The Bandon Marsh is a large marsh restoration project located in southwest Oregon. The land has been previously used for dairy farms and much of the marsh was used for cattle grazing. The goal of the restoration is to recreate a natural habitat for shorebirds and fish. The purpose of this thesis is to gather and analyze data on the geology of the marsh and both ground and surface water quality to evaluate its ability to support biology. The US Fish and Wildlife Service began restoration on the Bandon marsh in 2005. Research on the Ni-les'tun Unit began in that same year with Geoprobe coring, Ground Penetrating Radar (GPR), piezometer well installation, and surface water analysis. Twelve cores were taken with the Geoprobe in 2006 and later analyzed for this thesis. A total of 50 GPR lines were run across the marsh ranging from ~25 m to 1,200 m in length for a total of about 9,700 meters (9.7 km). The ten piezometer wells were sampled to collect groundwater pH, oxidation-reduction potential, salinity, temperature, and well water elevation. Twelve channel sampling sites were chosen to be sampled in three field runs taking place in winter 2011, summer 2011, and winter 2012. The summer 2011 and winter 2012 sampling runs were done after tide gates had been removed from the marsh. The Geoprobe core samples were retrieved for extensive lab analysis. Analyses included; bulk density, porosity, permeability, and grain size studies. Lab results revealed a low permeability fine grained upper sedimentary unit ranging from ~0.5 to ~2.5 meters depth with a coarser grained higher permeability lower unit. GPR lines were analyzed for groundwater surface depths and buried channel cut and fill features (Peterson et. al., 2004). GPR results showed a fairly shallow groundwater surface around 0.5 meters in the north down to 2.0 meters in the south. The piezometer wells showed a seasonal variation in well water elevation with higher elevations in the winter and lower in the summer. Well water chemistry showed both seasonal and spatial variations. Values for pH and dissolved oxygen were lower in the north side of the marsh and higher in the south side, and pH was higher in the summer than in the winter. Wells that were more proximal (within ~400 m) to the Coquille River Estuary showed higher conductivity (salinity) values than those that were more distal (greater than ~400 m). Channel sampling results showed similar trends as the groundwater results. The lower pH values tended to be in the north side of the marsh as well as lower dissolved oxygen values. Again, there was a spatial variability in conductivity with the higher values found closer to the Coquille River. Conductivity in the channel water showed a great seasonal variability with the highest values occurring in the summer time and much lower values in the winter. As expected, channel water increased about 10°C on average from winter to summer. The results of all of these observations and analyses are combined to put together a generalized flow model showing the different water inputs into the channel waters on the Ni-les'tun Unit. Once these inputs were derived, conclusions can be made on the quality of the water and its ability to support small aquatic life.

This dissertation presents the results of a study that I undertook to better understand the breeding biology of Eastern Kingbirds (hereafter, kingbirds) at Malheur National Wildlife Refuge in southeastern Oregon from 2003 to 2009. Kingbirds are long-distance migratory songbirds that breed across much of North America. This species is socially monogamous but, via frequent extra-pair copulations, is genetically polygamous. Kingbirds exhibit relatively high breeding site fidelity, often returning to the same tree to nest in subsequent years. Both members of a pair provide parental care, but there are often specific duties performed by both male and female kingbirds. For example, males typically perform much of the vigilant nest defense that this species is well known for and contribute, to an extent, to the feeding of nestlings. Females, on the other hand, are entirely responsible for incubating and brooding. Beginning in 2002, most adults within the population of kingbirds at Malheur had been banded with a unique combination of three colored plastic leg bands and an aluminum USFWS band. Also, as many nests were located in each year as possible, and any young that survived to fledging age were banded with a unique combination of leg bands. Considerable effort was then spent each year to locate as many banded individuals as possible, which allowed me to document adult and juvenile survival. Annual survival rates of adult male and female kingbirds in the population at Malheur did not differ and were relatively high at approximately 0.65. Juvenile survival rate was approximately 0.29, indicating that slightly less than one-third of nestlings survive the interval between fledging and their first potential breeding season. Resighting probability was high for both sexes, although higher for adult males (0.94) than for adult female (0.84). The latter finding is consistent with the higher site fidelity of males than of females. Resighting probabilities for juveniles were much lower at 0.68 and 0.40 for males and females, respectively. Again, this was expected because natal site fidelity is typically much lower than breeding site fidelity. Compared to most other reports, resighting probability and return rates of juvenile kingbirds was high, presumably because the riparian habitats where kingbirds breed at Malheur function as an ecological island surrounded by, for kingbirds, unsuitable high desert habitat. Thus, unless they disperse very long distances, the only option for juvenile kingbirds is to begin nesting on the refuge. The collection of blood samples from birds has become an increasingly common practice in ornithology. The data that can be gained from these samples allow a number of interesting questions to be asked such as understanding the genetic mating system of a species, patterns of hormone secretion, and discerning migratory pathways via stable isotope analysis. The volume of blood collected is usually small and was assumed to cause no long-term negative effects on sampled individuals. However, few studies have rigorously examined the effect of blood sampling on survival. I used a multistate mark-recapture analysis to assess the effect blood sampling on annual survival of kingbirds by combining the annual survival data described above with whether or not individuals were subjected to blood sample collection. The results of this analysis indicated that blood sampling had no effect on annual survival rates of kingbirds. Whether or not this is the case for other species remains to be seen. However, my results support the assumption that when done correctly, blood sampling has little to no long-term negative effects on birds. Comparative analyses of many bird species show that testosterone secretion exhibits fairly predictable patterns among breeding birds of different mating systems. Monogamous species reach a peak during mate attraction and the period of female fertility which is then followed by a sharp decline when young are in the nest. By contrast, males of polygynous species tend to maintain higher levels of testosterone throughout the breeding season to, presumably, maximize opportunities for acquiring additional mates. Kingbirds are socially monogamous but cryptically polygamous. However, because of high rates of extra-pair paternity, variation in reproductive success among males is much higher than what is expected for a monogamous species, and, instead, is similar to what has been reported for polygynous species. Therefore from 2005 to 2009 I measured testosterone concentrations from the blood samples collected from male kingbirds to determine the breeding season profile of testosterone secretion in kingbirds and to understand the factors that influence testosterone variation among individuals. Contrary to expected for a monogamous species, the testosterone profile of kingbirds did not exhibit the brief peak in testosterone followed by a precipitous decline. Testosterone peaked early in the season, but declined very slowly as the nest cycle progressed. I attributed this gradual decline in testosterone to the cryptically polygynous nature of the kingbird extra-pair mating system. I also found substantial variation in testosterone concentration among male kingbirds and was able to identify several factors contributing to this variation. As expected, testosterone declined as the breeding season progressed (independent of stage in nest cycle) and nest density increased, while increases in testosterone were correlated with the number of fertile females within the population. This suggests that male kingbirds were capable of modulating testosterone concentrations to both cope with an increase in conspecific density, but also to ready themselves for times when extra-pair copulations were more likely.

We have been studying the population of Eastern Kingbirds breeding in riparian habitats in Malheur National Wildlife Refuge (eastern Oregon) since 2002. These efforts have created an ideal research environment wherein most adults in the population have been color banded and DNA sampled and, as part of other research projects, nearly all broods since 2003 have been paternity tested. I decided to use behavioral video recordings of parental nest behavior undertaken between 2003 and 2010 for two unrelated projects. First, I tested the effectiveness of video sampling nesting behavior (see below and chapter 2). Second, I tested whether male kingbirds were able to affect their level of paternal investment in accordance with their level or realized paternity (see below and chapter 3). Chapter 2 was split into three distinct questions: 1) are parental nesting behaviors repeatable, 2) is a one hour sample sufficient to capture variability in these behaviors, and 3) is the first hour of recording sufficient to capture variability in these behaviors. I found overwhelming evidence that the behaviors I measured were repeatable. This is truly important, for if repeatability was disproven, it would call into question the use of sampling throughout the field of animal behavior. I similarly found strong evidence that a one hour sample was sufficient to capture variability in parental behaviors. From this I was able to suggest that further sampling effort would be better spent increasing sample size rather than observation length. Testing whether the first hour of recording was sufficient to capture variability in parental behavior found more muddled results. While there generally was correlation between behavioral values in the first hour and those over a longer observation period, most behaviors were found to have significantly lower values in the first hour. I tested whether this was the result of a lingering observer effect or a natural effect of time of day and concluded that an observer effect was the more likely explanation. In chapter 3, I ran one of the more in depth and complicated tests for a relationship between paternal investment and realized paternity that I was able to find in the literature. I used the standard male feeding rate as a measure of male investment as well as a far more nuanced measure derived from the first Eigenvector of an analysis of six different paternal behaviors. These were both tested using Akaike's Information Criterion against a number of variables likely to affect parental behavior, including realized paternity. Ultimately, I found no evidence that kingbirds were able to affect paternal investment in response to lost paternity, and conclude that they likely had no means to assess realized paternity within the nest of their social mate. However, from the other parameters tested I was able to find that males increased their investment as brood size increased and as the female spent more time attending, but not feeding, the nestlings. Also, male investment decreased as territory density increased and flight feather length increased. Theory surrounding the variability of extrapair paternity in birds can be used to account for these results. In a denser population males are predicted to reduce paternal effort for one of two reasons: increased likelihood of paternity loss in his own nest and an increased investment in pursuing extrapair copulations of his own. Longer flight feathers are a sign of age and quality in kingbirds, so males with this trait are expected to spend more time pursuing extrapair copulations as they are more likely to be successful in this endeavor.

A substantial portion of the remaining coastal pine savanna in the southeastern U.S. is burned periodically to maintain habitat for the endangered Mississippi Sandhill Crane (Grus canadensis pulla). However, the effects of this burning on other species of birds are unknown. Therefore, a one-summer study was conducted to determine the short-term response of non-target bird species to changes in vegetation structure due to winter prescribed burning of coastal pine savanna. Eight 25-ha study plots were censused using the spot-mapping technique from May - July 1995 at the Mississippi Sandhill Crane National Wildlife Refuge. More species were observed on old burn sites (burned 1.5 - 3.5 yr prior to the study) than new burn sites (burned the previous winter). Gross vegetation features did not differ between treatments. A total of 17 breeding species were recorded during the study. Although there were no significant differences within individual species' densities, seven species were observed only on old burn sites. Winter prescribed burning affected the presence of only shrub-characteristic species. / Department of Biology

I studied selenium contamination at Cibola National Wildlife Refuge (Cibola NWR) in the lower Colorado River Valley, California and Arizona. My objectives were to: (1) determine whether local irrigation practices resulted in exposure of fish to toxic levels of selenium; (2) assess the risks to humans of consuming fish from Cibola NWR; (3) assess whether diversity and abundance of fish were related to selenium concentrations or other water quality variables. Water, sediment, fish, crayfish, and aquatic plants were collected from sites which received irrigation return flows and sites which did not. Selenium was below toxic levels at sites receiving irrigation return flows. Selenium was at the toxicity threshhold for fish at two sites receiving water directly from the Colorado River. Concentrations of selenium in fillets of largemouth bass (Micropterus salmoides) from one lake exceeded levels that elicit consumption advisories in California. Most people would be unlikely to consume toxic amounts of fish, but an advisory should be posted to inform people about potential risks. Gill nets were used to determine species diversity and abundance. There were no strong correlations between selenium levels and indices of species richness and equitability. There was a consistent inverse relationship between selenium levels and catch-per-unit-effort for all species combined. This inverse relationship suggested that selenium may be one of a suite of factors limiting abundance of fish. Correlations between indices of species diversity and abundance and values of water quality variables generally were not significant. Temperature and salinity may have limited abundance of some species seasonally. Increases in selenium levels at sites that are already at the toxicity threshhold could impair reproduction of sensitive species. Population declines and concern about edibility of fish could impair the recreational fishery. Therefore, site-specific ways to reduce selenium accumulation should be studied and implemented. Site-specific limnological conditions may play a role in accumulation of selenium to toxic levels, but major sources of selenium appear to be upstream in the Colorado River basin rather than local agricultural practices. Therefore, selenium input to the Colorado River from sources throughout the basin should be monitored and reduced where possible.

Shallow water availability coupled with anthropogenic degradation of seagrass beds limits wading bird food resources in dynamic coastal ecosystems. Identifying prey species critical to wading bird reproductive success and the environmental drivers of key prey species abundance is important for understanding how environmental stressors influence prey and change the quality of foraging patches. Little Blue Herons (Egretta caerulea) are reportedly generalists eating insects, crustaceans, and fish; however, the proportions of prey items in the diet may shift spatially and temporally from freshwater to marine systems during breeding and non-breeding periods. I investigated prey selection by Little Blue Herons in Florida at the Great White Heron National Wildlife Refuge and the western Florida Bay, during 2016 and 2017 breeding seasons by investigating prey availability at low-tide locations along mudflats compared to stomach regurgitate samples collected from Little Blue Heron chicks 1 to 4 weeks old. Little Blue Herons selected Gulf toadfish (Opsanus beta) and prawns (Farfantepenaeus spp.) from the estuarine environment, but also consumed terrestrial prey (e.g. tree crabs) suggesting Little Blue Heron foraging habitat is not restricted to tidal flats. Additionally, these results support the characterization of Little Blue Herons as a generalist. After identifying important prey species, I modeled the associations of selected prey species with submerged aquatic vegetation density and abiotic variables to better understand habitat preferences and important habitat characteristics that drive prey density. Models support total seagrass density and algal density as having the greatest effect on prey selected by Little Blue Herons. Prawn density has a strong positive association with seagrass density. Gulf toadfish (Opsanus beta) and prawns (Farfantepenaeus spp.) had strong positive association with algae while pipefish (Syngnathidae) had a strong negative association with algae suggesting algae density in seagrass meadows should be considered when assessing the quality of seagrass meadows for Little Blue Heron prey and habitat suitability. My results varied from previous studies where prawns and gulf toadfish were associated with specific seagrass species. Therefore, some Little Blue Heron prey species in south Florida may not be affected by changes in submerged aquatic vegetation community composition if submerged aquatic vegetation densities remain constant. Studies are needed that clarify the complex interactions between prey and specific habitat metrics to validate the strength of landscape scale drivers of wading bird prey densities in dynamic coastal ecosystems and to determine how these communities will respond to anthropogenic environmental change. / Includes bibliography. / Thesis (M.S.)--Florida Atlantic University, 2018. / FAU Electronic Theses and Dissertations Collection

A declining population of ring-necked pheasants (Phasianus colchicus) was studied at Tule Lake National Wildlife Refuge (TLNWR) from the summer of 1990 through the spring of 1993. Pheasant densities/50 ha at TLNWR in 1989, 1991, and 1992 were considerably lower (16.86, 8.49, and 6.81) than the >62 density seen in the mid-1950s. Mean body weight of hen pheasants at TLNWR was significantly lower than hens at nearby Lower Klamath National Wildlife Refuge (LKNWR) which was not intensively farmed. Mean tarsal lengths of hens at TLNWR were also significantly shorter than hens at LKNWR, suggesting reduced skeletal growth and potential nutritional problems. The lack of adequate cover was responsible for poor early nest success at TLNWR. Later in the season, spring planted crops provided adequate cover to conceal nesting hens; however, only 0.21 young in 1991 and 0.02 young in 1992 were produced per radio-equipped hen. These rates are extremely low compared to rates required to maintain a stable population. Most adult mortality occurred during the spring and early summer months at TLNWR before crops provided adequate cover, and long before pesticide applications. The main predator of the pheasants was the golden eagle (Aquila chrvsaetos). Of special concern at TLNWR was pheasant and other wildlife exposure to anticholinesterase (antiChE) insecticides used on agricultural croplands at the refuge. Direct toxicity of antiChE compounds (in this case methamidophos) killed 2 young pheasants, but no adult radio-equipped hens died as a direct result of insecticide intoxication. This finding was of particular interest because 15% of the adult pheasants collected in and around potato fields had 55% brain ChE inhibition. The extent of the effects of insecticide exposure on the survivorship of pheasant young was uncertain as they were not radio-equipped. The overriding factor impacting the pheasant population at TLNWR and to a lesser extent LKNWR was poor habitat, especially in the spring when most mortality occurred. The poor habitat also resulted in extremely low recruitment (up to 1 September). Nearly all adult mortality and most of the low recruitment occurred before the insecticide spray season. The population was nearly extirpated during the severe winter of 1992-93. / Graduation date: 1995

Depressions in the prairie pothole region (PPR) are commonly referred to as sloughs and were formed during the most recent glacial retreat, ~10-17 kyrs ago. They are hydrologically isolated, as they are not permanently connected by surface inflow or outflow channels. Extreme thunderstorms are common across the prairies and the hydrologic response of isolated wetlands to intense rainfall events is poorly understood. The purpose of this study was to compare the response of different landscape/ecological elements of a prairie wetland to snowmelt and extreme rainstorms. Comparisons were completed by investigating the spring snowmelts of 2005 and 2006 and the rainstorm event of June 17 - 18, 2005, in which 103 mm fell at the St. Denis National Wildlife Area (NWA) Saskatchewan, Canada (106°06'W, 52°02'N). The wetland was separated into five landscape positions, the pond center (PC), grassed edge (GE), tree ring (TR), convex upland (CXU), and concave upland (CVU). Comparison of the rainfall of June 17 18, 2005 with the spring snowmelts of 2005 and 2006 indicates that the hydrologic consequences of these different events are similar. Overland flow, substantial ponding in lowlands, and recharge of the groundwater occur in both cases. Analysis of this intense rainfall has provided evidence that common, intense rainstorms are hydrologically equivalent to the annual spring snowmelt, the major source of water for closed catchments in the PPR.

Depressions in the prairie pothole region (PPR) are commonly referred to as sloughs and were formed during the most recent glacial retreat, ~10-17 kyrs ago. They are hydrologically isolated, as they are not permanently connected by surface inflow or outflow channels. Extreme thunderstorms are common across the prairies and the hydrologic response of isolated wetlands to intense rainfall events is poorly understood. The purpose of this study was to compare the response of different landscape/ecological elements of a prairie wetland to snowmelt and extreme rainstorms. Comparisons were completed by investigating the spring snowmelts of 2005 and 2006 and the rainstorm event of June 17 - 18, 2005, in which 103 mm fell at the St. Denis National Wildlife Area (NWA) Saskatchewan, Canada (106°06'W, 52°02'N). The wetland was separated into five landscape positions, the pond center (PC), grassed edge (GE), tree ring (TR), convex upland (CXU), and concave upland (CVU). Comparison of the rainfall of June 17 18, 2005 with the spring snowmelts of 2005 and 2006 indicates that the hydrologic consequences of these different events are similar. Overland flow, substantial ponding in lowlands, and recharge of the groundwater occur in both cases. Analysis of this intense rainfall has provided evidence that common, intense rainstorms are hydrologically equivalent to the annual spring snowmelt, the major source of water for closed catchments in the PPR.

The Lower Rio Grande Valley (LRGV) of Texas is an ecologically diverse region in the United States and marks the northernmost extension for many tropical species of plants and animals. Since the early 1900s, 95% of the native Tamaulipan brushlands have been cleared due to agricultural practices and urban development. The plain chachalaca (Ortalis vetula) is a medium sized bird endemic to the native brushlands of the LRGV. In 2003, I trapped and radio-tagged 29 birds (16 males, 13 females) to evaluate the effects of fragmentation on the population dynamics (i.e., survival, mortality, and movements) of this brushland species. My study objectives were to estimate (1) seasonal survival of chachalacas by sex, and (2) ranges, core areas, and movements using radio telemetry. Mammalian predation (43%, n = 6) and unknown (43%, n = 6) deaths accounted for the majority of mortality observed. I found no difference (P > 0.05) in estimated 8-month survival (December 2003-July 2004) between males (S = 0.364, SE = 0.132) and females (S = 0.405, SE = 0.153). In comparing seasonal survival for all birds (males and females combined), I observed a difference (P < 0.05) in survival between the nesting (S = 0.414, SE = 0.103) and breeding seasons (S = 0.917, SE = 0.079). Female ranges (x‾ = 117 ha, range = 42–177 ha) and core areas (x‾ = 23 ha, range = 5–46 ha) during the nesting season were larger than male ranges (x‾ = 41 ha, range = 31–46 ha) and core areas (x‾ =10 ha, range = 7–14 ha) during the same period. During the breeding season, female ranges (x‾ = 59 ha, range = 10–188 ha) and core areas (x‾ = 9 ha, range = 2–33 ha) were similar to male ranges (x‾ = 48 ha, range = 4–130 ha) and core areas (x‾ = 9 ha, range = 1–23 ha). Mean distances between seasons were similar for both sexes (females, nesting, x‾ = 486, breeding, x‾ = 345; males, nesting, x‾ = 184, breeding, x‾ = 292), though females distances generally were greater. Dispersal defined as movement off the Santa Ana National Wildlife Refuge was observed for 3 birds. In 2 cases, a radio-tagged female and male were observed crossing the Rio Grande River (approximately 100-m wide) to habitat in Mexico. Study results suggested mammalian predation may limit the growth of chachalaca populations. Though land use changes such as agricultural uses may not directly limit chachalaca populations in providing cover and food, concentration of populations in remnant native brushlands may serve as ecological “sinks” to the species. Greater range and movement data observed in my study may be attributed to suboptimal habitat (i.e., increased fragmentation) for plain chachalacas.