Systemic risk and systemic opportunity: Alternative realities in ecological and social systems

Kogtraveler.com Simon Levin HistoricalCollapse2019 With thanks to Increasingly, we are hearing news about impending catastrophes

1ggye33lc4653z56mp34pl6t.wpengine And there are historical precedents 20 years ago, John Steele called attention to ecological regime shifts

February 1998 REGIME SHIFTS IN MARINE ECOSYSTEMS S33

Ecological Applications, 8(1) Supplement, 1998, pp. S33–S36 ᭧ 1998 by the Ecological Society of America

REGIME SHIFTS IN MARINE ECOSYSTEMS

JOHN H. STEELE Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543 USA www.whoi.edu Abstract. Time scales, and the trophic relations between these scales, are very different in the sea from those on land. In particular, marine systems are much more responsive to decadal scale alterations in their physical environment but are also much more adaptable. Thus it is difficult, and probably counterproductive, to try to define a baseline state for marine ecosystems. Further, regime shifts in fish communities can have major economic consequences without being ecological disasters. Climatic changes at decadal scales, from natural or anthropogenic causes, are likely to produce or enhance regime shifts. There are different management issues in different sectors. The coastal zone demands our intervention to assure integrated management of the land and sea components. At the other extreme, our understanding of open ocean systems is an essential element of climate prediction and so of eventual management. Between these two environments, our use of resources in continental shelf seas requires an ability to distinguish between human and natural causes of long-term change. Key words: ecological fungibility; fisheries; management; regime shifts; time scales; sustain- ability.

INTRODUCTION this reason, ocean systems are more amenable to longer term predictions: witness our successes with El Nin˜o Ifinditdifficulttoapplytheconceptofsustainability (Chen et al. 1995). Yet, these climate forecasts are still to life in the sea. In part, this is because the idea of on time scales of months to years rather than decades. sustainability eludes strict definition, especially in such When we turn to time scales of biological systems achangeableenvironmentastheocean.Yetthereisa on land and in the sea, the relation tends to be the common element in the different expressions for sus- reverse of the physics. On land, perennial grasses and tainability: our obligations to future generations trees are longest lived, with lifetimes that can be mea- (Brundtland 1987). There is the strong implication that sured in centuries. In the open sea, at the base of the we have some ability to forecast the consequences of food chain, are microscopic plants—phytoplankton— present actions or policies at decadal time scales. Pre- with lifetimes of days for individual organisms. Near diction is one of the objectives of scientific research the top of the open sea food chain are fish that we and it is from this viewpoint that I shall consider our harvest, like cod or tuna, with life-spans of years. Thus, understanding of the ocean. It is in this aspect of sci- the impact of our activities at time scales on the order ence—prediction—that the scientist and the policy of generations (i.e., decadal) can be very different on maker usually overlap and sometimes disagree. land and in the sea. From the human perspective, we Our ability to say something about the future is close- are in between, looking in different temporal ‘‘direc- ly related to the time scales involved. In the natural tions’’ at the land and sea. This, necessarily, has an world, this depends on the time scales of the processes effect on our concept of sustainability at our own in- we are interested in. I have argued (Steele 1985, 1991) tergenerational scales in the two environments. that a fundamental feature of the oceans is the differ- There is another difference between land and sea. ence in time scales compared with the land and at- Because of the similarity in time scales of the physical mosphere. Here, I will try to relate these differing and biological processes in the ocean, we oceanogra- scales to issues of management, especially of fisheries. phers see marine ecosystems as closely coupled to the As we know too well, weather forecasts are limited dynamics of advection and dispersion. Thus, marine to about a week, because this is the usual lifetime of communities are much more responsive to decadal cyclonic systems. In the sea, the corresponding feature, scale alterations in their physical environment than ocean eddies, can have durations of up to a year. For those on land, but are also much more adaptable to such changes (Steele 1991).

Manuscript received 20 February 1996; revised 15 July OCEAN SECTORS 1996; accepted 15 September 1996; final version received 18 February 1997. For reprints of this Special Issue, see footnote Given these generalizations about time scales in 1, p. S1. ocean processes, there is still tremendous diversity of

S33 But interest in alternative states, thresholds and transitions was not a new occupation in ecology BNL 19244 WOODWELL, 1974

THE THRESHOLD PROBLEM IN E C O S Y S T E M S George M . W o o d w e l l

The biotic resources of t h e e a r t h , most of w h i c h exist in n a t u r a l o r lightly managed ecosystems, a r e p r e s e n t l y in r e t r e a t undsr a barrage o f different types of d i s t u r b a n c e s that range from simple mechanical displace- ment to c u m u l a t i v e toxification of e n v i r o n m e n t . The r e s o u r c e s are i r r e p l a c e - able by t e c h n o l o g y , no m a t t e r the a b u n d a n c e of e n e r g y . The l o s s e s are r a p i d , accelerating, cumulative, and l a r g e l y irreversible; at s o m e point these changes must be r e c o g n i z e d as r e d u c i n g the h a b i t a b i l i t y of t h e e a r t h f o r man. Biotic change has n o t a l w a y s done so, o f c o u r s e , but t h e c h a n g e s a r e now worldwide and a f f e c t supplies of f o o d , fiber, fuel, a n d w a t e r as w e l l as the q u a l i t y of w a t e r and a i r . We a s k n o w w h a t can b e d r a w n from t h e science of e n v i r o n m e n t to m i t i g a t e or r e v e r s e the c u r r e n t trends? Is it possible to a l l o w limited worldwide changes in t h e p h y s i c s and c h e m i s t r y o f environment without aggravating these trends? Is i t r e a s o n a b l e to a s s u m e that thresholds for e f f e c t s of d i s t u r b a n c e exist in n a t u r a l ecosystems? Or are a l l d i s t u r b a n c e s effective, cumulative, and d e t r i m e n t a l to t h e n o r m a l *Biology Department, Brookhaven National Laboratory, Upton, New Y o r k 11973. Research carried out a t B r o o k h a v e n National Laboratory under the a u s p i c e s of the U . S . A t o m i c Energy Commission. MISTER -NOTICE- This report was prepared as an account of work sponsored by ihe United States Government. Neither the United Stales nor the United States Atomic Energy Commission, nor any of their employees, nor any of their contrauors, subcontractors, or their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, com- pleteness or usefulness of any information, apparatus, product or process disclosed, or represents that its. use would not infringe privately owned rights.

TED © 1977 Nature Publishing Group Ecological and socio-economic systems indeed can undergo sudden transitions

Scheffer, M. 1997. Ecology of shallow lakes. Kluwer Academic Publishers, Dordrecht, NL. And one transition can lead to another

NETWORK ANALYSIS Seeing a global web of connected systems Social-ecological shifts may often be causally linked

By Marten Scheffer and Egbert H. van Nes niques to mine the wealth of qualitative els of complex systems remain difficult to understanding of what affects what in the verify and inevitably leave out potentially he Arab Spring, the invention of pen- world (5, 6). Rocha et al. took such an ap- important processes (9). An emerging field icillin, and the recent mass bleach- proach to analyze a database of transitions of research therefore aims to complement ing of coral reefs are reminders that in intertwined socioecological systems (re- the simulation approach by extracting cau- much of the change in nature and gime shifts) on the basis of more than 1000 sality directly from the increasing flood T society happens in just a tiny por- scientific papers. They combined causes of time series data (10, 11). This is tricky tion of time. Understanding why and and effects of each of the regime shifts to because of the problems of spurious corre- when such critical transitions happen re- build a larger web of potential links be- lations and the chicken-and-egg situations mains notoriously difficult. On page 1379 tween them. They report that almost half that are inherent in characterizing causal of this issue, Rocha et al. (1) mine a data- of the theoretically possible links could loops. For example, it may be tempting

base of shifts in social and ecological sys- plausibly be causal. to interpret the observation that Earth’s Downloaded from tems and conclude that about half of them A major advantage of such a qualitative temperature started to rise somewhat RESEARCH may be causally linked on different scales. causal web analysis is that it allows the before atmospheric CO2 concentration Their results highlight the importance of use of heterogeneous collections of narra- started to increase as evidence that the unraveling hidden connections in the web tives. One may, for instance, build a cau- first caused the latter at a glacial termina- https://www.techwyse.com/ of ecological and social systems on which sality web simply from expert insights to tion. However, such lags are meaningless

we depend. analyze potential feedback loops and sta- in nonlinear dynamical systems in which http://science.sciencemag.org/ It makes intuitive sense that the great bility consequences (7). However, the use only sophisticated approaches may help Syrian exodus, the global financial crisis, of qualitative information inevitably has to probe the multiple simultaneous direc- ◥ Brexit, and the election of Pres- tions of causality (12). An emergent challenge for science and prac- ident Trump are not entirely “Understanding complex systems as a whole There is no silver-bullet ap- unrelated. Yet, revealing the proach to unraveling the big RESEARCH ARTICLE tice is that regime shifts can potentially lead to web of causalities behind such is crucial if we want to understand resilience scheme of things in nature transitions remains difficult. and regime shifts, be it in the biosphere or in and society, but that should Similarly, we have difficulties not stop us from leaving our subsequent regime shifts. We define a regime understanding connections the human body.” comfort zone and addressing between the fates of ice sheets, the questions that matter. Un- monsoons, tropical forests, and other tip- limitations too. Most importantly, it can derstanding complex systems as a whole is on January 12, 2019 shift as cascading when its occurrence may af- ping elements of Earth system that may only provide a catalog of the possible. The crucial if we want to understand resilience CRITICAL TRANSITIONS shape the cascade of transitions toward a actual stability properties and the net ef- and regime shifts, be it in the biosphere fect the occurrence of another regime shift. A hothouse Earth (2). The challenge of see- fects along causal chains depend critically or in the human body (13). The way that ing the bigger web of connections is a on the relative strength of the different Rocha et al. extracted insights from a web common thread in our struggle to under- mechanisms involved. For example, snow of qualitative narratives, and the emerging variety of causal pathways connecting regime stand transitions in complex systems. For cover reduces heat absorption, causing approaches to unravel causality from big example, after the collapse of the financial temperatures to drop; this allows more data are reasons for optimism. j services firm Lehman Brothers, the global surface to freeze. However, a potential run- shifts have been identified (table S1). For ex- REFERENCES cascade of events ran through a network away of this destabilizing feedback toward 1. J. C. Rocha et al., Science 362, 1379 (2018). of connections between banks and other a “snowball Earth” is prevented by coun- 115 ample, eutrophication is often reported as a 2. W. Steffen et al., Proc. Natl. Acad. Sci. U.S.A. , 8252 Cascading regime shifts financial institutes that was largely hidden teracting forces. At the same time, even (2018). from the public eye (3). Another example is weak destabilizing feedbacks can produce 3. S. Battiston et al., Science 351, 818 (2016). the Arab Spring, which precipitated from a a tipping point if they become aligned (8). 4. J. Brownlee, T. Masoud, A. Reynolds, Middle East Law and regime shift preceding hypoxia or dead zones 7 complex web of drivers that encompassed Thus, although analyzing webs of qualita- Governance , 3 (2015). 5. R. Levins, Ann. N.Y. Acad. Sci. 231, 123 (1974). social factors, climate-induced crop fail- tive insights can produce an inventory of 6. G. Giordano, C. Altafini, Sci. Rep. 7, 11378 (2017). in coastal areas (16). Similarly, hypoxic events ures, and the rise of the biofuel market— the theoretically possible secondary effects 7. A. S. Downing et al., Ecol. Soc. 19, 31 (2014). these connections could be reconstructed and causal loops, there is the question of 8. I. A. van de Leemput, T. P. Hughes, E. H. van Nes, M. within and across scales 35 only in hindsight (4). what we can do to reveal which of those Scheffer, Coral Reefs , 857 (2016). have been reported to affect the resilience of 9. T. Stocker, Climate Change 2013: The Physical Science This raises the question of what scien- are plausibly important. Basis: Working Group I Contribution to the Fifth tists can do to unravel the dazzling web The most obvious approach is to build Assessment Report of the Intergovernmental Panel on coral reefs to warming and other stressors in of connections that weaves the biosphere quantitative simulation models to study Climate Change (Cambridge Univ. Press, 2014). 1,2 1 1 1,2,3,4 and its human participants together. One the relative importance of different pro- 10. J. Runge, Chaos 28, 075310 (2018). Juan C. Rocha *, Garry Peterson , Örjan Bodin , Simon Levin 11. G. Sugihara et al., Science 338, 496 (2012). approach is to use smart analytical tech- cesses. This can help to constrain uncer- 12. E. H. van Nes et al., Nat. Clim. Chang. 5, 445 (2015). the tropics (17). If, why, and how a regime shift tainty in the vast catalog of the possible, 13. M. Scheffer et al., Proc. Natl. Acad. Sci. U.S.A. 115, 11883 showing, for example, that variation in (2018). Department of Environmental Sciences, Wageningen somewhere in the world could affect the oc- Universiteit, 6708 PB Wageningen, Netherlands. solar activity has a negligible effect on re- Email: [email protected] cent warming (9). However, detailed mod- 10.1126/science.aav8478 Regime shifts are large, abrupt, and persistent critical transitions in the function and currence of another regime shift remain largely SCIENCE sciencemag.org 21 DECEMBER 2018 • VOL 362 ISSUE 6421 1357 structure of ecosystems. Yet, it is unknown how these transitions will interact, whether the Published by AAAS open questions and a key frontier of research occurrence of one will increase the likelihood of another or simply correlate at distant (18, 19). places. We explored two types of cascading effects: Domino effects create one-way Research on regime shifts is often confined dependencies, whereas hidden feedbacks produce two-way interactions. We compare them to well-defined branches of science, reflecting em- with the control case of driver sharing, which can induce correlations. Using 30 regime pirical, theoretical (20), or predictive approaches shifts described as networks, we show that 45% of regime shift pairwise combinations Downloaded from (10, 21). These approaches require a deep knowl- present at least one plausible structural interdependence. The likelihood of cascading edge of the causal structure of the system or a effects depends on cross-scale interactions but differs for each type. Management of regime high quality of spatiotemporal data. Hence, re- shifts should account for potential connections. search on regime shifts has generally focused on the analysis of individual types of regime shifts egime shifts occur across a wide range of regimes when exposed to shock events (such as rather than potential interactions across systems.

social-ecological systems (1–3). They are hurricanes) or the action of external drivers (such We took another approach and instead explored http://science.sciencemag.org/ difficult to predict and reverse (4, 5) and as fishing) (10). More than 30 different regime potential cascading effects among a large set often produce sustained shifts in the avail- shifts in social-ecological systems have been of regime shifts. We investigated two types of R ability of ecosystems services (6). When a documented (3), and similar nonlinear dynam- interconnections: domino effects and hidden system undergoes a regime shift, it moves from ics are seen across societies, finance, language, feedbacks. Domino effects occur when the feed- one set of self-reinforcing processes and struc- neurological diseases, and climate (11, 12). As hu- back processes of one regime shift affect the tures to another (2, 7–9). Changes in a key var- mans increase their pressure on the planet, re- drivers of another regime shift, creating a one- iable (for example, temperature in coral reefs) gime shifts are likely to occur more often and way dependency (10, 19, 22). A feedback mech- often make a system more susceptible to shifting more severely (13–15). anism is a self-amplifying or -dampening process

Fig. 1. Method scheme. Pairs of regime shift on December 20, 2018 causal networks were merged to create a response variable matrix that accounted for drivers shared, domino effects, or hidden feedbacks. In all examples, two minimal regime shifts are depicted as causal diagrams, drivers are red, and variables belonging to feedbacks are purple. For driver sharing, the joint network is simplified as a two- mode network that allows us to study the co- occurrence of drivers (in red) across regime shifts (in blue). Driver a is shared by regime shifts 1 and 2, but driver b is not. The response variable matrix counts the number of drivers shared by all pairwise combinations of regime shifts. For domino effects, two regime shift networks are joined together, where driver c in regime shift 2 is also part of a feedback process in regime shift 1, creating a one-way dependency (orange link) between the two regime shifts. The response variable matrix counts all the one-way causal pathways between pairwise combinations of regime shifts. For hidden feedbacks, two minimal regime shifts, when joined together, give rise to a new unidentified feedback (orange circular pathway). The response variable matrix counts all hidden feedbacks that arise when merging pairwise combinations of regime shifts. The 30 causal networks used and the labeled matrices of the resulting response variables are shown in figs. S1 and S3.

Rocha et al., Science 362, 1379–1383 (2018) 21 December 2018 1 of 5 Ecosystems and the Biosphere are Complex Adaptive Systems Heterogeneous collections of individual units (agents) that interact locally, and evolve based on the outcomes of those interactions.

NOAA 11 So too are the socio-economic systems with which they are interlinked

www.suite101.com 12 There is an obvious need for early warning indicators of collapse

Or recovery Home.netcom.com Critical transitions occur in physiological states

http://www.edmontonneurotherapy.com/treatment_of_migraine.html Are there early warning indicators?

www.aesnet.org Can we read the tea leaves?

REVIEW

points. The basic ingredient for a tipping point is a positive feedback that, once a critical point is passed, propels change toward an alternative Anticipating Critical Transitions state (6). Although this principle is well under- stood for simple isolated systems, it is more chal- Marten Scheffer,1,2* Stephen R. Carpenter,3 Timothy M. Lenton,4 Jordi Bascompte,5 lenging to fathom how heterogeneous structurally William Brock,6 Vasilis Dakos,1,5 Johan van de Koppel,7,8 Ingrid A. van de Leemput,1 Simon A. Levin,9 complex systems such as networks of species, Egbert H. van Nes,1 Mercedes Pascual,10,11 John Vandermeer10 habitats, or societal structures might respond to changing conditions and perturbations. A broad Tipping points in complex systems may imply risks of unwanted collapse, but also opportunities range of studies suggests that two major features for positive change. Our capacity to navigate such risks and opportunities can be boosted by are crucial for the overall response of such sys- combining emerging insights from two unconnected fields of research. One line of work is tems (7): (i) the heterogeneity of the components revealing fundamental architectural features that may cause ecological networks, financial and (ii) their connectivity (Fig. 1). How these markets, and other complex systems to have tipping points. Another field of research is uncovering properties affect the stability depends on the na- generic empirical indicators of the proximity to such critical thresholds. Although sudden ture of the interactions in the network. shifts in complex systems will inevitably continue to surprise us, work at the crossroads of these Domino effects. One broad class of networks emerging fields offers new approaches for anticipating critical transitions. includes those where units (or “nodes”)canflip between alternative stable states and where the probability of being in one state is promoted by bout 12,000 years ago, the Earth sud- emerging research areas and discuss how excit- having neighbors in that state. One may think, for denly shifted from a long, harsh glacial ing opportunities arise from the combination of instance, of networks of populations (extinct or Aepisode into the benign and stable Hol- these so far disconnected fields of work. not), or ecosystems (with alternative stable states), ocene climate that allowed human civilization to or banks (solvent or not). In such networks, het- develop. On smaller and faster scales, ecosystems The Architecture of Fragility erogeneity in the response of individual nodes occasionally flip to contrasting states. Unlike grad- Sharp regime shifts that punctuate the usual fluc- and a low level of connectivity may cause the net- ual trends, such sharp shifts are largely unpre- tuations around trends in ecosystems or societies work as a whole to change gradually—rather than dictable (1–3). Nonetheless, science is now carving may often be simply the result of an unpredict- abruptly—in response to environmental change. on November 13, 2012 into this realm of unpredictability in fundamental able external shock. However, another possibility This is because the relatively isolated and differ- ways. Although the complexity of systems such is that such a shift represents a so-called critical ent nodes will each shift at another level of an en- as societies and ecological networks prohibits ac- transition (3, 4). The likelihood of such tran- vironmental driver (8). By contrast, homogeneity curate mechanistic modeling, certain features turn sitions may gradually increase as a system ap- (nodes being more similar) and a highly connected out to be generic markers of the fragility that may proaches a “tipping point” [i.e., a catastrophic network may provide resistance to change until a typically precede a large class of abrupt changes. bifurcation (5)], where a minor trigger can invoke threshold for a systemic critical transition is reached Two distinct approaches have led to these in- aself-propagatingshifttoacontrastingstate.One where all nodes shift in synchrony (8, 9). sights. On the one hand, analyses across networks of the big questions in complex systems science This situation implies a trade-off between lo- and other systems with many components have is what causes some systems to have such tipping cal and systemic resilience. Strong connectivity www.sciencemag.org revealed that particular aspects of their structure determine whether they are likely to have critical thresholds where they may change abruptly; on the other hand, recent findings suggest that cer- tain generic indicators may be used to detect if a State system is close to such a “tipping point.” We high- State light key findings but also challenges in these Downloaded from

1Department of Environmental Sciences, Wageningen Univer- sity, Post Office Box 47, NL-6700 AA Wageningen, Nether- 2 lands. South American Institute for Resilience and Sustainability Stress Stress Studies (SARAS), Maldonado, Uruguay. 3Center for Limnology, University of Wisconsin, 680 North Park Street, Madison, WI Modularity Connectivity https://www.twinings.co.uk 4 53706, USA. College of Life and Environmental Sciences, ++ University of Exeter, Hatherly Laboratories, Prince of Wales Heterogeneity Homogeneity Road, Exeter EX4 4PS, UK. 5Integrative Ecology Group, Estación 16 Biológica de Doñana, Consejo Superior de Investigaciones 6 Científicas, E-41092 Sevilla, Spain. Department of Economics, Adaptive capacity Resistance to change University of Wisconsin, 1180 Observatory Drive, Madison, WI ++ 53706, USA. 7Spatial Ecology Department, Royal Netherlands Local losses Local repairs Institute for Sea Research (NIOZ), Post Office Box 140, 4400AC, ++ Yerseke, Netherlands. 8Community and Conservation Ecology Gradual change Critical transitions Group, Centre for Ecological and Evolutionary Studies (CEES), University of Groningen, Post Office Box 11103, 9700 CC 9 Groningen, Netherlands. Department of Ecology and Evolu- Fig. 1. The connectivity and homogeneity of the units affect the way in which distributed systems with tionary Biology, Princeton University, Princeton, NJ 08544–1003, local alternative states respond to changing conditions. Networks in which the components differ (are USA. 10University of Michigan and Howard Hughes Medical Institute, 2045 Kraus Natural Science Building, 830 North Uni- heterogeneous) and where incomplete connectivity causes modularity tend to have adaptive capacity in versity, Ann Arbor, MI 48109–1048, USA. 11Santa Fe Institute, that they adjust gradually to change. By contrast, in highly connected networks, local losses tend to be 1399 Hyde Park Road, Santa Fe, NM 87501, USA. “repaired” by subsidiary inputs from linked units until at a critical stress level the system collapses. The *To whom correspondence should be addressed. E-mail: particular structure of connections also has important consequences for the robustness of networks, [email protected] depending on the kind of interactions between the nodes of the network.

344 19 OCTOBER 2012 VOL 338 SCIENCE www.sciencemag.org Many such transitions have characteristic early warning signals

• Critical slowing down REVIEW

points. The basic ingredient for a tipping point is a positive feedback that, once a critical point • is passed, propels change toward an alternative Increasing variance Anticipating Critical Transitions state (6). Although this principle is well under- stood for simple isolated systems, it is more chal- Marten Scheffer,1,2* Stephen R. Carpenter,3 Timothy M. Lenton,4 Jordi Bascompte,5 lenging to fathom how heterogeneous structurally William Brock,6 Vasilis Dakos,1,5 Johan van de Koppel,7,8 Ingrid A. van de Leemput,1 Simon A. Levin,9 complex systems such as networks of species, Egbert H. van Nes,1 Mercedes Pascual,10,11 John Vandermeer10 habitats, or societal structures might respond to changing conditions and perturbations. A broad Tipping points in complex systems may imply risks of unwanted collapse, but also opportunities range of studies suggests that two major features for positive change. Our capacity to navigate such risks and opportunities can be boosted by are crucial for the overall response of such sys- • Increasing autocorrelationcombining emerging insights from two unconnected fields of research. One line of work is tems (7): (i) the heterogeneity of the components revealing fundamental architectural features that may cause ecological networks, financial and (ii) their connectivity (Fig. 1). How these markets, and other complex systems to have tipping points. Another field of research is uncovering properties affect the stability depends on the na- generic empirical indicators of the proximity to such critical thresholds. Although sudden ture of the interactions in the network. shifts in complex systems will inevitably continue to surprise us, work at the crossroads of these Domino effects. One broad class of networks emerging fields offers new approaches for anticipating critical transitions. includes those where units (or “nodes”)canflip between alternative stable states and where the probability of being in one state is promoted by • bout 12,000 years ago, the Earth sud- emerging research areas and discuss how excit- having neighbors in that state. One may think, for Flickering between states denly shifted from a long, harsh glacial ing opportunities arise from the combination of instance, of networks of populations (extinct or Aepisode into the benign and stable Hol- these so far disconnected fields of work. not), or ecosystems (with alternative stable states), ocene climate that allowed human civilization to or banks (solvent or not). In such networks, het- develop. On smaller and faster scales, ecosystems The Architecture of Fragility erogeneity in the response of individual nodes occasionally flip to contrasting states. Unlike grad- Sharp regime shifts that punctuate the usual fluc- and a low level of connectivity may cause the net-

ual trends, such sharp shifts are largely unpre- tuations around trends in ecosystems or societies work as a whole to change gradually—rather than on October 21, 2012 dictable (1–3). Nonetheless, science is now carving may often be simply the result of an unpredict- abruptly—in response to environmental change. into this realm of unpredictability in fundamental able external shock. However, another possibility This is because the relatively isolated and differ- ways. Although the complexity of systems such is that such a shift represents a so-called critical ent nodes will each shift at another level of an en- as societies and ecological networks prohibits ac- transition (3, 4). The likelihood of such tran- vironmental driver (8). By contrast, homogeneity curate mechanistic modeling, certain features turn sitions may gradually increase as a system ap- (nodes being more similar) and a highly connected out to be generic markers of the fragility that may proaches a “tipping point” [i.e., a catastrophic network may provide resistance to change until a typically precede a large class of abrupt changes. bifurcation (5)], where a minor trigger can invoke threshold for a systemic critical transition is reached Two distinct approaches have led to these in- aself-propagatingshifttoacontrastingstate.One where all nodes shift in synchrony (8, 9).

sights. On the one hand, analyses across networks of the big questions in complex systems science This situation implies a trade-off between lo- www.sciencemag.org and other systems with many components have is what causes some systems to have such tipping cal and systemic resilience. Strong connectivity revealed that particular aspects of their structure determine whether they are likely to have critical thresholds where they may change abruptly; on the other hand, recent findings suggest that cer- tain generic indicators may be used to detect if a State system is close to such a “tipping point.” We high- State

light key findings but also challenges in these Downloaded from

1Department of Environmental Sciences, Wageningen Univer- sity, Post Office Box 47, NL-6700 AA Wageningen, Nether- 2 lands. South American Institute for Resilience and Sustainability Stress Stress Studies (SARAS), Maldonado, Uruguay. 3Center for Limnology, University of Wisconsin, 680 North Park Street, Madison, WI Modularity Connectivity 53706, USA. 4College of Life and Environmental Sciences, ++ University of Exeter, Hatherly Laboratories, Prince of Wales Heterogeneity Homogeneity Road, Exeter EX4 4PS, UK. 5Integrative Ecology Group, Estación Biológica de Doñana, Consejo Superior de Investigaciones 6 Científicas, E-41092 Sevilla, Spain. Department of Economics, Adaptive capacity Resistance to change University of Wisconsin, 1180 Observatory Drive, Madison, WI ++ 53706, USA. 7Spatial Ecology Department, Royal Netherlands Local losses Local repairs Institute for Sea Research (NIOZ), Post Office Box 140, 4400AC, ++ Yerseke, Netherlands. 8Community and Conservation Ecology Gradual change Critical transitions Group, Centre for Ecological and Evolutionary Studies (CEES), University of Groningen, Post Office Box 11103, 9700 CC 9 Bardy, B.; Oullier, O.; Bootsma, R. J.; Groningen, Netherlands. Department of Ecology and Evolu- Fig. 1. The connectivity and homogeneity of the units affect the way in which distributed systems with tionary Biology, Princeton University, Princeton, NJ 08544–1003, local alternative states respond to changing conditions. Networks in which the components differ (are USA. 10University of Michigan and Howard Hughes Medical Institute, 2045 Kraus Natural Science Building, 830 North Uni- heterogeneous) and where incomplete connectivity causes modularity tend to have adaptive capacity in Stoffregen, T. A.;J. Exp. Psych. Vol versity, Ann Arbor, MI 48109–1048, USA. 11Santa Fe Institute, that they adjust gradually to change. By contrast, in highly connected networks, local losses tend to be 1399 Hyde Park Road, Santa Fe, NM 87501, USA. “repaired” by subsidiary inputs from linked units until at a critical stress level the system collapses. The *To whom correspondence should be addressed. E-mail: particular structure of connections also has important consequences for the robustness of networks, 28(3):499-514. [email protected] depending on the kind of interactions between the nodes of the network.

344 19 OCTOBER 2012 VOL 338 SCIENCE www.sciencemag.org But such signals are not uniform Theor Ecol (2013) 6:255–264 Current caveats DOI 10.1007/s12080-013-0192-6

ORIGINAL PAPER

Early warning signals: the charted and uncharted territories

Carl Boettiger Noam Ross Alan Hastings · ·

Received: 19 March 2013 / Accepted: 23 May 2013 / Published online: 21 June 2013 ©SpringerScience+BusinessMediaDordrecht2013

Abstract The realization that complex systems such as down, statistical detection is a challenge. We review the ecological communities can collapse or shift regimes sud- literature that explores these edge cases and highlight the denly and without rapid external forcing poses a serious need for (a) new early warningbehaviorsthatcanbeused challenge to our understanding and management of the nat- in cases where rapid shifts do not exhibit critical slowing ural world. The potential to identify early warning signals down; (b) the development of methods to identify which that would allow researchers and managers to predict such behavior might be an appropriate signal when encountering events before they happen has therefore been an invaluable a novel system, bearing in mind that a positive indication for19 discovery that offers a way forward in spite of such seem- some systems is a negative indication in others; and (c) sta- ingly unpredictable behavior. Research into early warning tistical methods that can distinguish between signatures of signals has demonstrated that it is possible to define and early warning behaviors and noise. detect such early warning signals in advance of a transition in certain contexts. Here, we describe the pattern emerging Keywords Early warning signals Regime shifts · · as research continues to explore just how far we can gener- Bifurcation Critical slowing down · alize these results. A core of examples emerges that shares three properties: the phenomenon of rapid regime shifts, a pattern of “critical slowing down” that can be used to detect Introduction the approaching shift, and a mechanism of bifurcation driv- ing the sudden change. As research has expanded beyond Many natural systems exhibit regime shifts—rapid changes these core examples, it is becoming clear that not all sys- in the state and conditions of system behavior. Examples tems that show regime shifts exhibit critical slowing down, of such shifts include lake eutrophication (Carpenter et al. or vice versa. Even when systems exhibit critical slowing 1999), algal overgrowth of coral systems (Mumby et al. 2007), fishery collapse (Jackson et al. 2001), desertification of grasslands (K´efi et al. 2007), and rapid changes in climate (Dakos et al. 2008; Lenton et al. 2009). Such dramatic shifts Carl Boettiger and Noam Ross contributed equally. have the potential to impact ecosystem health and human well-being. Thus, it is important to develop strategies for C. Boettiger (!) adaptation, mitigation, and avoidance of such shifts. Center for Stock Assessment Research, Department of Applied Math and Statistics, University of California, Mail Stop SOE-2, The idea that complex systems such as ecosystems could Santa Cruz, CA 95064, USA change suddenly and without warning goes back to the e-mail: [email protected] 1960s (Lewontin 1969;Holling1973; May 1977). Such early work revealed that even simple models with the appro- N. Ross A. Hastings Department· of Environmental Science and Policy, priate nonlinearities were capable of unpredictable behav- University of California Davis, 1 Shields Avenue, ior. The only way to predict the transition was to have Davis, CA 95616, USA the right model—and that meant having already had the Caution is needed…mechanisms need to be identified

http://en.wikipedia.org/wiki/Catastrophe_theory http://en.wikipedia.org/wiki/Ren%C3%A9_Thom In physical systems, phase transitions provide a model

Ising Model http://www.icmp.lviv.ua/ising/galam.html Phase transitions don’t all show same features • First-order: Discontinuity in first derivative of free energy (Ehrenfest). – Melting of ice, boiling of water • Second-order: Discontinuity in second derivative. – Ferromagnetic transition – Infinite correlation length – Power-law decay of correlations near criticality Synchrony can reduce robustness by reducing dimensionality

www.naturephotosociety.org.sg 23 Synchrony is an indicator of market stress

Quantifying the Behavior of Stock Correlations Under Market Stress SUBJECT AREAS: Tobias Preis1,2,3, Dror Y. Kenett2,4, H. Eugene Stanley2, Dirk Helbing5 & Eshel Ben-Jacob4 STATISTICAL PHYSICS, THERMODYNAMICS AND NONLINEAR DYNAMICS 1Warwick Business School, University of Warwick, Coventry, CV4 7AL, United Kingdom, 2Center for Polymer Studies, Department PHYSICS of Physics, Boston University, Boston, MA 02215, USA, 3Department of Mathematics, University College London, London, WC1E 4 5 INFORMATION THEORY AND 6BT, UK, School of Physics and Astronomy, Tel-Aviv University, Tel-Aviv 69978, Israel and, Chair of Sociology, in particular of COMPUTATION Modeling and Simulation, ETH Zurich, 8092 Zurich, Switzerland. STATISTICS Understanding correlations in complex systems is crucial in the face of turbulence, such as the ongoing financial crisis. However, in complex systems, such as financial systems, correlations are not constant but Received instead vary in time. Here we address the question of quantifying state-dependent correlations in stock 29 June 2012 markets. Reliable estimates of correlations are absolutely necessary to protect a portfolio. We analyze 72 years of daily closing prices of the 30 stocks forming the Dow Jones Industrial Average (DJIA). We find the Accepted striking result that the average correlation among these stocks scales linearly with market stress reflected by 25 September 2012 normalized DJIA index returns on various time scales. Consequently, the diversification effect which should Published protect a portfolio melts away in times of market losses, just when it would most urgently be needed. Our 18 October 2012 empirical analysis is consistent with the interesting possibility that one could anticipate diversification breakdowns, guiding the design of protected portfolios.

Correspondence and ild fluctuations in stock prices1–8 continue to have a huge impact on the world economy and the requests for materials personal fortunes of millions, shedding light on the complex nature of financial and economic systems. should be addressed to For these systems, a truly gargantuan amount of pre-existing precise financial market data9–11 com- W 12–15 T.P. (mail@tobiaspreis. plemented by new big data ressources is available for analyses. de) The complex mechanisms of financial market moves can lead to sudden trend switches16–18 in a number of stocks. Such sudden trend switches can occur in a synchronized fashion, in a large number of stocks simulta- neously, or in an unsynchronized fashion, affecting only a few stocks at the same time. Diversification in stock markets refers to the reduction of portfolio risk caused by the investment in a variety of stocks. If stock prices do not move up and down in perfect synchrony, a diversified portfolio will have less risk than the weighted average risk of its constituent stocks19,20. Hence it should be possible to reduce risk in price of individual stocks by the combination of an appropriate set of stocks. To identify such an appropriate set of stocks with anti-correlated price time series, the assumption mostly used is that the correlations among stocks are constant over time21–26. This widely used assumption is also the basis for the determination of capital require- ments of financial institutions that usually own a huge variety of constituents belonging to different asset classes. Recent studies building on the availability of huge and detailed data sets of financial markets have analyzed and modeled the static and dynamic behavior of this very complex system27–39, suggesting that financial markets are governed by systemic shifts and display non-equilibrium properties. A very well known stylized fact of financial markets is the leverage effect, a term coined by Black to describe the negative correlation between past price returns and future realized volatilities in stock markets. According to Reigneron et al.40, the index leverage effect can be decomposed into a volatility effect and a correlation effect. In the course of recent financial market crises, this effect has regained center stage, and the work of different groups has focused on uncovering its true nature40–46). Reigneron et al. analyzed daily returns of six indices from 2000 to 2010 and found that a downward index trends increase the average correlation between stocks, as quantified by measurements of eigenvalues of the conditional correlation matrix. They suggest that a quadratic term should be included to the linear regressions of the dependence of mean correlation on the index return the previous day. Here, we will expand on these results utilizing 72 years of trading of the 30 Dow Jones industrial average (DJIA) components (see also47,48). Using this financial data set we will quantify state-dependent stock market correlations and analyze how they vary in face of dramatic market losses. In such ‘‘stress’’ scenarios, reliable correlations are most needed to protect the value of a portfolio against losses.

SCIENTIFIC REPORTS | 2 : 752 | DOI: 10.1038/srep00752 1 Dimensional Reduction as an Early Warning Indicator of Transition • Power laws • Correlations in financial markets • Housing market variations

James Watson and https://www.express.co.uk/news/science/81 George Hagstrom 9741/Wormholes-in-Milky-Way-galaxy- interstellar Issues of critical transitions and sustainability cut across disciplines

http://www.sickchirpse.com/the-darwin-awards- Thank you

Source unknown