A graph-theoretic analysis of relationships among ecosystem stressors

The development of strategies to reduce risk to ecosystems has been propose d as a centerpiece for environmental protection. An essential prerequisite to the development of such strategies is the identification and evaluation of those factors (called stressors) which cause stress to ecosystems. Stres s refers to external forces which threaten an ecosystem's viability or inte grity, that is, the ability to maintain its organizational structure and fu nction. New analytical tools are needed to assist with the difficult task o f ecosystem risk assessment. These tools must be sophisticated enough to en able an analyst to do more than focus on single stresser-single assessment endpoint pairs. By an assessment endpoint is meant an environmental value o r ecosystem service to be protected or restored. In this paper we develop a nd describe analytical tools based on the mathematical subdiscipline of gra ph theory which can be used to gain insights into the interactive relations hips which exist among the stressors themselves. The manner in which these stressors are connected to each other and interactive paths among the stres sors which tend to augment or diminish the impacts of certain stressors upo n the ecosystem can be delineated. When used in conjunction with analytical methods which deal more directly with the impacts of stressors upon the ec osystem itself, the results derived with the aid of these tools can be used to more clearly identify clusters of stressors which should be the focus o f environmental protection activities. Applications of these analytical too ls to Green Bay, Lake Michigan and the St Croix National Scenic Riverway ec osystems in the United States are discussed. (C) 1999 Academic Press.