Globally, floods have become more frequent in recent decades, a finding that is consistent with observations of anthropogenic climate change (Stott, 2016). Meanwhile, floods are becoming increasingly severe, leading to enormous economic loss and death (Lehner, Döll, Alcamo, Henrichs, & Kaspar, 2006; Luo et al., 2019). Despite substantial flood prevention efforts, the resultant loss of human life and property persist at high levels, with floods accounting for 34 and 40%, respectively, of all global natural disasters in quantity and losses (Lyu, Shen, Zhou, & Yang, 2019; Petit-Boix et al., 2017). Flood hazard assessment (FHA) is considered an important means of decreasing flood disaster losses and ensuring the healthy and sustainable development of human society. Therefore, research in this area is imperative to meet human needs (Wang et al., 2015).
With growing interest in FHA, it is essential to keep track of the key trends and the direction it is heading. FHA, a qualitative or semi-quantitative method, is focused on the combined influence of disaster-inducing factors and environments (Hallegatte, Green, Nicholls, & Corfee-Morlot, 2013; Mu et al., 2021). Comprehensive flood hazard assessment has generally been used in flood insurance, floodplain management, disaster evacuation, disaster warning, disaster evaluation, flood influence evaluation, the improvement of the public's flood risk awareness and understanding of flood disasters (Luo et al., 2020; Zou, Zhou, Zhou, Song, & Guo, 2013). In recent years, FHA research has increased in quantity and sophistication because more new technologies are integrated, such as remote sensing (RS), GIS, and open data (including social media data, volunteer geographic information) (Fang, Hu, Shi, & Zhao, 2019; Sajjad, Chan, & Kanwal, 2020; Thaler & Hartmann, 2016).
Most FHA-related studies were focused on modeling and proposing a new hydrological risk framework. For example, some scholars developed and examined a novel approach for assessing flood risk in river catchments in a spatially consistent way (Falter et al., 2015; Zhu, Luo, Zhang, & Sun, 2020). Their research was based on a set of coupled models that generated long time series of spatially consistent meteorological fields, through the subsequent models, into long time series of flood damage deriving flood risk estimates directly from the simulated damage. A similar study proposed a univariate and copula-based bivariate hydrological risk framework for advancing traditional flood risk analysis (Guo, Chang, Wang, Huang, & Zhou, 2018). Their research considers the design flood of downstream hydraulic structures and instream flow requirements for sediment transport.
Some scholars also offer a multi-criteria index approach to classify potential flood risks at the river basin scale (Toosi, Calbimonte, Nouri, & Alaghmand, 2019). Their research spatially analyzes seven parameters (runoff coefficient, elevation, slope, distance from the drainage network, rainfall intensity, soil erosion and land use), combining the information in the Flood Hazard Index. In another study, based on long-term monthly streamflow data from 9 gauging stations covering the period of 1960–2014, nonstationary analysis was used to quantify flood frequency and flood risk (Sun et al., 2018).
A case study introduced an approach that employs a future land-use simulation model for scenario-based 100-year coastal flood risk assessment (Lin, Sun, Nijhuis, & Wang, 2020). The research explores the possible implications of future land-use changes due to the ongoing urbanization on projected environmental changes (sea-level rise, storm surge, and land subsidence). Moreover, some scholars provided a risk assessment framework for flash floods induced by tropical cyclones under climate change scenarios (Zhang, Wang, Chen, Liang, & Liu, 2019).
In addition, several recent comparative studies reported promising results using machine learning methods in FHA-related research. Some scholars presented a semi-supervised machine learning model (the weakly labeled support vector machine) to assess flood-prone areas with limited flood inventories (Huo et al., 2020; Zhao, Pang, Xu, Peng, & Xu, 2019). Their study is a novel attempt to introduce a semi-supervised machine learning model in urban flood susceptibility assessments and proves that it has good performance in FHA. Finally, some studies provided state-of-the-art ensemble models of boosted generalized linear model and random forest, and Bayesian generalized linear model methods for higher performance modeling (Hosseini et al., 2020).
As comprehensive research is limited, it is vital to analyze global trends in FHA-related publications. This work starts with the three main pillars of scientific publication: who writes this article (authors), who publishes magazines on the topic (sources) and what was written (documents) (Rodriguez-Soler, Uribe-Toril, & Valenciano, 2020). This study provides a detailed analysis performed for FHA conducted during 2000–2020 using the bibliometrix R-tool. The bibliometric analysis entails several steps, including numerous analyses and mapping software tools, usually available only under commercial licenses, resulting in a very complex process (Guler, Waaijer, & Palmblad, 2016). The complexity of the process reduces the potential and the possibilities of bibliometrics, especially for researchers without general programming skills (Aria & Cuccurullo, 2017). In recent years, bibliometrics has assembled specialized software into a comprehensive and organized data flow through automated workflows, suitable for multi-step analyses using different types of software tools (Guler et al., 2016; Munim, Dushenko, Jimenez, Shakil, & Imset, 2020). Consequently, we can investigate, in detail, the leading edge of academic articles, publication data, and information about authors, journals, countries, and institutions with significant contributions in FHA using the bibliometrix R-tool (Shikoh & Polyakov, 2020).
Compared with previous literature reviews, the literature search process detailed here is robust, repeatable and transparent. Published studies related to FHA and the analysis of mapping the conceptual structure of this field and determining future research avenues were reviewed. The remainder of the article is structured as follows. First, the methods and databases are described. Second, the investigation results are obtained through descriptive and bibliometric analysis. Finally, limitations and suggestions for future research are provided in the conclusions section.