The costs of uncertainty: How the resolution of environmental data affect the predictions of mechanistic models

Background/Question/Methods

Recently, models of population dynamics have been used to infer the impacts of climate change on the distributions of species. The predictions of these models depend greatly on parameters that characterize food and temperature. Currently, the temporal and spatial resolution of these parameters differs greatly from that experienced by individuals. For example, records of monthly minimal and maximal temperatures fail to capture diel variation that might have a large impact on organismal performance. Similarly, the abundance of food is assumed to be unlimited or constant throughout the range, primarily because prey densities are usually unknown. Using an individual-based model of population dynamics, we explore the effects of thermal variation and food abundance on the life-history and geographic range of a widespread lizard, Sceloporus undulatus. Simulations were run using different temporal resolutions of temperature (ranging from hourly to monthly) and different levels of prey abundance (ranging from minimal to maximal values reported in the literature). 

Results/Conclusions

We found that the climate resolution and food availability greatly impacted the outcome of our simulations: Embryos had a lower chance of surviving to the hatchling stage in simulations using hourly climate data than they did in simulations using mean monthly climate data. Moreover, lizards were less likely to survive the winter in simulations with hourly data, probably because of a decrease in the potential duration of activity during the summer. Food availability affected the age of maturity; lower food availability resulted in less energy for juvenile growth and delayed maturation. Our results indicate that a greater resolution of data for environmental temperature and food availability should improve our understanding of life histories, population dynamics, and geographic ranges.