Abstract
Investigation of time trends in contaminant levels has been one of the highest priority tasks within the International Council for the Exploration of the Sea (ICES). Univariate procedures of analysing for temporal variations (provided in the ICES guidelines) are of severely restricted scope. Additional problems, which arise due to multicollinearity when two or more covariates are used, are identified. A procedure which has a more extended scope than the univariate one and which overcomes the multicollinearity situation is presented. Canadian Atlantic cod ( Gadus morhua) data of the eight-year period, 1977–1985, are analysed for time trends by this procedure. Whereas the true mathematical form of temporal variations was not identifiable and will require additional time points, linear trends were significant and provided an effective description of these variations.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
