Search

Analysing and interpreting long-term phytoplankton time series present a number of challenges, arising from potential historical inconsistencies in data collection and taxonomic identification of organisms. In a previous paper, Pomati et al. (2012) found a remarkable increase in phytoplankton diversity that coincided with oligotrophication and warming of Lake Zurich over a 32-year period. These findings were recently challenged on the basis of potential biases in detection limits and taxonomic classification over the time series (Straile, Jochimsen & Kümmerlin, 2013). We agree that being cautious with long-term phytoplankton data series is extremely important, but argue that the increase in richness detected in Lake Zurich cannot be due only to methodological bias., Following additional analysis of the Lake Zurich phytoplankton dataset, we found that the shift in taxon detection limits reported by Straile et al. (2013) is not supported by the data and stems from a rounding error in the calculation of density in the dataset available to those authors. We found a decline in the proportional abundance for common taxa, an increase in the annual prevalence of taxa and reduced community turnover over the time series. Taken together, the data clearly indicate a trend of decreasing dominance, while more taxa coexist simultaneously. We also argue that the taxonomic classification has been robust (at least at the family level) and propose a diagnostic plot that can help detect an unbiased signal of change in plankton richness through time., Straile et al. (2013) observed perfect synchrony in species occurrence between Lake Zurich and nearby Lake Walen. While we agree that such perfect synchrony in community composition can reflect a bias in the database compilation, it can also be an important ecological signal of changes in regional species pools that deserves further analysis., We conclude that the results of Pomati et al. (2012) are robust and not substantially undermined by the criticisms of Straile et al. (2013). More generally, it is indeed possible to extract meaningful signals of biodiversity change from long-term phytoplankton monitoring datasets, provided there is a clear understanding of how the data have been sampled, recorded and analysed over the history of the time series. [ABSTRACT FROM AUTHOR]