| Popis: |
A problem common to all large area analysis is how to deal with large data sets which have evolved over a period of years. Such data sets often have inconsistent (in time and space) parameters, sampling location, sampling methodology, and analytical procedures. This is the case for the nearshore water quality data for the Canadian Great Lakes which have been collected by the Ontario Ministry of the Environment since 1967. As a result, a methodology has been developed to reduce (filter) the data set to a manageable size and provide a substantive summary of nearshore water quality for the period 1967 to 1973. Initially, the data can be divided temporally to reflect major changes in laboratory techniques and sampling methodology (timeframes). Subsequently, the data base is partitioned into geographical regions reflecting sampling station configurations and subjective criteria regarding expected homogeneous water quality. Surface and subsurface data are distinguished among limnologically defined seasons. Descriptive Statistics are generated for each cell of the 2 × 3 timeframe matrix (rows = surface and subsurface, columns = spring, summer, and fall). One and two-way analysis of variance techniques are used to determine Significant Statistics for each timeframe matrix. These are assumed to be conservative estimates of average water quality for a given parameter within a particular region and for specified time periods and depths or combinations thereof. Selected water qualtiy data for Lakes Erie and St. Clair and Lake Ontario, including Toronto and Hamilton Harbours and the Bay of Quinte, are presented and discussed as examples of the usefulness of these data for providing (i) period of record composite comparisons among water bodies; (ii) spatial comparisons within any one time frame and for any selected cell; and, (iii) temporal trends determined by weighted linear regression for a selected cell through the three time frames. It is concluded that this data base is useful for evaluating broad spatial and temporal trends and assessing the effect of remedial programs and warrants more comprehensive analysis in conjunction with loadings information and open lake data. |
