Development of a common conceptual framework for the adaptation of coastal flood protection systems

ABSTRACT

In the context of climate change, sea level rise and land planning, systemic approaches based on resilience principles are required for the adaptation of coastal areas. In this paper, a new conceptual framework is proposed in which an area and its flood protection system are considered to be two nested social-ecological systems (SESs). Other types of networks of structures, especially transportation infrastructure (ports and waterways, roads and railways) can also be represented, in relation to the areas where they are located, as nested SESs. This conceptual framework, based mainly on the concept of social-ecological resilience, is recommended to develop a common adaptation strategy shared between infrastructure managers and land use planners. It also allows the complementary use of the concepts of technical, ecological and social resilience by different actors to meet specific objectives, which may concern the management of natural or anthropogenic structures or the revision of institutions.

KEYWORDS:

• Resilience
• flood protection systems
• conceptual framework
• climate change adaptation
• system analysis

1. Introduction

Coastal areas are attractive to people, as they are suitable for residential, commercial, and recreational activities, but they are also highly exposed to natural hazards. Many coasts are impacted by the spread of urbanization (Güneralp, Güneralp, & Liu, 2015; Muis, Verlaan, Winsemius, et al., 2016; Neumann, Vafeidis, Zimmermann, et al., 2015) with increasing international trade and globalization. They are also subject to the effects of climate change, including sea level rise (Cazenave & Le Cozannet, 2013; IPCC, 2019; Rovere, Stocchi, & Vacchi, 2016), causing an intensification of coastal hazards (Evans, Ashley, Hall, Penning-Rowsell, Saul, et al., 2004a; Vitousek, Barnard, Fletcher, et al., 2017; Vousdoukas, Mentaschi, Voukouvalas, et al., 2018).

These changes may lead to situations of local concern in terms of livability and exposure to disasters, in particular caused by marine flooding (Hallegatte, Green, Nicholls, et al., 2013; Pycroft, Abrell, & Ciscar, 2016; Tiggeloven, de Moel, Winsemius, et al., 2020). Marine submersion is caused by storms or hurricanes (Bertin, Bruneau, Breilh, et al., 2012; Dietrich, Bunya, Westerink, et al., 2010; Harris, 1963) or by tsunamis (Intergovernmental Oceanographic Commission, 2012; Margaritondo, 2005; Varsoliwala & Singh, 2021). The impacts of these events are exacerbated by human-induced drivers, such as land use changes, land subsidence caused by groundwater and/or hydrocarbon extraction (Erkens, Bucx, Dam, et al., 2015), disturbance of sedimentary movement in rivers and along coasts (Syvitski, 2008), coastal erosion (Luijendijk, Hagenaars, Ranasinghe, et al., 2018; Masselink, Castelle, Scott, et al., 2016; Pontee, 2013; Ruggiero, Buijsman, Kaminsky, et al., 2010) and coastal habitat degradation (Gardner, Barchiesi, Beltrame, et al., 2015).

While in a stable environment, analytical approaches to flood risk management were preferred (in particular, through the ‘Source-Pathway-Receptor-Consequence (SPRC)’ conceptual model (Sayers, Hall, & Meadowcroft, 2002)), systemic approaches based on the principles of resilience, are now often applied to consider multiple causes and effects and to respond with collective actions (Cassel & Hinsberger, 2017; Evans, Ashley, Hall, Penning-Rowsell, Sayers, et al., 2004b; Igigabel, Nédélec, Bérenger, et al., 2022; Krob, 2008). Thus, there are many examples where cities, counties or states put in place comprehensive adaptation measures to consider the local infrastructure, social and ecological context (e.g., 100 Resilient Cities Network (The Rockefeller Foundation, 2015), Build Back Better (Clinton, 2006)). Ecosystem-Based Adaptation practices are often applied (Culwick & Bobbins, 2016; United States Environmental Protection Agency [US EPA], 2020).

The existence of a levee system introduces the additional challenge of managing a set of structures. Their cumulative length often reaches tens of kilometers, and they interact with natural and built environments. This complex system can only be managed properly by specialists from dedicated institutions (CIRIA et al., 2013). These specialists and the organizations to which they belong will be referred to as the ‘protection structure manager’. Projections of climate change raise the question of how to adapt levee systems deployed over a millennium (Guében-Venière Servane, 2015; Welch, Nicholls, Lázár, et al., 2017), especially since these systems contribute to the degradation of the protection provided by natural features (Syvitski & Yoshiki, 2007; Syvitski, Kettner, Overeem, et al., 2009).

It is thus necessary to consider how levee systems evolve in relation to the areas in which they are positioned, but first, it is necessary to define what is meant by a levee system. Without distinguishing between types of infrastructure, Markolf, Chester, Eisenberg, et al. (2018) correctly point out that infrastructure systems are not simply technological systems and should be understood as complex and interconnected social, ecological, and technological systems (SETS). Hamstead, Iwaniec, McPhearson, et al. (2021) provide many examples of how the concept of SETS can be applied in urban areas. Furthermore, Thomas et al. (2019) showed that human and technical resilience capacities are interconnected, interrelated, and interdependent, necessitating a systemic management approach. However, the application of these principles to complex and nested systems is challenging, which is linked to the knowledge of the system and the differing points of view of each actor. In the case of coastal flood protection systems, structure managers and land use managers (generic term for people involved in more general land use spatial planning issues) often have diverging views because their own objectives lead them to assign different meanings to the concept of resilience applied to these systems (ImdR, 2018).

This article aims to understand better these differences and to present a new conceptual framework to foster dialogue between actors and to develop a common and comprehensive point of view. This type of objective has been pursued previously by seeking a single definition of resilience suitable for all systems and actors. For example, Meerow, Newell, and Stults (2016) propose a single definition that combines the multiple objectives of stability, adaptability and transformability for urban resilience. However, the objective in the current work is to obtain a more general and adaptable conceptual framework using the concept of resilience, which can be applied to a wide range of contexts. This will require specifying the different meanings of the word ‘resilience’ and identifying the types of systems to which each of these definitions may apply.

After the presentation of the conceptual framework, its validity is assessed by studying its applicability for the analysis and adaptation of geographical areas and their infrastructure networks. This last step, beyond validating the conceptual framework, aims to improve the operational character of the concept of resilience when multiple, diverse and highly interdependent systems coexist (e.g., infrastructure, ecosystems, human communities).

Following a review of operational methods and tools for managing and adapting flood protection systems (Igigabel, Nédélec, Bérenger, et al., 2022), the need for a common conceptual framework was highlighted. To achieve this goal, this paper is the result of a thorough study of the existing literature, bringing together concepts from different scientific domains, of operational practices, and of unique field experiences. This work consists of: (i) the identification and clarification of resilience concepts applicable in the interdisciplinary field of flood protection; (ii) the construction of a new conceptual framework showing how the various concepts of resilience can be applied to the studied systems; (iii) the identification of concepts used consciously or unconsciously, implicitly or explicitly, in the existing literature reporting concrete case studies of risk analysis or adaptation strategies; and finally (iv) the demonstration of how the proposed conceptual framework can be used in specific case studies.

2. Resilience: one word, several meanings

The concept of resilience has entered the lexicons of a wide range of disciplines, including not only natural sciences, but increasingly social sciences as well (Cote & Nightingale, 2012; Brown, 2014). Davidson, Jacobson, Lyth, et al. (2016) argue that attempts to apply the concept of resilience to many different fields have impeded the continued improvement of its ‘operationalization’ within policy making and its implementation because of:

• a lack of consensus on the meaning of resilience, as a result of multiple definitions;

• differing interpretations of the same definition by policy makers, preventing the development of a common set of goals being generated; and

• difficulties in measuring progress in improving resilience.

These challenges appear clearly in water governance, where systems may be exposed to climate change impacts, hydrological variability and uncertainties associated with different dimensions of global environmental change (Rodina, 2018). Several authors therefore used resilience typologies to clarify the key concepts. For example, based on Carpenter, Walker, Anderies, et al. (2001), Wied, Oehmen, & Welo (2019) provided a review of resilience, asking of what, to what, and how? However, their conclusions only concern engineering systems. This is why it seems preferable to searching for a typology established for different types of systems (e.g., infrastructure, ecosystems, human communities). Davidson, Jacobson, Lyth, et al. (2016), through broader investigations in the fields of applied research, distinguish the following categories: ecological resilience, engineering resilience, social-ecological resilience, urban resilience, disaster resilience, and community resilience. Here, this list is simplified by:

• excluding disaster resilience, since it is common to all systems, and we want to take into account the diversity of these systems;

• preferring social-ecological resilience over urban resilience, as advocated by Davidson, Jacobson, Lyth, et al. (2016), arguing that the definition and theoretical foundation are very similar (Pickett, Cadenasso, & Grove, 2004 ; Pike, Dawley, & Tomaney, 2010 ; and Davoudi, Brooks, & Mehmood, 2013).

Consistent with this preliminary analysis and following Rodina (2018) and Disse, Johnson, Leandro, et al. (2020), the definitions of resilience will be distinguished depending on whether they relate to engineering systems, ecosystems or human communities (systems defined by sectoral approaches) or systems encompassing all of these components (social-ecological systems).

2.1. Engineering, ecological and social resilience

Engineering resilience is the ability of a system to return to an equilibrium state after a temporary disturbance (Holling, 1996). The more rapidly it returns, and with the fewest fluctuations, the more stable it is. The emphasis is on the return time : for a « fail-safe » engineering design, efficiency, constancy, and predictability are all sought-after qualities. When considering anthropogenic structures in levee systems, the concept of technical resilience applies. The objective of maintenance activities is to keep structures in a predefined state to maintain specified levels of protection. When a disturbance occurs, recovery work aims to return the structures to their previous state as quickly as possible (Disse, Johnson, Leandro, et al., 2020). Adaptability and, a fortiori, transformability are therefore not commonly sought characteristics. However, there are now recommendations for adaptive management of flood protection assets, including an integrated, pragmatic, flexible and innovation-friendly governance (Sayers, Gersonius, den Heijer, et al., 2021).

Ecological resilience determines the persistence of relationships within a system and is a measure of the ability of a system to absorb changes and persist (C. S. Holling, 1996). Ecological resilience is not based on a single equilibrium paradigm, but instead considers that multiple equilibrium states may be possible (Gunderson, 2000). Hence, resilience is not necessarily focused only on maintaining the current state (or returning to a previous state), but more on developing new trajectories and evolving to a new system (Folke, 2006). When considering natural formations, the concept of ecological resilience is relevant since the protection objectives can be achieved using biological or physical dynamics. For example, in the case of a dune, plant growth results in the deposition of sand by slowing the wind. Therefore, revegetation will not only develop the resistance of the dune to erosion, but will also encourage the accumulation of sand.

For systems including human communities, social resilience is used to describe the ability of groups or communities to cope with external stresses and disturbances as a result of social, political, and environmental changes (Adger, 2000). Moberg and Galaz (2005) argue that social resilience can be distinguished from ecological resilience because of the ability of humans to anticipate and therefore influence the future. The concept of social resilience can be applied to each of the organizations in charge of a territory or an infrastructure system.

2.2. Social-ecological resilience

To address the issues arising in the context of global changes, such as climate change and other major anthropogenic disturbances, the concept of resilience has been extended to cover social and ecological aspects more broadly. Social-ecological resilience is defined as the capacity of interconnected social, economic, and ecological systems to cope with a hazardous event, trend, or disturbance, responding or reorganising in ways that maintain their essential function, identity, and structure. Resilience is a positive attribute when it maintains the capacity for adaptation, learning, and/or transformation (Walker, Holling, Carpenter, et al., 2004). Resilience can also be a negative attribute when decisions lead to the perpetuation of undesirable situations, for example: degradation of biodiversity, maintaining certain parts of the population in poverty or in areas exposed to hazards.

This definition of resilience refers to a social-ecological system (SES): an integrated system that includes human societies and ecosystems, whose structure is characterised by reciprocal feedbacks, emphasising that humans are part of, not apart from, nature (Arctic Council, 2016; Berkes & Folke, 1998).

It should be emphasized here that even though social-ecological systems may encompass technical systems, ecological systems and social systems, the concept of social-ecological resilience does not necessarily include other resilience concepts that take into account the particular properties that characterize technical systems, ecosystems and human communities.

Protection systems can be considered as SESs for two main reasons. First they are integrated sets of technical and natural elements that do not provide protection separately but in combination. Second, continuous action of managers of natural or anthropogenic structures (as well as other actors involved in regulation or financing) is necessary for the proper functioning of these systems.

The same reasoning can be applied to transportation infrastructure networks, such as ports, waterways, roads and railroads (and even extended to networks consisting of pipelines or cables, whose ecological component is less clear). Similarly, a territory with a human population, natural features and infrastructure can be considered as a SES, and the concept of social-ecological resilience can therefore be applied to this system. This view will be adopted more spontaneously by those responsible for spatial planning than by those directly involved in the management of anthropogenic and natural structures, since it is difficult for humans to imagine themselves as belonging to the same system as the objects with which they interact. Lastly, the management of anthropogenic and natural structures is often carried out by different organizations, which makes it difficult for each individual actor to consider the system as a whole.

3. Common conceptual framework for protection systems and the associated areas

It is clear from the preceding section that the adaptation of protection systems through the development of their resilience may be considered with significantly different approaches depending on the actors: the levee system and natural feature specialists apply the concept of resilience to the protection structures with the main objective of maintaining their performance, while planning specialists apply the concept of resilience to a geographical area with the objective of its sustainable development (considering less the performance of the protection system and the risks associated with its use). The utilization of two conceptual approaches – social-ecological resilience and technical or ecological resilience – applied to two distinct systems – the geographical area and its protection system –, results in partitioning the risk analysis and the formulation of adaptation strategies employed in both fields (Igigabel, Diab, & Yates, 2021).

Here, a new conceptual framework is proposed promoting the convergence of views on the adaptation strategies of geographical areas and their protection systems. First, a geographical area and its protection system are considered to be nested SESs. This approach favors the concept of social-ecological resilience, and allows complementary and coordinated application of other resilience regimes to achieve specific objectives. In addition, the highest-level SESs at a regional or global scale may occasionally also be considered, in particular for issues related to ecosystem preservation, resource management and climate change mitigation.

The use of resilience concepts at the protection system level and then at the geographical area level are presented below.

3.1. Resilience concepts at the protection system level

A protection system can be represented as an SES comprising a technical component (anthropogenic structures), an environmental component (natural structures) and a social component (mainly managers of anthropogenic or natural structures).

The concept of social-ecological resilience has been developed to address the interactions between these three components (Figure 1). The application of this concept should be favored in order to understand the functioning of the entire system and its evolution, and to define long-term strategic plans.

Figure 1. Components of a protection system and identification of the applicable resilience concepts (color bar).

In a complementary way, managers of anthropogenic and natural structures may use the concepts of technical and ecological resilience in their daily activities to monitor and maintain the components of the protection system. These concepts allow considering the modes of operation and evolution particular to engineering structures and ecosystems.

Moreover, the structure management activities should be studied by considering the institutional organizations involved, with regard to the social resilience regime. The vulnerability of an SES depends not only on the state of natural and anthropogenic structures but also on the ability of institutions and humans to manage them. The management of protection systems is already very complex, without considering added stressors generated by the effects of climate change or land use planning. A multitude of skills are required for system management (Figure 2), including asset inspection, performance and risk assessment, and asset management planning. If, under normal circumstances, these activities can be carried out for most structures on the basis of visual inspections and simplified analyses, further investigations may be necessary when analyzing meteocean events generating significant hydraulic conditions.

Figure 2. Asset performance tool propeller (Environment Agency, 2011).

Unless adaptation measures are taken, a large part of the tropics will be exposed annually to the present-day 100-year extreme sea level by 2050, and by the end of this century this may apply to most coastlines around the world, implying unprecedented flood risk levels (Vousdoukas, Mentaschi, Voukouvalas, et al., 2018). Thus, the capacity of the institutions to manage the protection systems by mobilising the multiple necessary skills mentioned above could become insufficient soon.

3.2. Resilience concepts at the geographical area level

A geographical area often has in general multiple functions. It is therefore necessary to represent the protection system at this level in relation to other infrastructures organized in networks, in particular transportation infrastructure (ports and waterways, roads and railways, etc.) and water, energy and telecommunications networks. These infrastructure networks are characterized by reciprocal feedbacks between people and nature; therefore, they can also be considered as SESs. Figure 3 shows the protection SES in relation to another infrastructure network within a given geographical area. This figure synthesizes the proposed new conceptual framework. For the sake of clarity, only two infrastructure networks are represented on this two-dimensional diagram. A three-dimensional representation could represent an unlimited number of infrastructure networks, all sharing the same ecological component (natural structures), and all related to the social component at the level of the geographical area considered (territorial institutions and other stakeholders).

Figure 3. Proposed new conceptual framework, representing a protection system in relation to other infrastructure networks within a territory and identifying resilience concepts used for different systems (color bar).

The nature of the systems should lead to a preference for the concept of social-ecological resilience in order to understand their functioning, predict their evolution, and define strategic orientations. Similar to the protection system level, other resilience concepts may also be used for issues specific to ecosystems, engineering structures and social organizations. In each of these three areas, the conceptual framework allows for the integration of specific issues, as indicated below.

Ecosystems interact with infrastructure supporting multiple activities. Managers should ensure that cumulative effects do not lead the ecosystem to an undesirable regime. Changes in slow variables and feedbacks can indeed lead to nonlinear changes or regime shifts in the SES if certain thresholds are exceeded, with substantial impacts on the set of ecosystem services produced by the SES (Scheffer, Carpenter, Foley, et al., 2001). For example, a regime shift in a lake may result in eutrophication and loss of biodiversity (S. R. Carpenter, 2003). This kind of problem can be addressed by referring to the concept of ecological resilience.

Regarding the engineering components, there are interdependent relationships between networks. For example, a power grid can power a drainage system. Interruption in operation of one drainage system may lead to flooding and interruption of another. As Lhomme et al. (2013) stated, networks then act as vectors of risk propagation. This kind of problem can be addressed by referring to the concept of engineering resilience.

Lastly, the resilience of protection systems is strongly influenced by social and economic factors: the maintenance of the system is conditioned by access to material and financial resources, and the existence of an appropriate mode of governance (Schleussner, Donges, Donner, et al., 2016). In this regard, it is necessary to revisit the process highlighted by Bhowmik (2017), which concerns not only poor regions but also rich regions: with multiple and amplified shocks, the organization and governance of the system becomes more and more complex. It may then reach a point where adaptation measures lose effectiveness, which can lead to a collapse of the system. This kind of problem can be addressed by referring to the concept of social resilience.

4. Application of the conceptual framework

In order to reduce the risk of disaster, progressive or even more radical changes will have to be considered, not only in response to extreme events, but also as part of a planned approach (P. Sayers, Moss, Carr, et al., 2022; Zevenbergen, Veerbeek, Gersonius, et al., 2008). Meeting this challenge requires shared methodological frameworks to coordinate decision-making at different levels of governance. Thus, the adaptation of coastal flood protection systems must be considered at multiple spatial scales: the protection structures, the geographical area directly concerned, and also larger spatial scales (e.g., a catchment). At each of these scales, constraints exist for the adaptation of protection systems: at the geographical area and structure levels, the expansion and densification of the urban fabric reduce the possibility of moving or even widening the structures or granting greater mobility space for natural formations. On a broader scale, the objectives of water management, land use planning and environmental protection must be considered in conjunction with security objectives.

The adaptation of geographical areas and their protection systems is part of an ongoing process of adjusting to changes, with no end. This means that defining successful adaptation is more about the sustainability of processes and the principles of fairness and equity than it is about measuring outcomes at any given point in time (Barnett, Evans, Gross, et al., 2015; Hurlimann, Barnett, Fincher, et al., 2014; Stafford-Smith, Horrocks, Harvey, et al., 2011). Successful adaptation is therefore a matter of ‘socially and environmentally sustainable development pathways, including both social justice and environmental integrity’ (Clark & Harley, 2020; Eriksen, Aldunce, Bahinpati, et al., 2011; Sayers, 2017). The conceptual framework proposed here suggests taking the innovative approach of representing multiple nested SESs. It is intended to facilitate the understanding of the evolution of geographical areas and their protection systems (diagnosis and prognosis), as well as to define adaptation strategies. The choice to favor the social-ecological resilience regime to define a common framework allows benefiting from the adaptive cycles model (Holling, Gunderson, & Ludwig, 2002) for SES analyses, and from recommendations for strengthening their resilience, synthesized by Biggs, Schlüter, Biggs, et al. (2012) for SES adaptation. In addition, the relevance of other resilience regimes to address specific issues will be sought.

4.1. Adaptive cycles model application to geographical areas and their protection systems

The Adaptive Cycles Model (Figure 4) was derived from a comparative study of the dynamics of ecosystems. It focuses on the processes of destruction and reorganization, resulting in the release of energy and material, which are often neglected in favor of growth and conservation (e.g., the slow accumulation and storage of energy and material in a forest).¹ This cycle consists of four phases (Walker, Holling, Carpenter, et al., 2004). Two of the phases – the growth and exploitation phase (r), which merge into a conservation phase (K) – comprise a slow, cumulative forward loop of the cycle, during which the dynamics of the system are reasonably predictable. As the K phase continues, resources become increasingly locked up, and the system becomes progressively less flexible and responsive to external shocks. Eventually, this is inevitably followed by a chaotic collapse and release phase (Ω) that rapidly gives way to a phase of reorganization (α), which may be rapid or slow. During this phase, innovation and new opportunities are possible. The Ω and α phases together comprise an unpredictable backloop. The α phase leads into a subsequent r phase, which may resemble the previous r phase or may be significantly different (Figure 4).

Figure 4. The adaptive cycle (from Panarchy, edited by Lance H. Gunderson and C.S. Holling: Figures 2–1 (page 34). Copyright © 2002 island press. Reproduced by permission of island press, Washington, DC).

Adaptive cycles occur at different spatial and temporal scales, and SESs exist as a ‘panarchy’, or a set of nested adaptive cycles, as shown in Figure 5 (Gunderson & Holling, 2002).

Figure 5. Panarchy of nested adaptive cycles (from panarchy, edited by Lance H. Gunderson and C.S. Holling: Figure 3-10 (page 75). Copyright © 2002 island press. Reproduced by permission of island press, Washington, DC).

According to this theory, systems with larger spatial scales are considered to have greater inertia and to evolve over longer time scales. Thus, the interconnectedness of hierarchical scales in a panarchy contributes to system resilience because disturbances at one scale can be absorbed by other scales in the system (Nash, Allen, Angeler, et al., 2014). The adaptive cycle model offers the possibility to represent the joint evolution of geographical areas and their protection systems (considered as nested SESs) over the long term. In the following two sections, this model is applied first at the level of the protection system, then at the level of the geographical area, and conclusions are made for the protection system as a subset of the geographical area.

4.1.1. Adaptive cycles model application to protection systems

In order to assess how the adaptive cycle model can be applied to protection systems and what lessons can be drawn from it, it is necessary to give a preliminary overview of the life cycles of its components (anthropogenic and natural).

A structure, from its construction until its eventual decommissioning, is subject to multiple actions of monitoring, maintenance, and in the case of deterioration, repair. If the protection objectives evolve, structures can also be adapted. Figure 6 illustrates this type of approach for levees, which are generally the most common type of structure in flood protection systems.

Figure 6. Life cycle of a protection structure (CIRIA et al., 2013).

Several cycles appear in this diagram (Figure 6): a routine cycle (main cycle, thick line) corresponding to the long-term management, an internal cycle corresponding to an event requiring rapid intervention (thin line), and an external cycle corresponding to a change in the design (dashed line). The last two cycles lead to a return to the routine cycle (unless the structure is decommissioned).

When natural formations are integrated into a protection system, their fate is quite similar to that of anthropogenic structures. There is a routine cycle corresponding to seasonal variations (e.g., accretion-erosion of beach-dune systems or variations in mangrove vegetation), and human interventions take place either in emergencies (internal cycle) to maintain or restore the protective function, or in anticipation (external cycle) to reinforce a natural formation whose state is considered too precarious.

Thus, it appears that the life cycles of anthropogenic structures and natural formations have the main characteristics of adaptive cycles in which:

• the routine cycle corresponds to the operational and conservational phases (r and K in Figure 4);

• the inner cycle corresponds to the chaotic phase of collapse and release (Ω in Figure 4). The materials and the natural or artificial order that made it possible to perform the function of protection, are partially or totally missing. It is a short phase corresponding to a marine-weather event or a tsunami, and it includes any repairs carried out under emergency conditions;

• the external cycle corresponds to the reorganization phase (α in Figure 4), which can be fast or slow, and can create new opportunities and foster innovation.

However, the analogy with the adaptive cycle of Holling, Gunderson, and Ludwig (2002) is valid only if the anthropogenic structure or natural feature is sufficiently damaged so that the system collapses and a new cycle can be considered initiated. Otherwise, the successive cycles are more part of the operational and conservational phases of the adaptive cycle.

As marine submersions occur on large scales and simultaneously affect a multitude of structures and natural formations, this analysis conducted on a component of the protection system is also valid on the global scale of a protection system.

The description of the conservation phase by Holling, Gunderson, and Ludwig (2002) is representative of the situation of levee systems facing both an increase in biophysical and urbanization constraints:

• the system becomes progressively less flexible and responsive to external shocks: the saturation of land use by urbanization prevents the repositioning of levees, or even just widening and raising them; and

• resources become increasingly locked up: the difficulty of changing the system leads to an increase in maintenance expenditures to the detriment of capital expenditures.

Holling, Gunderson, and Ludwig (2002) predict that this conservation phase will inevitably be followed by a chaotic collapse and a reorganization phase. At the level of the protection system, these predictions are all the more justified since increased urbanization and intensification of phenomena generating hazards often occur at regional or larger spatial scales. This is detrimental to the ability to allocate additional resources as a priority to protection systems whose condition would be particularly degraded. It should be noted that the precariousness of protection systems is not linked to changes affecting the systems themselves, but rather to external changes at higher levels of the panarchy (in particular urbanization, subsidence, and sea level rise that can be considered as ‘slow variants’ for protection systems). Therefore, adaptation paths should be sought primarily at higher levels of the panarchy.

4.1.2. Adaptive cycles model application to geographical areas

As for protection systems, the applicability of the adaptive cycles model to geographical areas can be assessed with the modes of representation used for this type of system.

The implementation of integrated flood risk management policies can be represented as a cycle (Figure 7). Provided that a cycle profoundly alters the protection strategy, it may be considered as an adaptive cycle of an area considered as an SES. Otherwise, if the actions taken during this cycle correspond only to minor adaptations, they are more closely related to the operational and conservational phases of the same adaptive cycle.

Figure 7. Cycle of flood management. Credit: Conitz, Zingraff-Hamed, Lupp, et al. (2021) inspired by Thieken (2004). Reproduced with permission from Annegret Thieken.

Figure 7 shows that protection strategies are generally organized around phases of risk analysis and preparedness (i.e. disaster risk reduction), crisis management, and recovery relative to major events. These phases can be linked to the adaptive cycle of Holling, Gunderson, and Ludwig (2002):

• the growth and conservation phases are represented by risk analysis and disaster risk reduction measures;

• the chaotic phase of collapse and liberation is represented by the management of a major crisis; and

• the reorganization phase is represented by a recovery phase which may, depending on the description of the adaptive cycles, result in a system fundamentally different from the initial system.

Even more so than in the case of protection systems, the « conservation » phase leads to questioning the complexification of governance systems (which responds to the complexification of the areas to which they apply).

4.1.3. Protection systems in the panarchy of social-ecological systems

The panarchy chosen for the study of protection systems reveals a paradoxical situation: while the higher-level systems of greater inertia (the planet, the region and the geographical area) should have a regulatory role according to the theory of adaptive cycles, they are evolving in general at such a speed that not only do they no longer play this role, but on the contrary, they cause a destabilization of the systems at lower levels.

At the global level, destabilization is primarily caused by the multiple effects of climate change. In parallel, extensive and rapid global changes, including urbanization, growing human populations, rising consumption, and increased global connections, have led to a large and growing demand for provisioning services. Meeting these needs has resulted in large-scale conversion of natural ecosystems to cropland, which has reduced the capacity of ecosystems to produce other ecosystemic services essential to human health and security – especially regulating services (Raudsepp-Hearne, Peterson, & Bennett, 2010).

In addition, considering the regional and the local area as systems of intermediate ranks in the panarchy, actions such as deforestation, the disturbance of sedimentary flows, or the acceleration of urbanization are likely to generate instabilities as well for protection systems. For example, the degradation of coral reefs or urbanization in wetlands and mangroves is increasing significantly the risks of marine submersion through the degradation of natural protections and increased exposure to hazards.

4.2. Application to geographical areas and their protection systems of the recommendations for SES adaptation

In this section, if the territory and its protection system are considered as separate (though nested) SESs, they will no longer be treated separately since the definition of adaptation strategies generally requires joint consideration. The definition of the adaptation strategies of the protection systems should be integrated into a territorial project in order to avoid forms of maladaptation that invariably result in environmental degradation and higher risks. In particular, strategies should avoid two biases. The first bias is the increasing dependence on engineering structures, such as levees (Welch, Nicholls, Lázár, et al., 2017; Wong, Losada, Gattuso, et al., 2014), pumping and drainage systems (Aerts, 2018; Bloetscher, Heimlich, & Meeroff, 2011), and storm barriers in estuaries (Burdick and Roman, 2012). The second bias is unsustainable resource management. In particular, fresh water, sand, armourstone, and wood are multiple-use resources. Clarification is needed on the state of these resources (Brakenridge, Syvitski, Niebuhr, et al., 2017; Day, Agboola, Chen, et al., 2016; Renaud, Syvitski, Sebesvari, et al., 2013) and their uses (Peduzzi, 2014; Torres, Brandt, Lear, et al., 2017).

Beyond the identification of these biases, general recommendations to strengthen the resilience of SESs can help positively define guidelines for the adaptation of protection systems. The synthesis (Figure 8) established by Biggs, Schlüter, Biggs, et al. (2012), shows that adaptation principles relate to the generic properties of SESs (Diversity and redundancy, Connectivity, Slow variables and feedbacks) and to the key attributes of SES governance (Learning and experimentation, Participation, Understanding SES as CAS, Polycentricity).

Figure 8. The seven principles for managing the resilience of an SES (Biggs, Schlüter, Biggs, et al., 2012).

For the application of these principles, an SES should be considered as a Complex Adaptive Systems (CAS), i.e. a system of interconnected components characterized by emergent behavior, self-organization, adaptation, and substantial uncertainties about system behavior (Biggs, Schlüter, Biggs, et al., 2012). Building on this critical idea, Biggs, Schlüter, Biggs, et al. (2012) define resilience as ‘the capacity of an SES to sustain a desired set of ecosystem services (ES) in the face of disturbance and ongoing evolution and change’. According to Millennium Ecosystem Assessment (2005), SESs produce a « bundle » of ecosystem services (ES), categorized as provisioning (e.g., freshwater, crops, and meat), regulating (e.g., flood and climate regulation), and cultural services (e.g., recreation and spiritual values).

In the following, the adaptation measures are defined as a result of these principles, and they are positioned within the proposed conceptual framework, specifying the system concerned and the associated resilience regime.

4.2.1. Application of the principles relating to generic properties of SESs

Based on the framework established by Biggs, Schlüter, Biggs, et al. (2012), the three principles corresponding to the generic properties of SESs are to maintain diversity and redundancy, manage connectivity and manage slow variables and feedbacks.

Maintaining diversity and redundancy are two essential principles of protection.

Diversity appears in the very definition of protection systems, which can be composed not only of various man-made structures, but also of natural structures. These physical components of a protection system can also be supplemented by regulatory measures concerning the development. Synergies can be sought between the protection system and regulatory measures. For example, the prohibition of construction in a low-land zone can be both a protection measure (the retention basin thus created can be considered as an integral part of the protection system) and a vulnerability reduction measure (by avoiding exposing buildings to the most unfavorable conditions). The first interpretation of this protection measure is at the level of the protection system, while the second is at the level of the geographical area. In addition, the restoration of degraded ecosystems or the creation of new natural buffer areas are important measures to increase the natural protection against storm surges and wave action generated by tropical storms and hurricanes (Constanza, Mitsch, & Day, 2006; IPCC, 2019. Vulnerability reduction measures contribute to the protection of critical infrastructure and individual buildings (Zhu, Linham, & Nicholls, 2010). In most cases, nature-based solutions should be encouraged. In urban areas, these solutions are called « green infrastructure », which may be defined as « The interconnected set of natural and constructed ecological systems, green spaces, and other landscape features. It includes planted and indigenous trees, wetlands, parks, green open spaces, and original grasslands and woodlands, as well as possible building- and street-level design interventions that incorporate vegetation. Green infrastructure provides services and functions in the same way as conventional infrastructure (Culwick & Bobbins, 2016) ». This definition, while generic, emphasizes both the systemic nature and diversity of potential contributions.

Protection systems are also based on the principle of redundancy. The most obvious examples are systems consisting of dikes installed in multiple rows, which both provide staggered barriers to the progression of water, but also, between their rows, provide retention basins (Dupuits, Schweckendiek, & Kok, 2017; Marijnissen Richard, Kok, Kroeze, et al., 2021). Other studies also show the effectiveness of hybrid solutions combining land-based structures (for example levees) with the natural environment, such as coastal wetlands and mangroves (Vuik, Borsje, Willemsen, et al., 2019; Vuik, van Vuren, Borsje, et al., 2018). These modes of protection combining several natural and anthropogenic structures reveal the difficulties in distinguishing between the protection system and the area in which it is located. They also show that it is possible to combine the principles of diversity and redundancy in the same zone.

Ultimately, the principle of maintaining diversity and redundancy applies well to coastal flood protection. However, there may be opposing logic for adaptation, depending on whether redundancy or diversity is preferred. Redundancy encourages preserving the system (regime of engineering resilience), while diversity encourages innovation, thus further transforming the system (regime of social-ecological resilience).

The second principle stated by Biggs, Schlüter, Biggs, et al. (2012) is managing connectivity, which involves two key issues for coastal flood protection that affect both the natural and urban environments: (1) water flow and sediment displacement, and (2) the preservation of ecosystems.

Water flow and sediment displacement are clearly major issues in coastal flood protection in different environments. For example, in deltas, reduced freshwater and sediment inputs from river basins are critical factors that determine sustainability (Day, Agboola, Chen, et al., 2016; Renaud, Syvitski, Sebesvari, et al., 2013). In these areas, where engineering operations have often constrained flows by channelling them, it is necessary to restore larger natural flows of water and sediment through the system in the long term and at the scale of the watershed. This requires improved river-floodplain connectivity, if necessary by the removal of levees (Brakenridge, Syvitski, Niebuhr, et al., 2017). In general, these adaptation measures should be defined at larger spatial scales than the area that benefits from the protection system. The social-ecological resilience regime should be preferred to address all issues concerning water flow and sediment displacement.

Preserving ecosystems is another key issue since the ability of coastal SESs to cope with and adapt to sea level rise depends on coastal ecosystem health and habitat connectivity (IPCC, Pörtner et al., 2019). The importance of connectivity to the resilience of SESs and the ESs they produce underlies many conservation initiatives, such as the conservation of levees (which are often the last green corridor in urbanized areas) and the design of networks of protected areas (e.g., the Great Barrier Reef Marine Protected Area network (McCook, Ayling, Cappo, et al., 2010)). These measures can be carried out at very different scales, sometimes involving only selected natural elements or anthropogenic structures belonging to the protection system, and sometimes including several components of the territory that may not be part of the protection system. In the first case, it may be more relevant to refer to the ecological and engineering resilience regime. In the second case, it may be more appropriate to refer to the social-ecological resilience regime.

The third principle stated by Biggs, Schlüter, Biggs, et al. (2012), managing slow variables and feedbacks, is also relevant for protection systems.

Slow ecological variables are often linked to regulating ESs, e.g., climate regulation and flood regulation (Millennium Ecosystems Assessment, 2003). Concerning the coastal flood protection ES, the rate of soil sealing, relative sea level rise, sediment transport, and other drivers influencing erosion phenomena (sea states, the wave climate and marine currents) are slow variables whose monitoring is necessary to understand the system evolution and the associated risks. These monitoring measures can be implemented for all types of systems and are not associated a priori with a particular resilience regime (by default, the social-ecological resilience regime can be adopted).

Finally, feedbacks occur when a change in a particular variable, process, or signal either reinforces (positive feedback) or dampens (negative feedback) subsequent changes of the same type (Biggs, Schlüter, Biggs, et al., 2012). Feedbacks between human interventions and delta morpho-sedimentary responses show that management interventions that weaken stabilizing feedbacks underlying coastal flood protection can erode the resilience of SESs (Welch, Nicholls, Lázár, et al., 2017). Conversely, strengthening the stabilizing feedbacks in a system can help maintain a particular SES regime and the associated ESs in the face of external stresses (e.g., climate change (Mumby & Hastings, 2008; Thrush, Hewitt, Dayton, et al., 2009)). The study of feedbacks may be relevant for all types of systems and is not associated with a particular resilience regime (by default the social-ecological resilience regime can be adopted).

4.2.2. Application of the principles relating to key attributes of SES governance

Climate change and associated extreme events demand increasing complexity within the modes of governance, i.e. formation of new social institutions to engineer, mitigate, and adapt to climate problems, which, in turn, increase related expenditures. Bhowmik (2017) warned that this process may lead potentially to a collapse of societies.

The adaptive capacity of human communities depends on assets (financial, physical, and/or ecological), capital (social and institutional), knowledge and technical know-how (Klein, Midgley, Preston, et al., 2014). However, it should be stressed that if these factors determine the adaptation potential, they do not necessarily translate into effective adaptation if awareness of the need to act, willingness to act, and/or cooperation is lacking. For example, the ability to adapt to sea level rise depends both on the coastal elevation and on the social and political ability to develop protection or relocation measures (IPCC, 2019).

To address these issues, the SESs should be considered as complex adaptive systems, emphasizing the substantial uncertainties surrounding these systems. Thus, it is necessary to continually learn and experiment, and adaptively manage uncertainty, disturbance, and surprise rather than attempt to eliminate them (Biggs, Schlüter, Biggs, et al., 2012). This principle, coupled with that of participatory management of resources and links between communities and governments, forms the basis of a recommended mode of governance that includes engaging transformational forms of adaptation: adaptive co-management (Olsson, Folke, & Berkes, 2004; Ruitenbeek & Cartier, 2001).

For « coastal area » SESs, an expanded network of actors is important for different reasons, including for example:

• in the definition of town planning rules, a consultation phase may involve all stakeholders in the decision-making process. Perherin (2017) described in detail how the debate between the actors can be organized around the production of hazard maps carried out as part of the elaboration of coastal risk prevention plans. Greater openness to multiple stakeholders allows sharing hazard knowledge with and involving the community, while also promoting the integration of risk reduction into territorial public action;

• citizen participation in data collection and analysis to increase community awareness and to encourage citizens to be active participants in the development of flood risk management plans. For example, the European FloodCitiSense research program (McCrory & Veeckman, 2017), with study sites in Birmingham, Brussels, and Rotterdam, worked on reducing the vulnerability of urban areas to rainfall flooding by developing an urban pluvial flood warning service that was co-created by and for citizens and city authorities. The project aimed to integrate crowd-sourced hydrological data of local stakeholders through a citizen science approach, making use of low-cost sensors and web-based technologies. It followed the guidelines of the European Flood Risk Directive (Directive 2007/60/EC) that encourages the participation of local stakeholders in the development of flood risk management plans, in which citizens are expected to take active responsibility at the beginning of the process (e.g., flood monitoring and mapping), rather than passively receiving a service.

A variety of approaches exist to involve stakeholders in the definition and implementation of strategies at the territorial level. To succeed, it is necessary to provide tools (e.g., methodology for the diagnosis of geographical areas and their protection systems (Igigabel, Nédélec, Bérenger, et al., 2022), development of databases and geographical information systems (Cerema, 2020)) and to create opportunities for sharing knowledge and projects (e.g., open-information websites and community meetings).

The principle of polycentricity should also be applied through concerted decision-making and coordinated action at different levels of governance. For example, the European Flood Risk Directive (Directive 2007/60/EC) also respects this principle by refocusing institutions on the geographical perimeters of emerging risk (e.g., river basins). The institutional and administrative restructuring of governance, by developing more integrated visions of land use planning through combining risk prevention and management of aquatic environments is another notable advance (Cerema, 2020). Following the principle of polycentricity, a research project created a learning network extending across multiple scales, from the national, to local, and finally street/building scale in the city of Dordrecht (Netherlands) to improve the practical management of urban floods, raising awareness of the residual risk (Zevenbergen, Veerbeek, Gersonius, et al., 2008). Local, regional, and national authorities, a water board, a developer, a housing corporation, and two research institutes participated in the project.

4.2.3. Summary of the application of SES management principles

Ultimately, as summarized in Table 1, coastal flood protection requires addressing multiple issues simultaneously. In each case, to build an effective dialogue between the stakeholders, it is necessary to identify the systems concerned and the management principles to be implemented. Depending on the case, different resilience concepts may be applied.

Table 1. Identification of key coastal flood protection issues and the main associated resilience regimes.

Principles for managing the resilience of an SES	Issues associated with coastal flood protection	Main resilience regime to consider
Diversity	Composition of the protection system and spatial planning: reducing hazard exposure by controlling urbanization, in particular in low-lying areas, and using nature-based solutions to reduce the vulnerability of critical infrastructure and buildings can also foster diversity.	Social-ecological resilience
Redundancy	Composition of the protection system and spatial planning: coastal flood protection systems may include multiple lines of defense. Between these lines, as well as on the sea or land side, natural or landscaped buffer zones can be important and sustainable measures against marine-weather events.	Engineering resilience
Connectivity	Flow management: The control of water circulation and the preservation of exchanges of all kinds (including humans and other living species, materials, energy, and information) are essential, in particular in the urban environment.	Ecological resilience or social-ecological resilience
Slow variables and feedbacks	Understanding the evolution of SESs: Parameters such as the rate of soil sealing, relative sea level rise, and sediment transport should be monitored at the scale of a structure (e.g., levee) or an area (e.g., delta).	By default, social-ecological resilience
Learning and experimentation, Participation, Understanding SES as CAS, Polycentricity	SES management: Governance can be improved by applying the principles of adaptive co-management and polycentricity at the level of the protection system and the associated area	Social resilience (or social-ecological resilience)

Although there are some conceptual elements that appear in most interpretations of resilience, other elements are domain specific. These differences make it difficult to ‘operationalize this concept’ (Davidson, Jacobson, Lyth, et al., 2016). The application of the proposed conceptual framework helps to meet this challenge by specifying the definitions applicable to the various issues.

4.2.4. Case studies illustrating the importance of properly identifying systems and resilience regimes

SES management principles can be applied provided that the relations between the identified systems (protection system, local area, or higher-level systems) are fostered by the adoption of the same concept of resilience. In general, the social resilience regime is relevant to governance issues, but if technical or ecological issues are also involved, then it is necessary to adopt the social-ecological resilience regime. The following three case studies illustrate the importance of properly identifying the type of system and the resilience regime to apply, showing the impacts that these choices have on the proposed management plans.

By studying urban planning processes that have led to adaptation to sea level rise in Lakes Entrance in Victoria (Australia), Hurlimann, Barnett, Fincher, et al. (2014) concluded that to be sustainable, urban planning needs to (i) facilitate local ownership of adaptation responses, (ii) build collective action within local communities and between local communities, as well as between different branches and levels of government, and (iii) be fair in its application in space and time. The authors added that addressing these social dimensions of adaptation takes time, but they are a sine qua non of sustainable adaptation to RSL rise. This case shows that before addressing flood risk management through technical or ecological considerations, information sharing, exchanges and decisions based on solidarity may be indispensable within communities. The concept of social resilience promotes the development of collective capacities (economic, social, cultural, spiritual and political) and collective processes (governance and engagement in planning and decision-making), as demonstrated by Berkes and Ross (2013).

Another study, carried out in Italy on the Lambro river basin (in the metropolitan city of Milan), shows the need to accompany the paradigm shift of a change in governance (Vitale, Meijerink, Domenico, et al., 2020). In this case, engineering resilience emphasizes the use of flood protection structures, whereas social-ecological resilience advocates river restoration and spatial strategies to reduce flood risk. Bottom-up initiatives support social-ecological resilience, but national policies and funding support engineering resilience via the construction of hard structures. In this case, the concept of social-ecological resilience should be used to bring stakeholders’ points of view together.

Finally, the case of the Netherlands illustrates an integrated approach where the governance promotes the strengthening of social-ecological resilience: the Delta programme concerning protection systems is scientifically supported by local studies and by the national research program Knowledge for Climate (KfC), which runs in parallel. This addresses the majority of adaptation challenges related to climate change, for both floods and droughts, urban and countryside environments and their functions, and integrating a wide range of scientific fields, from climatology to decision support and governance (Klijn, Kreibich, de Moel, et al., 2015). In this case, flood risk is managed at the scale of a country, taking care to adapt the mode of governance to the multiple technical, ecological and social issues arising in the different areas, which leads to mainly using the concept of social-ecological resilience.

5. Conclusion

The implementation of the concept of resilience in the field of flood protection is generally part of collaborative (between the main actors) and even participatory (by including the population more broadly) approaches. However, the views, objectives and interests of the stakeholders may be different, or even opposing, which may lead to the adoption of definitions of resilience providing contradictory recommendations. In particular, the managers of protection systems tend to favor the concepts of technical or ecological resilience, since their main objective is often to keep anthropogenic structures or natural elements in a predefined state to maintain specified levels of protection. On the other hand, land use managers are more likely to focus on social and ecological resilience with the goal of adapting or even transforming their territory.

To help reconcile contrasting points of view, different definitions of resilience, including technical, ecological, social, and socio-ecological, are presented in direct relation to the nature of the multiple systems concerned (protection systems, natural and anthropogenic structures, as well as territories, human communities and their institutions). Then, a new conceptual framework is proposed, based on these definitions and on the theory of social-ecological systems organized in a panarchy. A key point is to consider infrastructure networks and their associated territories as nested SESs.

The conceptual framework was assessed by studying its applicability to the analysis (diagnosis and prognosis) and adaptation of territories and their infrastructure networks.

The new conceptual framework is compatible with the adaptive cycle model (C. Holling, Gunderson, & Ludwig, 2002) and contributes to improving the understanding of the joint evolution of protection systems and the associated areas, considered as SES. It is particularly clear that this model is consistent with the main modes of representation of the life cycle of structures and the adaptation cycle of the geographical areas. It is also relevant for studying the relationships between the area and its protection system, as well as the influences (anthropogenic climatic or non-climatic factors) of higher-level systems, especially at regional or global scales.

For the definition of adaptation measures, the proposed conceptual framework calls for a systematic examination of the nature of the systems and the resilience regime adopted. Recommendations for strengthening the resilience of SESs (synthesized by Biggs, Schlüter, Biggs, et al. (2012)) can make a significant contribution to addressing the main issues associated with coastal flood protection.

Finally, the application of this new conceptual framework to the analysis and adaptation activities shows that the concept of social-ecological resilience appears to be in general the most relevant for developing shared adaptation strategies and establishing a long-term dialogue between protection structure managers and land use planners. In addition, the concepts of technical, ecological and social resilience can be used by the different actors to meet more specific objectives. For example, the day-to-day management of flood protection structures could be addressed using the concept of technical resilience, while the cumulative effects of multiple infrastructure on the natural environment can be better addressed through the concept of ecological resilience. The challenges faced by institutions in charge of maintaining systems in the context of rising sea levels could be better addressed through the concept of social resilience.

Laws and regulations

Directive 2007/60/EC of the European Parliament and of the Council on the assessment and management of flood risks.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The authors confirm that the data supporting the findings of this study are available within the article.

Additional information

Notes on contributors

Marc Igigabel

Dr Marc Igigabel is a project director at the Cerema (which stands for Centre for Studies and Expertise on Risks, the Environment, Mobility and Urban Planning), in the Technical Division for Risk, Water and Sea. His professional background is marked by engineering activities and by studies lead on coastal hazards, harbour development, renewable energies and marine pollutions. In the last years, his research focuses on coastal hazard changes and the definition of adaptation strategies.

Marissa Yates

Dr Marissa Yates is a researcher at the Ecole des Ponts ParisTech working in the Saint Venant Hydraulics Laboratory, after having passed 12 years at the Cerema. Her research focuses on improving the understanding of coastal hydrodynamics and morphological evolution using observations, laboratory experiments, and numerical models.

Youssef Diab

Pr Youssef Diab is scientific director at the EIVP School of Paris Urban Engineering and Professor at the University Gustave Eiffel. He is a recognized expert on cities’ resilience, governance and climate adaptation. His research work is related to the field of sustainable and resilient cities, including urban policy and decision-making tools.

Notes

1. https://www.resalliance.org/panarchy