Natural Hazards Review

Published by American Society of Civil Engineers
Print ISSN: 1527-6988
Contribution to benefit-cost ratio by factor for: ͑ a ͒ earthquake; ͑ b ͒ wind; and ͑ c ͒ flood 
Sensitivity of benefit to uncertainties ͑ grants for earthquake project mitigation activities ͒ 
Mitigation ameliorates the impact of natural hazards on communities by reducing loss of life and injury, property and environmental damage, and social and economic disruption. The potential to reduce these losses brings many benefits, but every mitigation activity has a cost that must be considered in our world of limited resources. In principle benefit-cost analysis (BCA) can be used to assess a mitigation activity’s expected net benefits (discounted future benefits less discounted costs), but in practice this often proves difficult. This paper reports on a study that refined BCA methodologies and applied them to a national statistical sample of FEMA mitigation activities over a ten-year period for earthquake, flood, and wind hazards. The results indicate that the overall benefit-cost ratio for FEMA mitigation grants is about 4 to 1, though the ratio varies according to hazard and mitigation type.
We test the extent to which growth in the 11 CIS countries (excluding Russia) was associated with developments in Russia, overall, as well as through the trade, financial and remittance channels over the last decade or so. The results point to the continued existence of economic links between the CIS countries and Russia, though these links may have altered since the 1998 crisis. Russia appears to influence regional growth mainly through the remittance channel and somewhat less so through the financial channel. There is a shrinking role of the trade (exports to Russia) channel. Russian growth shocks are associated with sizable effects on Belarus, Kazakhstan, Kyrgyz Republic, Tajikistan, and, to some extent, Georgia.
Numerical experiments using a mesoscale meteorological model (MM5) are performed to evaluate the mountainous orographical effects on the heavy rainfalls brought by Typhoon 0514 (NABI), which caused the flood disaster in the southeast Kyushu area of Japan. The terrain conditions considered in the numerical model are three folds: first, a flat terrain with the altitude 1m above mean sea level; second, an idealized line-shaped mountain terrain; third, a complex terrain using GTOPO30. Although an accumulated rainfall due to Typhoon 0514 is recorded higher than 1,000 mm, a calculated one using the flat terrain is 250-300 mm. The calculated rainfall using the complex terrain becomes 200-300% (500-900 mm) in comparison to flat terrain case. This discrepancy is found to cause by blocking and evolving the convective cells, which are generated by lifting up the water vapor along the mountain slope in the windward areas. A ratio of the forecasted rainfall with/without orography provided an important index for the risk of the heavy rain in the tropical cyclone.
Through five systematic, large-scale mail surveys conducted since 1993, the Disaster Research Center (DRC) has obtained data on hazard awareness, preparedness, disaster impacts, and short- and long-term recovery among 5,000 private-sector firms in communities across the United States (Memphis/Shelby County, Tennessee, Des Moines, Iowa, Los Angeles, California, Santa Cruz County, California, and South Dade County, Florida). This paper summarizes findings from those studies in three major areas: factors influencing business disaster preparedness; disaster-related sources of business disruption and financial loss; and factors that affect the ability of businesses to recover following major disaster events. Implications of the research for business contingency planning and business disaster management are discussed.
In order to motivate flood insurance purchase and promote flood hazard awareness and mitigation, the Community Rating System (CRS) of National Flood Insurance Program (NFIP), credits floodplain management activities and awards flood insurance premium discounts. CRS, however, has been marked by a lack of active participation since its inception. The objective of this study is to provide empirical evidence related to community decisions involving incentive-based flood risk mitigation projects. We test a number of hypotheses offered by previous researchers regarding factors that motivate local hazard management initiatives through an examination of patterns in CRS participation across all 100 North Carolina counties from 1991 to 2002. Specifically, we examine the influence of flood experience, hydrological risk, local capacity, and socioeconomic factors on county hazard mitigation decisions. Results indicate that flood history and physical risk factors increase likelihood of local hazard mitigation adoption. We find evidence that the probability of CRS participation is lower in counties with a greater proportion of senior citizens and greater level of education, and that flood hazard mitigation activities at the county level are more likely when a greater number of nested of municipalities participate in CRS.
This paper examines the history of St. Louis, Missouri in coping with flood risk over the past 15 years, with a focus on flood insurance. Six challenges to the continued management of riverine flood risk are identified and discussed. They are (1) many property owners don't buy flood insurance, (2) people underestimate flood risk, (3) we need better flood maps, (4) we have a "love affair" with levees, (5) flood risk is increasing over time, and (6) we take deep pride in rebuilding after a disaster. Recommendations for how to improve flood risk management in light of these challenges are offered. Focused attention is given to the possibility of long-term flood insurance contracts tied to long-term loans for risk-mitigating activities in overcoming the six challenges.
US and international officials have sought ways to effectively use mobile technology to communicate health information to help thwart the spread of coronavirus disease 2019 (COVID-19). This study offers a preliminary exploration into the use of the state-level (N=6) and local-level (N=53) Wireless Emergency Alert (WEA) for notifications regarding COVID-19 in the US. The study compares changes in reported rates of infections and deaths between states and localities that issued WEA messages in March and April of 2020 with states that did not. Small sample sizes and differences in the rates of COVID-19 spread prohibit robust statistical analysis and detection of clear effect sizes, but estimated effects are generally in the right direction. Combining statistical analysis with preliminary categorization of both WEA message content and social media themes suggests that WEA messages might play an important role in instructing people to take protective in response to COVID-19. These initial lines of inquiry will be further developed in subsequent research.
The COVID-19 pandemic resulted in significant social and economic impacts throughout the world. In addition to the health consequences, the impacts on travel behavior have also been sudden and wide ranging. This study describes the drastic changes in human behavior using the analysis of highway volume data as a representation of personal activity and interaction. Same-day traffic volumes for 2019 and 2020 across Florida were analyzed to identify spatial and temporal changes in behavior resulting from the disease or fear of it and statewide directives to limit person-to-person interaction. Compared to similar days in 2019, overall statewide traffic volume dropped by 47.5%. Although decreases were evident across the state, there were also differences between rural and urban areas and between highways and arterials both in terms of the timing and extent. The data and analyses help to demonstrate the early impacts of the pandemic and may be useful for operational and strategic planning of recovery efforts and for dealing with future pandemics.
Because annual death and destruction from hurricanes in the United States can vary by >4 orders of magnitude, common logarithms are convenient measures. If the data are stratified into years when at least one major hurricane (maximum winds >49 m/s) made landfall and those without a major-hurricane-landfall, logarithmic mortality and damage in both subsets appear to be normally distributed. Combined log-normal distributions accurately represent the data. From 1900 through 2008, total hurricane mortality decreased with a halving time of 26years. Consistent with previous analyses, historical damage normalized for population growth, increasing individual wealth, and inflation did not exhibit significant trends in either subset. Comparison between the complete 1900-2008 record and a truncated version spanning 1900-2000 shows that the effects of hurricanes since 2000, especially Katrina, doubled the expected loss of life and increased the expected damage by 10%, consistent with the relatively large statistical uncertainty in estimation of the model parameters. In the context of the model, the expected frequency of disasters that cause >$100 billion in damage is approximately three times a century and for those that claim >1,000 lives is approximately once a century.
Earthquake prediction is by all means controversial and challenging, given the fact that some recent catastrophic earthquakes went unpredicted. Not surprisingly, statistical approaches have been utilized to model earthquake randomness in time or space. One of the suggestions is that the earthquake's temporal probability distribution should follow the Poisson model, which is suitable for rare events by definition. As a result, the customarily used hypothesis should be largely associated with the prior judgment that earthquakes are rare, but not as a result of abundant quantitative evidence or theoretical derivation. Therefore, this study aims to offer new empirical evidence to the hypothesis based on 110-year-long earthquake data around Taiwan. From the series of statistical tests, the first statistical inference is indeed in line with the model's proposition: the level of fitting between observation and theory is better for earthquakes with a lower mean rate. To be more specific, it shows that the Poissonian hypothesis applied to local magnitude (ML)3.0 earthquakes around Taiwan with a mean annual rate as high as 1,600 is clearly rejected, but as far as ML7.0 earthquakes with a mean rate of 0.35 per year are concerned, the same hypothesis is statistically accepted for modeling their temporal randomness. Also, according to the tests on a variety of conditions, the annual rate of approximately 0.1 per year (or 10-year return period) was suggested as a reasonable empirical estimate for Poissonian rareness. Accordingly, from a practical point of view, it should be a robust analytical presumption to use the Poisson model in daily earthquake engineering analyses because the return period of design earthquakes is longer than 10 years, if not much longer.
Recent research indicates that the energy generated by hurricanes follows a power law distribution. The authors hypothesize that economic damages caused by hurricanes also follow a power law distribution. Using yearly hurricane damage data from 1900 to 2005 the authors show that the distribution of yearly damages due to hurricanes in the United States may follow either a power law or lognormal distribution. Furthermore, if the distribution of damages follows a power law, then for the best-fit distribution, the tail of the distribution is so fat that the variance of damages, conditional on being in the tail, is potentially unbounded. (C) 2014 American Society of Civil Engineers.
Tornadoes are one of the most devastating natural hazards that occur in the United States. Although the average number of tornadoes per year across the country is approximately 1,200, the annual likelihood of experiencing a tornado at a particular location is quite small due to their relatively small size. However, the high consequence of a tornado strike necessitates the determination of geographic tornado hazard. This paper presents a methodology to estimate the annualized probabilistic tornado hazard over the contiguous United States based upon the most recent 38 years of climatological tornado data. Furthermore, with the use of detailed damage surveys after the April 3-4, 1974 and April-May 2011 tornado outbreaks, an empirical method was developed and applied to account for the gradient of wind speed along a tornado's path length and path width. From this, a probabilistic tornado hazard index can be developed across the United States and can eventually be extended to tornado hazard for any location. (C) 2014 American Society of Civil Engineers.
In recent years, claims have been made in venues including the authoritative reports of the Intergovernmental Panel on Climate Change (IPCC) and in testimony before the U.S. Congress that economic losses from weather events have been increasing beyond that which can be explained by societal change, based on loss data from the reinsurance industry and aggregated since 1980 at the global level. Such claims imply a contradiction with a large set of peer-reviewed studies focused on regional losses, typically over a much longer time period, which concludes that loss trends are explained entirely by societal change. To address this implied mismatch, this study disaggregates global losses from a widely utilized reinsurance data set into regional components and compares this disaggregation directly to the findings from the literature at the regional scale, most of which reach back much further in time. The study finds that global losses increased at a rate of $3.1 billion/year (2008 USD) from 1980-2008 and losses from North American, Asian, European, and Australian storms and floods account for 97% of the increase. In particular, North American storms, of which U.S. hurricane losses compose the bulk, account for 57% of global economic losses. Longer-term loss trends in these regions can be explained entirely by socioeconomic factors in each region such as increasing wealth, population growth, and increasing development in vulnerable areas. The remaining 3% of the global increase 1980 to 2008 is the result of losses for which regionally based studies have not yet been completed. On climate timescales, societal change is sufficient to explain the increasing costs of disasters at the global level and claims to the contrary are not supported by aggregate loss data from the reinsurance industry. (C) 2014 American Society of Civil Engineers.
In this study, the characteristics of tropical cyclones (TCs) and their economic losses in China for a 30-year period (1984-2013) are analyzed at provincial scale. The TC parameters are quantitatively analyzed on inherent trends and cycles using the Mann-Kendall test and the Fast Fourier power spectrum. Different normalization methods are applied to attribute socio-economic factors to the increasing original economic losses. These losses are normalized with the consumer price index, the conventional as well as the alternative normalization method. The frequency of maximum economic losses from TC events is expressed in return periods as calculated with the Generalized Extreme Value distribution function. The results show a noticeable shift to stronger TCs within detected strong cycles of 5 and 11-13 years, but no significant trends in the climatic parameters related to TCs. Especially the numbers of affected population and the TC intensities in inland provinces have amplified. Results of the different normalization methods show that by considering population growth, urbanization, and economic development the economic losses per capita did not increase, but rather stabilized. The original economic losses from TCs increased during the period of 1984-2013, with 2013 as the most costly TC-year. In comparison, the conventional and alternative normalized economic losses have not increased and rather see 1996 as the most costly TC year. Based on the return periods of probable maximum losses per TC, it is shown that the TC-event with the highest losses in 2013, i.e. “Typhoon Fitow”, is a 23-year event if the losses are adjusted with the consumer price index, but a more realistic 3.5-year event if the conventional normalization method is considered. In conclusion, with the economic development and urbanization, China’s vulnerability and exposure to tropical cyclones has increased, while the normalized economic losses from TCs did not rise.
Between June 5 and 9, 2001, Tropical Storm Allison dropped upwards of 50 cm (20 in.) of rain on the Texas Medical Center (TMC) in Houston, causing the worst urban flood in U.S. history. The unprecedented rainfall event flooded hospitals, labs, underground tunnels and garages, and power stations, and resulted in excess of $1.5 billion in damages and the loss of decades of medical research. Tropical Storm Allison served as a severe wake-up call to management at the TMC. In response, they developed a hazard mitigation plan (HMP) to minimize the impact of natural and artificial hazards on the TMC campus and its member institutions in the future. Today, the TMC is the premier example of a world-class institution that has a working hazard mitigation plan. This paper discusses the impacts of Tropical Storm Allison (2001) to the TMC and the measures officials have taken to protect and upgrade the flood infrastructure as an example of hazard management for other large, vulnerable institutions. (C) 2014 American Society of Civil Engineers.
Building Materials by Housing Types
Descriptive Statistics for Independent Variables in the Linear Regression Model
Subjective Recovery Models
Percentage of Households with Access to Basic Amenities before and after Tsunami Households'
While much has been written about postdisaster housing reconstruction, few have explored the effects of certain types of building designs or materials in improving the quality of the housing stock, basic utilities, and amenities. This study examines housing improvements at two time points (sixmonths and three-and-a-halfyears) after the 2004 tsunami and discusses structural recovery to the built form as well as households' perceptions of recovery at the individual and community levels. The analysis is made on the basis of surveys conducted in 558 rural households in the coastal district of Nagapattinam in Tamil Nadu, India. Findings suggest that the reconstructed core-housing units are structurally good and have improved because of a stricter adherence to building standards and materials. Despite assistance and improvements to their housing stock, respondents from the backward classes and scheduled castes, lower income groups, and those pursuing nonfishing livelihoods expressed lower levels of perceived recovery compared with those from households from the most-backward caste, those with higher incomes, and fishing folks. Findings suggest that postdisaster housing recovery programs need to emphasize sustainable development goals that ideally not only contribute to physical well-being but also address a wide range of social problems as well. This will help to ensure that long-term physical, social, and economic recoveries are equitable to a broad range of beneficiaries.
Reinforced-concrete-frame buildings, particularly midrise multifamily condominium structures, experienced significant damage in the 2009 L'Aquila, Italy earthquake. The most frequently observed damage was cracking and failure of nonstructural masonry infill walls, but some buildings also experienced structural damage, including column shear failure and collapse. Fieldwork conducted approximately 3 weeks after the earthquake was used to collect data from approximately 483 concrete frame structures in L'Aquila, including building location, characteristics, damage, and postearthquake occupancy. A second survey 1 year later investigated whether buildings had been repaired or reoccupied. These results show that the building damage is correlated to height, usage, elevation irregularities in strength and stiffness, and the ground-shaking intensity at each site. This information provides the basis for empirically derived fragility functions for this typology of concrete frame. Approximately 0.2% of concrete structures collapsed. The study also shows that occupants of tall and older condominium structures in areas north and west of the city center were particularly affected; single-family homes were less damaged and more likely to be occupied during fieldwork. Damaged concrete buildings have led to significant disruption of the community and social fabric, causing the temporary or permanent closure of small businesses, medical offices, restaurants, churches, and schools.
Obtaining accurate spatial details of the parameters involved in landslides has been a major challenge in determining the risk of landslide following an extreme event of rainfall, earthquake, or a combination of both. In recent decades, advances in remote sensing with high resolution satellite imagery and digital elevation models have permitted very detailed mapping and analysis of landslide hazards; however, there has been little work verifying the reliability and precision of these techniques as compared to traditional field surveys. This paper seeks to improve this situation by assessing the feasibility of using remote sensing to determine landslide vulnerability. This has been carried out in two parts. Firstly, global positioning system (GPS) coordinates collected in the field after the September 30, 2009, Padang earthquake in Sumatra, Indonesia were compared to advanced spaceborne thermal emission and reflection radiometer (ASTER) and Google Earth digital elevation model (DEMs) and Satellite Pour L' Observation de la Terre (SPOT-5) satellite imagery. They showed reasonable spatial and elevation differences, which demonstrates the suitability of remote sensing for landslide hazard assessments. Secondly, results from a geographic information system (GIS) analysis carried out with these data showed that remote sensing is capable of producing practical landslide hazard maps that reflect an accurate measure of landslide risk during the 2009 Padang Earthquake. Inclusion of a water saturation contribution map in the conventional slope stability has proven able to better identify areas susceptible to landslides. Prior to the disaster, many of these landslide locations were demarcated as moderate risk regions in the local hazard map. Considering the high lethality of these events, this underestimate of the risk is a strong argument for a review of landslide risks using remote sensing to aid in assessing the combined effects of earthquake and rainfall on such landslides in this region. (C) 2014 American Society of Civil Engineers.
The boundary organization concept has been used to establish that collaborative arrangements and outputs across science and policy domain boundaries need to be credible, relevant, and legitimate in order to be to be effective. Although widely accepted in other issue-driven fields, this concept does not have equivalent currency in the natural hazard and disaster risk reduction context. This paper uses the development of the New Zealand Natural Hazards Research Platform during a recent earthquake disaster to assess the utility of the concept in this topic area. Lessons are also identified concerning the use of larger consortium organizations to increase policy and other end-user involvement in the management and coordination of research funding, and the impact of a major disaster on this research-funding initiative. Mapping the Platform’s collaborative arrangements in relation to boundary tensions over time makes it possible to distinguish disaster effects from preexisting and ongoing structural effects and incentive regimes. Largely based in the research domain, this organization was well placed to resist the negative pressure of postdisaster time compression on research quality. The lack of balancing policy input at all levels made it difficult to resist the effect of this pressure on the networking required to integrate disciplinary, organizational, and higher-level science/policy domains, and thus build the legitimacy of the larger collaboration. The utility of the boundary organization concept stemmed from the emphasis on balance across domains and scales. The focus on effects, trends, and patterns serves as a counterweight to the blame attribution common after high-profile disasters.
The March 11, 2011, Mw 9.0 Tohoku earthquake caused widespread liquefaction, lateral spreading, and dynamically induced settlement in the Kanto plains north of Tokyo, Japan. In April 2011, the authors conducted postearthquake engineering field reconnaissance of the mesoseismal region as part of an ASCE-sponsored investigation of the earthquake. The field investigation revealed that numerous levees, roads, and structures built in proximity of the Tone River and its tributaries were damaged severely by the Tohoku earthquake owing to the liquefaction and associated lateral spreading. More importantly, the investigation revealed that much of the liquefaction and lateral spread damage was directly related to anthropogenic (man-made) land use changes rather than natural conditions. These historical changes generally were related to major fluvial transportation improvements made along the Tone River that started in the seventeenth century and have continued into the present era. These findings have implications for levees found along many rivers in seismically active areas around the world, including those along the Sacramento River in California and the Mississippi River near the New Madrid seismic zone.
The 2011 Joplin, Missouri, tornado was one of the deadliest and costliest tornadoes in US history, damaging approximately 8,000 structures and causing more than $2 billion in economic damages. As with most extreme events, reports following the tornado documented the widespread damage, including complete destruction of one of the local hospitals along with various schools. However, recovery processes have not been documented and evaluated at the same level of spatial detail. Following the 2011 Joplin tornado, researchers periodically revisited neighborhoods at 6-month intervals for the first 2 years, then yearly for the following 3 years, with the goal of collecting spatial video data in order to document when structures were fully repaired or rebuilt. This case study documents the building repair time (time to reach full building functionality) patterns based on that data set for the first 2 years following the devastating Enhanced Fujita (EF) 5 tornado that struck Joplin, Missouri. The preliminary results comprehensively show a longer average building repair time for older (pre-1970) buildings and areas of lower population count, while rebuild times were quicker for areas where a relatively small amount of the population did not have access to a vehicle and the median age was lower. Within this Joplin recovery study, however, the year built for the structures was concluded to be a stronger factor in delaying recovery time than income, which is typically considered a primary contributor to repairing and rebuilding a structure following an extreme event. Overall, buildings ranging from minimal to severe damage were typically fully repaired within the first year, while buildings that were completely destroyed reached full functionality evenly across 6-month, 1-year, 1.5-year, 2-year, and greater than 2 years recovery times.
The State of California's 2012 Central Valley flood protection plan (CVFPP)-a large-scale reconnaissance-level assessment of flood management problems and potential solutions-formulated a set of broad management alternatives and compared them by assessing risk reduction attributable to each. Both economic risk and life risk were assessed and compared. Economic risk was evaluated with standard U.S. Army Corps of Engineers methods and software. Inputs that described the flood hazard, flood-management system performance, property exposure, and vulnerability to damage were developed by the California Department of Water Resources (DWR) using state-of-practice models (hydrologic, hydraulic, geotechnical, and economic) and data. After considering life-risk analysis methods proposed in the literature, for the CVFPP reconnaissance-level analysis, DWR developed and applied an innovative method that leveraged the economic-risk analysis method and its inputs. Measures of hazard and performance remained the same as for the economic-risk analysis, but exposure in the flood-plain was expressed in terms of population at risk, and vulnerability was expressed as a relationship between hazard and life-loss. The new analysis yielded a description of risk in terms of probability of various magnitudes of life-loss, which could be integrated to estimate an expected value. Although approximate, the method provided systematic, repeatable, timely estimates of life-risk for a variety of alternatives considered, allowing fair comparison of the management alternatives for high-level planning studies. However, this method can only broadly support real-time emergency response decision making and other activities that require more detailed conceptual accounting of complex flooding and human response. (C) 2014 American Society of Civil Engineers.
Material convergence represents the flow of in-kind donations, specifically supplies and equipment, which travels to an area in the aftermath of a disaster. Unfortunately, unsolicited low-priority or nonpriority donations can disrupt the ability of responders to effectively manage items that arrive, thereby impeding the distribution of high-priority items to those affected by the disaster. Previous literature has emphasized the enormity of this problem and identified the necessity of addressing it. With this issue in mind, therefore, a series of interviews was conducted in the aftermath of the September 2013 flooding in Colorado to determine the extent and impact of material convergence on the relief efforts for this particular disaster event. The results of the case study show that relief organizations observed material convergence to be less of an issue in this particular event than in previous disasters. To better understand this result and to provide support for helping to improve future response and relief efforts, practices and methods that were considered to have contributed to reducing material convergence during these particular flooding relief efforts are discussed.
This study analyzes the news media coverage of floods in a major flood-prone state of India, Bihar. Although news media helps shape public perception and political actions, little analysis of this means of communication has been conducted in India. Major topics and issues discussed during the coverage of Bihar floods in years 2013, 2017, and 2019 were provision of food, shelter, and health facilities, failure of the transportation system, waterlogging in urban areas, and failure or management of embankments. There was no skepticism on whether climate change was real. Political parties took contradictory positions: the ruling party attributed floods to a changing climate but other parties, and news media, emphasized the lack of disaster mitigation actions and were uninterested in climate change. This study suggests that it is more important to prepare for disaster mitigation actions around the major issues discussed and communicate them to the public. Media should become a major stakeholder by questioning the authorities about disaster preparations prior to the monsoon season and communicating mitigation actions to the public once disaster has struck, and help both public and government to better manage and mitigate the disaster.
Relative location of Brays Bayou watershed and a digital elevation model (DEM) of the watershed
2015 Memorial Day Maximum Stage
Observed Streamflow Compared with Modeled Results for May 2015 Conditions
Observed Stages Compared with Modeled Results for May 2015 Conditions
Hydrologic Comparisons for the Three Watershed Conditions
In this study a hindcast analysis was performed of the 2015 Memorial Day flood for Brays Bayou in Harris County, Texas. Point rainfall totals as high as 27.84 cm (10.96 in.) occurred in this watershed in a 12-h period, causing severe flood damage to 1,185 residential properties. By validating hydrologic and hydraulic models representative of watershed conditions at the time of the storm, this study provides a comprehensive spatial perspective and understanding of the flows and stages, floodplain extents and depths, and residential damages that occurred. In addition, Project Brays, one of the largest urban watershed-scale flood reduction projects in the nation, was nearly halfway complete at the time of the 2015 Memorial Day storm. Subsequently, this study evaluates and compares the flood response to this rainfall event for different watershed conditions representing preproject conditions (2001), May 2015 conditions (current), and postproject conditions (2021 at the earliest). The flood reductions that were provided by Project Brays at the time of the storm and the flood impacts that would have occurred even if the project had been finalized prior to the storm are assessed in detail.
The 2017 hurricane season was a costly multi-billion-dollar season for the United States. Communicating the real-time risk of these events to the public is usually performed through explanations from meteorologists via media coverage. However, despite meteorologists’ expert knowledge, the public will sometimes not heed warnings because they have developed mistrust in the accuracy of the science used and sometimes believe that media coverage exaggerates. Additionally, the public relies heavily on the Saffir-Simpson Wind Scale even when the most dangerous factors typically are not wind speed. Recently, a new hurricane impact level ranking system was developed using an artificial neural-network model in order to communicate the impending risk of an event from its multiple hazards. In the last three hurricane seasons, the real-time use of this neural network–based model has been tracked, evaluated, and proven relatively accurate. This study assesses what changes in the meteorological forecasts are causing changes in the predicted impact level and how the change in the impact level compares with the language used to communicate risk within the advisories themselves. The results show that the changing language in the National Hurricane Center advisories matched the 2017 forecasted impact levels and that the impact level forecast is most accurate approximately 30 h out from landfall. The most influential meteorological parameter that changes the predicted impact level for most hurricanes over the last three seasons was found to be the wind speed and population affected. For a track shift to result in an impact level change, the affected population is required to approximately double. For Hurricane Harvey, however, rapid intensification of all variables led to its eventual higher impact level forecast.
This study investigates the effectiveness of Reverse 911 warning systems compared with other evacuation warning sources. This study also investigates the impact of individual differences on evacuation behavior, and presents a regression model of evacuation behavior on the basis of empirical data. A 57 question survey was administered by telephone to people who had been affected by the 2007 San Diego wildfires, with 1020 usable responses (8.4% response rate). By signal detection theory, Reverse 911 warnings had the best performance compared with other evacuation warning sources, as indicated by the high influence rating (0.66) and hit rate (1.00). People who received the Reverse 911 warning also had a significantly higher rate of evacuation (0.80), as did those who received warnings from more than one source (0.78). Regression analysis shows that the Reverse 911 warning was critical in predicting whether an individual evacuated. Demographic factors, including knowledge and experience with previous wildfires, also played a significant role in evacuation rates, and therefore must be considered when designing a warning system.
(Color) Distribution of the reference points and the observation points 
Like most natural systems, landslides also have critical points at which a great change in displacement occurs. The displacement may recover increasingly slowly from small perturbations when landslides approach an abrupt change. The Yemaomian Landslide, located in the Three Gorge Region, China, is studied. A time series of displacement is modeled using an autoregressive (AR) model and a detrended fluctuation analysis (DFA) method. The coefficient and variance of AR(1) and the scaling exponent of DFA are estimated using a slide window. The results show that the DFA scaling exponent can indicate the abrupt change in displacement. The DFA scaling exponent increases when an abrupt change is approaching. If the variation in displacements is small, the scaling exponent declines accordingly. The variance of AR(1) may also indicate the displacement fluctuation. These early-warning signals are extracted directly from the observations. They can be used to detect abrupt changes in real time series. They have the potential to be used as early-warning signals for a wide range of landslides.
The accuracy of the Federal Emergency Management Agency's HAZUS general building stock (HGBS), widely used to estimate potential monetary damage to building structures from floods, earthquakes, or hurricanes, is quantified by comparing HGBS square footage and replacement-cost values to corresponding data from building cost inventories previously compiled for three Midwestern locations. The HGBS data (released in 2015 based on year 2010 census data) underestimates all building square footage by between 15 and 20% and overestimates replacement costs by between 31% (without considering depreciation) and 56% (including depreciation). The magnitude of HGBS inventory differences vary across structure types, particularly among different classes of commercial structures, but are markedly consistent across the three locations. Because the HGBS underestimates structure size and overestimates replacement costs, it severely overestimates replacement costs on a dollar per square foot basis, on average by between 51 and 81%. These large overestimates could potentially occur with a partial or hybrid Level 2 HAZUS analysis, in which supplemental structural square footage is obtained from a non-HGBS source, such as a tax assessor, while using default HGBS building replacement cost data.
Top-cited authors
David R Godschalk
  • University of North Carolina at Chapel Hill
Michael K Lindell
  • University of Washington Seattle
Carla Prater
  • Texas A&M University
Roger Pielke
  • University of Colorado Boulder
Christopher W. Landsea
  • National Oceanic and Atmospheric Administration