Factors influencing the temporal coherence of five lakes in the English Lake District

1. The lakes in the Windermere catchment are all deep, glacial lakes but th ey differ in size, shape and general productivity. Here, we examine the ext ent to which year-to-year variations in the physical, chemical and biologic al characteristics of these lakes varied synchronously over a 30-40-year pe riod. 2. Coherence was estimated by correlating time-series of the spring, summer , autumn and winter characteristics of five lakes: Esthwaite Water, Blelham Tarn, Grasmere and the North and South Basins of Windermere. Three physica l, four chemical and two biological time-series were analysed and related t o year-to-year variations in a number of key driving variables. 3. The highest levels of coherence were recorded for the physical and chemi cal variables where the average coherence was 0.81. The average coherence f or the biological variables was 0.11 and there were a number of significant negative relationships. The average coherence between all possible lake pa irs was 0.59 and average values ranged from 0.50 to 0.74. A graphical analy sis of these results demonstrated that the coherence between individual lak e pairs was influenced by the relative size of the basins as well as their trophic status. 4. A series of examples is presented to demonstrate how a small number of d riving variables influenced the observed levels of coherence. These range f rom a simple example where the winter temperature of the lakes was correlat ed with the climatic index known as the North Atlantic Oscillation, to a mo re complex example where the summer abundance of zooplankton was correlated with wind-mixing. 5. The implications of these findings are discussed and a conceptual model developed to illustrate the principal factors influencing temporal coherenc e in lake systems. The model suggests that our ability to detect temporal c oherence depends on the relative magnitude of three factors: (a) the amplit ude of the year-to-year variations; (b) the spatial heterogeneity of the dr iving variables and (c) the error terms associated with any particular meas urement.