The distributions of bird species are limited by factors of the environment. Different species have different ranges of environmental conditions that they tolerate. Some of the conditions that birds are constrained by will shift as climate changes. In fact, some species are already shifting their ranges toward the poles or to higher elevations. Researchers have used statistical models to relate spatial environmental data and the occurrence of bird species. They then change the data to reflect the way conditions may appear in the future and use the models to predict the ways in which bird species will shift their distributions as a result. This approach has been criticized because it ignores the interactions between species, that species can adapt to new environments, and that some species are better at dispersing than others. A new method that avoids these limitations will be used in this project. Statistical models will be used to predict the occurrence and relative abundance of birds across the United States that build on a long running survey of bird occurrences along with a suite of spatial data. The models will then be incorporated into a simulation model using an agent-based framework, where agents are the group of birds of an individual species that occupy a small part of the landscape. The models that describe the ways in which birds are distributed will be allowed to evolve in response to the environments available and the presence of other species. For example, small changes in the temperature range of a species may occur as it moves into new landscapes. Adjustments will then be made to the simulation model until the distribution of birds is similar to their real, current distributions. Finally, response surfaces that reflect climate change effects will replace those that represent current conditions and the simulations will be continued. The distributions of bird species will thus evolve in response to the new conditions and the results will reflect how bird distributions and communities may change under a future climate, considering interactions between species and the ability of the species themselves to adapt.
Models that represent how individuals respond to changing environmental conditions are rare in studies of how species are distributed. The method used here has the potential to address a variety of questions and to make analyses dynamic, with patterns of species distributions developing over time during the computer simulations. The effort will support a PhD candidate who will work with a new team, with new data sets, and learn new research skills. The results will be used in university courses and workshops and to train high school students. Results will be published in journal articles and summarized on a project web site.
