Eight decades of phenological change for a freshwater cladoceran: what are the consequences of our definition of seasonal timing?

Abstract

1. Changes in the seasonal timing of re-occurring biological events, or phenology, are one of the most widely reported ecological responses to environmental change. Previous studies have demonstrated that plankton populations have shifted their phenology in recent decades but there is a lack of consistency with respect to the phenological metrics that have been analysed. 2. A uniquely long-term perspective on phenological change for Daphnia galeata was adopted, by analysing an eight decade data set (1934-2009) on the seasonal dynamics of this species in the north basin of Windermere, UK. Ten different phenological metrics were used and rates of phenological change derived from each metric were compared. Furthermore, the evidence for effects of spring water temperature, phytoplankton phenology and over-wintering population size on D. galeata phenology was evaluated for analyses utilising each of the ten metrics. 3. Of the ten phenological metrics used, nine showed statistically significant trends towards earlier seasonal timing over the study period. However, rates of change varied widely (3.7 - 6.7 days per decade). Regression analyses showed a consistent effect of spring water temperature (frequently in the month prior to the average timing of D. galeata population development), and phytoplankton phenology on the timing of D. galeata spring population development. However, the amount of variability explained by these drivers differed markedly when different D. galeata metrics were considered. Furthermore, the importance of over-wintering D. galeata population size, and the precise phytoplankton phenological metric related most closely to D. galeata phenology, were inconsistent among models based upon different D. galeata metrics. 4. Hierarchal models, allowing grouping of the different phytoplankton phenological metrics by their conceptual class, showed that the seasonal timing of the phytoplankton peak had the most consistent effect upon D. galeata phenology. These models also confirmed that D. galeata phenology was influenced by temperatures in the month previous to the average timing of population development. 5. Phenological metrics differ mathematically and conceptually. As a result, they are indicative of different population dynamical processes and are influenced by different ecological mechanisms. Analyses that combine information from different phenological metrics will greatly improve mechanistic understanding of the factors influencing phenological change.