Times Series Analyses as a Means of Examining Long Term Biological Data Sets

Biologists, and other environmental scientists have traditionally used central tendency statistical analyses (eg. correlative analysis) to describe and quantify the variance between two variables. These analyses are best suited for situations where the dependent and independent variables are a priori defined. In marine systems, where the causal relationship is not always clear, the chance for misinterpretation is introduced. This is particularly so for monitoring programs of resource stocks when investigating the causes of fluctuations or trends in abundance, be they natural (climate), man-made (pollution) or due to fishing pressure. While much can be learned from these data using the traditional statistical approaches, more may be extracted using time series analyses. Time series models are also better suited for forecasting as they identify and partition trends and cycles that are poorly reproduced in linear correlative statistics.