A Systems Approach to Pelagic Ecosystem Dynamics in an Estuarine Environment

Book excerpt: Phytoplankton dynamics in Auke Bay, Alaska, were studied during summer, 1969. Nitrate, chlorophyll a, particulate organic carbon, phytoplankton and zooplankton species composition and hydrographic and meteorological data were collected and analyzed. Nitrate input into Auke Bay through freshwater runoff was negligible. A bloom of Thalassiosira aestivalis formed the spring bloom in Auke Bay in 1969. No phytoplankton bloom occurred during a June period when winds were light and variable. Two major blooms of Skeletonema costatum occurred after periods when Auke Bay surface layers were mixed by wind-induced turbulence. The water column became nearly isothermal after periods of high wind mixing although a pronounced density and halal structure persisted, a consequence of input of freshwater to the bay from Auke Creek. Examination of fecal pellets collected from the bay and results of laboratory grazing experiments suggested that Skeletonema costatum was not grazed by zooplankton living in Auke Bay. Nonlinear ordinary differential equations were written to describe phytoplankton and nitrate dynamics in Auke Bay. The phytoplankton dynamics equation included formulations for time-varying insolation and for time-varying wind mixing coefficients. Formulations for effects of nitrate concentration on the photosynthetic assimilation number and for effects of phytoplankton standing crop on the extinction coefficient of light in the water column were included. The nitrate dynamics equation included a formulation for effects of wind-mixing of nitrate-rich water into the euphotic zone from deeper layers of the water column, as well as a formulation for utilization of nitrate in phytoplankton growth. Computer simulation response of the equations reproduced the bloom pattern observed in the field data with some discrepancies caused by assumptions used in model development. The phytoplankton and nitrate model response was strongly coupled to the pattern of the wind-mixing coefficient, as required by the field data. Variations in model parameters had little effect on phase relations between model response and field data but strongly affected model response magnitude.