 In British Columbia, it's not uncommon to see all kinds of wildlife. Sometimes, even in places you don't expect them. Provincial wildlife biologists are tasked with managing wildlife populations. Wildlife is managed on the principle of conservation first, then meeting the sustenance and ceremonial rights of First Nations, and then licensed hunting opportunities. The crucial question is, how many animals are actually out there? In a province this big, fighting the answer is no easy task. For animals that form herds like caribou, mountain goats and sheep, we try to count all of the animals in an area. But for non-herding animals like moose, trying to count each animal would take a staggering amount of time and money. Instead, we do a survey. Because moose have a somewhat predictable use of habitat, we can count them in sample areas, which gives us a good estimate of how many would be found in the entire landscape. We need good snow coverage and cold weather for effective counts, as this makes moose more visible. One survey technique that's commonly used is called a stratified random block count. It sounds complicated, but it's easier to understand if we break it down. Biologists divide the province into population management units. For surveys, we divide these further into smaller blocks, which are usually about 25 square kilometers. Stratified means the blocks are ranked according to how many moose we think will be there. These classes, or strata, predict which blocks will have the most or the least moose. We use two methods to rank blocks. The first is a quick flyover of all the blocks, looking for moose and their tracks. This gives a rough idea of how many moose are there and what blocks fit into which strata. The second method is computer-based. Things like vegetation, elevation, snow load and steepness can help us determine where the moose are most likely to be and classify the blocks into strata. The selection of which blocks to count in each strata is random. Picking some at random is like taking an election poll. It gives you a picture of voter preference without having to contact every single voter. So, a stratified random block count means to divide an area into blocks, determine which blocks are likely to have high and low numbers of moose and randomly choose which blocks to survey. Then we fly over those blocks and count all the moose we can see. We try to find all the moose in the block, spending more time in areas with denser cover and less time in open areas, or circling back to where we've seen tracks. When someone spots a moose, we circle it to know its location, gender and age before continuing on. We often fly back and forth across a block in straight parallel lines to ensure we cover the whole area. The gold lines on this map represent the flight path taken during one survey. Our main goal is to get an accurate population estimate. We're more successful at this when we focus our search in the blocks where more moose will be and spend less time in blocks where we expect there to be fewer moose. When we've surveyed the randomly selected blocks, it's time to compile the results. Using moose counted from surveyed blocks, we can predict moose numbers for the rest of the blocks in the same strata that weren't counted, and we do the same for each strata. Add it all up and we get a total population estimate. Now, there are always some moose we can't see because the trees and bushes are blocking them from our view. We account for these hidden moose using something called a sightability correction factor. There are a few ways to calculate this, but they're all based on the idea that we're more likely to see moose in open areas than in dense vegetation. Therefore, most who can see in dense vegetation count for more than moose we see in open areas. For example, if we estimate 800 moose in the survey area, but we think we've missed seeing 20% of the moose that are actually there, we'd estimate there are actually 1,000 moose, we just couldn't see 200 of them. What we're looking for in a survey is an accurate estimate, not an exact number. Every step of the process, from stratifying blocks to the actual count to the sightability correction factor, has a little bit of uncertainty. We acknowledge that uncertainty by showing our final estimate as a range. So if we estimate 1,000 moose but calculate our uncertainty to be 20%, we would express that by saying the total population estimate is 1,000 plus or minus 200 moose. This tells us that the true number of moose in the area is between 800 and 1,200. We estimated 1,000, but it's give or take 200. These survey methods give us accurate estimates that are close to the true numbers of animals on the land without counting every single one. Watch the next episode to find out what we learn from the surveys and how that impacts management and hunting.