Appendix S1 The relationship between number of zooplankton individuals counted per sample and the number of species identified in four ELA lakes. Numbers on the panels indicate the lake number. Richness was calculated by the method of Dodson (1992), which excludes littoral species and is therefore slightly different from the one we use. A mixed-effects model with only an intercept and a random term for “Lake” provided a superior fit to models with a fixed effect for the number of individuals with either the same (= 0.0002) or different (= 0.0042) slopes for each lake. Thus, lakes differ in species richness and the number of animals counted had no discernable effect on the estimate of species richness.

Appendix S2 (a) The correlation between average daily and annual zooplankton species richness. (b) The best-fit models for average annual richness on the four time scales, and for all variables on all time scales.

Appendix S3 The scaling between the mean and standard deviation for each of the ten limnological variables on a log-log scale. The top rows are the relationships between the annual mean and the standard deviation among monthly averages, the bottom rows are for the long-term mean and the standard deviation of annual averages. The parameter values shown are for the mixed-effects models with the following random effects: none, Lake, or Lake nested within Dataset. A likelihood ratio test was used to select among the three models with different random terms included. The parameter values shown are for the fixed effects of the log(mean(x)) on log(StDev(x)). A slope of 1 indicates that a multiplicative model is appropriate form of variance decomposition, while an additive model is appropriate for a slope of 0 (Chatfield 2004). The slopes were significantly greater than zero in every case except for the relationship between the long-term mean and annual standard deviation for pH and Surface Temperature. We therefore applied a multiplicative variance decomposition model for all of the variables in our dataset.

Appendix S4 Summary statistics on environmental variables. Each column indicates the mean and 95% bootstrapped confidence intervals for each variable. Units are given in the text.

Appendix S5 Comparison of the best subset models containing between one and 11 of the variables included in any of the final models for each time scale (in Table  2). The first and second order terms for interannual variability in TP were removed together from the model. The models are ranked from best (top) to worst (bottom) based on AIC (indicated in the first column). Each row indicates a model, and the shaded rectangles indicate the variables included in the model. The colour indicates the time scale as shown by the top row. The number of models containing a variable is an indication of its importance as a predictor of average daily zooplankton species richness after accounting for variance explained by all of the others.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

ELE_1438_sm_appendixS1-5.doc173KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.