An integrated resource selection model to predict avian habitat use from grassland vegetation structure
Thursday, August 5, 2021
Link To Share This Poster: https://cdmcd.co/5mqQB9 Live Discussion Link: https://cdmcd.co/QM8kvn
Katy M. Silber and W. Alice Boyle, Division of Biology, Kansas State University, Manhattan, KS, Trevor J. Hefley, Department of Statistics, Kansas State University, Manhattan, KS
Presenting Author(s)
Katy M. Silber
Division of Biology, Kansas State University Manhattan, KS, USA
Background/Question/Methods Grassland birds are highly mobile; many disperse among sites within and between years, yet we know little about how they make post-dispersal settlement decisions. At broad scales, birds often occur in higher densities in areas with heterogenous vegetation, but identifying the vegetation structures that some birds select at the territory level has remained elusive. Grassland vegetation varies in space and time; fire, grazing, and weather variability create heterogeneity in plant communities. Existing methods that quantify resource selection often create spatio-temporal misalignment between habitat and demographic data. To investigate the extent to which vegetation drives grassland bird settlement decisions, we collected point-level data on vegetation height and composition once per month from May-July 2014-2020 at the Konza Prairie Biological Station. Konza is a Long-Term Ecological Research site in NE Kansas consisting of native tallgrass prairie, managed using experimental combinations of fire and grazing. We developed a predictive model for vegetation height and cover at any point in space and time. We integrated these predictions with avian mark-recapture data using an inhomogenous point process model to evaluate resource selection for Grasshopper Sparrows (Ammodramus savannarum), Dickcissels (Spiza americana), and Eastern Meadowlarks (Sturnella magna). Results/Conclusions We predicted grassland vegetation at each point and time present within our avian mark-recapture dataset and found each species selects different vegetation structure and composition based on their life histories and nesting behaviors. Grassland bird densities throughout the study site changed within and among breeding seasons to track changes in vegetation, suggesting vegetation drives settlement decisions in these species. Grasshopper Sparrows and Eastern Meadowlarks selected for areas with multiple cover types, and Dickcissels were more likely to occur in areas with woody vegetation than Grasshopper Sparrows and Eastern Meadowlarks. This study provides a framework for future research in integrated resource selection modeling, which can provide insight into land management for declining grassland birds.