Predicting Valley Fever Outbreaks: Novel Mechanistic Models Incorporating Climate and Ecological Interactions.

Area of Science:

• Environmental science
• Mycology
• Epidemiology

Background:

• Coccidioidomycosis (Valley fever) is an environmentally acquired fungal infection endemic to the Americas.
• Climate change is increasing exposure risks in endemic and non-endemic zones.
• Existing statistical models struggle with complex fungal life cycles and climate non-stationarity.

Purpose of the Study:

• To develop and compare mechanistic Ordinary Differential Equation (ODE) models for forecasting Coccidioidomycosis.
• To explicitly map environmental drivers to the fungal life cycle stages.
• To assess the impact of climate data and wildlife reservoirs on forecasting accuracy.

Main Methods:

• Developed successive ODE model iterations incorporating soil moisture, temperature, and wildlife dynamics.
• Calibrated models against human case data from Arizona.
• Derived a time-variant environmental reproduction number to assess transmission potential.
• Employed statistical tests (Diebold-Mariano, Modified Diebold-Mariano) for comparative forecasting analysis.

Main Results:

• Mechanistic models without climate data underperformed statistical baselines.
• Integrating climate data improved predictive power to levels comparable to statistical models.
• Explicitly incorporating a wildlife reservoir significantly enhanced forecasting accuracy over statistical baselines.
• Model performance was rigorously evaluated using various statistical metrics.

Conclusions:

• Mechanistic models offer a biologically grounded approach to forecasting Coccidioidomycosis.
• Climate data and wildlife reservoirs are critical factors for accurate disease burden prediction.
• This framework can guide public health interventions in response to climate-driven changes in disease risk.