FREQUENCY RESPONSE OF A MARINE ECOSYSTEM 



25 



> 



I 



O- 



O 



cc 

 O 



_i 

 X 



(a) 



3/15 " 4 4 I 4/24 ' 5/14 ' 6/3 ' 6/23 ' 7/13 ' 8/2 ' 8/22 ' 9/11 ' 10/1 

 3/25 4 '14 5 '4 5/24 6/13 7/3 7/23 8/12 9/1 9/21 



SAMPLING DATES 



(b) 



Fig. 2 Ammonia (a) and chlorophyll (b) concentrations in one of 

 several replicate microcosms (No. 2) receiving a 3-month sewage 

 input (between arrows). One representative control, microcosm 

 No. 5, is also shown (adapted from Oviatt, Perez, and Nixon, 1977). 



important for perturbations that temporarily or permanently disrupt 

 feedback pathways (e.g., clear-cutting of forests). 



In practice, defining system boundaries, inputs, and outputs can 

 be troublesome. An exogenous environmental input, by definition, is 

 one that affects, but is not affected by, system state. Solar radiation, 

 rainfall, and temperature fit this definition. An output can be any 

 compartment monitored by the investigator, or it can be transport of 

 material beyond the system boundary, where it no longer modifies 

 system functions. 



Thus the operational definition of ecosystem response has 

 depended to a great extent on what outputs researchers were 

 equipped to measure. As yet, not a great deal of thought has been 

 given to defining which outputs or compartments should be 

 monitored to elucidate important aspects of ecosystem stability. The 

 response linearization hypothesis of Patten (1975) may be correct 

 for some types of outputs but not for others. 



