The data show that out of the 223 publications, the vast majority of 210 (94%) were scientific articles. These were followed by six (3%) proceeding papers and four (2%) review articles, while only three (1%) were book chapters. In terms of article coverage, they were mainly devoted to the following four categories: 50.4% of the publications dealt with water resources; 44.4% environmental sciences; 37.1% multidisciplinary geosciences; 22.6% meteorology atmospheric sciences. The quantity and trajectory of the publications over the given time period indicate the importance of a certain research field within the scientific community. Natural logarithmic trend analysis of the number of publications indicates two phases (Figure 2).

There was a rapid growth in the number of publications from 2014 to 2019. The strong upward trajectory increased from near 0 to around 3.4, indicating exponential growth in publication activity. This phase represents the emergence and rising interest in flood-related research in the climate change context (Figure 2). On the other hand, there is a flattening of the curve and small fluctuations in values in the second phase. However, this phase may indicate that the research community were focusing on deeper specialisation or applications, rather than volume growth (Table 1). This two-phase development trajectory is also very useful for thematic evolution (Figure 10).

In the process of analysing the publications, a total of 1076 authors were identified. The eight most prolific ones (with at least four publications) are shown in Figure 3. These authors accounted for around 15% of the publications. Among them, Stoffel from the University of Geneva had the highest rating with six (almost 3%) articles. Costache, Li, Lu, Pham, Wang, Xiong and Zhang followed, with four scientific articles each. It should be noted that almost half of the leading authors were from China.

Figure 4 shows the ten most productive institutions by number of publications over the 10-year period. During this time, 573 different institutions collaborated, publishing 210 articles on climate change and floods. The results show that more six of the ten most productive institutions were from China, with the Chinese Academy of Sciences taking first place (12 publications), followed by Beijing Normal University (nine publications); the Egyptian Knowledge Bank (nine publications); China Meteorological Administration; Hohai University, China; University of Chinese Academy of Sciences; and University of Geneva (seven publications each); Indian Institutes of Technology, Sun Yat-sen University, and University of Texas System (six publications each).

Among all hydroclimatic hazards, floods are considered the most severe, often resulting in significant loss of human life and significant socio-economic damage. For this reason, countries are paying close attention to the socio-economic and environmental impact of flooding in the climate change context. Global research conducted by (Chen et al., 2023) shows that countries mostly located in the southern regions of the world, in Southeast Eurasia, Africa and South America, should reinvestigate their flood defence standards and mitigation strategies to prepare themselves for an increasing magnitude of 50-year floods from 16.4% to 30.6% at the current global warming level. Figure 5 shows the number of publications on climate change-induced flood research from the ten countries most concerned about this issue. China with 60 publications, the US with 52, India with 22, Spain with 20, England and Italy with 15 each, Pakistan with 12, and France, Iran and Switzerland with 11 are among the leaders.

The above countries have established funds to finance research on flood assessment, sufficient mitigation, and the development of early warning systems. The Natural Science Foundation of China, the National Basic Research Development Program of China, the European Union, and the United States Department of Energy are the leading sponsoring organisations. They both directly and indirectly funded around 35 percent of the research in the 223 publications. According to the authors, based on their affiliations, countries and sponsoring funds, China, Europe and the United States are leading regions in addressing the socio-economic and environmental impacts of climate change-induced floods.

A high number of citations often reflects the significance and originality of particular research. It suggests that the work has made a meaningful contribution to the field and has been widely referenced by other scholars. Total citation (TC) and average annual citation (AAC) serve as common indicators that have continuously changed over time. Table 1 lists the top 10 articles with an AAC of 13 or more.

These articles were published during the period 2017 to 2021. The article the most frequently cited is “Flood susceptibility modelling using advanced ensemble machine learning models “by (Towfiqul Islam et al., 2021). This novel research demonstrated the implementation and measurement of two hybrid ensemble models—Dagging and Random Subspace—integrated with three advanced ML techniques (Artificial Neural Network (ANN), Random Forecast, Support Vector Machine) for flood susceptibility mapping. Table 2 shows the top approaches and techniques used in the articles above, which may guide future research directions (Table 2).

The fact that an article receives the most citations indicates that this type of scientific research is becoming more popular. In the context of climate change, investigating “flood susceptibility” with hybrid modeling and “forecast streamflow” with physically based hydrological modeling are becoming more common topics. In studying these, it appears that the use of GIS, hydrological models, and the analysis of various climate databases through ML algorithms are yielding better results than measurement and observation data.

A high citation rate for an article is also directly related to its quality and novelty, with publication in a prestigious journal increasing the rate even further. According to the analysis results, the 223 publications were cited a total of 5,368 times across 99 journals. Figure 6 shows the top 14 journals that received more than 100 citations published between 2015 and 2024. Nearly 45 percent of the citations are cited in the top five journals: Science of the Total Environment (989 citations), Water (429 citations), Journal of Hydrology (355 citations), Natural Hazards (324 citations), and Geoscience Frontiers (318 citations) (Figure 6).

The journals Science of the Total Environment (Netherlands), Water (Switzerland), and Geoscience Frontiers (China), also shown in Table 1, demonstrate strong performance in both impact metrics and indexing. They are considered Q1 in their category.

Co-authorship and keyword co-occurrence maps were created using data downloaded from WoS via VOSviewer software (version 1.6.17). In bibliometric research, co-authorship analysis examines collaborative relationships between scientists, institutions and countries, as well as between those who work together on scientific projects and publications. It reveals patterns and relationships between authors based on the number of jointly published works and also illuminates the structure and developmental processes of these relationships.

Analysis of the input data resulted in a network of 1,054 authors. Those with at least two publications on the topic of climate change and flooding were included. Twenty-nine elements were distributed across five clusters: cluster 1 (seven items) in red; cluster 2 (seven items) in green; cluster 3 (six items) in blue; cluster 4 (five items) in black; and cluster 5 (four items) in yellow (Figure 7). The figure reveals that Huang, Yang, Lu, Chen and Wang may serve as network hubs with their interdisciplinary and inter-group collaboration, while the green cluster appears to be a central bridge linking several clusters, showing that its members may play key roles. The figure also reveals that Chinese academic communities have strong internal collaboration networks, as all the researchers who link clusters are from China: Pengnian Huang from Nanjing University of Information Science & Technology, China; Hui Lu from Tsinghua University, China; and Jing Wang from the Hydrological Survey Bureau of Tibet Autonomous Region, China.

As a result of the analysis, 1483 keywords were identified, and 51 elements with at least seven occurrences were selected from these. Low-value and rarely used elements were not taken into account. Based on the overall connection strength, four clusters were formed, each selected keyword was matched to a node, and a network map of all the keywords was constructed (Figure 8). The size of the nodes corresponds to how many times they appear, or their frequency. However, the proximity of a node to others indicates certain related topics that frequently occur alongside the subjects being studied.

In Figure 8, showing the co-occurrence of keywords, the network is heavily centralised around the keywords “climate change” and “flash flood”. This indicates that a significant portion of the research in the dataset focuses on these topics and how they relate to other concepts. These keywords are divided into 51 elements across four clusters: cluster 1 in red, with the main keyword being “flash flood”, includes the following words - catchment, China, climate change, disaster, dynamics, Europe, frequency, impact, land-use, management, risk, river, sensitivity, soil-erosion, variability. Cluster 2 in green is primarily focused on “river-basin” with the following topics - AHP, flood hazards, flood risk, flood susceptibility, FR, GIS, hazard, HEC-RAS, ML, RS, risk-assessment, special prediction). Cluster 3 in blue is the “model”, which includes the keywords basin, calibration, climate, flood, HEC-HMS, hydrology, runoff, soil-moisture, system, urbanization, vulnerability. Finally, cluster 4 in yellow, “precipitation”, includes the following: events, extreme precipitation, performance, prediction, simulation, streamflow, temperature, trends, uncertainty. This analysis shows that the main focus of climate change-induced flood research is on the interrelationship between river basins and precipitation, and the application of models (physically-based hydrological, statistical, and soft computing) to management (Figure 8).

The three-field plot provides a clear visualisation of the current academic landscape in flood-related research in countries and their institutions. The analysis confirmed that China is overwhelmingly the most significant contributor. The primary institutions associated with climate change-induced flood research are Beijing Normal University and Hohai University. In terms of geography, China is the dominant country, but there is also notable participation from researchers affiliated with institutions in the United States, Switzerland and Romania (Figure 9).

The three-field plot clarifies and confirms that the main research topics are strongly directed toward “climate change”, “flood”, “flash flood”, “ML”, “GIS”, “RS”, “flood hazard” and “flood risk”. Another method that helps researchers to establish the centre of academic attention is the evolution of research topics, which aids in identifying current and future research directions in their field.

A trend analysis of the number of publications revealed two phases: a rapid growth phase and a stabilisation phase (Figure 2). Referring to these trend phases, ten years of publications were examined in two stages (2015-2019 and 2020-2024) (Figure 10). The analysis revealed that the scientific community focus had moved from broader climate change impacts to more specific and localised issues such as flash floods, precipitation patterns, and the use of advanced modelling and forecasting techniques.

In the second phase of evolution, “precipitation”, “management” and “catchment” were central and well-developed topics, as there is strong interest in connecting “precipitation” to “catchment” and “management”, while “forecast”, “cover change” and “land use” were niche themes during this time period. However, “climate change”, “flooding” and “model” remained as basic themes during the 2020-2024 period. In addition, as an emerging topic, ANN has developed interest in applying ML algorithms to flood-related problems, primarily for forecasting and risk assessment (Figure 11).

Global climate change is affecting the duration and frequency of flash and river floods. A number of authoritative organisations, such as UNDRR and IPCC, and numerous research studies confirm that the negative impact of climate change on the magnitude of floods is inevitable (Nakaegawa & Kobashi, 2025). Research conducted in several of the largest river basins around the world shows that at a 3.0°C warming level, the magnitude of 100-year floods will increase in the Rhine (Europe), Mississippi (US), and Yellow River (China) basins by 13%, 16% and 31% respectively (Huang et al., 2018). Due to the complexity of flood forecasting, several assessment methods and techniques have been developed (Ali et al., 2025), with bibliometric analysis providing both quantitative and qualitative analysis to identify the most modern research directions and methodologies.

Based on the research methodology developed, a bibliometric analysis of works devoted to the topic of climate change-induced flood over a period of ten years was conducted. From 2015, there was a sharp increase in interest in such research, with the annual number of publications stabilising after 2019. This shows that researchers were focusing on deeper specialisation or application development rather than the number of articles over the last five years of the study period (Towfiqul Islam et al., 2021). Most of the publications in this area are scientific articles on water resources and environmental sciences. Among the authors, Costache, Li, Lu, Pham, Wang, Xiong and Zhang are also on the list of most productive researchers, but Stoffel from the University of Geneva has been making significant contributions to research on climate change impacts on floods (Gobiet et al., 2014). Notably, almost half of the lead authors are from China, and the Chinese Academy of Sciences is also a leader in the field, including water resources, earth and environmental science, although the United States and European countries also make significant contributions.

In the context of climate change, in-depth research on the topics “flood susceptibility” and “forecasting streamflow” is being conducted. Physically-based hydrological modelling ((Salas et al., 2018); (Zhou et al., 2019)) and soft computing models ((Bafitlhile & Li, 2019); (Rahman et al., 2021)) of streamflow forecasting research, and statistical (Wiel et al., 2017) and hybrid modelling ((Hong et al., 2018); (Towfiqul Islam et al., 2021); (Waqas et al., 2021)) in flood susceptibility assessment research are commonly used, giving reliable results. In the early stabilisation stage, the most frequently cited article was by (Towfiqul Islam et al., 2021). Their novel research generalised and demonstrated hybrid ensemble ML models for flood susceptibility mapping. This is one of the significant signs that using advanced soft computing and statistical models for flood susceptibility assessment has been introduced into the field. A high citation rate for an article is directly related to its quality and novelty; publication in a prestigious journal increases the citation rates even further. According to the analysis results, the journals Science of the Total Environment (Netherlands), Water (Switzerland) and Geoscience Frontiers (China) are foremost in popularising high-level scientific research on climate-induced flood science.

The network map of keywords shows that there are four clusters, all of which are centralised around flash flooding and climate change. The analysis shows that the main focus of research on climate change-induced flood is geared towards river basins, precipitation prediction, and the use of soft computing model applications (GIS, AHP, ML). Although the results of the three-field plot analysis show that “flood risk” and “flood hazard” are the main research topics, it also confirms that the use of tools such as ML, GIS and RS in the field of climate change-induced flood research is widespread. In addition, the time series evolution of research topics identified that the focus of the scientific community is moving from broader climate change impacts to more specific and localised issues such as changes in precipitation patterns, and the use of advanced models and forecasting techniques (ANN, GIS). A separate analysis of the research directions formed in the five years up to the end of 2024 indicates that in climate change-induced floods, modeling remains a basic theme, and that “precipitation”, “management” and “catchment” are also central and well-developed topics. However, the results also indicate that ANN should be introduced as an emerging tool for forecasting and risk assessment.