Factors influencing spatial patterns and trends in migratory landbird stopover across Texas and Louisiana
Date
2023
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
To optimize energy and speed each migration season, Nearctic-Neotropical landbirds must navigate the climate of the airspace during flight, and the heterogeneity of the landscape as they stopover to rest and refuel. With a diverse environment ranging from arid to temperate, proximity to the Gulf of Mexico, the presence of some of the brightest cities in the United States, and the existence of multiple major migration routes, Texas (TX) and Louisiana (LA) represent a unique landscape to study how climate and land cover factors influence where migrants choose to stopover, how these relationships vary seasonally, and how they change over time. I used 26 years of data (1995-2020) from 15 NEXRAD weather surveillance radars distributed across TX and LA to calculate bird densities and distributions in relation to environmental variables. I first fit boosted regression tree (BRT) models among an ensemble of 196 overlapping modeling frames (400 km x 400 km in extent) stratified across the study area to determine the relative importance and beta relationships between 26 environmental predictor variables and 13 years of seasonal mean bird density. I found a high degree of spatial and seasonal nonstationarity in variable importance and the directions of beta relationships, reflecting potential differences in species compositions and migration strategies, as well as complex interactions between abiotic and biotic factors influencing stopover distributions. Consistently important variables included temperature, precipitation, distance to artificial light at night (ALAN), and east-west winds. Interestingly, I found evidence of attraction to bright light at the broad scale and avoidance of bright light at the fine scale that varied depending on the mean distance to ALAN across the landscape. My results suggest that attraction to bright light was greater in areas that were on average further from ALAN, while avoidance of bright light was greater in areas that were on average closer to ALAN. To expand upon my findings and investigate how changes to the landscape and climate influence changes in stopover densities and distributions over time, I used generalized additive models (GAMs) to examine the relationships between 35 predictor variables, their top-ranking interactions, and trends in seasonal mean bird density from1995-2020. Similar to the results of the spatial BRT models, I found a high degree of seasonal and geographic variability across TX and LA. Generally, however, temperature, east-west winds, and their interactions were highly important variables in influencing bird density trends over time. In the fall, stopover density increased in warmer areas and areas that became warmer over time, while in the spring, cooler areas and areas that became cooler were correlated with increases in bird density. In addition, spring stopover density increased further from bright ALAN where winds were blowing fastest from the east, suggesting that easterly winds may be outweighing the effects of light attraction. Spring stopover density decreased further from bright ALAN where winds were blowing fastest from the west, suggesting that westerly winds may be contributing to the effects of light attraction. In the fall, easterly winds may be outweighing the effects of light attraction. Overall, quantification of the direction and magnitude of these relationships is novel for the south-central region of the U.S. and provides a strong foundation for future studies that investigate the regional relationships I found in more detail. Moreover, understanding the effects of climate trajectories and their combinations with factors like ALAN intensity and urban development, can aid in enacting preventative measures for landbird conservation.
Description
Keywords
Boosted regression tree, Artificial light at night, Landbird conservation, Migration season