Effect of detection heterogeneity in occupancy‐detection models: an experimental test of time‐to‐first‐detection methods

Imperfect detection can bias estimates of site occupancy in ecological surveys but can be corrected by estimating detection probability. Time‐to‐first‐detection (TTD) occupancy models have been proposed as a cost–effective survey method that allows detection probability to be estimated from single site visits. Nevertheless, few studies have validated the performance of occupancy‐detection models by creating a situation where occupancy is known, and model outputs can be compared with the truth. We tested the performance of TTD occupancy models in the face of detection heterogeneity using an experiment based on standard survey methods to monitor koala Phascolarctos cinereus populations in Australia. Known numbers of koala faecal pellets were placed under trees, and observers, uninformed as to which trees had pellets under them, carried out a TTD survey. We fitted five TTD occupancy models to the survey data, each making different assumptions about detectability, to evaluate how well each estimated the true occupancy status. Relative to the truth, all five models produced strongly biased estimates, overestimating detection probability and underestimating the number of occupied trees. Despite this, goodness‐of‐fit tests indicated that some models fitted the data well, with no evidence of model misfit. Hence, TTD occupancy models that appear to perform well with respect to the available data may be performing poorly. The reason for poor model performance was unaccounted for heterogeneity in detection probability, which is known to bias occupancy‐detection models. This poses a problem because unaccounted for heterogeneity could not be detected using goodness‐of‐fit tests and was only revealed because we knew the experimentally determined outcome. A challenge for occupancy‐detection models is to find ways to identify and mitigate the impacts of unobserved heterogeneity, which could unknowingly bias many models.     

The impact of a fox‐ and cat‐free safe haven on the bird fauna of remnant vegetation in southwestern Australia

Introduced predators are a serious threat to Australian vertebrates. However, the consequences of predation for an area's avifauna have rarely been quantified. We took advantage of the establishment of a 7,832 ha fox‐ and cat‐free safe haven at Mt Gibson, in Western Australia, to assess the consequences of excluding introduced mammal predators on the bird fauna. Bird surveys were conducted over 6 years, before and after the establishment of the introduced predator‐free safe haven. After 3 years, half the sites were enclosed by the fence that excluded introduced predators, while the remainder of sites remained outside the fence and were exposed to fox and cat activity. The sites were stratified by four major vegetation types. A total of 91 bird species were variously detectable with the survey approach, but were typically more detectable during morning surveys. Site occupancy varied considerably among species, but overall, occupancy by all species was most likely to be either not impacted or positively impacted by the safe haven. The most notable change was that avifaunal richness appeared to increase in woodland and shrubland habitats within, as compared to outside, the safe haven. We conclude that: (1) the safe haven had an overall positive impact on bird occupancy; and (2) there were no consistent trends with respect to the kinds of species whose occupancy was positively impacted, beyond them all being small‐ to medium‐sized birds and mostly insectivorous. However, these conclusions must be tempered by the poor detection probability of many species.     

Responses of Primates to Landscape Change in Amazonian Land‐bridge islands—a Multi‐scale Analysis

Large hydroelectric dams are one of the current drivers of habitat loss across Amazonian forests. We investigated how the primate community at a hydroelectric dam in Brazilian Amazonia responded to changes in the landscape and local habitat structure of land‐bridge islands after 21 yr of post‐isolation history. The Balbina Dam, constructed in 1986, inundated 3129 km² of primary forests and created more than 3500 variable‐sized islands. We conducted primate and habitat structure surveys on 20 islands from 5 to 1815 ha, and extracted forest patch and landscape metrics for each island. The number of primate species per island varied between 0 and 7 species. Primate composition varied substantially according to both island area and forest cover remaining within the landscape, whereas island area alone was the most significant predictor of richness. Locally, tree density and vertical stratification were the most significant explanatory variables of primate composition and richness. A model containing area effects had the most explanatory power regarding site occupancy for most species. Individually, each species responded differently, with howler and brown capuchin monkeys showing greater tolerance to cope with habitat changes. Body size was also an important predictor of primate occupancy. We recommend protecting large fragments and enhancing the suitability of surrounding habitats to ensure primate conservation in most Neotropical fragmented landscapes. Given the flat topography of hydroelectric reservoirs, which mainly favors the formation of small islands, and the escalating hydropower development plans in Amazonia, our findings provide evidence for pervasive detrimental impacts of dams on primate communities.     

A goodness‐of‐fit test for occupancy models with correlated within‐season revisits

Occupancy modeling is important for exploring species distribution patterns and for conservation monitoring. Within this framework, explicit attention is given to species detection probabilities estimated from replicate surveys to sample units. A central assumption is that replicate surveys are independent Bernoulli trials, but this assumption becomes untenable when ecologists serially deploy remote cameras and acoustic recording devices over days and weeks to survey rare and elusive animals. Proposed solutions involve modifying the detection‐level component of the model (e.g., first‐order Markov covariate). Evaluating whether a model sufficiently accounts for correlation is imperative, but clear guidance for practitioners is lacking. Currently, an omnibus goodness‐of‐fit test using a chi‐square discrepancy measure on unique detection histories is available for occupancy models (MacKenzie and Bailey, Journal of Agricultural, Biological, and Environmental Statistics, 9, 2004, 300; hereafter, MacKenzie–Bailey test). We propose a join count summary measure adapted from spatial statistics to directly assess correlation after fitting a model. We motivate our work with a dataset of multinight bat call recordings from a pilot study for the North American Bat Monitoring Program. We found in simulations that our join count test was more reliable than the MacKenzie–Bailey test for detecting inadequacy of a model that assumed independence, particularly when serial correlation was low to moderate. A model that included a Markov‐structured detection‐level covariate produced unbiased occupancy estimates except in the presence of strong serial correlation and a revisit design consisting only of temporal replicates. When applied to two common bat species, our approach illustrates that sophisticated models do not guarantee adequate fit to real data, underscoring the importance of model assessment. Our join count test provides a widely applicable goodness‐of‐fit test and specifically evaluates occupancy model lack of fit related to correlation among detections within a sample unit. Our diagnostic tool is available for practitioners that serially deploy survey equipment as a way to achieve cost savings.     

Statistical approaches to account for false‐positive errors in environmental DNA samples

Environmental DNA (eDNA) sampling is prone to both false‐positive and false‐negative errors. We review statistical methods to account for such errors in the analysis of eDNA data and use simulations to compare the performance of different modelling approaches. Our simulations illustrate that even low false‐positive rates can produce biased estimates of occupancy and detectability. We further show that removing or classifying single PCR detections in an ad hoc manner under the suspicion that such records represent false positives, as sometimes advocated in the eDNA literature, also results in biased estimation of occupancy, detectability and false‐positive rates. We advocate alternative approaches to account for false‐positive errors that rely on prior information, or the collection of ancillary detection data at a subset of sites using a sampling method that is not prone to false‐positive errors. We illustrate the advantages of these approaches over ad hoc classifications of detections and provide practical advice and code for fitting these models in maximum likelihood and Bayesian frameworks. Given the severe bias induced by false‐negative and false‐positive errors, the methods presented here should be more routinely adopted in eDNA studies.     

Strong but taxon‐specific responses of termites and wood‐nesting ants to forest regeneration in Borneo

Land use change is accelerating globally at the expense of biodiversity and ecosystem functioning. Invertebrates are numerically dominant and functionally important in old growth tropical rain forests but highly susceptible to the adverse effects of forest degradation and fragmentation. Ants (Formicidae) and termites (Blattodea: Termitoidae) perform crucial ecosystem services. Here, the potential effects of anthropogenic disturbance on ant and termite communities in dead wood are investigated. Community composition, generic richness, and occupancy rates of ants and termites were compared among two old growth sites (Danum Valley and Maliau Basin) and one twice‐logged site (the Stability of Altered Forest Ecosystems’ (SAFE) Project), in Sabah, Malaysian Borneo. Occupancy was measured as the number of ant or termite encounters (1) per deadwood items, and (2) per deadwood volume, and acts as surrogates for relative abundance (or generic richness). Termites had a lower wood‐occupancy per volume in logged forest. In contrast, there were more ant encounters, and more ant genera, in logged sites and there was a community shift (especially, there were more Crematogaster encounters). The disruption of soil and canopy structure in logged forest may reduce both termite and fungal decay rates, inducing increased deadwood residence times and therefore favoring ants that nest in dead wood. There is an anthropogenic‐induced shift of dead wood in ants and termites in response to disturbance in tropical rain forests and the nature of that shift is taxon‐specific.     

Scale‐dependent strategies for coexistence of mesocarnivores in human‐dominated landscapes

Identifying factors influencing the distribution of and interactions within carnivore communities is important for understanding how they are affected by human activities. Species differ in their ability to adapt to humans depending on their degree of specialization in habitat use and feeding habits. This results in asymmetric changes in the ecology of co‐occurring species that can influence their interactions. We investigated whether human infrastructures and free‐ranging domestic dogs (a species typically associated with humans) influenced the co‐occurrence and habitat use of mesocarnivores in a landscape of high human population density in Maharashtra, India. We used 40 camera trap locations during 233 trapping nights and used Bayesian co‐occurrence occupancy models to investigate the habitat use and coexistence of species at different spatial scales. Additionally, we investigated their temporal overlap in space use. Indian foxes altered their habitat use both spatially and temporally in order to avoid free‐ranging domestic dogs and other larger competitors. The use of human infrastructure by jackals and jungle cats was limited by the presence of dogs. Our results illustrate how habitat use of smaller carnivore species changes both spatially and temporally in order to avoid larger competitors. We also show that the presence of species associated with humans mediates the influence of human infrastructures on the habitat use of mesocarnivores. We highlight the importance of acknowledging the potential impact of urbanization not only on single species, but also on the interactions within the community.     

Identification of cell culture conditions to control N‐glycosylation site‐occupancy of recombinant glycoproteins expressed in CHO cells

The effect of different cell culture conditions on N‐glycosylation site‐occupancy has been elucidated for two different recombinant glycoproteins expressed in Chinese hamster ovary (CHO) cells, recombinant human tissue plasminogen activator (t‐PA) and a recombinant enzyme (glycoprotein 2—GP2). Both molecules contain a N‐glycosylation site that is variably occupied. Different environmental factors that affect the site‐occupancy (the degree of occupied sites) of these molecules were identified. Supplementing the culture medium with additional manganese or iron increased the fraction of fully occupied t‐PA (type I t‐PA) by approximately 2.5–4%. Decreasing the cultivation temperature from 37 to 33°C or 31°C gradually increased site‐occupancy of t‐PA up to 4%. The addition of a specific productivity enhancer, butyrate, further increased site‐occupancy by an additional 1% under each cultivation temperature tested. In addition, the thyroid hormones triiodothyronine and thyroxine increased site‐occupancy of t‐PA compared to control conditions by about 2%. In contrast, the addition of relevant nucleoside precursor molecules involved in N‐glycan biosynthesis (e.g., uridine, guanosine, mannose) either had no effect or slightly reduced site‐occupancy. For the recombinant enzyme (GP2), it was discovered that culture pH and the timing of butyrate addition can be used to control N‐glycan site‐occupancy within a specific range. An increase in culture pH correlated with a decrease in site‐occupancy. Similarly, delaying the timing for butyrate addition also decreased site‐occupancy of this molecule. These results highlight the importance of understanding how cell culture conditions and media components can affect the product quality of recombinant glycoproteins expressed in mammalian cell cultures. Furthermore, the identification of relevant factors will enable one to control product quality attributes, specifically N‐glycan site‐occupancy, within a specific range when applied appropriately. Biotechnol. Bioeng. 2009;103: 1164–1175. © 2009 Wiley Periodicals, Inc.     

Taking the bait: The influence of attractants and microhabitat on detections of fauna by remote‐sensing cameras

Autonomously triggered cameras are a common wildlife survey technique. The use of attractants and surrounding microhabitats is likely to influence detection probabilities and survey outcomes; however, few studies consider these factors. We compared three attractants (peanut butter‐based, tuna‐based and a control) in a Latin square design through a coastal shrubland with high microhabitat variability at Cape Otway, Victoria, Australia (38º50?S, 143º30?E). Deployments involved 36 cameras for four days in each of five years. The percentage cover of each vegetation structural type (low [no or sparse cover], moderate [grass] or high [shrubs]) within 20 m of each camera was calculated and reduced to a single variable using PCA. Dynamic occupancy modelling, with lure type and vegetation structure as covariates of detection probability, found that peanut butter attracted the greatest diversity of species (24 of 35 species, 69%) and yielded the greatest number of detections (50% of 319) when compared with tuna oil (66% and 24%, respectively) and the control (43% and 26%, respectively). Peanut butter attracted more Macropodidae (wallabies) and Muridae (rats and mice); however, vegetation structural variables were the greatest influence on Corvidae/Artamidae (raven/currawong) detections with higher detectability in more open areas. Vegetation structure also influenced Muridae detections. This study reinforces the critical choice of appropriate attractants and camera placement when investigating vertebrate groups and highlights the role of microhabitat in the detection of small mammals and birds. We suggest future large‐scale camera surveys consider different bait types and microhabitats in their designs, to control for any biases and enable future advice on ‘optimal’ methods.     

Species‐specific and temporal scale‐dependent responses of birds to drought

Global climate change is increasing the frequency and intensity of weather extremes, including severe droughts in many regions. Drought can impact organisms by inhibiting reproduction, reducing survival and abundance, and forcing range shifts. For birds, considering temporal scale by averaging drought‐related variables over different time lengths (i.e., temporal grains) captures different hydrologic attributes which may uniquely influence food supplies, vegetation greenness/structure, and other factors affecting populations. However, studies examining drought impacts on birds often assess a single temporal grain without considering that different species have different life histories that likely determine the temporal grain of their drought response. Furthermore, while drought is known to influence bird abundance and drive between‐year range shifts, less understood is whether it causes within‐range changes in species distributions. Our objectives were to (a) determine which temporal grain of drought (if any) is most related to bird presence/absence and whether this response is species specific; and (b) assess whether drought alters bird distributions by quantifying probability of local colonization and extinction as a function of drought intensity. We used North American Breeding Bird Survey data collected over 16 years, generalized linear mixed models, and dynamic occupancy models to meet these objectives. Different bird species responded to drought at different temporal grains, with most showing the strongest signal at annual or near‐annual grains. For all drought‐responsive species, increased drought intensity at any temporal grain always correlated with decreased occupancy. Additionally, colonization/extinction analyses indicated that one species, the dickcissel (Spiza americana), is more likely to colonize novel areas within the southern/core portion of its range during drought. Considering drought at different temporal grains, along with hydrologic attributes captured by each grain, may better reveal mechanisms behind drought impacts on birds and other organisms, and therefore improve understanding of how global climate change impacts species and the landscapes they inhabit.     

Co‐occurrence dynamics of endangered Lower Keys marsh rabbits and free‐ranging domestic cats: Prey responses to an exotic predator removal program

The Lower Keys marsh rabbit (Sylvilagus palustris hefneri) is one of many endangered endemic species of the Florida Keys. The main threats are habitat loss and fragmentation from sea‐level rise, development, and habitat succession. Exotic predators such as free‐ranging domestic cats (Felis catus) pose an additional threat to these endangered small mammals. Management strategies have focused on habitat restoration and exotic predator control. However, the effectiveness of predator removal and the effects of anthropogenic habitat modifications and restoration have not been evaluated. Between 2013 and 2015, we used camera traps to survey marsh rabbits and free‐ranging cats at 84 sites in the National Key Deer Refuge, Big Pine Key, Florida, USA. We used dynamic occupancy models to determine factors associated with marsh rabbit occurrence, colonization, extinction, and the co‐occurrence of marsh rabbits and cats during a period of predator removal. Rabbit occurrence was positively related to freshwater habitat and patch size, but was negatively related to the number of individual cats detected at each site. Furthermore, marsh rabbit colonization was negatively associated with relative increases in the number of individual cats at each site between survey years. Cat occurrence was negatively associated with increasing distance from human developments. The probability of cat site extinction was positively related to a 2‐year trapping effort, indicating that predator removal reduced the cat population. Dynamic co‐occurrence models suggested that cats and marsh rabbits co‐occur less frequently than expected under random conditions, whereas co‐detections were site and survey‐specific. Rabbit site extinction and colonization were not strongly conditional on cat presence, but corresponded with a negative association. Our results suggest that while rabbits can colonize and persist at sites where cats occur, it is the number of individual cats at a site that more strongly influences rabbit occupancy and colonization. These findings indicate that continued predator management would likely benefit endangered small mammals as they recolonize restored habitats.     

Economics of membrane occupancy and respiro‐fermentation

The simultaneous utilization of efficient respiration and inefficient fermentation even in the presence of abundant oxygen is a puzzling phenomenon commonly observed in bacteria, yeasts, and cancer cells. Despite extensive research, the biochemical basis for this phenomenon remains obscure. We hypothesize that the outcome of a competition for membrane space between glucose transporters and respiratory chain (which we refer to as economics of membrane occupancy) proteins influences respiration and fermentation. By incorporating a sole constraint based on this concept in the genome‐scale metabolic model of Escherichia coli, we were able to simulate respiro‐fermentation. Further analysis of the impact of this constraint revealed differential utilization of the cytochromes and faster glucose uptake under anaerobic conditions than under aerobic conditions. Based on these simulations, we propose that bacterial cells manage the composition of their cytoplasmic membrane to maintain optimal ATP production by switching between oxidative and substrate‐level phosphorylation. These results suggest that the membrane occupancy constraint may be a fundamental governing constraint of cellular metabolism and physiology, and establishes a direct link between cell morphology and physiology.     

Factors influencing the occurrence of the Southern Long‐nosed Bandicoot (Perameles nasuta Geoffroy) during a population irruption and decline

Modelling the occupancy of species is a key part of the discipline of ecology, with particular efforts often focused on identifying which environmental, vegetation and other factors influence why a given species occurs where it does. Here, based on data gathered between 2003 and 2016, we developed models of the environmental, terrain and vegetation factors associated with site occupancy of the Southern Long‐nosed Bandicoot (Perameles nasuta Geoffroy) at Booderee National Park in the Jervis Bay Territory, south‐eastern Australia. The Long‐nosed Bandicoot is a nocturnal omnivorous marsupial which feeds primarily on subterranean invertebrates and fungi. The species has undergone a major reduction in abundance and site occupancy following a peak irruption in 2006 with the percentage of the sites at which the species was present declining from 74.3% of sites trapped in 2006 to 10.5% of sites trapped in 2016. We found the Long‐nosed Bandicoot was distributed widely across Booderee National Park and occurred in all six broad vegetation types in the reserve (forest, heathland, woodland, shrubland, sedgeland and rainforest). Detection‐occupancy modelling revealed that the persistence and colonisation of sites by the species were negatively associated with the per cent cover of leaf litter, irrespective of broad vegetation type. Occupancy models are an important tool in identifying parts of landscapes most likely to support populations of particular species, such as the Long‐nosed Bandicoot over time, and they may assist management of protected areas to prioritise resources to manage the habitat of those areas.     

Improving the accuracy of population size estimates for burrow‐nesting seabirds

Seabird numbers can change rapidly as a result of environmental processes, both natural and anthropogenic. Informed management and conservation of seabirds requires accurate and precise monitoring of population size. However, for burrow‐nesting species this is rarely achieved due to spatial and temporal heterogeneity in burrow occupancy. Here, we describe a novel method for deriving more accurate population size estimates that employs mark‐recapture methods to correct for unknown variation in nest occupancy throughout a breeding season. We apply it to estimate breeding numbers of a colonial, burrowing seabird, the Little Penguin Eudyptula minor, on the Summerland Peninsula, Phillip Island, Australia. Estimates of active burrow numbers during the September 2008 to February 2009 breeding season were adjusted to numbers of breeding birds based on burrow occupation and modelled population demographics at six, fortnightly monitored reference sites. The population was estimated to be 26 100 (95% CI: 21 100–31 100) and 28 400 (23 800–33 000) breeding Penguins in two temporally separated surveys within one breeding season. We demonstrate using simulation that the method is robust to variation in burrow occupancy throughout the breeding season, providing consistent and more accurate estimates of population size. The advantage of using the corrected method is that confidence intervals will include the true population size. Confidence limits widened as burrow occupancy declined, reflecting the increased uncertainty as larger adjustments for low burrow occupancy were required. In contrast, the uncorrected method that uses burrow occupancy alone as a measure of breeding numbers was inconsistent and significantly underestimated population size across much of the breeding season. Although requiring considerably more survey effort, the corrected approach provides a more accurate means for monitoring population changes in colonially breeding animals while collecting demographic data that can help diagnose the drivers of population change.     

Declining old‐forest species as a legacy of large trees lost

Aim

Global declines in large old trees from selective logging have degraded old‐forest ecosystems, which could lead to delayed declines or losses of old‐forest‐associated wildlife populations (i.e., extinction debt). We applied the declining population paradigm and explored potential evidence for extinction debt in an old‐forest dependent species across landscapes with different histories of large tree logging.

Location

Montane forests of the Sierra Nevada, California, USA.

Methods

We tested hypotheses about the influence of forest structure on territory extinction dynamics of the spotted owl (Strix occidentalis) using detection/non‐detection data from 1993 to 2011 across two land tenures: national forests, which experienced extensive large tree logging over the past century, and national parks, which did not.

Results

Large tree/high canopy cover forest was the best predictor of extinction rates and explained 26%–77% of model deviance. Owl territories with more large tree/high canopy cover forest had lower extinction rates, and this forest type was ~4 times more prevalent within owl territories in national parks ( = 19% of territory) than national forests ( = 4% of territory). As such, predicted extinction probability for an average owl territory was ~2.5 times greater in national forests than national parks, where occupancy was declining () and stable (), respectively. Large tree/high canopy cover forest remained consistently low, but did not decline, during the study period on national forests while owl declines were ongoing—an observation consistent with an extinction debt.

Main conclusions

In identifying a linkage between large trees and spotted owl dynamics at a regional scale, we provide evidence suggesting past logging of large old trees may have contributed to contemporary declines in an old‐forest species. Strengthening protections for remaining large old trees and promoting their recruitment in the future will be critical for biodiversity conservation in the world's forests.