The
Desert Tortoise Model:
- A Spatially Explicit Simulation -
Overview
- Recent History of the Desert Tortoise -
The desert tortoise is a herbivorous reptile inhabiting the Mojave Desert
of California. Recently experiencing a population decline, it has consequently
been listed as a Federally Threatened Species. The species is difficult
to survey due to its patchy distribution, low population densities (20 tortoises/sq.
mi.), and widely spread population over millions of hectares in the Mojave
Desert. Natural resources managers in Federal and State agencies are in
critical need of technologies that can improve the accuracy and/or predictive
capabilities of assessment and monitoring programs for threatened, endangered,
and sensitive (TES) species.
The U.S. Department of Defense is a unique example of an organization working
to better understand and manage complex landscape systems. Possessing large
tracts of land that function as military training centers, the military
is mandated by guidelines in the Endangered Species Act to manage and preserve
ecosystems. To that end, a spatially explicit model was designed to help
manage and protect the desert tortoise (Gopherus agassizii ) at Fort
Irwin, a U.S. Army training center in the central Mojave Desert of California.
- A Dynamic Landscape Simulation -
The primary goal of this research was to introduce military installation managers
to dynamic landscape simulation (DLS) for predicting the results of human interaction
with the environment. A DLS is initialized with system state information, typically
captured in a geographic information system (GIS) and predicts future states
through simulation models that capture the system's dynamic interactions.
The Desert Tortoise model is designed to assist land management offices of the
U.S. Department of Defense in finding an appropriate balance between training
activities and preservation of desert tortoise habitat, for the sake of complying
with the Federal Endangered Species Act of 1990. Non-military land management
offices may draw benefit from our experience, using the general framework for
developing similar models.
A 57 x 57 cell grid of 1 sq. km resolution represents Fort Irwin's landscape
from the perspective of the tortoise, similar in size to its natural home range.
Rastor-based GIS maps input the initial physical conditions at the installation.
Variables include elevation, slope, aspect, soil types, soil compaction, soil
moisture, vegetation coverage, and tortoise densities. A one month time step
captures the dynamics of seasonal climatic changes, vegetation cover, military
operations, and tortoise population growth and migration.
Five STELLA submodels interact to form the Desert Tortoise Model. Each sub-model
was developed by a different research team made up of students with the appropriate
expertise.
Climate submodel
includes the monthly temperature, surface temperature, precipitation, evapotranspiration,
and soil moisture.
Vegetation submodel
includes satellite images from thematic mapper bands 1 through 7, elevation,
slope, watershed, and road buffer. The amount of vegetation on Fort Irwin was
quantified in units of percent aerial cover rather than as numbers of plants
or amount of biomass.
Training Impacts submodel
determines the indirect impacts of military training on tortoise populations.
Direct impacts were decidedly less important than those resulting from landscape
modification.
Tortoise Dispersal Movement
submodel simulates movements of tortoises between adjacent cells
as well as calculate changes in juvenile, adult, and elder tortoise densities
due to dispersal. Values generated in this submodel are utilized by the tortoise
population dynamics submodel to adjust tortoise densities within each cell.
Tortoise Population Dynamics submodel serves as the key component
of the overall model. Inputs from the four other submodels coalesce in this
section and determine the potential impacts that different levels of habitat
quality and human landscape management will have on the desert tortoise population
at Fort Irwin.
This simulation effort demonstrates the possibility of applying
Dynamic Landscape Simulations to the toolbox of military installation land use
managers. Recent developments in GIS and modeling software have made it possible
to model a non-homogenous landscape, exchanging diverse information among the
gridded landscape cells.
While models can aid in making sense of the various components of an ecosystem,
they should not be meant to replace field experimentation, as this is crucial
for verification of the models. Land use managers should play a key role in
design and implementation of a model to ensure its usefulness.
Future research for this particular model should include obtaining more
accurate tortoise dispersal and military training data. Additional simulation
scenarios could determine the sensitivity of various parameters, aiding the
design of optimal spatial and temporal patterns for different levels of miltary
training. There are two basic methods to use. The first involves using the same
sensitivity perturbation in all affected cells, but varying the proportion of
cells that are affected by the perturbation. The second approach involves the
use of multivariate permutation tests that could be used to determine if two
output maps were "significantly" different. Although these tests are
relatively new they are already available as computer programs and published
in peer-reviewed journals (Reich,1993).
