Open population maximum likelihood spatial capture‐recapture

Open population capture‐recapture models are widely used to estimate population demographics and abundance over time. Bayesian methods exist to incorporate open population modeling with spatial capture‐recapture (SCR), allowing for estimation of the effective area sampled and population density. Here, open population SCR is formulated as a hidden Markov model (HMM), allowing inference by maximum likelihood for both Cormack‐Jolly‐Seber and Jolly‐Seber models, with and without activity center movement. The method is applied to a 12‐year survey of male jaguars (Panthera onca) in the Cockscomb Basin Wildlife Sanctuary, Belize, to estimate survival probability and population abundance over time. For this application, inference is shown to be biased when assuming activity centers are fixed over time, while including a model for activity center movement provides negligible bias and nominal confidence interval coverage, as demonstrated by a simulation study. The HMM approach is compared with Bayesian data augmentation and closed population models for this application. The method is substantially more computationally efficient than the Bayesian approach and provides a lower root‐mean‐square error in predicting population density compared to closed population models.     

Bayesian model selection for spatial capture–recapture models

A vast amount of ecological knowledge generated over the past two decades has hinged upon the ability of model selection methods to discriminate among various ecological hypotheses. The last decade has seen the rise of Bayesian hierarchical models in ecology. Consequently, commonly used tools, such as the AIC, become largely inapplicable and there appears to be no consensus about a particular model selection tool that can be universally applied. We focus on a specific class of competing Bayesian spatial capture–recapture (SCR) models and apply and evaluate some of the recommended Bayesian model selection tools: (1) Bayes Factor—using (a) Gelfand‐Dey and (b) harmonic mean methods, (2) Deviance Information Criterion (DIC), (3) Watanabe‐Akaike's Information Criterion (WAIC) and (4) posterior predictive loss criterion. In all, we evaluate 25 variants of model selection tools in our study. We evaluate these model selection tools from the standpoint of selecting the “true” model and parameter estimation. In all, we generate 120 simulated data sets using the true model and assess the frequency with which the true model is selected and how well the tool estimates N (population size), a parameter of much importance to ecologists. We find that when information content is low in the data, no particular model selection tool can be recommended to help realize, simultaneously, both the goals of model selection and parameter estimation. But, in general (when we consider both the objectives together), we recommend the use of our application of the Bayes Factor (Gelfand‐Dey with MAP approximation) for Bayesian SCR models. Our study highlights the point that although new model selection tools are emerging (e.g., WAIC) in the applied statistics literature, those tools based on sound theory even under approximation may still perform much better.     

Estimation of demo-genetic model probabilities with Approximate Bayesian Computation using linear discriminant analysis on summary statistics

Comparison of demo‐genetic models using Approximate Bayesian Computation (ABC) is an active research field. Although large numbers of populations and models (i.e. scenarios) can be analysed with ABC using molecular data obtained from various marker types, methodological and computational issues arise when these numbers become too large. Moreover, Robert et al. (Proceedings of the National Academy of Sciences of the United States of America, 2011, 108, 15112) have shown that the conclusions drawn on ABC model comparison cannot be trusted per se and required additional simulation analyses. Monte Carlo inferential techniques to empirically evaluate confidence in scenario choice are very time‐consuming, however, when the numbers of summary statistics (Ss) and scenarios are large. We here describe a methodological innovation to process efficient ABC scenario probability computation using linear discriminant analysis (LDA) on Ss before computing logistic regression. We used simulated pseudo‐observed data sets (pods) to assess the main features of the method (precision and computation time) in comparison with traditional probability estimation using raw (i.e. not LDA transformed) Ss. We also illustrate the method on real microsatellite data sets produced to make inferences about the invasion routes of the coccinelid Harmonia axyridis. We found that scenario probabilities computed from LDA‐transformed and raw Ss were strongly correlated. Type I and II errors were similar for both methods. The faster probability computation that we observed (speed gain around a factor of 100 for LDA‐transformed Ss) substantially increases the ability of ABC practitioners to analyse large numbers of pods and hence provides a manageable way to empirically evaluate the power available to discriminate among a large set of complex scenarios.     

Long‐term panmixia in a cosmopolitan Indo‐Pacific coral reef fish and a nebulous genetic boundary with its broadly sympatric sister species

Phylogeographical studies have shown that some shallow‐water marine organisms, such as certain coral reef fishes, lack spatial population structure at oceanic scales, despite vast distances of pelagic habitat between reefs and other dispersal barriers. However, whether these dispersive widespread taxa constitute long‐term panmictic populations across their species ranges remains unknown. Conventional phylogeographical inferences frequently fail to distinguish between long‐term panmixia and metapopulations connected by gene flow. Moreover, marine organisms have notoriously large effective population sizes that confound population structure detection. Therefore, at what spatial scale marine populations experience independent evolutionary trajectories and ultimately species divergence is still unclear. Here, we present a phylogeographical study of a cosmopolitan Indo‐Pacific coral reef fish Naso hexacanthus and its sister species Naso caesius, using two mtDNA and two nDNA markers. The purpose of this study was two‐fold: first, to test for broad‐scale panmixia in N. hexacanthus by fitting the data to various phylogeographical models within a Bayesian statistical framework, and second, to explore patterns of genetic divergence between the two broadly sympatric species. We report that N. hexacanthus shows little population structure across the Indo‐Pacific and a range‐wide, long‐term panmictic population model best fit the data. Hence, this species presently comprises a single evolutionary unit across much of the tropical Indian and Pacific Oceans. Naso hexacanthus and N. caesius were not reciprocally monophyletic in the mtDNA markers but showed varying degrees of population level divergence in the two nuclear introns. Overall, patterns are consistent with secondary introgression following a period of isolation, which may be attributed to oceanographic conditions of the mid to late Pleistocene, when these two species appear to have diverged.     

Efficient super‐resolution volumetric imaging by radial fluctuation Bayesian analysis light‐sheet microscopy

Various computational super‐resolution methods are available based on the analysis of fluorescence fluctuation behind acquired frames. However, dilemmas often exist in the balance of fluorophore characteristics, computation cost, and achievable resolution. Here we present an approach that uses a super‐resolution radial fluctuations (SRRF) image to guide the Bayesian analysis of fluorophore blinking and bleaching (3B) events, allowing greatly accelerated localization of overlapping fluorophores with high accuracy. This radial fluctuation Bayesian analysis (RFBA) approach is also extended to three dimensions for the first time and combined with light‐sheet fluorescence microscopy, to achieve super‐resolution volumetric imaging of thick samples densely labeled with common fluorophores. For example, a 700‐nm thin Bessel plane illumination is developed to optically section the Drosophila brain, providing a high‐contrast 3D image of rhythmic neurons. RFBA analyzes 30 serial volumes to reconstruct a super‐resolved 3D image at 4‐times higher resolutions (~70 and 170 nm), and precisely resolve the axon terminals. The computation is over 2‐orders faster than conventional 3B analysis microscopy. The capability of RFBA is also verified through dual‐color imaging of cell nucleus in live Drosophila brain. The spatial co‐localization patterns of the nuclear envelope and DNA in a neuron deep inside the brain can be precisely extracted by our approach.     

Dispersal ability and habitat requirements determine landscape‐level genetic patterns in desert aquatic insects

Species occupying the same geographic range can exhibit remarkably different population structures across the landscape, ranging from highly diversified to panmictic. Given limitations on collecting population‐level data for large numbers of species, ecologists seek to identify proximate organismal traits—such as dispersal ability, habitat preference and life history—that are strong predictors of realized population structure. We examined how dispersal ability and habitat structure affect the regional balance of gene flow and genetic drift within three aquatic insects that represent the range of dispersal abilities and habitat requirements observed in desert stream insect communities. For each species, we tested for linear relationships between genetic distances and geographic distances using Euclidean and landscape‐based metrics of resistance. We found that the moderate‐disperser Mesocapnia arizonensis (Plecoptera: Capniidae) has a strong isolation‐by‐distance pattern, suggesting migration–drift equilibrium. By contrast, population structure in the flightless Abedus herberti (Hemiptera: Belostomatidae) is influenced by genetic drift, while gene flow is the dominant force in the strong‐flying Boreonectes aequinoctialis (Coleoptera: Dytiscidae). The best‐fitting landscape model for M. arizonensis was based on Euclidean distance. Analyses also identified a strong spatial scale‐dependence, where landscape genetic methods only performed well for species that were intermediate in dispersal ability. Our results highlight the fact that when either gene flow or genetic drift dominates in shaping population structure, no detectable relationship between genetic and geographic distances is expected at certain spatial scales. This study provides insight into how gene flow and drift interact at the regional scale for these insects as well as the organisms that share similar habitats and dispersal abilities.     

Comparing clustered sampling designs for spatially explicit estimation of population density

Spatially explicit capture–recapture methods do not assume that animals have equal access to sampling devices (e.g., detectors), which allows for gaps in the sampling extent and nonuniform (e.g., clustered) sampling designs. However, the performance (i.e., relative root mean squared error [RRMSE], confidence interval coverage, relative bias and relative standard error) of clustered detector arrays has not been thoroughly evaluated. I used simulations to evaluate the performance of various detector and cluster spacings, cluster configurations (i.e., number of detectors arranged in a square grid), sampling extents and number of sampling occasions for estimating population density, the relationship between detection rate and distance to a detector from the animal's center of activity (σ) and base detection rates, using American black bears (Ursus americanus) as a case study. My simulations indicated that a wide range of detector configurations can provide reliable estimates if spacing between detectors in clusters is ≥1σ and ≤3σ. A number of cluster configurations and occasion lengths produced estimates that were unbiased, resulted in good spatial coverage, and were relatively precise. Moreover, increasing the duration of sampling, establishing large study areas, increasing detection rates and spacing clusters so that cross-cluster sampling of individuals can occur could help ameliorate deficiencies in the detector layout. These results have application for a wide array of species and sampling methods (e.g., DNA sampling, camera trapping, mark-resight and search-encounter) and suggest that clustered sampling can significantly reduce the effort necessary to provide reliable estimates of population density across large spatial extents that previously would have been infeasible with nonclustered sampling designs.     

Differences in spatial synchrony and interspecific concordance inform guild‐level population trends for aerial insectivorous birds

Many animal species exhibit spatiotemporal synchrony in population fluctuations, which may provide crucial information about ecological processes driving population change. We examined spatial synchrony and concordance among population trajectories of five aerial insectivorous bird species: chimney swift Chaetura pelagica, purple martin Progne subis, barn swallow Hirundo rustica, tree swallow Tachycineta bicolor, and northern rough‐winged swallow Stelgidopteryx serripennis. Aerial insectivores have undergone severe guild‐wide declines that were considered more prevalent in northeastern North America. Here, we addressed four general questions including spatial synchrony within species, spatial concordance among species, frequency of declining trends among species, and geographic location of declining trends. We used dynamic factor analysis to identify large‐scale common trends underlying stratum‐specific annual indices for each species, representing population trajectories shared by spatially synchronous populations, from 46 yr of North American Breeding Bird Survey data. Indices were derived from Bayesian hierarchical models with continuous autoregressive spatial structures. Stratum‐level spatial concordance among species was assessed using cross‐correlation analysis. Probability of long‐term declining trends was compared among species using Bayesian generalized linear models. Chimney swifts exhibited declining trends throughout North America, with less severe declines through the industrialized Mid‐Atlantic and Great Lakes regions. Northern rough‐winged swallows exhibited declining trends throughout the west. Spatial concordance among species was limited, the proportion of declining trends varied among species, and contrary to previous reports, declining trends were not more prevalent in the northeast. Purple martins, barn swallows, and tree swallows exhibited synchrony across smaller spatial scales. The extensive within‐species synchrony and limited concordance suggest that population trajectories of these aerial insectivores are responding to large‐scale but complex and species‐ and region‐specific environmental conditions (e.g. climate, land use). A single driver of trends for aerial insectivores as a guild appears unlikely.     

Estimating Coyote Densities with Local,Discrete Bayesian Capture-Recapture Models

Recent advances in noninvasive genetic sampling and spatial capture-recapture (SCR) techniques are particularly useful for monitoring cryptic wildlife species such as carnivores. In southern Arizona, USA, coyotes (Canis latrans) are thought to negatively affect endangered Sonoran pronghorn (Antilocapra americana sonoriensis), although no estimates of coyote abundance or monitoring programs exist. Sonoran pronghorn are provided supplemental feed and water in this region, resulting in areas where pronghorn and other species are congregated. Because of the higher density of artificial water sources for Sonoran pronghorn on the Cabeza Prieta National Wildlife Refuge (CPNWR), we predicted that coyote density would be higher relative to the Barry M. Goldwater Range (BMGR), where artificial water sources are less dense. We used discrete Bayesian SCR models in a local evaluation approach to provide baseline estimates of coyote abundance and understand how coyote density varied between 2 contrasting areas of land use. We identified 106 individuals from scat samples across 3 sessions in 2013 and 2014 and achieved high genotyping and individual identification success rates (~78%). Encounter rates at water catchments were nearly 11 times higher compared to road and trail transects. As predicted, we found that coyote density was on average 2 times higher on the CPNWR (11.2 coyotes/100 km²) compared to the BMGR (5.3 coyotes/100 km²). The local evaluation approach significantly reduced computational time, making the discrete Bayesian approach more practical to implement across a large study area. Our study represents an important contribution towards developing a robust monitoring program for coyotes. We hope that our novel implementation of the local evaluation approach increases the ability of wildlife managers to understand the effects of land use and other ecological influences on large carnivore populations. © 2020 The Wildlife Society.     

Does the punishment fit the crime? Consequences and diagnosis of misspecified detection functions in Bayesian spatial capture–recapture modeling

Spatial capture–recapture (SCR) analysis is now used routinely to inform wildlife management and conservation decisions. It is therefore imperative that we understand the implications of and can diagnose common SCR model misspecifications, as flawed inferences could propagate to policy and interventions. The detection function of an SCR model describes how an individual''s detections are distributed in space. Despite the detection function''s central role in SCR, little is known about the robustness of SCR‐derived abundance estimates and home range size estimates to misspecifications. Here, we set out to (a) determine whether abundance estimates are robust to a wider range of misspecifications of the detection function than previously explored, (b) quantify the sensitivity of home range size estimates to the choice of detection function, and (c) evaluate commonly used Bayesian p‐values for detecting misspecifications thereof. We simulated SCR data using different circular detection functions to emulate a wide range of space use patterns. We then fit Bayesian SCR models with three detection functions (half‐normal, exponential, and half‐normal plateau) to each simulated data set. While abundance estimates were very robust, estimates of home range size were sensitive to misspecifications of the detection function. When misspecified, SCR models with the half‐normal plateau and exponential detection functions produced the most and least reliable home range size, respectively. Misspecifications with the strongest impact on parameter estimates were easily detected by Bayesian p‐values. Practitioners using SCR exclusively for density estimation are unlikely to be impacted by misspecifications of the detection function. However, the choice of detection function can have substantial consequences for the reliability of inferences about space use. Although Bayesian p‐values can aid the diagnosis of detection function misspecification under certain conditions, we urge the development of additional custom goodness‐of‐fit diagnostics for Bayesian SCR models to identify a wider range of model misspecifications.     

Inferring past demographic changes from contemporary genetic data: A simulation‐based evaluation of the ABC methods implemented in diyabc

Inferring the demographic history of species and their populations is crucial to understand their contemporary distribution, abundance and adaptations. The high computational overhead of likelihood‐based inference approaches severely restricts their applicability to large data sets or complex models. In response to these restrictions, approximate Bayesian computation (ABC) methods have been developed to infer the demographic past of populations and species. Here, we present the results of an evaluation of the ABC‐based approach implemented in the popular software package diyabc using simulated data sets (mitochondrial DNA sequences, microsatellite genotypes and single nucleotide polymorphisms). We simulated population genetic data under five different simple, single‐population models to assess the model recovery rates as well as the bias and error of the parameter estimates. The ability of diyabc to recover the correct model was relatively low (0.49): 0.6 for the simplest models and 0.3 for the more complex models. The recovery rate improved significantly when reducing the number of candidate models from five to three (from 0.57 to 0.71). Among the parameters of interest, the effective population size was estimated at a higher accuracy compared to the timing of events. Increased amounts of genetic data did not significantly improve the accuracy of the parameter estimates. Some gains in accuracy and decreases in error were observed for scaled parameters (e.g., N_eμ) compared to unscaled parameters (e.g., N_e and μ). We concluded that diyabc ‐based assessments are not suited to capture a detailed demographic history, but might be efficient at capturing simple, major demographic changes.     

Relating mesocarnivore relative abundance to anthropogenic land‐use with a hierarchical spatial count model

There is growing need to develop models of spatial patterns in animal abundance, yet comparatively few examples of such models exist. This is especially true in situations where the abundance of one species may inhibit that of another, such as the intensively‐farmed landscape of the Prairie Pothole Region (PPR) of the central United States, where waterfowl production is largely constrained by mesocarnivore nest predation. We used a hierarchical Bayesian approach to relate the distribution of various land‐cover types to the relative abundances of four mesocarnivores in the PPR: coyote Canis latrans, raccoon Procyon lotor, red fox Vulpes vulpes, and striped skunk Mephitis mephitis. We developed models for each species at multiple spatial resolutions (41.4 km², 10.4 km², and 2.6 km²) to address different ecological and management‐related questions. Model results for each species were similar irrespective of resolution. We found that the amount of row‐crop agriculture was nearly ubiquitous in our best models, exhibiting a positive relationship with relative abundance for each species. The amount of native grassland land‐cover was positively associated with coyote and raccoon relative abundance, but generally absent from models for red fox and skunk. Red fox and skunk were positively associated with each other, suggesting potential niche overlap. We found no evidence that coyote abundance limited that of other mesocarnivore species, as might be expected under a hypothesis of mesopredator release. The relationships between relative abundance and land‐cover types were similar across spatial resolutions. Our results indicated that mesocarnivores in the PPR are most likely to occur in portions of the landscape with large amounts of agricultural land‐cover. Further, our results indicated that track‐survey data can be used in a hierarchical framework to gain inferences regarding spatial patterns in animal relative abundance.     

Pattern,prediction and parsimony in continental‐scale synthesis of pyromes: a reply to Gosper et al.

We (Murphy et al., 2013; Clarke et al., 2015) have recently developed a framework to understand the spatial distribution of fire regimes and plant fire‐response traits at large spatial scales. We integrated a range of data sources to create a continental‐scale overview of Australian pyromes from which to infer pyrogeographic drivers. Gosper et al. (in press) have criticized our approach, based on our misclassification of a vegetation type (eucalypt woodland), with distinct fire regime, in the Coolgardie bioregion of Western Australia. We argue that the intention of our integrative approach was to develop and refine conceptual models of Australian pyrogeography, not to produce a predictive map of fire regimes, and certainly not to guide local‐scale fire management. Like all models, continental‐scale syntheses of pyromes are imperfect, yet they still represent powerful tools for understanding the drivers of the spatial distribution of fire regimes.     

Buprenorphine‐elicited alteration of adenylate cyclase activity in human embryonic kidney 293 cells coexpressing κ‐, μ‐opioid and nociceptin receptors

Buprenorphine, a maintenance drug for heroin addicts, exerts its pharmacological function via κ‐ (KOP), μ‐opioid (MOP) and nociceptin/opioid receptor‐like 1 (NOP) receptors. Previously, we investigated its effects in an in vitro model expressing human MOP and NOP receptors individually or simultaneously (MOP, NOP, and MOP+NOP) in human embryonic kidney 293 cells. Here, we expanded this cell model by expressing human KOP, MOP and NOP receptors individually or simultaneously (KOP, KOP+MOP, KOP+NOP and KOP+MOP+NOP). Radioligand binding with tritium‐labelled diprenorphine confirmed the expression of KOP receptors. Immunoblotting and immunocytochemistry indicated that the expressed KOP, MOP and NOP receptors are N‐linked glycoproteins and colocalized in cytoplasmic compartments. Acute application of the opioid receptor agonists— U‐69593, DAMGO and nociceptin— inhibited adenylate cyclase (AC) activity in cells expressing KOP, MOP and NOP receptors respectively. Buprenorphine, when applied acutely, inhibited AC activity to ~90% in cells expressing KOP+MOP+NOP receptors. Chronic exposure to buprenorphine induced concentration‐dependent AC superactivation in cells expressing KOP+NOP receptors, and the level of this superactivation was even higher in KOP+MOP+NOP‐expressing cells. Our study demonstrated that MOP receptor could enhance AC regulation in the presence of coexpressed KOP and NOP receptors, and NOP receptor is essential for concentration‐dependent AC superactivation elicited by chronic buprenorphine exposure.     

Complete mitochondrial genome of the black‐tailed hornet,Vespa ducalis (Hymenoptera: Vespidae): Genomic comparisons in Vespoidea

We sequenced the complete mitochondrial genome (mitogenome) of the black‐tailed hornet, Vespa ducalis (Hymenoptera: Vespidae). The genome was 15,779‐bp long and contained typical sets of genes [13 protein‐coding genes (PCGs), 22 tRNAs, and 2 rRNAs]. The V. ducalis A + T‐rich region was 166‐bp long and was the shortest of all sequenced Vespoidea genomes, including Vespa. The genome was highly biased toward A/T nucleotides—80.1 % in the whole genome, 77.8 % in PCGs, 83.4–85.6 % in RNAs, and 92.8 % in the A + T‐rich region. These values are well within the typical range for genes and regions of Vespoidea mitogenomes. Start and stop codons in several Vespa species—including V. ducalis—were diversified, despite these species belonging to the same genus. In comparison with the ancestral mitogenomes, Vespa mitogenomes—including that of V. ducalis—showed substantial gene rearrangement; however, we detected no gene rearrangement among Vespa species. We conducted phylogenetic reconstruction based on concatenated sequences of 13 PCGs and two rRNAs (12,755 bp ) in available species of Vespoidea—21 species in six subfamilies in two families (Vespidae and Formicidae). The Bayesian inference and maximum likelihood (ML) methods revealed that each family formed strong monophyletic groups [Bayesian posterior probability (BPP) = 1; ML, 100 %]. Moreover, V. ducalis and V. mandarinia formed a strong sister group (BPP = 1; ML, 94 %).     

Bayesian design of noninferiority trials for medical devices using historical data

Summary We develop a new Bayesian approach of sample size determination (SSD) for the design of noninferiority clinical trials. We extend the fitting and sampling priors of Wang and Gelfand (2002, Statistical Science 17 , 193–208) to Bayesian SSD with a focus on controlling the type I error and power. Historical data are incorporated via a hierarchical modeling approach as well as the power prior approach of Ibrahim and Chen (2000, Statistical Science 15 , 46–60). Various properties of the proposed Bayesian SSD methodology are examined and a simulation‐based computational algorithm is developed. The proposed methodology is applied to the design of a noninferiority medical device clinical trial with historical data from previous trials.     

The distribution of <Emphasis Type="Italic">Biomphalaria</Emphasis> (Gastropoda: Planorbidae) in Lake Victoria with ecological and spatial predictions using Bayesian modelling

Intestinal schistosomiasis is a global disease of enormous public health importance. Assessing the local risk of transmission of the parasite causing the disease requires an appraisal of the distribution of its intermediate snail hosts, Biomphalaria spp.. In East Africa, the Lake Victorian Basin is a major freshwater ecosystem and highly endemic for intestinal schistosomiasis, although detailed distribution data for Biomphalaria have not been collected. We used on-the-ground malacological surveys and conjoint measurement of environmental determinants to develop models for analysis of variables associated with distribution and predictive snail mapping into data-deficient areas. Four expeditions were undertaken to collect snails along the Lake Victoria shoreline, visiting 223 sites overall. Environmental measurements were recorded at the time of inspection and water samples taken for quantification of anion and cation concentrations. The spatial distributions of Biomphalaria choanomphala and Biomphalaria sudanica were modelled in two Bayesian multivariate models: one non-spatial and one spatial with random effects. The results showed that chloride, nitrate, sulphate and the number of sympatric snail species were significant predictors of B. choanomphala whereas habitat, water depth, pH and sulphate were significant predictors for B. sudanica. The range of spatial autocorrelation was large (572.9 km for B. choanomphala and 175.3 km for B. sudanica). Interpolating snail abundance data by kriging revealed two ‘hot-spots’ of high abundance of Biomphalaria. These areas should be targeted in future expeditions to ground-truth the model’s predictions. Our study is the first to use Bayesian methods for determination of biomedically important snail distributions and sets crucial limnological baseline data for assessment of future changes in snail biodiversity in the Lake Victoria Basin.     

Seasonal habitat selection and space use by a semi‐free range herbivore in a heterogeneous savanna landscape

Understanding factors that influence habitat selection in heterogeneous landscapes is fundamental for establishing realistic models on animal distribution to inform rangeland management. In this study, we tested whether seasonal variation in habitat selection within the home range of a large herbivore was influenced by constraints such as, distances from water and central place using semi‐free range cattle (Bos taurus) as a case study. We also tested whether shifts in space use over time were dependent on spatial scale and on the overall abundance of resources. We predicted that distance from water significantly influenced dry season habitat selection while the influence of the central place on habitat selection was season‐independent. We also predicted that shifts in space use over time were spatial scale‐dependent, and that large herbivores would include more diverse habitats in their home ranges during the dry season, when water and food resources are less abundant. Multinomial logit models were used to construct habitat selection models with distances from water and central place as habitat‐specific constraints. Results showed significant variations in habitat selection between the dry and wet season. As predicted, the effect of distance from central place was season‐independent, while the effect of water was not included in the top dry season models contrary to expectation. A diverse range of habitats were also selected during the dry season including agricultural fields. Results also indicated that shifts in space use were spatial scale dependent, with core areas being more sensitive to changes than the home range. In addition, shifts in space use responded to temporal changes in habitat composition. Overall, our results suggest that semi‐free range herbivores adopt different foraging strategies in response to spatial‐temporal changes in habitat availability.     

When mechanism matters: Bayesian forecasting using models of ecological diffusion

Ecological diffusion is a theory that can be used to understand and forecast spatio‐temporal processes such as dispersal, invasion, and the spread of disease. Hierarchical Bayesian modelling provides a framework to make statistical inference and probabilistic forecasts, using mechanistic ecological models. To illustrate, we show how hierarchical Bayesian models of ecological diffusion can be implemented for large data sets that are distributed densely across space and time. The hierarchical Bayesian approach is used to understand and forecast the growth and geographic spread in the prevalence of chronic wasting disease in white‐tailed deer (Odocoileus virginianus). We compare statistical inference and forecasts from our hierarchical Bayesian model to phenomenological regression‐based methods that are commonly used to analyse spatial occurrence data. The mechanistic statistical model based on ecological diffusion led to important ecological insights, obviated a commonly ignored type of collinearity, and was the most accurate method for forecasting.     

Using functional traits to predict species growth trajectories,and cross‐validation to evaluate these models for ecological prediction

Modeling plant growth using functional traits is important for understanding the mechanisms that underpin growth and for predicting new situations. We use three data sets on plant height over time and two validation methods—in‐sample model fit and leave‐one‐species‐out cross‐validation—to evaluate non‐linear growth model predictive performance based on functional traits. In‐sample measures of model fit differed substantially from out‐of‐sample model predictive performance; the best fitting models were rarely the best predictive models. Careful selection of predictor variables reduced the bias in parameter estimates, and there was no single best model across our three data sets. Testing and comparing multiple model forms is important. We developed an R package with a formula interface for straightforward fitting and validation of hierarchical, non‐linear growth models. Our intent is to encourage thorough testing of multiple growth model forms and an increased emphasis on assessing model fit relative to a model's purpose.