By Susan T.

Williams
Fall 2011

GEOG 483

Project 2:
Manipulating and Summarizing Attribute Data
Objective
This project involved utilizing ArcMap 10 software for various functions pertaining to the central Alabama
tornadoes of April 27, 2011, beginning with defining a projection and then reprojecting the map to a
specific coordinate system. Once this was accomplished, other objectives included joining separate
tables to incorporate additional data, adding and populating fields within those tables, then applying
thematic maps to the results in order to demonstrate various types of damage and destruction caused by
the tornadoes.

Introduction
This lesson focused on the real-world scenario of some very damaging tornadoes that ripped through
central Alabama earlier this year. Although the damage statistics are not real, they were estimated based
on compiled data from the American Red Cross, ESRI, and the National Severe Storm Laboratory (King,
2011). Using this data, students were asked to create thematic maps that demonstrate tornado paths by
total damage, property damaged, property destroyed (in both numbers and dollars), and housing unit
density (ibid).

Defining and Projecting Data

The shapefiles were not initially defined or projected, which would lead to a misalignment of data sets and
inaccuracy of any geographic calculations such as area or distance. To remedy this, I first defined the
projection as the NAD 1983 Geographic Coordinate System using the Projections and Transformations
tool within the Data Management Tools of ArcToolbox. After verifying within the metadata that the
projection had been successfully defined and then applied to all the shapefiles, the data was reprojected
as UTM Zone 16N.

Procedure and Thematic Maps

After applying the newly defined and projected data shapefiles to the map, we can see which Alabama
counties were affected by tornadoes. I added the layer for the tornado paths and their accompanying EFScale number. EF-Scale, or Enhanced Fujita scale, is a numeric rating for a tornado based on
observed damage, with EF-0 being weakest and EF-5 being strongest (ibid). When a thematic map is
created from this information, we can visualize not only the paths the tornadoes took, but also the
strength according to EF-Scale rating [see Figure 1]. A thematic map is a type of map that emphasizes a
particular feature or topic (DiBiase, 2011), such as the specific intensities of the tornadoes in Figure 1. As
you can see by the image, these tornadoes were incredibly strong and made their way through multiple
counties. I can only imagine what a terrifying experience this must have been to those who were there.

Figure 1: Thematic map showing the paths and intensities of the Alabama tornadoes on April 27, 2011. The
Enhanced Fujita scale, or EF Scale, which is used to classify a tornados intensity by observed damage, is
represented here by a color ramp where the lightest color represents the weakest tornado and the darkest
color represents the strongest tornado. Map produced with ESRI ArcMap v10.0. Used here for educational
purposes only.

I next grayed out counties that were untouched by this storm system to make it easier to focus on only
counties specifically affected. I could have removed them from the map altogether, but it can be a little
disorienting to view a cluster of disembodied counties with no border reference if one is not very familiar
with the state of Alabama.
Figure 2 is a choropleth map that displays Total Property Damaged. These statistics were acquired by
adding a field in the attributes table to total up all houses, apartments, and mobile homes that were
damaged in each county. The map is classified using natural breaks in the data rather than quantiles or
equal intervals, and results are displayed in a color ramp in which the lightest color represents the least
number of damaged housing units and the darkest color represents the greatest number of damaged
housing units. The paths and Enhanced Fujita Scale classifications of the tornadoes have also been
displayed. The county that was hardest hit is shown in the darkest green - Tuscaloosa County.

Figure 2: Thematic map showing total property damage of all housing units per affected county. One
county Tuscaloosa - is shown in the darkest shade of green, indicating the greatest amount of property
damaged. The Enhanced Fujita (EF) scale, used to classify a tornados intensity by observed damage, is
also shown. Map produced with ESRI ArcMap v10.0. Used here for educational purposes only.

Next I developed a thematic map to show total property destroyed. These statistics were acquired by
adding a field in the attributes table to total up all houses, apartments, and mobile homes that were
destroyed (rather than damaged) in each county. The map is classified using natural breaks in the data
rather than quantiles or equal intervals, and the results are displayed in a color ramp in which the lightest
color represents the least number of destroyed housing units and the darkest color represents the
greatest number of destroyed housing units. The paths and Enhanced Fujita Scale classifications of the
tornadoes have also been displayed. The county that was hardest hit is shown in the darkest green and
we can see that it is, again, the aforementioned Tuscaloosa County, with 2375 total housing units
destroyed (an amount almost equal to the housing units that were damaged in the same county).

Figure 3: Thematic map showing the total property destruction of all housing units per affected county. Again,
Tuscaloosa County is shown in the darkest shade of green, indicating the greatest amount of property
destroyed. The Enhanced Fujita (EF) scale, used to classify a tornados intensity by observed damage, is also
shown here. Map produced with ESRI ArcMap v10.0. Used here for educational purposes only.

The next thematic map (see Figure 4) shows total property destroyed in dollars rather than in numbers
of housing units as the previous map did. I changed the color ramp to a blue scheme so that any major
differences from the previous map would be sure to stand out. These statistics were acquired by
adding a field in the attributes table to multiply the total number of housing units destroyed by the
median home value. Again, the map is classified using natural breaks in the data rather than quantiles
or equal intervals, and the results are displayed in a color ramp in which the lightest color represents the
smallest dollar amount of destroyed property and the darkest color represents the greatest dollar
amount of destroyed property. The paths and Enhanced Fujita Scale classifications of the tornadoes
have also been displayed on the map.

Figure 4: Thematic map showing the total property destruction in dollars per affected county. Again, it is
Tuscaloosa County that is shown in the darkest shade of blue, indicating the greatest dollar amount of property
destroyed.

The Enhanced Fujita (EF) scale, used to classify a tornados intensity by observed damage, is also

shown here. Map produced with ESRI ArcMap v10.0. Used here for educational purposes only.

The county that had the greatest number of housing units damaged and destroyed Tuscaloosa County
- also had the greatest dollar amount of destroyed property, which is somewhat expected. Obviously
this might not always be the case, such as if a tornado had hit an area where the median value of homes
was higher than the average of surrounding areas. For example, Elmore County, which shows as the
second shade of blue, had previously displayed as the lightest green on the previous map representing
property destroyed by number of housing units. This average median home value in Elmore County is
$98,000, compared to a lower median value of $73,600 and $58,900 in Tallapoosa and Chambers
Counties respectively, both of which are just to the east of Elmore County and were also in the tornados
path.
Figure 5 is a thematic map showing housing density in the affected counties. These statistics were
acquired by
adding a field
in the table to
divide the
number of
housing units
per square
mile. I
selected a
purple
classification
color scheme
this time to
emphasize the
representation
of housing
density rather
than damage
or destruction.

Figure 5: Housing Density (housing units per square mile) per affected county. One county Jefferson
is shown in the darkest shade of purple, indicating the greatest housing density. The Enhanced Fujita
(EF) scale, used to classify a tornados intensity by observed damage, is also shown here. Map produced
with ESRI ArcMap v10.0. Used here for educational purposes only.

There are some interesting observations here. The most densely populated county is Jefferson, which
one would assume would lead to a higher number of homes being damaged or destroyed. Yet Jefferson
County did not rank highest in terms of Property Damaged, Property Destroyed in number of housing
units, or Property Destroyed in dollars. Although neighboring Tuscaloosa County has a higher median
value per home ($106,600 compared to $90,700 in Jefferson County), this would not account for
Tuscaloosa County experiencing more property damaged and destroyed per number of housing units.
Marshall County is another interesting example. Located northeast of Jefferson County, this unfortunate
county was hit by fifteen tornadoes, yet did not suffer a great deal of damage. I believe there are three
factors that account for this. First, when looking closely at the housing density map, one can see that
Marshall is not one of the more densely-populated counties. Secondly, the majority of the tornadoes
had shorter paths here, meaning that they did not travel as great a distance as many of the tornadoes in
other counties and may have been shorter in duration. Finally, looking at the color scheme classification
of the EF Scale reveals that most of the tornadoes in Marshall County were of lesser intensity than those
experienced in other counties. Specifically, out of the fifteen tornadoes, one was rated EF0 and eleven
were rated EF1. While Im sure that information would be of little comfort to those who experienced
the tornadoes, it is clear that the situation had the potential to be much worse.

Further Refinement
After downloading and importing the census tracts and demographic data shapefiles, which are based
on 2000 census data, I joined these two tables using STFID as the key common to both. I added a field
for housing density and used the field calculator to populate this column by dividing the number of
housing units per census tract by the areas square miles. The new column then displayed the housing
density for each individual census tract. Using this information, I created another thematic map
representing housing density (housing units per square mile) using census tracts rather than the county
level (see Figure 6).
A census tract is a small subdivision of a county delineated in a way that creates a relatively
homogenous statistical sample of the population contained within its boundaries, which averages about
4,000 inhabitants (U.S. Bureau of the Census, 2011). Using census tract data generally results in more
detailed, and therefore more accurate, information.

Figure 6: Housing Density (housing units per square mile) based on census tracts within the tornadoaffected counties.

The Enhanced Fujita (EF) scale, used to classify a tornados intensity by observed

damage, is also shown here. Although it is difficult to ascertain much at this scale, Jefferson Countys
higher density of population can be seen fairly clearly. Map produced with ESRI ArcMap v10.0. Used
here for educational purposes only.

Its fairly difficult to determine much on a map this scale, so I created a larger-scale thematic map which
focused on the aforementioned Tuscaloosa and Jefferson counties, which suffered the greatest
destruction and have the highest density respectively (see Figure 7).

Figure 7: A larger-scale map showing Housing Density (housing units per square mile) based on census
tracts within a few of the tornado-affected counties. The individual census tracts are more easily
distinguishable in this view. The Enhanced Fujita (EF) scale, used to classify a tornados intensity by
observed damage, is also shown here. The path of an EF4 tornado can be seen cutting right through the
most densely populated areas of Tuscaloosa County and skirting around the more densely populated
sections of Jefferson County to the northeast. Map produced with ESRI ArcMap v10.0. Used here for
educational purposes only.

When viewed at this scale, the individual census tracts are more distinguishable to the eye and there are
some interesting observations to be made. The same EF-4 tornado ripped through both counties
(identified as Tornado ID 47 in the attribute data). Tuscaloosa County, which had less population
density than Jefferson County and yet suffered greater tornado damage in both property and dollar
amount, was unfortunate enough to suffer Tornado 47 moving right through the center of the most
9

densely-populated areas of the county. In Jefferson County, however, Tornado 47 moved through some
of the countys least-populated areas. This would certainly account for the discrepancy in the higherdensity Jefferson county suffering from less damage than the less-dense Tuscaloosa county.

Summary
One would assume that a strong tornado passing through a county that has the highest housing density
will cause the most damage and destruction. However, we have seen that may not always be the case
depending on where the highest density of population is specifically located within those counties and
which path the tornado takes. Obtaining a greater level of data such as census tracts is very useful for
this type of analysis.
Thematic maps are very helpful in classifying and displaying geographic data, but one should be careful
to not rely too heavily on any one map or the analysis could be misleading. When a variety of thematic
maps representing various attributes are considered as a whole, users are far more likely to arrive at
accurate conclusions.

..

References
DiBiase, Davide. (2011) The Nature of Geographic Information, Chapter 3, Section 13. The Pennsylvania
State University World Campus Certificate/MGIS Programs in GIS.Retrieved October 14, 2011 from
https://www.e-education.psu.edu/natureofgeoinfo/c3_p13.html
King, E., & Walrath, D (1999-2011). Problem-Solving with GIS, Lesson x, /Part x, Section x/. The
Pennsylvania State University World Campus Certificate/MGIS Programs in GIS. Retrieved October 2,
2011 from https://www.e-education.psu.edu/geog483/node/1879
Report header [image of tornado] courtesy of Arthurs Nature Clipart. Used for educational purposes
according to the Terms of Use. Retrieved October 14, 2011 from

http://www.arthursclipart.org/nature/nature/page_04.htm
U.S. Bureau of the Census, Glossary. Retrieved October 18, 2011 from
http://factfinder.census.gov/home/en/epss/glossary_c.html
All maps produced with ArcMap v10.0 by Environmental Systems Research Institute (ESRI). (2011)
ArcGIS 10 Help. Esri, Redlands, California.

This document is published in fulfillment of an assignment by a student enrolled in an educational offering of the
Pennsylvania State University. The student, named above, retains all rights to the document and responsibility for
its accuracy and originality.

10