EIIP Virtual Forum Presentation — June 13, 2007
Estimating Disaster Resilience
The Social Vulnerability Index of the United States (SoVI)
Susan L. Cutter, PhD
Carolina Distinguished Professor of Geography
Director, Hazards & Vulnerability Research Institute
University of South Carolina
Amy Sebring
EIIP Moderator
The following version of the transcript has been edited for easier reading and comprehension. A raw, unedited transcript is available from our archives. See our home page at http://www.emforum.org
[Welcome / Introduction]
Amy Sebring: Good morning/afternoon everyone. Welcome to the EIIP Virtual Forum! Thank you for joining us today. Avagene is on travel and could not be with us, but sends her regards. Lori Wieber is here to help us out. Our topic today is "Estimating Disaster Resilience: The Social Vulnerability Index of the United States (SoVI)."
Now it is my pleasure to introduce today's guest. Dr. Susan Cutter is a Carolina Distinguished Professor of Geography at the University of South Carolina, and Director of the Hazards & Vulnerability Research Institute. Her primary research interests are in the area of vulnerability science--what makes people and the places where they live vulnerable to extreme events and how this vulnerability is measured, monitored, and assessed.
Dr. Cutter has also led post-event field studies of the role of geographic information technologies in rescue and relief operations in (September 11th World Trade Center attack) and studies of evacuation behavior from Three Mile Island (1979), Hurricane Floyd (1999), and the Graniteville, SC train derailment and chlorine spill (2005).
Most recently (2006) she has led a Hurricane Katrina post-event field team to examine the geographic extent of storm surge inundation along the Mississippi and Alabama coastline and its relationship to the social vulnerability of communities. Further biographic details, as well as links to her book " Hazards, Vulnerability and Environmental Justice" and related Web sites are provided on today's Background Page.
Welcome Susan and thank you for joining us today. I now turn the floor over to you to start us off please.
[Presentation]
Susan Cutter: Thank you Amy and welcome everyone.
Losses from natural disasters are escalating in the United States and have taken both an economic toll as well as a human one (lives lost and injuries). The underlying question is whether the severity and number of events is increasing or whether society is simply becoming more vulnerable to natural hazards.
At the moment, there is no systematic way to assess vulnerability and compare it from place to place, let alone determine whether it has increased or decreased. One starting point is to look at social vulnerability, that is, those characteristics of people and social groups that make them susceptible to the effects of natural hazards and also reduce their ability to adequately respond and recover.
Our collective research experience in the hazards and disasters field provides ample evidence for understanding the characteristics of people and social groups that make them more vulnerable.
[Slide 1]
For example, we know from our post-event field studies that race and ethnicity increase vulnerability in a couple of ways. First, minority populations tend to live in housing that is located in less desirous and often hazardous locations because it is affordable, and are therefore at a higher risk than non-minorities before an event. Second, there may be language and cultural barriers that inhibit the understanding of warning messages or affect access to post-disaster recovery resources.
Age also increases social vulnerability, specifically among the elderly and the very young. Finally, socioeconomic status both increases and decreases social vulnerability in the following ways. Wealthier communities have more capacity to absorb losses and recover because of insurance and access to other financial resources, but at the same time, there is more personal property to lose. Poorer communities do not have such capacity to absorb losses.
What is less well known is how these characteristics interact to produce the social vulnerability and more importantly how they change over space and time. Social vulnerability has many different dimensions and is not easily captured using a single variable. At the same time, social vulnerability certainly varies over time and across regions.
So, in order to assess social vulnerability, a methodology was required that used a consistent set of measures that could be replicated across time and space. This led to the development of the Social Vulnerability Index or SoVI. The Social Vulnerability Index (SoVI) provides a county-level assessment of those pre-existing social conditions that contribute to vulnerability to environmental hazards.
SoVI is based on US Census data and provides a socioeconomic and demographic profile of the county. It does not include any measures of exposure, that is, the threat agent or natural event (e.g. hurricane, flood). To construct SoVI, we collected 42 Census variables that represent what the research literature says contribute to a community's ability to prepare for, respond to, and recover from hazards. We used a statistical technique called principal components analysis to reduce the variables into a smaller set of indicators and from there we simply summed the indicator values to create the overall SoVI score.
The initial construction of SoVI used 1990 Census data, produced 12 simplified indicators or factors, and explained 78% of the variance in the original data set. We have subsequently calculated SoVI for 1960, 1970, 1980, and 2000 using the same methodology and variables (with some minor adjustments due to changing Census definitions).
We find that the number of factors remains relatively constant (11 or 12 for all decades, except 1960 which only had 9), and the percentage of explained variance ranges from 73%-78%. What this means is that SoVI is quite a robust metric for understanding and explaining social vulnerability. We also found that two factors were in the top three for every decade: socioeconomic status and age (both young and old).
To visually compare the SoVI scores for counties at the national level, we mapped them. Because we are interested in the extremes in vulnerability, we chose to highlight the most vulnerable populations (the top 20%), the least vulnerable populations (the bottom 20%), while still displaying the remaining counties in the country with average social vulnerability. In this visualization, you can see the geographic distribution by county for the entire US for 2000.
[Slide 2]
As you can see from the map, there are distinct clusters of high vulnerability counties (red). These are located along the US-Mexican border, in California, in the lower Mississippi River basin, and in Florida. The driving factors of social vulnerability vary geographically; in the border areas, immigration and ethnicity are the primary contributors to the social vulnerability, in the lower Mississippi basin it is race and poverty, while in Florida it is an elderly population.
Given the interest in SoVI, especially in light of the requirements for hazard vulnerability assessments under DMA 2000, we created a web application that provides the 2000 SoVI by state.
[Slide 3]
The website (http://www.cas.sc.edu/geog/hrl/sovi.html) also provides an overview of how SoVI was constructed. More importantly, we've added a frequently asked questions page to help users. You can query the 2000 SoVI by an individual state, or by FEMA region.
For each state (or FEMA region) we provide a PDF version showing two maps. The first map compares the social vulnerability of the counties within the state to the whole country (a simple zooming in from the national map from Slide 2), while the second map provides the SoVI based on a re-mapping of only those counties within the state, and therefore shows the most vulnerable of the most vulnerable within the state itself. We also provide a table that shows the national percentile ranking-the higher the percentile, the more vulnerable the county.
[Slide 4]
As you can see from the Florida example, Broward County is in the highest category on the national map (dark red), yet when compared to counties within Florida and re-mapped, it slips down to the second highest category. In this way, planners can get both a state and a national perspective. And, the SoVI scores and national percentiles are downloadable for further use.
SoVI can help determine which places may need more pre-impact preparedness based on their existing level of social vulnerability. It can also help assess where additional resources (human and financial) may be needed to facilitate longer term recovery after an event, again based on the social characteristics of the place.
The best example of this is New Orleans and Hurricane Katrina. After Hurricane Katrina, we developed SoVI for Orleans Parish at the census tract level to see where the most vulnerable populations were located. Not surprising, areas like the Lower Ninth Ward stand out in terms of high social vulnerability, while the French Quarter is low.
[Slide 5]
When compared to the level of flood inundation (a measure of exposure), we can see a slightly different pattern, where some sections of the city were severely flooded, while others had very little water or none at all. If you then overlay the exposure (flooding), with the pre-existing social vulnerability, you can begin to estimate which areas might take longer to recover from this disaster.
[Slide 6]
Those areas on the slide that are colored in dark grey illustrate not only the highest levels of social vulnerability, but also those communities that were the most flooded. Based on this we would predict that communities like the Lower Ninth Ward, Pontchartrain Park, Gerttown, or Venetian Isles (see the yellow stars) will take longer to recover from this disaster than the French Quarter or Garden District.
This is but one illustration of how SoVI can be used to understand social vulnerability and how that influences a community's ability to prepare for, respond to, and recover from hazards and disasters.
Now that I have provided this lengthy introduction it is time for questions and comments, and I will turn the floor back over to our Moderator.
Amy Sebring: Thank you very much Susan. Now, to proceed to your questions and comments.
[Audience Questions & Answers]
Question:
Rick Tobin: I'm stunned at this work. Do you see many changes once the new census comes in, especially with Katrina relocations?
Susan Cutter: For New Orleans, we don't have data yet but we're thinking that the out-migration will change the relative levels of vulnerability for the Parish.
Question:
Sophi Beym: Dr. Cutter, I have a MS in GIS and use GIS for my Emergency Management and Mitigation analysis. My question is does this interface with HAZUS Software?
Susan Cutter: Sophi, not yet but we are working with the HAZUS development team to incorporate a SoVI type measure.
Question:
Jaye Compton: As an emergency logistician it is important that plans are in place to provide emergency supplies to local populations following catastrophic disasters. Have you seen this SOVI data used in support of planning such as this? Are there any examples you can point me to where it has been used for something like this?
Susan Cutter: Jaye, not that I'm aware of, but this is precisely how we think it could be used.
Question:
Roger Fritzel: It appears that combining SoVI with other information on Critical Infrastructure could enhance regional vulnerability studies. Have you discussed such use with any people working on critical infrastructure protection?
Susan Cutter: Roger, we've begun to think about that within the context of urban places, but haven't gone much beyond that at the moment.
Question:
Ron Lopez: Our experience in the San Francisco public health department reveals that social vulnerability in disaster is inextricably tied to social support systems that often exist INDEPENDENTLY of socio economic status. Can the variable of social support networks (friends, care givers, neighbors, family, etc.) be factored into the statistical model you cite?
Susan Cutter: Ron, the short answer is yes, but we might consider those support systems as more reflective of resilience.
Question:
Scott Westlund: Who initiated the vulnerability assessment? Was including perspectives from individuals or groups within a community part of the process?
Susan Cutter: Scott, the formulation was an attempt to develop a standard way of measuring vulnerability so that we could compare places across the nation. There was no community involvement in this; rather, we worked from the disaster literature and field studies conducted over the past 50 years.
Question:
Ron Cohea: How aware are local and state emergency managers of this tool?
Susan Cutter: We think they are becoming more aware, hence our decision to make this available on the web. We work with a number of state and local managers and have assisted them in the past.
Question:
Barbara Fay: Has this been published to the LLIS site yet?
Susan Cutter: Barbara, not that I'm aware of.
Question:
Bill Thompson: I am curious if religious backgrounds were taken into account with the data. As we found at Nickel Mines, even when resources were available, beliefs took precedence and made the recovery much more difficult. Thus, when resources are available, if given the diverse backgrounds and beliefs that could interfere, and knowing this, could preplan it more concisely.
Susan Cutter: No, because we relied on Census information.
Question:
John Boyle: As a planner I recognize that Orleans Parish started developing on the high ground along the Mississippi River. Later the marsh area between Lake Pontchartrain and the river was drained and developed. Can SoVI include historic community development and environmental parameters to display "land development miscues," as well as help guide long term recovery?
Susan Cutter: John, we can look at SoVI historically and monitor changes in the social geography, and from there one could also look at historical land use patterns.
Question:
Rick Tobin: Have you shared this data with insurance companies and hospital systems/HMOs? I'll bet they'd be very interested to see this data. I know I can use it for some projects I'll be working on for mass care and shelter.
Susan Cutter: Rick, we haven't directly shared, but hope it is useful to all stakeholders involved in emergency management.
Question:
Amy Sebring: Susan did you see a trending up between the decades you studied?
Susan Cutter: We're finding that the range of vulnerability scores is narrowing than in the past, meaning that the extremes are becoming less pronounced. Yet, we're also seeing regional shifts in the most and least vulnerable.
Question:
Sophi Beym: Last question, thank you. Dr Cutter - would you categorize this as a form of Environmental Justice? Vulnerability Justice per se?
Susan Cutter: Oohhh, I like the term. Yes, this is very much in the EJ arena, and is an area where we've also done considerable work.
Question:
Davis Patterson: I'm curious about the narrowing of the gap, given the greater extremes in wealth distribution over time in the US. Why do you think the extremes in vulnerability are less?
Susan Cutter: While the income gap is widening, there are more people in the middle and lower wage categories, which might help explain this.
Question:
Lori Wieber: Have you developed, or do you plan to develop any risk analyses for the U.S. that overlays various hazards (e.g., flood hazard, earthquake hazard) with the SoVI?
Susan Cutter: We are doing this on a place by place basis. For example, we've done flooding for New Orleans, hurricane inundation modeling for Charleston, SC, tsunami risk for coastal Oregon and many more examples. We've not done a systematic analysis for the whole US since much of the hazards data are incomplete at present.
Question:
Jamie Mitchem: Based on the SOVI scores and past research, can you suggest some ways that social vulnerability could be reduced? In other words, what (if anything) can we do to mitigate social vulnerability?
Susan Cutter: Clearly some of the drivers are socioeconomic status and differential access to resources. This may mean that reducing social vulnerability to hazards means more resources placed in those communities that are more vulnerable, rather than evenly distributed resources. We could also begin to think about how to make the most vulnerable communities more resilient.
Question:
Ron Cohea: Is the American Red Cross also aware of this? It seems like it would help them in their relief efforts.
Susan Cutter: Ron, I'm not sure if the Red Cross is aware or not.
Question:
Amy Sebring: Susan, as shown on your slides on New Orleans, you mentioned that you developed data to the census tract level, which is the resolution most useful to local planners. Do you have tract level data for the remainder of the U.S. or instructions on how to do this oneself?
Susan Cutter: The methodology for SoVI is available on the web site along with a publication that goes into detail on its construction. To do this for the whole country at the tract level would be a HUGE task, but if someone is willing to pay, it could be accomplished. One side note is that it is now 2007 and much of the Census data is becoming old, especially in high growth areas like the coast.
Question:
Rick Tobin: I'm working on the ASTM Work Group developing the new international standard for school emergency preparedness. Your project can really help predict the vulnerability of school districts to recover after major impacts. I hope you will keep that in mind as part of your stakeholder outreach. Just so many possible uses. An amazing effort on your part. How do we keep in touch with you and your team?
Susan Cutter: Rick, we're always here and can be reached at http://www.cas.sc.edu/geog/hrl.
[Closing]
Amy Sebring: Good note to finish up on. Let's wrap it up for today. Excellent questions everyone. Thank you very much Susan for an excellent job. We hope you enjoyed the experience today.
Susan Cutter: Thank you all for your great questions.
Amy Sebring: Please stand by a moment while we make a couple of quick announcements Again, the formatted transcript will be available later today. If you are not on our mailing list and would like to get notices of future sessions and availability of transcripts, just go to our home page to subscribe. If your organization is interested in becoming an EIIP Partner, please go to our home page and click on "Partnership for You"
Thanks to everyone for participating today. We stand adjourned but before you go, please help me show our appreciation to Dr. Cutter for a fine job.