HR: 0830h
AN: B21B-0713 [PDF]
TI: Agent-Based Modeling of Physical Factors That May Control the Growth of Coccidioides immitis (Valley
Fever Fungus) in Soils
AU: * Gettings, M E
EM: mgetting@usgs.gov
AF: U.S. Geological Survey, 520 N. Park Ave, Rm 355, Tucson, AZ 85719 United States
AU: Fisher, F S
EM: ffisher@swfo.arizona.edu
AF: U.S. Geological Survey, 520 N. Park Ave, Rm 355, Tucson, AZ 85719 United States
AB:
A model of the spread and survival of the fungus Coccidioides immitis in soil via wind-borne spore transport has been
completed using public domain agent-based modeling software. The hypothetical model posits that for a successful new site to
become established, four factors must be simultaneously satisfied. 1) There must be transport of spores from a source site
to sites with favorable soil geology, texture, topographic aspect, and lack of biomass competition. 2) There must be
sufficient moisture for fungal growth. 3) Temperature of the surface and soil must be favorable for growth. Finally, 4) the
temperature and moisture must remain in favorable ranges for a long enough time interval for the fungus to grow down to
depths at which spores will survive subsequent heat, aridity, and ultraviolet radiation of the hot, dry season typical of the
Southwest U.S. climate. Using agent-based modeling software, a model was built so that the effects of combinations of these
controlling factors could be evaluated using realistic temperature, rain and wind models. The rain probability and amount,
temperature annual and diurnal variation, and wind direction and intensity were based on the weather records at Tucson,
Arizona for the 107-year period from 1894 to 2001. Favorable ground was defined using a fractal tree algorithm that emulates
a drainage network in accordance with observations that favorable sites are often adjacent to drainage channels.
Numerous model runs produced the following five conclusions. 1) If any property is not isotropic, for example wind direction
or narrow paths of rainstorms, parts of the favorable areas will never become colonized no matter how long the model runs.
2)The spread of sites is extremely sensitive to moisture duration. The amount of wind and temperature after a rain control
the length of time before a site becomes too dry. 3) The distribution of wind and rainstorm direction relative to that of
the favorable sites is a strong control on the spread of colonization. East-west winds across an area that has mostly
north-south favorable sites restricts spread strongly. 4) Soil temperature was the least sensitive control in the model,
although it does control the ultimate dormancy of a site. Fifth, the model results cover the spectrum of complete
colonization of all favorable sites from a few source sites to none, one, or two new sites in three years of model
simulation. This implies the probability of new sites depends on the four factors in a Bayesian way. These results indicate
that the complexity introduced in the model from site favorableness, temperature, moisture, and duration of favorable
temperature and moisture conditions is adequate to explain observed distributions of real sites.
DE: 0400 Biogeosciences
DE: 3210 Modeling
DE: 3220 Nonlinear dynamics
SC: Biogeosciences [B]
MN: 2003 Fall Meeting