1. FREQUENCY AND DISTRIBUTION OF FLOOD FATALITIES IN TEXAS COUNTIES PROBLEM STATEMENT The objective of this project is to utilize a geographic information system (GIS) to analyze the spatial distribution of historically-recorded fatalities attributed to flooding events in Texas counties and identify any existing patterns, areas of vulnerability and correlation to population growth. The circumstances contributing to flood fatalities are also examined to ascertain possible relationships between factors and determine the variables that characterize those most likely to be at risk. In addition, total dollar amount of event damages to crops and property are considered. DATA Essential flood event data and data layers utilized for analysis has been collected from a variety of sources. Flood events resulting in fatalities were collected from the Spatial Hazard Events and Losses Database (SHELDUS) for the United States. The primary data source for the events included in SHELDUS is the National Oceanic Atmospheric Administration (NOAA) National Climatic Data Center (NCDC) Storm Data publications. Therefore, qualitative and quantitative information regarding the flood fatalities has been obtained through the NCDC Storm Events Database. The NCDC supplies a collection of information on natural hazard events that have been consistently monitored since the 1950s. This data is received from the National Weather Service (NWS). However, some data is provided by other sources, such as the media, law enforcement, or other organizations and may not have been verified by the NWS. Therefore, the quality and accuracy cannot be guaranteed. The Mapping Website of the Texas Water Development Board (TWDB) provided data layers for the major rivers, major river basins and terrain in the state of Texas. The county data layer was obtained from Texas Natural Resource Information System (TNRIS). In addition, comparison maps of population growth in PDF format were acquired through the Texas State Data Center (TSDC). All data layers will be implemented in ArcGIS, a computer software program instrumental in producing static maps and performing both spatial and statistical analysis. METHODOLOGY The process used in collecting the necessary flood fatalities and event details is discussed first. To determine the number of flood fatalities and obtain details about these events, a search was first conducted in SHELDUS from 1950 to 2007. The events that resulted in at least one fatality were then compiled into an Excel spreadsheet and cross-referenced against the NCDC Storm Events Database. Only events that appeared in both databases were considered and used in analysis. The reports from the NCDC database were reviewed to determine qualitative and quantitative information for each of the flood events in the final dataset. Copies of the reports were saved in digital format and printed copies were compiled into a single binder. If available, the following information was recorded: year, county location, gender, age, type of incident death, latitude and longitude, other descriptive location information and event total dollar amount of property and crop damage. In addition, statistics were calculated for the recorded data and organized in tabular format. To conduct final analysis and map production, a number of GIS functions were employed. Data layers for Texas counties, rivers, river basins, and terrain were first imported into ArcGIS. Flood fatality frequency and damage totals for each county utilized keyboard entry. Fields were added to the county layer for total fatalities (FAT_TOT), fatalities that occurred during each decade (FAT_DECADE), property damage total (PROP_TOT), and crop damage total (CROP_TOT). The county layer was edited to input values for the newly created attribute data fields. Finally, graduated color maps based were created on the various attribute values for visual interpretation of flood fatality frequency and distribution and comparison to population growth maps. INTRODUCTION Flooding is a global phenomenon and is one of the most dangerous weather-related natural disasters in the world. More deaths are attributed to flash floods than any other weather-related disasters that occur in the United States and Texas accounts for the most flood-related deaths and damages (French 1983). There is an explicit need for a better understanding and knowledge about the magnitude, pattern, and circumstances surrounding fatalities caused by floods and in particular, flash floods. The International Flash Flood Laboratory (IFFL) is currently being developed under the James and Marilyn Lovell Center for Geographic Education and Hazards Research at Texas State University. The IFFL will function as a center of learning to improve flash flood research, knowledge dissemination, and risk communication. A comprehensive analysis based on available data regarding fatalities attributed to flooding in Texas counties may provide insight into these incidents, potentially reducing the occurrence of flood-related deaths. In addition, this study will play an integral role in establishing the IFFL. CriteriaFatalitiesPercentage Gender13574.2 Male8160.0 Female5440.0 Fatality Location10557.7 VE (Vehicle)8076.2 OU (Outside)54.8 OT (Other)11.0 IW (In Water)1615.2 MH (Mobile Home)32.9 Age9753.3 0 - 20 years3334.0 21 - 40 years2525.8 41 - 60 years1818.6 60 + years2121.6 Lat/Long Available189.9 RESULTS Based on the available data from SHELDUS and NCDC Databases, the final database includes a total of 182 fatalities, occurring in 112 separate events from 1993 to 2007. No fatality reported all of the criteria under study. Nearly all event records included descriptive information about fatality location, such as identifying water crossings and only 18 of the records had fatality location latitude and longitude. The number of flood events nearly doubled in the last two decades. However, frequency of fatality incidents only slightly increased, resulting in a decreased number of deaths per flood event. Property damage is reported most often with 71 of the 112 events reporting a total of over $225 million in damages. Crop damage reported only about $31.7 million in damages as a result of 14 flood events. Most flood fatalities are vehicle related, representing 76.92% or 80 deaths. Mobile home and outside/open area fatality location represented the fewest number of fatalities and accounted for only 2.80%. There is no real significant difference in age. However, males represented 60% of fatalities that reported gender. A distinct pattern exists in the spatial distribution of incidents and a number of physical environment factors can be identified. of By comparing decadal fatality maps to population growth maps, it is evident that a direct correlation exists between population growth and number of fatalities. Counties that have experienced high population growth also experienced high number of flood related deaths and high levels of property damage. In addition, comparison to the Texas reference map, reveal that counties located within close proximity of major rivers or in areas of elevation change account for the vast majority of flood related fatalities. The Central Texas region is the most vulnerable area susceptible to flood-related deaths and damage. CONCLUSION The use of the SHELDUS and NCDC databases provided for the study and analysis of flood fatality frequency, geographic distribution of fatality location, and the circumstances and associated factors that contributed to flood fatality. The variables that describe those most likely to be at risk are males under the age of 20 and over the age of 60 traveling by automobile in flooded areas. Those most likely at risk are located in areas of higher population growth and elevation changes in the Central Texas region. Texas has the largest road system of any state which possibly contributes to high flood fatality frequency. Another possible explanation for the pattern and frequency of flood fatality incidents is a lack of public awareness of the risks and dangers associated with flash floods in Texas. The results of this study are subject to limitations. First, there is a limited amount of accurate and complete data for all fatalities. Previous research suggests that a lack of accurate data and incomplete methods for collecting and publishing such data lead to inconclusive results that are not representative of the entire population (French 1983). Although flood fatality data does exist, efforts to collect and publish this information have not been thorough or consistent (Frech 2005). Better qualitative and quantitative data, especially with regards to exact geographic location, will provide insight into the causes and factors associated with flood fatalities. It is imperative that this information be made more readily available for areas at risk can properly implement mitigating and preventative measures. In addition, there is a need for the development of a standardized method for reporting the location and circumstances of flood fatalities. This study will prove to be an integral part of the International Flash Flood Laboratory to be interpreted by people in the academic, hazards, and planning communities. The results may be used to allow for responsible, safe, and sustainable development in lower water areas, and high growth areas. This study, serving as a foundation, may also be built upon by others in the future to map out the vulnerability of other geographic areas to the hazards posed by flash floods. Findings from this research may provide insight into the patterns of flood related events and can be used as predictors for future loss of life associated with such events. A better understanding of where these deaths are occurring can lead to better mitigating efforts so that mortality at low water crossings can be reduced, if not eliminated. CriteriaEventsFatalitiesPercentage 1990s Total408446.2 1993352.7 1994684.4 199562413.2 1996573.8 199711158.2 199872312.6 1999221.1 2000s Total729853.8 2000794.9 2001784.4 200210137.1 2003221.1 20048105.5 2005352.7 2006784.4 2007284323.6 Total112182100.0 Descriptive Data107 N/A 95.5 Property Damage7163.4 Crop Damage1412.5 Injuries2623.2 Texas State University – San Marcos GEO 4427 – Spring 2009 Department of Geography Dr. Giordano Nancy Tian EVENT TOTALS FATALITY CIRCUMSTANCE DATA County Total Fatalities 1990s2000s Property Damage Crop Damage Austin202 $ 8,000 $ - Bandera321 $ 5,020,000 $ 1,000,000 Bastrop202 $ 100,000 $ - Bell927 $ 467,500 $ - Bexar271611 $ 23,000,000 $ 105,000 Blanco202 $ 100,000 $ - Brown110 $ 250,000 $ 50,000 Burnet202 $ 137,000,000 $ - Chambers110 $ 50,000 $ 5,000 Comal101 $ - Comanche303 $ 42,500 $ - Cooke303 $ 28,000,000 $ - Coryell404 $ 75,000 $ - Dallas23212 $ 350,000 $ - Denton440 $ - Ector110 $ 300,000 $ - Edwards211 $ 85,000 $ - El Paso431 $ 250,000 $ - Ellis202 $ 20,000 $ - Fort Bend211 $ 35,000 $ - Freestone101 $ 12,000 $ - Gillespie303 $ 20,000 $ - Goliad110 $ - Grayson312 $ 20,000,000 $ - Grimes303 $ 50,000 $ 5,000 Guadalupe642 $ 5,080,000 $ 100,000 Hardeman110 $ - Harris716 $ 3,812,000 $ - County Total Fatalities 1990s2000s Property Damage Crop Damage Hays321 $ 2,500,000 $ 50,000 Howard202 $ 10,000 $ - Jefferson202 $ 5,010,000 $ - Kendall505 $ 300,000 $ - Kerr211 $ 25,000 $ - Lavaca110 $ 1,000,000 $ 50,000 Liberty101 $ 10,000 $ - Mason101 $ 15,000 $ - McLennan211 $ 20,000 $ - Medina110 $13,000,000 $ 100,000 Milam101 $ 10,000 $ - Navarro202 $ 10,000 $ - Nueces220 $ - Palo Pinto101 $ 10,000 $ - Parker211 $ 5,000 $ - Potter202 $ 73,000 $ - Real440 $ 1,066,667 $ 20,000 Scurry110 $ 50,000 $ - Shackelford101 $ 3,000,000 $ - Shelby110 $ - Somervell202 $ - Starr110 $ 50,000 $ - Tarrant826 $ 265,000 $30,000,000 Taylor101 $ 30,000 $ - Travis422 $ 3,050,000 $ 150,000 Uvalde211 $ 30,000 $ 90,000 Webb101 $ 1,000,000 $ - Williamson303 $ 500,000 $ - FATALITY TOTALS BY COUNTY AND DECADE – TOTAL CROP AND PROPERTY DAMAGE BY COUNTY Produced by the Texas State Data Center National Climatic Data Center (NCDC). 2009. Event Record Details. http://www4.ncdc.noaa.gov/cgi-win/wwcgi.dll?wwEvent~Storms. Last accessed 04.18.09. Frech, Marshall. 2005. Flood Risk Outreach and the Public’s Need to Know. Journal of Contemporary Water. Research and Education. 130 (61-69). French, Jean, R. Ing, S. Von Allmen, and R. Wood. 1983. Mortality from Flash Floods: a Review of National Weather Service Reports, 1969-81. Public Heath Reports. 98.6 (584-588). National Climatic Data Center (NCDC). 2009. Event Record Details. http://www4.ncdc.noaa.gov/cgi-win/wwcgi.dll?wwEvent~Storms. Last accessed 04.18.09. Texas Natural Resource Information Systems. “Data Search and Download.” http://tnris.org/datadownload.jsp. Last accessed 04.18.09. Texas Water Development Board. “Mapping & GIS Data Website.” http://www.twdb.texas.tx.us/mapping/gisdata.asp. Last accessed 04.18.09. Spatial Hazard Events and Losses Database for the United States (SHELDUS). 2009. s.v. “Flooding, 01/01/1950 – 12/31/2007, United States,” http://webra.cas.sc.edu/hvriapps/sheldus_setup/sheldus_results.aspx. Last accessed 04.18.09. Rebecca Whitton – Project ManagerCameron Howitt – Assistant Project Manager Justin Briseno – GIS AnalystMichael Stanley – Web Designer