Resilience issues and challenges into built environments: a review

10 This paper proposes a review of existing strategies and tools aiming at facilitating the 11 operationalization of the concept of resilience into built environments. In a context of climate change, 12 increased risks in urban areas and growing uncertainties, urban managers are forced to innovate in order 13 to design appropriate risk management strategies. Among these strategies, making cities resilient has 14 become an imperative. This injunction to innovation fits perfectly with the urban, economic, political, 15 social and ecological complexity of the contemporary world. As a result, the concept of resilience is 16 integrated into the issues of urban sprawl and the associated risks. However, despite this theoretical and 17 conceptual adequacy, resilience remains complex to integrate into the practices of urban planners and 18 territorial actors. Its multitude of definitions and approaches has contributed to its abstraction and lack 19 of operationalization. This review highlights the multitude of approaches and methodologies to address 20 the bias of the lack of integration of the concept of resilience in risk management. The limit is the 21 multiplication of these strategies which lead to conceptual vagueness and a lack of tangible application 22 at the level of local actors. The challenge would then be to design a toolbox to concentrate the various 23 existing tools, conceptual models and decision support systems in order to facilitate the autonomy and responsibility of local stakeholders in integrating the concept of resilience into risk management they use. Modules, created on the structure of the expert systems, combine information from different organizations in order to identify interdependencies between them. A time simulator also been

and functions". Cardona (2004) defined resilience as the capacity of the damaged 162 ecosystem or community to absorb negative impacts and recover from these. 163 o Adaptive capacity: Pelling (2011) defends the idea that resilience is the ability of 164 an actor to cope with and adapt to hazards stress. . It refers to the "ability of systems, 165 institutions, humans and other organisms to adjust to potential damage, to take 166 advantage of opportunities, or to respond to consequences" (IPCC, 2014). This 167 implies considering the entire pool of assets (social, physical, financial, natural, 168 human, and cultural) and resources (technological, knowledge and governance) 169 which can be mobilized to build resilience to climate change impacts. Socio-170 technical end ecological aspects are equally targeted in a systemic perspective 171 (Whitney et al., 2017), including consideration of trade-offs among them to avoid 172 social-ecological traps which can risk conditions (Carpenter and Brock 2008).

173
o Reaction capacity, linked to self-organization: Pickett et al. (2004) have defined 174 resilience as the "ability of a system to adjust in the face of changing conditions" 175 and Ahern (2011) has defend resilience as a "capacity of systems to reorganize and 176 recover from change and disturbance". 177 o Ability to rebuild using internal and external forces: Walker et al. (2004) developed 178 the idea that resilience is the capacity to "reorganize while undergoing change so 179 as to still retain essentially the same function, structure identity, and feedbacks" 180 o Learning capacity: The Resilience Alliance (Walker and Salt, 2006) defends that 181 resilience is a combination of three capacities, absorb and remain within the same 182 state, the capacity of self-organization and "the degree to which the system can 183 build and increase the capacity for learning and adaptation" (Carpenter et al.,184 2001; Klein et al., 2003) 185 o Ability to bounce back or reach a new state of equilibrium: to some authors, there 186 is one single-state equilibrium which implies to bounce back to equilibrium 187 previous disturbance (Holling, 1996). On the contrary, others consider that we can 188 observe multiple-state equilibrium which suppose that systems have different 189 stable states (Davoudi et al., 2012;Holling, 1996) 190 191 These different capacities can be self-sustaining or, on the contrary, contradict each other (such as 192 the capacities of resistance and adaptation). Faced with these different positions, the notions and 193 concepts associated with that of resilience accentuate the abstraction and incomprehension of the 194 concept. 195 196 1.3. Concepts associated to resilience perception 197 198 No doubt a victim of its multitude of disciplines and definitions, resilience has been continually 199 associated with or compared to related concepts. Resilience is regularly compared or associated with the 200 concepts of vulnerability and sustainable development (Romero-Lankao et al., 2016). The classic way of analyzing resilience and vulnerability is to contrast them: if you are resilient, you 205 are not vulnerable and vice versa (Folke et al., 2002). This clear opposition seems logical: if resilience 206 is the ability to adapt to a shock and vulnerability is defined as the propensity to damage, then the more 207 vulnerable a concept is, the less resilient it is (Pelling, 2003). So the equation is simple, reducing 208 vulnerability is the same as increasing resilience (Klein et al., 2003). 209 Yet this opposition has been widely contested. First of all, social vulnerability reflects the capacity 210 to face, anticipate and adapt to risks .These social capacities are largely integrated into the notion of 211 resilience (Cardona, 2004). Resilience can therefore be seen as an integral part of the concept of 212 vulnerability (Britton and Clark, 2000), being aimed "to not only restore functionality but also correct 213 existing social, political, and economic structures that may have increased exposure and constrained 214 capacity to cope with the crisis" (Patel and Nosal, 2016). Thus, the two concepts cannot be completely 215 opposed. Concerning the positioning aimed at qualifying the concept of vulnerability as "negative", 216 https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License. "positive" vulnerability provides a counter-argument. Vulnerability is considered positive when it leads 217 to a change that brings about a beneficial transformation (Gallopín, 2003) .For example, in a situation 218 of vulnerability in a dictatorial political system, its collapse is positive. Seeing the collapse or paralysis 219 of an urban system following a flood can raise awareness and allow it to evolve towards more 220 appropriate functioning. Indeed, while in risk assessment vulnerability is in general hazard-specific, 221 certain factors -such as poverty, the lack of social networks and social support mechanisms, inadequate 222 governance structures -will aggravate or affect vulnerability levels irrespective of the type of hazard 223 (Prowse, 2003;UNEP, 2003). Such dimensions of resilience, which involve society and ecosystems as 224 a whole, can be used to identify cross-cutting vulnerability aspects to be tackled as high-level policy and 225 governance issues, linked e.g. to limitations in access to and mobilization of the resources of individuals 226 and institutions, as well as to the incapacity to anticipate, adapt, and respond to absorb the socio-227 ecological and economic impact of hazards (Miller et al., 2010;UNISDR, 2011;Cardona et al., 2012;228 EEA, 2016). Under these conditions, vulnerability and resilience are no longer in opposition but are part 229 of a whole. They can then be approached along a continuum. This new stance leads to the notion of 230 resilient vulnerability (Provitolo, 2012). This notion reflects the idea that "vulnerability can be traversed 231 and modified by resilience considered from a global perspective, i.e. that this resilience can, on the one 232 hand, be directly linked to the vulnerability to which it applies and, on the other hand, have a positive 233 or negative effect depending on the scale at which the system is studied" (Provitolo, 2012). 234 In conclusion, resilience and vulnerability are not dichotomously opposed. The two concepts are 235 equally adaptable to technical and/or social systems. Resilience and vulnerabilities overlap in their 236 approach to systems to provide a vision of exhaustive of the elements composing this one. Addressing 237 the two concepts leads to an analysis of the question of long-term risks. It is therefore necessary to learn 238 to live with the change and uncertainty and not seek short-term control of risks. Analyze together the 239 two concepts is like learning from crises (vulnerability approach) and innovate to adapt to risks 240 (resilience Faced with increasing risks, stakeholders have identified two concepts (Saunders and Becker, 2015), 245 that of resilience (taking into account the management of disturbances) and that of sustainable 246 development (analyzing the balanced economic, social and environmental development of the territory). 247 For some, resilience is a necessary condition for sustainability (Folke et al., 2002;Klein et al., 2003). 248 For others, after studying the possible trajectories of ecosystems according to different initial states, 249 resilience is not sufficient, sometimes it is not even necessary. 250 However, Toubin et al. (2015) defend the fact that resilience can play a role in the realization of the 251 sustainable city (Elmqvist et al., 2019), an ideally functioning urban system. The urban resilience 252 enhancement approach is then defined as a means of managing the jolts of the urban system subjected 253 to numerous disturbances (short-time resilience) and maintaining it in the ideal trajectory of 254 sustainability (long-term resilience) linked to a system state indicator (economic growth, carbon 255 footprint, or demographics, for example). Resilience is thus presented as a means of achieving 256 sustainability (Toubin et al., 2015). Nevertheless, resilience may also "run counter to sustainability 257 goals: for instance, efficiency reduces diversity and redundancy, both of which are key features of 258 resilience. This conflict is illustrated by high-density urban areas, which can be more efficient to run in 259 terms of, say, energy distribution, communications and waste collection. However, these areas can also 260 be vulnerable to extreme events such as flooding because they are less diverse (with few green areas, 261 for example) and have few redundancies (in the form of back-up facilities and disaster-management 262 processes)" (Elmqvist, 2017).

264
Resilience as the capacity of a system to adapt to disturbances thus appears better able to satisfy the 265 need to operationalize the sustainable city. Indeed, the normative basis of sustainable development, 266 particularly in the expression of the major global principles, "freezes" the ideal model to be achieved, 267 while its subjective character raises many debates as to the -moral -values to be pursued. Conversely, 268 resilience seeks to free itself from norms in favor of descriptive magnitudes and ensure a better reactivity 269 of the urban system in the face of the unexpected. "Improving resilience increases the chances of 270 sustainable development in a changing environment where the future is unpredictable and surprise is 271 https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License.
likely." (Folke et al., 2002). Developing a sustainable territory and community cannot therefore be 272 envisaged without a long-term resilience strategy. 273 274 275 276 The concept of resilience is a multifaceted concept, involving a plurality of disciplines, definitions, 277 notions and associated concepts. This diversity can be interpreted both as a source of opportunity but 278 also as a difficulty in the operationalization of resilience and its lack of integration into risk management 279 strategies. In the face of new risks linked to climate change, the evolution of urban areas and the 280 concentration of issues (part 1), the concept of urban resilience represents both an innovative and 281 essential concept but also full of operational limits both at the international and local levels (part 2). 282 This is why a variety of methods, concepts and strategies have been developed to address the issue of 283 operationalization and appropriation of the concept by local actors in order to respond to these limits of 284 application (part 3). We will conclude on the notable advances in these approaches to integrate the 285 concept of resilience by presenting the next steps needed to respond to the limits still present in the 286 scientific and operational field. 287 288 2. Urban risks: over-urbanization, cascading effects and multi-risk approach 289 290 The current climate change context has led to an increase in natural disasters of about 2% per year 291 for the past 15 years (Catastrophes Naturelles-Observatoire permanent des catastrophes naturelles et des 292 risques naturels, 2016). At the same time, the increase in the number of people and goods in urban areas 293 is making it more fragile. considerably the cities. Today, nearly three out of five cities, with 500,000 294 inhabitants, are at risk. However, urban areas produce between 70 and 80% of the world economy and 295 are home to 55% of the world's population , with an increasing urban-rural drift expected to raise this 296 value up to 68% by 2050 (UNDESA, 2019; Zevenbergen et al., 2010). Such a concentration of stakes 297 increases the impact of disasters and raises questions on the future of cities. 298

Over-urbanization 299
In 2008, half of the world's population lived in urban areas. This concentration is likely to accelerate. 300 Projections show that urbanization, combined with overall world population growth, could add an 301 additional 2.5 billion people to urban areas by 2050 (United Nations, 2018). The unplanned expansion 302 of urban areas to face to this rapid growth, combined with inappropriate land-use planning, a 303 geographical location at risk (river mouth, swampy areas, major river bed, etc.) and difficult regulation 304 of building standards, contributes to the over-vulnerability of urban territories and populations. Urban 305 areas in coastal regions are particularly exposed to sea level rise. Low-lying coastal areas -less than 10 306 metres above sea level -account for just 2% of the world's land but are home to 13% of the world's 307 urban population. In 2007, Africa had 37 cities with more than 1 million inhabitants, half of which are 308 located -at least in part -in the low-lying coastal zone. 309 However, this tendency to focus on a specific area can be observed on a global scale: cities occupy 310 only 1% of the world's territory (Angel et al., 2018) . Developed countries have therefore never 311 concentrated more value added per km 2 than they do at present. This concentration of population on 312 such a small portion of the territory has increased spatial and social vulnerability through the exposure 313 of the issues. Indeed, it seems logical to consider that the more a population and its issues are 314 concentrated in a small area, the greater the damage will be. Flooding in an uninhabited area will not be 315 considered and apprehended in the same way as in a metropolis (Mitchell, 1999). 316 The example of storm Xynthia in France in 2010 is an example of the effects of over-urbanization 317 on the reality of the disaster. This storm is one of the deadliest disasters in France with 59 deaths. The 318 marine submersion, which reached 1.53 meters in La Rochelle, affected some communes up to 85% of 319 their surface area (Duvat, 2011). The magnitude of the disaster was due in particular to demographic 320 https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License. change and rapid urbanization in the area. Thus, between 1946 and2007, urbanization in the lower areas 321 doubled or even tripled in some communes, leading to significant vulnerability. Indeed, the decline in 322 agricultural activities has had several effects, including the disappearance of risk culture and the over-323 urbanization of land. Real estate developers and investors have seized land to build, on marshes or dunes, 324 subdivisions, which are vulnerable to the risk of flooding (Duvat, 2011). Some of these lands ended up 325 under a metre of water when the storm passed, trapping the inhabitants in buildings unsuited to the 326 hazards. The second factor of vulnerability is the progressive replacement of populations of farmers and 327 sailors by urban dwellers, tourists and pensioners. These populations live from discontinuously on the 328 territory and therefore lose the knowledge of the natural functioning of the territory, leading to a 329 vulnerability of the populations. Xynthia was thus such a dramatic event because of "the modes of 330 occupation of space (which) gradually neglected the hazards of submersion and flooding" (Duvat, 331 2011). It is therefore no longer only natural disasters that impact cities but urbanization that leads to 332 over-vulnerability, leading to a melting pot of opportunities for risk amplification (Mitchell, 1999).

333
Concentration is thus perceived as an aggravating factor in risk management. This concentration is 334 expressed by the density of population present in a given territory. It is established that the denser the 335 area, the more vulnerable it is, the greater the potential for loss. It is therefore established that in urban 336 areas, natural hazards tend to have more serious consequences (Mitchell, 1999). 337 Three risks have a particular impact on urban areas (CRED and UNISDR, 2018). 338  The earthquake is the most dreadful hazard, as it is responsible for the largest number of victims 339 worldwide, averaging 130,000 a year (Sigma, Swiss Re., 2011). However, the number of 340 victims depends very largely on the nature of the buildings and the nature of the preventive 341 measures. At the same magnitude, the disaster in Port-au-Prince claimed 222,000 victims, but 342 only 500 in Santiago de Chile. As for material damage, given the very unequal insurance 343 coverage, the official estimate is reduced to $10 billion in Haiti, but $30 billion in Chile. In 344 addition to its direct destructive effects, the earthquake can trigger either fires by breaking 345 energy networks, such as the one that ravaged Tokyo in 1923, or tsunamis.

346
 Flooding is also a major risk for large agglomerations that are located either in estuaries, on the 347 coast, in alluvial valleys or on slopes that have become unstable. Urban sprawl is often the most 348 vulnerable, due to poorly regulated urbanization, especially in areas where water is stagnant, 349 such as in Buenos Aires, Dhaka, Phnom Penh or New Orleans. It can also involve mudslides or 350 landslides on urbanized slopes such as the favelas of Rio de Janeiro. The urban dimension also 351 determines the extent of soil waterproofing, and therefore the extent of runoff. In addition to 352 these direct damages, there are also those related to the disorganization of services or the 353 degradation of equipment and industrial installations that are specific to any large urban area.

354
Climate change includes a new risk, that of the gradual rise in sea levels. As a consequence of 355 probable climate change, it threatens many of the world's port cities such as London, the Dutch 356 delta with Rotterdam/Amsterdam, but also Tokyo or New York.

357
 Wildfires, which can occur in periods of drought and heat waves, can cause immeasurable 358 damage, as in Australia in 2019, resulting in the destruction of 3500 homes, 5852 outbuildings, 359 34 direct deaths and 417 by excess from smoke inhalation (Borchers Arriagada et al., 2020). In 360 Europe, forest fires in Greece in 2007 and in Portugal 2017 claimed 80 and more than 100 lives, 361 respectively. In 2018, 99 lives were lost in Greece, 2,500 people were evacuated in Portugal and 362 Spain, 50 people evacuated in UK, while Sweden had to face the most serious series of forest 363 fires in its modern history, although with no fatalities. 364 365

366
The over-vulnerability of these urban areas in the face of natural risks also leads to the emergence 367 of "urban" risks. 368 https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License.

Fragile urban spaces confronted with cascading effects 369
Urban space is made up of several infrastructures, some more essential than others. Called Critical 370 Infrastructures (CI), these infrastructures concentrate all the functions (Pescaroli and Kelman, 2017) 371 necessary for the proper functioning of a community. The term critical infrastructure only appeared in 372 the United States in the 1990s following a succession of disasters, including the first attack on the World 373 Trade Center (1993), followed by that of Oklahoma City (1995) and the gas attack in the Tokyo subway 374 (1995). These infrastructures were then defined as vital to the point that "their incapacity or destruction 375 would significantly weaken (US) defense or economic security". Critical infrastructure is defined as 376 telecommunications, power generation systems, oil and gas storage and transportation systems, banking 377 and finance, passenger transportation, water supply and distribution, emergency services (medical, 378 police, fire), and those that ensure the continuity of government (Fekete et al., 2015). They are termed 379 "critical" because their potential destruction could weaken the entire defense and economic organization 380 (Serre and Heinzlef, 2018) of a country or city. Critical infrastructure can be natural; water supply, flood 381 water storage; or physical; energy networks, telecommunication networks, emergency services, 382 transport networks; or virtual systems such as cyber-information systems. However, these CIs interact 383 with each other and thus create interdependencies (Serre, 2018) within the urban space. These 384 interdependencies then play the role of a risk diffusion factor. According to the concept of the cascading 385 effect (Bach et al., 2013;Nones and Pescaroli, 2016;Pescaroli and Nones, 2016;Serre and Heinzlef, 386 2018) i.e. a chain reaction causing changes in a territory some areas come to be impacted by the disaster, 387 even if they were not located in the same area. directly in the flood hazard extension zone. As urban 388 areas are interconnected, infrastructure failure will impact the territories across geographic and 389 functional boundaries (Boin and McConnell, 2007). Because these components are connected at 390 multiple scales, CIs can have an impact on much larger territory than their first impact territory. For 391 example, floods can have an impact on a specific area, such as a road, but as the interconnected, the risk 392 will spread to other territories that should not have been interconnected. naturally be flooded (Lhomme 393 et al., 2013) by compromising power grids, supply of vital resources, etc. .

394
Therefore, some damages are not caused by direct physical damage, but by through business 395 interruption. A distinction is made between direct and indirect impacts. The direct impacts are the 396 tangible impacts and refer to the damage of the elements. physical (furniture, buildings, stocks, 397 equipment, etc.). Indirect impacts occur when they are not caused by the disaster itself. Indirect impacts 398 can be related to interruption or damage to critical infrastructure service. These may occur outside the 399 area directly affected by the disaster and extend into the time after the shock (OECD, 2014). 400

The contribution of urban networks to the spread of risks 401
The role of urban networks is a good example for understanding and measuring what a CI failure 402 can lead to. Urban networks are an essential part of the urban system. In an interconnected world, urban 403 networks connect more and more people and territories and offer a wide variety of resources and 404 opportunities. However, they also create complex situations of interdependence. Public transport, 405 electricity, gas, telephone, heating, waste, etc. make the management of the urban system more complex. 406 While they are essential for creating dynamics, relationships, and opportunities, they also create 407 complex situations of interdependence. In addition to being a key component of the economy, these 408 networks are also extremely vulnerable in the event of a crisis. Because of their interconnectivity, all 409 urban operations depend on them. A single failure can have cascading effects affecting the entire 410 network and, due to a reticular urban system, the whole city. Some examples illustrate these effects: effects (Pescaroli and Kelman, 2017). The hurricane in August resulted in the breaching of 413 protective dykes causing the destruction of 300,000 homes and 1,833 deaths (Knabb et al., 414 2005). The disaster was exacerbated by the domino effects that followed the destruction of the 415 dikes, which made relief operations more complex. Transportation such as highways and 416 https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License. bridges were affected, reducing the ability to deliver vital resources -such as water, food and 417 medical supplies. Medical facilities were, for the most part, damaged or destroyed. All of these 418 effects have made the territory and its inhabitants more fragile, making it more difficult for CIs 419 to be brought back into service, but also for social and spatial functioning to function properly. 111 homes and damaged 20 others (Kunz et al., 2013). Daily life was severely disrupted by the 427 interruption of the metro, the breakdown of the heating network, security systems and 428 telecommunication services. In addition, alternative solutions such as emergency generators 429 could not be operated, as refineries were insufficient in number and unable to provide the 430 necessary fuel. While direct damage was estimated at 32.8 billion in repairs and restoration, 431 indirect losses cost the city and its citizens much more. The unpreparedness of managers and 432 citizens has considerably increased the impacts of the crisis. For example, the late evacuation 433 order and misinformation have resulted in the impossibility of evacuating certain institutions. 434 In addition, the crisis has put the vulnerability of sewer systems, poor anticipation of sewer 435 system failures, and the lack of network, the absence of a plan B for access to generators and 436 relay antennas, and the installation of the resistant flood barriers (Le Haut Comité Français pour 437 la Défense Civile, 2013) . In this case, the over-connected territory and society have created 438 new risks and made crisis and post-crisis management more difficult and complex 439 440 Societies and territories are therefore deeply vulnerable to potential functional disruptions due 441 to a crisis (Boin and McConnell, 2007). If the hazard persists (earthquake, flood, hurricane, etc.), it 442 is transformed by "nature-society hybridization" (Reghezza-Zitt et al., 2012), i.e. by the actions and 443 practices of humans in their environment. Thus, while natural hazards are not new, their impacts are 444 evolving due to climate change, urban growth and urban structural changes. 445

The integration of the concept of multi-risk in the management of urban areas 446
Due to the interconnection of territories and the emergence of cascading risks, risk management 447 must evolve from a single-risk to a multi-risk approach (Kappes et al., 2012) in order to understand the 448 diversity and consequences of interactions and interconnections. Whether due to a combination of 449 several natural risks, "about 3.8 million km 2 and 790 million people in the world are relatively highly 450 exposed to at least two hazards, while about 0.5 million km 2 and 105 million people to three or more 451 hazards" (Gallina et al., 2016) or as a result of cascading effects following a specific risk, or of man-452 made disasters, territories and populations are exposed to a multitude of risks, forcing stakeholders to 453 innovate in traditional risk management. Furthermore, the climate change context increases the 454 likelihood of multi-risk exposure (Dilley et al., 2005;Komendantova et al., 2014). For instance, the 455 positioning of inter-tropical islands exposes them to a combination of risks such as storms, cyclones and 456 coastal erosion associated with the gradual rise of the oceans. If only one of these risks were analyzed 457 in a disconnected way from the others, the risk analysis, strategies and management established would 458 not be adequate and realistic to prepare these territories and their populations (Rosendahl Appelquist 459 and Balstrøm, 2014). 460

461
Risk management must therefore focus on integrated management in order to address the multitude 462 of interconnected risks. This comprehensive approach will allow considering their short-and long-term 463 impacts, which can have cascading effects, and to innovate in solutions adapted to an interconnected 464 world (Garcia-Aristizabal et al., 2012). Multi-risk assessments and all-hazards approaches need to be 465 strengthened, overcoming the limitation of single-hazards assessments in defining suitable and cost-466 effective resilience measures in regions potentially affected by multiple sources of natural hazards. From 467 an operational perspective, multi-risk and multi-level (vertical/horizontal) governance frameworks 468 shifting from a single (siloed) risk focus to embracing a multi-risk approach when working with 469 technical and political authorities should be co-developed and co-evaluated. 470

472
The context of over-urbanization has led to a situation of vulnerability of urban spaces to risks. At 473 present, half of all people live in urban areas, a rate that is expected to reach 70% by 2050. This 474 concentration of people and goods weakens territories in the face of the growing increase in urban risks.

475
Because of the inter-connected world, the interdependence between the different urban systems (virtual 476 and/or physical), accentuates the dependence and vulnerability of populations and spatial functioning. 477 Some infrastructures, essential to the proper functioning of the territory, are more targeted. Faced with 478 the potential disruption of one of these critical infrastructures, a chain reaction can occur and have a 479 lasting impact on territories that cross administrative borders. The city needs to be analyzed as "a system 480 of systems, with each of those systems (e.g. communications, water, sanitation, energy, healthcare, 481 welfare, law and order, education, businesses, social and neighborhood systems) potentially having 482 separate owners and stakeholders" (UNISDR, 2017). The collaborative process underlying an 483 assessment of systemic vulnerabilities emerging from such an interpretation lays the foundations for 484 expanding the risk assessment framework towards wider objectives linked to the resilience of urban 485 systems in a multi-risk perspective (UN-Habitat, 2017). In the face of these growing uncertainties, risk 486 management must evolve and provide local managers and decision-makers with the keys to solutions. 487 New concepts are therefore gradually being integrated into risk management in order to help territories 488 and populations adapt to climate change, growing risks and related uncertainties. 489 490

Urban resilience: advances and limits 491
Faced with these growing challenges related to risks in urban areas, risk management has therefore 492 evolved by adapting the concept of resilience to the analysis of risks in urban environments. 493

Urban resilience 494 495
Urban resilience can therefore be defined as the concept that studies urban systems, i.e. the 496 interactions between the different components that participate in the creation of the territory. Urban 497 resilience refers to a systemic approach that encompasses the multiple layers (built, social, political, 498 etc.) and structures that produce an integrated vision of the urban object. Urban resilience would 499 therefore be a tool for analyzing the complexity of the urban system and defining the different capacities 500 and capabilities of each element that defines this system in order to live and survive a disruptive event. 501 The ability to define what is meant by resilience is an essential prerequisite for reducing the 502 consequences of a disaster. Determining what is "at risk" in a specific area is an essential step in this 503 regard. But when we talk about urban resilience, aren't all elements are essential? Most research on 504 operationalizing resilience focuses on a technical-functional approach ( Table 1). As a result, it is mostly 505 the technical and material elements, such as urban networks, that are analyzed in these studies (Gonzva 506 and Barroca, 2017; Lhomme et al., 2013;Serre, 2018Serre, , 2016. However, an urban system is made up of 507 multiple components that are constantly interacting. There is no conceptual and theoretical consensus 508 in the scientific and policy community (Table 1) on the definition and objectives of urban resilience, 509 which reinforces the lack of clarity in establishing resilient risk management strategies. 510 511 https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License.

Sources
Systems Definitions

OECD Cities
Resilient cities are cities that have the ability to absorb, recover and prepare for future shocks (economic, environmental, social & institutional). Resilient cities promote sustainable development, well-being and inclusive growth C40 Cities Cities are at the forefront of experiencing a host of climate impacts, including coastal and inland flooding, heat waves, droughts, and wildfires. As a result, there is a widespread need for municipal agencies to understand and mitigate climate risks to urban infrastructure and services -and the communities they serve.

ICLEI Cities
A resilient city is prepared to absorb and recover from any shock or stress while maintaining its essential functions, structures and identity as well as adapting and thriving in the face of continual change. Building resilience requires identifying and assessing hazard risks, reducing vulnerability and exposure, and lastly, increasing resistance, adaptive capacity, and emergency preparedness.

Resilience Alliance
Cities A resilient city is one that has developed capacities to help absorb future shocks and stresses to its social, economic, and technical systems and infrastructures so as to still be able to maintain essentially the same functions, structures, systems and identity. Alberti et al., 2008 Cities The degree to which cities tolerate alteration before reorganization around a new set of structures and processes Campanella, 2006 Cities The capacity of a city to rebound from destruction Lamond and Proverbs, 2009 Cities Encompasses the idea that towns and cities should be able to recover quickly from major and minor disasters Lhomme et al., 2013 Cities The ability of a city to absorb disturbance and recover its functions after disturbance Urban Resilience Hub

Urban system
The measurable ability of any urban system, with its inhabitants, to maintain continuity through all shocks and stresses, while positively adapting and transforming toward sustainability https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License. Holling, 1973 System The persistence of relationships within a system, a measure of the ability of systems to absorb changes of state variables, driving variables, and parameters, and still persist

UNISDR System
The ability of a system, community or society exposed to hazards to resist, absorb, accommodate, adapt to, transform and recover from the effects of a hazard in a timely and efficient manner, including through the preservation and restoration of its essential basic structures and functions through risk management

100RC System
The capacity of individuals, communities, institutions, businesses, and systems within a city to survive, adapt, and grow no matter what kinds of chronic stresses and acute shocks they experience Pickett et al., 2004 System The ability of a system to adjust in the face of changing conditions Godschalk, 2003 Critical infrastructure networks A sustainable network of physical systems and human communities Serre et al., 2013 Critical infrastructure networks Urban resilience aims to maintain urban functions during the event and recover thanks to resistance capacities (assessing damages), absorption capacities (assessing alternatives) and recovery capacity (assessing accessibility)

Cimellaro et al., 2010
Critical Infrastructure Resilience is defined as a function indicating the capability to sustain a level of functionality or performance fora given building, bridge, lifeline networks, or community, over a period defined as the control time that is usually decided by owners, or society Ouyang et al., 2012 Critical Infrastructure Resilience as the joint ability of infrastructure systems to resist (prevent and withstand) any possible hazards, absorb the initial damage, and recover to normal operation Longsttaff, 2005 Community The ability by an individual, group, or organization to continue its existence (or remain more or less stable) in the face of some sort of surprise Adger, 2000 Community The ability of communities to withstand external shocks to their social infrastructure Ganor, 2003 Community The ability of individuals and communities to deal with a state of continuous, long term stress; the ability to find unknown inner strengths and resources in order to cope effectively, the measure of adaptation and flexibility https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License. Coles, 2004 Community A community's capacities, skills and knowledge that allow it to participate fully in recovery from disasters Wagner and Breil, 2013 Community The general capacity and ability of a community to withstand stress, survive, adapt and bounce back from a crisis or disaster and rapidly move on Asprone et al., 2014 Hybrid approach City resilience is based on the efficiency of hybrid networks composed by citizens and urban infrastructures.

Heinzlef et al., 2020
Hybrid approach The ability of populations, territories and infrastructures to put in place resources, skills and capacities in order to best experience a disruptive event so as to limit its negative impacts. Capacities can be both tangible (urban networks, supply of vital resources, etc.) and intangible (knowledge of risk, economic dynamics, institutional framework, etc.). One of the reasons for this lack of clarity is the complexity of current urban systems. The city 515 is a complex object to define, describe and analyze. The urban components are and the models are 516 struggling to analyze the urban system. Urban growth accompanied by urban, social, technical, political 517 and economic changes are leading to a fragmentation of urban space. This fragmentation and increasing 518 complexity does not and shared knowledge of urban space, which is a prerequisite for a global and 519 shared vision and knowledge of the urban space. complicates risk management. 520 Since the 1970s, systems thinking has emerged to address complex systems. The difficulty of 521 defining the city as an object emphasizes complexity and therefore suggests that we can consider the 522 city as a system. A system can be defined as a set of elements and interactions between elements that 523 form an organized whole, with the internal organization of the system constituting its structure and the 524 behavior of the interacting elements defining its dynamics. Each system is defined according to a 525 purpose and an objective. 526 This urban system (Bretagnolle et al., 2006) is defined by the interdependencies existing 527 between the various components of the city, due to the multiple networks of relationships they have with 528 each other. The systemic approach aims to observe, interpret and reconstruct the real world by putting 529 forward hypotheses on the organization of cities (Paulet, 2009). The analysis of a system is therefore a 530 construct and presupposes choices among the different variables of the system. The study of a city today 531 therefore implies understanding and considering the interdependencies between cities and their 532 components, and analyzing their connections. 533 The city thus first of all creates interweaving urban components, such as technical systems (such 534 as urban networks and/or critical infrastructures) or public infrastructures (governance, education, 535 health, police, justice, etc.) (Lhomme et al., 2013), but it also creates interrelationships with its 536 environment, due to its open system characteristic. As an open system, it both transforms itself through 537 intrinsic capacities but also receives resources through flows and information from their environment. 538 Since it is not self-sufficient, the relationships between cities and countryside, between cities and towns, 539 are essential and must be analyzed in the global study of an urban system. These interactions can 540 therefore be considered as a source of wealth, more (food) resources, knowledge, techniques, but also a 541 source of fragility (uncertainties, overproduction of waste, new urban risks, social risks, etc.). This non-542 exhaustive list nonetheless allows us to understand the fragility of urban environments, because their 543 sources of growth, expansion and wealth can, at its peak, also be synonymous with vulnerability. 544 The city is therefore a complex object to apprehend and study. Because of its construction 545 protean, the city is difficult to define and identify as a single object. The evolutions are constant and 546 https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License. vary according to the social, urban and technical components, environmental, political, economic, etc., 547 of the urban space. Focusing on the issues at stake challenged by risk, cities are also concentrating 548 resources to deal with it. This is so in these spaces that urban systemic resilience must be analyzed and 549 operationalized. 550 551 3.3. … Including some limits 552 553 Despite its growing importance and use in expert and policy discourse, the concept of resilience 554 faces many limitations. 555 556 3.3.1. A conceptual vagueness 557 558 The concept of resilience faces a conceptual confrontation in the multitude of definitions and 559 associated notions. This concept is today over-used, over-solicited in multiple fields (psychology, 560 ecology, political science, physics, geography, etc.) and related to many concepts (Emrich and Tobin,561 2018). This multitude of uses has turned it into a buzzword (Reghezza-Zitt et al., 2012), a word 562 "suitcase" (Rufat, 2015) that complicates its understanding. A resilient system is in turn defined as a 563 system capable of stability but also of adaptation and evolution (Hegger et al., 2016;Tempels and 564 Hartmann, 2014). We speak of both "bouncing back" to a (potentially anterior) equilibrium or "bouncing 565 forward" to a new state of balance and harmony. Faced with this ambiguity, or even contradiction, 566 among the objectives and guidelines of resilience, actors and experts come up against grey areas (Disse 567 et al., 2020). Beyond these two characteristics, Brand and Jax (2007) analyzed the studies, definitions 568 and methodologies addressing the concept of resilience over the past 35 years and pointed to the abstract 569 trend of resilience. According to them, resilience must therefore be perceived and understood as a 570 perspective (of planning, risk management, spatial and social development) rather than as a concept or 571 tool to be clearly and unanimously defined (Kim and Lim, 2016 This conceptual vagueness has contributed to the political reappropriation (Béné et al., 2018) 576 of the concept of resilience without resulting in clear strategies adapted to local actors and territories at 577 risk (Bahadur and Tanner, 2014; Béné et al., 2012;Cannon and Müller-Mahn, 2010;Duit et al., 2010) . 578 Many scientists and experts have denounced the tendency to overuse and abuse the term resilience. 579 Having become a political and management imperative, resilience has been transformed into a political 580 and crowd-unifying tool. Resilience can therefore be used more for political positioning or institutions 581 to strengthen their dominant governance model without necessarily leading to reflection on processes 582 of transformation or evolution that are generally necessary for the establishment of resilient systems 583 (Béné et al., 2018) .

585
Beyond limitations related to the lack of consensus on the concept of resilience, there are also 586 limits to its implementation in risk management strategies. 587 588

Financial limitations 589 590
The cost of a resilient approach or accommodations is often pointed out. Whether it is spatial 591 redevelopment (reworking urban density, refuge areas, critical infrastructures, risk areas, etc.) or the 592 purchase of so-called resilient development tools (Heinzlef et al., 2020), local managers and actors are 593 faced with a mismatch between the cost of this approach and their daily priorities. The fact also that 594 climate change and the associated risks are a more or less distant threat and hardly imaginable threat, 595 makes decision-makers less focused on the necessary evolution of risk management strategies through 596 the integration of resilience into the planning process (Leichenko et al., 2015). 597 598 https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License.

Cultural barriers 599 600
The local cultural dimension of risk management can also be seen as a barrier (Heinzlef et al.,601 2020b) to the implementation of the concept of resilience (Heinzlef et al., 2019a). This socio-cultural 602 dimension can be expressed at several levels. When resilience becomes an operational object, it often requires technical management tools. 615 However, the general bias for operationalizing resilience involves its quantification and representation. 616 Simply put, the tool must be able to conclude whether or not the territory is resilient. Numerous studies 617 have provided answers to this issue. After establishing the need for urban technical networks in the 618 functioning of urban territories, concluding that these networks contribute to the resilience of urban 619 areas becomes obvious. A great deal of research has therefore analysed resilience through the resistance Faced with the difficult consensus around the concept of resilience, its operationalization is regularly 632 questioned. The difficult formalization, linked to the multitude of interpretations and approaches, results 633 in a complex transition from theory to practice. However, this is the challenge posed by all studies on 634 resilience, in order to use this concept to build adequate risk management strategies. Several approaches 635 have therefore attempted to respond to these challenges by proposing methodologies that aim to 636 operationalize resilience. This operationalization translates into the design of tools for measuring 637 resilience, spatial decision support systems or approaches that promote collaboration between experts 638 and local stakeholders. In this section, we will analyze some of these works, frameworks, structures and 639 methodologies. Some authors have attempted to synthesize all existing models (Constas et al., 2010;640 Schipper and Langston, 2015) but the forty or so models mentioned (Bahadur et al., 2015) underline the 641 (over)abundance of approaches to resilience. We will attempt to scan the approaches aimed at assessing 642 resilience through the creation of indicators, models proposing a conceptual framework or decision 643 support systems, and then methodologies aimed at creating collaborative work in order to operationalize 644 resilience. 645 646 647 648 649 https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License.

Methods and tools for evaluation, modelling and integrating resilience into risk management 650 651
A large part of operationalization involves determining how a concept can be measured (Adger 652 et al., 2004) and determining which indicators will be used to measure the concept in order to 653 generate data about it. The assessment of resilience therefore essentially involves its measurement 654 (Heinzlef et al., 2020a)  Indicators are quantitative variables intended to represent a characteristic of a system or concept. 661 They have been used to inform decision making, improve stakeholder participation, build consensus, 662 explore underlying processes, etc. (Parris and Kates, 2003). The objective of an indicator is to provide 663 information that should help actor to steer the course of action towards the achievement of an objective 664 or to enable him to evaluate the result. The indicator can be a parameter, a value, a data or an observation. 665 Its objective is to give indications or describe a phenomenon, a situation, an environment or a process. 666 It is necessary to define a preliminary objective to which the indicators will tend. An indicator can be 667 composed of a single variable or a combination of variables (Birkmann, 2006). 668 Regardless of the word used, the indicator primarily defines the compelling relationship between 669 the information contained in the indicator and the object pointed to by the indicator (Birkmann, 2006). 670 The function of the indicator is therefore to show, to place in space, and it is this spatializing nature that 671 makes it interesting as a geographical tool (Freudenberg, 2003). If the indicator in the primary sense of 672 the term does not analyze or define, it takes on its full meaning through the observer's reading of it.

673
Because of its eminently operational nature, it enables observations and results to be anchored in 674 practical reality. It answers directly to the question asked by the user confirming or not the initial 675 hypothesis. The hypotheses and judgements made when choosing the questions and data relevant to the 676 development of the indicator, as well as the evaluation of the usefulness of the indicator, require the 677 existence of objectives, implicit or explicit. An indicator collects data and information in order to 678 aggregate knowledge, which is essential for making the right choices (Wisner and Walter, 2005). For 679 this reason, indicators are fully involved in the decision support process (Tate, 2012). 680 However, despite its operational nature, the indicator is only an experience of reality and not a 681 proper experiment. It is therefore necessary to bear in mind that it is a practical image of reality but that 682 it is not objectively the reality of the territory. The indicator merely reproduces or reconstructs an image 683 of the geographical space, which makes the choice of indicator, variables and treatments very subtle and 684 complex. However, this choice is itself built around representational a priori, a socio-cognitive paradigm 685 that cannot be denied. It is therefore necessary to make the construction of these variables and indicators 686 as objective as possible in order to claim that the results are real. There are several "formats" of 687 indicators. Multiple indicators, for example, can be combined with unstructured composite indicators, 688 or indices, which attempt to distil the complexity of an entire system into a single measure. Social 689 indicators have been used since the 1960s, with applications to the environment (1970s), sustainability 690 (1990s), and more recently vulnerability (Birkmann, 2006;King and Macgregor, 2000) and resilience 691 . The main global indices and recent regional studies that model various aspects of 692 vulnerability include the vulnerability index, the Human Development (UNDP), the Disaster Risk Index 693 (UNDP 2004)  Measuring resilience has become an international priority in order to build strategies for the 701 future. risk management (Winderl, 2014). The question of how to measure resilience is as old and as 702 important as the concept itself (Prior and Hagmann, 2014). Numerous indices and indicators of resilience 703 have been developed in various disciplines. In general, they are used for different purposes and, as a 704 https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License. result, they measure different things. An exploration of attempts to measure resilience reveals the 705 difficulty in establishing a measure that is both accurate and "fit for purpose" (Hinkel, 2011). 706 Measurement requires that a phenomenon be observable and allow for systematic attribution of value, 707 but the conceptual nature of resilience makes this difficult. Scientists do not have not yet agreed on 708 specific conventions for measuring resilience and, consequently, there is a substantial literature that 709 discusses both how and whether the phenomenon can and should be measured (Hinkel, 2011). 710 The identification of resilience requires planners to identify variables that trigger disturbances 711 in a city (a community, region or landscape), the frequency and intensity of these events, and the 712 mechanisms that enhance adaptability that can be activated to respond to (or avoid) these disorders. It 713 is need to assess the socio-economic dimensions of an urban area (Ahern, 2011). As established 714 previously, it is necessary to establish common denominators that induce vulnerability or strengthen 715 resilience (Gonçalves, 2013) . However, the difficulty essential is to measure these dimensions. The 716 significant challenges in measuring the resilience lead either to imperfect quantified measurements or 717 to a search for indicators of universal resilience (Hallegatte and Engle, 2019). Cutter et al. (2008) 718 highlight this difficulty in believing that "if we conceptually or sometimes intuitively understand the 719 vulnerability and resilience, the devil is always in the details, and in this case, the devil is measurement" 720 (Cutter et al., 2008b environmental -Cutter proposes to measure resilience (Cutter et al., 2014). Each indicator is divided 730 into sub-variables such as education, age, language proficiency, employment rate, immigration rate, 731 access to food, disaster training, social stability, access to health, access to energy, etc. (Cutter et 732 al., 2014). Each variable has a positive or negative effect on community resilience. Data acquisition 733 was an important issue. More than 20 data sets were obtained from the U.S. federal government 734 through online data portals. Four datasets were obtained from NGO websites, two through a contact 735 with the American Red Cross, and one from an open access data portal at a major press briefing. 736 One data source was the Dun and Bradstreet's Million Dollar Database and required a paid 737 subscription to acquire the data (Cutter et al., 2014). Once the data acquisition was completed, a 738 processing work, "cleaning" of the data was necessary. The chosen method of treatment was applied 739 to the Min-Max. The Min-Max normalization assigns a value of 0 to the minimum value and from 740 1 to the maximum value. All other values are scaled to between zero and one by subtracting the 741 minimum value and dividing by the range (minimum subtracted from the maximum). While this 742 method makes it much easier to compare between a large number of variables, the disadvantage 743 remains that the final score is not a measure absolute value of community resilience for a single 744 location, but rather a relative value of community resilience for a single location. in which several 745 locations can be compared. This is why the proposed work is done at the US level and not at a finer 746 scale or for an single year, not being a comparative work over several years. 747 This approach is a key work in the process of operationalizing the concept of resilience. In 748 addition to the definition of resilience criteria, it also makes it possible to locate more or less finely 749 the territories on which to focus efforts to increase territorial and social resilience. Its systemic 750 approach to the territory (considering the elements that make up the territory) is completely adapted 751 to risk analysis. 752 753  The DS3 Model (Spatial Decision Support System) (Serre, 2018)

755
The DS3 Model has defined three capabilities to assess the resilience of urban networks to flood risk.

756
Resilience is defined here as the ability of a system to absorb a disturbance and subsequently recover its 757 functions. Three capabilities are assumed to determine the degree of resilience of these networks: 758 759 https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License.
o The capacity to resist: this consists of determining the material damage following a risk. 760 It is considered that the more a network is damaged, the more likely it is that there will 761 be a slower and more complex return to service; 762 o The absorption capacity: it illustrates the fragilities and strengths of the network 763 allowing to build alternatives to it following a component failure; 764 o Recovery capacity: this represents the time required to return to service of the network 765 and its components. 766 767 These capabilities enable the resilience of a city's urban technical networks to be defined and 768 measured. The methodology was tested to assess the resilience capacity of the Hamburg district, Am 769 Sandtorkai/Dalmannkai. Each resilience capacity was analyzed according to the components of the 770 neighborhood, at its scale and then according to the interactions with its environment. Using this 771 technical resilience measurement tool, the case study analysis identified interdependencies and potential 772 domino effects at the neighborhood level. The definition of these three capabilities made it possible to 773 analyze resilience over a long period of time, before, during and after a disturbance. The systemic 774 approach here is defined by the analysis of inter-network interactions and interconnections in order to 775 assess cascading risks in urban environments. 776 777  An hybrid approach (Heinzlef et al., 2020a)

779
This research made it possible to develop three indicators for defining and measuring resilience in 780 order to gain a comprehensive and exhaustive understanding of the concept. These indicators analyze 781 the urban, social and technical resilience of a city (Heinzlef et al., 2019a).

783
o The social resilience indicator illustrates a population's ability to adapt and recover from 784 disruption (Hutter and Lorenz, 2018). Many factors contribute to social resilience, 785 including age , community and political investment (Voss, 2008), 786 socioeconomic status (Flanagan et al., 2011), knowledge and perception of risk, etc. 787 (Hutter and Lorenz, 2018). This methodology understands social resilience as 788 community resilience (Wilson, 2013) and not individual resilience (Hutter and Lorenz,789 2018).

790
o The urban resilience indicator includes all urban dynamics, such as physical elements 791 (Norris et al., 2008;Opach and Rød, 2013) (age of the building, urban density,

792
building functions, critical infrastructure, etc.) or more virtual elements such as 793 economic dynamics through the creation or suppression of businesses or touristic 794 dynamism (Tierney, 2014).

795
o The technical resilience indicator includes urban networks (Serre, 2016). It is used to 796 analyze the diversity and accessibility of these networks within a radius of 100m in 797 order to assess their resilience and their impact on the territory in the event of a crisis 798 (Heinzlef et al., 2020a).

800
This study has been tested and validated in Avignon (France), and built with 90% open data in order 801 to allow the reuse of this methodology on other national case studies. The systemic approach is 802 illustrated by taking into account the multitude of elements that make up the urban territory in order to 803 have a global vision and approach to the territory, its population and its potential resilience. As the concept of resilience is a complex subject to address and operationalize for local actors, 811 many tools have been created to simplify, define, measure and attempt to operationalize this concept.

812
The need to create decision-support systems makes sense in terms of the abstraction of the concept. In 813 risk management, taking is a complex combination of knowledge management and decision-making 814 https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License. processes. reasoning (Tacnet et al., 2014). Decision-support systems are defined as integrated computer 815 systems, designed for decision making. When territorial issues are addressed, these are referred to as 816 spatial decision Support System (DSS). They combine spatial and non-spatial data, functions analysis 817 and visualization of Geographic Information Systems (GIS) and decisions in order to construct, evaluate 818 and produce solutions (Keenan and Jankowski, 2019). These space-based decision support systems have 819 been developed to address the limitations of the GIS such as lack of modeling capabilities and lack of 820 flexibility of GIS for adapt to variations in the context or spatial decision-making process (Densham, 821 1991 Indeed, current tools such as GIS are often inadequate in the face of the complexity of the real 826 issues facing users (Andrienko et al., 2007). For individuals, the visual context favors the acquisition of 827 knowledge (Kwan and Lee, 2003). There are many forms of data visualization that are primarily 828 scientific and information visualization (Marzouki et al., 2017). If data have a combination of spatial, 829 semantic and temporal dimensions, they are referred to as geographic data/information and geo-spatial 830 data (Marzouki et al., 2017). The visualization of these data then becomes specific, and goes beyond 831 simple scientific and information visualization (Kurwakumire et al., 2019). The integration of 832 visualization in the analysis of geo-spatial data has led to a transformation of traditional mapping 833 through the digital era (Çöltekin et al., 2017) . This evolution of traditional mapping has led to 834 geovisualization, a "set of visualization methods and tools for interactively exploring, analyzing and 835 synthesizing location-based data for knowledge building" (Dykes and International Cartographic 836 Association, 2007). Geovisualization combines scientific visualization, information visualization, 837 mapping, geographic information systems (GIS), exploratory data analysis and many other methods to 838 explore, analyze, synthesize and represent geographic data and information (Nöllenburg, 2007) . As a 839 result, many spatial decision support systems have been equipped with visualization techniques and 840 dynamic interfaces to combine technological capabilities with local interpretations and knowledge. Map 841 production is accessible and understandable through a visual interface to enable exploration, 842 understanding, analysis and reuse of a complex, geolocalized and heterogeneous database. 843 Thanks to a dynamic interface and a technical power capable of processing complex data, 844 geovisualization tools allow to communicate information about complex data necessary for the decision 845 support process. In addition, the interactivity of these tools allows the users to be actors in front of the 846 tool, by navigating, by making a visualization request, by downloading data or displaying information 847 as needed. The tools of are therefore both communication tools and tools for the production of 848 geovisualisation knowledge by being an integral part of the "reflection/knowledge process".

849
( MacEachren et al., 2004). As Bishop et al.(2013) point out, neuroscience has been a major contributor 850 to the development of the human brain. and has already demonstrated that visualization techniques are 851 essential to cognitive processes. leading to decision making (Padilla et al., 2018). Geovisualization thus 852 integral part of spatial decision support systems, as it allows to meet both scientific and societal needs 853 to initiate a process of reflection and thereby build and produce knowledge.

855
Several methodologies have produced tools to clarify the concepts of resilience and 856 vulnerability. These tools are spatial decision support systems and have made it possible to dissect the 857 concept of resilience. The objective of each of these approaches is to make the concept accessible by 858 creating links between scientific advances and territorial reality. 859 860 4.2.3. Examples of spatial decision support systems 861 862  The DOMINO tool (Robert et al., 2008)

864
A tool for modelling the spatial and temporal propagation of domino effects between critical 865 infrastructures (CI) has been developed for the city of Montreal. It consists of a geographic database in 866 which organizations have entered relevant information about their dependencies on the critical resources 867 they use. Modules, created on the structure of the expert systems, combine information from different 868 organizations in order to identify interdependencies between them. A time simulator has also been 869 https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License. developed to visualize the propagation of potential domino effects following a failure. Being a geomatics 870 tool, it is It is possible to combine several layers of information in DOMINO. The partners' managers 871 (CI managers and emergency preparedness officials) have secure access to it, managed according to 872 levels corresponding to their user profile. Thus, each organization has access to its information, while 873 the results of the simulations are available to all the users. In terms of resilience, DOMINO allows 874 analyses on the different parameters of the resilience. Its establishment on a territory requires the various 875 organizations concerned to exchange information on their own disruption management capability. The 876 implementation of this information enables consistency analyses to be carried out on a given territory 877 and integrate broader community implications. The systemic dimension of this tool lies in the fact that 878 it analyses the interdependencies of critical infrastructures in an urban space and models potential 879 service disruptions. 880 881  The ViewExposed tool (Opach and Rød, 2013)

883
A Norwegian study addressed the issue of vulnerability of territories s response to climate change 884 (Opach and Rød, 2013). In order to avoid an increase of local and national vulnerability, the researchers 885 have developed a ViewExposed tool, including the aim is to inform local authorities about the most 886 vulnerable areas of the territory Norway and also the causes of this vulnerability. The methodology used 887 is based on the work of SoVi (Cutter et al., 2003) and the University of South Carolina (Tate, 2012 The objective of IntVI is to focus on a municipality's exposure to natural hazards and to put it into 898 perspective with regard to the local population's capacity to resist them. For PhyVI, the exposure of 899 municipalities to natural risks is expressed as a percentage and depends on the work of Norwegian 900 insurers (Norwegian Natural Perils Pool). Based on the data, the researchers were able to determine that 901 during the period 1980-2010, 60% of the damage was caused by storms, 26% by floods, 7% by 902 landslides and 5% by storm surges (Opach and Rød, 2013). Concerning SoVI, the objective was to 903 assess the adaptive capacity of municipalities with regard to physical exposure. Thus, in the next step, 904 the SoVI was calculated using the methodological framework constructed for Norway by (Holand et al., 905 2011). Finally, PhyVI and SoVI were compiled to create IntVI. To do so, the weights most correlated 906 with the Norwegian Natural Perils Pool claims data were used: 60% for PhyVI and 40% for SoVI. The 907 tool, which takes the form of an interface, has been created for professionals, local elected officials and 908 residents. It is the product of a collaboration between scientists and local experts through workshops 909 (Opach and Rød, 2013). The authors of the interface wished to answer two fundamental questions: Who 910 are the vulnerable people? And where do they live? The objective is therefore to identify regions with a 911 high level of social vulnerability to environmental risks in order to reduce it and thus help to improve 912 their resilience. In addition, the platform is open and scalable as any actor in the field can submit a 913 reflection using the "submit a comment" section of the interface. Although focused on the concept of 914 vulnerability, this tool also integrates the response of local managers and actors to natural disasters. It is 915 therefore both the vulnerabilities but also the resilience strategies that are integrated. In addition, this 916 tool proposes a collaborative and participatory approach between local actors and scientific experts. 917 918 919 4.3. Integrating resilience into urban management through collaborative approaches 920 921 The United Nations International Strategy for Disaster Risk Reduction (UNISDR) has developed 922 10 key points for creating resilient cities. The first point is to set up organizations or coordinations to 923 understand and reduce risks, based on the participation of citizens and civil societies (UNISDR, 2015). 924 https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License.
The objective is to build local actions and alliances to ensure that each actor understands his or her role 925 in reducing and preparing risk reduction and resilience strategies (Heinzlef et al., 2020b;Gupta et al., 926 2010). 927 928 4.3.1. Collaborative approaches, a key for operationalizing resilience 929 930 Involving "local" people or people directly concerned by the issues studied does not appear to be 931 new (Toubin et al., 2015) and even less original. The richness of having people from all walks of life 932 interact with each other facilitates an exploration of possibilities, enriching discussions, encouraging 933 cross-fertilization of views on the same subject, making it possible to be both more measured and more 934 incisive in a specific area. The contribution of "profane" knowledge in thorny social and societal issues, 935 as scientific knowledge cannot respond to all uncertainties, with the result that "expert" conclusions are 936 called into question. Resilience, a social and thorny concept, is therefore a subject that would require 937 the confrontation of views, knowledge, scientific and practical knowledge, perceptions and 938 interpretations. However, although the population is often the first to be impacted by natural hazards 939 and their inappropriate management, the fact remains that the inhabitants (Kuhlicke et al., 2011) and 940 also the urban services (Toubin et al., 2015), which are nonetheless first-rate actors, are not sufficiently 941 involved. The defended idea is that the creation of a hybrid knowledge (Djenontin and Meadow, 2018;942 Lemos and Morehouse, 2005;Schneider and Rist, 2014) allowing the involvement of all actors of the 943 territory, from the inhabitant to the manager via the scientist, would make it possible to operationalize 944 urban resilience thanks to an appropriation of the concept and stakes of urban risks. In fact, collaboration 945 is mainly based on the appropriation of the different stakeholders of the same subject of tension and 946 discussion. Collaboration therefore goes beyond the simple exchange of knowledge and information, 947 but makes it possible to "create a shared vision and articulated strategies for the emergence of common 948 interests that extend beyond the limitations of each particular project" (Chrislip, 2002 In her thesis, Marie Toubin develops a methodology to contribute to the improvement of conditions 955 of urban resilience and more particularly the resilience of urban networks. Her analysis of the 956 interactions and interdependencies of urban networks highlighted the intrinsic fragilities of urban 957 systems and their management in the event of a crisis. Faced with the challenges observed, the research 958 objective was therefore to develop methodological approaches and tools to help urban service managers 959 identify and characterize technical and organizational interdependencies in order to ensure service 960 continuity despite a disruption. The approach was built by integrating the main managers of the City of 961 Paris' urban services. The methodology made it possible to construct interviews to assess the criticality 962 of the resources required for the system to function properly. It was therefore possible to rank the 963 resources according to their importance and use. This research made it possible to draw up and analyze 964 the interdependencies of the Parisian urban networks. It highlights certain dependencies, particularly 965 those on electricity, telecommunications and travel. The collaborative approach made it possible to 966 involve managers in thinking about strategies to mitigate or at least manage these interdependencies. 967 Moreover, the collaborative process has illustrated the need to move beyond isolated approaches but 968 instead to foster a common vision. The interweaving of scales but also of services makes cooperation 969 and transparency between operators and decision-makers indispensable for the construction of a more 970 resilient city. 971  Resilience by design in Mexico City: A participatory human-hydrologic systems approach 972 (Freeman et al., 2020) 973 The study developed by Freeman et al. (2020) in Mexico City highlights issues of building a 974 freshwater resilience of urban systems among several territories and stakeholders. In order to find a way 975 to manage systems of feedbacks and tradeoffs between stakeholders, Freeman et al. have developed a 976 Resilience by Design methodology (Brown et al., 2020). The aim of this methodology is to identify with 977 https://doi.org/10.5194/nhess-2020-217 Preprint. issue, this methodology provides a planning and common framework to identify solutions and 979 compromises between urban managers, political stakeholders and decision-makers. In this case study, 980 Resilience by Design methodology revealed "consistent stakeholder preferences for social (such as 981 equity in water allocation among users) and economic performance", such as domestic, agricultural and 982 industrial sectors. These common solutions guide to persistence, adaptation and transformations.

983
Understandings and choices about how much "resilience of what, to what, for whom and at what cost" 984 require a shared narrow and adaptive approach (Freeman et al., 2020). Thinking jointly about issues and 985 related solutions helps to establish an understanding of the concept of resilience and established 986 strategies over time. Actors must therefore debate and envisage solutions in an egalitarian and united 987 manner in an evolving territory in order to tend to increase its resilience. 988

990
Several methodologies exist in order to operationalize resilience concepts and integrate it into urban 991 risks strategies. The main approaches are divided into three categories: (1) assessing resilience through 992 its measure with indicators, (2) modeling resilience with geovisualization techniques and (3) developing 993 collaborative approaches in order to lead to resilience understanding and adoption by stakeholders. 994 Indicators are helpful to define main resilience characteristics and to provide a measurement to 995 analyze resilience potentialities. These indicators might be specific (Serre, 2018) or exhaustive (Heinzlef 996 et al., 2020a). They have an important utility to urban managers to define low resilience areas and 997 concentrate their strategies on it. 998 Geovisualization techniques are used to unbuilt resilience abstraction thanks to tools, interfaces and 999 data which allow comprehension and facilitate resilience integration. Interactivity, communication, 1000 navigation, visualization lead to a precise resilience analyze. These tools are essential for knowledge 1001 construction and sharing and are part of the "reflection/knowledge process". 1002 Finally, collaborative approaches lead to local stakeholders' responsibilities to integrate resilience 1003 into risk strategies management. It is useful to create a shared vision on complex concepts and strategies 1004 between "experts" and "local actors". Their proper experiences (local risk management heritage and 1005 scientific knowledge) lead to a territorialized risk and resilience strategies. It is also a long-term 1006 guarantee to resilience strategies adoption. 1007 1008 5. Discussion 1009 1010 The multitude of existing models for operationalizing resilience indicates the growing importance 1011 of the concept. These models, as diverse and varied as they may be, are essential to the transcription of 1012 the concept into a concept tool . Going beyond the controversy over the 1013 exact definition of the concept, these models propose to operationalize resilience. The accuracy of their 1014 methodology then takes a back seat because what matters then is not that the model be rigorous, but that 1015 it be operational. However, not everyone has the same objective or goal (   The diversity of these models illustrates the interest and efforts developed to respond to the challenges of operationalizing the concept of resilience. While some apprehend urban resilience through the analysis of networks and through a technical-functional approach, others seek to develop hybrid, more exhaustive approaches that attempt to understand and analyze the diversity of the urban territory. The decision support approach also differs from one tool to another, with some advocating the usefulness of indicators, others justifying the need for visualization to lead to a process of understanding and decision making, and finally, some defending the need to integrate local actors at the beginning of any reflection on the concept of resilience. These models are neither exhaustive nor exclusive and it is necessary to use them jointly or at different times and phases in the construction of a resilience strategy. However, this multitude does not promote the understanding and appropriation of a concept that is still abstract for many local actors and managers. Whether it is due to the overly technical nature of the tools (such as for the DS3 Model), a lack of understanding of the concept or even a lack of knowledge of the tools themselves, local stakeholders have very little appropriation of the operationalization methodologies and therefore the concept of resilience.
A tool to define, measure, clarify and assist in decision-making would therefore be of significant interest. The objective of a new tool can be used as a basis for reflection and suggestions for further progressive implementation of the concept of resilience in risk management strategies. This prototype would be to promote an inclusive approach that would make it possible to bring together the different existing approaches around the concept of resilience and to develop a framework for reflection and action between local actors and scientific experts around the issue of operationalizing the concept. This type of tool could be achieved through the design of a resilience observatory. Observatories are key tools to support the observation, reflection, understanding and analysis of phenomena or territories. These tools, which are at the interface of reality and knowledge, are essential in the decision-making process, allowing the acquisition of knowledge and data while taking the necessary distance to have the most global vision possible of a phenomenon. Their usefulness in establishing monitoring of phenomena, territorial evolution and interaction, make them essential tools for apprehending events over https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License. the long term, which is essential for establishing resilience strategies. A team based at the Oceanic Island Ecosystems joint research unit (UMR EIO) of the University of French Polynesia (Heinzlef et al., 2020(Heinzlef et al., , 2019bSerre et al., 2019) has launched a prototype observatory on the islands of Tahiti and Moorea to analyze, measure and operationalize resilience. The objectives are multiple (Fig.1) and focus in particular on increasing knowledge of territorial risks, the acquisition, storage and enhancement of data related to risks and resilience and finally the integration of stakeholders in the process of reflection and implementation of resilience strategies. This prototype can serve as a basis for reflection and suggestions for further progressive implementation of the concept of resilience in risk management strategies.

Conclusion
This article has provided a review of the concept of resilience and its operationalization. Confronted with a conceptual vagueness and a multiplication of definitions, notions and associated concepts, resilience loses its relevance and usefulness in risk management strategies. Yet this concept, which encourages adaptability, evolution and flexibility, is perfectly in line with climate change and the associated risks and uncertainties.
The currently challenge, whether in the scientific community or in urban planners and decisionmakers sphere, is to work on its operationalization by promoting concept understanding and its adoption by local actors. This need has led to a multitude of scientific positions, tools and methodologies aimed at dissecting the concept of resilience and the concepts and capacities associated with it. These operationalization strategies can promote the design of indicators to define and measure resilience, develop spatial decision support systems to visualize territorial resilience or promote the implementation of collaborative approaches to involve local stakeholders in the integration of the concept in local risk management strategies. Although these methodologies in themselves provide opportunities for reflection or even initiatives for resilience strategies, their contribution remains modest and visible in a very short period of time.
Thinking about a new kind of tool for addressing resilience in the long term and an inclusive approach to the concept and associated methodologies would make it possible to respond to these current limitations. This tool, which would take the form of a resilience observatory, would make it possible to
develop a toolbox, bringing together conceptual and tangible advances related to the operationalization of resilience.

Acknowledgments:
This publication has received financial support from the CNRS through the MITI interdisciplinary programs and from the IRD. https://doi.org/10.5194/nhess-2020-217 Preprint. Discussion started: 6 July 2020 c Author(s) 2020. CC BY 4.0 License.