threshold formulation is used to determine which nutrient or light is limiting. 

 Phytoplankton respiration is considered to be the sum of a low maintenance rate 

 plus a term proportional to production and temperature. To maintain constant 

 nutrient s toichiometry within the plankton, nutrient excretion is proportional 

 to respired carbon. 



Zooplankton grazing is a temperature-dependent saturation expression based 

 on total food supply. The expression includes a minimum threshold for feeding 

 as well as size-selective resource partitioning. Feeding and assimilation 

 efficiencies are food-specific constants. Food ingested but not assimilated 

 is egested as detritus. Respiration is a temperature-dependent process that 

 is the sum of a low maintenance rate and a term proportional to the feeding 

 rate. Nutrient excretion is proportional to respiration. 



Transformation rates between detritus, dissolved organic nitrogen, ammonia, 

 nitrite plus nitrate, available phosphorus, and available silicon pools all are 

 temperature dependent and first order. Sinking rates of phytoplankton and 

 detritus are size and density dependent. 



The purpose of the model is to synthesize information on process kinetics 

 into equations describing the rate of change of ecosystem components as functions 

 of the components themselves, of environmental variables, and of group-specific 

 coefficients. Process equations posed by experimentalists over the past several 

 years are used to describe specific biological and chemical processes. For 

 some of the processes more than one theory or hypothesis has been espoused by 

 researchers as a control or regulator. Use of a particular process equation in 

 the overall framework is one way to test that theory or hypothesis in the 

 context of the larger system. As such, the model is merely a tool allowing a 

 quantitative evaluation of isolated process theories in a whole-system context. 



The model considers the lake to be homogeneous horizontally and to be 

 segmented vertically into two layers representing lakewide averaged epilimnion 

 and hypolimnion. A 1-dimens ional , 18-layer heat diffusion model calibrated to 

 temperature profiles measured in Lake Ontario is used to calculate variable 

 depth of the thermocline and average epilimnion and hypolimnion temperatures. 

 Values of the diffusion (or exchange) coefficient between the two layers are 

 then calculated from temperature changes in the two layers and from the extent 

 of thermocline displacement (fig. 2). 



The ecological model was calibrated with data collected during the Inter- 

 national Field Year for the Great Lakes (IFYGL) from March to November 1972. 

 Documentation of seasonal cycles of IFYGL data and model output, as well as 

 detailed discussions of model equations, coefficients, and sensitivity analyses, 

 are presented elsewhere (Scavia 1980a). 



Although the model includes detailed process equations describing the food 

 web and nutrient cycles in Lake Ontario it is still a crude representation of 

 reality. Therefore, before this model was used to examine lake-scale phytoplankton 

 production and phosphorus cycling, its adequacy as a representation of the 

 seasonal dynamics of relevant lakewide averaged properties in Lake Ontario was 

 tested. This was done by comparing measured and simulated properties. The 

 comparisons are best for those properties measured with least uncertainty (i.e., 

 chemical properties) and worst for those properties difficult to measure (i.e., 



61 



