Date of Award

5-15-2022

Document Type

Thesis

Degree Name

Master of Science in Coastal Marine and Wetland Studies

Department

Coastal and Marine Systems Science

College

College of Science

First Advisor

Christopher E. Hill

Second Advisor

Derek P. Crane

Third Advisor

John J. Hutchens

Abstract

Grassland birds, such as the loggerhead shrike (Lanius ludovicianus), are threatened by extensive land-use change, including urbanization. Loggerhead shrikes have been declining since the early 1900s and are rare or extirpated in portions of their former range. Obtaining reliable population estimates of loggerhead shrikes is important to identify demographic trends and ensure conservation decisions are based on an accurate assessment of their current status. Imperfect detection, or the inability to detect every shrike during a survey, can result in biased population estimates. Estimating detection probability is labor-intensive and requires assumptions that are difficult to satisfy in most field studies. Understanding a species' detectability before beginning research is thus ideal. In chapter one, I calculated the detection rates of long-term resident loggerhead shrikes along a highway corridor dominated by commercial and industrial development in Horry County, South Carolina. Using a generalized linear mixed model, I evaluated factors that could affect detection probability. All 47 shrikes included in the study were color-banded and monitored about twice per month year-round, from January 2019 through December 2021, allowing observers to determine occupancy and sex and track the status of each bird. Collectively, observers conducted 2,208 qualifying surveys on occupied territories. The best-fitting models showed that all predictor variables, including observer identity, season, sex, year, and the interaction between sex and season, affected detectability. The average detection rates for males and females had overlapping 95% confidence intervals, as did the years 2019, 2020, and 2021. Loggerhead shrikes had an average detection rate of 0.58 (range = 0.28–0.75, SD = 0.10) and were detectable year-round, with an average detection rate of 0.67 (SD = 0.17) during the breeding season, 0.56 (SD = 0.17) during winter, and 0.51 (SD = 0.16) during the molting season. The breeding season was the only season where the 95% confidence intervals for male and female detection rates did not overlap, with males having a detection rate of 0.70 (SD = 0.17) and females 0.60 (SD = 0.19). Females likely had a lower detection rate during the breeding season because they were unavailable for detection while incubating. Observers had detection rates ranging from 0.47 to 0.68 (SD = 0.08), suggesting variation among observers could affect population estimates. The results of chapter one can be used to improve the accuracy of population estimates by providing correction factors to account for the imperfect detection of urban shrikes. The correction factors can be used to identify the appropriate number of repeat surveys, increasing the efficiency of survey protocols and decreasing the costs of demographic studies. In chapter two, I identified habitat characteristics associated with loggerhead shrike occurrence along urban gradients in Charleston, Florence, Georgetown, and Horry counties, South Carolina. I surveyed 300 2.25-ha sites with impervious surface levels ranging from 3.2 to 96.3%. Using a generalized linear model, I assessed how impervious surface (%), highway density (m/ha), other road density (m/ha), total road density (m/ha), canopy cover (%), and powerline density (m/ha) are related to loggerhead shrike occurrence. Because loggerhead shrikes have been documented responding to habitat characteristics at multiple spatial scales, I measured predictor variables at the 2.25-ha, 20.25-ha, and 1-km scales.

Download

Share

COinS