We research a method to extract information, from sampled data, about the degree and the spread of contamination in a marine environment. The data set consists of heavy metal concentrations in sediments sampled in Liverpool bay. As a benchmark of a "clean" environment, Morecambe bay was used for comparison in a preliminary study. Based on fuzzy set theory, we formulated a mathematical approach and apply it on both environmental regions. Before applying the fuzzy set approach, it was first necessary to investigate subjective (e.g., human error, physical limitation of field equipment) and statistical evidence for the existence of missing and outlying observations. Two statistical tests were used for the detection of outliers: Grubb's test and Chebyshev's inequality. Using the concept of fuzzy sets, an appropriate function was chosen to present the concept of contamination with respect to increasing rates of metal concentrations. Two methods for designing and generating n-place aggregation operators were suggested and a variety of fuzzy aggregation operations were then explored to formulate a set of possible indices of contamination, called here "loading indices". The concept of similarity measures was used to quantify the degree of similarity between the various loading indices. The availability of annually sampled data from Liverpool bay between the years of 1986 until 1992 inclusive, presented us with the opportunity to compare the annual fluctuation of metal contamination levels over seven years. In addition to this, the temporal changes of clean and highly contaminated areas, in a sampled region, were studied. The results were presented in the form of coloured plots, with overlayed contours, separately for each loading index. Grouping of similarity measures between all thirteen loading indices was done using hierarchical relational clustering and a similarity tree. Five distinct groups of indices were found, from which one index representative, from each group, was chosen. These were the Minimum, Maximum, Average, Fuzzy integral and Weighted average. Therefore, by using different aggregation operators, we can highlight different aspects of the spatial distribution of contamination by overall heavy metal loading in a marine environment.
Finally, our analysis lead us to conclude that there is no evidence to support that the area of Liverpool bay, during the period from 1986-1992, was not contaminated overtime with heavy metal loading.