“…Similar to a RDA, the first canonical AEM axis represents the dominant pattern in the data and the second AEM axis the next most important pattern. In essence, AEM modelling facilitates the identification of different temporal scales (i.e., from seasonal, to decadal to millennial scales) in complex time series data (Baho, Futter, Johnson, & Angeler, ) and has been widely applied to different ecosystems and taxa (invertebrates: Angeler, Allen, & Johnson, ; birds: Kampichler et al, ; paleolimnology: Spanbauer et al, ; microscopy‐based phytoplankton: Angeler, Allen, & Johnson, ; insects: Blanchet et al, ), well supporting its explanatory power. The fitted site scores (i.e., linear combination of explanatory variables; lc scores) of the first and second canonical AEM axis were plotted along time to identify the temporal variability of communities at different temporal scales.…”