GEOG 483: Problem Solving with GIS
Brian G. Buchanan
Project 3: Locating Tornado Relief Sites
January 29, 2008
On May 3, 1999, a devastating series of tornados smashed their way through the central counties of Oklahoma. Using spatial analysis within a geographic information system (GIS), it is possible to graphically display which public facilities close to the affected areas represent candidate relief sites. In addition, it is possible to use a GIS to determine which of the affected areas are likely to have the greatest need for emergency relief. This information could be invaluable for relief workers because they could prioritize where they are needed the most.
In order to accomplish the above goals, I began by adding the County, TornadoPaths, and OkPlaces layers to an empty ArcView template. Population data and specific information on each tornado were added to the Counties and TornadoPaths shape files; creating two additional shape files labeled countiesnew and tornados. Based upon this data, a population density field could be added to the countiesnew attribute table. The population density was essential to determine the areas that would have the greatest need for emergency relief, which will be demonstrated later. The additional tornado information included data on the width of each of the tornados.
Figure 1: Screenshot showing the attribute table of the CadidateReliefSites showing the population density field. Produced using ArcView 9.2 on January 28, 2008.
In order to examine which of the public facilities (OkPlaces) were candidate relief sites, buffer zones were created around the paths of the tornados. These buffer zones were polygons that represented the width of each tornados path, and were saved as TornadoBuff. I then created an additional buffer around the TornadoBuff using the Buffer Wizard built into ArcView. This one mile buffer zone around each TornadoBuff was labeled as Emerg_rel_zones. I selected all the churches, hospitals, and schools within the Emergency Relief Zone. These were the cadidate relief sites and are are labeled and shown on the figure below.
Figure 2: Overview of the 1-mile Emergency Relief buffer zones around the tornados and the Candidate Relief Sites. Produced by ArcView 9.2 on January 28, 2008.
The second goal of this project was to determine the areas that were likely to have the greatest need for emergency relief (King 2007). To do this, I created a spatial intersection between the Emerg_rel_zones and CountiesNew layers. Specifically, I was interested in combining the Emergency Relief Zones with the population density attribute field that I had previously created within the CountiesNew attribute field. This intersection created a new layer labeled ReliefPriorityZone. I created a thematic map legend within this layer using graduated colors. This color ramp demonstrated the severity of the tornados and theoretically, the areas that were most likely to have the greatest need for emergency relief.
Figure 3: Close-up view of Oklahoma showing in detail the Relief Priorty Zones around each tornado and the Cadidate Relief Sites. Produced using ArcView 9.2 on January 28, 2008.
In order to refine the map, I combined the 2000 census data of Oklahoma with the countiesnew attribute table. I again performed a spatial intersection between the Emer_rel_zones and CountiesNew layers, creating a new layer (ReliefPriorityZone2). I repeated my steps to create a thematic map and created the image below. It is interesting to note that the 2000 census data created a more refined map, with the relief priority zones divided into smaller portions. This map could aid relief workers because the relief priority zones are better delineated due to a more accurate population density derived from the census data.
Figure 4: Close up view of Oklahoma showing the relief priority areas partly based upon the 2000 US Census. Produced using ArcView 9.2 on January 28, 2008.
By conducting this exercise, I determined that there were 33 candidate relief sites located close to the path of the tornados that struck central Oklahoma on May 3, 1999. In addition, I was able to generate a map showing the areas having the greatest need for relief work based upon the width of the tornado and the population density within the width of the tornado. This information could be invaluable to aid workers trying to determine where they were needed most after the tornado event that struck Oklahoma. A GIS was able to catalog and calculate the various forms of data about the torndaos and the population in order to produce maps that could aid in the relief work after the storms.
Unfortunately, there are some limitations of this analysis for “real-world” emergency planning and emergency relief. The maps were produced based upon the assumption that a greater population density and a greater width of the tornados equals greater damage and a greater need for relief work. Data on the ratings of each tornado on the Fujita Scale (F-Scale) intersected with the Emergency Relief Zones could more accurately demonstrate where aid was needed. The Fujita scale “is determined based on damage observed (by the tornados)” (King 2007). By adding this data, a thematic map could be produced to show areas that needed the greatest relief not just due to the width of the tornado and population density, but also due to the severity of the tornado as measured by the damage it caused.
This project was a good example of how a GIS can be used to correlate and calculate complex data to achieve certain goals. In Project 3, I was able to determine there were 33 ideal candidate relief sites within Oklahoma for people to go to after the tornados struck on May 3, 1999. These sites were located within a 1-mile buffer of the width of the tornados. In addition, by comparing the width of the tornado as well as the population density, a reasonable estimate of the areas that had the greatest need for relief work was approximated within a thematic map.
Source
King, Beth (2007) Problem Solving with GIS, Lesson 1. The Pennsylvania State University World Campus Certificate Program in GIS. Accessed 11 January 2008.
This document is published in fulfillment of an assignment by a student enrolled in an educational program of The Pennsylvania State University. The student, named above, retains all rights to the document.