Large amounts of laboratory data have been generated as part of environmental investigations, and particularly ground-water investigations conducted during the 1970s and 1980s. These data present a historical water-quality record at a facility by representing, in many cases, the baseline ground-water quality, and changes in water quality due to degradational processes and remedial activities.
Ground-water chemistry data collected during the 1970s and 1980s were often generated using laboratory quality assurance/quality control (QA/QC) procedures considered inadequate by today's standards. In addition, the documentation required to support the quality of the laboratory data may be unavailable. The absence of the supporting QA/QC documentation or the use of different QA/QC procedures than those used today should not be the basis for removal of the data from the facility's ground-water quality database. In a large number of cases, the historical data are “good data.” Historical water-quality data are useful and should be available for use in the design of ground-water monitoring programs, for the selection and design of remedial measures, and to monitor the effectiveness of past and anticipated future remedial activities.
This paper presents a series of techniques that, when used together, can demonstrate the usability of a historical water-quality database where laboratory QA/QC data are unavailable. This paper makes an important assumption that there is a difference between data usability and data validation. The assessment of the usability of a database is performed as a two-part exercise. The first step is to conduct a qualitative pattern recognition procedure and the second is to use more stringent statistical procedures to identify if the data are presenting a consistent description of the ground-water system. This data-usability technique has been successfully applied at a site in California. The results of this case study are also present in this paper.