Ecosystem Reality – Modelling: Reflections Pt 5

The second advance produced by our series of studies of large scale ecosystems was a set of deep case studies with modeling efforts that could be used in a comparative analysis of ecosystems behavior and ecosystems management. Those examples included some 20-30 examples of crisis-ridden histories of forests, fisheries, agriculture, human diseases and water resource development.

One theoretical study suddenly helped significantly, when my eyes were opened to the essential way to understand and display the (relatively simple) causes of complex behavior (Ludwig, Jones and Holling, 1978). It was Don Ludwig and Dixon Jones who taught me the way, using the essence of qualitative differential equation theory.

It all started when Don took a half page I wrote explaining the essence of the causes of forest changes mediated by spruce budworm in eastern Canada. He then turned that into a coupled, three differential equation model that expressed the interacting dynamics of budworm, foliage and trees. Meanwhile Dixon, with help from Bill Clark and I, had been developing the big simulation model of the system that emerged out of a series of workshops with the scientists and policy people in New Brunswick. As part of our philosophy of economy in modeling, I had been careful to leave out the effects of avian predation, relying on an eventual check with measured behavior of the whole system in nature to tell us what essentials we had missed. When we discovered that the behavior of the simulation model simply did not match the field behavior, we used it and our ecological knowledge to discover the “missing process”, as a kind of interactive, diagnostic procedure.

The missing piece turned out to be one with certain specific nonlinearities at low densities of budworm and low volume of foliage. The only process we could discover to fill the bill was predation by the 35 different species of insectivorous birds. That linked us back to my earlier set of predation discoveries and we added the effect using the predation equations and parameter data from the field. The effect added progressively stronger predation as budworm densities rose from low levels, and faded thereafter as budworm populations increased- that is, a domed shaped response. Since the densities of birds were essentially constant, that predation effect gradually weakened as the forest aged and the increasing volume of foliage dispersed the searching by birds. The result was periodic outbreak of the insect in older forests.

When these same bird predation effects were then added to Don’s differential equations, that too began to reflect what occurred in nature. So it was a beautiful example of the power of linking three key methodological concepts; Don’s qualitative differential equation approaches, Dixon’s scientifically infused simulation modeling and my general process analysis modeling (Ludwig et al. 1978). The advance led to a clear way to understand and compare the 20-30 examples of complex ecosystem behavior in totally different kinds of situations (Holling, 1986).

The results appeared in the second paper discovered by the students i.e. in Holling 1986. It is a chapter in the first (and maybe only) significant book that deals with sustainability in a fundamental, interdisciplinary way. That book was Bill Clark’s inspiration and creation. My chapter for the first time developed the theoretical discoveries emerging from the comparison of those ecosystem studies. Some of the key features of ecosystems popped out: e.g. there had to be at least three sets of variables, each operating at qualitatively different speeds. There was an essential interaction across scales in space and time covering at least three orders of magnitude. Non-linearities were essential. Multi-stable states were inevitable. Surprise was the consequence.

And a puzzle emerged concerning what seemed to be an inevitable pathology of resource management. In case after case, the same pattern appeared. An economic or social problem was identified as being present or looming in the near future. It was then narrowly defined and treated in a least cost manner for fast corrective response. Then, unknown to all, the system evolved.

First, the problem seemed to disappear. Budworm outbreak populations became controlled, forest fires were suppressed before spreading, water was stored and irrigation became possible for agriculture, fisheries were augmented with hatchery stocks, and so on. Second, industry expanded: pulp mills, tree harvesting, agriculture, fisheries and with that, regional economic and social development.

Third, slow, unappreciated changes occurred that meant that resilience was restricting, was declining. In most cases, the resilience declined because spatial heterogeneity shifted to a more homogeneous state. A “spark”, once initiated, could therefore spread up scale. That is, conditions for outbreaks in healthy forests spread, forest stands became more homogeneous in age and became fuel rich, salt accumulated in soil as soil water levels rose, natural fish stocks gradually went extinct leaving fisheries precariously dependent on a few enhanced stocks. All became disastrous surprises waiting to happen.

Slowly decreasing resilience faced fast increasing economic and social dependencies that made retreat and redesign extremely difficult. Working with nature was rarely conceived. Instead, the response to correct the surprises, started or continued a sequence that maintained the evolving system with more and more costs. The classic example of that is the Everglades, which, after over 80 years of four crises, now is launched into an eight billion dollar restoration, with little active adaptive design. In contrast, the Columbia River system is deeply involved in a policy that indeed does exploit natural forces in an interesting adaptive scheme.

Other examples of “command and control”, of passive and active adaptation in regional social/ecological systems have been recently described in Olsson et al 2006, leading to a set of considerations and actions we identified for successful transformation toward adaptive governance,

This universal pattern represented one of the social traps later discovered as a potential for panarchies. Subsequent avoidance of the trap can occur through learning and actions to enhance resilience by reintroducing spatial heterogeneity at appropriate scales. But often the remedial responses simply continued and extended the process, protected by gradually increasing investments of money to monitor, subsidize and control.

Adaptive cycle

And I used the paper to present the first big theoretical synthesis. That was the place where the Adaptive Cycle was first described and presented. That is, there are four components of change in ecosystems, the traditionally known and slowly evolving exploitation and conservation phases and the newer, fast, unpredictable creative destruction and renewal phases. The first two are when capital and skills are slowly accumulated, but resilience is typically gradually lost. The last two are when unpredictability explodes, capital is freed for other roles and novelty can become implanted. Moreover, those same four components seemed to provide a general metaphor for all systems, and examples were discussed from economics, technology, institutions and psychology. In fact, I discovered that the creative destruction phase had already been posited decades earlier by an economist, Joseph Schumpeter, for international businesses. Maybe economists were not all so narrow!


Holling, C.S. 1986. The resilience of terrestrial ecosystems; local surprise and global change. In: W.C. Clark and R.E. Munn (eds.). Sustainable Development of the Biosphere. Cambridge University Press, Cambridge, U.K. Chap. 10: 292-317.

Holling, C.S. and A.D. Chambers. 1973. Resource science: the nurture of an infant. Bioscience 23(1): 13-20.

Ludwig, D., D.D. Jones and C.S. Holling. 1978. Qualitative analysis of insect outbreak systems: the spruce budworm and forest. J. Animal. Ecol. 44: 315-332.

Olsson, P., L. H. Gunderson, S. R. Carpenter, P. Ryan, L. Lebel, C. Folke and C. Holling 2006. Shooting the Rapids: Navigating Transitions to Adaptive Governance of Social-Ecological Systems. Ecology and Society 11 (1): 18. [online] URL:

Walters, C.J. 1986. Adaptive Management of Renewable Resources. MacMillan, New York.

Walters, C., and Martell, S. 2004. Fisheries Ecology and Management. Princeton Univ. Press, Princeton, NJ.

2 thoughts on “Ecosystem Reality – Modelling: Reflections Pt 5”

Leave a Reply

Your email address will not be published. Required fields are marked *