AN INVERSE MODEL ANALYSIS OF PLANKTONIC FOOD WEBS IN EXPERIMENTAL LAKES

We used inverse methods to reconstruct carbon flows in experimental la kes where the fish community had been purposely altered. These analyse s were applied to three years of data from a reference lake and two ex perimental lakes located in Gogebic County, Michigan. We reconstructed seasonally averaged flows among two size groups of phytoplankton, het erotrophic bacteria, microzooplankton, cladocerans, and copepods. The inverse analysis produced significantly different flow networks for th e different lakes that agreed qualitatively with known chemical and bi ological differences between lakes and with other analyses of the impa ct of fish manipulations on food web structure and dynamics. The resul ts pointed to alterations in grazing pressure on the phytoplankton tha t parallel changes in the size and abundance of cladocerans and copepo ds among lakes. Estimated flows through the microbial food web indicat ed low bacterial production efficiencies and small carbon transfers fr om the microbial food web to the larger zooplankton. This study demons trates the use of inverse methods to identify and compare flow pattern s across ecosystems and suggests that microbial flows are relatively i nsensitive to changes at the upper trophic levels.