Predicting the spread of invasive species and responses of species to climate change both face the challenge of predicting responses in climates and regions for which data on species performance is often unavailable. Moreover, in each case different types of data are available to inform how species may respond to variation in climate, from simple occurrence data to demographic experiments under controlled conditions. Ideally we would use all available information about species’ tolerances and niche boundaries to build cohesive predictions of expected performance under changing conditions. However, it proves challenging to integrate multiple data sources because they may be spatially disjunct and reflect different stages of population dynamics (e.g. recruitment of seedlings, abundance, fecundity). In this study, we explore how we might make better use of available data for improving predictions of species responses to climate change. We develop a cohesive analysis that integrates multiple data sources. In particular, we use hierarchical (multilevel) models to integrate different data sources informing the suitability of different climatic conditions for a species. We use a case study of an invasive species, Celastrus orbiculatus, in the northeastern U.S. to evaluate this approach.
Results/Conclusions
We present a general framework for integrating diverse data into a best understanding of species’ responses to climatic gradients. We use a case study of one invasive species (Celastrus orbiculatus) for which we combine data sources to improve predictions of potential invaded range. We then use this case study to show how models of habitat suitability using a combination of distribution, abundance and demographic data alter predictions of expected responses to climate change. Invasive species are also shown to be useful model organisms for predicting responses to climate change due to their recent spread across fragmented landscapes. Our results emphasize the utility of hierarchical, process-based models for understanding and forecasting changes in species distributions under scenarios of climate change. Although a lack of data on species’ responses to future climates will remain a challenge, we argue that efforts to assimilate data from multiple sources will aid prediction under future scenarios.