Improved prediction of climate impacts on migratory pathways through machine learning, hydrologic modeling, and network theory

Timothy Keitt, University of Texas at Austin

Abstract

Climate change is predicted to have profound effects on precipitation regimes worldwide. Wetlands and wetland-dependent spaces may be strongly impacted by increases and decreases in rainfall. We propose to model historical and projected changes to wetland distributions across the United States to assess future threats to wetland biodiversity in general, and migratory wetland birds specifically. Our long-term goal is to contribute to the understanding of mobile organisms’ population ecology in dynamically changing landscapes and to use this knowledge to improve management of species and ecosystems impacted by climate change. The objectives of this project are to advance the state-of-the-art in modeling the impacts of climate change on wetland biodiversity by integrating both climatic and hydrodynamic predictions into models of species distributions and population connectivity. Specifically we will build habitat-driven climate change forecasts for wetland birds utilizing the output of advanced land-surface/sub-surface components of the latest global climate models.

To meet our objectives, we will focus on two main areas. First, we will develop climate and habitat driven models for historic breeding, wintering and migratory distributions of wetland birds in North America. Second, we will assess potential climate change impacts on migratory wetland bird species under varying climate scenarios.

During the first phase of this project, we will create a historical model of wetland distribution across the United States. Simulation outputs of the Weather Research and Forecasting Model (WRF; a regional climate model) with boundary conditions set by the outputs from the Community Climate System Model, will be used to develop proxies for increasing and decreasing wetland saturation. Wetland areas will be identified using a digital elevation model and other geological inputs. This phase of modeling will be validated using historic wetland field observation across regions and wetland delineations provided by the National Wetland Inventory.

The second phase will be to model the historical breeding and wintering ranges. We will use data from the North American Bird Banding Lab and Wetland Breeding Population and Habitat Survey. The next stage will be to estimate the migratory pathway model. Finally, we will create the future projections of the migratory pathway model. Future predictions will be driven by the climate model outputs. We intend to use the output from multiple runs that varying according to climate change scenario.

The deliverables of this project will be contributions to both the scientific knowledge base of biogeographical predictions of climate change and human resource assets. The proposed research will increase the ability to more robustly predict species range shifts. This project uses state-of-the-art climate models in a completely novel manner to improve methods on species range predictions. This project will also forge a new relationship between the ecological and climate modeling communities. Current research in both field is intensively pursuing questions related to climate change; yet almost completely independently.