Use of ciliate and phytoplankton taxonomic composition for the estimation of eicosapentaenoic acid concentration in lakes

Abstract

The polyunsaturated fatty acid eicosapentaenoic acid (EPA) plays an important role in aquatic food webs, in particular at the primary producer–consumer interface where keystone species such as daphnids may be constrained by its dietary availability. Such constraints and their seasonal and interannual changes may be detected by continuous measurements of EPA concentrations. However, such EPA measurements became common only during the last two decades, whereas long-term data sets on plankton biomass are available for many well-studied lakes. Here, we testwhether it is possible to estimate EPA concentrations from abiotic variables (light and temperature) and the biomass of prey organisms (e.g. ciliates, diatoms and cryptophytes) thatpotentially provide EPA for consumers.2. We used multiple linear regression to relate size- and taxonomically resolved plankton biomass data and measurements of temperature and light intensity to directly measured EPA concentrations in Lake Constance during a whole year. First, we tested the predictability of EPAconcentrations from the biomass of EPA-rich organisms (diatoms, cryptophytes and ciliates). Secondly, we included the variables mean temperature and mean light intensity over the sampling depth (0–20 m) and depth (0–8 and 8–20 m) as factors in our model to check for large-scale seasonal- and depth-dependent effects on EPA concentrations. In a third step, we included the deviations of light and temperature from mean values in our model to allow for their potential influence on the biochemical composition of plankton organisms. We used the Akaike Information Criterion to determine the best models.3. All approaches supported our proposition that the biomasses of specific plankton groups are variables from which seston EPA concentrations can be derived. The importance of ciliates as an EPA source in the seston was emphasised by their high weight in our models, although ciliates are neglected in most studies that link fatty acids to seston taxonomic composition. The large-scale seasonal variability of light intensity and its interaction with diatom biomass were significant predictors of EPA concentrations. The deviation of temperature from mean values, accounting for a depth-dependent effect on EPA concentrations, and its interaction with ciliate biomass were also variables with high predictive power.4. The best models from the first and second approaches were validated with measurements of EPA concentrations from another year (1997). The estimation with the best model including only biomass explained 80%, and the best model from the second approach including mean temperature and depth explained 87% of the variability in EPA concentrations in 1997.5. We show that it is possible to predict EPA concentrations reliably from plankton biomass, while the inclusion of abiotic factors led to results that were only partly consistent with expectationsfrom laboratory studies. Our approach of including biotic predictors should be transferable to other systems and allow checking for biochemical constraints on primary consumers.