Path analysis of a reservoir ecosystem model

Abstract

A path analysis of a static ecosystem model for a cove in Lake Texoma, Texas-Oklahoma, U.S.A. was performed based on algebraic manipulations of the adjacency (structure) matrix and a corresponding Markov one-step transition (function) matrix. The analysis revealed:(1) overwhelming predominance of indirect over direct effects, (2) very long path lengths required to approach total effects, (3) self effects of compartments diminish with increasing path lengths, (4) interaction effects between compartments initially increase and then decrease with increasing path length, (5) each compartment's structural and functional relationship within the model can be characterized in terms of numbers of incoming and outgoing paths of different lengths and influences associated with these paths, and (6) compartments can be classified according to strengths, patterns and rates of change of structural and functional input-output coupling with increasing path length, and whether they are net producers or consumers of influence in the system network. The significance of these findings for general ecology and reservoir science is discussed, with the conclusion that larger scale models are likely to be required to account realistically for the dominant indirect effects in ecosystems. © 1982 Taylor & Francis Group, LLC.