Diatom species variation between lake habitats: implications for interpretation of paleolimnological records

A sample of a sediment record contains diatom species that have grown in disparate habitats and eventually accumulated in a deep part of the lake. The original habitats may differ in substrate, depth location, and availability of resources. Identifying the species characteristic of each habitat should improve our ecological and environmental interpretation of the sediment record by distinguishing habitat specific responses. With this aim, we studied the benthic diatom communities of a deep oligotrophic lake across several habitats. The main source of variation in the diatom composition was the substrate type; particularly, sediment biofilms. Depth was the second factor. The thermocline defined a shift in diatom communities that also included changes in the dominant lifeforms. A third factor was the mesoscale heterogeneity (i.e., rock sides). Although most species were present in many habitats, characteristic species were identified for all the main habitats and used for an improved interpretation of the deep sediment record. Appropriate standardization showed increasing species richness and diversity from epilimnetic epilithic samples to hypolimnetic sediment samples. We estimate that more than 5000 valve counts are required for appropriate comparisons. Consequently, in sediment records with lower counts per sample, one has to amalgamate samples—losing temporal resolution—to achieve reliable analyses of diversity changes over time. Deep sediment samples are representative of the gamma-diversity of the lake diatom metacommunity, which result from the local alpha diversity of the habitats and the beta-diversity of the variability in composition among them. This double source of diversity has to be taken into account when using the sediment record for estimating lake biodiversity changes. On the other hand, we show that an estimation of the spatial (habitat) heterogeneity of a reconstructed environmental variable can be achieved using subsets of species characteristic of each habitat. We demonstrate the procedure by reconstructing the pH fluctuations during the last 200 years in several habitats from a single sediment record. The results are coherent with the expected differences between predominantly trophogenic or tropholithic habitats.