Spectral Analysis of Selected Lake Erie Levels

Abstract

Monthly Lake Erie levels recorded at Cleveland and at Buffalo were analyzed for trend, periodic, and autoregressive components. Trend components might be derived from changing channel conditions, increasing consumptive use of water, from climatic change, or from isostatic adjustments. The periodic component will contain the 12-month seasonal cycle but may also contain other cycles. Autoregression is caused, in this case, by the self-regulating reservoir effect of large lakes. It was found that, as might be expected for a large lake, autoregressive components dominate the time series and that there are also significant periodicities and minor trend components. No trend components were found that might correspond to the effects of a climatic change. Isostatic adjustment was measured between Cleveland and Buffalo but was found to be heavily masked by a periodicity in the differences in levels.