Preliminary Summary Report of Hurricane Katrina Deceased Victim Recovery Locations in Louisiana

Transcription

1 Preliminary Summary Report of Hurricane Katrina Deceased Victim Recovery Locations in Louisiana Ezra Boyd Graduate Research Assistant LSU Hurricane Public Health Center Note: The information contained in this preliminary report remains subject to future revision. Hurricane Katrina made landfall along the Mississippi-Louisiana state line on Monday, August 8, 5. The massive storm surge stretched from La Place, La to Pensacola, Fl, inundating an estimated miles of coast. While the levees that protect New Orleans were not overtopped, multiple failures in the federally constructed hurricane protection system led to catastrophic flooding there. 8% percent of Orleans parish and % of St. Bernard parish was inundated by water that flowed through breeched floodwalls and eroded earthen levees. This flood necessitated the complete evacuation of New Orleans, left a half-million people homeless, and took the lives of approximately,5 Louisiana residents. However, the true loss-of-life of this disaster remains uncertain. One year later, the Louisiana Department of Health and Hospitals (DHH) reports nearly missing and the New Orleans Fire Department continues to recover the remains of deceased victims from the rubble of destroyed houses. This report summarizes what is currently known about the deceased. Current Report of Deceased and Missing The State Medical Examiner, Dr. Louis Cataldie, has spearheaded DHH s effort to reunite the living and recover and identify the deceased. Initially, the Disaster Mortuary Operational Response Team (DMORT) and local parish coroners worked in conjunction with the State Medical Examiner s office. However, on March, 6 DMORT ceased operation and closed their $7 million dollar morgue and identification center less than four months after it was opened. Currently, the DHH Family Assistance Center handles reports of missing persons, while victim identification has returned to the parish coroners. DHH provides the official figures on dead and missing on their Katrina Missing website ( As of

2 August, 6 this site listed,464 deceased victims and notes that the cases of 5 missing have been turned over to law enforcement. Of the confirmed dead,,8 victims perished within Louisiana, while 46 victims perished outside of the state of Louisiana. Of the,8 in state deaths, 875 victims were recovered and examined by the Medical Examiner s Office. The remaining victims were handled by parish coroners: by Orleans parish coroner and by other parish coroners. On August, the cases of 5 people who remain listed as missing were turned over to local law enforcement agencies. At least five persons have been located alive. Recovery Locations Dataset In mid-october, the LSU Hurricane Center (HC) established a collaborative effort with the State Medical Examiner s office. As part of this collaboration, DHH provided the HC with data on the recovery locations for the deceased victims. While continuously updated, a version of the dataset obtained on April, 6 lists the recovery locations of 87 victims through out the State of Louisiana. This corresponds to 7% of the total number of victims recovered within the State. Of the 87 victims listed in the dataset, 786 have been geocoded. This corresponds to 7% of the victims recovered within the state. The dataset includes the following information: Date of Recovery, Type of Facility, Facility Name, Street Address, Nearest Cross Street, City, State, Zip, Parish, Comments, Transported By, and Scene. Each entry in the dataset describes the recovery of one victim. In some cases, multiple victims are recovered from a location. This current dataset is based on Receipts of Remains provided to DHH by DMORT and it replaces an early dataset provided by Kenyon International Emergency Services that has been deemed corrupt by the State Medical Examiner s office. The dataset indicates that most victims were recovered from Orleans parish and most victims were recovered from residences. Of the 85 recoveries for which a parish is listed, 58 list Orleans parish, followed by 8 in St. Bernard parish. 5 victims were recovered from East Baton Rouge parish and 6 from Jefferson parish. The dataset lists a structure type for 857 victims. Of these, 44 were recovered from residences, 5 from hospitals and other health care facilities, 6 from nursing homes, 57 from street locations, from the Convention Center, and 8 from the Superdome. The following tables summarize the recovery locations dataset and are based solely on the information provided in the dataset, while the appendix provides various maps depicting recovery locations and aggregate results.

3 Tables & : Total number of victims recovered per City and per Parish as listed on the dataset. City ALEXANDRIA ARABI BASTROP BATON ROUGE CHALMETTE CLINTON GONZALES HARVEY HOUMA KENNER LAFAYETTE LUTCHER MERAUX METAIRIE NEW ORLEANS OPELOUSAS PLAQUEMINE SAINT BERNARD UNKNOWN VIOLET Grand Total Total Parish ASCENSION EAST BATON ROUGE EAST FELICIANA IBERVILLE JEFFERSON LAFAYETTE MOREHOUSE ORLEANS RAPIDES ST JAMES ST LANDRY ST. BERNARD TERREBONNE UNKNOWN Grand Total Total

4 Table : Total number of victims recovered per structure type as listed on the dataset. Structure Type AIRPORT (*) BUSINESS CANAL CHURCH CONVENTION CENTER CORONER OFFICE EOC FUNERAL HOME HEALTH CARE HOSPICE HOSPITAL HOSPITAL HOSPITALIERE LAKE MEDICAL CENTER MORGUE NURSING HOME PARKING GARAGE PARKING LOT RESIDENCE SCHOOL SHERIFFS OFFICE STREET SUPERDOME TEMP MEDICAL CLINIC UNKNOWN Grand Total Total * The Louis Armstrong International Airport on the Western Edge of Greater New Orleans served as a field hospital and shelter during the post-storm emergency evacuation.

5 As noted, 786 of the recovery locations have been geocoded. The tables below were derived by merging the recovery locations shapefiles with other shapefiles. Parish shapefiles was obtained from the LA GIS Data CD, while the Greater New Orleans Data Center provided a Orleans Parish Neighborhood shapefile. Note that there are minor discrepancies with the above table for recovers per parish. Table 4: Victims recovered per parish as determined by overlaying the geocoded recovery locations with Census Tiger shapefile. Parish Orleans St. Bernard East Baton Rouge Jefferson Lafayette Terrebonne Ascension Morehouse Rapides East Feliciana Iberville Livingston St. James St. Landry Grand Total Total

6 The New Orleans Community Data Center provided shapefiles for planning districts and neighborhoods. The two tables below only apply to Orleans parish. Table 5: Total number of victims recovered per planning district as determined by overlaying the geocoded recovery locations with planning district shapefile. Planning District Algiers Central City/Garden District English Turn French Quarter/CBD Gentilly Lakeview Lower Ninth Ward/Holy Cross Marigny, Bywater, St. Claude, St. Mid-City New Orleans East New Orleans East Area Uptown/Carrollton Village de L'Est Area Grand Total Total Recovered

7 Table 6: Total number of victims recovered per neighborhood as determined by overlaying the geocoded recovery locations with neighborhood shapefile. Neighborhood Lower Ninth Ward Neighborhood Freret Mid-City Neighborhood Touro West Lake Forest Little Woods Milneburg Central City Neighborhood Lower Garden District Neighborhood Lakeview Neighborhood Seventh Ward Fillmore Tulane/Gravier St. Anthony St. Roch Central Business District St. Claude Viavant/Venetian Isles Read Blvd East West End Dillard Florida Area Hollygrove Pontchartrain Park Gentilly Terrace Broadmoor Fairgrounds Gentilly Woods Plum Orchard Holy Cross Milan Read Blvd West Treme'/Lafitte Bywater Neighborhood Desire Area Pines Village Bayou St. John Total Recovered Neighborhood Total Recovered Dixon French Quarter East Carrollton East Riverside Florida Development Gert Town Iberville Development Marlyville/Fountainebleau Navarre Village de l'est West Riverside Audubon B.W. Cooper Black Pearl City Park Fischer Project Irish Channel Lake Catherine Lakeshore/Lake Vista Leonidas Marigny Old Aurora St. Bernard Area St. Thomas Development Whitney Algiers Point Behrman Desire Development Garden District Lake Terrace & Oaks Lakewood McDonough New Aurora/English Turn Tall Timbers/Brechtel U.S. Naval Base Neighborhood Uptown Neighborhood Grand Total

8 Causes and Circumstances Unfortunately, very little can presently be said about the causes and circumstances of those who died during Hurricane Katrina. Operationally, DHH utilizes a broad definition of deceased individuals for whom they provided identification and other services: anyone from the affected parishes that died between August 8 and October for which the cause of death is obviously not Katrina related. Only in a handful of cases, mostly those related to legal concerns, have autopsies been completed, and these results have not yet been made public. In very rough terms, it appears that approximately one third of the fatalities were due to direct exposure with flood waters. Overlaying the recovery locations with a flood depth raster layer shows that approximately 5 victims were recovered from locations that flooded. However, approximately of these victims were recovered from hospitals, multistory apartment buildings, or the second story of homes. For most of these victims, the death did not result from exposure to flood waters. These figures indicate that roughly 5 of the victims died as a result of direct exposure to the flood hazard. Presumably, drowning is the cause of death for a majority of these victims while trauma due to flood debris impact is a cause for some of them. For the remaining two-thirds of the victims, the cause of death varies greatly, but the circumstance fall into one of two categories. For some victims, unmeet basic needs during the emergency phase, such as food, water, medicine, medical treatment and security, resulted in the death. For these victims, causes of death include starvation, dehydration, stroke, heart attack, and violence. For other victims, the circumstances created by long term displacement resulted in the death. Many of these deaths resulted most directly from accidents while others resulted from long term medical conditions. For these victims, the State Medical Examiner has determined that circumstances of displacement strongly influenced the conditions of the death. At this time, we do not have sufficient data to precisely determine totals for each category. Speaking in very rough terms, approximately 5 victims fit into each of the categories. Recovery Locations and Levee Breeches Proximity to a breech location influences basic flood hazard characteristics such as the rate of rise and flow velocity of the flood waters. In turn, these flood hazard characteristics influence the survivability of a person trapped in the flood waters. Since we do not yet have reliable data on these on these flood characteristics, buffer zones around the breach locations serves as a proxy. The rational is that the average flow velocity will be larger for a smaller buffer than for a larger buffer. At this point in the analysis, no real relationship is apparent in the result below.

9 Table 7: Victim recovery locations and proximity to breaches. Buffer Size Total Recovered 5 ft / Mile Mile 6 48 Percent of Total (*).76% 4.5%.55% Area of Buffer Zone (m^) Total Recovered / Area 4.E-7.6E-6.76E-7 Real-Time, Remote Estimation of Flood Fatalities During the emergency phase, the LSU Hurricane Center provided support for State and Local efforts to protect the lives of those trapped in the flooded regions and at the various evacuation centers. As part of this effort, we were tasked to provide a real-time estimate of the number of expected flood fatalities. Early efforts suffered from a lack of reliable flood data and resulted in the controversial estimate of,-,. However, by September, a more reliable estimate of 5- was made available. Estimating the number of fatalities during a flood event requires three key pieces of information. First, we need to know the number of people exposed to the flood. In practical terms, this is the number of residents living inside the flooded region minus the number that evacuate and the number that adequately shelter-in-place. Second, we need to know the distribution of the flood depths. Both of these variables vary geographically, so GIS is an important tool in flood fatality estimation. Finally, using a flood fatality function (also termed a dose-response relationship,) we estimate the flood fatality rate based on the depth of water. Multiplying the flood fatality rate by the exposed population gives an estimate of the number of flood fatalities. For better or worse, Hurricane Katrina provided an initial test of this approach to flood fatality modeling and the model passed the test. Once accurate flood data was included in the analysis, the model produced results that are accurate to the order-of-magnitude. Naturally, given the complexity of the processes involved (such as the effects of search and rescue,) order-of-magnitude accuracy is a reasonable expectation. As this research moves forward, the accuracy of the model will improve and the flood fatality model will prove an indispensable tool as America prepares for the impacts of global sea level rise and increased hurricane activity. Next time a major American city floods, the flood fatality model will do more than estimate the total number. It will help response officials save lives by identifying areas of greatest potential loss-of-life.

10 Table 8: Estimated number of flood fatalities from September, 6. Potential fatalities by parish: Jefferson: Orleans: St. Bernard: St. Tammany: Plaquemines: Min 57 6 Max Total: Future Research The tragedy in New Orleans left us with a dataset that we hope will help minimize the number of lives lost during future hurricanes and flood disasters. Research on the causes and circumstances of Katrina related fatalities in Louisiana aims to fulfill this goal. However, one year after the disaster, a number of important uncertainties remain: ) The dataset presented here represents the 7% of the victims whose bodies were recovered and examined as part of the joint state-federal effort. We know that a number of victims were processed by parish coroners or out-of-state and their recovery information is not included in this dataset. Obtaining this data has proven to be a challenge. ) The DHH dataset includes a number of unknowns and erroneous entries. For these victims, determining the location of death is difficult or impossible. ) We will never have medically determined cause-of-death data on the majority of the victims. Regardless of these uncertainties, a number of important results are still possible: ) How do the observed patterns in the victim recovery locations relate to what we know about the flood hazard characteristics? ) How do the observed patterns in the victim recovery locations relate to the underlying social-demographic characteristics of the exposed population? The questions will be addressed in future research using this dataset.

11 Appendix: Maps of Victim Recovery Locations The DHH dataset consists of a list of addresses, each listing corresponding to the recovery of one victim. Using ArcMap. and Street Maps USA, this list of addresses was geocoded to produce a point shapefile. Around recovery locations had to be geocoded manually, because the provided information, for example victim recovered from a field next the Almonaster Bridge, was not suitable for the software. Some of the maps below depict just the points, while others aggregate the points using various polygon shapefiles. Note that in the maps below, some points denote multiple recoveries and some locations remain subject to uncertainty. Figure : Locations of victims throughout the State of Louisiana.

12 Figure : Locations of victims throughout metropolitan New Orleans.

13 Figure : Deceased victims map zoomed on the flooded areas.

14 Figure 4: Victim Recovery locations and the levee system. Note that only one flood death in Louisiana occurred outside of the levee system.

15 Figure 5: Victim recovery locations, levees, and levee breaches. The preliminary levee breach shapefile provided by Hampton Peele and DeWitt Braud.

16 Figure 6: Total victims recovered per parish.

17 Figure 7: Total victims recovered per zip code.

18 Figure 8: Total victims recovered per New Orleans neighborhood. Neighborhood shapefile provided by the Greater New Orleans Data Center.

19 Figure : Total victims recovered per Census Blockgroup.