Finding the elusive eastern spotted skunk

Screen Shot 2018-08-22 at 11.28.58 AM

Documentation of the mid-century eastern spotted skunk population decline by Gompper and Hackett (2005).

The eastern spotted skunk is an elusive, potentially rare and endangered species of skunk native to much of the eastern US between the Rockies and the Appalachian Mountains. The species was common throughout its range at the beginning of the twentieth century and people often saw eastern spotted skunks on family farms. During the 1940s and 1950s however, eastern spotted skunk populations crashed. The population decline is well documented, but reasons for the crash remain unclear. Hypotheses for the decline range from the expansion and modernization of agriculture to overharvest to disease. Likely, a combination of several concurrent factors lead to the decline. Eastern spotted skunk populations never recovered, remaining at low levels across much of their historic range.

Today, researchers are working with state wildlife agencies to identify where eastern spotted skunks are and determine which resources they need to maintain healthy populations. In some states, large-scale surveys for eastern spotted skunks resulted in no sightings, suggesting the species is locally extinct in parts of its historic range. Other states have identified populations and are working to understand whether the populations are at a healthy level.

In Arkansas, eastern spotted skunks were historically present across the entire state and recent surveys have revealed the species still has strongholds in the Ouachitas, or the western region of the state. It was with this knowledge the Arkansas Game and Fish Commission funded my research to determine whether eastern spotted skunks are present in the Ozarks, and if so, which resources they’re using. I conducted a large-scale camera trap survey in north-central Arkansas to answer these important questions. Although I recorded eastern spotted skunks at some camera trap sites, preliminary results suggest the species occurs at extremely low population levels in this part of the state.


An eastern spotted skunk visits a camera trap site in north-central Arkansas.

Using the information gathered from my camera survey, I decided to produce a species distribution model. This type of model uses presence-only data to evaluate where a species is most likely to be present based on characteristics of locations where we know eastern spotted skunks spend time. Using presence-only data means that I will only use camera trap locations where eastern spotted skunks were recorded. For example, from approximately 75 camera trap locations, eastern spotted skunks were photographed at only 4 sites. Failure to record an eastern spotted skunk at a camera trap site doesn’t necessarily mean the species is absent at that site; it simply means we don’t know for sure that eastern spotted skunks use that area. Thus, the locations where I recorded eastern spotted skunks on camera traps are “known locations.” I will use the 4 known locations where eastern spotted skunks were confirmed and exclude the remaining 71 camera trap locations for my species distribution model.

In addition to the 4 known locations from my camera trap survey, the eastern spotted skunk species distribution model will use an additional 72 known locations from eastern spotted skunk surveys by other researchers in Arkansas and southern Missouri. I will determine what the environment was like at the known locations, including how close they are to roads and other infrastructure, how close they are to water sources, and how dense the forest is at each location. Using this information, the species distribution model will predict where eastern spotted skunks are most likely to be across all of Arkansas and southern Missouri. For example, if most of the known locations are in areas where the forest is thick and dense, the model will predict that eastern spotted skunks are most likely to be in other thickly forested parts of the state and less likely to be in open fields.

Although the large-scale camera trapping survey I conducted resulted in limited eastern spotted skunk photographs, the species distribution model approach allows me to use these data. The final product will be a heat map of Arkansas and southern Missouri, with warm tones suggestive of eastern spotted skunk populations and cool tones meaning eastern spotted skunks are not likely to occur in those areas. The map will be useful for state wildlife agencies as they continue to determine where the species is and create management plans to prevent further population decline of this unique mammal.


Will you be at The Wildlife Society Annual Meeting in October 2018? Come to my talk on Tuesday, October 9 to see the results of the species distribution model.


These Aren’t The Poops You’re Looking For

Researchers have been using animal scat (read: poop) for decades, and for good reason.  Bits of identifiable food not fully ingested by the animal offer insights into diet, while many parasites pass eggs through feces.  More recently, physiologists have defined methods for extracting stress hormones from fecal samples, providing information on when animals become stressed and whether that stress is chronic.  Collecting scat is a great way to answer basic ecological questions on a given species.  Cool story, right?  Well, we’ve got a bit of a problem.

If you read through old research papers on carnivore diet, you’ll find that in many diet studies, researchers collected scats from live traps.  Say the target species is a grey fox.  The researchers set live traps, capture a grey fox, and discover a scat inside the trap.  It’s pretty clear that scat was made by the grey fox.  Easy identification.

field-guide-scatOne of the major benefits to studies utilizing scat today, however is that they are “non-invasive.”  That means researchers don’t need the animal in hand to conduct the study.  In turn, studies are cheaper and logistically easier.  The data (poops) are out there on the landscape, scientists just have to find them.  As long as they know the animal exists in the study area, they know its scat exists there too.

Until recently, a typical researcher conducting one of these non-invasive studies would carry a field guide on her poop collecting journeys.  When she wandered upon a scat sample, she could observe its shape, size, odor, and any associated tracks to determine who made the poop.  That’s right, a handy dandy field guide could tell her whether she was looking at the digested dinner of a bear, coyote, bobcat, red fox, grey fox, or domestic dog.  Or could it?

A team of researchers at Virginia Tech decided to test how handy that guide really was when it came to assigning appropriate species identifications to scat samples.  In other words, do carnivores leave scat samples different enough that a researcher can tell them apart, or is the researcher simply making an educated guess?  If the latter is true, how much are those guesses altering the study results?

They started by – you guessed it – collecting scats.  Every time they found a sample, they gave a species identification based on field guide descriptions of the most common carnivores found in their study area: the Virginia mountains.  Then, partnering with a lab at the University of Idaho, they used DNA left behind by the predator on the outside of each poop to confirm identifications.  These genetically based identifications were reliable; they identified the true pooper, and it wasn’t always the same as the field guide suggested.

Now the scientists decided to test if those incorrectly identified samples mattered.  Were they altering study results?  To find out, they conducted a diet study.  They looked at bobcat, coyote, and bear diet when the samples were identified using only field guides, then looked at diet for the three carnivores using the true, genetically confirmed identifications of the scat samples.  Details on using scat to discover dietary patterns in carnivores can be found in my Scoop on Poop series.

The researchers found they weren’t too good at assigning correct scat identifications using only field guides.  Coyote scats were only identified correctly in the field 54% of the time and bobcats had a similarly dismal field accuracy rate at only 57.1% of true bobcat samples identified correctly in the field.  Black bear scats, on the other hand, were easier to identify; almost all (95.2%) bear samples were identified correctly in the field, likely because they are much larger in size when compared to coyote and bobcat samples.

Sometimes the researchers incorrectly called bobcat scats coyote scats and vice versa…so what?  They’re probably after the same prey anyways, right?  As it turned out, that “sometimes” really influenced the results of the diet study.  When bobcat scats were misidentified, they were classified as coyote scats 98% of the time.  Similarly, bear scats called something other than bear in the field were called coyote 75% of the time.

Because they were classifying some bobcat scats as coyote in the field, it appeared that coyote diet was similar to bobcat diet (0.95 niche overlap, where 1 means identical diets and 0 means completely different diets).  In contrast, coyotes and bears appeared to have quite different diets (0.5 niche overlap) when using the field identification method.  In reality, bobcats and coyotes were tapping into some of the same prey resources, but not at the same frequencies.  Their true niche overlap, calculated based on those reliable genetic identifications, was 0.73; bears and coyotes actually shared more diet items than it seemed with a true niche overlap of 0.69.  The incorrectly identified scat samples provided a picture of how the carnivores were interacting on the landscape, just not the right one.

Scientists make a living on asking questions, and sometimes that means questioning their own methods.  In this case, it’s a good thing they did!  Moving forward in the realm of scat studies, the authors of the study suggest always corroborating field identifications of scat samples with genetic methods in the lab.  Read the complete study here.  The data are strong with this one.