fractured earth

Home » 2007 » January

Monthly Archives: January 2007

Resilience Thinking

it is a litany of familiar glooms. the earth is in a mess. once productive agricultural lands are dogged by unforeseen problems — salinisation, erosion, desertification, pollution. lake systems and rivers everywhere are experiencing algal blooms and a host of other problems associated with nutrient oversupply. productive rangelands are turning into unproductive expanses of woody shrubs. many marines fisheries have collapsed. several others are fully exploited and vulnerable.

there is an interesting question around here. right at the end of this paragraph. a part of this global crisis can be traced back to poverty. people, once in reduced circumstances, have no option but to overuse their resource base. another part of it can be attributed to wilful neglect – corruption, perverse subsidies, et al. but that doesn’t explain the whole picture. why are we seeing a lot of these collapses in areas that are under scientific management – like fisheries, plantation forestry, agriculture? what is going wrong?

and so, ladies and gentlemen, i upload my term paper. take a look. and, yes, as you will notice, i am still struggling to write like an academic.

matter begins: The answer lies in the Earth’s natural systems. All of them, organic and inorganic, are complex adaptive systems. They display emergent behaviour – a surprisingly coherent aggregate behaviour that arises as the outcome of decisions taken by the system’s individual components as they act and react to their fellow components and their surroundings. Now, one of the properties of complex adaptive systems is they can exist in multiple equilibriums. That is to say, they have more than one steady state. A disturbance, if large enough to overwhelm the defences of the system, can move it over a threshold into a new steady state. One where its function and structure can be very different. This is how the unpleasant surprises – the lack of recovery in fish populations, lakes abruptly shifting from clear water to persistent murkiness – can be explained.

It is a thought that was first introduced by CS Holling in a seminal paper called ‘resilience and stability of natural systems’. The ecosystems he was studying were throwing up the same pattern — new species were becoming dominant, old ones were disappearing, or were seeing alarming drops in numbers, and not showing any signs of a recovery. that showed no signs of ever recovering. Commenting on that, he wrote: “It’s useful to distinguish two kinds of behaviour. One can be termed stability, which represents the ability of a system to return to an equilibrium state after a temporary disturbance; the more rapidly it returns and the less it fluctuates, the more stable it would be. But there is another property, termed resilience, that is a measure of the persistence of systems and of their ability to absorb change and disturbance and still maintain the same relationships between populations or state variables.”

This property, resilience, is the most exciting thought in environmental systems right now. If we can identify the breaking point of a natural system, then we can surely manage it much better. A resulting question, what determines the resilience of natural systems, is the subject matter of this enquiry.

Understanding regime shifts

Before understanding resilience in more detail, it is important to understand how regime shifts take place.

Briefly, it appears to call for a combination of diminished internal resilience and heightened external forces. It has been concluded that socio-ecological systems (SES) follow a four-stage cycle. Two of them – a growth phase merging into an conservation phase – comprise a slow, cumulative period in which the dynamics of the system are reasonably predictable. As the second phase continues, resources become increasingly locked up and the system becomes less flexible and adaptive, and more vulnerable to external shocks. This, inevitably, leads to a collapse and then, the final phase, reorganisation. This leads on to the growth phase of a new adaptive cycle which might or might not resemble the previous growth phase.

In the first phase, all players are pioneer species that can grow rapidly and withstand harsh conditions. They so ameliorate the conditions so as to allow the entry of less hardy but more competitive species. these species in turn inhibit the pioneers but set the stage for their replacement by even more effective competitors. Throughout this process, biomass accumulates, regulation of biological, chemical and physical processes gets tighter, the species themselves adapt to this increasingly predictable environment.

This tradeoff between efficiency and flexibility has its costs. Ecosystems are shaped by fast, intermediate and slow variables. And, towards the end of the second phase, this system, while efficient, is not capable of handling changes in slow variables. For instance, if all the grass species in a rangeland were replaced by one species with the highest growth rate, production would be higher and more efficient in normal years, but the system itself would be much more vulnerable to droughts, pests, fire and so on. which is what happens. the system might be stable. but that stability is local and narrow.

It is useful to apply that perspective to the study of natural resource management. It begins with an implicit assumption that any changes in the system will be linear and incremental. Next, it tries to optimise output — maximise grain yields, fish catch, timber harvests — by reducing the probability of events that are seen as socially or economically undesirable. In the instance of grain, say, that will take the form of monocropping backed by fertilizer and pesticide use. What all that ends up doing, however, is extend the second phase. And then, when a slow variable changes, the system crashes. It is also pertinent to add here that the rate of change in slow variables is picking up. Human activity has modified several global environmental systems – the atmosphere is just one instance. Thus, adding to the external variability experienced by ecosystems.

As it were, these ecosystems are also more prone to flip anyway. As a system move from one phase to another, the path it follows is determined by three variables – resilience, adaptability and transformability. To understand this, it will be useful to think of a socio-ecological system (SES) as a three-dimensional plate with basins pressed into it. Each “basin of attraction” represents a steady state that the ecosystem could be in. Some of these are desirable, others aren’t. Further, this topology is dynamic – driven by external (environmental, political) or internal (the phase in the adaptive cycle that the SES is in) factors. Over time, as biophysical and social attributes of the system change, the basins get larger or smaller, appear or disappear.

In this “stability landscape, there is a ball, lying in one of the basins. That ball is an SES. As external factors change, it moves around. If disturbed enough, it might roll into another basin altogether. Given that context, resilience can be defined as the capacity of a basin to retain the SES within itself. Then, there is adaptability — the capacity of actors in a system to manage resilience. And, there is transformability. This is still something of a theoretical construct. It refers to the ability to construct a new stability landscape.

Now, in evolved systems that have been subjected to strong selection pressures for millennia, these three aspects of resilience (latitude, resistance and precariousness) are strongly interrelated, with strong feedbacks preventing threshold crossings — a function of co-evolution. That is not the case when it comes to recently developed SES like managed fisheries, agro-ecosystems, have very short co-evolutionary histories. In the absence of strong feedback controls, the likelihood of them crossing thresholds, as we push them towards higher performance and efficiency, is greater.

Take what happened in Lake Michigan. Between 1898 and 1940, the catch of lake trout was high, around 6 million pounds annually. Then, for four years, catches increased noticeably and then abruptly collapsed to near extinction by the 1950s. Reproduction had completely collapsed. At that time, it was suggested that an introduction of lampreys, which fed on the trouts, was to blame. In his paper, Holling disagrees. The lamprey might have played a role, he wrote. But it was just one of several unexpected inevitables that could have happened. Fishing had reduced resilience. And if not the lampreys, some other change, perhaps in the temperature, or in the weather, would have annihilated the trouts.

Or take the Florida Everglades. An internationally recognized wetland ecosystem, these are the lower third of a large hydrologic system that also includes the Kissimmee River and Lake Okeechobee. During the 20th century, the system was partitioned into agricultural, recreational, and conservation land uses by means of a water control system of levees, canals, and pumps. Since then, the nature of vegetation in the Everglades has changed from oligotrophic, e.g., sawgrass, to eutrophic, e.g., cattails. A slow variable phosphorus runoff from agriculture – had changed. An increase in that raised the nutrient count enough to make the everglades a suitable habitat for cattails.

Such are the threats that an increasing part of the world faces today. Erosion of resilience heightens vulnerability. Creating situations where even the tiniest disturbance can push an ecosystem over a threshold into a new regime state.

Measuring resilience

All this, however, leads to altogether trickier questions — measuring and managing resilience. While the conceptual development of the field has been going on for a while now, there is still a lot that we don’t know. Even today, the difficulty with resilience is that we often don’t know how much of it we have until we have lost it. That certainly is one part of the problem. Given that it is not a physical component of the system per se, but an emergent property, resilience has been hard to understand and measure. As Cummings et al say in an exploratory framework for the empirical measurement of resilience, “The multidimensional nature of the concept of resilience makes it difficult to operationalise.”

In the absence of ways to measure resilience, a lot of attention has been focused on surrogates. In this final part of the paper, I look at two of the more promising approaches being tried out. Like slow variables.

The logic of focusing on these is immaculate. Every system has drivers that try and move it from one set of mutually reinforcing structures and processes to another. For the most part, these are triggered by the action of slow variables. It is a tough approach. The number of slow variables that act on an ecosystem can be myriad – from the chemical composition of the soil to atmospheric conditions to the local biota. How do you decide which ones to track?

In their paper, Bennett et al outline a four stage process for identifying them. One, identify the aspect of the system that one wants to make more resilient. Also answer, they say, what changes this system should be resilient to. Next, identify the positive feedback loops trying to maintain this condition, and the negative feedback loops trying to nudge the ecosystem towards an alternative condition. These can be identified by asking questions like: what variables are changing?

What processes and drivers are producing those changes? And what forces control the processes driving those changes? The answers will define the variables that need to be measured, the processes (internal and external to the system) that produce these changes, and the connections between these changes. And, once a system model based on all these findings is ready, resilience surrogates can be identified. The first three focus on the state variable – how far is it from the threshold? How fast is it moving towards or away from the threshold? And what are the external controls or shocks that can change the direction or the rate of change of this state variable? The other two focus on the movement of the threshold itself.

How are these values arrived at? That calls for experimental data. Take the case of a freshwater lake that might be toggling from clearwater to murky. This mainly happens when the amount of phosphorus in the lake increases. The threshold is the concentration of P at which the system shifts. From there, identifying the rest of the surrogates is not too hard.

Another surrogate is redundancy. Quite simply, if an ecosystem has several functional groups, and there are plenty of different actors performing the same functional role, that ensures it is more resilient to sudden changes than systems where a fewer actors are responsible for performing important ecosystem functions. The critical element, add Allen, Gunderson and Johnson, is a combination of within-scale diversity and cross-scale redundancy (in this context, scale can be likened to a trophic level. In such a structure, the loss of a species at a particular scale can be absorbed by a similar species that interacts with the environment at a higher level. An example is provided, they say, by the response of different insectivorous birds, ranging from individual chickadees responding to small-scale infestations to flocks of crows responding to large scale outbreaks. That is an instance of cross-scale redundancy.

An instance of this can be seen in the coral reefs of the Caribbean and the great barrier reef. The former are in trouble. There has been an 80% decline in their area, mainly because of overharvesting, pollution, disease, and climate change. This collapse was preceded by long periods of dwindling fish stocks and increased nutrient and sediment runoff from land. By the 1950s, when modern studies of reef ecology began, a single species of sea urchin was keeping macro-algal blooms at bay. In the 1970s, a disease outbreak spread throughout the Caribbean, and sea urchin numbers fell, precipitating macro-algal blooms that still persist.

Part of the problem is because, in the Caribbean, several critical functional groups are missing or represented by only a handful of species. There are, for example, “no three-dimensional bottlebrush species and just one staghorn (Acropora cervicornis) and one tall, tubular coral (Acropora palmata). These are the dominant habitat-creating functional groups on healthy reefs. Until recently, the two species comprised more than 30–50% of the total coral cover. Today, many areas have effectively lost not only these two species, but also two critical functional groups and two major shallow-water reef habitats. In contrast, the great barrier reef many more species of corals and fishes, and is doing better.”

That gives us another way to measure resilience. If losses in function and changes in the distribution of function within and across scales are quantifiable, they can be incorporated into a measure of relative resilience.

update: that is where the paper ended. a tad abruptly, i grant you. the inevitable consequence of leaving work till too late. was left with no time to add that resilience is still something of an incomplete science, that huge questions remain unresolved. like the slow variables — is it easy or hard to identify them? i was under the impression that, given the myriad forces that can act on an environment, it would be notoriously tough to id the slow variables. and yet, fernandez (his paper is mentioned below in the bibliography) says there never are more than 3-4 (or 5) slow variables acting on an environment.

the other huge question is about operationalising all of this. read about deploying resilience thinking and you will keep coming across references that say things like, “to increase resilience, we must increase the latitude (diameter of the basin in the stability landscape), deepen the basin, push the threshold further back, and so on. huh? are they just extrapolating wildly from the metaphor here, or can such tacks indeed be adopted? and if so, then, how?

finally, there is a question on scale here — at what level does one pitch interventions? if i wanted to make a forest more resilient, i could introduce more species along every trophic level,and in every functional group. but is that prudent? see the post on invasive species

Advertisements