A Comprehensive Assessment of Buildings for Post-Disaster Sustainable Reconstruction: A Case Study of Beirut Port

Abstract

Natural and man-made disasters constitute a considerable threat to humans, especially when intertwined with complex geopolitical situations. Effective decision-making and management during post-disaster reconstruction projects should be based on an effective assessment of damages caused by disasters by considering social, economic, and environmental factors. An analysis of prior research on post-disaster assessment reveals a predominant focus on physical indicators. However, recognizing the crucial role of socio-economic factors in the post-disaster reconstruction process, this paper introduces a comprehensive methodology for evaluating disaster-related damages by considering both physical and socio-economic factors. The proposed method is initiated by identifying relevant physical and socio-economic indicators. These indicators are then synthesized based on the local context and experts’ opinions to derive the Physical Priority Index (PPI) and Socio-Economic Priority Index (SEPI). These indices subsequently guide the prioritization of reconstruction efforts, aligning with the decision-makers’ strategic vision. This method was employed to assess damages stemming from the Beirut port disaster, utilizing three physical indicators and nine socio-economic indicators. The examination of the PPI and SEPI of a major area of Beirut port did not unveil a straightforward correlation between these two indices. The low correlation between these indices increases the complexity of decision-making. However, given the profound socio-economic challenges in Lebanon, this paper recommends placing a higher emphasis on SEPI in the decision-making process. Nevertheless, stakeholders retain the flexibility to tailor their approach by combining PPI and SEPI indicators according to their policies. This adaptive approach ensures a nuanced and contextually relevant decision-making framework.

1. Introduction

In recent decades, there has been a notable increase in the frequency and severity of disasters [1]. This increase can be attributed to a myriad of factors, including climate change, rapid urbanization, and the expansion of conflicts [2]. The repercussions of these disasters resulted in extensive damages and losses for communities and infrastructure. Moreover, they triggered significant disruptions in essential services and decision-making processes. The aftermath of such catastrophes raises intricate inquiries about the subsequent reconstruction phase. Determining reconstruction priorities becomes a critical concern, especially given the constraints on resources and time [3,4,5]. These challenges also offer an opportunity to bolster both community and infrastructure resilience. However, conflicts further amplify the intricacy of the reconstruction endeavor, stemming from issues like weak governance, resource scarcity, and the breakdown of institutions and social cohesion.

Scholars discussed the complexity of the post-disaster reconstruction process [6], in particular the assessment of the damages for the formulation of effective reconstruction strategies [7]. This paper presents a comprehensive approach for the assessment of damages resulting from disasters occurring within intricate geopolitical settings. Such an assessment holds paramount significance within the reconstruction process, as it empowers both governing bodies and local communities to foster a holistic and collaborative comprehension of physical infrastructure and social dynamics. This shared understanding further facilitates the delineation of priorities and the establishment of a strategic framework for the subsequent reconstruction endeavor

The methodology outlined herein was formulated in the context of comprehensive research for rebuilding the Beirut port after the devastating explosion on 4 August 2020.

The significance of this paper is threefold. First, it deals with combining two complex factors, namely destructive disasters and complex geopolitical contexts. Second, it proposes an index to assist decision-makers in prioritizing sustainable post-disaster reconstruction. Third, it is applied to the reconstruction of the Beirut port, which combines a multitude of difficulties, including complex social, economic, and geopolitical issues.

Scholars have proposed various indicator systems for post-disaster damage assessment, covering different sectors. Ref. [8] presented a seismic screening of buildings in Canada using the Seismic Priority Index. Ref. [9] assigned five risk priority levels to facilities in Izmir for a seismic assessment procedure that assists decision-makers with strategic planning and buildings’ rehabilitation. Ref. [10] analyzed the performance of low-rise reinforced concrete structures in Ecuador after the 2016 Earthquake using the structural priority index that includes columns and walls indices. Ref. [11] proposed a multi-dimensional damage assessment framework for decision-making including social, economic, and environmental sides. Ref. [12] developed a framework for post-disaster building damage assessment via effective crowdsourcing, involving non-expert citizens and including physical damage to buildings without considering community features.

The state-of-the-art shows that scholars focus on physical indicators for buildings’ post-disaster assessment. Ref. [13] argued that building back better would remain a major challenge if the post-disaster reconstruction process does not integrate vulnerability. The Sendai Framework for disaster risk reduction focused on the importance of sustainability in reducing the risks of natural and man-made disasters and accelerating post-disaster recovery while improving the community’s livelihoods [14]. A new Build-Back-Better conceptual model was developed to analyze recovery after the Changning earthquake in China. The model used 35 indicators, including sustainable livelihoods and community resilience [15]. Ref. [16] proposed a prototype for post-disaster housing reconstruction that responds to the demands of the disaster victims by being affordable, wind-resistant, and sustainable. Several authors showed that community participation in rural areas can promote rapid recovery and sustainable development. Nevertheless, none of the suggested frameworks considered geopolitical factors that could influence the recovery process. Furthermore, frameworks encompassing elements beyond the structural state of buildings, such as the PEOPLES framework [17], were too complicated to implement, and lacked community engagement.

This research aims at the construction of a comprehensive framework that accounts for the following:

• Covering the buildings’ physical structures and the social and economic issues.
• Taking into account the complex geopolitical context.
• Involving the post-disaster construction stakeholders, particularly the local community.

This framework is applied to the reconstruction of the Beirut port, which resulted in huge damage to buildings and around 200 fatalities and thousands of injuries [18].

The paper starts by describing the research methodology, which includes the study’s objective, the procedure followed to define the Priority Index, the categories, the indicators, and the method of calculating the Priority Index. Then comes the application to the Beirut port disaster, including a brief description of the local context, data collection, data analysis, an estimation of indices priority, and the recommendation for post-disaster reconstruction.

As a matter of convenience, all of the nomenclature used in this study is listed in Abbreviation.

2. Methodology

This research aims to establish a comprehensive framework for assessing building damage resulting from disasters. The framework uses a PI as a decision-support tool for policymakers. This index considers both the physical features of buildings and the socio-economic factors in the conflict area. Figure 1 illustrates the methodology used to determine this index. It started with a literature review to determine the first set of indicators for assessing the post-disaster damage of buildings. This set was then adapted to the local context using experts’ opinions. The SWARA method, detailed in Appendix B, was used to rank indicators and determine their weights according to experts’ opinions. This method was preferred to the AHP method because the AHP requires n(n − 1)/2 pairwise criteria comparisons, and it becomes practically unfeasible to perform such huge consistent pairwise comparisons [19].

2.1. Literature Review

Scholars proposed different indices for post-disaster damage assessment using digital and technological tools. Some indices were used to evaluate resilience or vulnerability. Some scholars focused on only one issue, such as physical or structural aspects of constructions or social and economic patterns. Others combined several problems, such as physical, socio-economic, management, and governance. Ref. [20] conducted a critical review of tools proposed for post-disaster physical damage assessment in lifeline networks and buildings. Tools included satellite imagery, remote inspection, field inspection, multi-sensor networks, social media, LIDAR sensors, strain gauges, computer vision, and empirical relationships. Ref. [14] identified the data required for post-disaster assessment, ranging from pre-disaster mitigation to post-disaster recovery and reconstruction. This included diverse data sources such as cadastral maps, statistical data, information on damaged buildings and assets, lifelines, insurance data, and community resources. A study by [21] on the development of Learning Cities emphasized the contribution of social and economic status factors to the socio-economic well-being of residents in deprived areas, which is essential for the sustainability of cities. Governments created tools for post-disaster damage assessment to manage the emergency and recovery phase. The US government has used active crowdsourcing via an online reporting form of storm damage and flooding. It included only the building damage level on a four-level scale and the flood level (street level or inside the home) [22]. In the multi-dimensional damage assessment (MDDA), Ref. [11] calculated the total damage as the sum of social, economic, and environmental damages. The social damage comprised the community’s physical and mental damage. The economic damage included the value of the lost assets such as infrastructure and business. Environmental damage referred to the losses in ecosystems, wildlife, biodiversity, etc. Ref. [23] conducted background research to present a theoretical and practical approach in preparing a damage assessment survey. The study led to a list of indicators for post-disaster damage assessment in a rural context with a human-based approach. The indicators were classified into five categories: general information, physical evaluation of the building, social and economic evaluation of the residents, and overall evaluation, such as trust in the state and level of expectations. Ref. [24] developed an index system for disaster damages assessment based on four categories: disaster properties, social impact, economic influence, and natural environment. The list included sub-indicators in each category such as the number of people affected, the number of people to be resettled, direct and indirect losses, the unemployment rate of the affected population, and many others. Ref. [12] designed a survey questionnaire for building damage assessment for public participation, with a framework for the quantification and reduction in uncertainty in the outcome of the assessment. The survey only included the building’s physical indicators, such as the conditions of wall structures, roof structures, wall, cladding, roof decks and covers, windows and doors, and level of building damage. The survey did not address other aspects of the post-disaster situation. Ref. [25] identified the post-disaster needs for assessing damages in Greater Houstonians due to Harvey. They used a set of social vulnerability and disaster impact indicators that have not been well integrated within US post-disaster needs assessment protocols. These indicators included a comfortable place to sleep, adequate drinking water, electricity, adequate food, access to health care or medical services, separation from family members, and health-related issues. Ref. [26] argued that damage assessment methodologies should be tailored to the diverse information needs in post-disaster contexts. The housing damage assessment included (i) damages, such as houses fully or partially destroyed or household goods destruction; (ii) losses, such as demolition and rubble removal and loss of rental income; and (iii) needs, such as housing reconstruction, repair, retrofit, transitional shelters, settlement planning, and training. Ref. [7] proposed an integrated approach based on multi-criteria decision-making (MCDM) techniques to aid decision-makers in prioritizing post-disaster projects based on 21 sustainability criteria. Authors used MCDM for emergency and recovery phases in the disaster management cycle. Ref. [27] used four multicriteria decision-making techniques—AHP, TOPSIS, COPRAS, and BORDA—to evaluate the selection and suitability of emergency assembly areas for the Gölbasi district of Ankara, Turkey, a critical disaster management issue.

2.2. Beirut Context

Beirut is the capital of Lebanon, has existed for millennia, and is located across the Mediterranean Sea coast. The Lebanon Crisis Response Plan data estimate Beirut’s total population at 1,291,280 [28], with a total area of 146 km². Beirut has rich culture and history that can be observed in the city’s urban pattern, architectural expressions, and sociocultural fabric. However, it is marked by significant socioeconomic inequalities, with compounded crises from the financial collapse, the COVID-19 pandemic, and the explosion causing adverse effects on its economic activity and individuals’ livelihoods. Poverty levels in Lebanon have rapidly risen, with estimates revealing that over 55 percent of the country’s population were trapped in poverty in 2020 [29]. The current situation is characterized by depleted business activities, high unemployment rates, food insecurity, and increased poverty [28]. Moreover, Beirut City suffered from decades of civil unrest, conflict, underinvestment, lack of reliable data, unmanaged city growth, and poor governance [28]. According to [30], during the Lebanese civil war (1975–1990), building construction was affected by poor building code design provisions and a lack of material quality control. Despite being an area of moderate seismicity, most of the buildings in Beirut were designed to resist gravity loads only during that era, with little or no consideration for lateral loads. Seismic standards were introduced in the 1990s. However, they were not strictly enforced until 2013. Beirut also faces vacancies in building units because of factors such as unaffordability, real estate development, inappropriate land use, and urban planning policies.

On the 4th of August 2020, the massive explosion at the port of Beirut resulted in the loss of more than 200 lives and more than 6500 injuries. The explosion left about 300,000 people homeless, and approximately 25,000 buildings were affected, varying from minor damage to total collapse. The massive destruction resulted in a national and international wave of solidarity. However, the coordination among the stakeholders and the management of the process was often ineffective, which led to conflicts between owners, organizations, donors, and public authorities, as well as to redundancy and duplication in the renovation process. The continued presence of a significant number of damaged buildings, even after nearly three years since the explosion, serves as tangible evidence highlighting the significance and relevance of this research. The study area is limited to four of the most damaged zones near the explosion’s epicenter (Figure 2). It includes around 300 buildings with various architectural and structural types. The distance between the epicenter and the study area ranges from 500 m to 900 m.

2.3. Scales of Indicators

Before presenting the list of indicators, it should be mentioned that most of the buildings in Beirut include many units. On the other hand, buildings with only one unit are rare. This means there can be many owners in the same building or one owner and many residents or renters. Therefore, the decision-making includes three levels: the overall building level, the shared or common space level, and the personal unit level.

The overall building level encompasses works that target the entire building, such as exterior facade renovation, whole demolition and reconstruction, and so on. The intervention at this level requires the participation of the owners and residents as well as national and local officials and donors.

The shared space level includes shared spaces such as elevators and stairs. It also requires the participation and coordination of different stakeholders and all the owners and residents in decision-making. These two levels reveal the necessity of having a building committee for decision-making and managing the reconstruction process.

Personal unit levels mainly concern the owner and the occupant. The owner is responsible for the renovation works.

2.4. List of Indicators Based on Experts’ Opinions

The first list of indicators was proposed to 33 experts in Lebanon. Then, it was updated and reduced to 12 indicators that were classified into two categories: buildings’ physical characteristics and socio-economic indicators. The final list of indicators is presented in

Table 1. List of indicators for the PI (Source: Author).

Scales of Indicators	Physical Characteristics	Socio-Economic Profile
Building Scale	Building Type	Occupancy ratio before and after the disaster
	Damage Level	Estimated rehabilitation costs
	Cultural Heritage
Shared Space Scale		Capacity of the owners to afford renovation costs
		Estimated rehabilitation costs
		Compensation from funding
Personal Unit Scale (For N Units)		Capacity of households to afford renovation costs
		Estimated rehabilitation costs
		Presence of temporary shelter
		Compensation from funding

.

The indicators are classified into two categories:

(1)

Physical Characteristics Indicators

The physical indicators encompass the architectural and structural features, including the following:

• the building’s purpose and characteristics;
• the level of damage.

These factors help decision-makers identify the most affected and damaged areas and the building profiles, which are vital for calculating the cost of renovation or reconstruction.

The building type or destination is an important indicator in the physical characteristics category: residential, industrial, commercial, public institution, cultural or social, health, education, or mixed-use. It has a high role in prioritizing the reconstruction process.

The damage level greatly impacts the reconstruction project, from the renovation techniques to the material required, the labor, and the costs [23,31]. The third indicator concerns cultural heritage [32]. As mentioned before, Beirut has a rich cultural heritage. Buildings that are considered cultural heritage are subjected to specific architectural requirements.

The damage level scale ranges from 1 to 4:

Level 1: no damage

Level 2: non-structural damage (Structural elements such as beams, columns, and walls are intact; damage is limited to non-structural elements such as windows, partitions, and facades elements)

Level 3: between 25% and 50% of the building was damaged (structural elements were damaged)

Level 4: more than 50% of the building damaged (total collapse of the whole building, or some floors)

This assessment could be performed visually, via a crowdsourcing survey form, for example, or any other available data source. Images and videos are also used in the process.

(2)

Socio-economic indicators

This research focused on the following socioeconomic indicators:

• buildings’ occupancy;
• renovation costs, owners’ capacity, and availability of aids;
• existence of temporary shelters.

This set of indicators aims to identify the most vulnerable groups that require additional assistance, allowing for a more effective and equitable aid allocation.

Recovery efforts and reconstruction projects are designed to restore livelihoods and improve the well-being of affected communities. Therefore, it is important to track the displacement of the affected residents to coordinate their return or relocation. To consider this issue, the framework uses the occupancy ratio before and after the disaster [31]. Additionally, the renovation cost is an important issue for post-disaster reconstruction. Estimating the rehabilitation costs is a critical factor in the reconstruction process [32].

At the shared space and personal unit levels, additional indicators are used to measure owners’ capacity to afford rehabilitation costs [17,23,31] and funding compensation [17,23,31]. The significant socio-economic inequalities in Beirut should be considered by decision-makers. The affected community can receive aid from many public authorities, insurance companies, national and international NGOs, or other organisms. However, the recovery process should follow the principle of “Leaving no one behind”; therefore, the aid allocation program must consider funding received by the community. These indicators target the SDG10 “Reduced Inequalities”, and along with the other indicators, establish an understanding of the socio-economic issues and vulnerabilities.

For personal units, the availability of temporary shelters must also be considered [17,23,31]. Although the strong family ties in Lebanon worked well after the disaster, a lot of affected people suffered from a lack of temporary shelter.

The capacity to afford the rehabilitation costs is measured using the Likert scale:

• 1: All the costs can be covered
• 2: The majority of the costs covered
• 3: Half covered/Half not covered
• 4: Majority not covered
• 5: Cannot cover any cost

Table 2. Description of the indicators for the PPI and SEPI.

Indicator	Description	Units
Physical Characteristics of the Building
Building Type	Residential, healthcare, education, industrial, commercial, public institution or governmental, cultural or social, or mixed-use	NaN
Damage level	1—Total collapse: more than 50% of the building damaged; 2—Major damage: between 25% and 50% of the building damaged; 3—Minor damage: only non-structural damage; 4—No damage	NaN
Cultural Heritage	Is the building considered a cultural heritage or not?	YES/NO
Socio-Economic Profile of the Residents
Occupancy ratio before and after the disaster	(Number of occupied units/Number of total units) × 100 calculated before and after the disaster	Percentage
Building’s estimated rehabilitation costs	Estimation of the rehabilitation costs	$
Capacity of owners to afford shared space renovation costs	1—All the costs can be covered; 2—Majority of the costs covered; 3—Half covered/Half not covered; 4—Majority not covered; 5—Cannot cover any cost	NaN
Shared space estimated rehabilitation costs	Estimation of the shared space rehabilitation costs	$
Compensation from funding for the shared space	Percentage of not funded shared space renovation costs	Percentage
Capacity of households to afford their personal unit renovation costs	1—All the costs can be covered; 2—Majority of the costs covered; 3—Half covered/Half not covered; 4—Majority not covered; 5—Cannot cover any cost	NaN
Personal Unit estimated rehabilitation costs	Estimation of the unit rehabilitation costs	$
Presence of temporary shelter	Do the households have a temporary shelter that is near their principal house?	YES/NO
Compensation from funding for the personal unit	Percentage of not funded personal unit renovation costs	Percentage

describes each indicator with the method of calculation and the unit.

2.5. Estimation of the PI

To calculate the priority indices, numerical values were assigned to qualitative indicators. For the building type indicator, the healthcare centers and hospitals and residential buildings are the most critical to renovate. The scores of each building type are listed as follows:

• Healthcare Center: 5
• Residential and Mixed-use Commercial/Residential, Education Center, Public Institution: 4
• Commercial and Industrial: 3
• Cultural and Social: 2
• Not occupied or No data: 1

The cultural heritage status is a YES/NO question, so it needs to be quantified. The state of being a cultural heritage building is more prioritized in reconstruction projects because there is a crucial need to save the cultural aspect of the damaged city and prevent these buildings from demolition and real estate interests. Therefore, the cultural heritage status is given a score of 1, while a score of 0 is assigned to other buildings.

For the temporary shelter, affected residents with no temporary shelter should be prioritized in the reconstruction process. Therefore, a score of 1 was assigned for the absence of a temporary shelter and a score of 0 for others.

The PI is composed of two components, PPI and (SEPI):

$$PPI=w_{BT}\times BT+w_{DL}\times DL+w_{CH}\times CH$$

(1)

$$SEPI=w_{OR}\times OR+w_{BRC}\times BRC+w_{SOC}\times SOC+w_{SRC}\times SRC+w_{SCF}\times SCF+w_{POC}\times POC+w_{PRC}\times PRC+w_{PCF}\times PCF+w_{TS}\times TS$$

(2)

It is important to mention that for each building, there can be many personal units. However, the PI calculated is for the whole building. Therefore, the indicators related to personal units are obtained by calculating the mean of all the personal units in the building.

The estimation of each indicator was conducted as follows:

• The survey had two-level questions, one for the entire building and shared/common spaces, and the other for each personal unit in the building. As a result, two tables were created that were related to each other with a one-to-many relationship. The first step consisted of merging these tables based on the global ID of the building.
• The second step involved cleaning the data and addressing missing or NaN values.
• The third step consisted of handling nominal or categorical data by assigning numerical values to them.
• The fourth step involved calculating the mean of each field within the data corresponding to personal units in the building, for each building.
• The fifth step was data normalization, as each field had its scale, unit, and measure. The Min–Max method was used for normalization, which is expressed using the following equation

$$Z_i=\frac{x_i-x_{min}}{x_{max}-x_{min}}$$

(3)

6.

The final step was to multiply each indicator by its weight and sum the results for each building to obtain the PPI and SEPI, as in Equations (A2) and (A3).

3. Application to the Beirut Port Disaster

This section presents the application of the proposed methodology to the post-disaster buildings damage assessment in Beirut.

3.1. Indicators’ Weights Estimation

Estimating the indicators’ weights is based on the opinions of a panel of experts in the building sector, urban planning, disaster management, and reconstruction. Experts’ opinions were obtained through a questionnaire composed of two sections. The first focused on the experts’ profile, including their field of expertise, experience, and working sector, while the second concerned the indicators’ priority based on a 5-level score. High scores indicated high relevance.

Table 3. Description of experts’ profiles.

Expert’s Field	Nb of Experts	Work Sector	Nb of Experts	Years of Experience	Nb of Experts
Civil Engineering	20	Academic Field and Research	12	16 to 25	11
Project Management	4	Construction and Engineering Company	12	More than 25	8
Urban Planning	1	NGOs	2	5 to 10	6
Social Sciences	1	International Organizations	2	11 to 15	5
GIS Coordinator	1	Lebanese Armed Forces	1	Less than 5	3
Disaster Management	1	UN-Agency	1
Humanitarian and Emergency Response Professional (educational background in engineering)	1	Lebanese Red Cross	1
Governance and Quality Management Consultant	1	Consultant and Training Company, Academic, collaboration with International Organizations and Lebanese Armed Forces	1
Architect	2	Consulting	1
Environment Expert	1

summarizes the profiles of the 33 experts. In total, 36% of the experts worked in the academic sector, while 36% worked in construction and engineering companies. The remaining experts were distributed across organizations (national and international), consulting, and the Lebanese Armed Forces.

Table 4. Weights of indicators following the SWARA method application.

Attribute	Total Score	sj	kj	qj	Weight
Damage Level	150		1	1	0.37
Building Type	133	0.17	1.17	0.8547	0.32
Cultural Heritage	132	0.01	1.01	0.8462	0.31
Occupancy ratio before and after the disaster	131		1	1	0.115
Estimated rehabilitation costs of the building	131	0	1	1	0.115
Compensation from funding at the shared space level	131	0	1	1	0.115
Presence of temporary shelter at the personal unit scale	128	0.03	1.03	0.971	0.111
Estimated rehabilitation costs at the shared space level	127	0.01	1.01	0.961	0.11
Estimated rehabilitation costs at the personal unit scale	127	0	1	0.961	0.11
Compensation from funding at the personal unit scale	127	0	1	0.961	0.11
Capacity of owners to afford renovation costs at the shared space level	124	0.03	1.03	0.933	0.107
Capacity of households to afford renovation costs at the personal unit scale	124	0	1	0.933	0.107

summarizes the experts’ opinions and the results of applying the SWARA method. It shows that damage is the most important indicator of the building’s physical characteristics (Global score = 150). It is followed by the building type and its status of being a cultural heritage or not, which are nearly equal in importance (133 and 132, respectively). The occupancy and estimated rehabilitation costs are equally important in the residents’ socio-economic profile category (131). The compensation from funding is the most important at the shared space level (131), followed by the estimated renovation costs (127).

At the personal unit level, the presence of a temporary shelter is the most important, with a score of 128, followed by the estimated renovation costs (127). The capacity of the owners is the least critical indicator (124). The last column shows that the weights of the indicators are close, revealing the importance of selected indicators.

3.2. Data Collection Method

Data were collected using a questionnaire created with the ArcGIS Survery123 application.

Table 5. List of questions in the survey used for data collection in Beirut.

No	Question
1	Geolocation of the building
Physical Characteristics of the Building
2	Select the building type
3	What is the number of floors?
4	What is the damage level?
5	Is it cultural heritage or not?
6	What is the current status of the building? (recovered, still damaged, recovery ongoing)
7	Upload an image of the building
Socio-Economic Profile of the Residents
Building Scale
8	What is the estimation of the renovation costs of the building in dollars?
9	To what level are the owners capable of affording renovation costs?
Shared Space Scale
10	What is the estimation of the renovation costs of the shared space in dollars?
11	What is the value of the funding?
Personal Unit Scale
12	To what level is the owner capable of affording renovation costs?
13	What is the estimation of the renovation costs of the personal unit in dollars?
14	What is the value of the funding?
15	Do the households have a temporary shelter?

summarizes the list of questions.

According to the Disaster Risk Management Unit of the Lebanese Government, the Lebanese Army assessed 37,044 units [33]. The assessment included physical characteristics of the building such as the number of floors, units, the damage level, and the NGOs working on the building. However, the assessment did not take into account the social and economic aspects of the buildings.

This study included the assessment of 159 buildings in four main severely impacted zones: Karantina, Mar Mikhael, Rmeil, and Gemmayzeh, (Figure 3). This residential zone is the nearest to the explosion’s epicenter within a range of 500 m to 900 m. It includes a high proportion of cultural heritage buildings.

3.3. Data Descriptive Analysis

Figure A1 shows the distribution of the assessed buildings according to their type. As mentioned, most buildings are either residential or mixed-use (commercial/residential). The area is also characterized by a high number of cultural heritage buildings: around 46% of the buildings are classified in the category of cultural heritage (Figure A2).

As this area is close to the explosion’s epicenter, 133 buildings (84%) experienced a major collapse, whereas 24 buildings (15%) faced non-structural damage. Only two buildings (1.25%) were destroyed (Figure A3). Although the current status of the building is not an indicator for the PI, this information was collected to track the recovery process. In January 2023, 61% of the assessed buildings were recovered, 29% were still damaged, and 9 % were under renovation (Figure A4).

Figure A5 represents the distribution of the buildings’ renovation costs. Half are below 231,000 US $, while the average is 273,100 US $, with a standard deviation of 207,667 US $.

Figure A6 illustrates the distribution of the pre and post-disaster occupancy ratios. It shows that the most frequent value is 0%, which means that the number of occupied units remains unchanged after the disaster. Following that, the next value is 100%, indicating that all previously occupied units are currently vacant. The remaining values range between 0 and 100%. Notably, only one building shows more occupied units after the disaster. This building attracted residents due to its significantly improved aesthetic appeal after renovation. Several buildings, marked by an occupancy ratio of 100%, remain unoccupied even three years after the disaster. The increase in vacancy can be attributed to the fact that many residents were renters who were not responsible for the renovation expenses. Consequently, when building owners did not proceed with renovations, the renters vacated their apartments. Among the 98 renovated buildings, there are 39 in which all the previously occupied units are now vacant. In these cases, the residents, who were also renters, chose to relocate elsewhere after the disaster. It is worth mentioning that several owners opted to convert their apartments into Airbnb rentals rather than renting them to individuals or families, as it proved to be more financially advantageous.

Figure A7 and Figure A8 show that the majority of the owners cannot afford the estimated renovation costs of the shared space and the personal units. Figure A9 and Figure A10 reveal that for the majority of the renovated buildings, the costs were covered by NGOs. However, internal renovation varied between one organization and another. Most residents found a temporary shelter thanks to strong social bonds (Figure A11). Despite the risks, many individuals, particularly men, opted to stay home during the reconstruction process.

4. Results

4.1. Physical Priority and the Socio-Economic Priority Indices

Python in the ArcGIS API was used to determine the Physical Priority (PPI) and the Socio-economic Priority (SEPI) indices for each building.

Figure 4 illustrates the variation in the PPI with the building type. The inclination of residential or mixed-use buildings exhibits an increase in PPI. Figure 5 demonstrates that the severity of damage correlates with a higher PPI, while Figure 6 highlights a distinct rise in PPI based on the cultural importance of the buildings. It is worth noting that the PPI is not only dependent on this factor, but the other three factors influence it. As a result, certain non-cultural heritage buildings may possess a higher PPI than cultural heritage buildings, and buildings with minor damage can have a higher PPI than those with major damage. The discrete values of PPI are due to the influence of the three indicators on the PPI. The PPI values range between 0 and 1, with an average of 0.597, a standard deviation of 0.2, and a median of 0.505.

Figure 7 illustrates the variation in the SEPI with the total renovation costs. We observe a tendency for an increase in the SEPI with a building’s estimated renovation costs. Figure 8 does not indicate a clear relationship between the SEPI and the occupancy ratio. Figure 9 and Figure 10 do not show a clear variation tendency with the owner funding capacity for the shared and individual spaces.

Figure 11 indicates that the SEPI tends to increase with the increase in personal unit renovation costs. The SEPI values range between 0.052 and 0.68, with an average of 0.36, a standard deviation of 0.12, and a median of 0.357.

4.2. PPI and SEPI Spatial Distribution

Figure 12 shows the heatmap of the PPI. It indicates that the Mar Mikhael and Karantina zones have the highest PPI, followed by Rmeil and Gemmayzeh. In this area, most damages are in the categories of major damage or structural damage. Since the study area is residential, the building type with the highest frequency is residential and mixed-use (residential and commercial). It is important to mention that the mixed-use building type is common in Lebanon, especially in Beirut. Stores and markets usually occupy the ground floor and residential units on the other levels. In this area, the PPI does not vary significantly because most buildings are residential or mixed-use and severely damaged. The status of cultural heritage makes the difference. Figure 13 shows the heatmap of the SEPI. Since this index considers nine indicators, the results vary widely. It indicates that Mar Mikhael and Karantina also record the highest SEPI.

4.3. Discussion and Limitations

Analyses of the PPI and SEPI priority indices variation for the selected case showed the complex relationship of these indices with the related indicators. Figure 14 shows the distribution curve of the PPI. It indicates a concentration around two intervals [0.5, 0.625] and [0.75, 0.875]. Figure 15 shows a concentration of the SEPI in the interval [0.29, 045]. Figure 16 illustrates the values of SEPI and PPI for all the buildings. It shows a low correlation between these indices (Standard correlation = 0.065). This result shows all the difficulties in establishing a global PI. Since the social issue in the Beirut port disaster is very important, the SEPI could be dominant in determining the post-disaster reconstruction.

The highest SEPI index concerned a three-story cultural heritage residential building (0.683). This building became completely vacant after the disaster. The owners could not afford the reconstruction costs; some renters were left without housing.

Another example of a building with a SEPI index of 0.6081 is a severely damaged mixed-use. Like the previous case, the owner could not afford the necessary renovations, leaving the renters without shelter. As a result, the building remains vacant while the renters have found alternative housing solutions.

The building with the lowest SEPI (0.052) was abandoned before the disaster. Despite having the maximum PPI value due to its cultural heritage and intended use for housing, the building suffered a total collapse. No one lived in it during the disaster, so it recorded the minimum SEPI.

Data analysis shows that Mar Mikhael is the most vulnerable zone in the PPI and SEPI assessments. Cultural heritage buildings mark this area, which suffered extensive structural damage. The population relocation in this zone has been relatively slow. Furthermore, the area experienced many casualties, intensifying the trauma and making it even more challenging for the affected families.

The data collection process encountered several limitations, which can be outlined as follows:

• Certain residents declined cooperation with this research.
• Some buildings were vacant, making collecting details about their owners difficult.
• The rising trend of converting apartments into Airbnb rentals further complicated data collection, given the transient nature of the residents.

To overcome these limitations, we need (i) more coordination among local authorities, communities, and NGOs, and (ii) training programs for people involved in damage assessment, disaster prevention, and emergency operations.

5. Conclusions

This paper proposed a methodology for a sustainable assessment of buildings as a first step in post-disaster reconstruction. It considers both the buildings’ physical characteristics and the affected community’s socio-economic context. The PPI considers the physical characteristics, which combines the building type, damage level, and cultural heritage. The socio-economic context is considered through the SEPI, which combines the occupancy ratio, renovation costs, public aid, temporary shelter, and the capacity of the owners. The study of the socio-economic profile is a pillar in building sustainable cities and communities.

This methodology was applied to four zones of the port of Beirut. Data were collected from 159 damaged buildings. Analyses of both PPI and SEPI did not indicate a clear tendency with the related input indicators. They also showed a low correlation between PPI and SEPI, which makes their combination complex for decision-making.

Considering the dramatic socio-economic context in Lebanon, it is more effective for public authorities, international donors, and NGOs to privilege the SEPI in decision-making. However, each stakeholder could combine the PPI and SEPI indicators according to this policy. However, a successful reconstruction requires effective coordination among stakeholders involved in the reconstruction and the involvement of the affected community in the reconstruction process.