Graduation Date

Spring 5-9-2026

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Programs

Environmental Health, Occupational Health, and Toxicology

First Advisor

Jesse E. Bell

Second Advisor

Hongying Dai

Third Advisor

Azar Abadi

Fourth Advisor

Abraham Mengist

Abstract

Coccidioidomycosis is an emerging infectious fungal disease caused by inhaling airborne spores of soil-dwelling Coccidioides, whose lifecycle is thought to be shaped by environmental conditions. However, most prior studies relied on meteorological conditions—primarily precipitation and air temperature (AT)—that might not accurately reflect soil hydroclimatic patterns where the fungus grows. This dissertation advanced mechanistic understanding and predictive capability for coccidioidomycosis through two progressive studies. Study 1 provided the first assessment of concurrent and lagged topsoil (0–10 cm) soil moisture (SM) and soil temperature (ST) effects on coccidioidomycosis incidence, showing that multiyear alternating wet–dry and cool–warm cycles, along with concurrent dry and dusty conditions, were associated with higher incidence across seasons. Study 2 extended this foundation by developing a mechanism-informed multilayer environmental framework—the first applied to any soilborne mycosis—spanning one dust-dispersion layer (PM10, wind speed) and four soil–climate layers: meteorological (precipitation, AT), topsoil, midsoil (10–40 cm), and deepsoil (40–100 cm) SM and ST. Distributed lag non-linear models revealed nonlinear, season-dependent exposure–lag–response associations across all four soil–climate layers, characterised by multiyear alternating wet–dry and cool–warm oscillations that attenuated progressively from topsoil through midsoil to deepsoil. The observed depth-dependent lag structure and vertical divergence across layers provided the first quantitative evidence in the vertical dimension for the prevailing soil-sterilisation and grow-and-blow hypotheses. PM10 exhibited the most consistent concurrent positive associations across seasons and ranked as the most important predictor in the ensemble forecast. A novel two-stage stacked ensemble prediction framework developed in this dissertation demonstrated that combining all five layers outperformed every single-layer model in forecasting incidence; the selected pipeline relied solely on environmental data available within one week of the target month, enabling near-real-time nowcasting. Collectively, this dissertation provides the first evidence linking multilayer environmental exposures to coccidioidomycosis incidence across both temporal and vertical dimensions, and demonstrates that a multilayer framework could advance both mechanistic understanding and predictive performance. Future work might benefit from adopting and refining this multilayer framework and evaluating its applicability in other endemic settings and, potentially, for other environment-sensitive soilborne mycoses.

Rights

The author holds the copyright to this work and any reuse or permissions must be obtained from the author directly.

Download

Available for download on Sunday, April 30, 2028

Share

COinS