Using Remote Sensing and GIS Techniques to Identify Watersheds for Brush Control to Maximize Water Y

1
Using Remote Sensing and GIS Techniques to
Identify Watersheds for Brush Control to Maximize
Water Yield
Jason D. Afinowicz, E.I.T.1 Clyde L. Munster,
Ph.D., P.E.1 Brad P. Wilcox, Ph.D.2 Ronald E.
Lacey, Ph.D., P.E.1 1Department of Biological and
Agricultural Engineering, Texas AM
University 2Department of Rangeland Ecology and
Management, Texas AM University

SELECTION BT TOPOGRAPHY Wilcox (2002) suggests
that recharge occurring through stream channels
is enhanced by the ability of surface water to
run off before being taken up by local
vegetation. By decreasing interception by woody
species on hill slopes (approximately 5 or
greater) it is surmised that more water will be
allowed to enter the stream channel and
contribute to overall water yield. Selection of
steep slope regions is performed with the use of
the National Elevation Dataset for the Upper
Guadalupe, available at a spatial resolution of
30 meters.



THE UPPER GUADALUPE RIVER
ABSTRACT Shrub encroachment has drastically
transformed the landscape of arid and semi-arid
rangelands of the southwestern U.S. over the past
century. It has been hypothesized that this
replacement of natural herbaceous growth with
woody species has reduced the volume of
streamflow and aquifer recharge in these
landscapes. Several small scale field
experiments and computer-based hydrologic
simulations have been conducted to support this
theory. In recent years, the State of Texas has
begun to subsidize brush control as a means of
augmenting streamflow. However, if brush control
is to be regarded as a viable option for
increasing water availability in arid or
semi-arid populated areas, a technique must be
determined for targeting locations where brush
management funding will provide the greatest
addition to water yield. The working hypothesis
is that water yield from brush control will be
maximized in watersheds where 1. Precipitation
is greater that 450 mm, 2. Shrub coverage is
greater than 20, 3. Soil depth measures less
than 1 m, and 4. Limestone aquifers underlie the
soil profile. Locations that meet the listed
criteria for candidate brush control sites can be
easily identified by querying a number of data
sources using Geographic Information Systems
(GIS). A classification scheme was developed for
remotely sensed multi-spectral imagery of arid
and semi-arid areas that allows for the
recognition of regions meeting the required
criteria for brush cover. Selection of optimum
sites based on soil and climate characteristics
was conducted using high resolution databases.
The methods utilized throughout the entire
process were designed to be easily adapted to a
variety of locations.
Elevation dataset showing the regions meeting the
5 slope criteria (Shown in Red).

The Upper Guadalupe River watershed encompasses
over 374,000 ha of the Texas Hill Country, north
of San Antonio. The watershed boundaries are
contained within Bandera, Blanco, Comal,
Gillespie, Kendall, Kerr, and Real counties. The
watershed
SELECTION BY CLIMATE CHARACTERISTICS Hibbert
(1983) established a guideline for the necessary
precipitation required for a region to
demonstrate an increase in stream flow. Though
Hibberts threshold value of 450 mm/year was
selected from studies in Arizona and California,
this values has been applied to Texas rangelands.
This amount of average annual rainfall is
considered a requirement for vegetation change to
influence water yield. By using data from Oregon
States Parameter-elevation Regressions on
Independent Slopes Model (PRISM) datasets for
annual rainfall, it was verified that the Upper
Guadalupe met this criteria.

lies above the Edwards Aquifer recharge zone and
ultimately contributes to the groundwater supply
for the entire region, including San Antonio.
Springs
Regions meeting the 450 mm rainfall requirement
(Entire Upper Guadalupe).
fed by the aquifer host a variety of unique
species and contribute to the ecology of this
unique ecosystem.

• SELECTION CRITERIA
• For purposes of site selection for brush control,
  we will propose that the optimum sites as those
  meet the following criteria
• Slopes greater than 5
• Average annual rainfall greater than 450 mm
• Soil depth of less than 1 m
• Brush cover greater than 20
• Underlying limestone aquifers

SELECTION BY SOIL CHARACTERISTICS Wilcox (2002)
suggests that thick soil layers may prevent water
from infiltrating the profile and contributing to
aquifer recharge. The karst geology of the Upper
Guadalupe River watershed is optimum for taking
advantage of the numerous shallow soil regions in
the area. A threshold soil thickness of 1 m was
selected to locate regions where infiltration
past the root zone of overlying vegetation would
be allowed. Analysis was conducted using NRCS
SSURGO databases where possible, which represent
the most accurate and detailed digital depiction
of soil characteristics available. Where SSURGO
coverage was not complete, STATSGO data was used.

Map derived from electronic soil datasets showing
regions with soil depth less than 1 m.

BACKGROUND AND MOTIVATION Encroachment of woody
species, most typically mesquite (Prosopis
glandulosa) and juniper (Juniperus ashei and
Juniperus pinchotii), has dramatically changed
the landscape of semi-arid regions due to a host
of human and environmental factors (Van Auken,
2000) and is believed to have contributed to the
decrease of useful water yield (Wu et al., 2001).
Small scale experiments indicate that the
removal of juniper species, for instance, may
reduce transpiration by as much as 40,000 to
100,000 gallons per acre per year (Owens, 1996).
This shortage of water adds additional stress to
semiarid systems which operate in a
soil-water-deficient manner, meaning that the
potential ET rate far exceeds the annual
precipitation rate (Wilcox, 2002). Given the
rate at which the population of these semi-arid
regions is expanding, it is of the utmost
importance that efforts are made to minimize
wasteful losses of water (TWDB, 2002). It is by
this means that brush control has entered the
public focus as a possible way to mitigate the
problem of reduced water supplies in relatively
dry regions. Currently, there is no method for
determining regions where the greatest increase
in water yield will result from the
implementation of brush management practices.
Studies that have been conducted tend to focus on
a watershed based approach to determine the
overall effects of brush control. Research such
as this has focused on hydrologic simulation of
entire basins to add merit to the cause of brush
management, rather than the finer points of
implementing these practices (Bednarz at al.,
2001 Red River Authority, 2000). This is
especially important since Texas has begun the
subsidization of these techniques as a way of
compensating land owners who make the effort to
increase the overall availability of water for a
region (TWDB, 2002). If this process of state
funded brush control is expected to be
economically viable, there must be a way of
ensuring that the public funds expended in the
name of environmental, agricultural, and
anthropocentric well-being are being spent to
rehabilitate locations where the most benefit may
be gained.

SELECTION BY LANDCOVER The amount of shrub cover
is also indicated to be a major contributor to
the success of brush control by Wilcox (2002).
Though extensive landcover datasets exist, the
information presented fails to give any
indication of the amount of woody cover present.
For this reason, a new landcover dataset for the
Upper Guadalupe was created with the use of
Landsat Enhanced Thematic Mapper Plus (ETM)
imagery, high resolution (1 meter) digital
orthophoto imagery, and land cover information
from the 1992 Multi-resolution Landcover
Characteristics (MRLC) National Land Cover Data
(NLCD) set. Since the future aims of the project
look toward the creation of an easily computed
index of prime brush control locations an
emphasis was placed on the creation of a
landcover set with the minimum need for
ground-truth data for its creation. Minimizing
this time and money consuming step benefits the
systems use by water planners and engineers.
Landsat ETM Image of the Upper Guadalupe River
watershed, taken from Landsat Row 39, Paths 27
and 28.
The MRLC landcover grid was used as a basis to
train the Maximum Likelihood classifier for the
task creating a general landcover
Following classification of the satellite and
orthophoto imagery, a final classification was
performed using a visually determined threshold
for the
Dataset representing the regions meeting the
brush cover greater than 50 criteria.
set. The new product represented a new dataset,
contemporary to the 1999 satellite imagery that
was used. This provided a coarse method for
recognizing general brush covered regions and to
provide additional land cover information to be
used in later hydrologic analysis.
resampled DOQ classification. This allowed
regions recognized as brush during the satellite
image classification to be defined as either
light (0-20), moderate (20-50), or heavy
(50-100) brush cover.
MRLC Land Cover Grid for the Upper Guadalupe
River watershed. Provided in 30 meter resolution.
Dataset representing the regions meeting the
brush cover greater than 20 criteria.
Representative 1-m resolution DOQ showing detail
of brush cover. Mosaics were formed for the
entire watershed.
The same DOQ classified into regions containing
woody growth and no brush using a Maximum
Likelihood classifier.
Upon classification, the image was resampled to
30m and assigned a digital number based on amount
of cover.

• SECONDARY BRUSH CONTROL SITES
• Chosen using the following criteria
• Average annual rainfall greater than 450 mm
• Soil Depth of less than 1 m
• Brush cover greater than 20
• The selected regions covered an area of 142,282
  ha, or approximately 38.0 of the entire
  watershed.

• PRIORITY BRUSH CONTROL SITES
• Chosen using the following criteria
• Slope greater than 5
• Average annual rainfall greater than 450 mm
• Soil Depth of less than 1 m
• Brush cover greater than 50
• The selected regions covered an area of 36,330
  ha, or approximately 9.7 of the entire watershed.

Future Research Beyond the selection of potential
brush control target sites, several research
activities have been planned to continue the
advancement of GIS techniques in brush control
for water resources planning
REFERENCES Bednarz, S.T., T. Dybala, R.S.
Muttiah, W. Rosenthal, W.A. Dugas. 2000. Brush
management/water yield feasibility studies for
eight watersheds in Texas. Texas Water Resrources
Institute, CollegeStation, TX. Hibbert, A.R.
1983. Water yield improvement potential by
vegetation management on western rangelands.
Water Resources Bulletin. 19375-381. Owens,
M.K. 1996. The roll of leaf and canopy-level gas
exchange in the replacement of Quercus virginiana
(Fagaceae) by Juniperus ashei (Cupressaceae) in
semiarid savannas. American Journal of Botany.
83617-623. Red River Authority. 2000.
Assessment of Brush Management/Water Yield
Feasibility for the Wichita River Watershed Above
Lake Kemp, Hydrologic Evaluation and Feasibility
Study. Red River Authority of Texas. Wichita
Falls, TX. TWDB. 2002. Water for Texas 2002.
Texas Water Development Board, Austin, TX. Van
Auken, O.W. 2000. Shrub invasions of North
American semiarid grasslands. Annual Review of
Ecology and Systematics. 31197-215. Wilcox,
B.P. 2002. Shrub control and streamflow on
rangelands A process based viewpoint. Journal
of Range Management. 55318-326.
Validation of Landcover Set The landcover set
derived in this research will be verified by GPS
survey to validate its use in further research
and to evaluate the proposed methodology for use
by professional planners.
Calibration and Validation of the SWAT Model Data
from an in progress brush study at Honey Creek in
the Guadalupe watershed will be used to calibrate
and validate the Soil and Water Assessment Tool
(SWAT) for brush control analysis.
Simulation of Target Sites The SWAT model will be
used with the described calibration to verify the
validity of the proposed selection methodology by
electronically simulating brush in place and
brush removed conditions for the Upper Guadalupe.
http//ghg.tamu.edu