A Spatial Analysis of Wildlife's Ten-Year Cycle
Document Type Article
When this article was written Charles H. Smith was with the Department of Geography, University of Illinois, Urbana, Illinois 61801, and Jerry M. Davis was with the Department of Geosciences, North Carolina State University, Raleigh, North Carolina 27650, USA.
The population cycles that occur in several species of Canadian wildlife have been the subject of numerous studies in the twentieth century. To date, however, the emphasis of attention has been at the biological level. This paper introduces a rationale for a strictly geographic approach to the matter and presents the results of an initial statistical analysis of a spatially grouped time series: Canada lynx furs traded in the nineteenth and twentieth centuries. It is argued that the 10-year cycle represents a potential subject for study by geographers for two reasons. First, excellent sources of temporally and spatially-organized data exist. Second, causal theories now entertained include a strong geographic emphasis which has been poorly explored. In an effort to initiate spatial-temporal analysis of the problem, bivariate spectral time series analysis is applied to the lynx fur data to elicit information on regional and provincial lags in the spread of the population fluctuations across Canada. Earlier results reported in the literature are confirmed: a nodal region exists in west central Canada with the extremes of the country lagging 2-4 years behind. In addition, the results suggest that this nodal region has shifted some 500 or 600 miles to the southeast over the span of the last 100 years. This fact is interpreted as a lead favouring an approach to the study of 10-year cycles emphasizing macro-scale events.