I examined small mammal demography and habitat associations in 4 forest cover types on Quantico Marine Corps Base, VA. Study sites included clearcut, shelterwood, mature-riparian, and mature-upland habitats. My primary objective was to determine whether the abundance, species composition and demographic characteristics (density, survival, reproductive effort) of small mammals varied with respect to forest cover type. Secondarily, I was interested in identifying patterns of small mammal habitat selection and the factors that influence those patterns at micro- and macroscales. Small mammals were captured from May 1997-January 1999 on 10 sites (2 clearcut, 4 shelterwood, 2 riparian, and 2 mature) using a combination of Sherman live-traps and pitfall traps. Small mammal abundance and demographic characteristics were examined across forest cover types using a combination of statistical analyses, including Chi-square tests, Kruskal-Wallis tests and repeated measures ANOVA. I surveyed microhabitat features at individual trap stations (n=1000) using variable sized plots and the point quarter method and used these data to determine macrohabitat characteristics for sites (n=10). I examined species-habitat relationships at micro- and macro-spatial scales using Kruskal-Wallis tests, Wilcoxon Rank Sum tests, simple linear regression, stepwise multiple regression and stepwise logistic regression. Fourteen species of small mammals were captured over 7 trapping occasions. Five species including white-footed mice (Peromyscus leucopus), northern short-tailed shrews (Blarina brevicauda), eastern chipmunks (Tamias striatus), meadow voles (Microtus pennsylvanicus), and woodland voles (Microtus pinetorum), accounted for approximately 95% of the total number of individuals captured. Overall relative abundance and species composition of small mammals differed significantly across forest cover types. Catch per unit effort was greatest in shelterwoods followed by riparian, clearcut and mature forest cover types. I found significant differences in the abundances of white-footed mice and eastern chipmunks across forest cover types but not in those of northern short-tailed shrews, meadow voles or woodland voles. Shelterwood stands provided the highest quality habitat for white-footed mice and eastern chipmunks while clearcut stands provided high quality habitat for northern short-tailed shrews. Assessments of habitat quality for other species were based on weak evidence or inconclusive. Microhabitat preferences for the 5 small mammal species with > 100 captures were generally consistent with those previously reported in the literature. White-footed mice, eastern chipmunks and northern short-tailed shrews were associated with woody debris and brushy microsites, while meadow voles were associated with grassy vegetation and woodland voles with the presence of soft mast and woody stem densities. For each of these species, microhabitat was able to predict presence at individual trapping stations at a level better than expected by chance. For white-footed mice and woodland voles, however, habitat selection was found to be dependent upon macrohabitat, suggesting that habitat selection for these species is dynamic. Macrohabitat features were related to the abundance of several small mammal species. In general, the habitat characteristics important to individual species at microscales tended to be important at macroscales as well. The results of this study suggest that current even-aged forest management practices on Quantico Marine Corps Base are compatible with the maintenance of native populations of small mammals. Disturbances created by harvesting, at least temporarily, resulted in favorable microhabitat conditions for a variety of small mammal species. Species such as southern flying squirrels, however, were sensitive to disturbance, although it is likely that successional changes allow rapid recolonization of disturbed sites. / Master of Science

The Clinch Dace Chrosomus sp. cf. saylori is a species of minnow known from only two counties in Virginia. Prior surveys established the species' presence in just eight tributaries to the upper Clinch River. A management plan, which should include both population monitoring and habitat restoration, is still lacking for the species. Population monitoring must balance high detection probability with low risk of injury to captured individuals. I sampled 98 sites in 2014 and 2015 within the putative range of Clinch Dace to refine estimates of its distribution. I sampled 70 of the 98 sites with multiple gears and replication in an occupancy modeling framework. Clinch Dace occupied low-gradient headwater streams with relatively low conductivity in forested watersheds. My surveys uncovered two new tributaries occupied by Clinch Dace, and I was unable to find Clinch Dace in two historically occupied streams. Species detection probability was higher with backpack electrofishing than minnow trapping. N-mixture models suggest that Clinch Dace are more abundant in watersheds with high forest cover although forest cover is highly correlated spatially in the nested stream network. Density estimates from mark-recapture sampling suggest that Clinch Dace occur at low densities in approximately 31.5 km of headwater streams. The mean estimate of global population size was 6,706 individuals. Some populations could be affected by low genetic diversity. I conclude by developing a prioritization framework for restoration and protection of 15 candidate conservation areas. Managers should work with private landowners to implement best management practices in high priority watersheds. / Master of Science

Doctor of Philosophy / Division of Biology / Keith B. Gido / This dissertation consists of three research-based chapters which focus on habitat association of prairie stream fishes and how these fish communities influence stream ecosystem properties. Chapter one introduces important concepts used throughout the chapters, and describes my study streams. In chapter two, I identify local habitat factors associated with the diversity and density of fishes in two protected prairie watersheds. Specifically, the relative importance of habitat factors associated with fish communities were evaluated along a stream-size gradient and across multiple seasons and years. I found that species richness was positively associated with pool area and discharge. Redundancy analyses showed common prairie fish species exhibit ontogenetic habitat associations, with adults in deep and juveniles in shallow pools. Chapter 3 addresses how fish species richness in small prairie streams affects whole-stream metabolism and biomass distribution of benthic organic matter, algal and macroinvertebrates. This study was conducted by stocking experimental stream mesocosms that included pool-riffle habitats with three different communities that represent a gradient of species richness of headwater prairie streams from one to three common prairie stream fish species. I illustrated how species influence ecosystems across multiple spatial scales and found that different communities altered the distribution of algal biomass from benthic surfaces to floating mats and from pools to riffles. The objective of the fourth chapter was to quantify how two size classes of herbivorous prairie stream fish species, central stoneroller Campostoma anamolum and southern redbelly dace Chrosomus erythrogaster differentially affect stream ecosystem properties. This study was also conducted in experimental stream mesocosms, where each unit consisted of one riffle and one pool. Using ANOVAs, I found large dace were associated with longer filaments (F = 7.5, P = 0.002, df = 4) and small fishes with less benthic organic matter (F = 4.2, P = 0.02, df = 4). There was no evidence for ontogenetic shifts in diet and likely differences in energetic requirements and behavior drove the differences among treatments. My research finds that small-bodied prairie stream fishes have predictable habitat preferences and effects on stream properties are dependent on species identity, richness and size structure.

Habitat associations of sea otters during resting and feeding were investigated in Simpson Bay, Prince William Sound, Alaska during the summer months of 2001-2003. Sea otter locations collected during boat surveys were overlaid on bathymetry and sediment maps and water depth, sediment type, distance from shore, and position in the bay (peripheral vs. central) was determined for each. Logistic regression analysis was used to determine whether sea otter habitat use was non-random according to any of these habitat variables. Water depth was the most significant habitat association for feeding behavior, with the majority of feeding dives occurring in shallow water less than 20m deep. Position in the bay was the most significant habitat association for resting behavior, with more otters resting in the center of the bay. In addition, digital images taken of the sea otters during the boat surveys of 2002 and 2003 were used to examine the potential of using nose scars to photo-identify individual sea otters. Both male and female sea otters bore nose scars. Forty-five percent of all individuals encountered were considered identifiable from nose scars and a total of 114 individuals were identified. This compares favorably with the results of photo-identification studies of other marine mammals, suggesting that photo-identification may be a useful tool for the individual identification of sea otters as well.

Grassland birds are the most rapidly declining bird guild in North America, due in large part to extensive loss and fragmentation of grassland habitat resulting from the spread of agriculture and other human-dominated landscapes. Over the past several decades, grassland birds have increasingly become a guild of high conservation and management interest as their populations continue to decline and suitable grassland habitat becomes continually scarcer. Although studies investigating grassland bird responses to management actions and habitat structure are common, few studies are concerned with clarifying the mechanisms through which habitat structure may affect grassland birds. Filling this knowledge gap is critical for increasing our understanding of grassland bird ecology and improving the effectiveness of management and restoration actions for grassland birds. To address this knowledge gap, I took a uniquely holistic approach to traditional grassland bird-habitat studies by concurrently gathering data on multiple wildlife taxa that may interact with birds to examine how these different taxa respond to habitat characteristics across multiple spatial scales, and how those responses may in turn impact grassland birds. Research was conducted on 10 grassland sites at Burning Star State Wildlife Management Area in northeast Jackson County, Illinois, during the bird breeding season (May-July) from 2018 to 2021. My first five objectives were to determine the grassland habitat characteristics that were most highly associated with the following taxa: arthropods, an important food source for adult and nestling grassland birds; small mammals, which are occasional nest predators and an alternate prey source for more prolific nest predators; snakes; raccoons; and grassland birds. My final objective was to estimate the extent to which grassland bird habitat associations were mediated through nest predator abundance, alternate prey abundance, and food availability. In Chapter 2, I examined associations between grassland arthropod communities and habitat characteristics representing three spatial scales: local (within-patch vegetation structure and composition), patch (size, shape, edge composition), and landscape (landcover composition within a 400 m buffer). In addition to their relevance for grassland birds, arthropods play important functional roles in grasslands and are useful indicators of grassland health. I collected arthropod samples using pan traps in grassland patches at Burning Star, and used generalized linear mixed models to relate variation in arthropod biomass and diversity to habitat predictor variables. I found that arthropod biomass increased with vegetation height at the local scale and proportion of forest/shrub edge at the patch scale, while arthropod diversity responded only to local-scale variables, including a negative association with vegetation height and woody vegetation cover, and a positive association with forb cover. I conclude that local vegetation structure and composition are the main drivers shaping arthropod communities at Burning Star, and that limiting woody encroachment and increasing forb cover and variation in vegetation height within grassland patches may encourage arthropod abundance and diversity in tallgrass prairies. In Chapter 3, I estimated associations between small mammal abundance and habitat variables, again representing three spatial scales. I surveyed small mammal communities using a grid of 100 Sherman traps set out for three nights at each study site. I identified all captured individuals to genus, individually marked them with ear tags, and estimated abundance using a combination of Huggins P and C models in Program Mark and generalized linear mixed models in Program R. I found that small mammal abundance was positively related to vegetation density and negatively related to plant diversity, though variation in plant diversity affected Microtus voles more strongly than Peromyscus mice. At the landscape scale, small mammal abundance was positively associated with the amount of water surrounding a patch, and negatively associated with the amount of grassland surrounding a patch. Variation in small mammal community composition (proportions of Microtus vs Peromyscus) was mostly governed by differences in habitat structure at the landscape scale, rather than differences in vegetation structure at the local scale. I suggest that managers interested in influencing small mammal abundance in grasslands encourage dense vegetation growth by limiting disturbance if increased small mammal abundance is desired, or reduce vegetation density by increasing disturbance frequency to reduce small mammal abundance. Additionally, increasing plant diversity by sowing a high diversity of seeds may be an effective way to control Microtus vole populations. In Chapter 4, I estimated the habitat associations of snakes at Burning Star, focusing on the relative abundance of snakes among different grassland sites, as well as snake diversity and species-specific occupancy. Although snakes are prolific nest predators of grassland birds, they are also integral components of grassland systems, and there may be instances where managers and decision-makers wish to increase, rather than decrease, their abundance in grasslands. I found that snake community metrics were strongly and positively related to an increase in woody plant cover at the local (within-patch) scale. Snake relative abundance was also positively related to an increase in grass cover and a decrease in forb cover, though my occupancy results suggest that this was primarily driven by an increase in black kingsnakes (Lampropeltis nigra). At the patch scale, snake relative abundance and diversity were both positively related to the proportion of patch edge composed of roads. Habitat structure at the landscape scale had the smallest impact on snakes in this study, though the proportion of trees in the landscape was positively related to snake diversity. I suggested that managers and conservationists interested in manipulating snake abundance in grasslands focus on within-patch vegetation structure and composition. Decreasing woody cover in grasslands, or increasing the ratio of forbs to grasses, may reduce the presence of snakes, while maintaining a woody component could encourage both snake abundance and diversity. In Chapter 5, I estimated the habitat characteristics that were most strongly associated with raccoon abundance estimates in grassland patches at Burning Star. Raccoons have become increasingly important avian nest predators in midwestern grasslands due to rampant habitat fragmentation. I estimated raccoon abundance using an occurrence index from a series of baited trail cameras located in grassland sites. I found no convincing evidence of raccoon abundance being influenced by local scale habitat structure, beyond a weak association with vegetation height. At the patch scale, raccoon abundance was positively related to the proportion of patch edge composed of roads. At the landscape scale, raccoon abundance was negatively related to grassland, and positively related to water, within 400 meters of a grassland patch. I recommended that managers concerned with minimizing the presence of raccoons in grasslands should limit roads along grassland perimeters, maximize the proportion of grassland in the landscape surrounding a grassland patch, and avoid planning a grassland restoration in close proximity to open water if possible. In Chapter 6, I estimated the habitat characteristics that were most strongly associated with daily nest survival, nest density, and abundance of Dickcissels (Spiza americana), Field Sparrows (Spizella pusilla), and Common Yellowthroats (Geothlypis trichas), but also considered responses of all grassland bird species combined. I considered habitat characteristics representing four spatial scales: nest site, within-patch, patch, and landscape, though the nest site scale was only considered for nest survival analyses. I found that Dickcissels, an obligate grassland species, exhibited the strongest response to fire, as nest density drastically improved after previously undisturbed grasslands were burned. Dickcissel abundance was positively related to agriculture at the landscape scale and negatively related to woody cover. Field sparrows demonstrated a preference for woody cover and proximity to forests and shrublands, and Common Yellowthroats were positively associated with forb cover. Both Field Sparrow and Common Yellowthroat nest survival increased with greater distance from an edge, though no edge effect was detected for Dickcissel nest survival. All bird species benefitted from increased plant diversity and greater patch size. All species also responded negatively to vegetation height or litter depth, suggesting that fire, which reduces vegetation biomass and litter, may indirectly benefit the facultative grassland birds of Burning Star in addition to Dickcissels. In Chapter 7, I used structural equation models and data gathered in the previous five chapters to estimate whether the effects of habitat structure on breeding Field Sparrows is mediated through changes in predator (snake and raccoon) abundance, alternate prey availability, or arthropod biomass. I used Field Sparrows as the focal species for this chapter because they were the most common grassland bird in my dataset. I found no evidence of nest survival or nest density of Field Sparrows being directly influenced by nest predator abundance, alternate prey, or arthropod biomass, although habitat characteristics associated with increased nest survival were also associated with greater arthropod biomass and reduced predator abundance. I suggested that habitat structure at Burning Star may primarily impact breeding Field Sparrows through direct means, such as influencing nest concealment or foraging efficiency. These results also suggest that nest success and nest density are decoupled in this study area, so Field Sparrows may be preferentially selecting nest sites with structural characteristics that do not increase nest survival. Ultimately, my findings from this study indicate that while predator avoidance and food provisioning likely play an important role in determining nest survival for grassland birds, predator abundance and arthropod biomass may not necessarily predict predation risk and foraging efficiency to the extent that is often assumed.

Due to the global loss of tidal marsh area, potential anthropogenic and natural disturbances to these systems, and coastal marshes’ affinity for trapping environmental pollutants, understanding how marsh birds inhabiting these ecosystems will adapt to these changes is paramount. To quantify future changes, I first needed to have something to compare against – baseline estimates. To this end, I performed distance sampling line transect surveys during the nonbreeding season to estimate species-specific population abundance, density, and habitat associations and captured two species of marsh bird, Clapper Rail (Rallus crepitans) and Seaside Sparrow (Ammospiza maritima), to estimate baseline ingestion of one such environmental pollutant – microplastics. By providing the first baseline population and density estimates for numerous nonbreeding marsh birds, my findings suggest that the tidal marshes of Mississippi provide critical habitat for many of these species. Additionally, I was able to document the first evidence of microplastic ingestion by resident tidal marsh birds.

Many amphibian populations are rapidly disappearing throughout the world. An important issue for ecologists is why some amphibian species are more susceptible to decline than others. Here I present five experiments that compare the performance of an endangered (Rana sevosa) and a common (Rana sphenocephala) frog in changing habitats, to determine why these two species differ in their persistence. I include additional studies investigating the habitat requirements and behavior of R. sevosa. I found that habitat change in the form of canopy closure over breeding ponds negatively affects both species, making them smaller as tadpoles and at metamorphosis. The magnitude of size differences was greater for R. sevosa and this species was less likely to survive in closed canopy ponds. Larval survival was not affected in R. sphenocephala and this is likely a key reason for the persistence of this species in habitats where R. sevosa has been extirpated. The introduction of fish to breeding ponds would also differentially affect the two species. R. sevosa did not display behavioral defenses to the threat of fish predation while R. sphenocephala did. R. sevosa displayed a preference for certain characteristics in its habitat including open canopy ponds, grassy terrestrial habitats and an abundance of burrows. An association with the chemical cues of burrow making organisms suggests that these organisms may be important for R. sevosa. Therefore, specialized habitat requirements and behaviors which may be contributing to its decline. I found some evidence of asymmetric competition in the larval stage with R. sevosa negatively affecting R. sphenocephala. It does not appear that larval competition with R. sphenocephala has contributed to the decline of R. sevosa. The primary difference between the two species was in responses to habitat change. R. sevosa appears to be rigid in its habitat requirements and behavior and its inability to respond and adapt to change is a key component of its rarity. By contrast R. sphenocephala showed an ability to cope with habitat changes. Conservation of R. sevosa will require suitable management of the aquatic and terrestrial habitats, primarily through the instigation of an appropriate fire regime.

Apparent long-term declines of white-winged and surf scoter (<i>Melanitta fusca </i> and <i>M. perspicillata</i>) populations in the northern boreal forest have raised concern for these sea duck species. Reasons for population declines are not well understood but some evidence suggests that factors associated with events on the breeding grounds may be responsible. Breeding ground changes could adversely affect abiotic or biotic characteristics of upland or wetland habitats or key food sources for breeding females or ducklings, which in turn may lower productivity or recruitment. Like most boreal-nesting ducks, virtually nothing is known about wetland habitat preferences of scoters. Determining habitat features that scoters need to breed successfully, and how habitat changes in the boreal forest affect scoters, is an important step in understanding their ecology and developing conservation initiatives. Thus, my overall goal was to look for evidence of habitat selection in scoters at two spatial scales by characterizing biotic and abiotic features of areas used by scoter pairs and broods, and comparing these features with those of areas not used by scoters. Habitat characteristics and scoter use of wetlands in recently burned forest was also contrasted with unburned forest to determine whether habitat change caused by fire could affect patterns of habitat use by scoters.<p> I used remote sensing data as a tool to delineate coarse-scale patterns of habitat use by scoter pairs and broods. Results indicate that although scoters may not settle on wetlands in areas dominated by burned vegetation two years following the fire, three years after the fire I found no difference in scoter pair or brood use between wetlands in burned and unburned upland. I found that surf and white-winged scoter pairs often co-occurred on wetlands. I was unable to find any evidence to support the prediction that scoters prefer wetlands with irregular shorelines that might enhance pair isolation and offer greater protection to ducklings from severe winds and wave action. <p> Based on fine-scale wetland habitat characteristics, scoter pairs and broods used wetlands with more abundant food, a finding that is consistent with many other waterfowl studies. However, unlike some previous waterfowl studies, I did not find a consistent correlation between total phosphorus levels and amphipod abundance or wetland use by scoters. Very high total nitrogen to total phosphorus ratios in sampled wetlands lead me to speculate that wetlands in my study area may be phosphorus limited. I did not detect a difference in fine-scale features of wetlands surrounded by burned versus unburned vegetation. This study of scoters in the northern boreal forest was among the first to determine why scoters use specific wetlands or areas and not others.

Apparent long-term declines of white-winged and surf scoter (<i>Melanitta fusca </i> and <i>M. perspicillata</i>) populations in the northern boreal forest have raised concern for these sea duck species. Reasons for population declines are not well understood but some evidence suggests that factors associated with events on the breeding grounds may be responsible. Breeding ground changes could adversely affect abiotic or biotic characteristics of upland or wetland habitats or key food sources for breeding females or ducklings, which in turn may lower productivity or recruitment. Like most boreal-nesting ducks, virtually nothing is known about wetland habitat preferences of scoters. Determining habitat features that scoters need to breed successfully, and how habitat changes in the boreal forest affect scoters, is an important step in understanding their ecology and developing conservation initiatives. Thus, my overall goal was to look for evidence of habitat selection in scoters at two spatial scales by characterizing biotic and abiotic features of areas used by scoter pairs and broods, and comparing these features with those of areas not used by scoters. Habitat characteristics and scoter use of wetlands in recently burned forest was also contrasted with unburned forest to determine whether habitat change caused by fire could affect patterns of habitat use by scoters.<p> I used remote sensing data as a tool to delineate coarse-scale patterns of habitat use by scoter pairs and broods. Results indicate that although scoters may not settle on wetlands in areas dominated by burned vegetation two years following the fire, three years after the fire I found no difference in scoter pair or brood use between wetlands in burned and unburned upland. I found that surf and white-winged scoter pairs often co-occurred on wetlands. I was unable to find any evidence to support the prediction that scoters prefer wetlands with irregular shorelines that might enhance pair isolation and offer greater protection to ducklings from severe winds and wave action. <p> Based on fine-scale wetland habitat characteristics, scoter pairs and broods used wetlands with more abundant food, a finding that is consistent with many other waterfowl studies. However, unlike some previous waterfowl studies, I did not find a consistent correlation between total phosphorus levels and amphipod abundance or wetland use by scoters. Very high total nitrogen to total phosphorus ratios in sampled wetlands lead me to speculate that wetlands in my study area may be phosphorus limited. I did not detect a difference in fine-scale features of wetlands surrounded by burned versus unburned vegetation. This study of scoters in the northern boreal forest was among the first to determine why scoters use specific wetlands or areas and not others.

ABSTRACT Eastern Equine Encephalitis virus (EEEV) is an alphavirus with high pathogenicity in both humans and horses. Florida continues to have the highest occurrence of human cases in the USA, with four fatalities recorded in 2010. Unlike other states, Florida supports year-round EEEV transmission. This research uses Geographic Information Science (GIS) to examine spatial patterns of documented sentinel seroconversions and horse cases in order to understand the relationships between habitat and transmission intensity of EEEV in Florida. Sentinel sites were categorized as enzootic, periodically enzootic, and negative based on the amount of chicken seroconversions to EEEV. Sentinel sites were analyzed based on land classification data d using the Kruskal-Wallis test to determine which habitats were associated with disease transmission. Cluster analyses were performed for the horse cases using density-based spatial clustering of applications with noise (DBSCAN). Ecological associations of EEEV were examined using compositional analysis and Euclidean distance analysis to determine if the proportion or proximity of certain habitats played a role in transmission. The research in these studies provides evidence of ecological associations for EEEV transmission in Florida that hasn't been previously analyzed. Furthermore, these studies provide the groundwork for better understanding of why there is a disproportionate number of horse and human cases of EEEV in Florida than in any other state.