Coarse‐Scale Distribution Surveys and Occurrence Probability Modeling for Wolverine in Interior Alaska

Gardner, Craig L.; Lawler, James P.; Hoef, Jay M. Ver; Magoun, Audrey J.; Kellie, Kalin A.

doi:10.2193/2009-386

Cited by 19 publications

(34 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, to model spatially correlated unit effects, Hooten et al (2003) and Chelgren et al (2011) used a geostatistical model on a continuous spatial domain, whereas, Magoun et al (2007), Gardner et al (2010), andAing et al (2011) used a conditionally autoregressive (CAR) model on a discrete spatial domain. For the models developed here, we consider only the discrete spatial domain as occupancy study areas often represent tessellations of areas for which a prediction of the occupancy process is desired (e.g., Hooten et al 2003, Sargeant et al 2005, Magoun et al 2007.…”

Section: Spatial Regression Modelsmentioning

confidence: 99%

“…As with previous occupancy models containing complex autocorrelation structures (e.g., Hoeting et al 2000, Sargeant et al 2005, Magoun et al 2007, Gardner et al 2010), we adopt a Bayesian perspective and use Markov chain Monte Carlo (MCMC) for inference. Bayesian inference is based on the joint posterior distribution, which, in our case, may be written as ½z; z;ỹ; g; b; c; s j y } ½y jỹ ½ỹ j b; z ½z jz ½z j c; g ½g j s ½b ½c ½s: ð4Þ Owing to our judicious choice of link function, all the necessary full conditional distributions are available in closed form, and we are able to sample from them directly (i.e., Gibbs sampler).…”

Section: Bayesian Inferencementioning

confidence: 99%

“…Although this approach yields unbiased inference when occupancy observations at nearby units are conditionally independent given any available covariates, residual spatial autocorrelation may lead to biases and overestimated precision. As a result, a few researchers have modeled occupancy within a hierarchical spatial framework, which explicitly acknowledges the presence of autocorrelation in the occupancy process (e.g., Hoeting et al 2000, Hooten et al 2003, Sargeant et al 2005, Magoun et al 2007, Royle and Dorazio 2008, Gardner et al 2010, Aing et al 2011, Chelgren et al 2011, Karanth et al 2011. Incorporating spatial structure also reflects the intuitive notion that adjacent survey units are likely to exhibit occupancy rates more similar than those far apart.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Spatial occupancy models for large data sets

Johnson¹,

Conn²,

Hooten

et al. 2013

Ecology

157

251

View full text Add to dashboard Cite

Abstract. Since its development, occupancy modeling has become a popular and useful tool for ecologists wishing to learn about the dynamics of species occurrence over time and space. Such models require presence-absence data to be collected at spatially indexed survey units. However, only recently have researchers recognized the need to correct for spatially induced overdisperison by explicitly accounting for spatial autocorrelation in occupancy probability. Previous efforts to incorporate such autocorrelation have largely focused on logitnormal formulations for occupancy, with spatial autocorrelation induced by a random effect within a hierarchical modeling framework. Although useful, computational time generally limits such an approach to relatively small data sets, and there are often problems with algorithm instability, yielding unsatisfactory results. Further, recent research has revealed a hidden form of multicollinearity in such applications, which may lead to parameter bias if not explicitly addressed. Combining several techniques, we present a unifying hierarchical spatial occupancy model specification that is particularly effective over large spatial extents. This approach employs a probit mixture framework for occupancy and can easily accommodate a reduced-dimensional spatial process to resolve issues with multicollinearity and spatial confounding while improving algorithm convergence. Using open-source software, we demonstrate this new model specification using a case study involving occupancy of caribou (Rangifer tarandus) over a set of 1080 survey units spanning a large contiguous region (108 000 km 2 ) in northern Ontario, Canada. Overall, the combination of a more efficient specification and open-source software allows for a facile and stable implementation of spatial occupancy models for large data sets.

show abstract

Section: Spatial Regression Modelsmentioning

confidence: 99%

Section: Bayesian Inferencementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Spatial occupancy models for large data sets

Johnson¹,

Conn²,

Hooten

et al. 2013

Ecology

157

251

View full text Add to dashboard Cite

show abstract

“…Deriving population estimates for wolverines using methods such as mark‐recapture and density estimation (Golden et al , Royle et al ) is impractical across the vast North Slope. Alternatively, data on the detection or non‐detection of wolverines across large areas are easier and less costly to collect than abundance information (Magoun et al , Gardner et al ) and when combined with environmental variables, these data can be used in an occupancy modeling framework to estimate a species’ probability of occupancy with respect to habitat features hypothesized to be important to the species (e.g., land cover or land use types, elevation and terrain ruggedness, presence of other species; MacKenzie et al ).…”

mentioning

confidence: 99%

Distribution and occupancy of wolverines on tundra, northwestern Alaska

Poley

Magoun²,

Robards

et al. 2018

J Wildl Manag

Self Cite

View full text Add to dashboard Cite

Understanding wildlife distribution and habitat use is needed for effectively balancing resource development, wildlife conservation, and Alaska Native subsistence on the North Slope of Alaska, USA. This region includes the National Petroleum Reserve‐Alaska (NPR‐A), a 96,000‐km2 remote area of largely undeveloped lands that is important for wildlife, including caribou (Rangifer tarandus), wolves (Canis lupus), and wolverines (Gulo gulo). We focused our study on spring distribution and occupancy of wolverines in the NPR‐A because a baseline distribution estimate is required to understand current distribution and track changes over time. We conducted aerial surveys of wolverine tracks in snow during March and April of 2014 and 2015, surveying over 84,400 km2 using 100‐km2 hexagonal sampling units. We used hierarchical Bayesian occupancy modeling to determine wolverine distribution and estimate probability of occupancy within each hexagon, relative to measured covariates with potential to affect either detection or occupancy. Probability of wolverine occupancy increased as well‐drained soils increased, suggesting that wolverines prefer drier areas or habitat features associated with well‐drained soils. In addition, as standard deviation of elevation increased, wolverine occupancy also increased, indicating that wolverines may prefer areas with more rugged and variable terrain. Mean elevation was not retained as a covariate in the best‐fitting model, supporting the importance of terrain ruggedness rather than elevation on wolverine distribution within the NPR‐A. Spatially, areas of highest wolverine occupancy occurred within the southern and northeastern portions of the study area, with lowest occupancy in the northern portion of the study area west of Teshekpuk Lake. Based on the spatial pattern of wolverine probability of occupancy, we proposed 4 potential wolverine management zones with varying priorities for monitoring and managing wolverine populations. © 2018 The Wildlife Society.

show abstract

“…Our approach builds on recent applications of spatial occupancy models to track surveys (Magoun et al. 2007; Gardner et al. 2010; Hines et al.…”

Section: Introductionmentioning

confidence: 99%

A Bayesian hierarchical occupancy model for track surveys conducted in a series of linear, spatially correlated, sites

Aing¹,

Halls²,

Oken³

et al. 2011

Journal of Applied Ecology

View full text Add to dashboard Cite

Summary1. Natural resource agencies often rely on surveys of animal sign (e.g. scat, scent marks, tracks) for population assessment, with repeat surveys required to model and account for uncertain detection. Using river otter Lontra canadensis snow-track survey data as a motivating example, we develop a 3-level occupancy model with parameters that describe (i) site-level occupancy probabilities, (ii) otter movement (and thus, track availability) and (iii) recorded presence-absence of tracks (conditional on the availability of tracks for detection). 2. We incorporated several recent developments in occupancy modelling, including the presence of both false negatives and false positives, spatial and temporal correlation and repeated sampling across distinct observers. 3. We investigated optimal allocation of sampling effort (e.g. within and among snowfall events) using simulations. We also compared models that allowed site-level occupancy and track-laying processes to be spatially correlated with models that assumed independence among sites. 4. Both types of models (independence and spatial) performed well across a range of simulated parameter values, but the spatial model resulted in more accurate point estimates for detection parameters and credibility intervals with better coverage rates when data were spatially correlated. When applied to real data, the spatial model resulted in a higher estimate of the occupancy rate ðŴÞ than the baseline model (0AE82 vs. 0AE59). A minimum of 15-20 helicopter flights, distributed among at least three unique snow events, were needed to meet precision goals (standard errorŴ < 0Á05). 5. Synthesis and applications. We describe a flexible and robust occupancy modelling framework that accounts for heterogeneous detection rates in surveys of animal sign. The method allows for spatially correlated sites and should have broad relevance to surveys conducted by many natural resource agencies.

show abstract

Coarse‐Scale Distribution Surveys and Occurrence Probability Modeling for Wolverine in Interior Alaska

Cited by 19 publications

References 27 publications

Spatial occupancy models for large data sets

Spatial occupancy models for large data sets

Distribution and occupancy of wolverines on tundra, northwestern Alaska

A Bayesian hierarchical occupancy model for track surveys conducted in a series of linear, spatially correlated, sites

Contact Info

Product

Resources

About