STAGE DURATION ESTIMATION FOR CALANUS POPULATIONS, A MODELING STUDY

Population dynamics of Calanus finmarchicus have been modelled using very finely divided representation of the stock according to age-within-stage, in the manner of models developed by C. S. Davis, A. Sciandra, F. Carlotti and others. A key assumption of the model is that development rate is relatively insensitive to food-limitation, so that stage duration can be represented by a temperature function alone. We used the Belehradek function for this purpose, noting that better data are needed for fitting its parameters. The model closely simulates the timing of stage progression and relative stage abundances of C. finmarchicus in the Malangen fjord system (northern Norway) during the winter-spring generation. The model is sensitive to the resolution of the age-within-stage division, but it is fully stable at 0.5 h increments. Modifications of the model simulated several methods for field estimation of stage duration in Calanus (or other highly seasonal copepod populations). A method based on changes in stage proportions (the 'Heinle graph' method) is biased by confounding of the effects of developmental progress and mortality on stage proportions. However, the model shows that the bias is mild and the method gives useful estimates of stage duration. Simulation of a method based on molting rate determinations ('Kimmerer experiments') showed its unsuitability for highly seasonal stocks in which stage composition is changing rapidly. Differences in C. finmarchicus survivorship schedules between constant and continuously increasing temperatures were simulated, showing that such differences in pattern are critical to annual survival and stock production. Simple methods for fitting mortality rates to data using the model were extremely sensitive to sampling noise. More complex methods may succeed but remain to be developed.