Natural Hazards Review

Household displacement following disasters has become endemic in many areas worldwide, affecting at least 265 million people between 2008 and 2018. Although this figure includes short-term and potentially life-saving evacuations, there is ample evidence that not all households return after the emergency phase. Protracted displacement is associated with particularly negative consequences for the affected households and community. Yet, existing data on displacement duration are limited, and only a few disaster recovery models incorporate the multitude of factors beyond housing damage that are known to influence household return. This review synthesizes the current literature on disaster-induced displacement, including key terminology and context, the determinants of household return decisions, existing model-based approaches, and opportunities for future research. The identified key determinants of household return can be broadly organized into the following categories: physical damage to the built environment, psychological and social phenomena (e.g., attachment to place, social networks), household demographics (e.g., tenure, socioeconomic status), and pre- and postdisaster policies (e.g., housing reconstruction approach, design of aid programs).

Incarcerated people are some of the most vulnerable and at risk in our society. One understudied source of this high risk is the exposure of incarceration facilities to environmental hazards. This study addresses this gap by developing and applying spatial analyses to evaluate the exposure of incarceration infrastructure to wildfire, flood, extreme heat, and landslide hazards. First, we classified the exposure to these four hazards for all known incarceration facilities in the state of Colorado. Next, we statistically evaluated various demographic and infrastructure metadata to distinguish between elevated- and low-exposure facilities for each hazard type. Finally, we compared the exposure of incarceration infrastructure with the state’s kindergarten–12th grade (K-12) education infrastructure. We found that approximately 75% of studied incarceration infrastructure is in the elevated exposure group for at least one of the four categories studied. We did not find a significant correlation between landslide risk and the location of incarceration infrastructure. We found racial disparities—two of the hazards having greater exposure in facilities with higher rates of incarcerated people identifying as Black (in the case of extreme heat) and Hispanic or Latino (in the case of floods)—whereas wildfires and landslides did not appear to pose significant equity concerns. We found education and incarceration facilities to have statistically significant differences in exposure for one hazard, floods. The results of this study highlight the disproportionate risk exposure experienced by Black and Hispanic or Latino incarcerated people. The quantification of facility-level exposure to hazards is timely and poised to guide strategies that increase the resilience of the incarcerated population in Colorado and beyond.

Flooding has become a devastating natural hazard worldwide, posing a substantial hazard to civilization and leading to socioeconomic losses. Urban flooding in India is on the rise due to unplanned urbanization, uncontrolled population growth, and changes in land use. For this analysis, the urban region of Dhanera (Gujarat state, India) has been chosen as a study area to prepare a flood hazard map in a geographic information system (GIS) environment using a multiple-criteria decision making (MCDM) technique. This study includes qualitative analysis compound factor (CF) and analytical hierarchy process (AHP), and quantitative analysis (fuzzy logic) using eight flood-affected criteria (slope, distance from the main river, land-use land cover (LULC), soil, drainage density, elevation, geomorphology, and proximity to stream confluence). MCDM techniques examined the weight of each criterion. The historical flood depth data have been collected via field visits or published articles for validation. Flood hazard maps have been classified into very high, high, moderate, low, and very low hazard zones. In each MCDM method, Ward 1 and Ward 2 are very highly and highly hazardous, respectively. The flood hazard map using fuzzy logic and inundation map matches very high, high, and very low categories. So, fuzzy logic is the best method to quantify flood hazards for Dhanera. This study provides a low-cost and time-efficient tool that assists local administration in prioritizing the preservation and control of floods, further helping the decision makers in mitigation planning to reduce damages.

One of the most remarkable factors affecting the flood susceptibility in watersheds is the change in land use and land cover (LULC). Moreover, machine learning (ML) methods showed reliable efficiency in predicting LULC and preparing flood susceptibility maps. Therefore, this study aims to predict LULC of the future using land change modeler (LCM), and then investigate the effects of LULC changes on the flood susceptibility map using ML algorithms in the Talar watershed, Iran. In this regard, LCM model based on artificial neural network (ANN) and Markov chain analysis used to predict LULC in the future. Then, the effect of LULC changes on flood susceptibility was investigated using some ML algorithms, including ANN, logistic regression (LR), random forest (RF), and support vector machine (SVM). These four algorithms were used for the first time simultaneously to generate flood susceptibility maps in the present and future. Nine influencing factors, including altitude, slope, distance from rivers, land use and land cover, topographic wetness index, stream power index, profile curvature, rainfall, soil type, and 200 flood and nonflood locations, were utilized to create the flood susceptibility map. The selected ML models were assessed using root mean square error (RMSE) and coefficient of determination (${\mathrm{R}}^2$). RMSE values showed high performance of SVM, LR, RF, and ANN models with values of 0.125, 0.175, 0.1920, and 0.2834, respectively. Furthermore, ${\mathrm{R}}^2$ values of those models were computed as 0.9373, 0.8773, 0.8667, and 0.8525, respectively. Overall, the findings demonstrated the performance of the LCM method in forecasting the LULC map and the capability of the selected ML algorithms to produce flood susceptibility maps in the study area. In cities and watersheds all over the world that are dealing with LULC changes and a possibility of flooding, the findings and algorithms used in this study can be applied. This aids officials and decision makers to make more effective and efficient decisions in flood prevention and management.

The reasonable spatial allocation of shelters in key urban areas and the improvement of the evacuation efficiency are critical agendas in the field of earthquake mitigation and resilient cities. Based on the conventional shelter location model, which ignores blocked roads and has deficiencies on the computational accuracy of evacuation demands and evacuation environment, this study proposes an advanced methodology for the spatial allocation of shelters considering the blockage effect on roads under earthquakes. First, based on the elastic-plastic analysis of building groups, the damage condition of each building under different seismic intensities and the influence range of falling components on evacuation roads are determined. Second, the evacuation demands are identified with an accuracy at the building scale. Third, this study proposed the hybrid cellular automata evacuation model, which combines the merits of the microscopic evacuation and macroscopic evacuation models. The proposed model not only reflects the evacuation behavior of refugees and but also has acceptable computational efficiency. Fourth, the shelter quantity is taken as the optimization objective to complete the spatial allocation of shelters via a cyclic verification and simulations for regional evacuations. Finally, Yangbi City, Yunan Province, China is employed as a case study to prove the effectiveness of the proposed method.

There is a limited but growing body of literature on the impacts of the COVID-19 pandemic on academic faculty, scholars, and researchers in the US and around the world. In this article, we present findings of a US-based study conceived in 2020 at the onset of COVID-19. In it, we focus on ways in which the pandemic has affected the professional and personal lives of hazard and disaster researchers from the social sciences, public health, engineering, and other fields who started to examine social dimensions of COVID-19 at the beginning of the pandemic. Data are drawn from a systematic, qualitative, longitudinal study in which we gathered data at two points in time across approximately 18 months between the summer of 2020 and spring of 2022. Thirty interviewees in the first phase of the study and 18 interviewees in the second phase shared their experiences navigating the challenges of conducting pandemic-related research while themselves working in the pandemic environment. Through their rich narratives, study participants provided a range of perspectives. With respect to their professional lives, they described issues associated with social isolation, working from home, using digital platforms to conduct research and business, and shifts to online teaching, among other things. At a personal level, they discussed challenges of childcare and caregiving, as well as living in a generally stressful environment. We conclude by offering a number of suggestions for policy-makers and decision-makers to consider in the event of future events such as COVID-19.

Infrastructure projects require billions of dollars, but they are destroyed in seconds due to strong ground motions during earthquakes. The disaster mitigation cell of any nation seeks sustainable and resilient infrastructure schemes to reduce the catastrophic effects of any seismic hazard. This study attempted to carry out detailed seismic microzonation of the Jammu Region (JR), which is located in the northwestern part of the Himalayas. To accomplish this, seismic hazard analysis was carried out with a primary focus on site characterization using geophysical field testing. For the proposed seismic zones, the performance and vulnerability of tunnels were evaluated in various seismic environments. Further, the seismic risk of the 345-km-long Udhampur Srinagar Baramulla Rail Link (USBRL) project was assessed for serviceability in post-seismic conditions. The proposed zonation, risk matrices, and route maps will be useful for auditioning existing and proposed tunneling projects in this area. This will pave the way for earthquake-resistant design of sustainable infrastructure projects in seismically active areas like the Himalayas.

Abstract

Earthquakes are catastrophic natural disasters, and the built environment requires a quick recovery after such events. Estimating the loss from an earthquake can be complex as unique factors, in addition to conventional repair work, that need to be considered during the postperiod stage. Previous research has been able to identify 11 factors that impact the postearthquake cost estimation for the earthquake damage repair work (CEEDRW) process. Current loss estimation methods such as PACT, SLAT, and SP3 do not incorporate the impact of all these factors. This paper identifies the factors that impact different stages of traditional CEEDRW: precontract estimates, estimates during repair work, and cost calculations after construction. Discussion around methods to adopt these modeling factors into CEEDRW is also presented. A two-stage data collection method was adopted for this paper. Data were collected through semistructured interviews using 19 participants and a questionnaire survey using 92 participants who have experience in the CEEDRW process. The interview data analysis emphasized the usage of 11 factors at the three stages of estimation. Furthermore, survey results identified the most suitable methods of incorporating each factor in CEEDRW. Existing estimation tools used for CEEDRW do not consider a number of factors. This paper suggests the most suitable methods of considering these into an estimation tool that aims to enhance the accuracy and reliability of the CEEDRW process. The methods identified in this paper can be used in developing more reliable future cost estimation models used to determine postearthquake damage repair work. The data were collected from professionals who are experienced in housing repair work in New Zealand. The results vary for other types of geographic regions and construction methods.

Practical Applications

Currently used postearthquake cost estimation processes for earthquake damage repair work have a lot of drawbacks. To achieve the final aim of developing a standard postearthquake cost estimation model for earthquake damage repair work that is more accurate and faster than current ones, this research achieved one of the objectives of identifying methods that can be used to mitigate problems of the current cost estimation process. The research identifies methods that can be used to include factors that were not fully considered in current postearthquake cost estimation processes. The results of the research will help with increasing the accuracy of the upgraded model. The next step in this process is the develop the upgraded standard postearthquake cost estimation model for earthquake damage repair work. Even with the current findings, cost estimators, builders, and insurance agencies can improve their cost estimation process by including the impact of the factors through the suggested methods. Governments can implement changes to standard construction contracts and policies so that the construction industry can positively manage the cost-varying dynamics of earthquake damage repair work.

Abstract

Landslides often cause significant economic and human losses, and therefore landslide susceptibility mapping (LSM) has become increasingly important. Accurate assessment of LSM is important for appropriate land use management and risk assessment. The aim of this study is to define and compare the results of applying the random forest (RF) and logistic regression (LR) models for estimating landslide susceptibility, and also to confirm the accuracy of the resulting susceptibility maps in the Ordu-Bolaman River micro-basin. The study area was selected because it is one of the most landslide-prone areas in Türkiye. First, a total of 231 landslide locations were identified. Then 12 landslide-influencing factors were selected to generate landslide susceptibility maps. These maps were produced using the landslide influencing factors based on the RF and LR models in a geographical information system (GIS) environment. Finally, area under the curve (AUC) analysis, sensitivity, specificity, and accuracy were considered to assess and compare the performance of the two models. In addition, the maps were retested with large landslides not included in the training and test data sets, using general accuracy criteria. The results of the present study will be helpful for future landslide risk mitigation efforts in the research area.

Practical Applications

Landslide susceptibility mapping is crucial in adequately mitigating hazards and provides guidelines for landslide-prone areas to avoid hazards in the future. Therefore, landslide susceptibility assessment is of the utmost significance to ensure the safety of human life, mitigate the negative impacts on the economy, and prevent landslide hazards. Government agencies, policymakers, local authorities, and urban planners can use landslide susceptibility maps to plan effective management strategies for landslide prevention and mitigation and to make informed decisions regarding land use zoning and development. By identifying areas with very high and high landslide susceptibility, constructing critical infrastructure and buildings in hazardous zones can be avoided, reducing the risk of damage and loss during potential landslide events. Furthermore, landslide susceptibility maps play a crucial role in disaster risk management and emergency response planning. Authorities can use these maps to prioritize areas for early warning systems, evacuation routes, and disaster response teams. Farmers and landowners can benefit from the maps by being made aware of landslide-prone areas on their property. In addition, insurance companies can use landslide susceptibility maps to assess the risk of landslides in certain regions and adjust their insurance policies accordingly. By implementing these practical applications, landslide susceptibility maps can have a significant impact on reducing the vulnerability of communities and infrastructure to landslides, ultimately contributing to safer and more resilient environments. Consequently, this study is an example of landslide susceptibility mapping efforts for an agriculturally important area.

Emergency managers and alert and warning originators require tools to build effective warning messages. The purpose of this research is to develop one such tool—a Warning Lexicon—that can be used to construct consistent warning messages for Wireless Emergency Alerts. Specifically, the Warning Lexicon systematically establishes a common set of statements about hazard impacts and their associated recommended protective actions that can be used to quickly write effective warning message contents. Thus, the Warning Lexicon allows practitioners and risk communicators to (1) write effective warning messages at the time of the threat, (2) reduce message issuance delay, and (3) develop templated messages as part of their preparedness process. We built the Warning Lexicon through a theoretically informed, multiphased mixed methods process of content analysis and subject matter expert review to verify the accuracy of lexicon statements and validate the language used to instruct message receivers about protective actions. The resulting content incorporated into the Warning Lexicon includes 48 hazards, their associated impacts, and 112 protective action statements spanning atmospheric, technological, biological, and human-induced events. The result of this project is a comprehensive, theory-based, expert-informed, and practitioner-reviewed resource to support the composition of warning messages for imminent threat events.

Metro construction projects are accompanied by numerous unpredictable risk factors and risk events. Traditional safety risk analysis of metro construction projects focuses on risk factors of a particular risk event, such as collapse and falling. However, safety risk factors interact with each other and cause multiple risk events, forming a large transmission network. To reveal the influential relationship between risk factors and events in metro construction projects, the study proposes a safety risk transmission network model based on a factor–event network. First, safety risk factors and risk events were identified based on the reports of metro construction accidents in China. Then, a complex network containing 64 risk factors, 56 risk events nodes, and 533 interacted relationships was built. Finally, critical safety risk factors and risk events were obtained through quantitative assessment of the network topological feature quantity, and crucial transmission paths were identified. It was found that interactions between nodes were quite frequent risk factors of rainfall, geological voids, and illegal operations; and risk events of subsidence or cracks, earth and rock collapse, and death were crucial nodes in the network. These crucial nodes have a large influence on the network and play a pivotal role in risk transmission. Thus, project managers can develop a comprehensive long-term risk management plan based on critical factors and events and take measures to reduce losses by analyzing the risk transmission chains and taking measures to safeguard the neighboring nodes. It is of great significance for accident prevention, control, and improving safety risk management.

Abstract

Tornadoes represent a significant threat to life and property and tend to evoke protective action in most people. Studies have suggested that many people travel to the nearest storm shelter or flee the area, rather than sheltering-in-place as recommended by the National Weather Service (NWS). While shelter-in-place is the recommendation of the NWS, for tornado safety, few studies have quantified the risk reduction when compared to traveling to a storm shelter or fleeing the area. To address this knowledge gap, we developed an agent-based model, the tornado warning-induced shelter, travel, and evacuation response agent-based model (TWISTER ABM), to simulate protective action behaviors in the city of Norman, Oklahoma, under eight protective action scenarios including: (1) everyone who responds to the warning (responders) seeks refuge in the nearest sturdy building (seek refuge), seeks shelter in a FEMA-rated shelter (seek shelter), or flees the area, (2) all responders flee the area, (3), all responders seek refuge (shelter-in-place), (4) all responders seek shelter, (5) all agents flee the area, (6) all agents seek refuge, (7) all agents seek shelter, and (8) all agents do nothing. We found that, for an EF5 tornado hitting Norman at rush hour, the overall fatality rates by protective action type were 0.6% for those who took no action, 0.3% for those who sought refuge, 1.5% for those who sought shelter, and 1.1% for those fleeing the area. We also found that fatality rates were reduced by a factor of 6.6 for Scenario 6 (shelter-in-place) over Scenario 7 (travel to a FEMA-rated shelter). We believe that models such as TWISTER ABM can be used by the NWS and emergency managers in their attempts at communicating the effectiveness of shelter-in-place.

Practical Applications

Tornadoes are dangerous windstorms that can cause serious injury or death to people who do not take protective action. The NWS states that sheltering-in-place is the safest form of protective action, but no studies to date have shown how much it can reduce casualties. We developed an agent-based model to study how changes in protective action type can influence the fatality rate (fatalities per 1,000 residents) caused by a tornado in the city of Norman, Oklahoma. We found that, for an EF5 tornado hitting Norman at rush hour, the overall fatality rates, for all model runs, were lowest for agents who sheltered-in-place (0.3%) and highest for those who traveled to public shelters (1.53%). We also found that fatality rates were lowest when all agents sheltered-in-place (0.24%) and highest when every agent responding to the warning traveled to public shelters (1.54%), a 6.6 times reduction for shelter-in-place. We believe that models such as TWISTER ABM can be used by the NWS and emergency managers in their attempts at communicating the effectiveness of shelter-in-place.

Business interruption after disasters is an important metric for community resilience planning because has both economic and social consequences. Each additional day that a business is nonoperational further compounds lost revenue, wages, and lack of access to goods and services needed for recovery. Therefore, the use of surveys has grown in the literature as a way to capture the diverse information needed for modeling business disaster outcomes. However, variable inclusion and measurement can vary widely across studies, and there is a lack of guidance on how to structure surveys most effectively to facilitate this effort. This study fills these gaps through an analysis of variable choice, variable measurement, and measurement timing using data from an interdisciplinary field study in Lumberton, North Carolina after 2016 Hurricane Matthew. We found that empirical business interruption models can be improved significantly by using a comprehensive set of utility and damage variables; integrating damage information based on damage states for building, contents, and machinery; and capturing recovery-time dynamics by using business downtime and utility outage durations, rather than binary measurements. The results suggest that making these relatively small changes to survey design in future studies can yield large returns in empirical business models for community resilience research.

This study investigated the adverse impact of climate change and climate-related uncertainties on urban waterlogging. Based on multiclimate models, the future extreme precipitation and urban waterlogging under uncertainties was assessed for Beijing. Future rainfall scenarios under two representative concentration pathways (RCP4.5/RCP8.5) were generated by the Delta change method combined with Chicago design storm. The InfoWorks ICM was used to simulate the extent of urban waterlogging. Results showed that urban waterlogging in Beijing would increase, which would be induced not only by more severe climate changes, but also by the limited capacity of drainage systems. In addition, the model uncertainties were found to be increased with both the precipitation intensity and the level of climate change, potentially reaching up to 109%. Therefore, this study highlights the importance for taking climate change and the consequent uncertainties into consideration when designing waterlogging prevention measures.

Abstract

This study aims to examine the relationship between earthquake experience and preparedness over time, specifically investigating the statistical associations between direct and indirect experiences of infrequent and moderate earthquakes and various preparedness measures. Using survey data from 2018 to 2020, this study identified the relationship between direct/indirect experiences and compliance with emergency preparedness measures, seismic reinforcement, and buying insurance through two-way analysis of variance and regression analysis. Direct experience was significantly associated with all types of preparedness measures, with a substantial improvement in compliance with emergency preparedness measures. In contrast, indirect experience had only a minor association with level of compliance and intention to buy insurance. The so-called dread factor demonstrated a statistically significant relationship with the level of compliance in individuals who had experienced earthquakes directly. These findings provide important insights into the statistical associations between earthquake experiences and various preparedness measures and may have implications for effective policies related to preparedness for infrequent and moderate earthquakes.

Practical Applications

This study looks into how people’s personal experiences with earthquakes affect their level of preparedness over time. The researchers examined the effects of both direct experiences and indirect experiences using survey data between 2018 and 2020. In this study, direct experiences refer to feeling high-intensity shaking and trembling in close proximity to the epicenter, which may result in physical, financial, or psychological damage. Indirect experiences, on the other hand, involve feeling weak tremors far from the epicenter, where there is no material or physical damage. The study’s results indicate that individuals who have directly experienced an earthquake are more likely to take preparedness actions, especially regarding emergency guidelines. Conversely, those who have had indirect experiences show less willingness to follow emergency protocols and consider insurance. Furthermore, the study highlights the significant role of fear in influencing people’s compliance with emergency protocols, particularly among those who have directly experienced earthquakes. In conclusion, the research findings demonstrate a clear link between earthquake experiences and preparedness measures, providing valuable insights for effective strategies in managing infrequent and moderate seismic events.

Many empirical relationships have been proposed to relate the sediment volume to various influencing factors. However, the accuracy of such empirical relationships generally requires a large number of observation data, which is difficult to guarantee at a specific site. Moreover, based on the limited investigation data, a complicated empirical model with more input factors may be an overfitted equation. Therefore, how to develop a reliable prediction model of debris-flow sediment volume still remains a great challenge. This paper develops a probabilistic method to establish the most appropriate empirical model for predicting the debris-flow volume based on Bayesian inference. First, the limited site investigation data are preprocessed by a series of multicollinearity analysis to select the candidate input variables. Then, a Bayesian framework is developed to select the most appropriate model among alternatives and identify its corresponding model parameters based on the site investigation data and prior knowledge. To address the multidimensional issues in Bayesian inference, a multichain method, specifically the DREAM_(ZS) algorithm, is used to obtain the posterior distribution of model parameters of a candidate model to overcome the inefficient sampling problems of single-chain Markov chain Monte Carlo methods (e.g., Metropolis–Hastings algorithm). The DREAM_(ZS) algorithm is subsequently integrated with Gaussian copula to calculate the evidence of a candidate model, making it feasible in the model selection problem. Results show that compared with the preexisting empirical relationship, the proposed approaches can provide a more accurate and simpler empirical model by reasonably considering the balance between data fitting and model uncertainty.

As an increasingly urbanized country, China faces significant disaster risks. Social vulnerability influences the ability to prepare for, respond to, and recover from disasters. Quantifying social vulnerability is a worthwhile study because it can provide information vital to understanding the interactions among individuals, communities, and society and can further help us improve disaster risk mitigation. Therefore, this paper presents a case study on the social vulnerability assessment of the inner area of Jiaxing, China. With household questionnaire data, we calculated the social vulnerability index (SoVI) through categorical principal component analysis (CATPCA). We also presented a holistic and detailed profile of social vulnerability that links three different levels (e.g., households, apartment blocks, and communities) and contains four aspects of SoVI, such as statistical information, spatial pattern, dimension attributes, and risk state combinations of variables. This profile also responded to criticisms about the index method’s reductionist nature. Our results retained as much valuable information as we could, enabling a more holistic understanding of demographic, environmental, and social conditions that influence social vulnerability to disasters in the inner city of Jiaxing.

In recent years, several places all over the world have experienced devastating wildfire events that have resulted in exorbitant losses. With the climate warming every year, the trend of destructive fires is expected to continue in the future as well. To be better prepared for such events, an understanding of interaction between wildfires and the built environment is required. In this study, a graph-based approach is utilized to predict the damage to individual buildings in a test area selected from the 2022 Marshall Fire–affected region. The validity of the graph model is first shown by comparing the predicted damage with the observed damaged patterns, showing 72% accuracy. The model is further utilized to investigate the sensitivity of the damage patterns to wind conditions to better understand the correlation between wind characteristics and the resulting damage. For two wind speeds, 12 and $25\mathrm{m}/\mathrm{s}$, polar fragilities are developed by evaluating the damage for different wind directions. The results showed the variation in damage distribution among all wind directions for $25\mathrm{m}/\mathrm{s}$ to be more than twice that for $12\mathrm{m}/\mathrm{s}$, highlighting the fact that the impact of wind conditions is nonlinear and needs to be carefully considered for wildfire risk assessment.