Mary H. Hayden

False alarms and close calls: A conceptual model of warning accuracy

Abstract The false alarm rate (FAR) measures the fraction of forecasted events that did not occur, and it remains one of the key metrics for verifying National Weather Service (NWS) weather warnings. The national FAR for tornado warnings in 2003 was 0.76, indicating that only one in four tornado warnings was verified. The NWS’s goal for 2010 is to reduce this value to 0.70. Conventional wisdom is that false alarms reduce the public’s willingness to respond to future events. This paper questions this conventional wisdom. In addition, this paper argues that the metrics used to evaluate false alarms do not accurately represent the numbers of actual false alarms or the forecasters’ abilities because current metrics categorize events as either a hit or a miss and do not give forecasters credit for close calls. Aspects discussed in this paper include how the NWS FAR is measured, how humans respond to warnings, and what are alternative approaches to measure FAR. A conceptual model is presented as a framework for...

Barriers and Bridges to Prevention and Control of Dengue: The Need for a Social–Ecological Approach

This article critically examines how programs for the prevention and control of dengue fever have been conducted in the absence of an integrated approach, and considers the social and ecological factors influencing their effectiveness. Despite recognition of dengue fever as the most important arboviral disease affecting humans, and in spite of a greater emphasis on community-based control approaches, the burden placed on the communities, countries, and regions affected by this disease continues to rise. In considering historical experience in the Americas and the Asia-Pacific region, as well as the global forces that are exerting new pressures, the important elements of successful control programs are identified as community ownership, partnership with government, leadership, scalability, and control of immature mosquitoes. The key barriers to the exchange of knowledge and the transdisciplinary cooperation necessary for sustainable dengue control are rooted in differences in values among policy-makers, citizens, and scientists and are repeatedly expressed in technical, economic, cultural, geographic, and political dimensions. Through consideration of case studies in Cuba, Guatemala, Singapore, Thailand, Indonesia, and Vietnam, the limitations of control approaches that fail to take into account the complexities of ecological and social systems are presented. Bridges to effective control are identified as the basis for adaptability, both of control programs to the mosquito vector’s changing behavior and of education programs to public, regional and local particularities, as well as transdisciplinarity, community empowerment, the ability to scale local experiences up to the macro-level, and the capacity to learn from experience to achieve sustainability.

Storm Surge and ''Certain Death'': Interviews with Texas Coastal Residents following Hurricane Ike

Hurricane Ike made landfall near Galveston, Texas, on 13 September 2008 as a large category 2 storm that generated significant storm surge and flooding. This article presents findings from an empirical case study of Texas coastal residents’ perceptions of hurricane risk, protective decision making, and opinions of hurricane forecasts related to Hurricane Ike. The results are based on data from interviews with 49 residents affected by Hurricane Ike, conducted approximately five weeks after landfall. While most interviewees were aware that Ike was potentially dangerous, many were surprised by how much coastal flooding the hurricane caused and the resulting damage. For many—even long-time residents—Ike was a learning experience. As the hurricane approached, interviewees and their households made complex, evolving preparation and evacuation decisions. Although evacuation orders were an important consideration for some interviewees, many obtained information about Ike frequently from multiple sources to evaluate their own risk and make protective decisions. Given the storm surge and damage Ike caused, a number of interviewees believed that Ike’s classification on the Saffir‐Simpson scale did not adequately communicate the risk Ike posed. The ‘‘certain death’’ statement issued by the National Weather Service helped convince several interviewees to evacuate. However, others had strong negative opinions of the statement that may negatively influence their interpretation of and response to future warnings. As these findings indicate, empirical studies of how intended audiences obtain, interpret, and use hurricane forecasts and warnings provide valuable knowledge that can help design more effective ways to convey hurricane risk.

The Dengue Virus Mosquito Vector Aedes aegypti at High Elevation in México

México has cities (e.g., México City and Puebla City) located at elevations > 2,000 m and above the elevation ceiling below which local climates allow the dengue virus mosquito vector Aedes aegypti to proliferate. Climate warming could raise this ceiling and place high-elevation cities at risk for dengue virus transmission. To assess the elevation ceiling for Ae. aegypti and determine the potential for using weather/climate parameters to predict mosquito abundance, we surveyed 12 communities along an elevation/climate gradient from Veracruz City (sea level) to Puebla City (∼2,100 m). Ae. aegypti was commonly encountered up to 1,700 m and present but rare from 1,700 to 2,130 m. This finding extends the known elevation range in México by > 300 m. Mosquito abundance was correlated with weather parameters, including temperature indices. Potential larval development sites were abundant in Puebla City and other high-elevation communities, suggesting that Ae. aegypti could proliferate should the climate become warmer.

Information sources for flash flood warnings in Denver, CO and Austin, TX

Abstract This research examines sources of information for flash floods in two large metropolitan areas, Denver, CO, and Austin, TX. Previous research has noted that information delivery systems for weather forecasts are geared toward the cultural majority and suggests that inadequate warnings are a primary contributor to deaths and injuries from hazards. This investigation used chi-square analysis to determine the prime warning source preferences and preferred time of day for receiving different media. Results indicate that successful warning messages need to be targeted toward specific sub-populations if the warning is to be received, understood, and responded to properly.

The potential for the use of Open Source Software and Open Specifications in creating Web-based cross-border health spatial information systems

Globalization is contributing to the blurring of borders making irrelevant the distinctions between domestic and international health problems. Cross‐border and global health spatial information systems (CBHSIS) are required to address the new global health challenges. There is a need to build and document alternatives for addressing the technological, economic, and sociocultural–political challenges encountered in the creation and deployment of these systems. This paper documents the building of a prototype Web‐based multimedia GIS system for use in a public health context using Open Source Software and Open Specifications and its deployment across the US–Mexico border. These technologies offer advantages in addressing several of the challenges previously mentioned. We highlight the technological and sociocultural–political issues important in successful collaboration across borders and cultures and in the creation of interoperable CBHSIS.

Decision Making by Austin, Texas, Residents in Hypothetical Tornado Scenarios*

Abstract One of the goals of the Warning Project is to understand how people receive warnings of hazardous weather and subsequently use this information to make decisions. As part of the project, 519 surveys from Austin, Texas, floodplain residents were collected and analyzed. About 90% of respondents understood that a tornado warning represented a more serious and more likely threat than a tornado watch. Most respondents (86%) were not concerned about a limited number of false alarms or close calls reducing their confidence in future warnings, suggesting no cry-wolf effect. Most respondents reported safe decisions in two hypothetical scenarios: a tornado warning issued while the respondent was home and a tornado visible by the respondent while driving. However, nearly half the respondents indicated that they would seek shelter from a tornado under a highway overpass if they were driving. Despite the limitations of this study, these results suggest that more education is needed on the dangers of highway o...

CORRIGENDUM: False Alarm Rate or False Alarm Ratio?

Two items need to be clarified from an earlier work of the authors. The first is that the layout of the 2 3 2 contingency table was reversed from standard practice, with the titles of ‘‘observed event’’ and ‘‘forecast’’ transposed. The second is that FAR should have represented ‘‘false alarm ratio,’’ not ‘‘false alarm rate.’’ Unfortunately, the terminology used in the atmospheric sciences is confusing, with authors as early as 1965 having used the terminology differently from currently accepted practice. More recent studies are not much better. A survey of peer-reviewed articles published in American Meteorological Society journals between 2001 and 2007 found that, of 26 articles using those terms, 10 (38%) used them inconsistently with the currently accepted definitions. This article recommends that authors make explicit how their verification statistics are calculated in their manuscripts and consider using the terms probability of false detection and probability of false alarm instead of false alarm rate and false alarm ratio.

A Model for Defining West Nile Virus Risk Perception Based on Ecology and Proximity

West Nile virus (WNV) has made considerable impact as an emerging infectious disease, spreading from coast to coast across North America since 1999. The disease has exhibited great spatial variability in North America, making an ecosystems approach to understanding the local human and vector ecology critical to prevention and control. This study underscores the importance of employing both personal prevention and community participatory approaches to create messages that have been adapted to the local ecology and are designed to reduce the risk of human infection with this mosquito-borne virus. As the virus spreads into new areas, underlying attitudes toward mosquitoes and the local perception of environment/ecology can affect the success of control and prevention measures. This work presents results from focus group discussions conducted in two Colorado counties in 2003, a year of significant WNV activity in the state. Issues addressed include residents’ assessment of risk and how this perception varied by age group and location, use or nonuse of repellents, and community attitudes toward mosquito control in areas with different ecologies and histories of mosquito-borne disease. The need to address individual components of personal prevention, to target prevention to specific audiences, and to disseminate prevention messages through local channels is discussed. The authors propose including aspects of ecology and disease proximity in understanding risk perception and addressing emerging diseases with a prominent ecological component.

Meteorological Conditions Associated with Increased Incidence of West Nile Virus Disease in the United States, 2004-2012

West Nile virus (WNV) is a leading cause of mosquito-borne disease in the United States. Annual seasonal outbreaks vary in size and location. Predicting where and when higher than normal WNV transmission will occur can help direct limited public health resources. We developed models for the contiguous United States to identify meteorological anomalies associated with above average incidence of WNV neuroinvasive disease from 2004 to 2012. We used county-level WNV data reported to ArboNET and meteorological data from the North American Land Data Assimilation System. As a result of geographic differences in WNV transmission, we divided the United States into East and West, and 10 climate regions. Above average annual temperature was associated with increased likelihood of higher than normal WNV disease incidence, nationally and in most regions. Lower than average annual total precipitation was associated with higher disease incidence in the eastern United States, but the opposite was true in most western regions. Although multiple factors influence WNV transmission, these findings show that anomalies in temperature and precipitation are associated with above average WNV disease incidence. Readily accessible meteorological data may be used to develop predictive models to forecast geographic areas with elevated WNV disease risk before the coming season.