Evolutionary Study of Three Decades Literatures Related to Disaster Management Agency and Public Policy: A Bibliometric Study

This study is aimed at examining scientific literatures written by international authors focusing on disaster management agency topic that are related to public policy from the first paper published in 1995 to 2022 using bibliometric study employing a database from Scopus. To visualize the results, this study utilizes VOSviewer version 1.6.17. The results showed that this topic emerged in Thailand which focused on roles of local level disaster management agency in carrying out rehabilitation and reconstruction activities in district level. However, Federal Emergency Management Agency, or the United States Disaster Management Agency, was considered to be the most researched disaster management agency. Besides, open-access social science articles dominated the type of publications in this research area. Moreover, the United States and its authors were perceived as the most productive and influential country and authors. Meanwhile, the United Kingdom was recognized as the most influential sources title with Journal “Disasters” served the most productive and influential source title. Besides, the affiliation among Austria, Nigeria, South Africa, and the United States were acknowledged as the most productive and influential affiliations. The analysis also indicated that this topic developed and strongly connected to disaster and disaster management research areas. Considering the comprehensive nature of bibliometric analysis, this study may help scholars get a broad view of the relevant issues, while at the same time identifying reputable publications worth studying for reference.

those two approaches and the significant roles of central as well as local level disaster management agencies have led to curiosity about how scholars view issues related to disaster management agency as well as public policies governing the issue and reflect them in their scientific publications. Consequently, a comprehensive review of the relevant studies becomes essential. In this case, evolutionary studies of the field will enable scholars as well as researchers to examine and investigate such type of study (Hou et al., 2021).
Meanwhile, research on disaster management has been widely expanding from time to time and thus encouraged scholars to carry out studies on disaster management and its related issues using bibliometric studies. This is because bibliometric studies allow researchers to portray and visualize a large amount of scientific data in a specific field which will enable researchers to reveal evolutionary variation and investigate the emerging issues in the concerned area (Donthu et al., 2021). Those disaster management-related bibliometric studies, among others, focus on the use of big data to manage disasters (Akter & Fosso Wamba, 2017), disaster management policy in China (Zhang et al., 2018), the urgency of simulation to better prepare for natural hazards (Barnes et al., 2019), the use of emerging technology in disaster management (Du et al., 2017) (Vermiglio et al., 2022), community resilience (Demiroz & Haase, 2019)(Q. Yang et al., 2021), epidemics , the relationship of law and disaster management (Gurpur et al., 2021), as well as decision making in an emergency situation (Hou et al., 2021).
Besides, although disaster management has witnessed rapid development and there have been a large number of publications in the past decades, there are only a few studies emphasizing disaster management agency and policy so far. For example, Husna et al. (2022) elaborate on the preparedness of disaster mitigation agencies and the policies they implement in high-vulnerability zones in Aceh, Indonesia using a cross-sectional study. Clark-Ginsberg et al. (2021) emphasize how inequalities in disaster management agency function and policy may result in creating disaster risks for vulnerable and marginalized people. Meanwhile, Grove (2014) critically analyzes the participatory disaster resilience program and policy in Jamaica.
All those publications have provided valuable insights and research agendas worthy of academic attention.
However, none of the above disaster management-related bibliometric studies focuses on disaster management agency and public policy. Moreover, the studies carried out by Grove (2014), Clark-Ginsberg et al. (2021), and Husna et al. (2022) also lack a bibliometric analysis. Thus it is interesting to navigate the study development of the area of disaster management agency and public policy using bibliometric study since the field remains unexplored. Therefore, the purpose of this article is to review disaster management agency and public policy literature with the help of bibliometric analyses in a systematic way.

Methods
The bibliometric study deals with incorporating different structures, methods, and tools to learn and investigate a huge amount of scientific data while at the same time generating high-impact research (Ponce & Lozano, 2010) (Donthu et al., 2021). Bibliometric allows researchers to scrutinize trends in the retrieved data based on specific keywords the researchers utilize as well as visualize the emerging trends and knowledge distribution from various detail (Barnes et al., 2019)(Q. Yang et al., 2021. Therefore, bibliometric studies may deliver a one-stop overview, identify unfilled space in certain research areas, acquire research novelty, and place a study in a certain area of research (Donthu et al., 2021). Besides, a bibliometric study on literature related to the policy will enable researchers to uncover policy fundamental shifts and trajectories (Zhang et al., 2018). Chen & Xiao (2016) state that in a bibliometric research, when the research is aimed at revealing the details of a domain's major research topics and their relations at the micro-level, the selection of a few keywords to represent important research themes in the domain is necessary. Consequently, since this research is intended to reveal discourses on disaster management agency and public policy, three important keywords are selected, i.e. "disaster management", "agency", and "policy". Although considered as being a popular keyword, "disaster 129 management" was chosen as the first keyword since this kind of general keyword will be useful to provide a rough overview of a scientific discipline (Chen & Xiao, 2016). Further, we add two keywords "agency" and "policy" to explore how international publications reflect the relations between agencies operating in disaster management and policies related to the issue. In this case, we use disaster management terminology developed by the United Nations Office for Disaster Risk Reduction, i.e. the organization, planning, and application of measures to prepare, respond to, and recover from disasters (UNDRR, 2015). Further, agency refers to a government agency, both at the central and local levels, that is responsible for the management of disasters, including preventing disasters, mitigating disasters, recovering the impact of disasters, and rebuilding the society of an occurred disaster (J. Yang et al., 2009) . Meanwhile, policy is simply defined as an answer to public issues (Cooper in Islamy, 2017).
This bibliometric study is carried out with a certain protocol that is intended to conduct science mapping. The When carrying out this bibliometric study, we extract a literature search from Scopus since it comprises the most extensive structured set of data of citations and abstracts from peer-reviewed scientific and written literature, including articles from reputable scientific journals, book reviews, and conference proceedings (López-Muñoz et al., 2018). Scopus also covers a broad range of fields, such as social science, environmental science, medicine, business, management, accounting, energy, and so on. As seen in Figure 1, the search query was carried out using the keywords "disaster management", "agency", and "policy" in the title, abstract, and keywords fields of the database. We used all year period until 2022 to select data since this was valuable to reveal the very first and latest publications. To this step, we got 315 documents. Then, we refined the literature screening by limiting the subject areas into social science, environmental science, earth and planetary science, and decision science. The document types were also refined to only articles, conference papers, and book chapters. Only those three types of documents published in journals, conference proceedings, and books using English were considered in the analysis. To ensure the relevance of the documents to the analysis, a manual screening that took into account titles and abstracts was carried out. This further screening resulted in 206 documents. Then, we chose VOSviewer to help visualize and analyze the retrieved data since to date VOSviewer had been extensively utilized for the bibliometric study of several areas of research (van Eck & Waltman, 2010) (Hou et al., 2021). To comprehend the inquiry and visualization, metrics from Scopus were also employed.

General overview
Open-access indicates that peer-reviewed journals and articles are online and available without restriction. The documents belong to the "green" category (approved to be published or have been published and can be found in the repository), 18 documents belong to the "gold" category (ready in only open-access journals), 14 documents belong to "bronze" category (issued version of written documents which are accepted for publication), and the rest eight documents belong to "hybrid gold" category (available in journals that offer authors option of publishing open access) as seen in Figure 2.
Further, although concern related to disaster and disaster management studies grows rapidly, not much research can be found in the field of disaster management agency and public policy from Scopus. We figured out that "disaster management" was a popular or general keyword and had become one of the most researched themes in international journals. We found 23,733 publications applying "disaster management" as a keyword in the Scopus database. This number decreased to 1,957 publications when the keyword "agency" was added. Finally, when the last keyword "policy" was joined, there were only 315 publications. Furthermore, the analysis of the publications shows that the research in the field of disaster management agency and public policy was initiated in 1995 through an article entitled "Rehabilitation planning for flood-affected areas of Thailand: Experience from Phipun District". This article was published in Disasters, Volume 19, Number 4, pages 348 -355, and focused on a disaster management agency's tasks at a very small area (i.e. district level) and its challenges that wrecked the rehabilitation activities in disaster-affected areas. However, this topic seems to have failed to attract researchers' attention to this field since the second publication indexed by Scopus was just released three years later which The publication on the field of disaster management agency and public policy in Scopus is also diverse. Articles dominate this research area (81.1%) followed by conference papers (14.1%) and book chapters (4.9%). Those documents are published in various publications, including journals (167 documents), conference proceedings (29 documents), and books (10 documents) as seen in Table 1.  Figure 4. Year

Citation analysis
This type of analysis presumes that citations portray a conceptual relationship between publications when there is a citation from another or other publications (Appio et al., 2014). Hence, a publication is considered influential by the total of citations it gets. Meanwhile, the quality of the paper is derived from the number of citations Consequently, the United States is considered the most influential and productive country in this research area.
However, South Africa gets the first rank in the quality of paper (see Figure 5).
Studies carried out by countries around the globe also connected and create a network. The citation network analysis resulted by VOSviewer shows that there are three clusters of the country of origin. The first cluster This cluster connects to the second cluster which comprises the United States, Brazil, Netherlands, and Romania where the United States leads the cluster. Moreover, the third cluster exists and only consists of Australia. Overall, the United States dominates all those three clusters. This is relevant to the discussion on the most productive and influential country of origin seen in Figure 5. Besides, the citation overlay visualization also revealed that research areas on disaster management agency and public policy emerged from Thailand and then spread to the United Kingdom, the United States, the Netherlands, and Australia. Meanwhile, this field of study is relatively a new discussion for Nigeria, Brazil, and Romania as seen in Figure 6.
Meanwhile, in terms of source titles, the research area of disaster management agency and public policy is involved in 121 source titles comprising journals, conference papers, and books. Disasters from the United Kingdom has taken the first place in the number of publications (16 documents) and citations (522 citations), and thus, is considered to be the most productive and influential source title in this research area. Whilst, World Development ranks first for quality of paper (188) as seen in Table 2  In addition, the analysis also reveals that there are 581 authors, both working individually or in a team, who have published their documents related to disaster management agency and public policy discourse during 1995-2022.
Those authors produce one to three documents during that period. Among those authors, there are 105 authors whose publications have not been cited by another publication. Coleen Vogel's publications are considered to be the most influential publications in this research area since they are cited most frequently (see Table 3). Moreover, although producing the most publication (three documents), Laurie A. Johnson gets fewer citations (58) and thus does not belong to the top ten of the most influential authors. Additionally, the United States authors also dominate the field.  Table 4. Moreover, within the top ten cited articles, there were only two articles published in the last ten years, entitled "Agency, affect, and the immunological politics of disaster resilience" (Grove, 2014) with 104 citations and "Resilience and Justice" (Fainstein, 2015) with 100 citations. This indicates that the research area develops slowly, and thus, scholars tend to cite old publications more than the latest ones.

Co-authorship analysis
Initially, scientific work is individual work. However, as the increasing of the complexity of the research topic, research area, and methodology as well as the rapid development of theory and science, collaboration or affiliation among academics becomes more and more popular, especially when preparing a joint paper, which is known as co-authorship (Acedo et al., 2006). And thus, this type of analysis reveals interaction among researchers in a certain research area (Donthu et al., 2021).
In terms of country of origin, the bibliometric analysis of co-authorship shows that there are nine clusters of countries of origin. Cluster 1 consists of eight countries comprising Australia, Finland, New Zealand, Romania, consists of three countries comprising China, Oman, and Pakistan. In terms of productivity and citation, cluster 5 is considered the most productive and influential affiliation. Cluster 5 affiliation also has the highest quality of the paper. The strength of each country cluster is seen in Figure 7.

Co-occurrence analysis
Scholars carry out keyword co-occurrence analysis to disclose underlying research communities of a certain field of the research area and the potential relationship between those fields (Emich et al., 2020). Meanwhile, Zou et al. (2018) state that keywords convey the central idea of an academic publication. Further, Pesta et al. (2018) add that keywords are considered to be the most significant terms in an academic publication. Therefore, cooccurrence analysis can help scholars investigate the content of a document (Donthu et al., 2021).
There are all 1,486 keywords from 206 publications related to disaster management agency and public policy. The VOSviewer co-occurrence analysis is carried out using the full counting method, all keywords, and the lowest limit of occurrences of a keyword is 10 times. The analysis shows that there are four clusters with 27 nodes (see Figure   8). Cluster 1 (red) includes 12 items and focuses on the keywords "disaster management (152 occurrences) and risk assessment (33 occurrences)". Cluster 2 (green) consists of nine items and focuses on the keywords "disaster (34 occurrences) and human (26 occurrences)". Cluster 3 (blue) consists of three items and focuses on the keywords "disasters (50 occurrences) and disaster prevention (30 occurrences)". Cluster 4 (yellow) consists of three items and focuses on the keyword "flood" (13 occurrences). Moreover, the keywords "disaster management" and "disasters" dominate all keywords. Meanwhile, keywords related to "policy" and "organization and management" belong to cluster 2. Further, the overlay visualization reveals that the keywords "organization and management", "relief work", "article", and "human" have been employed since early 2011. Whilst keywords "policy", "local participation", "mitigation", and "risk management" emerged after 2013. On the contrary, the keywords "disaster prevention", "climate change", and "adaptive management" emerged after 2016.
In terms of author keywords, there are 640 keywords employed by authors in 206 publications related to disaster management agency and public policy. The analysis employs a full counting method, author keywords, and the lowest limit of occurrences of a keyword is four times. The analysis shows that there are sevens clusters with 25 nodes (see Figure 9). Cluster 1 (red) encompasses five items and focuses on the keyword "climate change" (12 occurrences). Cluster 2 (green) involves four items and focuses on the keyword "disaster recovery" (6 occurrences). Cluster 3 (dark blue) includes four items and focuses on the keyword "disasters" (9 occurrences).
Cluster 4 (yellow) consists of four items and focuses on the keyword "vulnerability" (10 occurrences). Cluster 5 (purple) covers three items and focuses on the keyword "disaster management (28 occurrences). Cluster 6 (light blue) consists of three items and focuses on the keyword "resilience" (8 occurrences). Last, cluster 7 (orange) consists of two items and focuses on the keyword "disaster" (19 occurrences).
(a) (b) Figure 8. Overlay and network visualization of all keywords co-occurrence related to disaster management agency and public policy. Before 2010, the authors' keywords represented area study of "bangladesh" and "developing countries". During 2010-2012, the authors' keywords focus on "india" and "risk". Further, during 2012-2014, the authors' keywords emphasize "disaster response" and "disaster planning". Meanwhile, during 2014-2016, the authors' keywords provide concern on "vulnerability", "adaptation", "resilience" as well as "disaster management". Then, during 2016-2018, the authors' keywords focus on "mitigation", "fema", "disaster risk reduction", and "disaster preparedness", in addition to "policy". Furthermore, from 2018 onwards, the authors' keywords focus on "emergency management", "social vulnerability", and "preparedness" as seen in Figure 10. Moreover, the United States' disaster management agency or known as Federal Emergency Management Agency (FEMA) has been included in several authors' publications and indicates that its existence is widely recognized and thus attracts scholars' concerns to investigate its roles and policy in disaster management.

Conclusions
Academics' concern for the disaster management agency and public policy is reflected in the indexed Scopus database during 1995-2022. Hence, this bibliometric study has revealed several captivating phenomena. First, initiated in Thailand in 1995, discourse on the topic stumbled in its development. A growing concern on the issue then re-emerged in 2008 and continuously developed so far. Second, although dominated by social sciences, this topic was widely discussed and thus covered diverse subject areas. Third, albeit the United States' first rank in productivity and influence in this research area, South Africa ranked first for the quality of the paper. Fourth, notwithstanding that the United States authors were the most influential and productive ones, the United Kingdom's journals defeated the United States in terms of source titles. Fifth, citations from old publications still dominate the citations of the articles. Sixth, the affiliation of the United States, Austria, Nigeria, and South Africa results in the most influential and productive publications as well as the highest quality of the paper. Seventh, although several publications have discussed disaster management agencies, FEMA was the only disaster management agency mentioned in the keywords. Last, the development and evolution of keywords related to this research area are also relevant to the shifting concern and paradigm of disaster management.