- Fort Hood Avian Simulation Model (FHASM) -
Two federally endangered bird species, the black-capped vireo (Vireo articapillus) and the golden-cheeked warbler (Dendroica chrysoparia), breed at Fort Hood, Texas. Presently, efforts are being made to ensure the long-term viability of these species. To support management efforts, FHASM has been developed by integrating state-of-the-art computer technology, ecological theory, field-based data, and opinions of regional biologists.
The purpose of this spatially-explicit simulation model is to provide a tool to aid land managers, biologists, and military personnel make policy decisions for Fort Hood, Texas. It allows the user to request various land use and management practices for the post and then simulates how these decisions affect the population dynamics of two endangered avian species.
- Introduction to Fort Hood,Texas -
Fort Hood is important to the Army's mission of maintaining a trained and ready force. The post, established in 1942, encompasses an area of approximately 88,000 ha (340 sq. miles). It is located in Bell and Coryll Counties in central Texas, between Waco and Austin. Long, hot summers and short, moderate winters are typical climate; average annual precipitation is 81 cm. The landscape is alternating flat valleys and limestone escarpments, which rise up to an elevation of 379 m.

- Introduction to the Endangered Species -
The black-capped vireo (Vireo articapillus ) is a small, greenish
songbird that breeds in Oklahoma and Texas and overwinters along the western
coast of Mexico. This species became listed as federally endangered in 1987.
Historically, the breeding range of black-capped vireos extended as far north as
Kansas, but has become restricted to the present range following loss of
appropriate breeding habitat. Black-capped vireos are habitat specialists that
require hardwood scrub habitat that consists of patchy shrubs and thickets
interspersed with live and dead trees. Breeding territories of black-capped
vireos pairs range from 1.5 to 3 ha. Individuals return to the post fom their
overwintering range in mid to late March, and leave the post for overwintering
ranges in August and September.
There appear to be two major threats to
the survival of the black-capped vireo, loss of habitat and nest parasitism by
cowbirds. Loss of existing breeding habitat, as well as potential for future
habitat, occurs following conversion to pasture for cattle and urbanization.
Cowbirds are obligate brood parasites that lay their eggs in the nests of other
species, including black-capped vireos. Black-capped vireos cannot successfully
fledge both their own young and the young of cowbirds; when parasitised, they
fledge only cowbirds or abandon the nest.
The golden-cheeked warbler
(Dendroica chrysoparia ) is small, brightly-colored songbird that breeds
only in Texas and overwinters in mountainous areas of Central America. This
species became listed as federally endangered in 1990. Golden-cheeked warblers
are habitat specialists that require mixed hardwood stands that consists of ashe
juniper and a variety of oak species. Breeding territories of golden-cheeked
warbler pairs range from 2 to 20+ ha. Individuals return to the post from their
overwintering range in March, and leave the post for overwintering range in June
and July.
The major threat to the survival of golden-cheeked warbler
appears to be loss of habitat following urbanization or conversion to pasture.
While the impact of cowbirds on the reproduction of golden-cheeked warbler is
relatively unknown, it is expected to become a greater problem as habitats
become more fragmented.
This introduction is a brief synopsis of
information presented by Tazik et al. (1992). Extensive information about the
endangered plant and animal species of Fort Hood can be found in this
source.

Breeding range in Texas for black-capped vireos

Breeding range for golden-cheeked warblers
The approach used to develop this spatial model incorporates four components:
- development of a single-cell model using STELLA simulation environment
- development of raster initialization maps using GRASS (Geographic Resource Analysis Support System) geographic information system
- translation, compilation, and multi-celluar simulation using SME (Spatial Modeling Environment) program
- sensitivity analysis conducted on supercomputers found at NCSA (National Center for Supercomputing Applications)
These components are summarized in the diagram below. Detailed descriptions of this approach, also used to develop simulations for other endangered species, can be found by clicking the diagram.

FHASM is based on a quarterly (i.e., 3 months) time step. The population
dyamics of black-capped vireo and golden-cheeked warblers are adjusted in
the second quarter (April through June), which approximates the breeding
season for these species. In constrast, processes related to weather, plant
community succession, impacts, and management practices are updated in each
time step. Each simulation is run for 100 years (i.e., 400 time steps)
FHASM incorporates a gridded landscape in which information is exchanged
among multiple cells. In this model, each cell is 4 ha in area (200 X 200
m) and Fort Hood is represented by a total of 21,540 cells. Each simulation
spans the entire post, an area of approximately 550 sq. km (340 sq. miles).

- Simulated Management Scenarios -
We simulated changes in the amount of effort spent controlling cowbirds and conducting land management (i.e., enhancing habitat for black-capped vireos, maintaining buffers around golden-cheeked warbler habitat, and reducing juniper encroachment). Nine different management scenarios were simulated using FHASM. By comparing results of these scenarios, it is possible to evaluate the relative effects of management activities on populations of endangered species.
![]()
- present-day effort ("baseline")
- double present-day effort ("increase")
- no effort ("none")
- present-day effort ("baseline")
- effort needed to manage all cells with potential ("increase")
- no effort ("none")
All possible combinations of the above levels of management were captured as nine different management scenarios, each simulated over a 100-year time period. Because each simulation run required two to four days on a Sparc1000 server, only a single simulation run of each scenario has been conducted to date. Because there are stochastic processes within FHASM, additional simulations are required to evaluate fully the effects of management strategies on populations of endangered species.
- Output Movies Generated by Simulations -
Results of each simulation were saved as MPEG movies. Viewing these animations provides a sense of qualitatitive changes that occurred over time and also places results in spatial context.
See list of available movies
- Graphical Comparison of Results -
Results of management simulations were also summarized in graphical form
to allow visualization of the changes in the numbers of endangered birds
over time. Although they do not convey spatial information, these graphs
do allow results of different simulations to be compared quickly.
Responses of black-capped vireos
- Effects of changing effort of cowbird control
- Effects of changing effort of land management
Responses of golden-cheeked warblers
- Effects of changing effort of cowbird control
- Effects of changing effort of land management
Principal Researchers
- Steven J. Harper
- Ann-Marie Trame
- James D. Westervelt
Consulting Researchers
- Jocelyn Aycrigg
- John D. Cornelius
- Emmett Gray
- Bruce Hannon
- Timothy J. Hayden
- Leslie Jette
- Thomas Maxwell
- Robert H. Melton
- William F. Seybold
- Michael Shapiro
- Dale Thurber
- Howard Weinberg
University of Illinois Student Researchers
- Leonardo Chapa
- Robb Diehl
- Tracy Galarowicz
- Rob Getz
- Ben Halperin
- Teresa Johnson
- Ryan Lindberg
- Denny Park
- Georgia Sebesta
- Ben Wang
- Mosheh Wolf
- Becky Zerlentes
Avian Species Submodel
The avian species submodel incorporates the habitat suitability and breeding site preferences of each species (see Introduction to FHASM). Habitat suitability is calculated as a function of the plant community type and the time the community has been in the current stage. In addition to habitat suitability, the allocation of birds to breeding sites is influenced by landscape features (patch size and distance to nearest large patch) and past occupancy, which accounts for site fidelity exhibited by these species.
Population processes within a breeding season are influenced by reproductive success and mortality. Seasonal fecundity is calculated using a model developed by Pease and Grzybowski (1995, The Auk, 112:343-363). The number of young successfully fledged is influenced by the probability of parasitism by cowbirds and by the relative quality of the breeding site. Mortality within the breeding season is a function of the probability of fire.
Population processes during periods when the neotropical migrants are not at Fort Hood are influenced by over-winter losses. Losses represent both deaths and permanent emigration from the site.

Submodel Results
Below are maps indicating where birds (left: black-capped vireos; right: golden-cheeked warblers) are allocated on a landscape of varying suitability. Birds in breeding territories are indicated by black marks, and suitable breeding habitat is indicated in red. Results of this simulation indicate that suitable habitat is not uniformly utilized, reflecting the role of other factors such as site fidelity.

Additional results, including changes in population density over time, can be found at Simulations of Alternative Management Strategies.

Fort Hood is within the Lampasas Cutplains physiographic region, and has a landscape of alternating flat valleys and limestone escarpments. Three types of plant communities can be distinguished:
- Edward's Plateau is characterized by broad, fertile valleys between limestone hills. The uplands have thin soils overlying hard limestone, which limits plant growth to hardy species such as shin oak (Quercus sinuata ), scrub liveoak (Quercus fusiformis ), and ashe juniper (Juniperus ashei ). Warm-season grasses dominate the valleys.
- Blackland Praire occurs on clay and silt soils containing a large amount of calcium carbonate. Deep soils can support dense, scrub timber but usually these soils are associated with tallgrass prairie dominated by Indiangrass (Sorghstrum nutans ) and little bluestem (Schizachyrium scoparium ).
- Cross-Timbers and Native Prarie areas are characterized by an overstory of post oak (Quercus stellata ) and blackjack oak (Quercus marilandica ), which ranges from open savannah to dense stands. The understory is composed of tallgrass prairie species. The native plant communities of this area were structured by regular fires before European settlers and cattle ranching influenced the landscape.

Fire suppression and grazing have led to changes in the plant communities. Due to increased encroachment by woody species, presently there are fewer prairies and savannahs and more scrubby woodlands than in historical times. In addition, grasslands have become dominated by Texas wintergrass (Stipa leucotrichia ) and prairie dropseed (Sporobolus asper ) due to a long history of grazing and mechanized military activity.
Plant Community Submodel
The plant community submodel incorporates transitions among 15 unique community stages. Stage transitions reflect community changes that occur due to natural succession, land use policies, and management activities.
Communities are influenced by soil type and topography, and communities are differentially susceptible to invasion by ashe juniper. Stage transitions occur in response to long-term land use policies (grazing and intensity of military activity), short-term impacts (fire and militiary activities), and management activities (prescribed burning, juniper clearing, and habitat enhancement).

Submodel Results
Below are maps resulting from a simulation conducted to investigate effects of current land use policies (left: initial plant communities; right: plant communities in 100 yrs). Results of this simulation indicate increased habitat fragmentation for some, but not all community types due to grazing, training, and fires.

Management

Land managers at Fort Hood manipulate plant communities to maintain habitats that are suitable for endangered species as well as for military training. There are four major management concerns at Fort Hood:
- Black-capped vireos require habitat of mid-successional stage, which occurs in some areas following disturbance. To maintain appropriate habitat, managers disturb vegetation using prescribed burns and mechanical clearing.
- Juniper is an aggressive invader at Fort Hood, especially in intensively grazed areas. Sites that are valued for training activities must be cleared of juniper, typically through prescribed burns and manual removal.
- Golden-cheeked warblers require mature forests to breed. Because this habitat is susceptible to accidental fires caused by military training, managers use prescribed burns to reduce fuel loads in adjacent areas.
- Nesting success of black-capped vireos, and possibly golden-cheeked warblers, is impacted dramatically by cowbird parasitism. Managers use large livetraps to control the number of cowbirds.
Management Submodel
The management submodel determines if individual cells should be considered for management activities. Criteria include time of year, slope, distance to endangered species habitat, time since last occupied by endangered species, and amount of juniper cover.
The dispersion of cells identified with potential for management is examined at a regional scale. Only areas that contain a defined number of cells with potential for management are considered further. Management activities are then simulated for only a small subset of cells within these areas. Preventing management in all cells with potential for such activities reflects constraints on the time and finances of managers at Fort Hood.

Submodel Results
Below are maps resulting from a simulation conducted to identify locations for manual removal of juniper (left: all cells with potential to be managed; center: areas with required number of adjacent cells; right: cells actually managed). Results indicate that the user-defined criteria used in this simulation constrained juniper removal to 15 cells, or 8% of cells identified as having potential for management.

Impacts to Ecosystems at Fort Hood

There are three major sources of impacts to the ecosystems at Fort Hood:
- Military maneuvers. The post provides facilities for training the III Corps, which includes one Cavalry Division with thirteen brigades and several hundred smaller military units. Maneuvers provide realistic experiences for troops, including dismounted infantry training, large-scale tactical exercises, emergency deployment training, and airborne operations. In addition to direct effects on plants, training can impact communities indirectly through compaction and erosion of soil.
- Cattle grazing. Grazing leases provide ranchers open range to graze 3500 animal units over most of the post. Cattle do not utilize the landscape evenly, allowing some areas to become overgrazed.
- Accidental and natural fires. Artillery rounds, tracers, and flares regularly cause fires on the post, and lightening serves as an additional source of fires. Present policy requires immediate suppression of fires; however, it is impossible to control all fires. In a dramatic example, approximately 5,700 ha (14,000 acres) were burned in wildfires that occurred during drought conditions in the spring of 1996.
Impacts Submodel
The impacts submodel determines the frequency and location of fire ignition sources. The pobability of ignition is calculated for each cell based on historical patterns of fire, and is influenced by time of year, location on the post, and type of plant community. The spread of fires to adjacent cells is stochastic and influenced by the relative flammability of each plant community.
This submodel also includes maps that represent impacts caused by military training and cattle grazing. The frequency, location, and intensity of impacts are not determined dynamically, but rather are provided by the user to reflect land use policies.
Submodel Results
Below are maps that indicate locations of 47 accidental fires (left: ignition sources; right: burned cells). Results of this simulation indicate that most fires were extinguished rapidly, while a few fires spread to adjacent cells and burned relatively large areas.

Below are maps that indicate intensity of impacts to plant communities caused by present day land use policies (left: military training; right: cattle grazing; note: increased intensity indicated by darker shades). Patterns indicate greater disturbance due to training on the west side of the post, and concentrated areas of disturbance due to cattle grazing over most of the post.
