A hybrid double-observer sightability model for aerial surveys

Griffin, Paul C.; Lubow, Bruce C.; Jenkins, Kurt J.; Vales, David J.; Moeller, Barbara J.; Reid, Mason; Happe, Patricia J.; McCorquodale, Scott M.; Tirhi, Michelle J.; Schaberl, Jim P.; Beirne, Katherine

doi:10.1002/jwmg.612

Cited by 26 publications

(97 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Previous research has shown that the detectability of ungulates is influenced by group size, canopy cover, individual activity (bedded, standing, moving), ground terrain, and light conditions (Thomas and Gray 2002;Gilbert and Moeller 2008;Patterson et al 2014;Peters et al 2014). Of these factors, vegetation cover, which is largely dependent on the habitat, is the most frequently reported factor known to affect the detectability of target wildlife during surveys (Gasaway 1985;Samuel et al 1987;Gilbert and Moeller 2008;McIntosh et al 2009;Jarding 2010;Griffin et al 2013). For this reason we chose to focus on the effect of habitat on the detection of targets.…”

mentioning

confidence: 99%

Evaluation of an unmanned aircraft system for detecting surrogate caribou targets in Labrador

Patterson

Koski²,

Pace³

et al. 2016

J. Unmanned Veh. Sys.

View full text Add to dashboard Cite

Regular, standardized population inventories have been suggested as an important component to the recovery of declining populations of boreal caribou (Rangifer tarandus caribou). Current survey methods typically employ manned aircraft, which can be noisy, expensive to operate, and dangerous for the people conducting the surveys. Small unmanned aerial systems (UAS) have garnered attention as a promising alterative to conducting aerial surveys in manned aircraft. Our research investigates the feasibility of using an UAS to conduct aerial surveys and determine which factors affect the detection of surrogate caribou targets, and hence may affect detection of real caribou. In the fall of 2013, we tested the capabilities of the Brican TD100E, a small, electric-powered fixed-wing UAS, to fly beyond visual line of sight near Goose Bay, Labrador. Seven surveys were done using different flight paths to collect aerial images of 26 surrogate caribou targets placed in six different habitats. Mixed effects logistic regression models were used to evaluate how habitat type, distance of the target from the image centerline, photo analysts’ experience level, flight time, and the target contrast against the landscape influenced the detection of surrogate caribou targets. We found that habitat type, target contrast, and the flight time affected target detection. Overall, 77.5% of the targets were detected; the odds of a photo analyst detecting a target in open habitat were roughly 10.5 times higher than in burned habitat and 42 times higher than in heavy forest. Target detection was influenced by the contrast of the target against the landscape, where a higher corrected integrated density was associated with greater target detection. The detection of targets was 87% during evening flights and 75% for morning flights. This study was the first of its kind to successfully fly a UAS beyond line of sight over land for non-military applications in North America and the findings of our research will provide an evaluation for using UAS to survey caribou in the future.

show abstract

mentioning

confidence: 99%

Evaluation of an unmanned aircraft system for detecting surrogate caribou targets in Labrador

Patterson

Koski²,

Pace³

et al. 2016

J. Unmanned Veh. Sys.

View full text Add to dashboard Cite

show abstract

“…(1) A simple Lincoln-Peterson estimate (i.e., not including any covariates) based on the 2-occasion aerial observations (Graham and Bell, 1989) relies on all groups being equally likely to be seen, but the vast difference in elk group size along with other factors affecting sighting probability were expected to make this assumption implausible and resulting estimates unreliable. (2) We discuss the contrast between our methods with simple sightability models (Steinhorst and Samuel, 1989), but these offer no improvements over the other methods presented here, as has been previously argued and demonstrated empirically (see model M S in Griffin et al, 2013), so we do not present estimates based on this method.…”

Section: F -Raw Count By Front Seat Observers In the Helicoptermentioning

confidence: 77%

“…Our models M D and M H are structured similarly to models in Griffin et al (2013), which should be consulted for additional details on model structure and fitting methods. We fit these same models to alternative subsets of the data, giving rise to M D + and M H * estimates for comparison purposes.…”

Section: F -Raw Count By Front Seat Observers In the Helicoptermentioning

confidence: 99%

“…There are several methods for estimating group-specific detection probabilities in aerial surveys, including sightability models (Steinhorst and Samuel, 1989), distance models (Burnham and Anderson, 1984;Buckland et al, 2004), mark-resight models (White, 1996;Skalski et al, 2005), double-observer models (Graham and Bell, 1989), and methods that combine two or more such techniques often referred to as ''hybrid'' models (Quang and Becker, 1997;Buckland et al, 2010;Burt et al, 2014). These techniques all have individual strengths and weaknesses (Griffin et al, 2013). By combining several techniques to form a hybrid model, the biases of each individual technique can be minimized, while strengthening their collective precision (Griffin et al, 2013).…”

Section: Introductionmentioning

confidence: 98%

“…These techniques all have individual strengths and weaknesses (Griffin et al, 2013). By combining several techniques to form a hybrid model, the biases of each individual technique can be minimized, while strengthening their collective precision (Griffin et al, 2013).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Application of a hybrid model to reduce bias and improve precision in population estimates for elk (Cervus elaphus) inhabiting a cold desert ecosystem

Schoenecker

Lubow

2016

Journal of King Saud University - Science

Self Cite

View full text Add to dashboard Cite

Accurately estimating the size of wildlife populations is critical to wildlife management and conservation of species. Raw counts or ''minimum counts'' are still used as a basis for wildlife management decisions. Uncorrected raw counts are not only negatively biased due to failure to account for undetected animals, but also provide no estimate of precision on which to judge the utility of counts. We applied a hybrid population estimation technique that combined sightability modeling, radio collar-based mark-resight, and simultaneous double count (double-observer) modeling to estimate the population size of elk in a high elevation desert ecosystem. Combining several models maximizes the strengths of each individual model while minimizing their singular weaknesses. We collected data with aerial helicopter surveys of the elk population in the San Luis Valley and adjacent mountains in Colorado State, USA in 2005 and 2007. We present estimates from 7 alternative analyses: 3 based on different methods for obtaining a raw count and 4 based on different statistical models to correct for sighting probability bias. The most reliable of these approaches is a hybrid double-observer sightability model (model M H ), which uses detection patterns of 2 independent observers in a helicopter plus telemetry-based detections of radio collared elk groups. Data were fit to customized mark-resight models with individual sighting covariates. Error estimates were obtained by a bootstrapping procedure. The hybrid method was an improvement over commonly used alternatives, with improved precision compared to sightability modeling and reduced bias compared to double-observer modeling. The resulting population estimate corrected for multiple

show abstract

Leveraging hunters as citizen scientists for monitoring non‐target species

DeCesare,

Podruzny,

Gude

2023

Ecol Sol and Evidence

View full text Add to dashboard Cite

Hunter populations can provide a tremendous workforce of citizen scientists afield when queried for data. Soliciting incidental observations of non‐target species from hunters may be a relatively important but untapped population monitoring resource in systems where hunter effort is common and widespread. During 2012–2016, we queried hunters of deer and elk for observations of a non‐target species, moose, across their statewide distribution in Montana. We analysed data in an abundance‐detection framework with n‐mixture models and evaluated the effects of covariates such as hunter effort, survey response totals, weekly session and forest cover on detection probability before using models to predict moose abundance. We collected an average of 3409 moose observations per year and our best n‐mixture model included effects of week, year (number of responses), site (proportionate forest cover) and site‐year (hunter effort) on detection probability, as well as an effect of site (area of forest and shrub habitat) on abundance. Density estimates averaged 0.099 (range 0.002–0.439) moose/km2 across sites or 0.200 (range 0.017–0.799) moose/km2 when limited to density within shrub and forest cover specifically. Statewide abundance totals across the 5‐year study period averaged 10,755 (range 9925–11,620). Goodness‐of‐fit tests showed that models were identifiable and overdispersion of the data was low, yet some caution is still warranted when extrapolating these data to abundance estimates. Querying a sample of deer‐elk hunters for observations of a non‐target species yielded thousands of spatially georeferenced detections per year and analysis in a temporally structured framework yielded estimates of both detection probability and abundance. Abundance estimates at this scale are unprecedented for moose in Montana and are encouraging for long‐term monitoring over space and time.

show abstract

A hybrid double-observer sightability model for aerial surveys

Cited by 26 publications

References 41 publications

Evaluation of an unmanned aircraft system for detecting surrogate caribou targets in Labrador

Evaluation of an unmanned aircraft system for detecting surrogate caribou targets in Labrador

Application of a hybrid model to reduce bias and improve precision in population estimates for elk (Cervus elaphus) inhabiting a cold desert ecosystem

Leveraging hunters as citizen scientists for monitoring non‐target species

Contact Info

Product

Resources

About