MODELING ENVIRONMENTAL STRESS 11 



calculated in a manner similar to Goh's (1975) vulnerability matrix 

 (Ulanowicz and Kemp, 1978). 



When Kemp and Homer's transformation is performed on 

 energetics data from two comparable marsh ecosystems, one of 

 which is impacted by thermal effluent, the results are striking. 

 Energy flows through the lower trophic levels remain almost 

 unchcQiged, but those through the higher levels fall off drastically 

 under the thermal stress. 



The observations of Kerr (1974), Golley (1974), and Kemp and 

 Homer (1977) lead to the common conclusion that stress tends to 

 result in more-simplified ecosystems. A great deal of debate has been 

 devoted to the converse of this proposition, i.e., that more-diverse 

 ecosystems are more resilient to stress. This proposition is properly 

 macroscopic; i.e., diversity and stability are legitimate community 

 properties. Few of the papers addressing this issue treat stability as a 

 calculatable characteristic of the system, however. Two exceptions to 

 this trend are MulhoUand (1975) and Jorgensen and Mejer (1977). 



Mulholland related the conditional entropy of an ecosystem to 

 its buffering capacity. The conditional entropy, which comes from 

 information theory, is the difference between the now-familiar 

 Shannon— Wiener diversity index and the average mutual informa- 

 tion, i.e., the amount of uncertainty about the distribution of energy 

 from the various sources resolved by observing the behavior of the 

 systems over a given time interval. Rutledge (1974) applied this 

 measure to two short-grass prairie ecosystems, one under low- 

 moisture stress. Surprisingly enough, the stressed system had higher 

 conditional entropy (effective choice of pathways). Mulholland 

 resolved this apparent contradiction by hypothesizing that the 

 "ecological resilience of a system [which has not undergone 

 transition] is maximum when conditions are harshest." 



Jorgensen and Mejer defined the buffering capacity of a 

 freshwater lake as the ratio of total phosphorous in the lake plus 

 sediments to the steady-state value of soluble phosphorous in the 

 lake. They found a tight correlation of this quantity with the exergy 

 of the system (exergy is a measure of the mechanical energy 

 equivalent of the distance from thermodynamic equilibrium). 



In a third study of change in macroscopic variables in response to 

 stress, Lane, Lauff, and Levins (1975) described the changes in 

 several of Levins's (1968) community niche values in response to the 

 eutrophication of a freshwater lake. Mean niche overlap, average 

 competitive success, and mean number of organisms per unit of 

 ecological space if no competitors were present, all increased 

 significantly with increasing nutrient loading. 



