Machine Learning Methods and Visual Observations to Categorize Behavior of Grazing Cattle Using Accelerometer Signals.

Accelerometers worn by animals produce distinct behavioral signatures, which can be classified accurately using machine learning methods such as random forest decision trees. The objective of this study was to identify accelerometer signal separation among parsimonious behaviors. We achieved this objective by (1) describing functional differences in accelerometer signals among discrete behaviors, (2) identifying the optimal window size for signal pre-processing, and (3) demonstrating the number of observations required to achieve the desired level of model accuracy,. Crossbred steers (Bos taurus indicus; n = 10) were fitted with GPS collars containing a video camera and tri-axial accelerometers (read-rate = 40 Hz). Distinct behaviors from accelerometer signals, particularly for grazing, were apparent because of the head-down posture. Increasing the smoothing window size to 10 s improved classification accuracy (p < 0.05), but reducing the number of observations below 50% resulted in a decrease in accuracy for all behaviors (p < 0.05). In-pasture observation increased accuracy and precision (0.05 and 0.08 percent, respectively) compared with animal-borne collar video observations.

Maize Elongin C interacts with the viral genome-linked protein, VPg, of Sugarcane mosaic virus and facilitates virus infection.

The viral genome-linked protein, VPg, of potyviruses is involved in viral genome replication and translation. To determine host proteins that interact with Sugarcane mosaic virus (SCMV) VPg, a yeast two-hybrid screen was used and a maize (Zea mays) Elongin C (ZmElc) protein was identified. ZmELC transcript was observed in all maize organs, but most highly in leaves and pistil extracts, and ZmElc was present in the cytoplasm and nucleus of maize cells in the presence or absence of SCMV. ZmELC expression was increased in maize tissue at 4 and 6 d post SCMV inoculation. When ZmELC was transiently overexpressed in maize protoplasts the accumulation of SCMV RNA was approximately doubled compared with the amount of virus in control protoplasts. Silencing ZmELC expression using a Brome mosaic virus-based gene silencing vector (virus-induced gene silencing) did not influence maize plant growth and development, but did decrease RNA accumulation of two isolates of SCMV and host transcript encoding ZmeIF4E during SCMV infection. Interestingly, Maize chlorotic mottle virus, from outside the Potyviridae, was increased in accumulation after silencing ZmELC expression. Our results describe both the location of ZmElc expression in maize and a new activity associated with an Elc: support of potyvirus accumulation.

A critical examination of indices of dynamic interaction for wildlife telemetry studies.

Wildlife scientists continue to be interested in studying ways to quantify how the movements of animals are interdependent - dynamic interaction. While a number of applied studies of dynamic interaction exist, little is known about the comparative effectiveness and applicability of available methods used for quantifying interactions between animals. We highlight the formulation, implementation and interpretation of a suite of eight currently available indices of dynamic interaction. Point- and path-based approaches are contrasted to demonstrate differences between methods and underlying assumptions on telemetry data. Correlated and biased correlated random walks were simulated at a range of sampling resolutions to generate scenarios with dynamic interaction present and absent. We evaluate the effectiveness of each index at identifying different types of interactive behaviour at each sampling resolution. Each index is then applied to an empirical telemetry data set of three white-tailed deer (Odocoileus virginianus) dyads. Results from the simulated data show that three indices of dynamic interaction reliant on statistical testing procedures are susceptible to Type I error, which increases at fine sampling resolutions. In the white-tailed deer examples, a recently developed index for quantifying local-level cohesive movement behaviour (the di index) provides revealing information on the presence of infrequent and varying interactions in space and time. Point-based approaches implemented with finely sampled telemetry data overestimate the presence of interactions (Type I errors). Indices producing only a single global statistic (7 of the 8 indices) are unable to quantify infrequent and varying interactions through time. The quantification of infrequent and variable interactive behaviour has important implications for the spread of disease and the prevalence of social behaviour in wildlife. Guidelines are presented to inform researchers wishing to study dynamic interaction patterns in their own telemetry data sets. Finally, we make our code openly available, in the statistical software R, for computing each index of dynamic interaction presented herein.

SNAPSHOT USA 2020: A second coordinated national camera trap survey of the United States during the COVID-19 pandemic.

Managing wildlife populations in the face of global change requires regular data on the abundance and distribution of wild animals, but acquiring these over appropriate spatial scales in a sustainable way has proven challenging. Here we present the data from Snapshot USA 2020, a second annual national mammal survey of the USA. This project involved 152 scientists setting camera traps in a standardized protocol at 1485 locations across 103 arrays in 43 states for a total of 52,710 trap-nights of survey effort. Most (58) of these arrays were also sampled during the same months (September and October) in 2019, providing a direct comparison of animal populations in 2 years that includes data from both during and before the COVID-19 pandemic. All data were managed by the eMammal system, with all species identifications checked by at least two reviewers. In total, we recorded 117,415 detections of 78 species of wild mammals, 9236 detections of at least 43 species of birds, 15,851 detections of six domestic animals and 23,825 detections of humans or their vehicles. Spatial differences across arrays explained more variation in the relative abundance than temporal variation across years for all 38 species modeled, although there are examples of significant site-level differences among years for many species. Temporal results show how species allocate their time and can be used to study species interactions, including between humans and wildlife. These data provide a snapshot of the mammal community of the USA for 2020 and will be useful for exploring the drivers of spatial and temporal changes in relative abundance and distribution, and the impacts of species interactions on daily activity patterns. There are no copyright restrictions, and please cite this paper when using these data, or a subset of these data, for publication.

Effectiveness and Efficiency of Corral Traps, Drop Nets and Suspended Traps for Capturing Wild Pigs (Sus scrofa).

Strategic control and eradication programs for wild pigs (Sus scrofa) are being developed to help curtail the expanding populations of this invasive, alien species. Drop nets and corral traps have a long history of capturing a multitude of wildlife species, so we evaluated the effectiveness and efficiency of these traps for controlling wild pigs in southern Oklahoma. We also developed and evaluated a suspended metal trap that provided real-time monitoring and deployment to capture animals. Effectiveness of each trap type was estimated as the proportion of pigs removed from the total population, whereas efficiency was calculated based on catch per unit effort (CPUE) (i.e., the number of person hours per pig removal). During 3 years of study (2010-2012), we removed 601 pigs, 296 using drop nets, 60 using corral traps, and 245 using suspended traps. Suspended traps removed 88.1% of the estimated population, whereas drop nets removed 85.7% and corral traps removed 48.5%. CPUE was 0.64 person hours/pig using suspended traps followed by 1.9 person hours/pig for drop nets and 2.3 person hours/pig for corral traps. Drop nets and suspended traps were more effective at removing a large proportion of the population (>85%), mainly through whole sounder removal, but the suspended trap with real-time notifications was the most efficient trap type, requiring fewer person hours to operate.

Prevalence of Neospora caninum Exposure in Wild Pigs (Sus scrofa) from Oklahoma with Implications of Testing Method on Detection.

Neospora caninum is a protozoan parasite, reported as a leading cause of cattle abortions and reproductive failure worldwide, costing the cattle industry approximately $1.3 billion annually. With wild pig (Sus scrofa) populations estimated at over six million in the United States, contact between wild pigs and livestock is inevitable, mainly because of the widespread geographic co-occurrence of the two species. As a known reservoir for numerous fungal, bacterial, viral and parasitic diseases, wild pigs are of particular importance for human and veterinary health relative to the prevention of infectious diseases. The seroprevalence of N. caninum in wild pig populations was previously documented in the United States, raising the question as to their exposure point of prevalence. This research screened 116 individual wild pigs for N. caninum using a variety of available assays. Using two different commercially available ELISA test kits, seroprevalence ranged from 12.5% to 67.8%. The Indirect Fluorescent Antibody Test resulted in our highest percent seroprevalence for these samples, at 84.1%. However, none of our samples showed any presence of N. caninum or associated pathologies via histological evaluation of representative tissues. Importantly, the assays used in this study were not congruent with all duplicate samples or between the test types used. The implications of these non-congruent results demonstrates that currently available testing assays produce variable results, underscoring the need for more reliable testing kits and a standardized methodology when assessing disease prevalence in wildlife, particularly for N. caninum in wild pigs, which impacts prevalence and comparability across studies.

SNAPSHOT USA 2019: a coordinated national camera trap survey of the United States.

With the accelerating pace of global change, it is imperative that we obtain rapid inventories of the status and distribution of wildlife for ecological inferences and conservation planning. To address this challenge, we launched the SNAPSHOT USA project, a collaborative survey of terrestrial wildlife populations using camera traps across the United States. For our first annual survey, we compiled data across all 50 states during a 14-week period (17 August-24 November of 2019). We sampled wildlife at 1,509 camera trap sites from 110 camera trap arrays covering 12 different ecoregions across four development zones. This effort resulted in 166,036 unique detections of 83 species of mammals and 17 species of birds. All images were processed through the Smithsonian's eMammal camera trap data repository and included an expert review phase to ensure taxonomic accuracy of data, resulting in each picture being reviewed at least twice. The results represent a timely and standardized camera trap survey of the United States. All of the 2019 survey data are made available herein. We are currently repeating surveys in fall 2020, opening up the opportunity to other institutions and cooperators to expand coverage of all the urban-wild gradients and ecophysiographic regions of the country. Future data will be available as the database is updated at eMammal.si.edu/snapshot-usa, as will future data paper submissions. These data will be useful for local and macroecological research including the examination of community assembly, effects of environmental and anthropogenic landscape variables, effects of fragmentation and extinction debt dynamics, as well as species-specific population dynamics and conservation action plans. There are no copyright restrictions; please cite this paper when using the data for publication.

Ecotypic and genotypic effects on regrowth and heading date in switchgrass (Panicum virgatum).

Switchgrass (Panicum virgatum L.) is a native perennial grass species with great potential for bioenergy and forage. However, knowledge about its genetics and biology related to breeding is still in its infancy. Studying the diversity of switchgrass germplasm will shed light on variability, response to environmental conditions, adaptability, breeding, etc. Thirty-six switchgrass accessions/cultivars were used to study the ecotypic and genotypic effects on regrowth, heading date, and vegetative growth period. The R-360 honeycomb design was used for planting these accessions in 2007. Data on regrowth and heading dates were recorded in 2008, 2010, and 2011. Vegetative growth period was calculated by subtracting the regrowth date from the heading date. It was found that the lowland started regrowing earlier (77 ± 0.4 days of the year, DOY) than the upland ecotype (82 ± 0.3 DOY). The upland had earlier heading date (160 ± 0.4 DOY) than the lowland ecotype (173 ± 0.5 DOY). Vegetative growth period was about 18 days longer in the lowland (89 ± 0.6 days) than the upland ecotype (71 ± 0.4 days). For switchgrass (i.e., all accessions), biomass yield was related positively to growth period and heading date; however, biomass was only weakly related to regrowth. Therefore, when targeting biomass in the breeding program, growth period may be a quick and reliable reference in both ecotypes to quickly estimate biomass potential while regrowth and heading date may be better used as a parameter for accessions within an ecotype.

Blood parasites in wood storks (Mycteria americana) from the southeastern United States.

Wood storks (Mycteria americana) are endangered throughout their breeding range in the United States. Because of this, researchers have had little opportunity to thoroughly examine the continental wood stork population for endoparasites. The blood protozoan Haemoproteus crumenium has been identified from several populations of wood storks in North America. However, there have been no reports of wood storks being infected with species of Plasmodium or Leucocytozoon. During 2003, 42 nestling wood storks were captured at sites in Georgia and 27 free-ranging wood storks were captured at postbreeding dispersal sites in Mississippi and Louisiana. Two thin blood smears were made from each bird, and they were examined for parasites. Haemoproteus crumenium was found in 5 wood storks (7.2%), whereas Leucocytozoon sp., Plasmodium sp., and microfilariae were not observed. Mean intensity of H. crumenium was 4.4 +/- 0.7 (SE) per 2,000 erythrocytes counted. All infected wood storks were subadult or adult and were from postbreeding dispersal sites in Mississippi.

Pseudorabies Virus and Brucella abortus from an Expanding Wild Pig ( Sus scrofa ) Population in Southern Oklahoma, USA.

Wild pigs ( Sus scrofa ) are causing increasing ecologic and economic damage at a global scale. Because wild pigs can carry ≥65 diseases that affect livestock, their widespread expansion threatens native wildlife and livestock. We screened wild pigs from south-central Oklahoma, US for antibodies against Brucella abortus , pseudorabies virus (PRV), and porcine reproductive and respiratory syndrome virus (PRRS). These pathogens were chosen because they are part of eradication programs in the US and could have large economic impacts on domestic livestock if transmitted from wild animals. We tested 282 serum samples during spring 2010 (n=149) and 2011 (n=133) and found an overall exposure rate to PRV of 24.1% (n=68); PRV was detected at two of three study sites. Two wild pigs had detectable antibody to B. abortus , and one had detectable antibody to PRRS. On average, 27% of wild pigs within a sounder were positive for PRV antibody, with 44% of the sounders (16/36) having at least one positive individual. These data highlight that wild pigs could carry pathogens that affect domestic livestock. Because the US is free of these pathogens in commercial livestock operations, continued surveillance and vaccination of domestic livestock are needed. Commercial livestock producers at the wildlife-livestock interface may benefit from spatial prioritization of risk zones to facilitate strategic control efforts.

Mapping areas of spatial-temporal overlap from wildlife tracking data.

BACKGROUND: The study of inter-individual interactions (often termed spatial-temporal interactions, or dynamic interactions) from remote tracking data has focused primarily on identifying the presence of such interactions. New datasets and methods offer opportunity to answer more nuanced questions, such as where on the landscape interactions occur. In this paper, we provide a new approach for mapping areas of spatial-temporal overlap in wildlife from remote tracking data. The method, termed the joint potential path area (jPPA) builds from the time-geographic movement model, originally proposed for studying human movement patterns. RESULTS: The jPPA approach can be used to delineate sub-areas of the home range where inter-individual interaction was possible. Maps of jPPA regions can be integrated with existing geographic data to explore landscape conditions and habitat associated with spatial temporal-interactions in wildlife. We apply the jPPA approach to simulated biased correlated random walks to demonstrate the method under known conditions. The jPPA method is then applied to three dyads, consisting of fine resolution (15 minute sampling interval) GPS tracking data of white-tailed deer (Odocoileus virginianus) collected in Oklahoma, USA. Our results demonstrate the ability of the jPPA to identify and map jPPA sub-areas of the home range. We show how jPPA maps can be used to identify habitat differences (using percent tree canopy cover as a habitat indicator) between areas of spatial-temporal overlap and the overall home range in each of the three deer dyads. CONCLUSIONS: The value of the jPPA approach within current wildlife habitat analysis workflows is highlighted along with its simple and straightforward implementation and interpretation. Given the current emphasis on remote tracking in wildlife movement and habitat research, new approaches capable of leveraging both the spatial and temporal information content contained within these data are warranted. We make code (in the statistical software R) for implementing the jPPA approach openly available for other researchers.

Annual survival and site fidelity of free-ranging white-tailed deer (Odocoileus virginianus): comparative demography before (1983-1992) and after (1993-2005) spatial confinement.

Survival and movement are important demographic variables influencing the dynamics of large herbivores with implications for management and evolution of life-history strategies. Management practices such as spatial confinement and harvest regulation attempt to control survival and movement for the sustainability of harvested deer populations, but a paucity of long-term data exists on these management practices. We examined annual survival and site fidelity of free-ranging white-tailed deer (Odocoileus virginianus) over 10 years (1983-1992) to compare demographic parameters after spatial confinement (1993-2005). We used capture records (n = 174; 104 females, 70 males), marked deer recaptures (n = 42), and dead recoveries (n = 68) to estimate sex-specific, age-specific and time-specific parameters. We found that annual female survival was 50% from 1983-1987 during a period of intense harvest, but increased to 93.7% after intense harvesting was eliminated. Prior to spatial confinement, annual survival of marked male deer averaged 36.7%-42.5%. After spatial confinement, annual survival increased on average for males (58%-99%) and females (77%-98%). Females showed high levels of site fidelity (>99%) prior to spatial confinement, whereas males showed much less site fidelity (≤4.5% for the 2 top-ranking models). During spatial confinement, the semi-impermeable fence effectively increased site fidelity of males (≥56%). These results stem from long-term study (23 years) of a large herbivore experiencing changes to life-history, resulting from changes in management that were applied to the population and aimed at altering population demographics, for sustainability of a harvestable population of deer.

Hematozoa and a new haemoproteid species from Cathartidae (New World vulture) in South Carolina.

A survey of turkey vultures (Cathartes aura) in South Carolina revealed the presence of a novel haemoproteid. Haemoproteus catharti n. sp. is described from the slides developed from the blood of this vulture. The new species is a thick, halteridial form with complete margins, and is considered distinct from the other recognized haemoproteids from the diurnal raptors in shape, pigment number, parasite outline, and host-family specificity. While reviewing blood films, a series of unusual immature schizonts of Plasmodium sp. were also observed, and these are illustrated, along with another distinct haemoproteid from old slides of poor quality produced from turkey vulture blood.

Prioritizing conservation of ungulate calving resources in multiple-use landscapes.

BACKGROUND: Conserving animal populations in places where human activity is increasing is an ongoing challenge in many parts of the world. We investigated how human activity interacted with maternal status and individual variation in behavior to affect reliability of spatially-explicit models intended to guide conservation of critical ungulate calving resources. We studied Rocky Mountain elk (Cervus elaphus) that occupy a region where 2900 natural gas wells have been drilled. METHODOLOGY/PRINCIPAL FINDINGS: We present novel applications of generalized additive modeling to predict maternal status based on movement, and of random-effects resource selection models to provide population and individual-based inference on the effects of maternal status and human activity. We used a 2×2 factorial design (treatment vs. control) that included elk that were either parturient or non-parturient and in areas either with or without industrial development. Generalized additive models predicted maternal status (parturiency) correctly 93% of the time based on movement. Human activity played a larger role than maternal status in shaping resource use; elk showed strong spatiotemporal patterns of selection or avoidance and marked individual variation in developed areas, but no such pattern in undeveloped areas. This difference had direct consequences for landscape-level conservation planning. When relative probability of use was calculated across the study area, there was disparity throughout 72-88% of the landscape in terms of where conservation intervention should be prioritized depending on whether models were based on behavior in developed areas or undeveloped areas. Model validation showed that models based on behavior in developed areas had poor predictive accuracy, whereas the model based on behavior in undeveloped areas had high predictive accuracy. CONCLUSIONS/SIGNIFICANCE: By directly testing for differences between developed and undeveloped areas, and by modeling resource selection in a random-effects framework that provided individual-based inference, we conclude that: 1) amplified selection or avoidance behavior and individual variation, as responses to increasing human activity, complicate conservation planning in multiple-use landscapes, and 2) resource selection behavior in places where human activity is predictable or less dynamic may provide a more reliable basis from which to prioritize conservation action.

Identifying and prioritizing greater sage-grouse nesting and brood-rearing habitat for conservation in human-modified landscapes.

BACKGROUND: Balancing animal conservation and human use of the landscape is an ongoing scientific and practical challenge throughout the world. We investigated reproductive success in female greater sage-grouse (Centrocercus urophasianus) relative to seasonal patterns of resource selection, with the larger goal of developing a spatially-explicit framework for managing human activity and sage-grouse conservation at the landscape level. METHODOLOGY/PRINCIPAL FINDINGS: We integrated field-observation, Global Positioning Systems telemetry, and statistical modeling to quantify the spatial pattern of occurrence and risk during nesting and brood-rearing. We linked occurrence and risk models to provide spatially-explicit indices of habitat-performance relationships. As part of the analysis, we offer novel biological information on resource selection during egg-laying, incubation, and night. The spatial pattern of occurrence during all reproductive phases was driven largely by selection or avoidance of terrain features and vegetation, with little variation explained by anthropogenic features. Specifically, sage-grouse consistently avoided rough terrain, selected for moderate shrub cover at the patch level (within 90 m(2)), and selected for mesic habitat in mid and late brood-rearing phases. In contrast, risk of nest and brood failure was structured by proximity to anthropogenic features including natural gas wells and human-created mesic areas, as well as vegetation features such as shrub cover. CONCLUSIONS/SIGNIFICANCE: Risk in this and perhaps other human-modified landscapes is a top-down (i.e., human-mediated) process that would most effectively be minimized by developing a better understanding of specific mechanisms (e.g., predator subsidization) driving observed patterns, and using habitat-performance indices such as those developed herein for spatially-explicit guidance of conservation intervention. Working under the hypothesis that industrial activity structures risk by enhancing predator abundance or effectiveness, we offer specific recommendations for maintaining high-performance habitat and reducing low-performance habitat, particularly relative to the nesting phase, by managing key high-risk anthropogenic features such as industrial infrastructure and water developments.