1. Introduction
Due to global warming and urbanization, torrential rains and floods events have become more frequent [
1]. Among the recorded natural hazards globally, the frequency of flooding events accounts for nearly half of all major events [
2]. In 2020, the continuous rainy season caused a significant increase in precipitation, leading to large-scale mountain torrents and urban waterlogging in China [
3]. According to media reports, 433 rivers in China were flooded above the warning line, and nearly 38 million people in 27 provinces and cities were affected [
4]. The devastating consequences of the flood disaster make the humanitarian supply chain, which provides timely relief especially important [
5]. A successful humanitarian supply chain should reduce the losses and casualties caused by disasters as far as possible [
6]. However, the characteristic of sudden disasters is that the time, place, and intensity are unpredictable [
7]. This characteristic requires decision-making not only on time but also adjusted with the disaster situation [
8]. In addition, it is difficult for the stakeholders involved in the humanitarian supply chain to achieve coordination and build the necessary trust due to their different expectations and goals [
9]. Hence, it is essential for the humanitarian supply chain to be resilient to reduce the losses caused by disasters in an unstable environment. Simultaneously, it is necessary to determine the key factors and their interaction to maintain the resilience of the humanitarian supply chain.
In order to reduce the devastating consequences of disasters, research on the humanitarian supply chain has gradually increased in recent years [
10]. Kittisak compared the humanitarian supply chain with the commercial supply chain and proposed that understanding the similarities and differences between the two is the basis for applying the commercial supply chain management concept to the humanitarian supply chain [
11]. Singh argued that the fundamental chain of humanitarian supply chain and commercial supply chain are similar, and the research concept of the commercial supply chain can be used to improve the performance of the humanitarian supply chain [
12]. Zarei argued that transportation capacity is critical to the sustainability of the humanitarian supply chain [
13]. Iman cited regulatory uncertainty, inexperienced employees and high costs as key reasons why humanitarian supply chains are not sustainable [
14]. Paula emphasized that human resources management is of great significance to disaster preparedness and rescue during disasters [
15]. Hossein proposed that a flexible humanitarian supply chain can handle disasters more effectively. He also pointed out that the standardization of data and distribution is the most important for the flexibility of the humanitarian supply chain [
16]. Kumar pointed out that for non-profit supply chains such as the humanitarian supply chain, Radio Frequency Identification (RFID) technology can effectively reduce costs and reduce the severity of errors [
17].
Humanitarian supply chains must be resilient because of the instability of the environment and the complexity of the stakeholders [
18,
19]. Resilience is defined as the ability of the system to return to normal operations after a shock [
20]. This study focuses on the development of humanitarian supply chain resilience during flood events and asks the following questions: (1) What factors influence the development of humanitarian supply chain resilience? (2) Is there a correlation between these factors? (3) What are the most important factors that affect resilience? In order to answer the above questions, we integrated the Delphi method and literature review findings to determine the main influencing factors. After the first round of the Delphi survey, this study identified 18 key factors found to resilience in the humanitarian supply chain and further classified these into 6 main criteria. Data from three rounds of Delphi surveys was obtained and the interrelationships between the various indicators were identified according to the fuzzy-DEMATEL method. The interrelationships between indicators were transformed into a network structure diagram. Finally, using the ANP method, the weights of each factor based on the network structure diagram were identified. Combining the fuzzy-DEMATEL method and the ANP method, the correlations between the 18 indicators and the weights of each indicator were calculated. Moreover, this research emphasizes that five of these indicators are the key directions for building the resilience of the humanitarian supply chain, which can be ranked as: capability and strategic planning (C1), coordination and cooperation (B2), transparency of information (B1), risk awareness (E2), and agility in processes (D1).
The structure of this paper is as follows: In the following section, we review the relevant literature and identify the criteria and indicators that affect the resilience of the humanitarian supply chain.
Section 3 then explains the research design and methods in detail.
Section 4 analyses the evaluation of humanitarian supply chain resilience based on real data from the study area. In
Section 5, we discuss the research results and end with
Section 6 which presents the conclusions and proposed recommendations for future research.
5. Discussion
The humanitarian supply chain is of great significance to alleviating the loss and suffering of flood-affected people. However, due to the uncertainty of flood events and the unpredictability of demand, it is necessary to develop a resilient humanitarian supply chain. This study used the fuzzy-DEMATEL method combined with the ANP method for analyzing humanitarian supply chain resilience. The interrelations and weights among the indicators were determined as shown in
Table 5 and
Table 6, respectively. It can be seen from these two tables that capability and strategic planning (C1) is the key indicator. Agarwal discussed the main obstacles faced by the humanitarian supply chain and proposed corresponding solutions. Their research results showed that capabilities and strategic planning directly positively impact the humanitarian supply chain. This result is consistent with the results of this study [
21]. In the humanitarian supply chain, long-term capability and strategic planning, including logistics deployment and resource allocation, are critical. Moreover, capacity and strategic planning should include different stages of work meetings and work update links to eliminate the uncertainty. It is vital to note that the lack of forward-looking strategic planning has a significant impact on the resilience of humanitarian supply chains. Therefore, the managers of the humanitarian supply chain should be careful in making plans.
Coordination and cooperation (B2) among stakeholders is the second most important resilience indicator. Ergun’s research emphasized that good coordination between stakeholders is beneficial to improving the resilience of the humanitarian supply chain, which is consistent with the results of this research [
60]. The humanitarian supply chain involves many stakeholders. Hence, in the recovery process of flooding events, it is often seen that the rescue work is slow even with sufficient funds and materials. John’s research showed that many supplies were donated by the National Crisis Committee during the Chennai flood relief operation. Nevertheless, many poor communities did not receive the materials due to poor coordination [
35]. In addition, it is indispensable to consider the expectations of different stakeholders when coordinating the response. For example, volunteers usually do not seek rewards for their rescue behavior, while companies and NGOs pay more attention to the reciprocal results brought about by cooperation [
61]. Further, the expectations of companies and NGOs for cooperation are quite different. Companies expect their brand image to be promoted. The focus of NGOs is to ensure that the affected communities and individuals receive the donated materials. Generally speaking, when NGOs donate materials, they hope that cooperative organizations can deliver materials to the affected communities and individuals. It seems unreasonable for NGOs to cooperate with other organizations and expect something in return. However, given that NGOs face increasing scrutiny and fierce competition for donor funding, this expectation is understandable.
Transparency of information (B1) is the third-ranked resilience indicator. Dubey pointed out that information sharing and supply chain transparency are the keys to a resilient humanitarian supply chain [
34]. Prasanna’s research results revealed that information sharing among stakeholders is the main factor in the success of the humanitarian supply chain [
62]. Their views coincide with the results of this research. Due to the large number of stakeholders involved in the humanitarian supply chain and the lack of formal information-sharing agreements between agencies, information exchange in the humanitarian supply chain is often quite problematic. Thus, achieving timely and effective transmission of information is an inevitable prerequisite for realizing the coordination among the humanitarian supply chain. Dubey’s research suggested that blockchain technology in disaster relief operations can effectively guarantee the exchange of information [
33]. Nagendra pointed out that a cloud computing platform based on satellite big data can also improve the collection of real-time information [
5].
Risk awareness (E2) is the fourth most important indicator of resilience. Patil et al. emphasized that the lack of risk awareness is an obstacle to the operation of the humanitarian supply chain. They also proposed that the cultivation of risk awareness is a long-term process [
45]. This view is consistent with this research. Risk awareness includes the training of rescuers and the understanding of flood-affected community members. The temporary and voluntary nature of disaster relief operations leads to uneven risk awareness among rescuers. Therefore, basic training and practical guidance should be provided for rescuers. In the meanwhile, flood-affected people often demonstrate wait-and-see behavior as they are often reluctant to evacuate their homes. Kensuke Takenouchi pointed out that disaster prediction models may lead to information dependence and wait-and-see behavior of the public [
63]. Strahan argued that wait-and-see behavior is influenced by the level of warning and the psychology of protecting one’s property [
64]. The appearance of wait-and-see behavior will bring great trouble to rescue operations. Hence, it is essential to cultivate risk awareness among people in disaster-prone areas. Agility in processes (D1) is the fifth-most weighted indicator. Richard discussed the importance of agility for emergency disaster relief activities. Moreover, they emphasized that the realization of agility is positive for the resilience of the humanitarian supply chain, which is consistent with the results of this research [
65]. Changes in the needs of the affected people and the uncertainty of floods require an agile and flexible humanitarian supply chain that improves handling emergencies in the rescue period.
Among the top 5 indicators in terms of weight, three resilience indicators, transparency of information (B1), coordination and cooperation (B2), and capability and strategic planning (C1), are the cause indicators. This shows that the impact of these three indicators on other indicators in the evaluation system is greater than other indicators. On the other hand, agility in processes (D1) and risk awareness (E2) are effect indicators, which shows that these two indicators are easily affected by other indicators. When decision-makers implement optimization strategies for these two indicators, they should pay full attention to other indicators that impact these two indicators.
During the flood event in Hechuan District, Chongqing established an emergency command system, including two committees, eleven special offices, and four special commands. Nevertheless, the participation of multiple humanitarian agencies, companies, and more than 600 volunteers made coordination and information exchange difficult. Differences in the core interests of various organizations have rendered the relief efforts of many actors ineffective and marginalized. Many volunteers arrived with relief supplies but knew little about the situation. Furthermore, relief materials donated from all over the country flooded into Hechuan District due to the spread of the disaster on social media. However, the lack of coordination and information exchange resulted in the donated relief materials generally converging in one location, causing unnecessary waste. The Municipal Flood Control and Drought Relief Headquarters of Chongqing advocates the normalization of consultations on flood relief and conducts emergency consultations based on the trend of important periods and the important nodes. Such a model not only guarantees the combination of strategic planning and actual conditions but also realizes the agility of the process. This consultation model is very worthy of reference for humanitarian supply chain decision-makers when similar incidents occur. Risk awareness, especially the risk awareness training of disaster-affected people, needs to be improved. Due to the lack of accuracy of the early warning system, there was an early warning of the flood peak crossing in 2019, and many manufacturers chose to transfer their activities in anticipation of the flooding, which ultimately did not occur. When facing the flood peak transit early warning again in 2020, this has caused some people affected by the event to wait and see. This kind of wait-and-see behavior has caused serious losses to many small and medium-sized enterprises.
These research results are based on the flood relief operations in Hechuan District and can also be applied to any large-scale disaster relief activities. Although various disaster situations are different, this model can be followed to operate resilient humanitarian supply chains in different locations and for a range of hazards. This study proposes that transparency of information (B1), coordination and cooperation (B2), capability and strategic planning (C1), agility in processes (D1), and risk awareness (E2) are key indicators for building a resilient humanitarian supply chain. Decision-makers should take optimization measures based on a full understanding of the interrelationships between such indicators.
6. Conclusions and Future Research
Nowadays, frequent disasters have brought great trauma to the economy and people all over the world. However, the humanitarian supply chain, which is of great significance to post-disaster rescue, has to operate in a turbulent environment with uncertain demand. It is therefore apparent that any humanitarian supply chain must be resilient. The main contributions of this research are as follows. First, through literature review and the Delphi method, 18 indicators for building the resilience of the humanitarian supply chain are summarized. Secondly, according to the fuzzy-DEMATEL method, the interrelationships between these 18 indicators are determined, as shown in
Figure 4. Finally, based on the ANP method, this research has derived the key indicators of these 18 indicators, including transparency of information (B1), coordination and cooperation (B2), capability and strategic planning (C1), agility in processes (D1), and risk awareness (E2). The finding of this research will help decision-makers involved in humanitarian supply chain management to visualize the interrelationships between the key indicators of resilience. These results will be of great significance for managers to have a deeper understanding of resilience indicators. Moreover, the approaches developed in this study can be applied not only to flood risk management but also to other disaster situations and in different locations.
There are some limitations to this study. First of all, sudden flood events can quite easily turn affected areas into “islands of information” [
5]. For example, even if a hot air balloon with a mobile base station is launched as soon as possible, there will still be a period when information exchange becomes difficult, if not impossible. This brings a level of stochasticity to the planning of the humanitarian supply chain in emergencies. How to best plan for this stochasticity when seeking to improve the resilience of humanitarian supply chains in such emergencies should be explored in future research. Secondly, rescue workers often have to face different religious beliefs and regional cultures due to the different countries and regions where disasters occur. Humanitarian supply chain managers should also consider the differences in problem handling caused by these backgrounds. Thirdly, rescue workers come from very diverse backgrounds with, for example, different education levels and experiences. How these characteristics impact the resilience of the humanitarian supply chain is an area needing further inquiry and investigation. Finally, this research has focused on the development of the humanitarian supply chain resilience mainly during the response stage. Further research is also needed in different phases of the flood risk management cycle such as the recovery phase.