Line-Transect Surveys

Credit: Kristin Laidre

The two most common methods to estimate polar bear population abundance are capture-mark-recapture (involving the physical or genetic marking of animals) and aerial survey methods. The two methods have different strengths and weaknesses and rely on different assumptions. Physical and genetic capture-recapture studies are usually part of a program where samples are collected from individuals over multiple years and thus provide a lot of additional data besides estimates of population size, such as information on survival and reproduction, along with condition and age of individual animals (see separate page on capture-mark-recapture). However, in some areas, the assumptions required for estimating population size with these methods can be difficult to meet and in other areas, access to physically mark polar bears has been logistically challenging. Further, these approaches can be stressful to animals, requiring pursuit by helicopter, and chemical immobilization in the case of physical capture-recapture. However, it is important to realize that aerial surveys provide an estimate of the number of bears in a specific region at a specific time, whereas capture-recapture analyses produce estimates of a superpopulation, defined as all animals with a non-negligible probability of using the sampling area over the course of the study. Because there are likely some temporary emigrations from a subpopulation (i.e. not all animals with fidelity to an area are available for sampling each year), it is expected that estimates of a superpopulation size are larger than estimates from an aerial survey (Laidre et al. 2023).

Wiig and Derocher (1999) advocated for an aerial survey approach, using distance sampling, for the Barents Sea subpopulation to estimate population size, due to the large extent and the lack of bases needed for capture-recapture based estimates. This led to the first polar bear aerial survey on a subpopulation-level scale, performed in August 2004 (Aars et al. 2009). A new aerial survey was conducted in Foxe Basin in 2009 (Stapleton et al. 2014) and in recent years, several polar bear aerial surveys, using a variety of methods, have been conducted in Canada, e.g. in the Foxe Basin, Western Hudson Bay and Southern Hudson Bay subpopulations (Stapleton et al. 2016; Obbard et al. 2015, 2018), in the Barents Sea in 2015 (Aars et al. 2017), and in the Kane Basin subpopulation (Wiig et al. 2022). The literature on aerial surveys to estimate animal abundance is extensive and has benefitted from decades of development by wildlife biologists and statisticians.

Description of Some Common Aerial Survey Methods

Aerial survey methods rely on the use of aircraft to detect and count animals. Following counting and distance, statistical methods are then used to either extrapolate to areas that were not surveyed, to estimate the number of animals that went undetected but were in the surveyed areas, or both. In some cases, complete or total counts of animals can be done. Total counts have no associated uncertainty, but are usually not possible logistically for populations that occupy large areas or for which detection of animals is not certain. An alternative is “strip transects” where observers define a strip of a certain width and count all individuals within the strip. Density is calculated by dividing the count by the strip area and then used to extrapolate to the uncovered areas, to gain a population estimate. This method requires the assumption that all individuals within the strip are detected, limiting the strip width. Further, this approach is inefficient, because animals outside the strip are often detected by observers but not able to be included in the count. Even within the narrow strip, the assumption that animals are detected perfectly is usually violated, so there are several approaches for estimating the proportion of animals within that are missed by observers. One such approach that has been used in polar bear surveys is visual mark-recapture, where observers within the aircraft are divided into front and rear observers, and it is recorded whether the front, rear or both observers detect each animal. The pattern of detections and non-detections can be used to estimate the proportion of animals observed within the strip. This method has been used in multiple surveys within the Southern Hudson Bay subpopulation (Obbard et al. 2015, 2018). This method requires that detection probability does not decrease with distance from the centre of the strip. All observations outside the strip have to be dropped from the analyses. These issues mean that only small areas of a subpopulation can be sampled, which reduces the reliability of extrapolations to unsampled areas. Strip transects have been used and evaluated for polar bears (e.g. Wiig and Bakken 1990).

Distance sampling has much in common with the strip transect method but allows detection probability to decrease with distance from a survey line, and uses the rate at which this happens, estimated using a detection function, to estimate the effective strip width. This is then combined with the number of animals detected to estimate density, which can be extrapolated to unsampled areas. Traditional distance sampling methods assume that all bears on the line (on the flight path) are detected by observers. However, the development of mark-recapture distance sampling allows for the modelling of imperfect detection on the transect line as well as the decline in detectability with distance from the line. Mark-recapture distance sampling uses a combination of distance sampling and the mark-recapture approach discussed above to adjust the distance detection function based on characteristics of the habitat in which bears are detected, or based on group size or bear activity when detected. Distance sampling is a highly transferable method that has been used in multiple population estimates for polar bears, applied both over the ice and during ice-free seasons in seasonal sea-ice systems.

Assumptions for distance sampling surveys are (Buckland et al. 2001):

  1. A large number of transects are randomly allocated in the study area independent of the distribution of the survey population
  2. All individuals on the line are detected with certainty (g(0)=1), or mark-recapture distance sampling is used to estimate the probability of detecting animals on the transect line.
  3. Animals are detected at their initial location- i.e., they do not move in response to the observers until after they are detected.
  4. Distances are measured without error

For more details on distance sampling methods, see this page. Literature on the subject is: Buckland et al. 2001, 2004, 2015, Borchers et al. 2002, Thomas et al. 2010.

Because of the highly heterogeneous distribution of polar bears in some subpopulations, often the best approach for estimating abundance will be a combination of multiple aerial survey approaches. In Southern Hudson Bay, Obbard et al. (2015, 2018) used a combination of distance sampling, strip transects with visual mark-recapture, and total counts to sample bears across the subpopulation during the ice-free season. Distance sampling can also be combined with physical biopsy sampling for a genetic capture-recapture study which was done in Kane Basin (Wiig et al 2022, Laidre et al 2023). Further, the aerial survey literature is vast and continually expanding. Thus, there are numerous additional approaches that have been applied to other species, which could be tested on polar bears.

Conclusion

Aerial surveys represent a robust and non-intrusive approach to estimating the number of bears within an area at a specific point in time. They offer some major advantages over physical and genetic capture-recapture. The main advantage is that an estimate of abundance can be developed after a short period of survey in a single year, whereas physical and genetic capture-recapture methods typically require multiple years of study. However, these estimates are necessarily not comparable because capture-recapture method might give the size of a superpopulation. Further, there is no physical contact with bears in aerial survey methods, limiting stress and potential injury to bears and providing for a safer survey approach for observers. However, in seasonal sea-ice systems, such as the Western and Southern Hudson Bay subpopulations, aerial surveys are highly effective also because of the large and easily detectable concentrations of bears during the ice-free season. For these reasons, aerial surveys have become more common for polar bear subpopulations in recent years.  However, these same advantages offer some drawbacks. In aerial surveys, typically we only can estimate abundance and get no information on immigration or emigration and only indirect information on survival and reproduction. This can be problematic when trying to understand why a population has increased or decreased across years. For this reason, aerial surveys typically must be repeated more regularly to provide sufficient information for managing polar bears. Further, with restricted ability to collect physical samples except from biopsies, aerial surveys greatly reduce what else can be learned about polar bears. For example, it is problematic to assess body condition, detailed sex (but from biopsies) and age structure, or diet using aerial surveys, but these are key pieces of information that come from physical and even genetic capture-recapture surveys. These drawbacks have led to development of methods for combining aerial surveys and mark-recapture (biopsies) approaches into integrated models that leverage the strengths of both approaches. These integrated models are becoming more common and likely hold great promise for application across the polar bear range.

Further reading

  • Aars J, Marques TA, Buckland ST, Andersen M, Belikov S, Boltunov A, and Wiig Ø. 2009. Estimating the Barents Sea polar bear subpopulation size. Marine Mammal Science 25: 35-52.
  • Aars J, Marques T, Lone K, Andersen M, Wiig Ø, Fløystad IMB, Hagen SB, and Buckland S.  2017. The number and distribution of polar bears in the western Barents Sea area. Polar Research 36: 1374125, https://doi.org/10.1080/17518369.2017.1374125.
  • Borchers DL, Buckland ST, and Zucchini W. 2002. Estimating Animal Abundance. Closed populations. Springer-Verlag, London.
  • Buckland ST, Anderson DR, Burnham KP, Laake JL, Borchers DL, and Thomas L. 2001. Introduction to Distance Sampling. Oxford University Press, Oxford.
  • Buckland ST, Anderson DR, Burnham KP, Laake JL, Borchers DL, and Thomas L. (eds). 2004. Advanced Distance Sampling. Oxford University Press, Oxford.
  • Buckland ST, Rexstad EA, Marques TA, and Oedekoven CS. 2015. Distance Sampling: Methods and Applications. Methods in Statistical Ecology. Springer International Publishing. https://doi.org/10.1007/978-3-319-19219-2
  • Laidre KL, Supple MA, Born EW, Regehr, EV, Wiig Ø, Ugarte F, Aars J, Dietz R, Sonne C, Hegelund P, Isaksen C, Akse GB, Cohen B, Stern HL, Moon T, Vollmers C, Corbett-Detig R, Paetkau, and Shapiro B. 2022. Glacial ice supports a distinct and undocumented polar bear subpopulation persisting in late 21st-century sea-ice conditions. Science 376: 1333-1338.
  • Miller DL. 2015. Distance: Distance Sampling Detection Function and Abundance Estimation. R package version 0.9.3. https://CRAN.R-project.org/package=Distance
  • Obbard ME, Stapleton S, Middel KR, Thibault I, Brodeur V, and Jutras C. 2015. Estimating the abundance of the Southern Hudson Bay polar bear subpopulation with aerial surveys. Polar Biology 38: 1713-1725.
  • Obbard ME, Stapleton S, Middel KR, Szor G, Jutras C, and Dyck M. 2018. Re-assessing abundance of Southern Hudson Bay polar bears by aerial survey: effects of climate change at the southern edge of the range. Arctic Science 4: 634-655.
  • Stapleton S, Atkinson S, Hedman D, and Garshelis D. 2014. Revisiting Western Hudson Bay: Using aerial surveys to update polar bear abundance in a sentinel population. Biological Conservation 170: 38-47.
  • Stapleton S, Peacock E, and Garshelis D. 2016. Aerial surveys suggest long-term stability in the seasonally ice-free Foxe Basin (Nunavut) polar bear population. Marine Mammal Science 32: 181-201.
  • Thomas L, Buckland ST, Rexstad EA, Laake JL, Strindberg S, Hedley SL, Bishop JRB, Marques TA, and Burnham KP. 2010. Distance software: design and analysis of distance sampling surveys for estimating population size. Journal of Applied Ecology 47: 5-14.
  • Wiig Ø, and Bakken V. 1990. Aerial strip surveys of polar bears in the Barents Sea. Polar Research 8: 309-311.
  • Wiig Ø, and Derocher A. 1999. Application of aerial survey methods to polar bears in the Barents Sea. Pp. 27-36 in Garner GW, Amstrup SC, Laake JL, Manly BFJ, McDonald LL, and Robertson DG. (eds). Marine Mammal Survey and Assessment Methods. Balkema, Rotterdam.
  • Wiig Ø, Atkinson SN, Born EW, Stapleton S, Arnold TW, Dyck M, Laidre KL, Lunn NJ, and Regehr EV. 2022. An on-ice aerial survey of the Kane Basin polar bear subpopulation. Polar Biology 45: 89-100.
© 2024 Polar Bear Specialist Group - A non-profit group dedicated to the research and management of polar bears. Read our Privacy Policy and Terms of Service.
↑