All posts by Steve Carpenter

Dead Ahead: Similar Early Warning Signals of Change in Climate, Ecosystems, Financial Markets, Human Health

What do abrupt changes in ocean circulation and Earth’s climate, shifts in wildlife populations and ecosystems, the global finance market and its system-wide crashes, and asthma attacks and epileptic seizures have in common?

According to a paper published this week in the journal Nature, all share generic early-warning signals that indicate a critical threshold of change dead ahead. Cheryl Dybas writing for covers a new paper on “Early Warning Signals for Critical Transitions” (Nature, 3 Sept 2009, 461: 53-59).

In the paper, Martin Scheffer of Wageningen University in The Netherlands and co-authors found that similar symptoms occur in many systems as they approach a critical state of transition.

“It’s increasingly clear that many complex systems have critical thresholds–‘tipping points’–at which these systems shift abruptly from one state to another,” write the scientists in their paper.

Especially relevant, they discovered, is that “catastrophic bifurcations,” a diverging of the ways, propel a system toward a new state once a certain threshold is exceeded.

Like Robert Frost’s well-known poem about two paths diverging in a wood, a system follows a trail for so long, then often comes to a switchpoint at which it will strike out in a completely new direction.

That system may be as tiny as the alveoli in human lungs or as large as global climate.

“These are compelling insights into the transitions in human and natural systems,” says Henry Gholz, program director in the National Science Foundation (NSF)’s Division of Environmental Biology, which supported the research along with NSF’s Division of Ocean Sciences.

“The information comes at a critical time–a time when Earth’s and, our fragility, have been highlighted by global financial collapses, debates over health care reform, and concern about rapid change in climate and ecological systems.”

It all comes down to what scientists call “squealing,” or “variance amplification near critical points,” when a system moves back and forth between two states.

“A system may shift permanently to an altered state if an underlying slow change in conditions persists, moving it to a new situation,” says Carpenter.

Eutrophication in lakes, shifts in climate, and epileptic seizures all are preceded by squealing.

Squealing, for example, announced the impending abrupt end of Earth’s Younger Dryas cold period some 12,000 years ago, the scientists believe. The later part of this episode alternated between a cold mode and a warm mode. The Younger Dryas eventually ended in a sharp shift to the relatively warm and stable conditions of the Holocene epoch.

The increasing climate variability of recent times, state the paper’s authors, may be interpreted as a signal that the near-term future could bring a transition from glacial and interglacial oscillations to a new state–one with permanent Northern Hemisphere glaciation in Earth’s mid-latitudes.

In ecology, stable states separated by critical thresholds of change occur in ecosystems from rangelands to oceans, says Carpenter.

The way in which plants stop growing during a drought is an example. At a certain point, fields become deserts, and no amount of rain will bring vegetation back to life. Before this transition, plant life peters out, disappearing in patches until nothing but dry-as-bones land is left.

Early-warning signals are also found in exploited fish stocks. Harvesting leads to increased fluctuations in fish populations. Fish are eventually driven toward a transition to a cyclic or chaotic state.

Humans aren’t exempt from abrupt transitions. Epileptic seizures and asthma attacks are cases in point. Our lungs can show a pattern of bronchoconstriction that may be the prelude to dangerous respiratory failure, and which resembles the pattern of collapsing land vegetation during a drought.

Epileptic seizures happen when neighboring neural cells all start firing in synchrony. Minutes before a seizure, a certain variance occurs in the electrical signals recorded in an EEG.

Shifts in financial markets also have early warnings. Stock market events are heralded by increased trading volatility. Correlation among returns to stocks in a falling market and patterns in options prices may serve as early-warning indicators.

“In systems in which we can observe transitions repeatedly,” write the scientists, “such as lakes, ranges or fields, and such as human physiology, we may discover where the thresholds are.

“If we have reason to suspect the possibility of a critical transition, early-warning signals may be a significant step forward in judging whether the probability of an event is increasing.”

Co-authors of the paper are William Brock and Steve Carpenter of the University of Wisconsin-Madison, Jordi Bascompte and Egbert van Nes of the Consejo Superior de Investigaciones Scientificas, Sevilla, Spain; Victor Brovkin of the Max Planck Institute for Meteorology in Hamburg, Germany; Vasilis Dakos of the Potsdam Institute for Climate Research in Potsdam, Germany; Max Rietkerk of Utrecht University in The Netherlands; and George Sugihara of Scripps Institution of Oceanography in California.

The research was funded by the Institute Para Limes and the South American Institute for Resilience and Sustainability Studies, as well as the Netherlands Organization of Scientific Research, the European Science Foundation, and the U.S. National Science Foundation, among others.

A Novel for the Long Now

Imagine that we wanted our descendants to persist for 10,000 years. How could we help that to happen? This question motivates most of the research on resilience, as well as initiatives such as Clock of the Long Now <> and policy-oriented initiatives such as the Millennium Ecosystem Assessment <>. Many insights about resilience have come from research on native cultures, such as an influential volume by Berkes, Colding and Folke on Navigating Social-Ecological Systems, and many other works cited in this blog and in the journal Ecology and Society.

In Girl With Skirt of Stars, Jennifer Kitchell draws a sharp contrast between modern society and a culture that has occupied the southwest of North America for thousands of years.

Lilli Chischilly is a Navajo lawyer with a full brief of problems. Someone arranged mutilated carcasses of sibling coyotes on the hood of her battered Dodge pickup truck – no doubt a message, but of what?. Her old flame has returned to Indian Country, yet somehow he is connected to an inexplicable murder. Then she is assigned to escort a powerful politician through the Grand Canyon for a publicity stunt – obviously a set-up for a hydropower dam in a national landmark that will drown sites sacred to her people. In the shadowy background a mysterious sniper, motivated by a century-old massacre, stalks the politician. This meticulously-crafted debut novel weaves Navajo ethnography, sexual tension, political power, and the beauty of Grand Canyon country into a fast-paced story. Kitchell’s voice is confident, reflecting her deep knowledge of Navajo culture and the physical beauty of the Southwestern US.  The novel’s ending foreshadows more stories to come. I’m eager to read them.

At one moment in the novel, Lilli brings the politician into an ancient cave with petrographs that hold the key to a culture that can last for ten millennia. Will it be drowned by the dam? This encounter with deep-time resilience is the key to the novel, and perhaps the key to human persistence through the current environmental crisis.

This novel is fun to read. It evokes questions that are central to resilience thinking. It will appeal to students who are interested in natural history, ancient cultures, and connections of native people to modern life. Once you open it you will read it all the way through.

Scenarios and Resilience

People or organizations can focus their effort on a narrow goal, or they can diversify the uses of resources to explore and innovate. It is hard to do both at the same time. This pattern arises in politics as well as in corporations, agencies or academic institutions. When politics of democracies begin to lock into a stationary state, party positions are caricatures, messages are simplistic, campaigns are tightly scripted, media events are rigidly coordinated, and big donors demand loyal candidates. These conditions do not encourage broad, creative, inventive discussions of the most important problems of the day. Such a political environment seems hopelessly incapable of addressing the multiple shocks of the present – the credit crisis, sharply rising prices of energy and food, shortage of arable land, declining capacity of ecosystems to produce the goods that people need, and the complex challenges of climate change, among others. These shocks are unprecedented, so the solutions are novel – the kinds of solutions that cannot emerge from gridlock politics.

Nonetheless, people need answers to complex questions. In a recent global survey, respondents were asked to identify the questions that were most important to them. Questions were then ranked in order of the number of respondents who identified them as important. All of the top-ranking questions were deeply complex. What does sustainability look like? How must humans adapt to survive the changes of this century? What economic structures best support a shift to sustainability? How can we re-invent politics so people feel that they have a voice? What kind of leadership does the world need now?

Complex questions can be addressed by scenarios – sets of stories about the future, derived from collaborative processes and models, designed to integrate diverse perspectives. The scenarios of the Millennium Ecosystem Assessment are a recent example.

Scenarios are a way of building resilience – the capacity to maintain useful features of nature and society, while inventing and implementing transformations to new ways of living. In a recent talk at Resilience 2008 I discussed some of the connections between scenarios and resilience. To break out of traps, people need positive stories of what the future could be, and blunt warnings of dangerous paths. Scenarios provide such motivating visions. Moreover, the process of scenario-building itself may create connections that enable transformation. Scenario projects form networks of people in settings that promote playful, inventive thinking at the margin of formal politics. The scenarios, the insights, the people, or the networks themselves are capable of infiltrating wider thinking, and thereby contributing to change when the conditions are right.

What could expand the use of scenarios to build resilience? We need more people trained in relevant skills such as collaboration, rapid prototyping, flexible fast modeling, synthesis, and use of art, music, science and stories together. Courses exist and a sizeable literature is available. Yet the best way to learn scenarios is by doing. Why not try scenario thinking the next time you face a complex problem with long-term consequences?

Endless Forms Most Beautiful

The closing words of Darwin’s Origin of Species are probably the best known passage in all of biology: “There is a grandeur in this view of life, with its several powers, having been originally breathed into a few forms or into one; and that whilst this planet has gone cycling on according to the fixed law of gravity, from so simple a beginning endless forms most beautiful and most wonderful have been, and are being, evolved”.

Darwin deliberately contrasted predictable cycles with the endless change of biological systems. How does novelty arise in evolution? Genetic variation alone seems insufficient. How do the nearly-identical genomes of bonobos, chimpanzees and humans give rise to such different organisms? The answer, as Darwin himself suspected, lies in the dynamics of development of organisms from embryos to adults.

In Endless Forms Most Beautiful (Norton, 2005), University of Wisconsin biologist Sean Carroll explains the mechanisms of animal development with breathtaking clarity. Some of the key discoveries come from his own work on developmental differentiation of the striking patterns of butterfly wings.


Carroll emphasizes three hallmarks of evolutionary innovation: (1) evolution works by modifying structures and processes that are already present, not by creation de novo; (2) structures and processes in organisms are multifunctional and partially overlapping, opening the possibility for differentiation through specializing or reorganizing the division of labor; and (3) organisms are modular, opening the possibility of changing the number of modules or the functions of individual modules. The underpinnings of modularity are modular geography of embryos, and modular genetic switches which allow evolutionary change to occur in one part of the organism, independent of other parts. He writes:

We have seen that insects, pterosaurs, birds or bats did not invent wing genes, butterflies a spot gene, or humans a bipedalism or speech gene. Rather, innovation in all these groups has been a matter of modifying existing structures and teaching old genes new tricks.

The key to innovation at the genetic level is the multifunctionality of tool kit genes. The multifunctionality of tool kit genes stems from their deployment at different times and places through batteries of genetic switches. In this manner, a protein such as Distal-less can act at one time to promote limb formation, and at another to promote eyespot development. The protein made each time is identical, so the difference in function is due to its action on different switches in these different contexts.

At an anatomical level, multifunctionality and redundancy are keys to understanding the evolutionary transitions in structures . . .The history of these structures also illustrates how “endless forms” evolve through cycles of invention and expansion. New structures open up new ways of living. The insect wing led to the evolution of dragonflies and mayflies, butterflies and beetles, fleas and flies, and more. The expansion of these groups was catalyzed in turn by a cycle of innovation and expansion by making modifications to the wings or body plan . . .

Why are existing body parts and genes the more frequent pathway to innovation? This is a matter of probability. Variation in existing structures and genes is more likely to arise than are new structures or genes, and this variation is therefore more abundant for selection to act upon.

The tool kit genes central to evolutionary innovation have been conserved through about 500 million years of animal evolution, and are found across the animal kingdom. Paradoxically, the fundamental mechanisms of evolutionary innovation have been strongly stabilized over eons of time.

Punctuated Equilibrium in Environmental Policy

Readers familiar with panarchy theory will find a rich set of relevant examples in a new book edited by Robert Repetto, Punctuated Equilibrium and the Dynamics of U.S. Environmental Policy.

In Chapter 2, Frank Baumgartner explains how U.S. environmental policy shows is static for long periods of time, reflecting stable institutional structures, shared understanding of goals, and balance of power among competing interests. Occasionally, however, there are bursts of innovation as public policies are radically restructured. These rare but crucially important bursts of innovation occur across a range of scales, from local to national. Baumgartner makes his case using statistical case histories of policy dynamics.

William Brock, in Chapter 3, explains social and economic mechanisms that cause long periods of stasis interrupted by bursts of enormous change in environmental policy. Brock uses minimal models grounded in well-established social and economic phenomena. The remainder of the book develops case studies in depth.

Some cases have undergone radical change: management of water in California, certain marine fisheries, and timber in the Pacific Northwest. Other systems seem locked in traps: greenhouse gas and climate policy, vehicular fuel economy standards, and livestock grazing on public lands. History suggests that these traps will eventually be broken.

In his introductory chapter, Repetto summarizes the positive and negative feedback mechanisms that underlie punctuated equilibrium. He writes:

Though they [the feedback mechanisms] are fully capable of explaining the observed patterns of stability and abrupt change, their workings are difficult to predict in particular policy struggles because of their complex interactions. The infrequency of policy breakthroughs suggests that most efforts to bring them about will fail. Entrenched interests and ideology will retain their dominance; challengers will be unable to gather sufficient resources, attention, and momentum. Nonetheless, such failed efforts may build a foundation for later success when conditions are more favorable by undermining the prevailing policy image, by mobilizing new interests, and by forming new coalitions. Even knowing that the odds are long, effective policymakers continue to work on their issues in order to be ready and primed when opportunities arise. Timing is crucial . . . . Across the broad range of resource and environmental policy issues, only a few, if any, are likely to have potential for significant change at any particular time. The ability to discern which ones these are is a vital strategic skill.

This pattern is well known to researchers familiar with adaptive cycles and panarchy. Surprisingly, this scholarly and well-documented volume has almost no references to research on adaptive cycles, resilience and panarchy. The lone exception is a citation to the 2002 Panarchy book edited by Gunderson and Holling. I hope that there is more exchange of ideas on punctuated policy dynamics and panarchy in the future.

Tremors and Tipping Points

Tipping points cause some important ecosystem surprises.  Examples include collapses of rangelands, water quality, and some fisheries.  The trouble with tipping points is that they are hard to anticipate in advance.  However, tremors may provide an advance warning of some tipping points.

The graphic shows a model of a pastoral system .  There is a tipping point when the stocking level of herbivores is about 5.  Above the tipping point, grassy vegetation disappears and the grazing system collapses.  As the tipping point is approached from low levels of herbivores, the standard deviation of grass biomass rises sharply before the tipping point is reached.  If the herbivore level is rising slowly enough, the rise in standard deviation could provide advance warning of impending collapse.  If the pastoralist was attentive to the warning, sheep numbers could be reduced in time to prevent the collapse.
 Pastoral Ecosystem

 Thomas Kleinen and colleagues have shown that reddening of the variance spectrum can anticipate rapid climate changes such as those that could result from a breakdown in ocean circulation.  Steve Carpenter and Buz Brock have analyzed water pollution, air pollution, and social systems that tremble before they tip.  They demonstrate increases in variance, which may be more easily detected than reddening of spectra.  Importantly, the variance increases can be detected with simple statistical filters using common time-series data.  No particular knowledge of the actual ecosystem dynamics is required.  Berglund and Gentz compare hard losses of stability in which an attractor vanishes (such as the pastoral system shown here) with soft losses of stability where an attractor divides like a braided river.  Hard losses of stability — the regime shifts that cause resource collapses — may provide stronger advance warnings than soft losses of stability — the regime shifts that gradually and imperceptibly create traps for ecosystem management.  Ludwig, Walker and Holling provide a more general discussion of hard and soft losses of stability in ecosystems.


Self-Organization of Ecosystem Lumpiness

niche evolutionWe have growing evidence that ecosystems are lumpy. Along an axis such as body size, for example, we find clusters of similar-sized species separated by intervals of body size in which no species are found. Multiple explanations exist for lumpy patterns, and causes are still debated. Scheffer and van Nes present a simple mathematical explanation for evolution of lumpy patterns in ecosystems. Their article appears in the Early Edition of PNAS on 3 April 2006. The abstract states

Here we show that self-organized clusters of look-a-likes may emerge spontaneously from evolution of competitors. The explanation is that there are two alternative ways to survive together: being sufficiently different or being sufficiently similar. Using a model based on classical competition theory, we demonstrate a tendency for evolutionary emergence of regularly spaced lumps of similar species along a niche axis . . . Our result suggest that these patterns may represent self-constructed niches emerging from competitive interactions.

Later, the authors comment

Finally, it is worth noting a remarkable link to Hotelling’s theory in social sciences suggesting that competition of companies or political parties will often lead to convergence rather than differentiation. In this field of research, the focus is on the problem that such convergence is not in the interest of the public. For instance, having more of the same kind of TV channels is not better. By contrast, the seeming redundancy of similar species in nature may be essential to ensure ecosystem functioning in the face of adverse impacts.

When Scheffer and van Nes’s article is published in the print version of PNAS, it will be accompanied by a commentary written by Craig Allen which places the new findings in the context of research on lumps dating to the original discovery by C.S. Holling in 1992 (Ecological Monographs 62: 447-502).

Greening of Cost-Benefit Analysis

The use of economic discounting for environmental decision analysis is often criticized. Discounting refers to the method of weighing present versus future benefits. Customarily, discounting has been calculated at a constant exponential rate, analogous to the interest rate on a loan. This can lead to absurd results. For example, living resources that grow more slowly than the discount rate (such as redwoods and whales) should be harvested to extinction, according to cost-benefit analyses using constant exponential discounting. Such outcomes have led some scientists and environmentalists to reject cost-benefit analysis for environmental decisions.

However, recent research shows that constant economic discounting is not supported by data for decisions with long time horizons. Proper approaches to discounting yield much greener decisions . Diverse economic models show that future interest rates are highly uncertain. Therefore environmental cost-benefit analyses must consider outcomes over a plausible range of models for future discount rates. When outcomes are averaged, models with relatively severe discounting (such as the constant exponential model) have negligible impact after a long period of time has elapsed. Instead, models that lightly discount the future have greatest impact on decisions. Thus optimal decisions are far greener than previously thought.

The greening effect is especially notable for decisions that involve environmental tipping points, such as species extinctions, freshwater quality, and climate change. Proper economic discounting leads to conservation-friendly decisions in these cases.

Constant exponential discounting has been a powerful obstacle to adaptive ecosystem management, because it downgrades the present value of future information. However, when discounting is done properly then learning has greater weight in decisions, favoring adaptive ecosystem management.

The Millennium Ecosystem Assessment has shown that properly computed cost-benefit analyses would often lead to conservation (rather than conversion) of ecosystems, even using constant exponential discounting. Proper discounting would favor conservation in an even wider variety of cases.

New Journals

Sustainability: Science, Practice, & Policy is a new peer-reviewed, open access journal that provides a platform for the dissemination of new practices and for dialogue emerging out of the field of sustainability. According to the journal’s web site, “The e-Journal fills a gap in the literature by establishing a forum for cross-disciplinary discussion of empirical and social sciences, practices, and policies related to sustainability. Sustainability will facilitate communication among scientists, practitioners, and policy makers who are investigating and shaping nature-society interactions and working towards sustainable solutions.”

Foresight: The International Journal of Applied Forecasting is an official publication of the International Institute of Forecasters.
The first issue contains a special feature on judgemental adjustment of statistical forecasts. These are methods for combining soft information or mental models held by individuals with statistical or mathematical models. The issue also includes a paper by John Boylan, “Intermittent and lumpy demand: a forecasting challenge”. In business, “slow items with intermittent and lumpy demand patterns may seem unimportant, but they can make up a substantial part of an organization’s inventory”. Boylan describes several methods for determining and forecasting regularities in lumpy time series. His discussion may be of interest to researchers studying lumpy series from a wide range of systems.

Positive Steps for Resilient Ecosystem Services

Although much of the mainstream press attention to the Millennium Ecosystem Assessment (see State of the World’s Ecosystems posted 31 March 2005) has emphasized the losses of ecosystem services and the adverse trends, a substantial fraction of the MA technical reports is devoted to positive, feasible steps that can be taken to improve ecosystem services in the future. All of these proactive steps are grounded in policies that are presently in place somewhere in the world today. A few examples:

• Increase the use of economic instruments and market-based approaches, e.g. assignments of property rights for ecosystem services, user fees for externalities, payment for ecosystem services, and mechanisms to express consumer preferences through markets (such as certification schemes)

• Explicitly include ecosystem services in poverty-reduction strategies

• Connect environmental management across ministries and sectors, instead of isolating it in a single ministry

• Create co-management systems to maintain reserves as part of regional mosaics

• Include local and indigenous knowledge, as well as technical knowledge, in decision-making

• Expand information available to individuals about how ecosystems affect them, and how their actions affect ecosystems

• Expand environment-friendly technology, especially in the areas of agriculture (water, nutrient and land use), urban design, and energy efficiency

Continue reading