Simulating the response to phosphate additions in the oligotrophic eastern Mediterranean using an idealized four-member microbial food web model

Elsewhere in this volume, observations of the natural microbial food web in the Cyprus Gyre, eastern Mediterranean, and its transient responses both to phosphate additions in situ and to phosphate and ammonium additions when enclosed in microcosm bottles, are reported. We here explore an idealized four-population model of the microbial part of the food web, containing features suggested in these reports to be essential for the observed responses. Such features include a steady state with P-limited growth heterotrophic bacteria and P-limited or N/P co-limited growth of phytoplankton a mechanism for luxury consumption and nutrient storage in the osmotrophs (phytoplankton and bacteria), a supply of labile organic carbon substrates in excess of bacterial carbon demand, a relatively small excess of bio-available nitrogen, and an assumption that heterotrophic bacteria are superior to phytoplankton in competing for dissolved organic nitrogen. From a P-limited steady-state dominated by heterotrophic organisms, the model responds to the in situ phosphate addition of the Lagrangian experiment with a decrease in chlorophyll, an increase in bacterial production and in bacterial biomass, and a decrease in uptake potential for phosphate. These modeled responses at the osmotroph level are qualitatively and quantitatively comparable to those observed, while detailed comparison of model and observations at the predator level appears more difficult. The model is also able to explain main traits of the dynamic patterns observed in microcosm experiments, both when different concentrations of phosphate were added to previously unperturbed water, and when water collected inside the patch of the Lagrangian experiment was enclosed and supplied with ammonia. We conclude that the idealized model contains sufficient elements to capture a useful first-order approximation to a presumably quite complex microbial food web. In this model, predator growth responds not only to food quantity, but also to food quality (stoichiometry). From steady states where also zooplankton are P-limited, the model thus has a potential for much more rapid zooplankton response to a phosphate-pulse than possible in models with fixed organism stoichiometry. The potential of transmitting a signal to the copepod level via food composition, rather than only via food abundance, is discussed. Implicit in both model and observations is a requirement of a large, unmeasured, production of degradable organic substrates for bacterial growth.