How Biologists Count a Herd That Never Stops Moving

The George River caribou herd is counted from the air, in July, by photographing the groups it forms a few weeks after calving, with GPS collars used to find those groups in the first place. The latest full count, from July 2024, put the herd at about 8,600 animals, a number that took four days of flying and years of collar data to produce.

Why can a herd like this not just be counted from the ground?

Because it does not sit still, and it does not sit together. For most of the year the George River herd is spread across a vast range of tundra and taiga straddling the Quebec and Labrador border, in small, scattered bands that a ground crew would never find, let alone count. A road census, the method that works for a deer population near a highway, is not available on a landscape without roads.

The one window when the herd behaves differently is the few weeks after calving, in early July, when cows with new calves gather into large, dense groups on open ground, partly for predator protection. That brief gathering is what makes an aerial photo count possible at all.

How do biologists find the groups before they can photograph them?

With collars fitted to a sample of animals in the herd, each sending two GPS locations a day and an individual VHF radio signal. In the 2024 survey, 62 of 65 active collars were sampled during the overflights. Each collared animal gives a location, and because caribou move in groups, a collared animal is a marker for the group around it. Crews fly to the collar locations first, then search the surrounding ground visually for every other group in the area, collared or not.

This is the step that turns a scattered herd into a countable one. Without the collars, a crew flying blind over that much ground would miss groups, double count others, and produce a number with no way to check it.

What actually happens during the flight?

Once a group is located, it is photographed from the air, not simply counted by eye. Caribou in a tight post-calving group are hard to count accurately from a moving aircraft; a photograph is not. Biologists later count every animal in every photograph, adults and calves, group by group. In 2024, 29 groups totalling about 8,200 caribou were counted this way.

The 2024 survey ran over four consecutive days, July 8 through 11, which gives a sense of the scale of the operation: enough good flying weather, across enough of the range, to find and photograph every group the collars pointed to.

How does a collar count turn into a population estimate?

Through a method called capture-mark-recapture, the same statistical logic used to estimate fish or bird populations from a sample of marked individuals. The proportion of collared animals actually seen and photographed during the survey, compared with the number of active collars known to be in the herd, lets biologists estimate how many uncollared animals are likely moving with them, and from that, the size of the whole herd.

The output is not a single clean number but an estimate with a 90 percent confidence interval, a range the true population is likely to fall within. The July 2024 survey, run this way, produced an estimate of about 8,600 animals.

What did the 2024 count actually find?

An increase, but a modest one against the herd's history. The 8,600 figure represents roughly a 19 percent rise over the previous full survey in July 2022, which had counted about 7,200. Set against the herd's low point of around 5,500 in 2018, the trend since then has been upward more often than not.

Set against the last decade, the number looks different. Biologists put the 2014 count at roughly 14,200, which means the 2024 estimate, despite the gain since 2022, remains about 40 percent below where the herd stood ten years earlier.

Why does the herd get counted this way instead of every year?

Full aerial photo surveys are expensive and weather-dependent, so they have come roughly every two years rather than annually: 2016, 2018, 2020, 2022, 2024, with the next one planned for July 2026. Between full surveys, the same collars keep working on their own, sending GPS locations twice a day and letting biologists follow seasonal range use without another full flight. The two-year photo count resets the baseline; the collar data fills in the years between.

What this tells a visitor, not just a biologist

A fly-in visitor will not see a survey crew, and will not see anything like the full herd on any single day in camp, since even 8,600 animals spread across the range Quebec and Labrador share rarely group up outside that one post-calving window the survey is built around. What a visitor sees on the ground, whether a dozen animals on a ridge or a loose line crossing a river, is a fragment of a number that only an aircraft, a set of collars and four days in July can put together properly.