You are on page 1of 14

Wesley Collins

Identifying Priority Wildfire Mitigation Areas in Hays County Using GIS


Introduction
Land use in the United States has changed dramatically over the past century leading to a
multitude of new and unforeseen challenges (Stohlgren et al. 1998). One area where these
changes are evident is the Texas Hill Country, or Balcones Canyonlands Geographically the area
is a thin line running along the eastern boundary of the Edwards Plateau that stretches from just
north of Austin to San Antonio and is characterized by karst geology and topographic relief
(Kastning Jr 1983). Historically, the region was a ranchland best described as savanna or
grassland with interspersed mottes of trees on the rolling hills and juniper thickets in some of the
steep canyons (Smeins 1980). Now, one of the fastest growing populations in the nation resides
in the region creating urban sprawl across the landscape (Kreuter et al. 2001). A consequence of
urban sprawl into formerly rural areas that humans are finding to be extremely difficult to cope
with is vegetation management; specifically, how to deal with the proliferation of woody
vegetation (Archer 1989).
When livestock were on the land on large ranches, they helped maintain the grasslands in
the central Texas region by eating woody plants when they were young and tender. Before
livestock carried out this task, the bison maintained the grassland along with periodic fire (Taylor
et al. 2012). Since there are now human populations in these regions, none of these processes are
able to take place, and some native woody trees such as Ashe juniper (Juniperus ashei) and
Honey mesquite (Prosopis glandulosa) proliferate to a point where they choke out other native
vegetation and become invasive.

Wesley Collins
One additional problem that has plagued the region recently is not a new foe. Drought has
been occurring often and with a vengeance since Europeans first stepped foot in Texas. Dealing
with drought is nothing new here, however; drought compounded with woody proliferation and
urban sprawl has led to some devastating new consequences in Texas such as wildfires in or near
urban areas. In communities where this type of disaster has not yet occurred, it may be time to
start preparing for the possibility that a wildfire may occur.
The target area for this study was Hays County, Texas. This area was chosen as the study
area for several reasons: A) Booming Population B) Current and Continued Urban Sprawl and C)
Major Land use Change over Past 50 Years from Farms/Ranch to subdivisions. The purpose of
this study was to delineate target areas across the entire county that should be the focus of brush
thinning and clearing efforts in order to diminish wildfire danger. While the potential for wildfire
cannot be eliminated completely, evidence has shown that the best way to reduce risk is by
diminishing available fuels (Graham et al. 1999). Target areas were delineated by overlaying the
following spatial data of Hays County in ArcGIS to reach an output of target areas: slope,
vegetation type, parcel size, distance from roads, and distance from waterways.
The spatial analyst extension was used extensively in this project and many of the
functions found within this tool box in ArcGIS were invaluable. Several of the functions used
from the spatial analyst extension included the reclassify tool; a tool used to reclassify attribute
data, and the slope function; a tool used to derive slope as a raster data set from a surface
elevation raster. Additional tools and functions used in the project included clip, erase, union,
buffer, intersect, create layer from selection, from raster to polygon. All functions used will be
discussed in more detail in the methods section.
Data

Wesley Collins
Data sources were primarily from federal or state government agency. Data used were as
follows: vector and attribute data of ecosystems in Texas from Texas Parks and Wildlife
(TPWD), hillshade raster of Texas from the Texas Water Development Board (TWDB), Texas
roadways vector data from TNRIS, rivers and streams vector data for Texas from TNRIS, central
Texas county outlines from Capital Area Planning Council of Governments (CAPCOG), and
parcel vector data from CAPCOG for Hays County.
The ecosystem data used in this project to identify the vegetation types to be targeted in
this effort was provided by TPWD. TPWD is the state governmental agency responsible for the
protection and stewardship of public land and wildlife resources, as well as extension services to
private landowners. The data used in this project contained extremely detailed information about
the vegetation types found in the area of interest.
The hillshade raster was downloaded from the TWDB, and agency responsible for the
monitoring, planning, and administration of Texas groundwater resources. TWDB maintains a
GIS inventory of several datasets and maps important to water in Texas. Hillshade data is among
these because it can be used in a number of functions in ArcGIS to derive elevation and slope, as
well as information on drainages. This particular hillshade data was created using digital
elevation models (DEM) which use point data on elevation to create a surface model of
elevation. In the case of this project, hillshade data was necessary to derive slope for the
planning of what areas should be targeted. The primary reason that slope is such a necessary
component of this project is because slope historically was the dividing line in distinct vegetation
zones and thus steeper areas need to be excluded from areas to target.

Wesley Collins
Data gathered from TNRIS included vector data on roads and waterways in Hays County.
Both of these were necessary for this project because they needed to be excluded from any
targeted brush removal.
The last two datasets came from CAPCOG, a regional planning council of local
governments in the region that also maintains an inventory of useful maps and GIS data for the
area. Hays County is included in the counties under CAPCOGs umbrella, and so they carried
several useful datasets for this project. The first set was simply a vector file of the county
outlines in Central Texas. The other data set gathered from CAPCOG was absolutely critical to
this project. The dataset used was a vector polygon file containing attribute data on every parcel
in Hays County. The data was meant primarily to be used for tax information, however I was
more interested in just the size of properties. This data set contained just that information,
specifically, attribute data of parcel acreage.
Methods
In order to aid in organization during both the analysis and reporting phase of this project,
models and flow charts were used within ArcGIS. Throughout the rest of this section, models or
flow charts may be referred to in addition to some of the resulting maps created during the
process. As stated in the data section, several steps were taken in order to organize data prior to
analysis. These steps included selecting and clipping the Hays County outline from the
CAPCOG Counties shapefile. Next, the Hays County extent was used to clip the hillshade,
waterways, and roads layers. These processes can be found in a flowchart in figure 1 of the
appendix. Once these preliminary tasks were complete, the true analysis began.

Wesley Collins
The first part of my analysis began with buffering all of the roads and streams/rivers
throughout the county. This was done in order to make sure that the final target areas did not
include any roads or streams and the areas around them. The reasoning for buffering roads was
to avoid having any possible confrontations with infrastructure such as power lines or other
anthropogenic structures around roads, and also to avoid affecting traffic when some of the
equipment needed for the project would be traveling from site to site. The rivers and streams in
the county were buffered so that riparian areas would be protected. These areas are very
important wildlife habitat and also are environmentally sensitive so it would be wise to steer
clear of disturbing these areas during this project (Hubbard 1977). Once these two buffer
functions were performed, the two resultant layers were united into one layer. The output was
polygons of all of the buffered streams and roads (Fig. 2)
In order to determine where in the county the invasive systems we wanted to target
occurred, we needed to cross reference the invasive ecosystems with parcels of land that were
adequately suited to be part of the project. The parcels in the vector file from CAPCOG were
selected based on their size attribute. We decided that for this project, properties should be at
least 200 acres of size so as to maximize the amount of area that could be treated while
minimizing the number of properties that would need to enroll in the project. Additionally, over
two-thirds of the hectares identified as potentially invasive ecosystems occurred on tracts of 200
acres or larger (20,051 hectares out of 27,718 hectares of invasive ecosystems, map 1 in
appendix).
Potentially invasive ecosystems across the county were selected based on attribute
information for vegetation type. If the vegetation common name included the name juniper,
mesquite, or invasive it was selected and a new layer was created from these attributes. These

Wesley Collins
were labeled as potentially invasive ecosystems. While there is a high probability that these areas
need to be treated, just because they were selected does not mean they are ideal for treatment.
Further individual on-site analysis will need to be done in all of these areas. Once both of these
new layers were created, an intersection was performed so that we could identify potentially
invasive ecosystems only on selected properties (Fig. 3).
Next, a model was created to determine the slope using a hillshade raster. The slope
function was used and the hillshade raster was used as the input. The slope function is a part of
the spatial analyst extension in ArcGIS. The spatial analyst extension contains many useful tools
which are unique to ArcGIS that are used to analyze raster data. The slope function is a function
which interpolates the slope based on the input DEM or hillshade raster. The function was run
with a raster layer showing the ranges of the slope in degrees across the county being the output
(see inset Map 2 in appendix).
The slope output was reclassified from a range of 0 to 90 degrees to a range of 1 to 10 by
1 using the reclassify function. The steepest slopes became a 10 and the lowest slopes became a
1. This output raster of reclassified slope was then converted into a polygon layer so that the
areas with the steepest slopes could be selected (Fig. 4). Slopes above 25.2 degrees were
classified as a 7, 8, 9, or 10 on the cost surface, and then selected and removed from the
converted polygon layer so that they would be excluded from the final target areas.
As previously mentioned, it was necessary to exclude the buffered roads and waterways
areas from the final target areas. In order to perform this, the erase tool was used to remove the
buffered polygons from the invasive ecosystems on selected properties (Fig 5). The final step
was to use the erase function again, using it this time to remove any areas of the invasive
ecosystems minus buffered areas layer that overlapped with areas containing the steepest slopes

Wesley Collins
(Fig. 6). The output of this analysis gave a very nice model of areas that should possibly be
considered for brush removal treatments (please see Map 3 in Appendix).
Discussion
The vegetation types of concern occur county wide across a broad range of property sizes
and ownership statuses, so we decided the best method to select properties would be to set a
cutoff for a minimum property size of 200 acres or approximately 80 hectares. In doing so, it was
thought that it would expedite progress by minimizing the number of landowners whom would
need to be contacted. As previously mentioned, 72 % of the ecosystems of interest occurred on
properties of 200 acres or greater, so it is projected that the project may have more success by
approaching this limited group of land owners for the time being. In the future, further expansion
of the project to smaller properties could easily be implemented with minimal augmentation of
the data used in this analysis.
Projects prior to GIS and GPS technology involving large scale brush clearing were often
inefficient and difficult to measure. Not only can the manager or planner use GIS for planning
their strategies, but also modeling responses to different management strategies (Afinowicz,
Munster and Wilcox 2005).
The spatial analyst extension was especially useful in this project because it contained the
slope tool. This tool gave an incredible advantage in planning a project of this scale because we
could easily derive slope of county wide exclude those areas from the target areas without ever
having to send a survey crew out on the ground.
Overall, this approach seemed to work extremely well at responding to the task of
identifying areas to target with brush removal and thinning as a way of reducing wildfire risk in

Wesley Collins
Hays County. The approach allowed for a very planned and pragmatic method of processing the
data, and the final finished product was a rendering that is very simple to understand despite
education or background. This is made possible through the tools available to manipulate,
analyze, and display spatial data in ArcGIS.
Conclusions
The final results for areas to be targeted for brush removal ended up being much larger
than anticipated in the beginning of the project totaling out at 19,827 hectares being
recommended for treatment. This approach to identifying target location to apply vegetation
treatments seemed to be very effective. Most of the data that was needed was easily attained.
While this project specifically is purely hypothetical and is probably not reasonable because of
the massive amount of public money that would have to be used to fund projects on private land,
it does show how useful GIS tools can be at analyzing this type of problem. The danger of
wildfire in Texas will be an increasing concern as the population continues to grow and weather
patterns remain unpredictable.
The application of GIS tools to this type of analysis could be used in many other
situations in which perhaps mitigation for wildfire could be done on public lands in the western
states, or even on individual properties if the data used is detailed enough and the property is
large enough to merit the use of GIS. It appears that there is already wide use of GIS towards the
application of fire management and prevention, and to a greater extent emergency management
and networking (Johnson 2000, Chuvieco and Salas 1996, Ambrosia et al. 1998). As data
continues to be gathered on spatial features and the technology gets more easily accessible, the
application for this type of analysis in GIS is likely to become common across all sectors of the
public and private sector.

Wesley Collins
Appendix
Figure 1

Figure 2

Figure 3

Figure 4

Wesley Collins

Figure 5

Figure 6

Wesley Collins
Map 1

Wesley Collins

Map 2

Wesley Collins

Map 3

Wesley Collins
Works Cited
Afinowicz, J. D., C. L. Munster & B. P. Wilcox. 2005. MODELING EFFECTS OF BRUSH
MANAGEMENT ON THE RANGELAND WATER BUDGET: EDWARDS
PLATEAU, TEXAS1. Wiley Online Library.
Ambrosia, V. G., S. W. Buechel, J. A. Brass, J. R. Peterson, R. H. Davies, R. J. Kane & S. Spain
(1998) An integration of remote sensing, GIS, and information distribution for wildfire
detection and management. Photogrammetric engineering and remote sensing, 64, 977986.
Archer, S. (1989) Have Southern Texas Savannas Been Converted to Woodlands in Recent
History? The American Naturalist, 134, 545-561.
Chuvieco, E. & J. Salas (1996) Mapping the spatial distribution of forest fire danger using GIS.
International Journal of Geographical Information Science, 10, 333-345.
Graham, R. T., A. E. Harvey, T. B. Jain & J. R. Tonn (1999) Effects of thinning and similar
stand treatments on fire behavior in western forests.
Hubbard, J. P. 1977. Importance of riparian ecosystems: biotic considerations. In Importance,
Preservation and Management of Riparian Habitat: A Symposium July 9, 1977, Tuscon,
Arizona, USDA Forest Service General Technical Report RM-43, 14-18.
Johnson, R. (2000) GIS technology for disasters and emergency management. An ESRI white
paper.
Kastning Jr, E. H. (1983) Geomorphology and hydrogeology of the Edwards Plateau karst,
central Texas.
Kreuter, U. P., H. G. Harris, M. D. Matlock & R. E. Lacey (2001) Change in ecosystem service
values in the San Antonio area, Texas. Ecological Economics, 39, 333-346.
Smeins, F. E. (1980) Natural role of fire on the Edwards Plateau. Junction, Texas, 4-16.
Stohlgren, T. J., T. N. Chase, R. A. Pielke, T. G. Kittel & J. Baron (1998) Evidence that local
land use practices influence regional climate, vegetation, and stream flow patterns in
adjacent natural areas. Global Change Biology, 4, 495-504.
Taylor, C. A., D. Twidwell, N. E. Garza, C. Rosser, J. K. Hoffman & T. D. Brooks (2012) LongTerm Effects of Fire, Livestock Herbivory Removal, and Weather Variability in Texas
Semiarid Savanna. Rangeland Ecology & Management, 65, 21-30.

You might also like