The Empirical Modeling of an Ecosystem

The authors have endeavored to create a verified a-posteriori model of a planktonic ecosystem. Verification of an empirically derived set of first-order, quadratic differential equations proved elusive due to the sensitivity of the model system to changes in initial conditions. Efforts to verify a similarly derived set of linear differential equations were more encouraging, yielding reasonable behavior for half of the ten ecosystem compartments modeled. The well-behaved species models gave indications as to the rate-controlling processes in the ecosystem.

Modeling Environmental Stress

The word stress when applied to ecosystems is ambiguous. Stress may be low-level, with accompanying near-linear strain, or it may be of finite magnitude, with nonlinear response and possible disintegration of the system. Since there are practically no widely accepted definitions of ecosystem strain, classification of models of stressed systems is tenuous. Despite appearances, most ecosystem models seem to fall into the low-level linear response category. Although they sometimes simulate systems behavior well, they do not provide necessary and sufficient information about sudden structural changes nor structure after transition. Dynamic models of finiteamplitude response to stress are rare because of analytical difficulties. Some idea as to future transition states can be obtained by regarding the behavior of unperturbed functions under limiting strain conditions. Preliminary work shows that, since community variables do respond in a coherent manner to stress, macroscopic analyses of stressed ecosystems offer possible alternatives to compartmental models.