Interactive effects of warming and species loss on model Antarctic microbial food webs

1. Predicting the effects of warming and species loss on ecosystems are two significant challenges currently facing ecologists. However, little is known of the interactive effects of these two factors. We hence tested whether or not warming and species loss interact to influence productivity and dissolved nitrogen concentrations in model Antarctic microbial food webs. Food webs, consisting of a uniform bacterial community and mixtures of six, four, two and zero bacterivorous flagellate species, drawn randomly from a pool of six flagellate species isolated from an Antarctic freshwater lake, were grown in soil extract suspension medium held in microcosms for 252 h. Half of the microcosms were kept at 4 °C and half were warmed to 8 °C over the first 36 h and then held at this temperature. 2. After 252 h there were significant interactive effects of flagellate species loss and warming on the abundance of bacterial prey and the concentration of ammonium in the medium: bacterial abundances were reduced by 75% and NH4-N concentrations were doubled in mixtures inoculated with six and four flagellate species, compared with those inoculated with two species, but only in warmed microcosms. This difference in response was apparently largely owing to the absence of Bodo saltans and Spumella putida, species with high grazing activities and growth rates, from most replicates of the warmed two species mixtures. 3. Evidence for an apparent complementarity effect was also found, with B. saltans and Spongomonas uvella growing more rapidly at 4 °C in mixtures of six species than in those of four species. 4. Data from a separate experiment, in which the flagellate species were grown in single-species culture under food-saturated conditions, confirmed that the logarithmic growth rates of B. saltans and S. putida were the highest of each of the six species at both 4 °C and 8 °C. 5. We broadly conclude from our data that random species loss from food webs or communities is likely to alter their responses to environmental change, largely owing to interspecific differences in responses to change.