Occupancy Estimation Research Articles

Long-term monitoring in transition: Resolving spatial mismatch and integrating multistate occupancy data

The success of long-term wildlife monitoring programs can be influenced by many factors and study designs often represent compromises between spatial scales and costs. Adaptive monitoring programs can iteratively manage this tension by adopting new cost-efficient technologies, which can provide projects the opportunity to reallocate costs to address new hypotheses, adapt to changing ecological conditions, or adjust sampling scale or resolution. If there is interest in longer time series of monitoring data, methodological transitions may necessitate integrated models to link newer data with historical data. However, data integration can be difficult if spatial or temporal scales are mismatched. Here, we develop an integrated multistate site-occupancy model and resolve sample unit spatial mismatch to link datasets from two northern spotted owl (Strix occidentalis caurina) monitoring schemes that broadly overlapped during a methodological transition. The first dataset was obtained from a decades-long spotted owl monitoring program using call-playback and mark-resight surveys on historical territories of varying size and shape. This monitoring program has recently transitioned to passive acoustic monitoring of randomly selected 5-km2 hexagons over larger spatial extents. Both monitoring datasets overlapped with areas in which barred owl (Strix varia), an invasive competitor that has played an important role in northern spotted owl declines, were being removed experimentally. Reconciling spatial mismatch substantially increased the representation of the call-playback dataset and integrating the two datasets increased precision of spotted owl use and paired occupancy estimates relative to single dataset estimates. Estimates of spotted owl pair occupancy across the study area were lower than previous territory-based estimates based on call-playback surveys. Our integrated model further showed that a concurrent barred owl removal experiment increased landscape use and site occupancy by pairs of spotted owls. Our empirical application of an integrated modelling approach demonstrates a useful analytical framework for long-term monitoring efforts undergoing methodological transitions (e.g. mark-recapture to non-invasive population monitoring). This framework allows monitoring programs to maintain continuity of monitoring objectives across methodological transitions, rigorously incorporate previous findings, and adaptively respond to changing ecological conditions.

Estimates of wildlife species richness, occupancy, and habitat preference in a residential landscape in New York State

Estimates of wildlife species richness, occupancy, and habitat preference in a residential landscape in New York State

Wetland occupancy by duck broods in cropland‐dominated landscapes of the United States Prairie Pothole Region

AbstractThe Prairie Pothole Region (PPR) is globally important for breeding waterfowl but has been altered via wetland drainage and grassland conversion to accommodate agricultural land use. Thus, understanding the ecology of waterfowl in these highly modified landscapes is essential for their conservation. Brood occurrence is the cumulative outcome of key life‐history events including pair formation and territory establishment, nest success, and early brood survival. We applied new technological advances in brood surveying methods to understand brood use of wetlands and how land use and wetland‐specific factors influenced brood use of 413 wetlands in crop‐dominated landscapes in the PPR of Iowa, Minnesota, North Dakota, and South Dakota, USA, during summers of 2018–2020. Dynamic occupancy models combining information from 2 visits throughout the year revealed no difference among the 4 states or between private and public lands, resulting in a region‐wide annual wetland occupancy estimate of 0.41 (95% credible interval [CrI] = 0.26, 0.58). We assessed aquatic invertebrate forage availability, wetland and upland vegetation communities, and various water chemistry metrics in a subset (n = 225) of these wetlands to evaluate how landscape and wetland‐specific factors influenced occupancy. The amount of grassland surrounding wetlands was the only variable to influence occupancy at a landscape scale, while wetland size, invertebrates, fish, and vegetation communities influenced occupancy at finer scales. Closer scrutiny of wetland area revealed occupancy was greater in small wetlands after controlling for total wetland area. Our results indicate the greatest constraint on brood occupancy across crop‐dominated landscapes of the PPR in the United States was the occurrence of semipermanent wetlands suitable for brood rearing. Other factors, such as wetland vegetation or surrounding land use, had minor intervening influences on duck brood use and ducks were distributed invariant of wetland ownership or broad spatial processes occurring among states. These results demonstrated wetland conservation and restoration strategies are likely to yield gains in annual duck broods across this vast, altered, and highly modified landscape.

To improve existing marsh bird survey protocols, we need to evaluate closure assumptions

AbstractOur collective understanding of secretive marsh birds has increased in the past decades due to the development and implementation of the North American Standardized Marsh Bird Monitoring Protocol (hereafter, Protocol). The Protocol proposes call broadcast surveys to increase vocalization and detection rates within 3 standardized survey periods aimed at surveying peak breeding activity for a suite of secretive marsh birds. We noted a trend in the literature linking occupancy modeling with the survey design from the Protocol, despite some evidence that vocalizations decline across survey periods, which could indicate lack of population closure. An underlying assumption of occupancy modeling is closure, and the Protocol was designed to focus on only birds which will remain in an area throughout the breeding season and not migrants that may only be present in the first survey period. Including migrating marsh birds, especially if a large percentage of marsh bird detections are migrants, can bias occupancy estimates and lead to erroneous density and population size estimates that may affect conclusions about habitat resource and bird associations. We urge researchers and managers to carefully consider the analytical and field techniques when designing studies for marsh birds and to not simply pair the biweekly survey design within the Protocol with occupancy modeling and ignore closure assumptions, turnover rates, and potential differences in resource use by migrating and breeding marsh birds. Specifically, we suggest that researchers consider short survey interval times (e.g., ~2 days rather than 2 weeks) or continuous call monitoring using automated recorded devices deployed for 1–2 weeks per survey location when using occupancy analysis on secretive marsh birds. We also call for future study of turnover rates, stopover duration, and vocalization rates during migration and breeding periods to better inform study designs and increase the appropriateness of statistical analysis.

Parrot population trends in Nicaragua revealed by long-term monitoring

SummaryMonitoring parrot populations is of high importance because there is a general lack of quantified population trends for one of the most threatened avian orders. We surveyed parrots in Nicaragua in 1995, 1999, 2004, and 2013 at a minimum of 227 points within 56 sites stratified among the Pacific, Central Highlands, and Caribbean biogeographical regions to assess population trends. From point-count data we calculated encounter rate, flock rate, and flock size metrics and we used presence/absence data to generate species-specific occupancy estimates. Encounter rate, flock rate, and flock size data suggested family-level declines from 1995 to 2004 with some recovery between 2004 and 2013. Patterns of parrot occupancy varied among species with four decreasing, five increasing, and two with no detectable change. Six species of conservation concern are identified, including the Critically Endangered Great Green Macaw and Yellow-naped Parrot, additionally Olive-throated Parakeet, Scarlet Macaw, Brown-hooded Parrot, and White-crowned Parrot, only listed as Least Concern. All six are likely suffering from deforestation and potential unchecked trade activity in the Caribbean. Differing population trends of the regionally disjunct Yellow-naped Parrot subspecies suggest a link to variable deforestation and trade pressure experienced between the Pacific and Caribbean. Our results highlight the importance of actively monitoring changing parrot populations, even when considered Least Concern, so that directed conservation actions can be taken if needed.

Non-Intrusive Room Occupancy Prediction Performance Analysis Using Different Machine Learning Techniques

Recent advancements in the Internet of Things and Machine Learning techniques have allowed the deployment of sensors on a large scale to monitor the environment and model and predict individual thermal comfort. The existing techniques have a greater focus on occupancy detection, estimations, and localization to balance energy usage and thermal comfort satisfaction. Different sensors, actuators, and analytic data methods are often non-invasively utilized to analyze data from occupant surroundings, identify occupant existence, estimate their numbers, and trigger the necessary action to complete a task. The efficiency of the non-invasive strategies documented in the literature, on the other hand, is rather poor due to the low quality of the datasets utilized in model training and the selection of machine learning technology. This study combines data from camera and environmental sensing using interactive learning and a rule-based classifier to improve the collection and quality of the datasets and data pre-processing. The study compiles a new comprehensive public set of training datasets for building occupancy profile prediction with over 40,000 records. To the best of our knowledge, it is the largest dataset to date, with the most realistic and challenging setting in building occupancy prediction. Furthermore, to the best of our knowledge, this is the first study that attained a robust occupancy count by considering a multimodal input to a single output regression model through the mining and mapping of feature importance, which has advantages over statistical approaches. The proposed solution is tested in a living room with a prototype system integrated with various sensors to obtain occupant-surrounding environmental datasets. The model’s prediction results indicate that the proposed solution can obtain data, and process and predict the occupants’ presence and their number with high accuracy values of 99.7% and 99.35%, respectively, using random forest.

The GENESIS-UV study on ultraviolet radiation exposure levels in 250 occupations to foster epidemiological and legislative efforts to combat nonmelanoma skin cancer.

Several hundred million of the 3 billion formally employed people worldwide are at risk from high levels of solar ultraviolet radiation (UVR). Chronic light damage to the skin can lead to nonmelanoma skin cancer (NMSC), especially when irradiation is too high and is acquired for decades. However, data with uniform metrics, high resolution over time and in-depth occupational profiles are not available. To build a worldwide usable matrix of UVR exposure in occupations and to show use cases for the data. One thousand test persons were recruited to wear electronic data logger dosimeters during their working time for 7 months each. The measurements yielded 3.7 billion data points for around 48 000 days with high-quality data capture covering more than 250 occupations and 650 activities. Scientific evaluation of the data included daily and half-hourly means, geographical transfer calculations to the world, threshold exceedance quotas, transcriptome effects, and occupational disease estimates. A compendium for global use is presented. In-depth analyses and the resulting implications for research have been elaborated to directly link exposure data to effects in the human body. Interestingly, the annual irradiances of the different occupations span a wide range of values: from about 650 to 50 standard erythemal doses, with different distributions over the months. Detailed exposure data per occupation were derived, and the risk on an occupational or activity basis assuming different exposure level quotas was quantified. This showed that, for example, in temperate latitudes, all activities with a duration of > 2 h outdoors are associated with increased NMSC risk. We offer our work to enable sound studies on the nature of ultraviolet-induced skin cancer, dose-response relationships, intermittency of skin exposure, and derivation of limit values. Sociological studies on prevention are now possible. Practitioners may use the findings for their daily work with employees.

Environmental DNA metabarcoding effectively monitors terrestrial species by using urban green spaces

Ecological corridors and urban parks are created to connect habitats that become isolated due to urban development and integrate ecological, social, and cultural functions. Establishing and implementing the ecological function of urban green structures and conservation plans for urban terrestrial species requires prioritizing developing an understanding of the community structure of urban terrestrial species. However, understanding the appearance and distribution of terrestrial species in urban ecosystems is often neglected or lacking; furthermore, the inaccurate detection of species distribution can create biased estimates of species occupancy. Recently, an applied research method for detecting terrestrial species such as mammals and birds has been proposed based on molecular biology such as environmental DNA (eDNA) metabarcoding. We installed artificial water tanks (mesocosm) in urban green spaces (Urban forests; UFs, Urban ecological corridors; UECs, Urban parks; UPs), applying eDNA metabarcoding as a tool to detect wild terrestrial species, and camera traps (CTs) and field surveys were performed in parallel to verify the results. We detected 22 urban terrestrial species only by eDNA metabarcoding and identified 43 species in total by performing eDNA metabarcoding, CTs, and field surveys in parallel. In addition, we identified 31 species of UFs, 13 species of UECs, and 28 species of UPs of urban terrestrial species using the three species survey methods in parallel. We confirmed that eDNA metabarcoding is very useful for detecting small species that are difficult to detect with the naked eye and internal species vulnerable to disturbance. However, since some species were only detected by CTs or field surveys due to DNA false-negative errors or the species’ life-history characteristics, it was judged that the species inventory can be effectively collected using several monitoring techniques.

Reconstructing compact building models from point clouds using deep implicit fields

While three-dimensional (3D) building models play an increasingly pivotal role in many real-world applications, obtaining a compact representation of buildings remains an open problem. In this paper, we present a novel framework for reconstructing compact, watertight, polygonal building models from point clouds. Our framework comprises three components: (a) a cell complex is generated via adaptive space partitioning that provides a polyhedral embedding as the candidate set; (b) an implicit field is learned by a deep neural network that facilitates building occupancy estimation; (c) a Markov random field is formulated to extract the outer surface of a building via combinatorial optimization. We evaluate and compare our method with state-of-the-art methods in generic reconstruction, model-based reconstruction, geometry simplification, and primitive assembly. Experiments on both synthetic and real-world point clouds have demonstrated that, with our neural-guided strategy, high-quality building models can be obtained with significant advantages in fidelity, compactness, and computational efficiency. Our method also shows robustness to noise and insufficient measurements, and it can directly generalize from synthetic scans to real-world measurements. The source code of this work is freely available at https://github.com/chenzhaiyu/points2poly.

Occupational demands associated with rotator cuff disease surgery in the UK Biobank.

Physically-demanding occupations may increase rotator cuff disease (RCD) risk and need for surgery. We linked a job-exposure matrix (JEM) to the UK Biobank cohort study to measure physical occupational exposures and estimate associations with RCD surgery. Jobs and UK Standard Occupational Classification (SOC) codes were recorded during the UK Biobank verbal interview. Lifetime job histories were captured through a web-based survey. UK SOC codes were linked to a JEM based on the US O*NET database. O*NET-based scores [static strength, dynamic strength, general physical activities, handling/moving objects (range=1-7), time spent using hands, whole body vibration, and cramped/awkward positions (range=1-5)] were assigned to jobs. RCD surgeries were identified through linked national hospital inpatient records. Multivariable Cox regression was used to calculate hazard ratios (HR) as estimates of associations with RCD surgery. Among those with lifetime job histories, associations were estimated for duration of time with greatest exposure (top quartile of exposure). Of 277 808 people reporting jobs, 1997 (0.7%) had an inpatient RCD surgery. After adjusting for age, sex, race, education, area deprivation, and body mass index, all O*NET variables considered were associated with RCD surgery (HR per point increase range=1.10-1.45, all P<0.005). A total of 100 929 people reported lifetime job histories, in which greater exposures were significantly associated with RCD surgery after >10 years of work (eg, HR for 11-20 versus 0 years with static strength score ≥4 = 2.06, 95% confidence interval 1.39-3.04). Workplace physical demands are an important risk factor for RCD surgery, particularly for workers with more than a decade of exposure.

Occupational characteristics associated with SARS-CoV-2 infection in the UK Biobank during August–November 2020: a cohort study

BackgroundOccupational exposures may play a key role in severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection risk. We used a job-exposure matrix linked to the UK Biobank to measure occupational characteristics and estimate associations with a positive SARS-CoV-2 test.MethodsPeople reporting job titles at their baseline interview in England who were < 65 years of age in 2020 were included. Healthcare workers were excluded because of differential access to testing. Jobs were linked to the US Occupational Information Network (O*NET) job exposure matrix. O*NET-based scores were examined for occupational physical proximity, exposure to diseases/infection, working outdoors exposed to weather, and working outdoors under cover (score range = 1–5). Jobs were classified as remote work using two algorithms. SARS-CoV-2 test results were evaluated between August 5th-November 10th, 2020, when the UK was released from lockdown. Cox regression was used to calculate adjusted hazard ratios (aHRs), accounting for age, sex, race, education, neighborhood deprivation, assessment center, household size, and income.ResultsWe included 115,451 people with job titles, of whom 1746 tested positive for SARS-CoV-2. A one-point increase in physical proximity score was associated with 1.14 times higher risk of SARS-CoV-2 (95%CI = 1.05–1.24). A one-point increase in the exposure to diseases/infections score was associated with 1.09 times higher risk of SARS-CoV-2 (95%CI = 1.02–1.16). People reporting jobs that could not be done remotely had higher risk of SARS-CoV-2 regardless of the classification algorithm used (aHRs = 1.17 and 1.20). Outdoors work showed an association with SARS-CoV-2 (exposed to weather aHR = 1.06, 95%CI = 1.01–1.11; under cover aHR = 1.08, 95%CI = 1.00–1.17), but these associations were not significant after accounting for whether work could be done remotely.ConclusionPeople in occupations that were not amenable to remote work, required closer physical proximity, and required more general exposure to diseases/infection had higher risk of a positive SARS-CoV-2 test. These findings provide additional evidence that coronavirus disease 2019 (COVID-19) is an occupational disease, even outside of the healthcare setting, and indicate that strategies for mitigating transmission in in-person work settings will remain important.

Conservation prioritization through combined approach of umbrella species selection, occupancy estimation, habitat suitability and connectivity analysis of kingfisher: A study from an internationally important wetland complex (Ramsar site) in India

Assessment of occupancy status, as well as projection of suitable habitats and connectivity of wetland indicator species, and thereby identification of potential conservation umbrella and projection of conservation priority areas are often considered important for wetland conservation. Kingfishers are wetland indicators and suffer from habitat degradation due to world-wide destruction of wetlands. Therefore, they can be considered potential candidates for conservation intervention. The present knowledge about the spatial distribution of suitable areas and habitat connectivity of kingfishers at a landscape level is non-existent. We conducted extensive surveys and recorded four kingfisher species in East Kolkata Wetlands (EKW; Ramsar site No. 1208; ∼125 km2). The occupancy estimates were highest for White-throated kingfisher (Halcyon smyrnensis, WTK), followed by common kingfisher (Alcedo atthis, CK), stork-billed kingfisher (Pelargopsis capensis, SBK) and lowest for pied kingfisher (Ceryle rudis, PK). WTK has the highest amount of suitable areas followed by CK, PK and SBK. The spatial overlap of suitable habitats showed that SBK is the potential umbrella species and therefore provides conservation benefits to other kingfisher species and eventually to the EKW. In addition to water areas, emergent vegetation, crop lands and tree cover are other important habitats for kingfishers. The connectivity analyses revealed that suitable habitats were disjunct and are under various anthropogenic threats. Therefore, we need to protect suitable habitats and connectivity between them. Finally, we identified conservation priority areas. Conservation intervention on these high priority zones will not only be beneficial for kingfishers, but also for other avifauna having similar resource requirements as well as the wetland parse.

First population estimates of two Critically Endangered frogs from an isolated forest plateau in Madagascar

AbstractIn the largely deforested areas of Madagascar, small forest fragments remain as last refuges for amphibian diversity. Isolated populations of the Critically Endangered Anodonthyla vallani and Anilany helenae persist in the fragmented forest of Ambohitantely but little information is available to inform their management and any conservation interventions. We generated estimates of population size and occupancy for both species in the largest fragment of Ambohitantely Special Reserve using acoustic survey data collected from 84 sites along 12 transects in December 2018. We used a single-season occupancy model to estimate detection and occupancy and a Royal–Nichols model to estimate abundance and population size. Anilany helenae and A. vallani had high occupancy rates (80 and 93%, respectively) whereas their detection rates differed (34 and 55%, respectively). Abundance and occupancy were best explained by vegetation structure whereas detection was influenced by time of survey and rainfall. For our sampled sites the estimated population sizes of males were 855 for A. vallani, with an estimated density of 52 individuals/ha, and 388 for A. helenae, with an estimated population density of 23 individuals/ha. Given their relatively low densities, small population sizes and restricted ranges, any further habitat loss could have drastic consequences for these populations. Our results provide guidance for future species-focused studies, and can inform conservation management at the local scale. Our work will help to improve species monitoring in Madagascar and elsewhere, especially for range-restricted non-charismatic amphibians.

A fusion framework for vision-based indoor occupancy estimation

Building occupancy information is essential for energy saving, comfort improvement, and security management. Existing vision-based indoor occupancy measurement methods have achieved remarkable progress; however, they mainly focus on single-vision situations, i.e., cameras at room entrances or interiors. These methods struggle to achieve high accuracy because of the complex indoor environments. For example, they often fail to detect occupants and generate many false positives. In this paper, to address these issues, we propose a novel fusion framework for occupancy detection and estimation based on two different perspectives. First, we design a head detection method combined with indoor scene knowledge to filter false positives and recover missed detection. Second, we propose a two-vision entrance counting method to refine the predicted results. Finally, we propose a cumulative error clearing strategy named dynamic Bayesian fusion (DBF), which integrates entrance counting and static estimation. Our framework achieves superior performance through ablation studies compared to existing methods on practical building surveillance videos, with occupancy estimation SCOREs of 99.2%, 98.5%, and 94.9%. Our framework can clear cumulative errors and stabilize estimation results. Practical experiments validate its potential for building energy saving and comfort improvement. The code is available at https://github.com/kailaisun/FFO.

Sensor Fusion for Occupancy Estimation: A Study Using Multiple Lecture Rooms in a Complex Building

This paper uses various machine learning methods which explore the combination of multiple sensors for quality improvement. It is known that a reliable occupancy estimation can help in many different cases and applications. For the containment of the SARS-CoV-2 virus, in particular, room occupancy is a major factor. The estimation can benefit visitor management systems in real time, but can also be predictive of room reservation strategies. By using different terminal and non-terminal sensors in different premises of varying sizes, this paper aims to estimate room occupancy. In the process, the proposed models are trained with different combinations of rooms in training and testing datasets to examine distinctions in the infrastructure of the considered building. The results indicate that the estimation benefits from a combination of different sensors. Additionally, it is found that a model should be trained with data from every room in a building and cannot be transferred to other rooms.

Designing drug occupancy studies with PET neuroimaging: Sample size, occupancy ranges and analytical methods

Molecular neuroimaging is today considered essential for evaluation of novel CNS drugs; it is used to quantify blood-brain barrier permeability, verify interaction with key target and determine the drug dose resulting in 50% occupancy, IC50. In spite of this, there has been limited data available to inform on how to optimize study designs.Through simulations, we here evaluate how IC50 estimation is affected by the (i) range of drug doses administered, (ii) number of subjects included, and (iii) level of noise in the plasma drug concentration measurements. Receptor occupancy is determined from PET distribution volumes using two different methods: the Lassen plot and Likelihood estimation of occupancy (LEO). We also introduce and evaluate a new likelihood-based estimator for direct estimation of IC50 from PET distribution volumes.For estimation of IC50, we find very limited added benefit in scanning individuals who are given drug doses corresponding to less than 40% receptor occupancy. In the range of typical PET sample sizes (5–20 subjects) each extra individual clearly reduces the error of the IC50 estimate.In all simulations, likelihood-based methods gave more precise IC50 estimates than the Lassen plot; four times the number of subjects were required for the Lassen plot to reach the same IC50 precision as LEO.

Estimating species richness with camera traps: modeling the effects of delay period, deployment length, number of sites, and interference imagery

AbstractBiologists commonly use camera traps for estimating species richness to inform conservation actions, steer land protection, and reveal effects of climate change. Long‐term studies using short delay periods (≤1 min) and numerous cameras produce voluminous amounts of redundant imagery. Thus, camera‐trapping procedures maximizing richness estimates while minimizing data collection need development. We used imagery of mammals spanning 4 deserts in the United States to model the effects of delay, deployment length (i.e., study duration), number of sampling sites, and interference events on the proportion of known species richness detected (Rp). We also determined the proportion of subsamples containing each species (SR) under different sampling conditions to inform subsequent occupancy estimation. We generated contour plots describing the optimal configuration of sites and deployment length that minimized the image acquisition required to estimate Rp = 0.9. The optimal configuration was independent of delay (requiring ~50 sites and 13 months). The shortest delay (10 sec) generated ~8 times more images than the longest (3600 sec) without substantially improving Rp and rare species detection. The shortest duration to acquire Rp = 0.9 was 10 months but required ~70 sites. The fewest sites needed were 22 and 29, depending on camera placement, requiring approximately 50 months of deployment. Simulated short, one‐month studies were only able to obtain Rp ~0.6 with 40–70 sites. Obtaining SR = 0.8 with a 3600 sec delay required between 1–12 months and 10 sites or 1–17 sites and 6 months for uncommon species. Adding interference imagery, even with long delays, produced SR ≥ 0.5 for rare species, generating data suitable for occupancy estimation. Overall, interference imagery had minimal effects on reducing SR estimates, unless the interference occurred continuously. Our guidance optimizes the number of sites, deployment length, and delay period while minimizing imagery acquisition to meet Rp and occupancy objectives with confidence.

Using static postures to estimate spinal loading during dynamic lifts with participant-specific thoracolumbar musculoskeletal models

Static biomechanical simulations are sometimes used to estimate in vivo kinetic demands because they can be solved efficiently, but this ignores any potential inertial effects. To date, comparisons between static and dynamic analyses of spinal demands have been limited to lumbar joint differences in young males performing sagittal lifts. Here we compare static and dynamic vertebral compressive and shear force estimates during axial, lateral, and sagittal lifting tasks across all thoracic and lumbar vertebrae in older men and women. Participant-specific thoracolumbar full-body musculoskeletal models estimated vertebral forces from recorded kinematics both with and without consideration of dynamic effects, at an identified frame of peak vertebral loading. Static analyses under-predicted dynamic compressive and resultant shear forces, by an average of about 16% for all three lifts across the thoracic and lumbar spine but were highly correlated with dynamic forces (average r2 > .95). The study outcomes have the potential to enable standard clinical and occupational estimates using static analyses.

The development of a convolutional neural network for the automatic detection of Northern Bobwhite Colinus virginianus covey calls

AbstractPassive acoustic monitoring using Autonomous Recording Units (ARUs) is becoming a significant research tool for collecting large amounts of ecological data. Northern bobwhite Colinus virginianus is an economically important game bird whose declining populations are of conservation concern, so efforts to monitor bobwhite abundance using ARUs are being intensified. Yet, manual processing of ARU data is time consuming and often expensive, so developing automatic call detection methods is a key step in acoustic monitoring. We present here the first single species convolutional neural network (CNN) developed purely for automatic bobwhite covey call identification and classification. We demonstrate the value of meaningful data augmentation by including non‐target calls and background noise into our training dataset, as well as evaluating alternative CNN score thresholds and model extrapolation performance. We trained our CNN on 6,682 manually labeled covey calls across three groups of sites within the southeastern USA. Precision and AUC from both CNN classification and individual call detection was high (0.80–0.99), and our model showed strong extrapolation ability across site groups. However, extrapolation performance significantly decreased for sites that were more dissimilar to the training data set if our meaningful data augmentation process was omitted. Our CNN detected significantly more covey calls than manual labeling using Raven Pro software, and processing time was greatly reduced: a single one hour wav file can be now analyzed by the CNN in roughly eight seconds. We also demonstrate using a simple case study that extremely high variability in estimates of bobwhite site occupancy and detection are obtained depending on the method of acoustic data processing (manual versus CNN). Our results suggest that our CNN provides robust and time‐saving analysis of bobwhite covey call acoustic data and can be applied to future research and monitoring projects with high confidence in the performance of the model.

Occupancy estimation in smart buildings using predictive modeling in imbalanced domains

Occupancy estimation in smart buildings using predictive modeling in imbalanced domains