INTRODUCTION

Moments of crisis offer an opportunity to learn about the resilience of social-ecological systems. That is, although we can theorize about what grants resilience in a well-functioning system, it is when things break down that we realize valuable lessons about the mechanisms that build and maintain resilience. Thus, there is much to be learned from observing society’s response to a crisis in terms of gaining a grounded perspective on what works and what does not when navigating turbulence. This is particularly important for resilience thinking, which as a field has developed sophisticated heuristics and principles on how to manage social-ecological systems (see Gunderson and Holling 2002, Walker and Salt 2006, Biggs et al. 2012), but rests on an uneven empirical foundation. The coronavirus pandemic that started in 2020 constitutes a rare chance to observe which resilience principles have been enacted on the ground, thus providing empirical backing to advance theory.

In March 2020, the World Health Organization officially declared the coronavirus crisis a global pandemic. COVID-19 belongs to the large family of coronaviruses that cause respiratory tract diseases, similar to SARS and MERS, and its symptoms are often mild but, for the elderly and those with underlying conditions, it can be fatal. At the time of writing, over 6.2 million people worldwide have died from coronavirus and more than 515 million have contracted the virus (WHO 2022). A vaccine was developed at the end of 2020 and over 11 billion doses have been administered throughout the world, although its deployment remains uneven (WHO 2022). The coronavirus pandemic quickly overwhelmed the capacity of health systems, disrupted supply chains, and prompted emergency lockdowns in most parts of the world. Although the management of the outbreak has varied widely over the course of the pandemic and among countries, no region has been spared, and some countries, such as the United States or India have been thrown into political turmoil as a result.

Beyond its medical dimensions, the coronavirus pandemic has exposed the vulnerability of highly interconnected, complex systems (Walker et al. 2020). In particular, the tightly coupled supply chains along which the economic system is globally organized, increased the susceptibility to shocks that spread quickly across scales and sectors, producing cascading failures (Bryce et al. 2020, Collins et al. 2020, Hynes et al. 2020, Kontogiannis 2021). We have witnessed a doubling of the risk of hunger (FSIN 2020) and disruptions throughout all sectors of the economy (Nicola et al. 2020). As countries navigate their way through the crisis, it has become evident that the multitude of overlapping actors, interests, and administrative boundaries make it difficult to coordinate an effective response and that impacts have and will continue to exacerbate existing inequalities. At a local scale, households and communities have been dealing not only with the burden of disease and grief, but also with a broad variety of shocks from recurrent lockdowns including loss of income, social isolation, and homeschooling.

Given its reach and severity, the COVID-19 pandemic presents an unusual opportunity to document individual and societal responses to a crisis event. We examined COVID coping strategies deployed within the first six months of the coronavirus pandemic to assess which resilience principles previously identified in the literature were actually employed in initial stages of the crisis with the following goals: (1) adding empirical evidence to the resilience principles identified in the literature; (2) offering lessons that can serve in preparation for the next crisis; and (3) exploring the temporal dimensions of navigating crisis and their implication for managing resilience. A growing number of studies have focused on assessing the effectiveness of the strategies initiated by governments and other organizations in their efforts to contain the pandemic by establishing the relationship between mitigation strategies and changes in the rate of transmission (see for example, Li et al. 2019, Haug et al. 2020), others have focused on early lessons from the experience of practitioners (see Linkov et al. 2021). We abstract from concrete strategies and experiences to consider the principles, or mechanisms, behind the strategies that have been commonly deployed. That is, whether a government imposes a school lockdown, or a curfew, or limits social gatherings above a certain number of people, all of these strategies boil down to one resilience principle, namely, managing connectivity. Similarly, whether a hotel is turned into a temporary hospital or a brewery starts producing hand-sanitizer, both strategies are examples that take advantage of existing redundancies. By focusing on the principles of resilience that were deployed, we can draw broader lessons that are generally applicable to social-ecological systems independent of context, thus solidifying the empirical basis for managing resilience.

Resilience remains difficult to define and implement. The term resilience as we use it here was introduced by Holling (1973) as an approach to understanding change and permanence in complex adaptive systems, which was initially applied to the management of natural systems but increasingly used to understand broadly defined social-ecological interactions. The evolution of resilience thinking as an approach has been outlined by others (see Walker and Salt 2006, Quinlan et al. 2015) but here we focus on the principles, or mechanisms, of resilience that are intended to guide implementation and monitoring.

Based on Biggs et al. (2012), we identified 10 resilience-building principles (Table 1), which we subdivided into principles pertaining to the structure of the system, principles that promote systems thinking, and principles relevant to the management of the social-ecological systems (after Quinlan et al. 2015). In terms of structuring a system, the following principles are understood to increase resilience:

• Increasing diversity, this is because having diverse elements in a system implies having alternatives for responding to a crisis, as well, increasing the chances that some parts of the system will be unaffected, or affected differently, by a shock;
• Building redundancy, refers to having elements of the system that are different but fulfill similar or overlapping functions, so that if one fails, other elements can still perform that function;
• Managing connectivity has to do with managing the links between the different parts of the system. Sometimes connecting parts of the system increases its resilience, as in when resources are needed, other times, such as during a health pandemic, isolating parts of the system is the best course of action.

The second set of principles pertains to adopting a systems lens in the framing of a situation. This translates more concretely into the following principles:

• Managing slow variables, which means identifying and tracking variables whose change is more gradual, often goes unnoticed, and yet is connected to thresholds that could lead to the reorganization of the system;
• Managing feedbacks, which has to do with understanding immediate as well as long-range reactions to an intervention in the system and can be amplifying or dampening;
• Framing issues in terms of complex adaptive systems, that is, considering the interactions between the social and ecological elements of the system and across scales, as well as its emergent properties.

Finally, the third set of principles pertains to the management of complex adaptive systems and includes:

• Inviting participation in decision-making processes so as to have a more complete picture and diversity of perspectives that will lead to a better understanding of the issue and trust-building among stakeholders;
• Providing opportunities for learning and experimentation, especially if these can reduce the inherent uncertainty of social-ecological systems;
• Implementing adaptive management, that formalizes an iterative learning-by-doing approach to decision making in which policies are understood as hypotheses; and
• Fostering polycentric governance, which involves having multiple decision-making centers that function in a semi-autonomous manner.

Although we recognize that resilience is above all an approach to a problem rather than a checklist of principles, we have chosen to focus on the principles as a necessary simplification that furthers the implementation and adoption of resilience thinking.

METHODS

Using snowball sampling, we administered an online survey to members of the Resilience Alliance and collaborators. The Resilience Alliance is an international, multidisciplinary research organization, established in 1999, that explores the dynamics of social-ecological systems and a referent of resilience thinking. We deployed a survey between May and July 2020 that asked participants to identify and characterize strategies used for dealing with the coronavirus pandemic with which they were familiar. Participants did not need to have first-hand experience implementing the strategies, just enough familiarity with the strategy to be able to characterize it in terms of: where the strategy was deployed, the scale at which it was deployed (individual, city, region, national), how quickly the strategy was deployed (it was pre-existing, within days, within weeks), and by whom (individuals, cities, national or state governments, businesses, grassroots, organizations, or partnerships). Last, participants were asked to identify which resilience principles were embodied in the strategy (see Fig. 1). For example, a participant may point out how turning an art gallery into a temporary emergency health shelter is an example of functional redundancy, or that lockdowns were a way of managing connectivity. At the end of the survey participants had the chance to suggest additional people to include in the survey. This study was approved by Arizona State University’s institutional review board (STUDY00011979) and all participants provided informed consent to take part in the study.

The data collected were reviewed by the authors although no effort was made to correct the participants’ characterization of their responses with regard to the resilience principles that participants assigned to the individual strategies (see supplemental materials). The first part of the results up to the correlation table uses descriptive statistics and relies on this data. A second analysis was done to identify and group strategies that formed a pathway out of the crisis. The pathways were derived by thematically coding the following open-ended questions from the survey “describe an intervention in response to the COVID-19 crisis that builds resilience (e.g., social distancing)” and “what are the strengths and limitations of this response?” Each response was inductively coded to derive insights directly from the data. Inductive coding is particularly relevant for exploratory studies seeking to derive new insights, as is the case here (Linneberg and Korsgaard 2019). Coding here refers to identifying segments of meaning in one’s data and summarizing those with a word or sentence (Saldaña 2013). A first order coding, which we called tags, identified a host of themes related to the interventions in response to COVID-19 (for e.g., new supply chains; going local). These themes were then aggregated into a higher-level set of coding to synthesize codes that were related to one another (Gioia et al. 2013, Linneberg and Korsgaard 2019). For example, the tags “government financial aid,” and “providing free healthcare” were related under a higher-level theme called “welfare state.” These higher-level themes represented the different pathways identified.

Limitations of the methods

Snowball sampling tends to exaggerate consensus because participants refer to people in their networks with whom they tend to share common values. This factor likely plays in our data because we chose to survey members of the Resilience Alliance, which form a somewhat cohesive group already. However, given that the survey required people to assign resilience principles to each strategy, it was important that the sample population would have enough knowledge and background on resilience thinking, and with the Biggs et al. (2012) paper, to be able to make this determination.

In addition, because the survey was deployed between May and July 2020, some of the respondents had already undergone the first wave of the pandemic, such as those in Asia, Africa, and Europe, whereas those in the Americas were in the thick of the first wave. Therefore, the responses capture different degrees of uncertainty and hindsight.

RESULTS

The survey was distributed to 124 resilience researchers and practitioners around the world and our response rate was 40%. The results contained 61 distinct strategies from 49 respondents (participants were given the option of submitting several strategies). Respondents were asked to submit a strategy with which they were familiar but there was no requirement for the strategy to have been implemented in the place where they lived, although this was the case for many of the submitted strategies. Overall, 17% of the submitted strategies were either global or regional in scope, e.g., European Union, pan-African, and 2% were online strategies. The United States and South Africa were the countries that had the most strategies identified (17% each), followed by Canada, Sweden, Spain (8% each), and Australia (6%). Other countries represented in the database included China, Singapore, Mexico, Germany, the Netherlands, New Zealand, Kenya, and Nigeria (see Fig. 2).

The database contains responses that were deployed at a variety of spatial scales and often at multiple scales simultaneously. Almost half of the strategies (49%) were implemented at the household or individual level, 41% were actions implemented at the city level, and 38% at the national level. About one-third of strategies were implemented at intermediate administrative scales, with 30% being implemented by states or provinces, and 28% being regional in scope, such as counties or Indigenous communities. International strategies accounted for 18% of the actions contained in the database; this meant strategies that were implemented in more than one country simultaneously. When it comes to who implemented the strategies identified, the majority originated from national and state governments (85%), although almost half of them were implemented by individuals and households (49%). Businesses were the implementers of 39% of the strategies in our database, the majority of which refers to strategies adopted by small businesses. City governments, grassroots, and large organizations implemented less than one-third of the strategies identified. Finally, strategies were deployed relatively quickly. Most of the strategies identified in our database were implemented within weeks (43%) or days (25%), with a small percentage of them having adaptations on something that was pre-existing (7%). There were only 11% of strategies that were implemented within months, but this reflects the fact that our survey was deployed in May and July 2020, so within months of the first coronavirus wave. Please refer to Figure 3.

Resilience principles

The principles of resilience ranked in order of frequency are shown in Figure 4 below. The principle that was most frequently cited in the set of strategies was managing connectivity, followed by managing feedbacks, and establishing processes for learning. Strategies that increased redundancy, participation, diversity, and the practice of adaptive management were also important. Last, and this may be a reflection of the timing of our survey, there were fewer strategies that demonstrated a systems understanding of the issues, such as those that tackle slow variables or promote polycentric governance.

Connectivity

Connectivity was the resilience principle most often identified in the strategies, which is not surprising given the nature of the crisis as a health pandemic. That is, viral illnesses spread through being in close proximity to an infected person, so it is sensible that early in the pandemic, strategies for reducing contact at all levels, from travel bans to sheltering-in-place to school closures, were broadly implemented as key to reducing contagion. Countering the broad move to isolate to reduce viral spread, participants also identified a number of ways that people devised to maintain, replace, or create new ties, presumably to satisfy the gap in our networks of relationships caused by the sudden removal of physical interaction.

In general, there were two tendencies with regard to the new forms of connectivity that emerged during the pandemic. One was to create virtual ties. For example, there were a multitude of authors that offered free online readings for children, and most white-collar work was adapted for teleworking from home. The second way in which connectivity changed during the pandemic was a renewed emphasis on local ties. Bottom-up solidarity networks that organized the delivery of essentials to populations at risk emerged, or were reinvigorated, in many communities around the world. For example, SOLIVID is an online platform that emerged in Barcelona (Spain) to map and connect grassroots solidarity efforts related to COVID-19, such as offerings of child care for working parents, sharing of classroom resources among teachers, or general mental health support. In Cuenca (Ecuador), people in need of food or supplies during lockdowns would hang a white flag on their door and neighbors responded by bringing what was needed.

Learning

Respondents identified two modes of learning taking place simultaneously during the pandemic: the first type of learning can be characterized as unidirectional transmission of information. Respondents identified a variety of ways and platforms that helped people be informed, for example, by listening to government bulletins to obtain updates on policy changes or by consulting a web dashboard on COVID cases. Resilience theory does not consider this form of passive learning as enhancing resilience because it is limited to the consumption of data, however others have indicated that appropriate management of information is important during health crises to “reduce illness, save lives, and maintain societal structures” (Reynolds and Quinn 2008:16S). Information for crisis and emergency risk communication needs to be prompt, accurate, credible, empathetic, respectful, and geared toward action (CDC 2018), although this is far from what we witnessed during the coronavirus pandemic, which was characterized by the constant spread of misinformation (Motta et al. 2020).

The second form of learning that participants pointed out was associated with activities that were experimental in nature and that captured immediate adaptive responses to changing conditions, such as updates of official guidelines as new information became available. That is, respondents recognized that there is a degree of learning in the initial experimentation that follows a crisis. For example, small businesses and restaurants had to quickly come up with creative ways of staying viable, which led to the development of delivery and pick-up services for most products, as well as ways of experimenting with the use of outdoor spaces, e.g., extending restaurant seating into sidewalks. Similarly, the elementary educational system used a variety of ways to continue teaching school children that included experimenting with online instruction platforms, hybrid models of learning, and revamped spaces and rules for in-person instruction. The latter form of learning has elements of adaptive management and, indeed, the two appear together in the co-occurrence matrix (Fig. 4). Although adaptive management is more structured, in both cases, learning occurs as an iterative process in response to a changing environment for the purpose of gaining new knowledge on how things work and improve future responses.

Feedbacks

Feedbacks are mechanisms to control the internal dynamics of the social-ecological systems by designing an intervention that either amplifies or dampens an initial response. The majority of feedbacks that participants identified in the response to the pandemic were dampening feedback loops, which are most useful for maintaining a system within bounds. In this case, dampening feedback loops were usually connected to reducing connectivity as noted in the co-occurrence table. Thus, the examples are similar to the ones cited earlier in relation to managing connectivity, such as the way in which bans on air travel, city-wide lockdown mandates, or school closures, which helped to reduce the numbers of in-person interactions where potential transmission could happen, which ultimately reduced the rate of contagion.

There were fewer examples of amplifying feedback loops. One worth noting was the way in which engaging in small solidarity efforts boosted morale. Mental health impacts due to grief, isolation, exhaustion, and anxiety soon became a concern paralleling the concern for contagion (Cullen et al. 2020). Participants identified that being engaged in actions that helped others, even in small ways, such as by delivering food to those in need or sending emails to check-in, were important ways of building positive energy that helped sustain them and allowed them to continue doing more voluntary actions. Indeed, others have pointed out how traumatic events create conditions for bonding and how the solidarity and social ties that emerge in the aftermath of a crisis play a key role in getting out of it (Elcheroth and Drury 2020).

Last, although feedbacks are considered mechanisms for managing the system and we have focused on those that were intentionally set up, they can also occur spontaneously, particularly amplifying feedbacks. For example, the panic buying that was observed at the beginning of the crisis emerged spontaneously, that is, people perceived a potential interruption in supply and began stocking up on essential items, causing store supplies to dwindle, which in turn increased the perception of scarcity and further induced more panic buying. Amplifying feedback loops can be problematic for maintaining resilience because they can cause a system to spiral out of control.

Redundancy

The types of measures that participants identified as increasing redundancy were associated with repurposing physical spaces and processes in which there was sufficient overlap in form or purpose. For example, open-air spaces such as parking lots and sidewalks in front of restaurants were converted into patios, large indoor spaces such as art galleries and hotels were used to house COVID patients temporarily as hospitals reached capacity. Examples of repurposing include the adaptation of manufacturing processes to supply health products in high demand, e.g., some distilleries began manufacturing hand sanitizer, large industrial manufacturers such as Honeywell in the United States began producing disposable facemasks.

Redundancy was also understood in a social sense as providing additional social safety nets to address the loss of income and purchasing power, which was one of the main challenges for households during periods of strict lockdown. Respondents highlighted aid programs sponsored by governments, for example, how Canada provided monthly relief cheques to a large swath of its population during the initial months of the crisis. As well, respondents identified bottom-up organizing efforts to create safety nets that had local reach and aimed to protect local businesses. For example, in Sweden as farmers lost revenues due to restaurant closures, REKO-rings (a scheme similar to community-sponsored agriculture) sprung up to ensure that these farmers were able to connect with buyers and sell their produce.

Participation

Within our data set, participation encompasses two forms of engagement. One understanding of participation was made up of strategies that facilitated people coming or acting together. That is, respondents identified strategies in which people partook in a collective activity, such as the opening of streets for pedestrian use or webinars for kids by children’s authors and illustrators, as examples of participation. Although in both cases participants were passive recipients, these strategies were perceived as participatory because they embodied a sense of collectivity and indeed, the matrix shows co-occurrence between the participation and connectivity principles.

By contrast, more academic understandings of participation presume some degree of power sharing in processes of decision making (sensu Arnstein 2019). The strategies that fall under this category demonstrate a higher degree of agency. In an example from South Africa, the nature-based tourism sector that was greatly affected during the pandemic reached out to donors to find ways of supporting their sector by switching to payments for ecosystem services as a means of financing conservation. In this case, it is clear that those involved in the tourism sector worked together to come up with solutions that benefited their group and that in doing so they claimed power to decide how best to organize their sources of financing. We also note in the co-occurrence matrix how participation appears together with learning and adaptive management and hypothesize that the pandemic created the conditions, or the necessity, for groups to self-organize and experiment together, for example, how bubble family arrangements organized to provide for childcare

Diversity

Roughly one-third of responses included taking advantage of diversity in some form. These ranged from a diversity of transportation methods to replace public transport during lockdown (e.g., bike and walk-friendly zones) to a diversity of food supplies and sources, (e.g., restaurants offering delivery, selling through farmers’ markets) to a diversity of governmental responses. Diversity builds resilience because it provides a repertoire of alternatives. Importantly, diverse elements will respond differently to the same shock, so even if some parts of the system are affected, others can still carry on. Thus, the critical characteristic that increases the resilience of a system is response diversity.

In the co-occurrence matrix, diversity coincides with learning, which reinforces the idea that the initial months of the pandemic created the conditions for quick experimentation and having, or creating, diverse options provided opportunities for comparison, whether it was governmental agencies trying different approaches at different scales (see polycentricity below), different approaches to teaching (in-person pods, online, or other), or different types of socialization (zoom happy hours, driveway gatherings, and zoom dating).

Adaptive management

The initial months of the coronavirus pandemic were characterized by both high uncertainty and a sense of urgency. Thus, many of the strategies identified by our survey respondents had the purpose of learning while managing, which is the essence of adaptive management. All sectors of society had to adapt quickly, from businesses to households to governments, while new information about the disease was still emerging. Because of the degree of uncertainty, most adaptations done in this context had a degree of experimentation and learning-by-doing, which has been mentioned before, e.g., restaurants modifying their business model from lockdown, to pick-up and delivery, to outdoor dining, to comply with changing regulations. One important aspect of adaptive management has to do with monitoring the impact of policies. This was clear in the strategies of governments and health agencies that continuously updated their directives based on new information, as well as in the phased responses according to levels of transmission, health dashboards, and contact tracing apps.

The following three principles appeared less frequently in the set of strategies that participants identified: Adopting a systems lens, which means approaching a situation considering its dynamic nature, inherent complexity, emergent properties, and nonlinear behavior. Among the few responses that suggested a systems lens was the way in which some federal governments chose to invest in Indigenous remote communities as a means of building general resilience in the face of COVID. Another highlighted direct financial help to citizens for the same reason. A third example was how contact tracing was digitized. Although what constituted taking a systems lens depended more on the framing of the respondents than perhaps other variables. All of these strategies seem to speak to interventions that considered larger temporal or spatial scales and second order effects.

Slow variables were generally poorly identified, i.e., most examples that identified slow variables as a feature of building resilience did not articulate what aspect of the intervention constituted the slow variable. One example is of traveling less and consuming less. Here, we can infer potential slow variables of behavioral shifts and lifestyle changes that reduce exposure (reducing detrimental types of connectivity). Another more direct example is from South Africa where mental health support has been provided for healthcare workers on the frontlines of combatting the virus. The idea being that the deterioration of mental health may be gradual and go unnoticed for a prolonged period of time before the person is at the point where they need help.

Finally, there were few examples of polycentricity in the response to COVID. Perhaps the most consistent way in which polycentricity shows up in the response to the coronavirus pandemic was in the way in which different levels of government were able to adapt their mandates to their jurisdictions as they saw fit. However, there is a thin line between polycentricity and simply an uncoordinated, politicized response. More productive examples of polycentricity at the grassroots scale include linking community action networks in Cape Town to each other horizontally and vertically to learn and share in a polycentric manner.

Pathways out of the crisis

Last, we also considered constellations of strategies that converged to create possible pathways out of the pandemic. All of the responses were categorized into broader pathways based on the similarities of strategies adopted across cases. Four main pathways were identified (see Table 2):

1. The securitization pathway emphasized top-down measures to control and restrict the movement of people to stop the contagion. Although physical distancing is warranted for reducing viral transmission, this pathway has undertones of authoritarianism that potentially impinges on broader civil liberties and grants undue power to authorities beyond the pandemic. Furthermore, prioritizing strict adherence to social distancing measures does not consider other important aspects of people’s livelihoods, such as the loss of income or the mental health impacts.
2. The grassroots pathway recognized the importance of bottom-up leadership. This pathway featured a strong mobilization of grassroot efforts, and in particular a restructuring of food supply chains and the delivery of essential goods to vulnerable people. COVID-19 responses also bring a restructuring of public spaces that support new collaborations to reimagine commons (urban green spaces, transportation systems) to provide safe social spaces and contribute to mental and physical health.
3. The online pathway illustrated the emergence of new ways of living and working online that allowed flexibility and social distancing but are also contributing to isolation and mental health issues. Some people pointed to a new emerging common that was global, rather than local, in nature which was the emergence of greater collaboration to solve the global pandemic, for example, by enhancing data sharing and international collaborations in dealing with the health crisis.
4. The welfare state pathway summed the idea that the way out of the pandemic required ramping up systems of social safety nets that extended care to the most vulnerable. Similar to the securitization pathway, this remains top-down but sees the role of government as a provider rather than enforcer.

These different pathways represent broad courses of action that emerged immediately after the pandemic that are not, however, mutually exclusive nor all-encompassing. Other pathways may emerge in the upcoming months and pathways often coexist simultaneously, especially because these are often driven by different sets of actors at different scales that focus on managing different aspects of the pandemic. Nonetheless, there are tensions between some of the pathways, for instance between increased securitization which limits the agency of individuals to increase safety and grassroots actions that often aim to activate people’s capacity to act.

Each of the pathways identified above is associated with activating a different set of resilience mechanisms (Fig. 5). The securitization pathway emphasized the two resilience mechanisms that are most prominent in the data overall, that is, the management of connectivity and feedbacks. The grassroots pathway also relied on the management of connectivity followed by diversity. The online living pathway focused on learning followed by connectivity. The welfare state pathway highlighted managing feedbacks and redundancy as its main resilience mechanisms. Thus, although all the pathways emphasize addressing system configurations (i.e., managing diversity, redundancy, and connectivity), the pathways also relied on managing the system (i.e., most frequent resilience strategy adopted in online living is learning) and adopting a system’s approach (i.e., primarily focus on feedbacks in the welfare state pathway) to increase resilience in light of the pandemic.

DISCUSSION

The purpose of our analysis was to understand the principles of resilience underpinning the initial responses to the COVID-19 pandemic, noting that although a pandemic constitutes a specific public health challenge, it can also be taken as a proxy for other large societal shocks including future pandemics, the impacts of climate change, and beyond. In doing so, we seek to provide empirical evidence for the principles of resilience that have been identified in the literature and also consider what lessons might be applicable for an upcoming crisis. Out of the resilience principles identified in the literature (Biggs et al. 2012), our survey identified three principles that shaped the initial response to the crisis, which were managing connectivity, enhancing learning, and the management of feedbacks. The principles of increasing diversity and redundancy, inviting participation, and practicing adaptive management were next in order of frequency. Finally, the principles related to adopting a systems lens, managing slow variables, and polycentricity were less prominent in our responses. We also note the formation of two predominant pathways out of the crisis that occurred simultaneously but are guided by opposite drivers: one relies on state-enforced mandates to stop transmission and the other relies on local, bottom-up approaches that look out for vulnerable groups. These two pathways were not mutually exclusive but coexisted highlighting the diversity of responses of different actors and at different scales.

The variation in the degree of implementation of the different resilience principles during the initial stages of the coronavirus pandemic suggests that some principles are more useful in a crisis situation than others. Namely, the most important resilience principle at the onset of the coronavirus crisis was the management of connectivity. Although the importance of connectivity can be readily ascribed to this being a health crisis, this principle likely applies more broadly because most crisis situations require both isolating the part of the system that has become dysfunctional while connecting it to resources that can help fix it. For example, when flooding occurs, it makes sense to cut traffic to the flooded area while also sending rescue helicopters to airlift survivors. The lesson here is that managing the connectivity between the elements of a system is critically important to maintaining its resilience while navigating a crisis, and that resilient systems need to have both shutoff mechanisms and alternative connection routes that can be activated quickly.

The other two principles of resilience that were implemented with higher frequency during the initial months of the coronavirus pandemic had to do with learning and managing feedbacks. Both of these speak to the need to navigate a highly uncertain and variable environment, which requires continuously reassessing what is known and then setting the proper incentives to guide action. In terms of learning, we remark that respondents included passive forms of learning as well as active, experiment-driven lessons. However, only the latter is usually considered relevant for building resilience. That is, traditional learning mechanisms and institutions are likely too slow during an evolving crisis. For example, during the early years of the AIDS/HIV crisis, gay men’s knowledge was more sophisticated in understanding what constituted safe sex practices than the medical establishment (Escoffier 1998). Part of this knowledge would have been learned through experimentation and channels that were not accessible to formal institutions. Thus, systems that are able to connect with vernacular and local knowledge sources can get key insights that will help them navigate uncertainty. This could also include setting up infrastructure to obtain signals from large, unofficial, and decentralized sources of information, for example, the Eyes on the Rise app developed in South Florida (US) that encourages citizens to report flooding events, or the HarrassMap app developed in Egypt for women to report incidents of sexual harassment. Both apps tackle very different crises but they capture crowdsourced knowledge that is anonymous and spatially explicit. Finally, to use learning mechanisms effectively to build resilience, it is necessary to understand the power dynamics associated with managing knowledge and information during a crisis. The coronavirus epidemic may have presented an extreme case of misinformation, particularly in countries like the United States where the crisis and its solutions have been highly politicized from the outset (Motta et al. 2020, Ratzan et al. 2020), however, in all cases it is important to recognize that the narratives spun by governments, health agencies, and media are responding to specific pressures and interests.

Focusing on feedbacks is important for two reasons: first, feedback mechanisms set up the incentive structure necessary for a coordinated response, which starts to get to the crux of the governance of crises. Here, there are important considerations not only on what prompts people to act collectively in an uncertain environment (see Elcheroth and Drury 2020) but also on how to balance speed with analysis, centralization with decentralization (Janssen and van der Voort 2020) and the underlying ethical considerations. The COVID-19 crisis revealed a variety of governance approaches across countries and cultures; from China to Sweden, we see variations in terms of the use of incentives or punishment as mechanisms for enforcement as well as the value placed on individual freedoms or collective actions (Yan et al. 2020). The second aspect of feedback mechanisms that is important for the management of crises is that, if set correctly, feedbacks provide an opportunity to understand the underlying workings of the system by observing the response to an intervention. This is the essence of adaptive management, where policies are understood to be testing hypotheses about the working of a complex system. The pandemic created conditions for rapid adaptation as people were actively experimenting, taking advantage of pre-existing diverse conditions, or setting up small experiments to understand what might work and why. There are a number of reasons why people are more willing to experiment during a crisis, for one, they are often faced with novel conditions for which they were not prepared. For example, working families had to think of alternative arrangements to provide childcare and education once schools were closed. At this point families were likely willing to consider a variety of arrangements, such as bubble families, because maintaining the status quo was not a viable option. As well, crises are moments of shifting baselines, of questioning and challenging core beliefs and assumptions that might no longer hold, thus opening the solutions space to experimental ideas (Smith and Elliott 2007).

At the other end of the spectrum were the principles that were implemented less frequently, namely, managing slow variables, adopting a systems lens, and polycentric decision making. This may be the case because all of these mechanisms require long lead times and need to be implemented before the actual crisis. That is, slow variables can be drawn upon during times of crises but they cannot be set up during one. For example, households that had savings were able to use these to shelter in place for longer and purchase supplies more easily. Critical slow variables need not be economic; Lugo (2020) identified social, ecological, and technological variables that played a role in the impact and subsequent recovery from hurricane Maria in Puerto Rico that ranged from government corruption to vegetation to the dependence of the island on fossil fuels. All of these factors were set up for success, or failure, prior to the actual hurricane. The takeaway is that because slow variables are tied to the overall resilience of the system, they can provide buffering capacity and options during turbulent times, but they need to have been in place well in advance.

Similar to the slow variables, a systems approach requires that it be present before the crisis strikes for it to be useful, which is not often the case. In particular, the pandemic revealed the fragility of living in a highly interconnected and complex world that also values maximizing efficiency. This is perhaps best exemplified by several sectoral analyses of supply chains during the crisis, e.g., food (Béné 2020, Love et al. 2021) or medical materials (e.g., Gereffi 2020). These analyses showed how the globalized economy that we live in relies on lean production, just-in-time delivery, vertical integration, and long-distance trading, which is the economic equivalent of the pathology of command-and-control (sensu Holling and Meffe 1996). That is, we have created a highly profitable (although unfair) economic system that performs efficiently, under a set of narrowly defined conditions. For example, to reduce inventory and storage costs, just-in-time delivery uses a set of logistics arrangements that require careful orchestration to ensure that all of the connections from manufacturing to delivery occur in a timely fashion. If one part of this system of logistics experiences a delay, the entire operation will be affected because there is little redundancy in this arrangement that can help cope with disturbances. Last, polycentric governance arrangements also need to be set up in advance but they are inherently difficult to implement, and it often emerges and self-organizes when conditions allow and institutional arrangements facilitate it. As well, polycentricity often serves as an enabler for other resilience-building principles including increasing participation, improving connectivity and diversity, and creating opportunities for learning and experimentation (Schoon et al. 2015).

What emerges from our analysis is that there is a strong sense of temporality that determines when each of the resilience principles is most useful. If we think of complex systems as having periods of stability punctuated by moments of rapid change as in Holling’s adaptive cycle (Gunderson and Holling 2002), then we can assign the principles of resilience identified in the literature as follows: first, navigating an evolving crisis (κ to Ω), as we have documented, requires managing connectivity in a way that isolates the issue but provides alternative routes to address the situation; activating learning mechanisms that emphasize the gathering and disseminating of information; and setting up appropriate feedbacks to direct action, resources, and people. Second, in the aftermath of the crisis as the society moves toward reorganization (Ω to a and eventually to r), there are key principles that need to be considered to enhance resilience for a future event. These include choosing pathways that build redundancy and diversity across a polycentric governance structure and that invite broader participation in the design process. Finally, as the new system consolidates (r to κ), it is important to resist simplification and maintain a systems lens (Abreu Saurin 2020, Hynes et al. 2020, Walker et al. 2020, Kontogiannis 2021), practice adaptive management, and monitor slow variables that are tied to the resilience of the overall system (see Fig. 6).

Future research

Our analysis of strategies deployed in the aftermath of the coronavirus crisis adds empirical evidence to the usefulness of resilience principles identified in the literature. However, by its very design, our work is the beginning of a larger conversation. We consider additional elements that impacted the first months of the initial wave of the coronavirus pandemic to identify possible areas of future study.

First, many of the responses were colored by a sense of solidarity and expressed a normative orientation that is usually mute in resilience thinking. For example, some answers spoke about prioritizing the needs of the most vulnerable. Also, answers tended to favor the local scale, highlighting horizontal linkages, e.g., checking on neighbors, rather than nestedness across scales. We are probably seeing the confluence of resilience and vulnerability ideas, which are often thought to be the inverse of one another but in fact have distinct lineages (Eakin and Luers 2006, Miller et al. 2010). Whereas resilience is rooted in systems science, vulnerability is rooted in disaster studies and has a stronger normative base. It seems that a clearer articulation between the two fields is necessary to explore both the systemic and the human implications of crises.

Second, our work identified key principles to navigating the pandemic in general, however, it would be interesting to compare different communities and their responses based on preexisting levels of resilience and adaptive capacity. For example, how did diversity factor in? That is, were some communities more resilient because they were more diverse and what type of diversity was most important? What were the conditions that allowed for quick learning and experimentation? Associated with that, as societies start to move toward recovery, how are decision makers planning for their desired resilience pathway? What are the key slow variables that need to be put in place so that communities build resilience for the known unknowns, i.e., whether pandemics, natural disasters, or political instabilities? And what are trajectories for recovery and how do they set things up for the next one? Finally, the role of polycentricity in governance remains difficult to assess. Can polycentric governance systems be planned or are they emergent properties of the social-ecological systems in which they are a part? Thus, important comparative work is still missing. Going forward, we can envision ways in which coordinated responses across municipalities and multiple scales of governance emerge, provide mechanisms to learn from successes and failures, and improve the overall system resilience to the shock of an epidemic.

CONCLUSION

We have examined initial responses to the coronavirus pandemic from a resilience perspective to understand the mechanisms that became activated as coping responses on the ground. We invited members of the Resilience Alliance and collaborators to identify the resilience principles behind these initial strategies. Although the sample is small, this was necessary to ensure that they could identify resilience principles because our study used deductive logic to interpret what was happening on the ground as the crisis unfolded. We found that the most important, or frequent, resilience principles enacted during the pandemic had to do with managing connectivity, which included isolating measures to reduce transmission while creating alternative ways of staying in touch. Similarly, learning was key in the early stages of the pandemic. Some of the learning was simply about gathering information. However, participants pointed to learning through experimentation, particularly as a way of figuring out how to proceed, as an important element to navigating the developing crisis. The third principle that was present in the early stages of the pandemic was the establishment of feedback mechanisms that helped guide the behavior and provide the infrastructure for governance. Our research also suggested two broad pathways out of the crisis that emphasized opposite qualities. Both occurred simultaneously but emphasized alternative routes, one being top down, the other grassroots. Our research suggests that there is an important temporal component that makes some principles of resilience more applicable than others, specifically during times of crisis. However, we say this knowing that our research reports the results of a small sample and that there is still critical comparative work that is missing.

LITERATURE CITED

Abreu Saurin, T. 2021. A complexity thinking account of the COVID-19 pandemic: implications for systems-oriented safety management. Safety Science 134:105087. https://doi.org/10.1016/j.ssci.2020.105087

Arnstein, S. R. 2019. A ladder of citizen participation. Journal of the American Planning Association 85:24-34. https://doi.org/10.1080/01944363.2018.1559388

Béné, C. 2020. Resilience of local food systems and links to food security - a review of some important concepts in the context of COVID-19 and other shocks. Food Security 12:805-822. https://doi.org/10.1007/s12571-020-01076-1

Biggs, R., M. Schlüter, D. Biggs, E. L. Bohensky, S. BurnSilver, G. Cundill, V. Dakos, T. M. Daw, L. S. Evans, K. Kotschy, A. M. Leitch, C. Meek, A. Quinlan, C. Raudsepp-Hearne, M. D. Robards, M. L. Schoon, L. Schultz, and P. C. West. 2012. Toward principles for enhancing the resilience of ecosystem services. Annual Review of Environment and Resources 37:421-448. https://doi.org/10.1146/annurev-environ-051211-123836

Bryce, C., P. Ring, S. Ashby, and J. K. Wardman. 2020. Resilience in the face of uncertainty: early lessons from the COVID-19 pandemic. Journal of Risk Research 23:880-887. https://doi.org/10.1080/13669877.2020.1756379

Centers for Disease Control (CDC). 2018. Crisis and emergency risk communication: introduction. Centers for Disease Control, Atlanta, Georgia, USA. https://emergency.cdc.gov/cerc/ppt/CERC_Introduction.pdf

Collins, A., M.-V. Florin, and O. Renn. 2020. COVID-19 risk governance: drivers, responses and lessons to be learned. Journal of Risk Research 23:1073-1082. https://doi.org/10.1080/13669877.2020.1760332

Cullen, W., G. Gulati, and B. D. Kelly. 2020. Mental health in the Covid-19 pandemic. QJM: An International Journal of Medicine 113:311-312. https://doi.org/10.1093/qjmed/hcaa110

Eakin, H., and A. L. Luers. 2006. Assessing the vulnerability of social-environmental systems. Annual Review of Environment and Resources 31:365-394. https://doi.org/10.1146/annurev.energy.30.050504.144352

Elcheroth, G., and J. Drury. 2020. Collective resilience in times of crisis: lessons from the literature for socially effective responses to the pandemic. British Journal of Social Psychology 59:703-713. https://doi.org/10.1111/bjso.12403

Escoffier, J. 1998. The invention of safer sex: vernacular knowledge, gay politics and HIV prevention. Berkeley Journal of Sociology 43:1-30.

Food Security Information Network (FSIN). 2020. 2020 global report on food crises. World Food Programme, Rome, Italy. https://www.wfp.org/publications/2020-global-report-food-crises

Gereffi, G. 2020. What does the COVID-19 pandemic teach us about global value chains? The case of medical supplies. Journal of International Business Policy 3:287-301. https://doi.org/10.1057/s42214-020-00062-w

Gioia, D. A., K. G. Corley, and A. L. Hamilton. 2013. Seeking qualitative rigor in inductive research: notes on the Gioia methodology. Organizational Research Methods 16(1):15-31. https://doi.org/10.1177/1094428112452151

Gunderson, L. H., and C. S. Holling. 2002. Panarchy: understanding transformations in human and natural systems. Island, Washington, D.C., USA.

Haug, N., L. Geyrhofer, A. Londei, E. Dervic, A. Desvars-Larrive, V. Loreto, B. Pinior, S. Thurner, and P. Klimek. 2020. Ranking the effectiveness of worldwide COVID-19 government interventions. Nature Human Behaviour 4:1303-1312. https://doi.org/10.1038/s41562-020-01009-0

Holling, C. S. 1973. Resilience and stability of ecological systems. Annual Review of Ecology and Systematics 4:1-23. https://doi.org/10.1146/annurev.es.04.110173.000245

Holling, C. S., and G. K. Meffe. 1996. Command and control and the pathology of natural resource management. Conservation Biology 10:328-337. https://doi.org/10.1046/j.1523-1739.1996.10020328.x

Hynes, W., B. Trump, P. Love, and I. Linkov. 2020. Bouncing forward: a resilience approach to dealing with COVID-19 and future systemic shocks. Environment Systems and Decisions 40:174-184. https://doi.org/10.1007/s10669-020-09776-x

Janssen, M., and H. van der Voort. 2020. Agile and adaptive governance in crisis response: lessons from the COVID-19 pandemic. International Journal of Information Management 55:102180. https://doi.org/10.1016/j.ijinfomgt.2020.102180

Kontogiannis, T. 2021. A qualitative model of patterns of resilience and vulnerability in responding to a pandemic outbreak with system dynamics. Safety Science 134:105077. https://doi.org/10.1016/j.ssci.2020.105077

Li, Y., H. Campbell, D. Kulkarni, A. Harpur, M. Nundy, X. Wang, and H. Nair. 2020. The temporal association of introducing and lifting non-pharmaceutical interventions with the time-varying reproduction number (R) of SARS-CoV-2: a modelling study across 131 countries. Lancet Infectious Diseases 29:193-202. https://doi.org/10.1016/S1473-3099(20)30785-4

Linkov, I., J. M. Keenan, and B. D. Trump. 2021. COVID-19: systemic risk and resilience. Springer, Amsterdam, The Netherlands.

Linneberg, M. S., and S. Korsgaard. 2019. Coding qualitative data: synthesis guiding the novice. Qualitative Research Journal 19:259-270. https://doi.org/10.1108/QRJ-12-2018-0012

Love, D. C., E. H. Allison, F. Asche, B. Belton, R. S. Cottrell, H. E. Froehlich, J. A. Gephart, C. C. Hicks, D. C. Little, E. M. Nussbaumer, P. Pinto da Silva, F. Poulain, A. Rubio, J. S. Stoll, M. F. Tlusty, A. L. Thorne-Lyman, M. Troell, and W. Zhang. 2021. Emerging COVID-19 impacts, responses, and lessons for building resilience in the seafood system. Global Food Security 28:100494. https://doi.org/10.1016/j.gfs.2021.100494

Lugo, A. E. 2020. Effects of extreme disturbance events: from ecesis to social-ecological-technological systems. Ecosystems 23:1726-1747. https:// doi.org/10.1007/s10021-020-00491-x

Miller, F., H. Osbahr, E. Boyd, F. Thomalla, S. Bharwani, G. Ziervogel, B. Walker, J. Birkmann, S. van der Leeuw, J. Rockstróm, J. Hinkel, T. Downing, C. Folke, and D. Nelson. 2010. Resilience and vulnerability: complementary or conflicting concepts? Ecology and Society 15(03):11. https://doi.org/10.5751/ES-03378-150311

Motta, M., D. Stecula, D., and C. Farhart. 2020. How right-leaning media coverage of COVID-19 facilitated the spread of misinformation in the early stages of the pandemic in the US. Canadian Journal of Political Science/Revue canadienne de science politique 53(2):335-342 https://doi.org/10.1017/s0008423920000396

Nicola, M., Z. Alsafi, C. Sohrabi, A. Kerwan, A. Al-Jabir, C. Iosifidis, M. Agha, and R. Agha. 2020. The socio-economic implications of the coronavirus pandemic (COVID-19): a review. International Journal of Surgery 78:185-193. https://doi.org/10.1016/j.ijsu.2020.04.018

Quinlan, A. E., M. Berbés-Blázquez, L. J. Haider, and G. D. Peterson. 2015. Measuring and assessing resilience: broadening understanding through multiple disciplinary perspectives. Journal of Applied Ecology 53:677-687. https://doi.org/10.1111/1365-2664.12550

Ratzan, S. C., S. Sommariva, and L. Rauh. 2020. Enhancing global health communication during a crisis: lessons from the COVID-19 pandemic. Public Health Research and Practice 30:3022010. https://doi.org/10.17061/phrp3022010

Reynolds, B., and S. C. Quinn. 2008. Effective communication during an influenza pandemic: the value of using a crisis and emergency risk communication framework. Health Promotion Practice 9(4_suppl):13S-17S. https://doi.org/10.1177/1524839908325267

Saldaña, J. 2013. The coding manual for qualitative researchers. Second Edition. Sage, Thousand Oaks, California, USA.

Schoon, M. L., M. D. Robards, C. L. Meek, and V. Galaz. 2015. Principle 7 - Promote polycentric governance systems. Pages 226-250 in R. Biggs (Oonsie), M. Schlüter, and M. L. Schoon, editors. Principles for building resilience: sustaining ecosystem services in social-ecological systems. Cambridge University Press, Cambridge, UK. https://doi.org/10.1017/CBO9781316014240.010

Smith, D., and D. Elliott. 2007. Exploring the barriers to learning from crisis: organizational learning and crisis. Management Learning 38(5):519-538. https://doi.org/10.1177/1350507607083205

Walker, B., S. Carpenter, C. Folke, L. Gunderson, G. Peterson, M. Scheffer, M. Schoon, and F. Westley. 2020. Navigating the chaos of an unfolding global cycle. Ecology and Society 25(04):23. https://doi.org/10.5751/ES-12072-250423

Walker, B., and D. Salt. 2006. Resilience thinking: sustaining ecosystems and people in a changing world. Island, Washington, D.C., USA.

World Health Organization (WHO). 2022. WHO coronavirus (COVID-19) dashboard. World Health Organization, Geneva, Switzerland. https://covid19.who.int/

Yan, B., X. Zhang, L. Wu, H. Zhu, and B. Chen. 2020. Why do countries respond differently to COVID-19? A comparative study of Sweden, China, France, and Japan. American Review of Public Administration 50:762-769. https://doi.org/10.1177/0275074020942445