Brian R. Mitchell, UVM Postdoctoral Research Associate



Coyote Vocal Communication Research at Dye Creek Preserve, California 


My Ph.D. research was primarily supported by the United States Department of Agriculture, Animal and Plant Health Inspection Service, Wildlife Services, National Wildlife Research Center (NWRC).  The main mission of this agency is to conduct research that helps solve conflicts between human interests and wildlife.  The California Field Station of the NWRC was focused on basic research aimed at getting a better understanding of coyote behavior, with the goal of developing more selective and effective methods of reducing coyote depredation of livestock.  Previous graduate students at the University of California, Berkeley had found that coyotes that killed sheep were almost exclusively the dominant territory holders, and they almost always killed sheep within their territory.  My research was aimed at developing a better understanding of the coyote’s long-distance communication system, and to evaluate whether this system could be used for selective coyote control.




Biodiversity in





NPS Northeast Temperate Network

USGS Vermont
Co-op Unit


My research began with a trip to the NWRC’s field station in Logan, Utah, where I recorded captive coyotes.  I needed a library of vocalizations for my later playback experiments, plus I was curious about the potential information content in coyote barks and howls.  For example, if coyote barks and howls are individually distinctive, I would need to use care when constructing playbacks to ensure that I was not broadcasting calls from many coyotes when I was trying to imitate one or two.
By analyzing spectrograms of howls and barks, I was able to determine that both of these vocalizations do indeed contain individually specific information.  Because of the tremendous advantage of being able to determine individual identities, I presume that coyotes use the information in barks to identify individuals they are familiar with.
Another interesting aspect of coyote barks and howls is that howls stably convey information for distances of at least one kilometer.  Barks, on the other hand, rapidly attenuated and did not appear suitable for transmitting information.  Barks likely serve other purposes, such as attracting information and providing information that listeners could use to estimate distance to the barking animal.

Jamie and coyote

I conducted two years of playback experiments on radio collared coyotes at the Dye Creek Preserve, a property managed by The Nature Conservancy in California’s northern Sacramento Valley.  To date I have only analyzed the first year of data (and am taking donations to fund analysis of the second year ;-).  The first year of playbacks used a variety of different stimuli (from a siren to human imitations of coyotes to actual coyote vocalizations) at different times during the course of an entire year.  The second year of playbacks consisted entirely of canid vocalizations broadcasted just before dawn.  During both years, my crew and I tracked coyotes before and after playback to determine how they responded.

Coyotes could respond by vocalizing or by approaching a playback.  Vocal responses were most likely to coyote group vocalizations when there was low wind.  Vocal responses were also more common just before dawn and during periods when the moon was up and bright.  The vocally responding animals were almost exclusively territorial alphas and betas; transients rarely vocalized.

Approach responses were most likely to playbacks of group vocalizations (although human imitations of coyotes generated similar levels of approach responses).  Approaches were most common when playbacks were within the responding animal’s home range, during the first half of the year, and at or before sunrise.  Territorial coyotes were twice as likely to respond as transients.

The results from the first year of playback studies strongly support the use of playbacks to locate or selectively attract territorial coyotes.  Coyote control actions based on the use of coyote vocalizations may be more selective than many existing techniques… but this needs to be tested with rigorous operational experiments.


Any large project generates a number of spin-offs, and mine was no exception.  Other coyote-related projects that I have been working on include a genetic analysis of the relatedness of individuals in my study population, a mark-recapture study using DNA found in scat, and a pilot study with GPS collars investigating how coyotes respond to human presence. 
The relatedness analysis is not yet complete, but preliminary findings indicate some evidence of infidelity among alpha pairs.  In addition, there is evidence of a tendency for coyotes to inherit their parents’ territory or settle in adjacent territories.  This could potentially lead to a situation where extended families benefit from living next to each other through reduced territorial aggression.
The scat DNA study is ongoing; Christen Williams of the NWRC is analyzing the samples, and we will conduct a mark-recapture analysis of the data to determine the size of the Dye Creek population.  Estimating the density of carnivores is particularly difficult, so we hope that scat DNA provides a cost effective tool for tracking carnivore numbers.
The advent of GPS collars small enough to be worn by coyotes provided me with an opportunity to look at how coyotes react to human presence.  Several coyotes were fit with GPS collars that recorded their exact location every 15 minutes.  While these coyotes were collared, my field crew and I recorded our locations as we traveled and worked on the study site; our activities were the majority of human activities on the preserve.  I hope to analyze this data to better understand how coyotes react to research activities.