The robustness of natural systems

Some important conclusions can be made about these model systems: first, an artificial system comparable in its size and complexity could hardly have this robust structure built into it; second, if grossly damaged, a natural system will usually be reduced to one of its viable subsystems.

Abstract

Although the relation between complexity and stability in natural systems has been widely investigated1, a number of other basic questions remain largely unstudied. For example: can we find fundamental properties which identify a persistent natural system? Are these properties due to the process by which such a system emerges? Is this process—natural selection—also responsible for that robustness which is so generally superior to that of an artificial system? This peculiar robustness of a natural system can show itself in a wide variety of ways; the resistance to shock and efficient nutrient conservation of a climax community2, for example, or the physical stability of proteinoid microparticles3. Can we find other system properties which help to explain this robustness—properties which transcend the particular type of system under consideration? Such a ‘universal’ set of requirements is indeed suggested by the results reported here. In the model studied, the survivor systems turn out to be composed, at every subsystem level, of subsystems each of which is independently viable. These viable subsystems are nested, overlapping, strongly connected and extraordinarily numerous, implying a ‘defence in depth’ for the main system against major collapse. Some important conclusions can be made about these model systems: first, an artificial system comparable in its size and complexity could hardly have this robust structure built into it; second, if grossly damaged, a natural system will usually be reduced to one of its viable subsystems. The components eliminated could, of course, include ones vital to human interests. Full restoration of the damage would not generally be practicable.