Lesson 3

Project 3: Locating Tornado Relief Site

In Lesson 2 we tracked the tornadoes of May 3 across the landscape of the state of Oklahoma.  In this lesson we went into considerably more depth to determine which public facilities (schools, hospitals, and churches) near the tornado paths were likely candidate emergency relief facilities.  We first created the destructive path of each tornado according to its strength.  Then we created an outside buffer of up to 1 mile.  Within that one mile buffer surrounding the tornado path, we located the candidate relief sites.  These sites are represented below by the small diamond boxes within the tornado buffer zones.  Further, we computed the geometric intersection of the affected counties by population density with the potential emergency relief zones to arrive at a prioritized listing of areas to which relief efforts could be sent, the implicit assumption being that the areas with the highest population density would require the most immediate attention.  Figure 1 below shows the results of this analysis.  Additionally, I color coded the affected counties by their respective population densities, as this was one of the factors contributing to the calculation of the Relief Priority values.

Figure 1:  Prioritized Relief Zones Showing Candidate Relief Sites

Figure 2:  Candidate Relief Sites Attribute Table

Limitations of This Exercise for Real World Analyses

While this was an interesting exercise in buffering and intersecting various layers of data, the utility of the analysis as is stands is somewhat limiting.  First, the Candidate Relief Site data contained only coordinate information, longitude and latitude.  Geocoded street information would provide much better information for locating these sites as well as for tracking the tornado paths with reference to familiar landmarks or street names.  Second, the implicit assumption is that the tornado relief would be prioritized by population density.  Given that this information might be valid for a large geographic area, i.e. the county level, it may not necessarily apply to the specific tornado path in one section of a county.  Finally, by establishing a buffer of only one mile around the tornado path, we have arbitrarily limited alternative candidate relief sites.  For example, if we set up a series of buffers, say at one mile intervals for up to five miles away from the tornado path, be could provide more, if not better,  information on alternative medical facilities.  In the list of candidate relief sites, only one hospital was identified, and that was well away from the areas where the tornadoes inflicted the most potential damage.  And finally, having information on fire and police stations might have increased the number of candidate relief sites as these facilities would include better first aid and trained first responders.

Try This

In an effort to obtain better and more detailed population data, the analysis to generate candidate relief sites was supplemented with the Census Tract data provided in Lesson 2.  A wider view of this resulting analysis is shown below in Figure 3 and a closer view is depicted in Figure 4.

Figure 3:  Tornado Paths and Relief Priority Zones by Affected Counties.

Figure 4:  Close Up of Relief Priority Zones and Candidate Relief Sites

As shown in Figure 4 above, the Candidate Relief Sites have been coded with appropriate symbols to designate churches, hospitals, and schools.  This more detailed view demonstrates the different Relief Priority Zones based upon the Fujita scale of the tornado and the population density of the census tract.