FREQUENCY RESPONSE OF A MARINE 

 ECOSYSTEM SUBJECTED 

 TO TIME-VARYING INPUTS 



ROBERT L. DWYER,*t SCOTT W. NIXON,* CANDACE A. OVIATT,* 

 KENNETH T. PEREZ,t and THEODORE J. SMAYDA* 

 *Graduate School of Oceanography, University of Rhode Island, Kingston, 

 Rhode Island; and tEnvironmental Protection Agency, Environmental Research 

 Laboratory, Narragansett, Rhode Island 



ABSTRACT 



Recent studies have drawn conclusions about the stability and linearity of 

 ecosystem response to environmental fluctuations, on the basis of analyses of 

 linear mathematical models of several different ecosystems. The validity of 

 model stability estimates can be verified only by an independent method that is 

 not based on assumptions implicit in the formulation of the model. An 

 independent method for deriving frequency response which uses time-series data 

 for an environmental input to and biological response from an ecosystem is 

 presented. Time series of solar radiation, water temperature, ammonia, and 

 phytoplankton abundance, from weekly samples taken in Narragansett Bay, 

 Rhode Island, over a 17-year period, and time series of ammonia and chlorophyll 

 a from a 6-month sewage-perturbation experiment in 150-liter microcosms 

 simulating Narragansett Bay were analyzed by spectral analysis. Spectra for the 

 environmental inputs to the bay ecosystem showed periodicities only at 1 

 cycle/year, whereas the spectrum for phytoplankton abundance showed addi- 

 tional significant periodicities at 2 and 4 cycles/year. These were judged to be 

 harmonics resulting from a nonlinearity in the ecosystem. Cross-spectral analysis 

 of inputs vs. phytoplankton abundance showed no evidence of linear frequency 

 response. Inadequacies in experimental design for the microcosm experiment 

 hindered quantitative estimation of microcosm frequency response, but the 

 advantage of microcosms in allowing full control during input — response 

 experiments is shown. 



The last two decades have been marked by a quantum jump in the 

 apphcation of dynamic systems theory to the analysis of ecosystems. 

 Mathematical modeling of ecosystems has been undertaken for two 



19 



