Mechanistic models to forecast the response of an insect fungal pathogen to global climate change

Background/Question/Methods

Fungal pathogens represent a group of diseases in which climate change is expected to have direct, dramatic impacts due to the strong influence of environmental factors on fungal life cycles and ecology. The effects of climate on pathogen transmission, however, are poorly studied, and consequently it is unclear whether climate change will cause epizootics of fungal pathogens to be more or less frequent. By using a combination of field experiments and model-fitting, we are constructing mechanistic models predicting how one such pathogen, Entomophaga maimaiga, which serves as a biological control for invasive gypsy moth caterpillars (Lymantria dispar), will respond to different climate change scenarios. To investigate how this disease functions under different climate conditions, we collected epizootic, experimental and weather data during natural E. maimaigaoutbreaks at four different sites along a 395.9 km latitudinal gradient weekly for three summers. We experimentally measured fungal force of infection by placing cages of healthy insects at each site for 24 hours each week. Using these data, we fit and compared weather-dependent epidemiological models with different functional responses to weather. We tested universal and site-specific models and estimated parameter values using maximum likelihood calculations. We compared models using AICc.

Results/Conclusions

A site-specific model in which fungal infection rate depended on total accumulated rainfall over the previous 4 to 9 days was the current best-fitting model. A universal model in which infection rates depended on accumulated rainfall over the previous 5 days received the next highest ranking (ΔAICc = 58.9). Between-site differences are likely due to inherent differences in soil quality and pathogen densities at each site. While meteorological models predict changes in mean precipitation levels will occur in some areas, an overall increase in precipitation variance is expected to occur globally. Our data suggest that E. maimaiga transmission is a strongly nonlinear function of rainfall. We therefore predict that the higher variance in rainfall will lead to more severe, but less frequent, pathogen epizootics. Because pathogenic fungi naturally help control pest-insects, understanding how these diseases will respond to climate change will allow us to construct more informed long-term pest-control strategies. Moreover, our results provide one of the first clear examples of the effects of weather on pathogen transmission. We plan to explore and test other spore germination models in the future and will incorporate our findings into more realistic models of within- and between-season disease transmission.