Mapping the Distribution of a Soil-Borne Human Pathogen: Coccidioides
Mark W. Bultman, U.S. Geological Survey
F.S. Fisher U.S. Geological Survey
Mark E. Gettings U.S. Geological Survey
Coccidioidomycosis is a public health issue of increasing importance to people in the southwestern United States and in parts of Central and South America. It is caused by Coccidioides, a dimorphic soil-inhabiting fungus. The saprophytic phase of the fungus is characterized by branching segmented hyphae that form a network of mycelium in the upper horizons of soils. As the fungus matures, it produces arthroconidia that can be separated by soil disturbance (natural or anthropogenic) and consequently dispersed by the wind. If airborne arthroconidia are inhaled by an appropriate host, primary infection may occur and the parasitic phase of the Coccidioides life cycle is initiated.
Habitat modeling of the saprophytic phase of the Coccidioides life cycle is difficult due to the limited number of known growth sites. This confounds the determination of statistical relationships among physical, chemical, and biological habitat parameters. Laboratory and site-specific field studies have determined many of these parameters. The modeling scheme therefore must use these parameters as input data and transform them into output that describes the favorableness of soil for hosting Coccidioides at all locations in the study area.The model should be able to deal with data of differing precision and accuracy and should reduce a potentially intractable number of model relationships to a smaller modeling framework with reduced dimension. The modeling technique chosen to do this is a fuzzy system. Because this modeling technique will be applied to a large number of spatially distributed cells within a raster Geographic Information System, the approach is referred to as a spatial fuzzy system.
In addition, a model of the spread and survival of Coccidioides in soil via wind-borne arthroconidia transport has been completed using public domain, agent-based modeling software. The model results mimic what is seen in nature and indicates that 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.