Movements, habitat use, and stopover duration of migratory songbirds in the western Lake Erie basin of northern Ohio
Ashley A. Buchanan, MS
Advisor: Paul Rodewald
Migration is often considered the most energetically demanding and stressful period during the life cycle of migratory landbirds. Because migrating birds spend the majority of their time in stopover habitats rather than in flight, it is important to our understanding of migration ecology to examine spatial and temporal patterns of bird movements during stopover periods. Stopover habitats are important because they provide migratory birds a place to rest, refuel, and avoid predators. The goal of my research was to evaluate movement behavior and habitat selection of forest patches by migrant songbirds during spring stopover in a highly fragmented, agriculturally dominated landscape during spring. My specific objectives were to 1) assess patterns of movement by these migratory landbirds within and among forest patches, 2) determine how local-level (patch area, habitat type) and landscape-level (habitat connectivity) attributes influence movement rate and patch residence time, and 3) examine factors influencing stopover duration. I used radio-telemetry to track individual Yellow-rumped Warblers (Dendroica coronata; YRWA) and Red-eyed Vireos (Vireo olivaceus; REVI) during spring migratory stopover in the western Lake Erie basin of northern Ohio from mid April to early June of 2006 and 2007. Monitoring migrants allowed me to document movement behavior, habitat selection, and stopover duration in this important region for migratory landbirds. In addition, I measured local (within patch) and landscape level habitat characteristics to evaluate their relative importance in habitat selection. Red-eyed Vireos and Yellow-rumped Warblers were captured, color-banded, and promptly transported to one of five pre-determined release sites. Birds were fitted with a small radio transmitter and released. Experimental relocations of migrants to different sites were used to test for habitat features that influence stopover behavior and to control for arrival timing. As a control for relocated birds, REVI and YRWA were also captured at each of the five release sites, processed, fitted with a radio transmitter, and released within the same site. Location data were collected at 30-minute intervals on days 1, 2, and 3 after capture (YRWA) and days 1 and 2 after capture (REVI). Birds were checked once per day thereafter to determine minimum stopover duration.
A total of 38 YRWA and 31 REVI were radio-tracked from 20 April to 1 June in spring 2006 and 2007. Movement data were analyzed using a generalized linear mixed model with maximum likelihood estimation. The experimental unit was the bird-patch-day, and the response variable was either residence time within patch or movement rate within patch for each bird-patch-day combination. The random effect was the individual bird in both sets of analyses. Fixed effects were patch habitat type, patch area, patch connectivity, year, and translocation. Minimum stopover duration was calculated as the difference (in days) between the initial release date and the date the bird was last relocated in the study area. Data on stopover duration were analyzed for biotic and abiotic effects using a generalized linear model for categorical analysis of Poisson count data. Following an information-theoretic approach, I used Akaike’s Information Criterion with a correction factor for small sample sizes (AICc) to rank candidate models for both movement behavior and stopover duration of each species.
Both species exhibited high variability in movement behavior. Mean total distance moved per day was 2.2 km for YRWA and 0.9 km for REVI. Mean movement rate (m min-1) per day was 3.5 for YRWA and 1.8 for REVI. Patch habitat type, area, and connectivity were important in describing variation in patch residence time and movement rate for YRWA. Patch habitat type, patch area, and translocation influenced patch residence time for REVI, but none of these factors was important in describing their movement rates. Mean minimum stopover duration for YRWA was 7.0 days (±3.6 S.D.) in 2006 and 4.8 days (±2.7 S.D.) in 2007. In contrast, mean minimum stopover duration was 1.6 days (±0.9 S.D.) in 2006 and 1.7 days (±1.6 S.D.) in 2007 for REVI. The model that best described stopover duration for YRWA included energetic condition, year, and release site. High variation in candidate models suggested that none of the models adequately described stopover duration for REVI. There was also significant variability in the movement behavior of individuals.
Spring migrants selected and spent more time in mature hardwood, beach ridge, and willow/cottonwood forest, suggesting a higher importance of these habitat types for conservation. Migrants also frequently used hedgerows as travel corridors and foraging locations, an indication that hedgerows may be an important landscape component for migrants and should be maintained to provide connectivity between habitat patches. Stopover duration in the western Lake Erie basin can last many days for some species suggesting that this study area may significantly contribute to the energetic balance of many migrants. My study suggests that land managers should conserve and restore these important habitats (mature hardwood, willow/cottonwood, and beach ridge forests), so that migrant songbirds have more opportunities to locate resources and efficiently refuel before continuing migration.