Interpretations of Polar Bear (Ursus maritimus) Tracks by Inuit Hunters: Inter-rater Reliability and Inferences Concerning Accuracy

1. What they did

Polar bear populations in Canada are mainly monitored with aerial capture-mark-recapture surveys, which are expensive, infrequent and sometimes lack local support. The authors asked whether Inuit hunters, through their tracking experience, could provide reliable, non-invasive estimates of polar bear characteristics from tracks.

Participants: nine Inuit from Gjoa Haven, Taloyoak and Cambridge Bay. Seven were active hunters and two were elders who no longer hunted. One hunter translated for the elders.

Interpreting tracks (11 to 15 May 2009): the group located polar bear tracks by snowmobile along routes chosen to avoid sampling the same bear twice. At each track, each participant independently estimated the bear’s sex, age in years and nose-to-tail size in feet, and how many days old the track was. There was no discussion, and “I don’t know” was allowed.

Checking the estimates: they measured how consistently the hunters agreed with each other (reliability); compared sex estimates with genetic sex determinations from hair, collected at baited barbed-wire hair-snag stations and along tracks in 2007, 2008 and 2009; and compared size estimates with stride lengths measured by three participants on nine tracks.

Interviews: short semi-structured interviews recorded each participant’s background, hunting and guiding experience, and how they interpret tracks. The authors explored whether experience related to reliability and accuracy.

2. What they found

Agreement between hunters

• The group met 99 tracks; all nine participants interpreted the same 78. No one answered “I don’t know”.
• The group was reliable in estimating sex, age and size of the bears. Intraclass correlations were 0.74, 0.74 and 0.91, above the usual benchmark of 0.7.
• They were not reliable in estimating the age of the track (0.35). This was mainly because of one participant. Without him, track-age reliability rose to 0.85, and the other measures changed little.
• Hunters still differed in their average estimates. Some participants consistently gave different sex ratios, ages, sizes or track ages from the rest.
• The group’s variability did not change over the course of the survey.

Sex compared with genetics

• Of 53 hair samples, 25 could be sexed and linked to single tracks. Tracks easily identified as females with cubs were excluded.
• Hunters’ sex estimates agreed with genetic sex 74.42% of the time (41 track–hunter pairs). That is significantly better than guessing (50%), though the sample was small.

Size compared with stride

• On nine tracks, size estimates by participants 1, 2 and 7, all experienced professional guides, were significantly correlated with stride length. Most other participants showed strong but not significant correlations. The elder (participant 8) and the youngest hunter (participant 6) correlated least.

How hunters read polar bear tracks

The criteria were widely shared:

• Sex: male footprints turn inward toward the centre of the track, which one hunter linked to males’ more developed shoulder muscles; male prints are larger, with wider toes; some hunters also used accompanying tracks.
• Age: older bears make larger, rounder prints; young males’ prints are similar in size to females’ but narrower; one hunter judged weight, and so age, from how deep the print sank in the snow.
• Size: footprint size.
• Age of track: weather and snow conditions. Soft prints are fresh; hard prints are older.

Many of these criteria match published scientific findings, for example that adult male footprints average 22 cm wide and adult females 16 cm.

Experience (exploratory, based on nine people)

• More guiding experience went with more reliable size estimates, and estimates that matched stride length better.
• Older participants were more reliable on bear age but less reliable on track age.
• More education went with more reliable track-age estimates.
• Preferring to hunt with others, and claiming to read a bear from only a few footprints, went with lower agreement or accuracy.
• Hunters learned tracking from elders, other hunters and family.

3. What they concluded

• Inuit hunters can make reliable, consistent estimates of polar bear sex, age and size from tracks. Their sex estimates agreed with genetics more often than chance.
• Individual hunters can interpret tracks reliably on their own, without group discussion.
• The shared techniques of hunters may explain their agreement. Individual experience and methods may explain differences, and the most reliable hunters’ techniques could be taught to others.
• Before track data can be used in monitoring, accuracy for age, size and track age must still be tested. Needed next steps: genotyping previously aged bears; calibrating stride against body size in captive bears; testing how substrate affects track ageing; and analysing digital footprint images to identify individuals.
• For now, track data should supplement, not replace, capture-mark-recapture surveys. Combined with genetic sampling, they could show changes in polar bear activity and signal when a new survey is needed to review harvest quotas.
• Inuit from any community with expertise like the most reliable hunters here could provide useful non-invasive estimates.

4. How tracking was or was not used

Inuit hunters interpreted polar bear tracks on sea ice to estimate each bear’s sex, age and size and the age of the track. Each hunter worked independently. Three participants measured stride lengths along suitable stretches of track where the bear had walked on flat ground. Hair was collected along tracks and at hair-snag stations for genetic sexing.

5. Tracker or observer certification

Not reported. No tracker certification is mentioned. The participants were seven active hunters and two elders. Some were professional polar bear hunting guides, and two were certified Canadian Rangers (a military reserve role, not tracker certification). Preliminary work in 2007 and 2008 suggested active, experienced hunters were the more reliable observers.

6. Other technologies, skills or innovations

• Non-invasive genetic sexing from hair. The authors refined the method so that female samples were no longer mistaken for male, then tested it on samples of known sex.
• Hair-snag stations: barbed-wire squares around a pole baited with seal meat.
• The authors suggest digital footprint images analysed with multivariate methods (as in footprint identification of rhinos) to identify individual bears.

7. Implications for conservation (from the paper)

• Inuit track interpretation could provide more frequent, cheaper and less invasive information on polar bears than aerial surveys alone, and could help reduce conflict between scientists and Inuit communities over harvest quotas.
• Track data could detect large changes in activity over wide areas and time, and help decide when new population surveys are needed.
• Integrating local expertise into monitoring requires first testing its reliability and accuracy.