Terrestrial carbon contribution to lake food webs: could the classical stable isotope approach be misleading?

Carbon stable isotope analyses have been widely used to estimate terrestrial carbon contribution to lake secondary production. In such approaches, phytoplankton is tacitly assumed as a single, isotopically homogenous source. Such assumption might be valid if (i) zooplankton do not feed selectively on specific algal taxa within bulk phytoplankton, or (ii) although zooplankton do feed selectively, the variability in the delta(13)C values amongst the different algal taxa is small compared with the variability between the delta(13)C values of bulk phytoplankton and terrestrial end-members. In a summer field study of six coastal lakes of British Columbia, Canada, we tested these assumptions using gut pigments and stable isotope measures on zooplankton and particulate organic matter. Results revealed that filter-feeding cladocerans positively selected cryptophytes, shown to be substantially (13)C-enriched compared with bulk phytoplankton and even with the terrestrial end-member. Comparing a classical two-source mixing model and a model that accounted for algal isotopic heterogeneity, we showed that the use of a classical two-source model can result in an overestimation of terrestrial contribution to zooplankton secondary production.