Meeting Abstract
Temperature is an important driver and constraint on life history traits of many organisms. Temperature can also impact disease systems, for instance by influencing traits like growth or reproductive rate of pathogens and hosts. For example, in the amphibian chytrid fungus system Batrachochytrium dendrobatidis , temperature is likely a key factor determining the infection rates of frogs. The role of temperature in disease systems has often been studied primarily with highly controlled experiments where the temperature is held constant. Extrapolating from the constant temperature experiments to temperature regimes more like those experienced under natural conditions is difficult and the theory relatively under-developed. One of the most common methods used to predict the effect of varying temperature on performance traits, based on constant temperature data, is rate summation. Recent work has indicated that rate summation performs poorly when tested directly in the lab. We seek to test this method in the amphibian chytrid fungus system while incorporating uncertainty in the fitting process. More specifically, we fit a Bayesian hierarchical logistic model to the optical density data that span 10 temperatures (13oC to 28oC), were the logistic growth rate is constrained by a Briere function. Posterior samples of parameters determining the shape the Briere function takes are then used to make predictions about how chytrid grows in varying temperature regimes. Data from the varying temperature experiments are then contrasted with the predictions. We find that, even including uncertainty in the parameter estimates, rate summation does a poor job of predicting growth under a time-varying temperature regime. This highlights the need for new theory to link constant and time varying temperatures.
