Significance of Secondary Infrastructure for Shelter Management in the Aftermath of Hurricane Michael

Juyeong Choi
Florida A&M University

Fehintola Sanusi
Florida A&M University

Mehmet Baran Ulak
Stony Brook University

Eren Erman Ozguven
Florida A&M University

Tarek Abichou
Florida A&M University

Publication Date: 2019


Diverse infrastructure sectors were engaged during Hurricane Michael to adequately support the operation of shelters and accommodate evacuee needs, including those who are vulnerable (e.g., seniors, children, and people with special needs) Supporting infrastructure consists of primary and secondary infrastructure. Specifically, primary infrastructure comprises service providers (e.g., utility services and local emergency agents) that are expected to serve shelters in a normal situation, while secondary infrastructure (e.g., nongovernmental organizations and local businesses) becomes available only when primary infrastructure is overwhelmed. Secondary infrastructure is significant for the operation of hurricane shelters, especially for prolonged recovery; however, information about availability is not readily available in the planning stage due to the ephemeral and ad hoc nature of resources typically associated with secondary infrastructure. Using focus groups of emergency agencies and documents (e.g., daily situation reports) as the major data collection method, this study describes secondary and primary infrastructure available for long-term shelter operations in Bay County, Florida, in the aftermath of Hurricane Michael. Through description and evaluation this research illuminates the emerging relationship between shelters and primary and secondary infrastructure during hurricane recovery.


Hurricane Michael was the strongest hurricane to land directly on the continental United States in the last two decades, devastating thousands of roadways and houses, and causing substantial infrastructure disruptions in the Florida Panhandle (Sellers, Begos, and Zezima, 20181). As a consequence of the hurricane, around 375,000 people in Florida were ordered to evacuate, resulting in emergency shelters being overwhelmed by the demand (Farrington and Lush, 20182). Once Hurricane Michael departed, emergency agencies considered closing their shelters and sending any remaining evacuees to county shelters. Many people were still unable to return to their homes for a variety of reasons such as utility services disruptions, roadway inaccessibility, and storm-related debris in their communities (Associated Press, 2018 3). In particular, the hurricane destroyed many homes in Mexico Beach and Panama City, Florida (Mahoney and Mower, 2018 4), thereby increasing the number of evacuees seeking long-term shelter. Long-term shelters were challenged to maintain acceptable operating levels and serve all segments of the remaining population, including those who are vulnerable (e.g., seniors, children, and people with special needs). However, managing such vital facilities is difficult for various reasons, including: diversity of evacuee needs, utility outages, and ongoing issues with hurricane-related debris.

During the planning stage, shelters and other infrastructure agents can prepare for such needs through capacity building of infrastructure (here called primary infrastructure). In addition, some service providers (here called secondary infrastructure) will emerge in order to supplement the functions of the primary infrastructure during post-disaster recovery in situations where primary infrastructure is overwhelmed. For instance, if a shelter exhausts its municipal water and stored bottled water (i.e., primary infrastructure), nongovernmental organizations (NGOs) and local businesses (i.e., secondary infrastructure) will often appear to provide the shelter with extra water (e.g., in the form of bottled water or a water truck). Despite the importance of secondary infrastructure for long-term shelter plans, especially in the aftermath of major hurricanes, the focus of existing data collection is typically limited to primary infrastructure. Since auxiliary support from secondary infrastructure is localized and emerges only during post-disaster recovery, information about its availability is ephemeral and is rarely made available to a disaster-prone community.

This report first discusses the diversity of potential shelter needs in Bay County and then details the research methodology and results. At the end, this study discusses what triggered secondary infrastructure to emerge in support of primary infrastructure and thus shelters and post-hurricane communities in Bay County. This understanding will help guide the engagement of secondary infrastructure in the planning stage.

Research Questions

  1. What secondary infrastructure allows a shelter to stretch its capacity when its primary infrastructure has reached its limits?
  2. How do the relationships between shelters and their primary and secondary infrastructure emerge and evolve during response and recovery in order to fix their operational problems, such as supply and debris management?
  3. How can we leverage any information collected to promote integrated capacity building planning approaches, thus yielding more resilient shelters in the future?

This research aims to collect information about the emergent relationships between shelters and their primary and secondary infrastructure, making this data available for hurricane shelter planning and management. This research focuses on available secondary and primary infrastructure for the operation of long-term shelters in Bay County, Florida.


Diversity of Potential Shelter Needs in Bay County, Florida

It is known that demographic features (e.g., age, race, health, income, and employment) affect the vulnerability of people to natural disasters (Cutter, 20035; Holand, 20146). For example, people with low incomes are more likely to live in areas where housing is affordable, but with less reliable access to utility services, while people with high incomes are more likely to have greater access to resources. Such inequalities in physical location make low-income people more vulnerable, and they are likewise more likely to be disadvantaged in terms of evacuation and sheltering (Ozguven et al., 20167). For example, aging populations may be vulnerable to natural disasters and need special attention due to physical limitations, limited sensory awareness, health conditions, and social and economic constraints (Ozguven et al., 2016). According to the U.S. Census Bureau, in 2017, the median household income for Bay County was $50,283 (in 2017 dollars), which was lower than the state average of $50,883 (in 2017 dollars; Figure 1a), and the number of people ages 65 years or older accounted for 16.5% of the entire population (i.e., the aging population [+65] was 30,838; Figure 1b). Based on Figure 1b, it is important to note that a relatively high number of older people live along the coast (i.e., in evacuation Zone A or B). This means the aging population most likely would be evacuated in the case of a hurricane. In addition, Figure 1c shows that people living along the coast may have less route options available during emergency due to an insufficient number of major roads. Figures 1a and 1c indicate that some populations with low median family incomes (ranging from $35,000 to $45,000) have limited access to the downtown area of Panama City (due to the limited number of major roadways) where recovery efforts and resources are likely to be concentrated in the aftermath of a hurricane. Considering the social vulnerability of people living in the area, it is perhaps not surprising that Bay County had a large number of evacuees needing to be accommodated in shelters as a result of Hurricane Michael.

Fig. 1. Demographic information and infrastructure facilities in the Bay County, Florida: (a) median family income, (b) aging population (65+) and evacuation zones and (c) major roadways (click to image enlarge).

Importance of Primary and Secondary Infrastructure for Successful Operation of Hurricane Shelters

In this research, primary infrastructure indicates any entity that is committed to providing services essential for communities to sustain their economic and social activities in a normal situation. Examples of primary infrastructure include municipal utility suppliers, hospitals, and government agencies. They are expected to provide utilities, health-care services, and public services, respectively, to communities in normal circumstances. Meanwhile, secondary infrastructure consists of entities that are not expected to serve communities in a normal situation, but may emerge to support them when the primary infrastructure is not sufficient. Examples include NGOs, local businesses, and other volunteer groups. According to the National Disaster Recovery Framework proposed by the U.S. Department of Homeland Security (2016), a local government is responsible for preparing for disasters and managing the response and recovery of its community. Therefore, this study considers primary infrastructure to be county-owned resources (e.g., the Bay County Emergency Operations Center [EOC] and municipal utility services) that are available to hurricane shelters for the operation and management of their facilities and for the overall recovery of affected communities. Secondary infrastructure comprises external supplies or supports that are generally ad hoc and not committed to community recovery. In the case of major hurricane events, the overall recovery process of affected communities is important for long-term shelter operations.

The recovery and restoration process for Bay County took a while due to the devastating impact of Hurricane Michael (Figure 2). Therefore, many evacuees had extended stays in hurricane shelters because it was not safe enough for them to return to their houses. Local emergency management agencies are responsible for providing secure accommodations for evacuees. Since hurricanes destroy houses, buildings, and infrastructure and generate an enormous volume of debris (Figure 2), affected communities should be recovered to some extent before evacuees return so that once they arrive they will have secure access to utility services and housing (Washington State Military Department, 20128). In this sense, the long-term operation of hurricane shelters is directly related to the overall recovery process of affected communities. The last shelter for Hurricane Michael closed on November 30, 2018 (McDonald, 20189), which was about 50 days after the first hurricane shelters opened. To understand the significance of primary and secondary infrastructure in the operation of hurricane shelters, we will also focus on their relationship with respect to the overall recovery process of Bay County.

Fig. 2. Impacts of Hurricane Michael. Image Credit: Eren Erman Ozguven and Mehmet Baran Ulak, 2019.


Answering the aforementioned questions requires (i) identifying interdependent infrastructure associated with the operation of hurricane shelters, (ii) categorizing infrastructure as either primary or secondary infrastructure, (iii) tracking the emergence of secondary infrastructure in relation to its relevant primary infrastructure throughout recovery, and (iv) analyzing the triggers of changes in the relationships between shelters and their supporting infrastructure. This study uses an interdisciplinary platform proposed by Choi, Deshmukh, and Hastak (2019) to identify critical infrastructure across multiple domains of systems important for community recovery and to examine interrelationships between these systems. In addition, two interviews were conducted with representatives of county emergency management agencies and state management agencies. Temporary documents (i.e., situation reports and incident action reports) were also collected and analyzed. Furthermore, qualitative coding was used to selectively extract any information relevant to the relationships between shelters and their primary and secondary infrastructure.

Interdisciplinary Platform to Capture Infrastructure Interdependencies

The purpose of the proposed interdisciplinary framework is to understand the relationships between primary and secondary infrastructure with regards to the operation of shelters and their operational contexts. Although such interplays are often necessary for the successful operation of critical facilities during post-disaster recovery (Deshmukh and Hastak, 201410), it is difficult to capture relevant information due to the breadth and complexity of such multidimensional systems (Flentge and Beyer, 200711; Ouyang and Dueñas-Osorio, 201412) To tackle this, we will use the framework of seven infrastructure layers proposed by Choi, Deshmukh, and Hastak (2019). According to this platform, complex interrelationships between multiple domains of critical infrastructure systems during post-disaster recovery can be modelled as seven interrelated layers of critical infrastructure: the civil, civic, social, environmental, educational, financial, and cyber layers.

In this research, we identified primary or secondary infrastructure in each layer according to its contribution to the post-disaster activities of long-term shelters. Using the seven-layer framework, the operational issues of shelters can be attributed to their key activities and associated infrastructure that lack adequate capacity to support the activities. Focusing on these operational issues and their relevant primary infrastructure facilitate capturing the functions of secondary infrastructure to overcome such challenges. For instance, when shelters lack enough water (i.e., an operational issue) due to the disruption of municipal water supply service (i.e., primary infrastructure in the civil layer), governmental groups such as county agencies (i.e., secondary infrastructure in the civic layer) and NGOs (i.e., secondary infrastructure in the social layer) may provide shelters and extra water resources (e.g., bottled water and water trucks) to compensate for the decline in water supply. Following this process of initial identification, relationships within the multilayer infrastructure are also categorized as either primary or secondary infrastructure based on the focus groups.

Focus Groups

Two focus group interviews were conducted for this study: one with the representatives of the Bay County EOC and one with the representatives of the Florida Division of Emergency Management (FDEM). The objectives of the interviews were to (i) identify the problematic activities in operating shelters and their associated primary and secondary infrastructure, (ii) measure the severity of such problematic activities on the operation and management of shelters and the overall recovery process, and (iii) identify solutions to overcome any operational and managerial issues related to facility management, operations, logistics, and debris management.

The first focus group interview took about 1.5 hours. This group comprised the representatives of the county departments related to emergency management such as the Division of Emergency Medical Services and the Department of Emergency Services. The focus group of the Bay County EOC officials was arranged to understand the problems shelters face in accommodating evacuee needs and the role of the Bay County EOC as the primary infrastructure supporting them.

The second focus group interview was conducted with the state-level emergency management personnel from FDEM and Florida State Emergency Response Team, who were directly involved in supporting the emergency responses of hurricane-affected counties, including Bay County. The respondents included 11 individuals/participants from the Bureau of Mitigation, the Infrastructure Branch, or Human Services (which provides statewide services such as mass care delivery, social services, and sheltering for hurricane-affected counties) and the emergency coordinating officers. This interview focused on understanding the challenges the state emergency agencies faced as the secondary infrastructure in supporting emergency operations at a county level. Thus, the questions were related to (i) the general roles of FDEM in supporting the operation of county shelters, (ii) the challenges FDEM had in supporting local government agencies during recovery, and (iii) how FDEM dealt with such challenges. All the questions were open-ended and representative(s) from each department provided responses based on their expertise and operational roles during the recovery process.

Use of temporary documents

The team also collected and analyzed situation reports and incident action plans from the Bay County EOC. Situation reports and incident action plans were prepared by county agencies two times a day and on daily basis, respectively. The incident action plans contain information regarding the actions, strategies, and objectives of the emergency managers during the emergency. Similarly, the situational reports document the day-to-day activities of overall shelter operations and the community recovery process in Bay County.

The situation reports were the primary focus for data collection because they were segmented into different sections that contain specific information about activities and operations related to different emergency support functions (ESFs). For instance, the reports provide broad information on the current status or progression of the ESFs (i.e., including ESF1 Transportation, ESF2 Communications, ESF3 Public Works, ESF4 Firefighting, ESF5 Planning, ESF6 Mass Care, etc.) and the priorities and operational objectives of the EOC. ESFs are the most critical functions for community recovery, and government support is granted under these groups. Each document reports on the county emergency agents’ corrective actions two times per day with a 12-hour interval, which yields the timeline of primary and secondary infrastructure for analysis. Other information documented includes the number of casualties; the jurisdictions that have been affected by the hurricane; the impacts of the disaster on individuals, schools, food supply systems, hospitals, and businesses; the current operational condition of utilities (e.g., water, electricity, communication, etc.) within the county, the status of the roadway infrastructure, the volunteer activities, the shelter operations, and the available human and material resources. This report also has information on what other external government agencies (i.e., state-level emergency managers such as FDEM and federal organizations such as the Federal Emergency Management Agency [FEMA]) were available to support community recovery.

Data analysis

Following the seven-layer framework, the information obtained from the focus groups was used to identify (i) issues related to the operation of hurricane shelters, (ii) the primary infrastructure that lacks sufficient capacity to support these functions, and (iii) secondary infrastructure that supplements the functions of the primary infrastructure in the context of shelter operations. Furthermore, auxiliary infrastructure used to support primary infrastructure operations were also identified. The data collected from the focus groups were primarily used to understand the operational situation of hurricane shelters and validate any findings from the result of the analysis of the ephemeral documents.

The ephemeral documents (i.e., situation reports) were then coded using a qualitative coding program known as Qualitative Data Analysis (QDA) Miner. QDA Miner can be used to code textual and photographic data and is also appropriate for mixed-methods data analyses (LaPan, 201313). All the supplemental documents were imported into QDA Miner and were then assigned particular combinations of codes to select text segments within the files (Provalis, 201614). The selected text segments were determined based on (i) the potential problematic key functions (e.g., ESF2 Communications, ESF6 Mass Care, ESF7 Logistics, etc.) and (ii) the problematic activities and (iii) their associated primary and secondary infrastructure that were identified during the interviews. Each selected text segment was defined as a variable (i.e., a problematic activity), and the texts related to the variable were defined under a category representing that variable. The categorized texts were then retrieved using the text analyzer. Subcategories were also defined for some variables representing primary or secondary infrastructure. For instance, “communication” (i.e., an operational issue of shelters) was defined as a parent category, while “mobile service providers” (i.e., primary or secondary infrastructure) was defined as a subcategory.

The retrieved text segments were presented along with their relevant time in a tabular format that shows the chronological order of each activity under the defined category. This information was then used to track the dynamics of the emerging interdependencies between the primary and secondary infrastructure during recovery. The emergence of secondary infrastructure support operations and the duration of their availability are presented in Figure 3. For example, Bay County experienced issues with diverse infrastructure service immediately after Hurricane Michael made landfall (10/10/2018), which triggered emergence of secondary infrastructure to support the operation of shelters. Specifically, some secondary infrastructure emerged immediately to help shelters accommodate diverse evacuee needs by supplying food, water, and other necessities (e.g., social groups [3c in Figure 3]; 10/12/2018) while others emerged some time later to help shelter residents meet their travel needs (e.g., Faith-Based volunteers [6c in Figure 3]; 10/24/2018).

Results and Findings


Many of the county and state agencies identified communication as a major challenge for recovery (especially during the operation of hurricane shelters). For instance, one of the challenges was the difficulty in disseminating post-disaster information (e.g., about shelter operations, food, and water supply, etc.) to the public. The communication infrastructure was severely affected during the first two weeks after the hurricane made landfall. While there was spotty communication available, it remained down for the whole operation of the EOC during this period. This disruption affected other emergency response activities as well. For instance, communication between ESF6 Mass Care, ESF7 Logistics, and other ESFs was not effective enough to supply resources to shelters in a timely manner. Due to the lack of communication, it was difficult to know how many people were in a shelter at a given point in time (1a in Table 1).

Table 1. Primary and secondary infrastructure important for recovery from Hurricane Michael (click image to enlarge).

Numerous secondary infrastructure entities emerged to fill the gap left by the communication infrastructure (i.e., primary infrastructure). For example, radio channels were used and operated by public information officers (PIOs; 1b in Table 1) to provide general information to the public on topics such as health concerns over the use of backup generators, curfew policy, shelter operations, and emergency reporting. Furthermore, social media platforms such as Facebook and Twitter (1c in Table 1) were used to connect people and the PIOs. Other traditional methods of communication such as flyers, press releases, and airplane message banners were also employed. Moreover, ad hoc communication services were deployed by mobile companies, such as Verizon, AT&T, and T-Mobile, to supplement the communication function (1d in Table 1). Even after communication was restored, as seen in Figure 3, the ad hoc services continued to provide reliable communication support to the shelters and the county. The communication equipment used by mobile companies included cells-on-wheels (COW) assets. Free mobile phones and mobile hotspots were also provided for EOC volunteers and disaster victims. Other supplementary systems such as lighting systems and gas generators were also provided to charge the ad hoc communication equipment.

Fig 3. Dynamic relationships between primary and secondary infrastructure (click image to enlarge).


Numerous prolonged power outages were reported throughout the Bay County area following Hurricane Michael. Approximately 96% of people in Bay County were without power after the hurricane struck. Power outages affect the provision of infrastructure services because entities rely on electricity to operate services. For instance, hospitals, businesses, traffic signals/lights, water distribution stations, water treatment plants, lift stations, wastewater treatment facilities, and airports all require a power supply to operate. The power outage caused a failure in one of the pump stations and water treatment plants in Bay County. Due to the health concern, a boil water notice was issued. One of the shelters was reported not to have a backup means of wastewater disposal, which challenged the operation of the shelter facilities and resulted in the relocation of the shelter’s evacuees. Communication also largely depends on power (e.g., internet access, mobile phones, cell phone towers, etc.). Specifically, the power disruption affected the internal communications of the Bay County EOC and its external communications with shelter operators (2a in Table 1). The restoration of the power supply system remained ongoing until the last date of the reports (December 26, 2018), although the number of customers without power continuously decreased during recovery. During the disruption, other critical facilities, such as two wastewater treatment facilities, had to rely on backup generators to continue operations.

The operators of public utility services took the lead in restoring disrupted services. More than 5,000 crew members worked to restore power in Bay County (2b in Table 1). As secondary infrastructure, other utility companies from outside the county deployed repair crews to support the power restoration process. For instance, Gulf Power deployed more than 4,000 repair personnel from 16 states to support the restoration process across multiple staging sites (Gulf Power Company, 201815). On December 26, 2018, it was reported that 99% of electricity service in Bay County had been restored.

Mass Care and Feeding Operations

In response to the hurricane, the Bay County EOC initially opened three shelters (i.e., Northside, Rutherford, and Bozeman) to accommodate a variety of population segments. The three shelters were relocated to two shelters on the beach of Panama City where electricity was available. Later, a shelter was opened at Arnold High School to consolidate the two shelters located on school sites so that they could be reopened for normal school operations. Note that these shelters provided not only safe places for evacuees to stay, but also essential services such as the distribution of water, food, and other necessities. In addition, points of distribution (PODs) were opened to hand out emergency supplies (3a in Table 1). PODs are places designated for people to pick up emergency supplies (e.g., food, water, etc.) following a disaster. There remained continuous demand for PODs as recovery proceeded; however, Bay County gradually closed its PODs (Figure 1), since continued POD operations could hinder the return of normal business activities.

Social groups and multiple organizations (i.e., secondary infrastructure) continued to provide feeding services and distribute supplies to affected communities. For example, the Salvation Army and the American Red Cross (ARC) operated mobile feeding vehicles throughout Bay County and roving vehicles to reach those who could not access feeding sites. ARC, Feeding the Gulf Coast, the Salvation Army, and World Central Kitchen (3c in Table 1) also opened several feeding locations and kitchens. As seen in Figure 3, these feeding activities by NGOs continued for up to 5 weeks in an effort to stabilize the community.

Damage Assessment, Search and Rescue, and Road Clearance Missions

The Urban Search and Rescue Task Force was responsible for locating, rescuing, and giving medical services to individuals who could not access health-care facilities in the aftermath of the hurricane. Meanwhile, state military supports (i.e., secondary infrastructure) were deployed to Bay County to assist in search and rescue operations and in road clearing missions. For instance, National Guard troops and the Coast Guard (4b in Table 1) helped to provide services related to security, the clearing of debris, and the securing of roads for repair personnel. In addition, the Florida Highway Patrol (FHP; 4c in Table 1) assisted in identifying roadways and bridges that were subject to disruption by debris and provided security escorts for emergency responders for road clearance missions, traffic control and security, and the staffing at PODs.

Local Law Enforcement

The Bay County Sheriff’s Office (BCSO; i.e., primary infrastructure) provided security and law enforcement and worked with the emergency managers to keep residents safe. In addition, the FHP (5b in Table 1) supported the local enforcement agency (i.e., secondary infrastructure) by guiding them on a patrol of the disaster-damaged areas. The BCSO and FHP together provided security escorts to city crews and helped to monitor traffic and update the status of interstate highways and bridges. Their collaboration helped to facilitate the effective deployment of humanitarian resources to various locations. The FHP was also tasked with assessing shelters for life safety (i.e., safety and security of personnel and evacuee) and critical issues. According to the situation reports, about 253 law enforcement officers were deployed during the response process.

Public Transportation

Regular transit services were resumed on October 26, 2018, approximately 15 days after Hurricane Michael made landfall. Transit services were provided for residents by the local transit agencies (i.e., Bay Town Trolley and Bay Area Transit, which constitute primary infrastructure) in order to help residents commute to various locations such as PODs, pharmacies, and appointments with FEMA. Shuttle service was also available from shelters to PODs. Public transportation in Bay County was supplemented by ride-share companies such as Uber (6b in Table 1) and faith-based volunteers (6c in Table 1). For instance, Uber provided services to first responders and shelter evacuees who were assigned to transitional long-term housing. Meanwhile, faith-based volunteers offered rides to assist evacuees in traveling to their new accommodations.


Primary infrastructure may have insufficient capacity, and as a result may not be able support the operation of hurricane shelters in the short-term or the longer-term recovery process. This can result in a disruption of critical functions thereby causing problems for community recovery. In such a trying situation, secondary infrastructure appears to fill the gap left by primary infrastructure. This study identified primary and secondary infrastructure in relation to operational issues facing critical facilities (i.e., shelters and their related critical infrastructure) in community recovery and tracked changing interactions in chronological order (Figure 3). Diverse entities (e.g., state agencies, NGOs, and other volunteer groups) across different infrastructure sectors were observed as secondary infrastructure (Table 1).

Furthermore, this study assumed such operational issues facing primary infrastructure were one of the triggers for activation secondary infrastructure. This research also examined how relationships changed over time between shelters and the primary and secondary infrastructure. The findings (Figure 3 and Table 1) corroborated the assumption about triggers. For example, public communication was one of the important functions for shelter operations and community recovery, and the disruption of the communication infrastructure immediately triggered the emergence of secondary infrastructure (i.e., the provision of other means of communication) such as the deployment of ad hoc communication methods. Interestingly, although the relationships between primary and secondary infrastructure were stimulated by the same mechanism as recovery efforts began (i.e., during the response and emergency phase), the duration of these relationships varied. This depended on the adequacy of the primary infrastructure’s capacity, with respect to the demand and urgency of corrective action. For instance, due to the large number of households experiencing power outages in Bay County, there was a pressing need to restore electricity service as soon as possible. As a result, out-of-state public utility companies deployed assets to expedite the restoration process until it was fully recovered. Meanwhile, Bay County closed its PODs early so that people would use local businesses to accommodate their needs (i.e., to promote the return of local business). Subsequently, state PODs (i.e., secondary infrastructure) were also closed. However, NGOs and faith-based organizations continued to operate feeding sites to meet the outstanding needs of evacuees and residents.

In addition to primary infrastructure, secondary infrastructure was also significant for the successful operation of shelters. As such, this reflects the diverse and complicated relationships that exist between primary and secondary infrastructure in supporting critical recovery functions (Figure 3 and Table 1). It is noteworthy that according to the developed seven-layer framework, secondary infrastructure consists of social and civic systems (Table 1). Although the county emergency agencies (i.e., the civic systems) had limited engagement from state and federal resources in the planning stage, there are ways to enhance engagement of secondary infrastructure through capacity building for social systems during the planning stage. Specifically, the availability and capacity of social systems during post-disaster recovery comes from community activities resulting from good will, which in turn depend on the social capital of a community (Hanifan, 191616). Social capital can be enhanced through trust, norms, and the networks of community members. Also, the identification of potential secondary infrastructure in the planning stage helps to enhance relationships with relevant entities through pre-disaster contracts or agreements. For example, some hospitals put an agreement with local water suppliers in place to ensure the availability of extra water during a water disruption (Choi, 201817).


This study examined the functions and availability of secondary infrastructure during post-disaster recovery. It also focused on understanding the changing nature of relationships between long-term shelters and their primary and secondary infrastructure. Results indicated that complex relationships between primary and secondary infrastructure exist, and may be utilized to support operation of shelters and overall community recovery. Although it is useful to understand these relationships for planning purposes, their complexity may impede the creation of an effective integrated response and mitigation plan. For example, ARC is well known to be a secondary infrastructure that supplements mass care functions. It is also likely that other humanitarian organizations will be deployed to support affected communities in an emergency. As a result, shelters may also be flooded with redundant supplies from multiple organizations while missing some key resources. If ARC knows where other NGOs and volunteer groups are deploying humanitarian teams and is able to coordinate with them, ARC can effectively coordinate humanitarian logistics plans. However, no data management platform exists to allow different infrastructure entities to communicate and coordinate their planning across the life cycle of disaster management (i.e., the preparation, response, recovery, and mitigation phases). In the future, we plan to further explore the diverse interactions between different infrastructure entities and develop a unified data management platform for integrated resilience planning.


  1. Sellers, F. S., Begos, K., & Zezima, K. (2018). After Hurricane Michael, Panama City residents cope with no power, cash-only transactions and baby-wipe showers. Retrieved 25 October 2018, from 

  2. Farrington, B., & Lush, T. (2018). Major Hurricane Michael bearing down on Florida Panhandle. Retrieved 25 October 2018, from 

  3. Associated Press. (2018). Hurricane Michael leaves trail of flattened homes, mud and harrowing tales of survival. Retrieved 25 October 2018, from 

  4. Mahoney, E. L., & Mower, L. (2018). Hurricane Michael update: hardest-hit areas still largely without power as death toll expected to rise. Retrieved 25 October 2018, from 

  5. Cutter, S. L. (2003). Social Vulnerability to Environmental Hazards. Social Science Quarterly, 84(2), 242–261. Retrieved from 

  6. Holand, I. S. (2014). The lifeline issue in social vulnerability indexing: A review of indicators and discussion on their application. Natural Hazards Review, 16(3), 140404032024009. Retrieved from 

  7. Ozguven, E. E., Horner, M. W., Kocatepe, A., Marcelin, J. M., Abdelrazig, Y., Sando, T., & Moses, R. (2016). Metadata-based Needs Assessment for Emergency Transportation Operations with a Focus on an Aging Population: A Case Study in Florida. Transport Reviews, 36(3), 383–412. Retrieved from 

  8. Washington State Military Department. (2012). Planning Considerations for Host Communities and Reentry in the Puget Sound Region. Retrieved from Camp Murray, WA: d.pdf 

  9. McDonald, Z. (2018). Last Bay County hurricane shelter closes. Retrieved 5 April 2019, from 

  10. Deshmukh, A., & Hastak, M. (2014). Enhancing post-disaster recovery by optimal infrastructure capacity building. International Journal of Research in Engineering and Technology, 3(28), 5–12. Retrieved from 

  11. Flentge, F., & Beyer, U. (2007). The ISE Metamodel for Critical Infrastructures. In Critical Infrastructure Protection (Vol. 253, pp. 323–336). Boston, MA: Springer US. Retrieved from 

  12. Ouyang, M., & Dueñas-Osorio, L. (2014). Multi-dimensional hurricane resilience assessment of electric power systems. Structural Safety, 48, 15–24. Retrieved from 

  13. LaPan, C. (2013). Review of QDA Miner. Social Science Computer Review, 31(6), 774–778. Retrieved from 

  14. Provalis. (2016). QDA Miner User Guide, 1–272. Retrieved from 

  15. Gulf Power Company. (2018). Gulf Power crews begin rebuilding after Hurricane Michael. Retrieved 5 April 2019, from 

  16. Hanifan, L. J. (1916). The Rural School Community Center. The Annals of the American Academy of Political and Social Science, 67, 130–138. 

  17. Choi, J. (2018). Dynamic strain capacity analysis and planning for critical infrastructure to improve community resilience to disasters. PhD Dissertation, Purdue Univ., West Lafayette, IN.