key: cord-0746604-9l8azl8o
authors: Yokomatsu, Muneta; Park, Hyejeong; Kotani, Hitomu; Ito, Hideyuki
title: Designing the building space of a shopping street to use as a disaster evacuation shelter during the COVID-19 pandemic: A case study in Kobe, Japan
date: 2021-11-11
journal: Int J Disaster Risk Reduct
DOI: 10.1016/j.ijdrr.2021.102680
sha: f0070435f9b056f274ade196109f8e1ab0f7c185
doc_id: 746604
cord_uid: 9l8azl8o

This study considers the risk of a natural hazard-induced disaster occurring during a pandemic, such as the novel coronavirus (COVID-19) pandemic, and develops the idea of utilizing a shopping street with disaster-proof buildings as a temporary evacuation shelter by incorporating countermeasures against the spread of infectious diseases. Using a case study of a shopping street in Kobe, Japan, we estimate shelter capacity by considering the requirement of 6 m(2) of space allotted for each person. The shelter can accommodate 1194 evacuees and provide them with food and drinks for one day, even in the worst case of lifeline disruption. This study proposes a method of designing shelter space, and demonstrates how non-homogeneous and noncontinuous spaces within shopping street buildings can be applied to prevent the spread of infection, through the classification of evacuee types and use of space and facilities designated for each type. The study further examines the liability issue of secondary infection at the shelter with reference to civic law and the roles of government in developing a distributed evacuation framework.

The novel coronavirus (hereafter COVID-19) has rapidly spread worldwide since the initial 32 outbreak in China in December 2019. The World Health Organization (WHO) declared a 33 pandemic on March 12, 2020, and presented the guiding principle of "Avoid the Three Cs" [1] , 34

indicating that COVID-19 spreads more easily in certain places: "crowded spaces with many 35 people nearby," "close-contact settings especially where people have close-range 36

conversations," and "confined and enclosed spaces with poor ventilation," and thus that built 37 spaces should avoid these characteristics. As other slogans and catchphrases, such as "social 38 distance," "stay home," and "new normal," infiltrated society, disaster management experts 39 and authorities were faced with the urgent need to modify evacuation strategies, including 40 expanding the shelter area required per person, from 3 m 2 to 6 m 2 . This was followed by a 41 heavy rainy season in Japan in 2020, during which designated evacuation shelters, such as 42 school gymnasiums and community centers, could only accommodate 30−50 % of the 43 evacuees they had before COVID-19 [2,3]. Accordingly, shelters could not admit almost half 44 of the evacuees, many of whom had to stay overnight in their cars or other places that were 45 not organized disaster shelters. This situation enhanced the Japanese government's 46 awareness to the fact that the total area of available shelter space needed to be increased to 47

provide safe havens for all potential evacuees. To address this issue, the national government 48 recommended that local governments encourage their residents to carry out "distributed 49 evacuation" [4, 5] , which is based on the principle that people should be asked to evacuate 50 not only to designated evacuation shelters but also to other, undesignated shelters available 51 in local communities or ad hoc for individual evacuees, such as hotels, welfare centers, homes 52 of relatives or friends, and religious buildings. It was even indicated that staying at home was 53 also an option, should it be safer than evacuating to another location [4,6−8]. 54

Under the distributed evacuation policy, residents have a greater chance of securing 55 temporary shelters; nonetheless, they may face other challenges, such as finding an 56 appropriate shelter, managing such a shelter, and preparing emergency and relief supplies, 57 by themselves or in cooperation with others. As for the management and operation of 58 evacuation shelters during a pandemic, some international and domestic organizations have 59 published guidelines for preventing the spread of COVID-19 in shelters [1, 4] , which share 60 several basic strategies such as classifying evacuees based on health status and providing 61 separate spaces and facilities based on those classifications (see Appendix A). Moreover, the 62 Japanese government requested each designated evacuation shelter to arrange for the 63 distribution of hygiene products, such as masks, alcohol-based hand sanitizers, and cleaning 64 supplies (see Appendix B), and recommended that evacuees, especially those in undesignated 65 J o u r n a l P r e -p r o o f shelters, prepare and take such hygiene products to shelters themselves. 66

Regarding the problems of insufficient space in designated shelters and difficulties in 67 providing disaster relief supplies to evacuees in undesignated shelters, the authors have 68

proposed the idea of using buildings on a shopping street as undesignated evacuation 69 shelters, and examined the potential to do so in the context of Japan and in particular the 70 case-study area, Taisho-Suji shopping street in Kobe [9] . The study estimated accommodation 71 capacity and available supply of food and drinks based on the available stockpiles of 72 restaurants and grocery stores. It was concluded that the shopping street could provide 73

shelter to approximately 2,300 persons, drinks for 1,700 person-days, and food for 1,200 74 person-days in the case of lifeline disruption, and for 2,800 person-days if substitute facilities 75 are available. Despite the idea's novelty, the risk of the spread of infectious diseases was not 76 considered in this estimation [9] . 77

The purpose of the present study is to develop the idea of using shopping-street evacuation 78 shelters during a pandemic such as COVID-19 from viewpoints of disaster management and 79

planning. Concretely, this study aims to: (i) update the estimation of a shelter's capacity to 80

accommodate evacuees and supply food and drink using the shopping street's inventory 81

considering the risk of the COVID-19 spread; (ii) investigate the possibility of providing 82 hygiene products from the inventory and the quantities required to accommodate all 83 evacuees; (iii) present example spatial designs for shelters as well as operational processes; 84 and, (iv) investigate implementation and governance issues, especially regarding liability 85 related to secondary infection and government attitudes. While the first aim above can be 86 achieved by applying the same method as was used in Kotani The remainder of this paper is structured as follows: Section 2 explains the background, target 105 area, and approach of the study; Section 3 provides the results of the case study; Section 4 106 discusses the applicability and implementability of the shelter plan; and, Section 5 107 summarizes the findings and discusses the potential of future studies. 108 109 2. Background, focus, and methods 110 2.1 Shelter space shortages before and after the COVID-19 pandemic 111

The shortage of designated evacuation shelters is not a problem that first emerged after the 112 outbreak of COVID-19, but one that Japanese society had already faced prior to 2019. In this study, we revisit the case study of the Taisho-Suji Shopping Street in the Nagata Ward  140 of Kobe, Japan (Figure 1), 1 which was proposed for use as an undesignated evacuation shelter  141 in Kotani et al. [9] . The shopping street was damaged by the 1995 Kobe earthquake (the Great 142

Hanshin-Awaji Earthquake). Before the earthquake, this street had the typical features of 143 downtown areas of the time, with its small space busy and crowded with people, goods for 144 sale, and so on (Figure 2a) . The shopping street suffered catastrophic damage during the 145 earthquake, and a large fire that occurred immediately after the earthquake caused further 146 extensive damage ( Figure 2b ); the narrow streets were densely packed with wooden houses, 147

preventing fire engines from entering the area. Approximately two months after the disaster, 148

to construct fire-and earthquake-proof buildings, Kobe proposed a recovery plan to 149

redevelop an area south of the Japan Railway Shin-Nagata Station [43] . In this project, the 150 street was transformed into a modern shopping arcade, with skyscrapers, high ceilings, and 151 wide streets ( shopping-street shelter. As mentioned above, the buildings currently on the street are fire-162 and earthquake-proof, the central passage is wide, and the arcade ceiling is high enough so 163 that fire engines can pass the street. Therefore, the buildings are suitable for use as 164 earthquake evacuation shelters. Further, this area is not subject to tsunamis or landslides, 165 and has been classified as inundation-free against floods on the scale of a one-hundred-year 166 return period. However, in 2020, updated flood simulation results indicated that the area 167 could be inundated by floodwaters of "less than 0.5 meters high" in the case of floods on the 168 scale of a one-thousand-year return period [45] . Since a flood's scale is predictable to some 169 extent before it reaches an area, in the case of huge-scale events such as those of the one-170

thousand-year return period, the floor plan is arranged so that the basement and first floor 171 are not used, while floors above the second floor will become more significant, owing to the 172 scarcity of tall buildings that can serve as evacuation shelters in the district. Considering such 173 needs for a flexible arrangement, in this study, we explore hazard scenarios except for a flood 174 of one-thousand-year-return-period scale, meaning that we include the basement and first 175 floor in shelter planning to ascertain the potential of the buildings to cover most disasters. 176

As previously mentioned, in Kotani The procedures to demonstrate these aspects are described below. 198 In the calculation of the total area of shared spaces, we exclude stairs, walls, and toilets, and 199 only focus on spaces where people could lie down; in calculating that of empty lots, we 200 assume that there is no furniture (e.g., shelving) there. Based on the area of available spaces, 201

we estimate the number of evacuees that can be accommodated in the buildings. In this 202 estimation, we assume that each evacuee would need 6 m 2 of space to avoid COVID-19 203 infection [33], instead of 3 m 2 /person, as is often used in non-pandemic situations. We also 204

illustrate what percentage of expected evacuees would be accommodated. 205

To estimate the quantities of food and drink supplied from the shopping street, we calculate 206 the average quantity of inventory from restaurants and shops, following the method and 207 inventory data in Kotani et al. [9] . The shops include convenience stores, supermarkets, 208 bakeries, and meat and fish shops. The calories (unit: kcal) and liter (unit: L) of supplied food 209 and drink, respectively, are calculated. Based on the recommended daily intake of calories 210 and water (2,000 kcal and 3 L per person for a day, respectively [9]), we estimate the number 211 of person-days that could be supplied with available food and drinks. In this estimation, as 212 the quantity of available food depends on the situation after the disaster, we consider the 213 following two cases: (1) the case of power and water outage (hereafter "the case without 214

utilities") and (2) the case of a substitute power and water supply (hereafter "the case with 215 substitute utilities"). 216

required to cover accommodated evacuees 218

The installation and use of hygiene products, as shown in Table A1 (b), are important in 219 evacuation shelters during a pandemic. Therefore, we examine the availability of these 220 products from the shopping street's inventory. We visit stores that are likely to sell these 221 products (e.g., supermarkets, convenience stores, and pharmacies) on the shopping street, 222 check whether they exist in the inventory, and determine the possibility of their being 223 supplied from the shopping street itself. 224

It is also recommended that citizens bring some hygiene products to the shelter themselves 225 (Appendix B); however, citizens may not do so, and thus, it is important to stockpile essential 226 hygiene supplies in shelters. We estimate the required quantities of four specific hygiene 227 supplies, which almost everyone needs-(1) masks, (2) alcohol-based sanitizer, (3) plastic 228 bags used for storing reusable materials, garbage etc. (i.e., "three types of polybags" in Table  229 A1(b)), and (4) wet wipes (i.e., "alcohol wipes for cleaning" in Table A1 (b))-to cover the 230 number of evacuees accommodated in the shopping street. As a benchmark, we assume that 231 each evacuee would wear one mask a day. Hand sanitizer requires at least 3 mL of liquid on 232 each use [46] . We assume that a person would use hand sanitizer 10 times a day, and thus 233 approximately 30 mL (3 mL X 10 times) would be used per person per day. We also assume 234 that a person would use one plastic bag for trash (20 L) and one pack of wet wipes (50 sheets) 235 a week. Evacuees would be expected to stay in the shopping street for one week. 236

In 2020, we could not estimate the average quantity of these hygiene products in the 237 inventory of each shop because the inventory had changed dramatically. For example, the 238 mask inventory decreased at the beginning of 2020, but later was overstocked, due to the 239 increased use of reusable masks and production expansion. Future studies should calculate 240 the exact numbers of hygiene products to be supplied from the shopping street; however, 241 clarifying the necessary numbers will contribute to the discussion on measures to increase 242 hygiene product stockpiles. to prevent the spread of infection. We begin by discussing the spatial arrangement by 249 obtaining floor maps of the building and information on the vacant stores and available 250 common spaces. [9], we estimate the total calories (unit: kcal) of food and the total quantity (unit: L) of drinks 278 supplied from the shops to be 2,404,820 kcal and 5,389,607 kcal of food in the case without 279

utilities and with substitute utilities, respectively, and 4,816 L of drinks regardless of the 280 utilities situation (the second row of Table 2 ). These supplies are equivalent to 1,202 and 281 2,695 person-days of food in the cases without utilities and with substitute utilities, 282 respectively, and 1,605 person-days of drinks (the last row of Table 2 ). If we were to provide 283 the available food and drinks to the number of evacuees who could be accommodated (i.e., 284 1,194 evacuees), these results indicate that we could provide meals on the first day of the 285 disaster, even in the case without utilities. The remaining food and drinks-8 person-days 286

(1,202 person-days -1,194 evacuees) of food and 411 (1,605 person-days -1,194 evacuees) 287

of drink, could be distributed to people who stayed in their homes or were in neighboring 288

shelters. If the food were cooked using power generators, gas cylinders, and well water, we 289 would be able to cover two days' worth of meals for the evacuees accommodated on the 290 shopping street. 291 

to cover accommodated evacuees 295

We visited stores that were likely to sell hygiene products (i.e., two supermarkets, one 296 convenience store, one drug store, and one electronics retail store) and checked their 297

inventory. As shown in Table 3 , we determine that most of the hygiene products, such as 298 masks, alcohol-based sanitizers, and thermometers, can be supplied from the shopping 299 street's inventory (In the table, ◎ denotes at least one store has the inventory; △ denotes 300 that the store visited do not stock them in their inventory.). 301

Following this, we estimate the required number of four specific hygiene supplies-(1) masks, 302

(2) alcohol-based sanitizer, (3) plastic bags, and (4) wet wipes-to cover a one-week stay for 303 evacuees accommodated in the shopping street (i.e., 1,194 evacuees), based on the 304 assumptions described in Section 2. It is found that 8,358 masks (1,194 people X 1 mask/day 305 X 7 days), 35,820 mL of hand sanitizer (1,194 people X 30 mL/1 day X 7 days), 1,194 plastic 306 bags, and 1,194 packs of wet wipes would be required (Table 4 ). These quantities of hygiene 307

products should be supplied from the shopping street's inventory and/or stockpiled as relief 308 supplies in the shopping street. 309 Table 3 . Hygiene products likely to be supplied from the shopping street's inventory 310

Possibility of being provided from the inventory Masks Type HN evacuees, with the letter "N" (not vulnerable) added. 338

The shelter space should be designed such that it meets three conditions. First, the entrance 339 needs to include a space to conduct health checks and determine each evacuee's type. gray areas indicate unavailable spaces, which are occupied by stores and used for purposes 362 other than disasters, and are therefore not expected to be used as shelter spaces during a 363 disaster. 364 Figure 5 illustrates the detailed procedures for classifying evacuees who come to the building. 366

At the reception, evacuees who are already known to be infected with COVID-19 are sent to 367 "the desk for evacuees with a confirmed infection" and classified as Type I. After checking 368 their support, evacuees would be identified as either Type IV or Type IN. 369

Evacuees who have been in close contact with people infected with COVID-19 would be sent 370

to "the desk for evacuees with close contact with the infected" and categorized as Type CS if 371 they report symptoms during their health status check, and Type C otherwise. Next, 372

depending on their support needs, Type CS evacuees would be categorized as either Type CSV 373 or Type CSN. The same procedure would be applied to Type C evacuees, who would then be 374 classified as either Type CV or Type CN. 375

Evacuees who have neither tested positive for COVID-19 nor made close contact with people  376 infected with COVID-19, but perceive some related symptoms, would be sent to "the desk for 377 evacuees with noticeable symptoms" and categorized as Type S. Then, their support needs 378 would be checked, and they would be classified as either Type SV or Type SN. 379

Evacuees who do not fit any of the above categories would be sent to the "desk for evacuees 380 without noticeable symptoms." If some symptoms are detected in the health status check, 381 those evacuees would be categorized as Type S; otherwise, they would be categorized as Type 382

H. Type H evacuees would then be classified as either Type HV or Type HN, depending on their 383 support needs. 384

After evacuees are classified as one of the above types, they would be requested to follow a 386 specific traffic line to access a specific space (Figure 4 ). The details of the assigned available 387 spaces on each floor and traffic lines with access to the spaces for each type are shown below. 388 Basement (Figure 4a ) 389

To reduce their contact with other evacuee groups as much as possible, we decided that only 390

Group I would use the basement. Group I would follow the purple route to go to the basement, 391

using only Gate 1 to avoid contact with other groups. Type IV evacuees would be prioritized 392 to use Elevator 1 (E1) to move to the basement. They would be allocated to one of the 393 partitioned Available Spaces 1, 2, and 3 (AS-B1, 2, and 3), which can accommodate six, two, 394 and seven people, respectively. Type IN evacuees would be asked to use Stairs 1 (S1) to reach 395

Available Space 4 (AS-B4), which can accommodate 264 people. Toilet BF1 (T-BF1) would be 396 used by Group I only. 397

In this example, we expect that the stay of Group I would be short, since they should be 398 moved to other appropriate facilities as soon as possible to receive medical treatment. Thus, 399

we allocated Group I to the basement, where they cannot see the outside and are likely to 400 get stressed to some degree. Thus, if significant space is available on the upper floors and 401

Group I could be kept separate from the other groups, it would be desirable to allocate them 402 there. 403 can also conceive of a case in which there are no Type IV or IN evacuees at the shelter because 470 they have already moved to hospitals due to their symptom severity. In such a case, Type CSV 471 and CSN evacuees with severe symptoms would be given the basement spaces with the 472 expectation of moving to a hospital as and when the situation allows. However, if hospitals 473 are busy and they need to spend several days in the shelter, they may find a small room on 474 the second floor to be more comfortable, as the sky would be visible. Another possible 475 arrangement would be one in which a small room on the second floor is allotted to one family 476 that includes Type CSV and CSN members with severe symptoms. 477

With some expectations on the proportions of the types based on the current state of 478 infection in society, the shelter can prepare for multiple situations to adapt to different 479 circumstances in advance. Thus, the shopping street shelter plan contributes to society by 480

increasing not only the total space for accommodating evacuees but also the number of 481 isolated rooms and the flexibility of spatial arrangement. 482

In Section 3.3, we described a shelter-space design method in which Group I-CS-C-S and Group 484 H are kept separate, with the aim of preventing the spread of infection among residents. 485

Moreover, unlike shelters with continuous large spaces such as school gymnasiums, which 486

are commonly used as designated shelters, the shopping street's arcade buildings have high 487 potential and enable a shelter design that incorporates the need for group divisions. evacuation centers could be attributed to the shelter manager, that possibility would be a 492 major obstacle to implementing this plan. Since this would be a non-designated shelter, the 493 plan would be implemented within the framework of community-based activities. The 494 "shelter manager" or "operator" is a committee of community members, possibly including 495 managers of the arcade buildings and shopkeepers from the shopping street. If the committee 496 is afraid of being sued after the spread of the infection, they are likely to discontinue the plan. 497

In this subsection, we will consider the issue of responsibility of the committee, that is, the 498 shelter manager. 499

Cases in which a COVID-19 carrier transfers the virus to a non-carrier are classified into the 500 following three types: 501

Case A: A case in which a Type IV, IN, CSV, CSN, CV, CN, SV or SN evacuee makes a false 502 declaration and is settled in a space meant for a Type HV or HN evacuee. 503

Case B: A case in which an evacuee does not follow the rules of the designated traffic lines 504 and communal facilities. 505

Case C: A case in which an evacuee who was identified as a Type HV or HN evacuee by not 506

having symptoms related to COVID-19 is actually a virus carrier. 507

There are no measures the shelter manager can take against Cases A and Case B. Accordingly, 508 the shelter manager is not liable for secondary infections caused by either case. 509

Likewise, even with Case C, there is no problem with the manager classifying such an evacuee 510

as Type HV or HN at reception, as the evacuee was asymptomatic at the time. In addition, if 511 symptoms such as fever appear later, and the manager moves and isolates the evacuee 512 promptly and not unattended, even if secondary infection occurs, the manager will not be 513 liable under the Civil Code of Japan. Article 698 (emergency administrative management) 514

states, "If a Manager engages in the Management of Business in order to allow a principal to 515 escape imminent danger to the principal's person, reputation or property, the Manager shall 516 not be liable to compensate for damages resulting from the same unless he/she has acted in 517 bad faith or with gross negligence" [50]. Thus, providing shelter to protect community 518 members from the risk of death or injury in the event of a disaster is an act of rescue from 519 urgent harm, namely, an act of first aid, and is considered to fall under the category of 520

"emergency administrative management" [51]. Therefore, managers will not be held civilly 521 liable unless they have caused damage to another person due to malicious intent or gross 522 negligence. Hence, as previously mentioned, as long as the manager checks the health status 523 of the evacuees and takes appropriate measures, such as the settlement and traffic divisions, 524

which consider the buildings' structural constraints, even if secondary infection occurs, a 525 lawsuit could not be filed and the manager would not be held liable. 526

Article 698 (emergency administrative management) of the Civil Code is considered to be a 527 provision for reducing liability with the intent to encourage people to provide first aid in an 528 emergency [52] . Therefore, its content is considered to support the practice of this shelter 529

plan. As such, potential managers should avoid hesitating to open non-designated shelters 530 due to concerns over possible liability. In Japan, based on the fact that Article 698 of the Civil 531

Code supports the opening of evacuation shelters, this provision should be understood and 532

shared by all community members when they are formulating their local disaster 533 management plans. 534

As previously stated, there has been a shortage of available evacuation shelters during a large-536 scale disaster, even without an ongoing pandemic, and non-designated evacuation shelter 537 use is increasing. If the required area per evacuee is doubled during a pandemic, then the 538 capacity of designated evacuation centers will be further reduced. Central and local 539 governments have already recommended distributed evacuation, and the need for non-540 designated evacuation centers is increasing. Therefore, non-designated evacuation centers 541 that are community-operated and have the above-mentioned capacity, such as the shopping 542 street evacuation shelter described in this study, can assist government-operated designated 543

shelters by reducing their population density. 544

In addition, shopping street shelters would also be valuable in that they would be pre-planned 545 by the community. Conventionally, most undesignated shelters have been cars in parking 546 areas for overnight stays or relatives' homes, where it is difficult for local governments to 547 estimate in advance the number of evacuees who will use non-designated evacuation shelters. 548

Furthermore, presumably some evacuees who ultimately stayed in their cars were refused 549 entry to a designated shelter because it had reached its capacity for accommodating evacuees. 550

In that case, in addition to the time loss and stress that evacuees experience during the 551 evacuation process, the population density of the designated shelter's reception space will 552

temporarily increase amid such confusion. In that respect, the shopping street shelter, whose 553 plan is shared by the community and government in advance, is superior to designated 554 shelters. 555

It is possible, however, that community residents, including those operating evacuation 556 shelters, will not be able to dispel a certain level of anxiety until it is demonstrated several 557 times that the shopping street non-designated evacuation shelter can actually perform well. 558

It is also likely that there will be differences in enthusiasm for the implementation of the plan 559 among shopkeepers who are responsible for providing food and drinks. Given such a situation, 560

even if it is positioned as an "undesignated evacuation shelter," it can benefit the local 561 community; thus, the local government should publicly express support for such a plan or 562 take action to guarantee the plan's validity. This will greatly increase the local population's 563 confidence in and motivation to use non-designated shelters. 564

Moreover, local governments must be motivated to provide disaster relief supplies to the 565 shopping street shelter in advance. Having these disaster relief supplies stored in the empty 566 spaces of shopping street buildings would also benefit the government. Such supplies would 567 include ones that do not normally exist as commodities in the shopping district, such as 568 blankets, partitions, and non-contact thermometers, and would also address a shortfall of 569 masks during a disaster. Without shopping street shelters, such supplies would only be 570 located and used in designated shelters and kept under administrative control. Relocating 571 them to shopping street buildings in advance will strengthen the policy of distributed 572 evacuation. Section 4. such proactive steps taken beforehand, they need to be transferred to a medical institution 605 as quickly as possible. In this respect, equivalent measures must be taken between designated 606

and non-designated evacuation centers. It is important for the government to take a uniform 607 approach toward isolation within evacuation shelters and in the process of moving evacuees 608

to medical institutions, assuming that many evacuation shelters do not have medical 609 professionals available. Furthermore, in 2020, the Japanese government published a manual 610 for operating shelters during the COVID-19 pandemic. This manual also points out the 611 importance of providing training to increase the number of individuals available to participate 612 in public health activities at evacuation shelters. 613

In this study, we have investigated the case of a shopping street in the Nagata Ward of Kobe, 614

Japan. In addition to using data from this area to calculate the available area for evacuation 615 and available quantity of food and drinks, the previous subsection developed a consideration 616 in accordance with the Civil Code of Japan. In this way, the topic was thoroughly explored in 617 line with the peculiarities and specificities of the targeted district. The results obtained from 618

such research procedures will raise questions regarding the finding's generalizability and their 619 applicability to other countries and regions. Some of the answers to such questions have 620 already been provided in our previous study, Kotani et al. [9] , which pointed out that shopping 621 streets in many countries and regions have the potential to become "shelters with food 622 stocks,"; shopkeepers often play a central role in the community's resident network and 623 possess the foundations to become members of a shelter management group. However, 624

environmental conditions, such as the location of a shopping street and the size and strength 625 of the buildings, vary depending on the location. Additionally, laws, institutions, and norms 626 regarding the concept of emergency administrative management and its liability-that is, the 627 issue of liability for secondary infections of infectious diseases, which were dealt with in this 628 study-vary by country. In that regard, it is beyond the scope of this study to offer an 629 international institutional comparison. Nonetheless, the "generality" of this study, in terms 630 of its contribution to research and practice regarding the utilization of shopping street 631 shelters in other countries is characterized by another dimension that is one level more 632

general. This study presented viewpoints, checkpoints, methods, and thought processes 633 when creating and characterizing evacuation shelters. Confirming the spirit implied by the 634 laws of each society is one of them. Including the above, a way that this study considers the 635 relationship between the individuality and generality of the research findings is in line with 636 many area studies. 637 638

In consideration of the increasing occurrence of large-scale disasters, the shortage of 640 designated evacuation shelters that had already been observed, and the seismic-resistant 641 buildings in shopping districts, we proposed in Kotani et al.

[9] that shopping street buildings 642 be utilized as disaster evacuation shelters that are perpetually stocked with food and drinks 643 as commodities and inventory. This time, after the outbreak of COVID-19 and after emergent 644 responses began to be provided for the purpose of preventing the spread of infection in 2020, 645

we quickly updated our idea of a shopping street evacuation shelter with conviction that this 646 type of shelter has enhanced comparative superiority against other, more typical shelters 647 during a pandemic. 648

Using the latest data from the Taisho-Suji Shopping Street in Kobe, Japan, we estimated the 649 capacity of the shopping street shelter; it can accommodate 2,388 evacuees under the non-650 pandemic condition, and 1,194 evacuees-which is equivalent to 23 % of the local 651

population-under the pandemic condition, where 6 m 2 of space is required for one person 652

to maintain enough distance from others. Moreover, we updated our estimation of supplies, 653

finding that the shopping street shelter could provide food for 1,202 person-days in the case 654 of lifeline disruption, for 2,695 person-days in the case of substitute utilities being available, 655

and drinks for 1,605 person-days. Additionally, we provided a list of necessary hygiene 656 supplies and the quantities required for 1,194 evacuees to stay in a shelter for a week. 657

The novelty of this study exists in how non-homogeneous and separate spaces in the buildings 658 are applied to prevent the spread of infection. Evacuees, who are classified by properties such 659 as being infected, having had close contact with infected individuals, being symptomatic, or 660 being vulnerable for reasons other than COVID-19 are separated, so that they do not share 661 space, toilets, or traffic lines. We exemplified the shopping street shelter design, 662 demonstrated its flexibility, and verified its physical competency. In addition, we investigated 663 the liability issue of secondary infection at a shelter and clarified that the Civic Law of Japan 664 supports community implementation of a shelter plan. Furthermore, we discussed the role of 665 the government, such as providing more convenient PCR test and increased beds in hospitals 666 and quarantine facilities, which can better promote its policy of distributed evacuation by 667 taking several actions in support of community-based shelter management. 668

While it is urged amid the ongoing pandemic, as of March 2021, that the shelter plan moves 669

to the practice stage, important issues to be addressed in future research remain. The 670 advantages and disadvantages of the arcade buildings of the targeted shopping street against 671 other big shopping malls should be examined. Their differences will include implementability, 672

where the local and small areas covered by the shopping street at the community scale may 673 manifest some strengths. For example, people can access it on foot and, moreover, evacuees 674 may be able to share some feelings of solidarity and concerns about the damage to their 675 residences based on the fact that they all live in the same district. Other topics that should be 676 addressed include: (1) We are deeply grateful to Shin Nagata Town Management Company Ltd. and Mr. Masayuki 687

Shishida, its former president and currently the auditor of Kobe Nagata TMO Company Ltd., 688

for comments and support. 689

Funding 690

This work was partly supported by JSPS KAKENHI (Grant No. 20H02272). 691 692 may be covered by the supplies stocked at the shelter. However, this would not be the case 731 for undesignated shelters. Evacuees who use undesignated shelters have to assume that, in 732 most cases, they need to procure both basic relief supplies and hygiene supplies. 733 734 

Several international and governmental organizations published guidelines in 2020 regarding 695 shelter management. WHO [1] suggests that evacuees maintain a physical distance

Depending on their health status, 697 evacuees would be classified into three groups: (i) evacuees without COVID-19 symptoms, (ii) 698 evacuees with COVID-19 symptoms, and (iii) evacuees confirmed to have COVID-19. In 699 addition, vulnerable people and high-risk groups should be separated from other groups and 700 given their own space. Each group is not allowed to interact with other groups. Further, 701 evacuees must use their specific assigned route in evacuation shelters. The groups should also 702 use separate public facilities

Three Cs" (i.e., closed spaces, crowded 705 spaces, close-contact settings) in evacuation shelters, evacuees and staff should use personal 706 protective equipment (PPE)

At the entrance of a shelter, operators should check evacuees' 710 health status and classify them into (i) infected evacuees, (ii) close contact with the infected, 711 (iii) evacuees who have symptoms, and (iv) general evacuees, including determining who is 712 vulnerable and in need of support. The concepts of separation of the settlement areas, public 713 facilities, and traffic lines, and of cleaning and ventilation, are almost the same as those 714 suggested by WHO [1]. It is emphasized that alcohol-based sanitizers should be located in 715 public spaces, such as at reception, the entrances to each room/staircase/toilet, restaurants, 716 and around places for trash. In addition, each evacuation shelter should prepare 717 approximately 25 hygiene items for evacuees (Table A1(b)). All evacuees and staff should also 718 use PPE

Necessary relief supplies during a disaster and pandemic

to prepare for natural disasters during COVID-19, the Japanese government 724 published a list of necessary emergency hygiene supplies that evacuation shelters are 725 expected to keep on hand, as shown in Table A1(b) [54]. Nevertheless, the government also 726 recognized that it is not easy for these supplies to be kept sufficiently stocked in shelters, and 727 thus requested citizens take some of their own simple hygiene supplies, such as masks, 728 alcohol-based sanitizer, and wet wipes when evacuating to shelters. Accordingly

Manila: WHO Regional Office for the Western Pacific

Evacuation shelter is not enough

Only 20 % can enter an evacuation shelter? Getting serious about a lack of 745 evacuation shelter

Cabinet Office, Regarding the further response to COVID-19 in evacuation shelters

Distributed evacuation-Emerged issued-successive fully booked 751 hotels

Patterns of emergency shelters in coastal plains a case 755 study after the great east Japan Earthquake and Tsunami in Higashi-Matsushima City

Recommendation related to evacuation, Japan Society for Disaster 758 Information Studies

How to deal with COVID-19 and emergency evacuation from now on

Potential of a shopping street to serve as a food 765 distribution center and an evacuation shelter during disasters: Case study of Kobe

University campus for a base of shelter for urban disaster prevention 769 preparedness: Note of promising project and coming plan in near future at Fukuoka 770 University

Rethinking the public building as post disaster shelters-in 772 the context of Old Dhaka

The role of regional public libraries and assistance system 775 for them in catastrophic disaster

Role of the Masjid in Disaster 778 management: Preliminary investigation of evidences from

Unnoticed but important: revealing 781 the hidden contribution of community-based religious institution of the mosque in 782 disasters

Consideration on functions of temples for 785 evacuation sites against tsunami attack

Designation of Buddhist temples as emergency place and/or 788 evacuation shelters and actual usage during widespread disasters

Current state and future issues of the development of roadside 791 stations as disaster prevention function-a case study of roadside stations in Fukushima 792 Prefecture

A study on the function of roadside rest area "Michinoeki" in disaster relief 796 operation, 6, The Bulletin of Graduate Engineering and Design Studies of Hosei 797 University

Utilization of cultural heritage buildings as evacuation spaces near Sendai after 801 the Great East Japan Earthquake

Potential of mosques to serve as evacuation 804 shelters for foreign Muslims during disasters: a case study in Gunma

A survey on temporary shelters for stranded victims 807 during disasters in commercial/business districts: Focus on Minato Mirai 21

Evaluation of the location 811 potential of commercial facilities as a place of refuge: A case study on facilities 812 providing food in the Kii Peninsula

Disaster response plans by the convenience stores and their 815 regional contribution map of Kawasaki City

Environmental transformation after disaster and 818 transfiguration process of practice in a community

A role of Ennichi and Jizobon in expanding interaction in a 822 local community -results of a social network survey in Nagata Ward

Quantitative evaluation of the roles of community events 827 and artifacts for social network formation: a multilayer network model of a community 828 of practice

Widespread outbreak of 832 norovirus gastroenteritis among evacuees of Hurricane Katrina residing in a large 833 "Megashelter

Infectious diseases 836 following natural disasters prevention and control measures

Shelter 839 crowding and increased incidence of acute respiratory infection in evacuees following 840 the Great Eastern Japan Earthquake and tsunami

Infection control after and during natural disaster

Outbreak of norovirus illness among wildfire 848 evacuation shelter populations-Butte and Glenn Counties

Managing disasters amid COVID-852 19 pandemic: Approaches of response to flood disasters

Implementation of evacuation 855 measures during natural disasters under conditions of the novel coronavirus (COVID-19) 856 pandemic based on a review of previous responses to complex disasters in Japan

Temporary sheltering, psychological 859 stress symptoms, and perceptions of recovery

World Health Organization, Disaster evacuation shelters in the context of COVID-19

Best practice for evacuation 865 centres accommodating vulnerable populations: a literature review

Shelters crammed with evacuees-30 people spent the night in cars, 868 enduring cold weather in Otsuchi Iwate 21 Mar

Cabinet Office, A report of the Kumamoto Earthquake

Administrative response to the 2016 Kumamoto earthquake: Lessons-875 learned from a failure

Regarding facilitation of evacuation in the future typhoon based on 879 typhoon No

Spread infection? Evacuation to hotels-at the scene of disaster issued the 882 highest level of warning during the Typhoon No

Typhoon No. 10-133 evacuation shelters were full, exceeded 886 expectations-verification required

Outline of the urban redevelopment projects after disasters in the South of 890 the Shin-Nagata Station

National 894 census: A result of population and other basic sets

Hazard map: Local response information to storm and flood

Hand coverage by alcohol-based handrub varies: 902 Volume and hand size matter

Cabinet Office, Reference for the response to COVID-19 in evacuation shelters

COVID-19 Infection: A helpful support book for life in evacuation

Survey report on the 2018 actual condition of 912 shopping streets: Overview

Article 698 Emergency administrative management

New annotated edition-The Civil Code

Catastrophes and the law of compensation for damages, Special issue on 919 earthquake disaster and civil law

Office administration

Article 697-711

Inter-ward, risk-diversified 924 allocation model for storing of disaster relief foods

Cabinet Office, Guideline for evacuation shelter establishment and operation training 927 considering COVID-19 response

694