The global system that we define as "ecology" is a complex web that is fragile due to its complexity, since a more interdependent system - or web - is less able to withstand external shocks than simple systems. This kind of systems view used in mathematical ecology is the basis of analysis that has gone on for several decades, and shows that systems have tremendous feed back sensitivities as they become more complex. Thus a grassland or savannah is more robust than a tropical rainforest, which collapses more easily under external stress, human impact, increased temperature gradients, and so forth. This is why forests - a highly complex and evolved system of life - turn into deserts.
A good discussion of mathematical ecosystems is here under "Science of Everyday Things: Ecosystems". It includes a discussion of Jared Diamond's book, Guns, Germs, and Steel as an accessible narrative of how complex systems produce more varied and responsive opportunites for diverse evolutionary strategies and production of new advantages that can dominate the ecologies of simpler systems which are their precursors. At the same time, these more complex systems are not as stable as the simpler ones that they have emerged from.
An included excerpt on this page is from the Encyclopedia of Public Health: Ecosystems talks about systems stress:
Stress from human activity is a major factor in transforming healthy ecosystems to sick ecosystems. Chronic stress from human activity differs from natural disturbances. Natural disturbances (fires, floods, periodic insect infestations) are part of the dynamics of most ecosystems. These processes help to "reset" ecosystems by recycling nutrients and clearing space for recolonization by biota that may be better adapted to changing environments. Thus, natural perturbations help keep ecosystems healthy. In contrast, chronic and acute stress on ecosystems resulting from human activity (e.g., construction of large dams, release of nutrients and toxic substances into the air, water, and land) generally results in long-term ecological dysfunction. Five major sources of human-induced (anthropogenic) stresses have been identified by D. J. Rapport and A. M. Friend (1979): physical restructuring, overharvesting, waste residuals, introduction of exotic species, and global change.
A short review of Systems Ecology on Wikipedia outlines the nature of the study of ecological systems and how humanity is part of this living system and must work within the framework of its laws.
An example of simple and evolving ecologies is famously the Galapagos, which requires a pristine environment to maintain its balance. It's a fundamentally simple system that can't support the demands of mamillian or human life on the islands, but yet has a highly interactive marine and rocky shore which provided Darwin with his clearest resource for his argument for evolution. Most of the third world countries have populations that subsist within a very narrow range their local ecology, and subsequently are more balanced within the system, but not living anywhere near Western standards. In India, this is true of much of the rural subcontinent, with some very interesting eco-village developments being established that do not demand more resources for human habitation than the environment is capable of balancing. More on that in a later post.
Meantime, further resources on environmental monitoring and mathematical ecology are here.