Epiphytic diatoms as flood indicators

The hydroecology of floodplain lakes is strongly regulated by flood events. The threat of climate warming and increasing human activities requires development of scientific methods to quantify changes in the frequency of short-lived flood events, because they remain difficult to identify using conventional paleolimnological and monitoring approaches. We developed an approach to detect floods in sediment records by comparing the abundance and composition of epiphytic diatom communities in flooded and non-flooded ponds of the Peace-Athabasca Delta (PAD), Canada, that grew on submerged macrophytes (Potamogeton zosteriformis, P. perfoliatus) and an artificial substrate (polypropylene sheets) during the open-water season of 2005. Analysis of similarity tests showed that epiphytic diatom community composition differs significantly between flooded and non-flooded ponds. After accounting for the pond effect, paired comparisons of the three substrates determined that variation in community composition between the artificial substrate and macrophytes was similar to that between the macrophyte taxa. Similarity percentage analysis identified diatom taxa that discriminate between flooded and non-flooded ponds. The relative abundance of 'strong flood indicator taxa' was used to construct an event-scale flood record spanning the past 180 years using analyses of sedimentary diatom assemblages from a closed-drainage pond (PAD 5). Results were verified by close agreement with an independent paleoflood record from a nearby flood-prone oxbow pond (PAD 54) and historical records. Comparison of epiphytic diatoms in flooded and non-flooded lakes in this study provides a promising approach to detect changes in flood frequency, and may have applications for reconstructing other pulse-type disturbances such as hurricanes and pollutant spills.