Past and Current Research

I use a variety of field- and laboratory-based approaches and focus primarily on the ecology and conservation of vertebrate populations. My past and current research has concentrated on the conservation of herpetofaunal (amphibians and reptiles) and avian (birds) species within the theme of global change ecology (i.e., disturbance, pathogens, climate change).

Please read further for a quick synopsis of my past and current research.

Vulnerability of Priority Amphibian and Reptile Conservation Areas (PARCAs) to Climate Change in the Northeastern United States

Collaborators: Clemson University (Dr. Kyle Barrett), Association of Fish and Wildlife Agencies (Priya Nanjappa), University of Maine (Allison Moody and Cynthia Loftin), and Maine Department of Inland Fisheries and Wildlife (Phillip DeMaynadier)


Projected Climatic Niche for the Green Salamander (Aneides aeneus) based on the CCSM4 and HadGEM global climate models (RCP 8.5). Values of 0 – 6 indicate areas of consecutively higher agreement of climatically suitable areas.

Global climate change represents one of the largest environmental challenges of our generation. Overwhelming evidence in the form of shifting species ranges and earlier breeding dates illustrate a small portion of global climate change impacts. Despite misinformation from local media sources and political groups, climate change remains a pervasive threat to all aspects of our surroundings including agriculture, native vegetation communities, and wildlife.

The complex interactions among atmospheric, oceanic, and geophysical processes create an immense challenge as it relates to predicting future climate patterns. Numerous climate labs stationed in different areas of the globe take on the Herculean task of modeling climatic processes to produce Global Circulation Models (GCMs) via a series of complex algorithms. Each lab is associated with a particular GCM (e.g., the Hadley GCM is managed by the Hadley Centre for Climate Prediction and Research in the United Kingdom). The large uncertainties associated with modeling global climate patterns warrant the need for multiple GCMs, and the eventual goal is to use these GCMs to predict future precipitation and temperature changes based on the additional impacts of anthropogenic disturbances. Representative Concentration Pathways (RCPs) provide this anthropogenic context and are essentially greenhouse gas trajectories that represent the concentration (PPM) dispersed in the atmosphere. Currently, four potential RCPs (i.e., 2.6, 4.5, 6.0, and 8.5) are recognized by the International Panel on Climate Change to represent potential greenhouse gas concentration trajectories. It is important to note that each RCP represents a realistic scenario to account for the uncertainty involved with future greenhouse gas emissions and provides a standardized method to provide context to researchers and policy makers.

Not all landscapes experience the effects of climate change equally.  Vulnerability, in the context of landscape ecology, is a relative measure of the susceptibility of a landscape to a given stressor. Landscape vulnerability is estimated by incorporating exposure (i.e., extent of an environmental stressor in a given landscape), sensitivity (i.e., degree to which landscape persistence may be affected by environmental stressors), and adaptive capacity (i.e., capacity of a landscape to cope with environmental stressors). Therefore, identification of species-rich and climate-resilient landscapes is of primary conservation importance. Currently, multiple collaborators (listed above) along with the North Atlantic Landscape Conservation Cooperative (NALCC) are implementing novel methods to determine high-priority landscapes (e.g., Priority Amphibian and Reptile Conservation Areas [PARCAs]) for amphibian and reptile conservation in the northeastern United States. The PARCA idea is loosely based on the Important Bird Areas (IBA) movement, which is one of the most successful conservation programs worldwide. The PARCA effort, similar to IBAs will focus on landscapes that harbor global, regional, and state priority amphibians and reptiles and/or areas that possess exceptional areas of amphibian and reptile biodiversity. Currently, PARCAs are being designated for landscapes within the NALCC and the South Atlantic Landscape Conservation Cooperative (SALCC).


Green Salamander (Aneides aeneus)


The PARCA designation process will provide data on landscapes that deserve continued or increased conservation efforts. However, we have no indication regarding long-term vulnerability of these landscapes to climate change and other stressors. Our vulnerability assessment will incorporate multiple spatially-explicit metrics, including projected temperature change, projected landuse change, priority amphibian and reptile species sensitivity, geographic context, patch size, and topographic relief in a GIS-based framework to assess climate change vulnerability of selected PARCAs. Our efforts provide a science-based structure to assess the long-term vulnerability of these habitats to climate change, which will aid in the allocation of conservation efforts to priority landscapes based on projected climate resiliency.


Response of Amphibians and Reptiles to Forest Management in Pine-Hardwood Forests of the Southeastern US 

Collaborators: Alabama A&M University (Yong Wang), USFS Southern Research Station (Callie J. Schweitzer), William B. Bankhead National Forest

Forest management has often been viewed negatively due to the large-scale and immediate  change to the landscape. In addition, forest management has largely been labeled as an extractive resource, driven primarily by financial gains. However, forest management can also be used to benefit the landscape through a restoration capacity. In many regions of the United States, forests have not been managed in  accordance with historical disturbance regimes. In these situations, forest management can be used in a synergistic capacity to stimulate  local economies and restore historical habitats and faunal  assemblages. However, the observed impacts of forest management  to the larger ecosystem can not be realized unless spatially-replicated manipulations are implemented across the landscape.

Picture 190

Recently thinned forest stand in the William B. Bankhead National Forest

My doctoral research formed one portion of a multi-disciplinary effort  to examine the response of the larger ecosystem to hardwood forest restoration in the William B. Bankhead National Forest (BNF). The BNF is a 72,00 ha, multi-use National Forest located in Lawrence, Winston, and Franklin counties in northwestern Alabama. In many regions of the southeast, including the BNF, Loblolly Pine (Pinus taeda) has been used to reforest lands previously used for agriculture. Although Loblolly Pine is native to the southeastern United States, it exists as a minor component in upland, forested areas of the Southern Cumberland Plateau.  In areas where Loblolly Pine was used as the dominant stocking tree, large outbreaks of the Southern Pine Beetle (Dendroctonus frontalis) have occurred throughout the region. Southern Pine Beetles, which are a native forest pest in the southeastern United States,  specialize on southern yellow pine species (e.g., Loblolly, Slash, Shortleaf Pines) and can result in large losses of standing timber. One of the obvious effects of Southern Pine Beetle outbreaks is the financial loss of quality timber, but an additional effect is the increase of standing and fallen timber, which may increase the risk of damaging wildfires. In the BNF alone, Southern Pine Beetle outbreaks have resulted in the loss of nearly 18,600 acres of Loblolly Pine.  In response to these timber losses, the BNF implemented a forest restoration plan (FRP) to restore pine-dominated forest stands to historical, upland hardwood communities. The FRP included using forest management in the form of prescribed burning and thinning to reduce the risk of Southern Pine Beetle outbreaks and aid in the restoration of hardwood-dominated forest communities.


Copperhead (Agkistrodon contortrix)

My research focused on evaluating the response of amphibians and  reptiles to prescribed burning and thinning. I used a variety of sampling  techniques including drift-fence trapping arrays, artificial cover objects,  radio-telemetry of Copperhead (Agkistrodon contotrix) snakes, and artificial breeding pools to evaluate the response of  the overall herpetofaunal community to forest management. We evaluated the response of herpetofauna over a four-year period and found a wide-range of species-specific responses to management practices. Overall, we found that lizards responded immediately to forest management with the responses tied directly to habitat conditions and thermal conditions of harvested stands (Sutton et al. 2013, 2014). We also found that three years is not an adequate amount of time to discern the impacts of forest management. Studies completed over a much longer time-scale (e.g., 10 years or longer) are necessary to better understand patterns in population changes due to a given landscape disturbance.

Tracking Long-term Changes in Plethodontid Salamander Communities

Collaborators: Marshall University (Thomas K. Pauley), Virginia Highlands Community College (Kevin Hamed), and the University of Tennessee (Matthew J. Gray)

Given the mounting evidence in support of global decline of amphibians, it is important to understand the potential impacts that changing environments have on amphibian populations. However, not all population declines occur on short-enough of a time-scale to be detected. This is particularly troublesome for species that do not vocalize or may be difficult to monitor effectively. One particular group of amphibians, the lungless salamanders (Family Plethodontidae), exemplify these difficulties with effective and accurate monitoring. These salamanders represent the most speciose family worldwide, and are found in highest abundance and species richness in forested environments of the eastern and western United States and Central America. Plethodontids are also found in parts of Europe and Asia, but with much lower species richness.  In fact, a recent discovery of a plethodontid salamander in Korea (Karsenia koreana) represents the only known plethodontid salamander species known to occur in Asia.

One of the primary difficulties with sampling plethodontid salamanders (along with many other wildlife species) is deriving accurate estimates of species detection. The term “detection” refers to the probability that a given individual or species is detected during a survey, given that the individual or species is present at the survey location. Many factors can influence species detection including weather conditions, seasonal effects, search experience, and even “abundance” of a given species (i.e., a more abundant species is often easier to detect compared to a rare species).  Given these conditions, it is often difficult to understand population trends over a long time-scale without considering potential factors that may affect the probability of whether a species or individual will be detected.

I am currently working with multiple researchers to use historically-collected plethodontid salamander abundance data to draw conclusions on current population trends for target salamander species. Pairing historical data with data collected from current conditions requires accurate record-keeping in terms of survey site locations, survey methods (e.g., transects vs. quadrats), and survey conditions (e.g., weather and seasonal conditions). Even factors such as researcher search experience can greatly influence perceived trends. For example, let’s say the researcher that completed the original surveys had nearly 2o years of salamander survey experience when these surveys were completed.  Now fast forward 30 years and say a graduate student comes along and wishes to re-survey these sites, but has relatively little experience surveying for salamanders. The original surveyor found that a given species was present at 87% (52 out of 60 sites) of all sites surveyed (based on a naiive occupancy estimate), whereas the graduate student found that occupancy for this species had decreased to 42% (25 out of 60 sites) across the same survey sites. Realize these estimates are based on a naiive estimate of occupancy (simply the percent of sites where a species was found compared to where it was not found). Although our graduate student found a large decrease in occupancy over the 30-year period, we have no idea whether this represents a true decline or is due to errors associated with our graduate student’s lack of survey experience. If we are able to examine how successful the two researchers are able to detect a focal species (i.e., detection probability) over multiple survey periods, we can use this detection probability to essentially “correct” the number of sites where “true” absences actually occurred. In terms of detection probabilities, let’s say the original researcher had an 90% chance of detecting a species given it was present, whereas the graduate student had a 45% chance of detecting the same species.  From these estimates alone we can make the assumption that sites where our experienced researcher did not find the focal species are more likely to be true absences compared to the absences observed by the graduate student. Now, if we take the original number of sites where a species was detected (52 for the skilled researcher and 25 for the graduate student) and multiply these values by the researcher detection probabilities (0.90 for the skilled researcher and 0.45 for the graduate student), we find that the number of sites found occupied increase much more for the graduate student (~56 sites) compared to the skilled researcher (corrected estimate: ~58 sites). If we use these detection-corrected estimates of occupied sites, we can derive a more accurate assessment of true occupancy (skilled researcher ~58/60 = 0.97 vs. graduate student ~56/60= 0.93).  So, we find that the original estimates produced by our graduate student were greatly influenced by his/her lack of detecting the species at occupied sites. Although the aPicture 019bove scenario represents a simplified example, it is meant solely to illustrate one of the many difficulties associated with pairing historical and current data to draw conclusions regarding the status of a given population. Another caveat with this example is the assumption that a reduction in salamander abundance does not influence our ability to detect our target salamander species. As one may imagine, this statement is a bit dubious given that an abundant species will be easier to detect compared to a less abundant species. The take home message is that many factors can affect the perceived trends in a target wildlife population. Consideration of the myriad factors impacting detection is essential to accurately detect population trends in long-term data sets.

Population Status and Richness of Wild Turtle Populations at the Tennessee State University Campus Wetland

I admit that I am very fortunate to have a job that allows me to work daily with the conservation of native wildlife. When I started at TSU, I was informed that we had a sizable wetland on our campus farm. Needless to say, I was pretty excited when I found this out. How better to teach students about wetland function and wetland-dependent wildlife conservation than hands-on experience?  Our campus wetland is a tremendous asset for students to participate in wetlands research and the TSU Wildlife Ecology Lab exploits this resource as much as possible.

Before I go further, I’d like to sprinkle in a little bit of history regarding our wetland. The current location of our campus wetland used to be an actively-farmed cornfield. Well, introduce the North American Beaver (Castor canadensis) and let them do what they do best. Truth be told, it is highly likely that the current location of our wetland was also historically a wetland that was drained for agricultural purposes, which is an all too common fate for many of our wetlands globally. Fortunately, Beaver populations have rebounded from the early 1900’s when many populations were trapped to the point of local extirpation. Aerial imagery throughout the years 2000 – 2005 revealed that Beavers were building active dams at the TSU wetland. By the year 2010, we had a wetland basin that is relatively comparable to what we have today.

nicole turtle

Nicole Witzel (Master’s candidate) with a fairly unagreeable snapping turle

Starting in spring 2016, the Wildlife Ecology Lab has completed weekly turtle trapping sessions to study the population status and species richness of turtle populations at the campus wetland. We bait hoop traps with a variety of baits, including mackerel, tuna, and dog food in attempts to sample the entire suite of turtles that inhabit the wetland. Each turtle capture is identified to species, measured, weighed, assessed for disease and/or injuries, and given an individual-specific shell notch and tagged with a Passive Integrative transponder (PIT) tag so we can identify individuals if and when they are captured later. Our goal is to continue weekly sampling throughout the summer and into late October. We will use the mark-recapture data to produce population estimates of turtles inhabiting the wetland. We also use our trapping opportunities to teach students about the value of wetlands and wetland-wildlife conservation.


Jeronimo Silva helping with wildlife outreach at TSU’s earth day celebration

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s