Conceptual Filters for Data Quality Assurance

Environmental information is the result of extremely complicated natural processes. This information enhances our understanding of the underlying phenomena. Collection and processing of oceanographic data involve a complex sequence of events. At every stage of this process, human, instrumental, analytical, data processing and retrieval errors of different nature and origins may occur. Data quality assurance should therefore be a basic step to safeguard the reliability of information. In the paper, some basic approaches of conceptual nature are introduced. Every data set passes through several tests. Examples of these are: basic laws of nature, prior conceptual knowledge of relevant processes, mass balances, extreme events and empirical cross-correlations between measured parameters. These tests of data credibility are formulated in fuzzy logic that constitutes the unifying method in the adopted approach. Brief introduction to fuzzy logic is also given.