Main

On the morning of 23 March 2021, the giant container ship Ever Given was passing through the Suez Canal on its way to Rotterdam when it suddenly ran aground diagonally, blocking the entire canal. Because the ship was one of the largest in the world, traffic was jammed in both directions for six days. Hundreds of vessels came to a standstill, and billions of US dollars’ worth of trade were lost given the lack of alternative routes and modes of transport. Disruptions at bottlenecks like this (Fig. 1) can have major consequences for billions of people, enterprises and nations, influencing food supplies, prices or access to spare parts, with potentially far-reaching social consequences1.

Fig. 1: Major maritime choke points and primary (solid blue) and secondary (dotted blue) shipping routes.
figure 1

Ocean shipping accounts for the bulk of all transportation in international trade (80% by volume and 70% by value)68,69. The numbers (%) are estimates of global food staple (wheat, maize, rice and soy) volumes passing maritime choke points in 2020. Choke points are rated as moderate (yellow, minimal delay for shipments), high (red, substantial cost due to transit time and shipping costs) and critical (purple, no obvious alternative maritime route is available). Many commodities pass several maritime choke points towards their final destination and must also pass coastal (ports) and inland (railway, waterway or road networks) choke points. Figure adapted with permission from ref. 70, Chatham House.

Good preparation to avoid and respond to disruption requires access to a broad set of options to face unanticipated disruptions2. The current paradigms of lean sourcing, just-in-time and optimization (‘efficiency’) are ill-suited in this regard as they are not designed to handle unexpected new situations, such as the Ever Given incident3 and the COVID-19 pandemic, particularly when occurring in tandem. Of course, widening the Suez Canal would increase the resilience of its traffic flow to incidents like the Ever Given but would be ineffective against other kinds of disruptions (for example, political disruptions or armed conflicts) that might interrupt traffic — or if ship sizes continue to grow. Alternative responses include increasing storage capacity at receiving ends of the traffic or diversifying how goods are transported (China’s silk railroad, for example). This example highlights that, typically, a wide range of potential options are available for escaping rigid, vulnerable and therefore unsustainable structures4,5.

The Ever Given incident is symptomatic of a global trend in which people, cultures and economies are increasingly linked across geographical locations and socio-economic contexts6,7, but with limited options for changing the links8,9. While this connectivity provides opportunities for humanity in terms of collective action to deal with global challenges (for example, climate, pandemics and conflicts) and sharing ideas, goods and information10, our capacity to understand and control global socio-economic networks (for example, trade and finance) is becoming progressively more limited as complexity and interdependencies increase11. Furthermore, humans have become a dominant global force with profound impacts on the Earth’s biosphere9,12,13,14. The world is witnessing an increasing frequency, magnitude and duration of extreme events, including pandemics, heatwaves, mega-fires, droughts, floods and storms15. The associated costs are substantial in terms of economic and ecological disruption, reduced health, civil unrest, increased risk of geopolitical conflicts, human migration and, ultimately, human lives16.

Increasing awareness of the many uncertainties that humanity faces has led to calls for building resilience17— most notably greater resilience to threats in general rather than to particular threats. Of the aspects of such general resilience18,19, the most crucial is having a diversity of responses to different kinds of disruptions. Though the value of diversification has long been recognized (“Don’t put all your eggs in one basket” — Cervantes, 1612 (ref. 20), the rapid increase in the frequency and severity of ecological, social and economic disruptions underlines its growing importance15,21,22.

In this paper, we suggest that, if we wish to build general resilience to disruptions that cannot be exactly determined in advance, society needs to strengthen its response diversity. Response diversity is a system’s variety of responses to disruptions of all kinds. While this term originates from ecology23, we argue that it is critical to improving the resilience of any complex system. It suggests keeping options open for unexpected situations, which is consistent with theories about optimal decision-making under uncertainty and irreversibility24,25,26,27.

As we work through the various facets of response diversity, an important point to bear in mind is that, like resilience, it is a property of a system and is neither ‘good’ nor ‘bad’ per se. It can help maintain the current state of a system no matter whether it is deemed desirable or undesirable. If the state of a system and its trajectory are clearly undesirable, the appropriate focus of response diversity should be on alternative transformational pathways. Because our reason for writing this paper is the serious loss of response diversity, the focus here is on when, where and how it is playing a positive role.

Despite the critical role that response diversity plays in nature and in society at large, insights that extend beyond single sectors and disciplines are currently lacking. In this paper, we aim to fill this gap by integrating the different ways in which the concept is used and applied and highlighting the interconnectivity between different types of responses across sectors and scales. In particular, we explore just what response diversity means, how it is expressed in all kinds of systems, how it can be built and lost, its costs and benefits, and its implications for policy and governance. We conclude with some suggestions for strategies and policies to maintain or enhance response diversity. Importantly, our aim is not to scrutinize individual strategies for implementing response diversity in particular sectors but rather to provide general guidelines relevant across disciplines, which can be explored in more detail within different specific contexts.

Response diversity

Living systems, from individual organisms to the global system, depend on having a set of processes (for example, in ecosystems, photosynthesis, decomposition and predation; or in an economy, production and exchange of goods and services, waste management and transport) that enable that system to function. To ensure that these processes can persist in the long term, agents in a system must have multiple ways by which they can respond to changes and disruptions. In other words, response diversity provides the raw material for adaptive behaviour (Fig. 2).

Fig. 2: A conceptual illustration of response diversity.
figure 2

A high-diversity system (top) (for example, an agroecosystem) is more likely to maintain system functions and processes when facing a disturbance, whereas a low-diversity system (bottom) (for example, a monoculture) is highly vulnerable to a specific disturbance. Symbols of different colours represent the diversity (n) of agents/structures in a system (for example, species, traits, reserves or strategies).

In ecosystems, there are different species that perform the same process but differ in the ways they respond to a particular disturbance23. Socio-economic systems have also developed a variety of ways for providing essential services with different coping capacities, such as different types of water storage and delivery infrastructure, different modes of transportation or different sources of various materials and products. Many small-scale irrigation systems have flexible institutions to manage environmental change — for example, by altering water allocation as water availability changes28. Such adaptive institutions provide a diverse repertoire of ‘software solutions’ for social organization and thus maintain critical response diversity. Some of these strategies emerged after existing services had failed to respond to some new kind of shock; others were planned in advance. These diverse ways in which actors respond to a variety of shocks enables the function concerned to continue, thereby helping the system as a whole to continue functioning in much the same way. This is how response diversity confers resilience23. However, responding in different ways is also likely to have consequences beyond a particular function or scale, as we illustrate.

We argue here that we need to identify different sources of response diversity, assess trends in those sources and understand the implications of responding differently. In many systems, response diversity is largely organizational and perhaps hierarchical, as discussed in Levin et al.29. In this paper, to help further unpack response diversity and make it more operational, we complement Levin et al.29 by focusing on spatial and temporal dimensions of response diversity, in addition to population and community responses (natural and human).

Spatial responses

In ecosystems, species operate at different spatial scales to avoid competition. This results in enhanced robustness over a wider range of environmental conditions30. In coral reefs, for example, small territorial fish and sea urchins keep algal proliferation under control. So do schools of larger fish species that move over much wider areas. If a local storm hits the reef and kills less mobile species, species that operate at larger scales act as an important component of response diversity. They can continue to regulate algal populations and ‘smooth’ them in time9,31. In a similar way, migratory birds vary the locations and sizes of their territories as a way to increase their resilience to lack of food or difficult weather conditions.

In social systems, international trade provides spatial response diversity for buffering against disruptions at a national or local scale by providing alternative food sources, alternative distribution lines or emergency supplies32. Trade from multiple sources, using various transport routes or modes, contributes to response diversity in the sense that if shocks to the availability of one exporter or importer occur, trade can continue with another. A recent example is the vulnerability to energy shortages that several European countries are currently facing with the Russian invasion of Ukraine. In cities with limited open space, strong planning traditions and highly formalized procedures, peri-urban areas can contribute with both alternative spaces for different activities and less rigid planning and decision-making processes. Placed between the urban and rural governance systems, such areas have often developed ways for circumventing legal barriers or entrenched urban governance that have adverse effects at the scale of the peri-urban system33. Peri-urban areas may thus offer both alternative spaces and alternative governance pathways, which may be explored in times of need. The adjacency to urban areas makes it a potential vital contribution to cities’ capacity to respond to different perturbations.

Whether ecological or socio-economic, spatial responses share a common feature: they integrate over space to smooth variation. These responses thus require mobility infrastructures that allow agents to move to resources or move the resources to them.

Temporal responses

We define a temporal response as a shift in when and how often something is done or in the amount of time invested in some activity. Such variation in resource use or extraction over time can be a necessary part of resilience — for example, to compensate for variation in the amount of resources available to be extracted in different periods, thereby avoiding periods of great shortages and smoothing the flow or supply of valued resources. Common examples in human societies include storage in granaries and reservoirs, as well as banks. Many animals use similar strategies and store some of their food to be able to consume it later. For example, the Eurasian jay (Garrulus glandarius) gathers oak (Quercus spp.) seeds that it buries in the soil for future consumption.

Insurance systems work in a similar way but add a scale dimension because present insurance payments from people currently not harmed can be used to compensate those who are harmed. In the future, the payment flows may go in other directions depending on who is harmed at the time. Importantly, insurance and financial systems work on trust: they are storages of commitments and require shared infrastructure.

Like spatial responses, temporal responses all have a common feature: they integrate over time to smooth variation. These responses require storage infrastructure to accumulate and release resources at different times — that is, to ‘move’ them in time.

Cross-scale interactions in responses to disruptions

While the dimensions of space and time are critical for response diversity, possible cross-scale interactions add a layer of complexity. For example, before the financial crisis in 2007–2008, individual banks used diversification to cope with uncertainty (that is, increasing their response diversity). However, since many banks deployed similar risk management models, homogeneity of responses emerged at the global scale such that response diversity was eroded within the sector as a whole34,35. In other words, building response diversity at smaller scales can erode response diversity at larger scales if local initiatives copy each other9,36 (Fig. 3). Examples of food systems and global supply chains illustrate this point (Boxes 1 and 2).

Fig. 3: How response diversity manifests at different spatial scales and how this can influence the propagation of risk.
figure 3

Circles of different colours represent the diversity (n) of agents/structures (for example, species, traits, reserves or strategies) in a system at local and global scales. Credit: globe icon, Freepik.com.

Challenges associated with response diversity

Numerical metrics have limited utility for measuring response diversity. As this paper makes clear, response diversity emerges from a complex of attributes and modes of operating, and any one of them may be limiting for any particular system when subjected to particular disturbances. Yet, there are approaches, frameworks and metrics that could help unpack this complexity. For example, one way to estimate response diversity in ecological systems is through functional metrics that allow us to measure the diversity and distribution of response traits — that is, functional characteristics that determine an organism’s response to perturbations — in a multi-dimensional functional trait space37.

In social(–ecological) systems, this could be translated into the diversity of livelihood, management and governance strategies that can be mobilized to cope with change38. However, it is important to note that strategies that provide response diversity for one type of disturbance may do nothing for another type of disturbance. For example, the diversification of livelihoods (for example, the selling and repairing of fishing gear or using alternative trade networks) in small-scale fisheries to cope with impacts of variable fish stocks or overfishing may have little effect if these alternative livelihoods still rely on abundant fish stocks and a disturbance affects the basic resource itself (that is, false response diversity). Subsequently, for each of the response diversity attributes and ways of operating, we should ask which aspects of diversity are the most critical for the system’s ability to respond to disturbances and in what ways these aspects could be increased.

There are two overall challenges in maintaining response diversity: (1) managing trade-offs between using resources in the best way for present conditions and using them to better deal with unexpected change in the future39 and (2) managing trade-offs between investments targeted at coping with different classes of potential shocks. Resolving the first challenge requires balancing costly investments into maintaining or building diverse ways of responding to shocks. Assessing the benefits of various investments is very difficult due to the uncertainty in future conditions, and equitably distributing the costs of managing variability to present generations raises difficult questions40.

There are intertemporal trade-offs in the benefits and costs of response diversity, and because it is an emergent property of a complex system, it is difficult to design in advance. In some circumstances, a particular kind of disruption can be envisaged (as in the Suez Canal case) and appropriate alternative responses planned. In most cases, investing in redundancy (back-up systems with some deliberate variation) and in modularity (to prevent the uncontrolled spread of unwanted phenomena) can provide some degree of response diversity.

The second challenge to the fostering of response diversity has to do with the nature, frequency and intensity of disturbances; the scale at which they operate and interact; and the disconnect between social and ecological responses. Specifically, a fundamental feature of feedback systems capable of coping with shocks and variability is that total fragility in a system is conserved — that is, there is a minimum intrinsic level of fragility that cannot be eliminated. This basic principle from modern control theory limits our capacity to cope with all possible disturbances: investments in feedback system architectures (such as response diversity) that are focused, for example, on disturbances of a certain frequency range necessarily make the system vulnerable or ‘fragile’ to disturbances at other frequency ranges41. This notion has been extended to biological systems to demonstrate hard robustness limits in systems with “highly optimized tolerance”42 and underpins general theories of biological robustness43 wherein systems must trade off optimality, robustness and evolvability44. Other work has extended the basic principle of robustness–fragility trade-offs to social–ecological systems45,46 and has illustrated, for example, trade-offs between increasing robustness to uncertainty in the economic domain and increasing vulnerability to uncertainty in the ecological domain. Such fundamental design considerations, the cost of response diversity and the necessary compromises in addressing the question of ‘resilience of what to what’ must play key roles in strategies for strengthening response diversity.

To serve its purpose, response diversity must maintain the agents and structures that ensure system stability over time. Insurance systems, for instance, were described above as a temporal response to dealing with unexpected disruptions. They typically cover situations where the expected consequences of a shock are high and the probability of its occurrence is low and uncorrelated among insured individuals. The insured bear the cost of response diversity in the form of the insurance premium. Sometimes, the cost can be too high even for insurance companies. In these situations, the reinsurance industry can help spread the risks over many insurance companies, in different parts of the world subject to different kinds of shocks, and in this way develop response diversity. This enables the insurance companies and their insured bodies to remain resilient to the range of shocks they can expect.

Nevertheless, situations where the probabilities of shocks or bad outcomes are strongly correlated at the global scale are harder to deal with through insurance and reinsurance systems, which often include force majeure clauses against them. Climate change dynamics, for example, are likely to trigger correlated shocks in large regions. Insurance might assist the victims of droughts, large forest fires or inundations even if these occur simultaneously (as witnessed in the summer of 2021 when central Europe suffered unusually large and severe inundations while wildfires across several continents were larger than in recorded history), as long as such events are sufficiently rare or low-cost, but less so if damage is overwhelmingly large or occurs simultaneously for almost everyone.

Climate change contributes to correlated hazard risks globally while also inducing synergy of multi-hazard risks. Reinsurance is important but falls short if risks are too strongly correlated globally47. Adaptation to risks by households and companies then becomes more relevant, and this requires diversity in itself because the best strategies in local situations to safeguard against hazards are not always clear. As an example, potential strategies for flooding hazards include flood protection, reduction of the peak flows, mitigation of vulnerability and relocation to safer areas48.

Another challenge is the current disconnect between ecological and socio-economic responses. Insurance against weather-related crop failure, for example, provides an opportunity for farmers to cover themselves against crop losses caused by droughts. These insurances are not based on directly measured loss of crops; payouts are instead triggered by an index, such as a predefined threshold in rainfall49. Farmers with access to this type of insurance seem more prone to invest in high-profit but riskier crops50. Since these insurances are often also coupled to the adoption of commercial inputs, they may reinforce the simplification of agricultural landscapes and the homogenization of practices9. In general, support to maintain functions in risky environments provides incentives to continue with increasingly risky behaviour and the associated loss of response diversity.

The world currently faces many serious problems — disease epidemics, climate change, economic meltdown, social turmoil, war and so on — and as this has unfolded, we seem to have moved from a social–ecological system with high resilience in its biosphere part but little in the social part to one with much more resilience in its social part (for now) at the expense of its biosphere, as illustrated by the example above. The lack of appropriate institutions for dealing with these problems has been identified as a major cause of the inability to act51. The process has undoubtedly been exacerbated by humanity’s overall success in increasing its short-term well-being to an unprecedented extent (for example, the number of people and the amount of welfare they enjoy on average).

Attempting to increase resilience in the social system without acknowledging the need to maintain it in ecosystems has led to a general decrease in social–ecological response diversity9. Hence, understanding the combined social–ecological responses across scales is crucial when evaluating intentionally designed response diversity and redundancy.

Finally, justice and equity issues loom large in the challenges facing programmes for response diversity. As we described above, there are often direct or indirect costs to responding, and these can be shared more or less equally. As was clearly shown by Elmqvist et al.23, maintaining high-level, aggregate functional performance is often a question of some responses being successful while others fail. Ecologically or evolutionary, this has no normative implications, but when expanding response diversity to people and social systems, it can. Response diversity often includes options that are exploitative or are degenerative in the long term. The roving bandits syndrome52, for example, illustrates a response option that is beneficial to powerful companies and their customers but disastrous for small-scale fishermen in the targeted regions. Less overt, any resources (financial or otherwise) invested in response diversity incur an opportunity cost because they could have been invested elsewhere to generate a future stream of benefits. Who bears that cost? Hence, improvements in response diversity in some dimensions could compromise social response diversity by increasing inequalities and putting more pressure on some vulnerable groups of people, which may increase the risks of social unrest.

Strategies to enhance response diversity

To address these challenges to building and maintaining response diversity, we suggest that a critical first step is to create widespread awareness of the meaning of response diversity and its crucial role in responding to unexpected change and sustaining long-run well-being. An appropriate second step would be to search for win–wins, where response diversity is enhanced as a by-product of other well-being-enhancing investments, accompanied by direct investments in response diversity, and we propose strategies to nurture it. While these strategies may be quite straightforward, they require some societal awareness about the role of response diversity and accompanying collective action, which may be more difficult to achieve. It is also important to notice that responding to multiple and compounded crises requires a combination of responses at both local and larger scales. Although many crises are local in nature, in a globalized and interconnected world, local communities are often deeply embedded within larger-scale dynamics. The fostering of strategies to enhance response diversity across multiple scales is therefore crucial.

Strategies for promoting response diversity logically build on the understanding of how it has evolved and been developed through temporal and spatial responses to the variability in natural and social environments, as described earlier. The strategies can be developed by individuals, organizations and governments, and in all cases a diverse portfolio can provide resilience by substituting, complementing or compensating for other elements or variables.

In situations where substitutable options exist, each one is likely to perform best in different contexts or situations. The performance and outcomes of these alternative responses are largely independent of each other, though they may be used in parallel as adaptations to the inherent unpredictability of the future. Examples include investment options in a diversified investment portfolio and duplication of production facilities in different parts of a country or the world. Parts of this response diversity may be lost if some aspects irrelevant to profitability or quality can influence investment decisions. For example, some types of investments (such as in genetically modified organisms) may be boycotted for ethical reasons, or some types of new technology may be at a disadvantage due to unintended restrictions from existing legislation. In ecological systems, an example of substitutable benefits in livestock production on rangelands53 showed that some of the minor grass species are analogues of, and can substitute for, the dominant, more productive species in terms of the ecosystem functions they perform. They differ in terms of their capabilities to respond to environmental stresses and disturbances, such as droughts and high grazing pressure, and can replace the dominants that are reduced or eliminated by such disturbances.

Response options can be complementary. Each is partial and limited in scope, and the outcome of the response depends on other responses. They are adaptations to the multi-dimensional nature of solutions to most problems. A strategy for harnessing this diversity may require simultaneous action on multiple fronts and cognitive capacity or coordination that may not exist, especially under stress or crisis. Responding to climate change in farming, for example, will probably include agricultural production practices, financial hedging and other pricing strategies, and political lobbying as complementary responses. Identifying such complementary options generally requires a systemic approach where the focus of investigations is on the whole picture rather than on specific details54. Such an approach combined with modelling of different options individually and in combination can help provide novel insights — for example, about what combination of policy instruments could address as many planetary pressures as possible55.

In compensatory responses, the failure or absence of one kind of response may require changing the strategy being used — for example, from using incentives (economic, ethical and so on) to using interventions (policy, technological or ecological). At the scale of an individual organism, response diversity to pathogens is reflected in the interplay between behavioural and physiological responses. Behaviours may reduce exposure to pathogens — for example, social distancing can decrease exposure, or varied diets and lifestyles can make the body more resilient to pathogens in general. But when this fails, the immune system must take over. It prepares in advance — it is adaptive and learns from its past. Immune memory (information storage) is “the ability of antigen‐specific cells of the immune system to recognize and “remember” pathogens previously encountered and to produce a qualitatively and quantitatively different response (i.e., faster or more robust) than the first encounter”56. Additionally, the immune system has many redundancies in case one defence fails57.

In complex systems such as cities, developing strategies for building and then harnessing response diversity is a challenge. Urban adaptation to extreme weather events, for example, has traditionally been seen as a problem best addressed through engineered infrastructure solutions (such as levees against flooding or air conditioning against heat waves). However, with changing disturbance regimes and an acknowledgement of the need to address multiple issues, nature-based solutions and hybrid approaches have gained traction58, combining different components and actors to offer alternative ways to implement and govern solutions.

Flood risk reduction in many cities, for instance, is shifting from reliance on highly engineered infrastructure to more integrated solutions with a diversity of designed living systems, such as reducing impervious surfaces, improving wetlands, and building bioswales and green roofs59. This hybridity diversifies the ways in which cities can respond to increasing climate variability60. All of this is further complicated by the fact that different parts of a city may need different responses, depending on social capital and the effectiveness of governance, both of which can vary across a city.

Beyond particular options as outlined above, strategies for building response diversity must involve diversity in goals and capabilities. Human (individual and collective) responses to natural and anthropogenic disruptions are needed not only because of the inherent unpredictability of future conditions or to build in redundancy that can compensate for local failures but also because people differ in their values, concerns and goals. Arguably, such heterogeneity at the individual or cultural level evolved as a strategy to ensure the response diversity of the population as a whole. Individualist versus collectivist societies (and the individuals within them) construct their reality in qualitatively distinct ways and see different classes of risk as actionable61. They not only pursue alternative meta-goals to different degrees (for example, personal utility versus social welfare) but do so by relying differentially on qualitatively disparate decision processes (for example, analytic, emotion-based or rule-based), providing response diversity at the process level62. At the population or group level, response diversity can also be expressed as heterogeneity in different agents’ capabilities. People trained in analytical thinking and social planning assess and use available information differently from those who rely mostly on intuition, personal experience and social networks. The two approaches provide diverse and complementary assessments of societal risks and appropriate responses63.

Traditional strategies to nurture diversity include compensating landowners for setting aside land and wetlands, planting crops to reduce soil erosion, encouraging local markets for locally grown products and providing labelling for traditionally made products, among others. Transforming society towards sustainability, however, requires more: a change of vision, goals and values that can guide system design and provide enough agency to influence institutions and policies4,64. Such norm shifts can be achieved through appropriate and timely supporting policies65, and this change must go beyond behavioural norms to deeper belief system elements. Hall and Lamont66, for example, argue that we need to move beyond the culture of “hard work” and consumption-based status that gives most of the rewards of the economic system to the few. Ideally, the question ‘How do I lead a meaningful life?’ should trigger a wide diversity of answers, not only variations of achieving status through a high consumption level.

Related to the need for norm shifts is a need to encourage variety in practices, rather than just the one ‘best’ way of doing things. Applying top-down control systematically combined with similar types of objectives, such as new public management, is likely to result in uniform solutions. These have proven to be often ill-adapted to disturbances like the COVID-19 pandemic. For example, many regions had rationalized away contingency stocks of medical supplies, which were suddenly needed. Balancing top-down approaches with greater bottom-up inputs could promote a greater variety of practices and solutions to problems.

In practical terms, two complementary areas require particular policies. First, each sector of concern (for example, health, economics, agriculture and industry) must ask and answer the question, ‘What are the likely or possible disruptions this sector might face, and what kinds of response diversity are needed to cope with them?’ Second, they must ask the complementary question, ‘How do the proposed changes aimed at increasing efficiency, savings and so on, also influence changes in response diversity, and what are the possible consequences of these changes in the short and long term?’ Proposed changes in development and operational procedures in governments, industries and corporations should include a formal obligation to explicitly answer these questions.

Given the trade-offs associated with nurturing response diversity (mainly in the form of foregone short-term efficiency), direct investment to nurture diversity in social and ecological systems will probably meet with pushback from special interest groups. While direct public investment will provide diversity, it is also necessary to actively search for spillovers from private investments and actions and to focus particularly on identifying those assets with positive spillovers — that is, positive unintended consequences rather than negative ones. In that context, it is important that public authorities maintain their role of gatekeepers and rule setters rather than trying to please particular industries of national economic importance. Putting in place and enforcing antitrust regulations is one way to ensure that diversity can be maintained.

Broad agreements are easy to reach, but real change requires working out the details of costs, benefits, winners and losers and actually implementing agreements. Identifying and addressing the trade-offs related to response diversity requires the capacity to investigate the consequences of actions in time and space. This will enable the chance to identify negative long-term trends and potential reinforcing feedback loops of concern as well as potentially correlated shocks. Hence, planning capacity focusing on systemic approaches is crucial to that end and can help identify win–win situations as well as shortcuts in an overly complex planning situation54.

Finally, it might be helpful to identify principles that societies can agree on that may contribute to response diversity. Accordingly, in Table 1 we conclude with seven tentative principles to develop policies across ecological, social and economic domains, from local to global scales, for building and maintaining response diversity and therefore resilience. We use the term ‘tools’ in a specific sense: we define tools as a set of tailor-made responses to a particular situation.

Table 1 Seven tentative principles to develop policies across ecological, social and economic domains

To further explore how these principles will translate in different contexts and how they could be implemented across local and regional policies, we foresee that the co-production of knowledge — that is, collaborative processes that convene academic and non-academic actors around problem framing and trust building, through knowledge generation67— can play an important role.