American Redstart Wintering Ecology

I recently finished 6 years of studying redstart wintering ecology for my Ph.D. at Tulane University. The main goal of this work was to determine how food availability and conspecific density combine to determine redstart body condition and migration timing. In pursing this goal, I first assessed and modified methods for estimating redstart food availability. Then, in order to better understand how redstarts use space in all three spatial dimensions I refined methods first used in eels to model 3D space use for the first time in birds. Next, to determine causal pathways between food availability, redstart space use, body condition, and migration timing, I carried out a territory-scale manipulation of food using an insecticide. Finally, I used four years of observational data to clarify how density-independent food limitation and density-dependent regulation interact to determine redstart body condition and migration timing.

Comparing the precision, accuracy, and efficiency of branch clipping and sweep netting for sampling arthropods in two Jamaican forest types

Estimating food availability is one of the most difficult aspects of avian ecology and there are dozens of methods. Sweep netting and branch clipping are two somewhat commonly used methods and both have been used in Jamaica to make inferences about food available to redstarts. I initially began sweep netting during my first field season but quickly noticed that field assistants of different sizes and strengths were collecting very different amounts of vegetation and presumably insects. Inter-observer bias is a problem in nearly all forms of measurement and to properly estimate this for sweep netting, I collaborated with several other researchers in Jamaica. We devised a study to directly compare inter-observer bias in both sweep netting and branch clipping. In addition, collecting these data would allow us to comapre these two methods in a number of ways to make inferences about their precision, accuracy, and efficiency. Ultimately, we found that both methods had considerable inter-observer bias.  Accuracy is notoriously hard to asses, but we did find evidence that sweep netting was more accurate in measuring habitat differences in food availability. Finally, we found that sweep netting was much more efficient in terms of collecting large amounts of insects. As a result of this study, sweep netting became my insect sampling method of choice for this system and I subsequently carried out all of the sweep netting myself to avoid problems with observer bias. This research was published in the Journal of Field Ornithology in 2012.
download pdf


Two neighboring American Redstart territories represented in 3 dimensions. Red and blue points represent individual locations used to create the 3D utilization distributions.

Modeling three-dimensional space use and overlap in birds

How animals use space in their environment has fundamental behavioral and ecological implications. Utilization distributions have become one of the most common methods for estimating space use in animals. Redstarts like nearly all species of birds also use space in 3D, as they are regularly observed foraging within 1 m of the ground and all the way to the top of the canopy. Many studies have noted the importance of the vertical dimension. For example, MacArthur's landmark 1958 paper found important vertical niche differentiation among warblers residing in coniferous forests. 

Until recently methods to simultaneously quantify space use in 3D were not available, despite the fact that most animals use space in 3D (e.g., flying, arboreal, burrowing, and aquatic animals). To help rectify this problem I used and refined available methods to quantify redstart space use in 3D for the first time.

I found that using 2D methods results in overestimation of spatial overlap between neighboring pairs of redstarts, when compared to 3D methods. This suggests that redstarts may partition their territories in all 3 dimensions. I also adapted several 2D indices of overlap for use with 3D data. Overall, I hope that demonstrating the creating of 3D utilization distributions will allow future researchers to ask new and innovative questions that could previously not have been asked. This paper will be published in the Auk in the October 2014 issue. I will upload a reprint as soon as possible.


Beginning of dry season (January) in Jamaican scrub forest. The photo below was taken from the same spot (note the pink flag in both trees).

Near the end of dry season (March) in Jamaican scrub forest. The photo above was taken from the same spot (note the pink flag in both trees).

Experimental reduction of winter food decreases body condition and delays migration in a long-distance migratory bird

Many tropical habitats experience pronounced dry seasons (see images to right), during which arthropod food availability declines. This potentially limits migratory bird populations just before they embark on one the most energetically demanding phases of their annual cycle; spring migration. In response to such declines in food, individuals may attempt to alter their space use to enhance access to food resources, but ultimately may be constrained from doing so by their neighbors. If these social constraints exist, food declines should then result in decreased body condition and delayed migration.

To determine if winter food availability affects space use, inter- and intra-specific competition, body composition (i.e., mass, fat, and pectoral muscle), and departure date, we experimentally reduced food available to wintering redstarts in high-quality mangrove habitat using an insecticide.

The food reduction was ultimately successful as food availability decreased by ~80%. This closely mimics natural declines in nearby poor-quality scrub habitats. While redstarts did not expand their territories, some did abandon territoriality and became floaters. Regardless of territorial status, food-reduced redstarts all deposited fat and lost pectoral muscle. Subsequently, food-reduced redstarts experienced on average a one-week delay in departure on spring migration, likely due to the loss of pectoral muscle. We know from previous work with redstarts that even a one-day delay in arrival on the breeding grounds can result in a significantly decreased chance of successfully reproducing. This experiment demonstrated for the first time that fluctuations in winter food availability can lead to fitness costs for migratory birds, and that the mechanism likely involves a fat-muscle tradeoff. This is concerning, especially for the Caribbean, given the long-term drying trend predicted in response to global climate change. This research is currently in review at Ecology.