms, and outdoor pond ecosystems. A statistical database management system had to be developed that would organize, document, retrieve, and analyze the large amounts of data that were generated during the study. The potential problems anticipated in developing the database included: (1) developing a logical file organization and documentation system; (2) designing consistent and unambiguous code sheets while satisfying the needs of different scientific investigators; and (3) creating and updating data files during the period of study. These problems were resolved by: (1) developing a hierarchical file structure and naming convention using partitioned data sets; (2) creating investigator specific and project specific code sheets while maintaining consistent variable names across all project components; and (3) employing the Statistical Analysis System's MACRO, INFILE, and SET statements for data set creation and update.