Quantifying methods to improve statistical power in grassland and passerine bird nesting studies

Abstract:

<p>Studying nests of grass- and ground-nesting birds is challenging because nests can be rare on the landscape and camouflage can make them difficult to locate. Yet, large sample sizes are needed understand complex interactions between nests and their environments. In this thesis, I present research on two technological approaches for addressing these challenges: thermal-imaging cameras for locating nests and thermal data loggers for monitoring nests.</p>

<p>In Chapter 2, I examine the use of a thermal-imaging camera for increasing the detection probability of warm bird nests against cooler background temperatures. Carrying the imager while searching for nests and using it when thermal conditions allowed did not result in a greater detection probability for bird nests. I believe this was due to prevailing sunny conditions obscuring the thermal signal from nests during the majority of the surveys, as well as dense vegetation blocking the thermal signal in much of the habitat we searched. Some study systems may benefit from use of a thermal imager, but I found the money used to purchase the device would have been better invested in additional technician hours searching for nests with traditional methods.</p>

<p>In Chapter 3, I describe a study where I tested (1) the impact of iButton® thermal data loggers on hatching success in grass- and shrub-nesting passerines overall and (2) as separate groups for species that either accept or reject brown-headed cowbird (<em>Molothrus</em> <em>ater</em>) eggs. I also tested (3) the difference between estimating daily survival rates of nests using data derived from in-person visits and data derived from iButtons. In addition, I present a simulation study to provide a general description of differences observed between in-person visits and iButtons for estimating daily survival rate. I found that iButtons did not affect hatching success probability for grass- and shrub-nesting passerines overall or for nest parasite egg-accepting and egg-rejecting species as separate groups. I did see a trend in the data to suggest nest parasite egg-rejecting species’ nests that received iButtons hatched at a lower rate, but the sample size was not sufficient to reach a definitive conclusion. Using iButtons to estimate success or failure dates in real nests of three species resulted in 0.3–0.9% higher estimates of daily survival rate, 12.2–15.3% larger sample sizes, and 2.0–6.3% reduced standard errors. The direction and trend of these measures was supported by the simulation study, although the magnitude of the effect was much smaller in the simulated data, possibly due to lower variation in the simulated data.</p>

<p>These studies of thermal-imaging cameras and iButton thermal data loggers provide guidance to researchers regarding available technological solutions for two common challenges in conducting bird nest studies. Research funds may be better spent on technician hours than on thermal-imaging cameras for most nest studies for ground-nesting birds, and iButton thermal data loggers represent an efficient method for increasing sample sizes and decreasing standard errors when estimating daily survival rates.</p>

File: 

stephenson_2017_msthesis.pdf