Geog483 Project #3

Locating Tornado Relief Sites

John Middendorf

Introduction

The project uses ArcGIS to do some fundamental spatial analysis to locate potential tornado relief sites in Oklahoma.  We are given three shapefiles: Oklahoma counties, Oklahoma places, and Tornado paths.  We are also given two database files: Tornado attributes, and population.  All data uses the NAD83 datum and the Oklahoma North State Plane coordinate system (Lambert Conformal Conic projection).

Though a series of joins, adding and calculating new fields, creating buffers (using both the Buffer Wizard and Arc Toolbox), and union and intersection operations, we are able to create a thematic map showing potential relief sites (from Oklahoma places) that are within 1 mile within the centerline of the tornado path  and not in the buffered tornado path.  We also create a graduated color Priority Zones based on population density within each county.  Priority Zones are also calculated based on census tracts population densities.

Maps

Above: a screenshot at the end of the analysis showing the union and intersected buffered priority zones and potential relief sites within the zones.

Above: Attribute data of potential relief sites (schools, churches, and hospitals) that are within one mile of the centerline of the tornado but not within the path of the tornado based on tornado widths.

Because the simple screenshots are not up to par for presentation purposes, the following maps refine the ArcMap output:

Above:  Map 1 showing the Relief Priority Zones and the candidate sites labeled.

Map 2: detail for the above map showing the highest priority relief zone with candidate sits marked.

Above: Results of “try this” exercise, showing census tract priorities.

Finally, the map below offers a single 8.5 x 11 inch page presentation with multiple views using the ArcMap layout view for a consise presentation showing all relevant information:

 Analysis Limitations for “real-world” planning.

The above analysis makes some assumptions based on the data provided.  The population database gives total population per county, thus the priority zones are based on population density per county.  Because of the large area each county covers, the tornados may have passed though a low density area within a high population density county, yet the priority of that region would still be high.  In this case, there may have been more damage and casualties in areas marked as lower priority using the above assumptions.  For a real world scenario, using population densities from census tract data would give much more relevant results. 

Also lacking is consideration of access to relief sites.  The chosen relief sites were chosen purely on spatial calculations, and not based on road information.   The selected sites may have been cut off by tornado road damage, or there could have been easier road access to candidate sites that were spatially more distant.