Importance of Data Mining in the Assessment of Anoxic Problem from Estuarine Monitoring Data

-A data retrieval procedure is used to process observed dissolved oxygen (DO) concentrations in the Chesapeake Bay to obtain information on the intensity or degree of anoxia, and to explore optimal DO-thresholds to define the degree of anoxia. The method consists of a series of processes, such as data interpolation, pattern mining, feature transformation, uncertainty management, and verification. A new metric, “anoxic intensity”, is an outcome from the data mining, which addresses the degrees of anoxia and improves over the traditional way using “anoxic volume” as the key metric of anoxia. The data retrieval method described overcomes the current methodology shortfalls that may lead to data misinterpretation. This paper emphasizes the importance of data mining for data utilization.