An Analysis of U.S. fatalities of “Superstorm” Sandy Using ... · An Analysis of U.S....
1
An Analysis of U.S. fatalities of “Superstorm” Sandy Using Socioeconomic and Exposure Metrics Craig A. Ramseyer, Y. Wang, J. Vanexel, J.M. Shepherd, and J.A. Knox Department of Geography, University of Georgia Top 10 Contributors PC1 PERCAP -0.751 QRICH -0.748 MEDAGE -0.736 QSERV 0.574 QINDIAN 0.681 QSPANISH 0.708 QRENTER 0.768 QPOVTY 0.775 QFHH 0.803 QED12LES 0.818 4% 13% 34% 20% 11% 5% 13% Cause of Death Hypothermia Blunt force trauma Drowning Tree fall Carbon Monoxide poisioning Fire related Other QINDIAN QASIAN QSPANISH QKIDS QPOP65O MEDAGE QFEMALE PPUNIT QRENTER QFHH NRRESPC QCVLUN PERCAP QRICH QPOVTY MC_RENT QMOHO QED12LES QFEMLBR QAGRI QSERV QSSBEN QMIGRA QURBAN HODENSTY RASTERVAL U SoVI -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 F2 (12.83 %) F1 (27.93 %) Variables (axes F1 and F2: 40.75 %) The authors find that 68 of the 105 deaths occurred within the storm surge buffer. 62 of those 68 fatalities occurred inside the NYC metro. The other 6 fatalities occurred along the New Jersey coastline. In addition, the three primary causes of deaths were drowning (34%), tree fall related (20%), and non-tree fall related blunt force trauma (13%). A principal component analysis (PCA) was completed to determine which social variables contributed the most to the social vulnerability of those communities that experienced fatalities. This showed that education, poverty, and age related variables were the most significant contributors to the variance. This study suggests that the climate scientist community must continue to improve public awareness of the lethality of storm surge. The skew toward elderly fatalities shows the need to target this vulnerable population subset during storm preparations. Multilingual weather watches and warnings need to be more readily available to help better prepare those non-English speaking populations. Future preparations also need to improve on strategies to educate low-income, lower educated populations of environmental hazards. 1) Develop a geographic database of Sandy related fatality data, 2) Determine the role that Sandy related exposures (wind, precipitation, storm surge) played in the spatial distribution of fatalities. 3) Determine the role that social vulnerability and variables used to determine social vulnerability played in the spatial distribution of fatalities. Research Objectives • Fatality data for the study were acquired from the New York Times (NYT). The NYT data included death location and cause of death fields. Death location was available for 105 fatalities. • This study used storm surge (FEMA MOTF), precipitation (NWS MPE), and wind reanalysis (NARR) data to infer exposure at each death location. • Social vulnerability data included the SoVI calculated for 2000 because of its availability at the census tract spatial scale. This study also analyzes each of the 32 socioeconomic variables used to calculate SoVI. • Fatality locations were spatially joined with the exposure metric data to determine the precipitation, wind, and storm surge impacts for each location. • Fatality locations were spatially joined with U.S. Census tracts in order to assess the socioeconomic makeup of the neighborhoods where fatalities occurred. • A PCA was conducted on SoVI, elevation (at fatality location), and the 32 socioeconomic variables used to calculate SoVI to determine which variables explain the most variance in the locations where fatalities occurred. Data & Methods Background On October 29, 2012 at 7:30 PM Post-Tropical Cyclone Sandy made U.S. landfall near Brigantine, NJ bringing substantial impacts to the highly populated I-95 corridor. Sandy damaged or destroyed at least 650,000 homes and led to power outages to 8.5 million customers. Damage estimates from 24 affected states from Sandy exceeded $50 billion. Vulnerability is widely used in the global change and variability literature, particularly to analyze effects of environmental hazards on populations and how vulnerability varies in both time and space (Cutter et al. 2003; Adger 2006; Myers et al. 2008). The vulnerability of a place to weather hazards can be defined as a function of exposure (extent, frequency, severity) and the sensitivity of the population. Sensitivity is determined by the capability of a population to prepare for, respond to, cope with, recover from, and adapt to hazards (Cutter 2000). Conclusions Fig. 1. Radar MPE 3 day total precipitation from Sandy. Fig. 2. NARR Reanalysis sustained wind speeds. Fig. 5. Causes of death for 105 Sandy fatalities with death location data. Drowning was the cause of death for nearly 1 in 3 fatalities related to Sandy. While blunt force trauma and tree fall related causes of death accounted for another 1/3 of fatalities. Fig. 4. The location of all fatalities in our database and the locations of the fatalities that occurred in the storm surge buffer (red triangles). Clustering of storm surge fatalities were noted for the coastal census tracts of Staten Island (analysis not shown). Sources: NYT, FEMA MOTF Fig. 3. The locations of each fatality, stratified by cause of death. Drowning and hypothermia are the primary causes of death in coastal areas while inland area fatalities were primarily caused by blunt force trauma and tree falls (Primary Source: NYT). Figs. 7 and 8. Results of principal component analysis show that principal component 1 accounts for 30% of the variance in the census tracts where fatalities occurred. The top contributors to PC1 were education related variables (QED12LES), economic status (PERCAP, QRICH), and poverty (QPOVTY, QRENTER, QSERV). Another top contributor was percent Spanish-speaking. Fig. 6. Age distribution of Sandy fatalities. The data show a skew towards elderly populations. Younger populations still were at risk despite having lower frequencies of fatalities. References • Adger, W.N., 2006: Vulnerability. Global Environ. Change, 16, 268-281. • Cutter, S. L., Boruff, B. J. and Shirley, W. L., 2003: Social Vulnerability to Environmental Hazards. Social Science Quarterly, 84, 242–261. • Cutter, S.L., J.T. Mitchell, and M.S. Scott, 2000: Revealing the vulnerability of people and places: a case study of Georgetown County, South Carolina. Annals of the AAG, 90, 713-737. • Fulton, R.A., 2002: Activities to improve WSR-88D radar rainfall estimation in the National Weather Service. In: Proceedings of the Second Federal Interagency Hydrologic Modeling Conference, Las Vegas, Nevada, July 28–August 1, 2002. • Jayakrishnan, R., Srinivasan, R., Arnold, G.J., 2004: Comparison of rain gauge and WSR-88D Stage III precipitation data over the Texas-Gulf basin. Journal of Hydrology, 292, 135–152. • Mesinger, F. and Coauthors, 2006: North American Regional Reanalysis. Bull. Amer. Meteor. Soc., 87, 343-360. • Myers, C.A., T. Slack, and J. Singlemann, 2008: Social vulnerability and migration in the wake of disaster: The case of Hurricanes Katrina and Rita. Popul. Environ., 29, 271-291. Contact: Craig A. Ramseyer, Climatology Research Laboratory, University of Georgia ([email protected]) Results
Transcript of An Analysis of U.S. fatalities of “Superstorm” Sandy Using ... · An Analysis of U.S....
An Analysis of U.S. fatalities of “Superstorm” Sandy Using Socioeconomic and Exposure Metrics
Craig A. Ramseyer, Y. Wang, J. Vanexel, J.M. Shepherd, and J.A. Knox Department of Geography, University of Georgia
Top 10 Contributors
PC1
PERCAP -0.751
QRICH -0.748
MEDAGE -0.736
QSERV 0.574
QINDIAN 0.681
QSPANISH 0.708
QRENTER 0.768
QPOVTY 0.775
QFHH 0.803
QED12LES 0.818
4%
13%
34%
20%
11% 5% 13%
Cause of Death Hypothermia
Blunt force trauma
Drowning
Tree fall
Carbon Monoxide
poisioningFire related
Other
QINDIAN
QASIAN
QSPANISH
QKIDS
QPOP65O
MEDAGE
QFEMALE
PPUNIT
QRENTER
QFHH
NRRESPC QCVLUN
PERCAP
QRICH
QPOVTY
MC_RENT
QMOHO QED12LES
QFEMLBR
QAGRI
QSERV
QSSBEN
QMIGRA
QURBAN
HODENSTY RASTERVAL
U
SoVI
-1
-0.75
-0.5
-0.25
0
0.25
0.5
0.75
1
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1
F2
(12.
83 %
)
F1 (27.93 %)
Variables (axes F1 and F2: 40.75
%)
The authors find that 68 of the 105 deaths occurred within the storm surge buffer. 62 of those 68 fatalities occurred inside the NYC metro. The other 6 fatalities occurred along the New Jersey coastline. In addition, the three primary causes of deaths were drowning (34%), tree fall related (20%), and non-tree fall related blunt force trauma (13%). A principal component analysis (PCA) was completed to determine which social variables contributed the most to the social vulnerability of those communities that experienced fatalities. This showed that education, poverty, and age related variables were the most significant contributors to the variance. This study suggests that the climate scientist community must continue to improve public awareness of the lethality of storm surge. The skew toward elderly fatalities shows the need to target this vulnerable population subset during storm preparations. Multilingual weather watches and warnings need to be more readily available to help better prepare those non-English speaking populations. Future preparations also need to improve on strategies to educate low-income, lower educated populations of environmental hazards.
1) Develop a geographic database of Sandy related fatality data, 2) Determine the role that Sandy related exposures (wind, precipitation, storm surge) played in the spatial distribution of fatalities. 3) Determine the role that social vulnerability and variables used to determine social vulnerability played in the spatial distribution of fatalities.
Research Objectives
• Fatality data for the study were acquired from the New York Times (NYT). The NYT data included death location and cause of death fields. Death location was available for 105 fatalities.
• This study used storm surge (FEMA MOTF), precipitation (NWS MPE), and wind reanalysis (NARR) data to infer exposure at each death location.
• Social vulnerability data included the SoVI calculated for 2000 because of its availability at the census tract spatial scale. This study also analyzes each of the 32 socioeconomic variables used to calculate SoVI.
• Fatality locations were spatially joined with the exposure metric data to determine the precipitation, wind, and storm surge impacts for each location.
• Fatality locations were spatially joined with U.S. Census tracts in order to assess the socioeconomic makeup of the neighborhoods where fatalities occurred.
• A PCA was conducted on SoVI, elevation (at fatality location), and the 32 socioeconomic variables used to calculate SoVI to determine which variables explain the most variance in the locations where fatalities occurred.
Data & Methods
Background On October 29, 2012 at 7:30 PM Post-Tropical Cyclone Sandy made U.S. landfall near Brigantine, NJ bringing substantial impacts to the highly populated I-95 corridor. Sandy damaged or destroyed at least 650,000 homes and led to power outages to 8.5 million customers. Damage estimates from 24 affected states from Sandy exceeded $50 billion. Vulnerability is widely used in the global change and variability literature, particularly to analyze effects of environmental hazards on populations and how vulnerability varies in both time and space (Cutter et al. 2003; Adger 2006; Myers et al. 2008). The vulnerability of a place to weather hazards can be defined as a function of exposure (extent, frequency, severity) and the sensitivity of the population. Sensitivity is determined by the capability of a population to prepare for, respond to, cope with, recover from, and adapt to hazards (Cutter 2000).
Conclusions
Fig. 1. Radar MPE 3 day total precipitation from Sandy.
Fig. 2. NARR Reanalysis sustained wind speeds.
Fig. 5. Causes of death for 105 Sandy fatalities with death location data. Drowning was the cause of death for nearly 1 in 3 fatalities related to Sandy. While blunt force trauma and tree fall related causes of death accounted for another 1/3 of fatalities.
Fig. 4. The location of all fatalities in our database and the locations of the fatalities that occurred in the storm surge buffer (red triangles). Clustering of storm surge fatalities were noted for the coastal census tracts of Staten Island (analysis not shown). Sources: NYT, FEMA MOTF
Fig. 3. The locations of each fatality, stratified by cause of death. Drowning and hypothermia are the primary causes of death in coastal areas while inland area fatalities were primarily caused by blunt force trauma and tree falls (Primary Source: NYT).
Figs. 7 and 8. Results of principal component analysis show that principal component 1 accounts for 30% of the variance in the census tracts where fatalities occurred. The top contributors to PC1 were education related variables (QED12LES), economic status (PERCAP, QRICH), and poverty (QPOVTY, QRENTER, QSERV). Another top contributor was percent Spanish-speaking.
Fig. 6. Age distribution of Sandy fatalities. The data show a skew towards elderly populations. Younger populations still were at risk despite having lower frequencies of fatalities.
References • Adger, W.N., 2006: Vulnerability. Global Environ. Change, 16, 268-281. • Cutter, S. L., Boruff, B. J. and Shirley, W. L., 2003: Social Vulnerability to Environmental Hazards. Social Science Quarterly, 84, 242–261. • Cutter, S.L., J.T. Mitchell, and M.S. Scott, 2000: Revealing the vulnerability of people and places: a case study of Georgetown County, South Carolina.
Annals of the AAG, 90, 713-737. • Fulton, R.A., 2002: Activities to improve WSR-88D radar rainfall estimation in the National Weather Service. In: Proceedings of the Second Federal
Interagency Hydrologic Modeling Conference, Las Vegas, Nevada, July 28–August 1, 2002. • Jayakrishnan, R., Srinivasan, R., Arnold, G.J., 2004: Comparison of rain gauge and WSR-88D Stage III precipitation data over the Texas-Gulf basin.
Journal of Hydrology, 292, 135–152. • Mesinger, F. and Coauthors, 2006: North American Regional Reanalysis. Bull. Amer. Meteor. Soc., 87, 343-360. • Myers, C.A., T. Slack, and J. Singlemann, 2008: Social vulnerability and migration in the wake of disaster: The case of Hurricanes Katrina and Rita.
Popul. Environ., 29, 271-291.
Contact: Craig A. Ramseyer, Climatology Research Laboratory, University of Georgia ([email protected])
Results
