Researchers can estimate grizzly bear numbers in British Columbia by collecting hair, using DNA to recognise individuals and checking how often the same bears turn up in repeated samples. Those observations help estimate how many animals were present but went undetected. A local survey can estimate density in its study area. A provincial estimate can combine field evidence with models for areas where surveys are limited. The result refers to a place, period and method. It differs from the number of sightings visitors report or the number of bears a guide sees along a river. BC provincial estimate report
Why biologists need an estimate
A grizzly can use forest, riverbanks and other places where it is out of a person’s view for much of the day. An observer at one river might see one bear twice and a second bear only once. Without reliable identification, it is hard to know whether the three sightings describe two bears, three bears or repeated appearances by the same one. Other animals can be present in the area and never enter the observer’s view.
Think of a wildlife study as an attempt to make those gaps measurable. The research team defines the area and period it intends to study. It then records which identifiable bears it detected on each sampling occasion. The results provide a minimum number of known individuals. The pattern of detections and redetections can also help estimate the number that did not leave usable evidence during sampling. That second step requires a statistical method. BC DNA inventory study
How a hair sample can identify an individual
Wildlife researchers can collect hair that a bear leaves behind and extract DNA from suitable samples. They compare genetic markers from separate samples to see which samples are consistent with the same individual. This allows them to identify an animal without physically capturing it during that survey. The collection design matters because researchers need to know when and where samples were collected, what effort went into obtaining them and how the sites represent the area.
A sample by itself cannot give the team a population estimate. It provides an individual detection if it contains usable genetic evidence. Several samples may belong to one animal, and a sample may contain too little usable material to establish identity. Laboratory quality checks are important because a mistaken match can merge records from separate bears, and a mistaken split can make one animal look like two. Either error can affect the number of individuals and how often the same animal appears in the record.
The resulting record links identities with detections. Suppose a bear’s DNA appears at one station early in a survey and later at a second station. The team can record both detections as evidence of the same animal at two places. If several samples at the first location belong to it, those samples add detail about the detection without adding several more bears to the population. Researchers use the distinction between individual animals and pieces of evidence throughout the analysis.
What monitoring in the Koeye watershed shows
In the Koeye watershed on BC’s Central Coast, Heiltsuk-led researchers used non-invasive hair sampling between 2006 and 2009. Their monitoring was guided by Heiltsuk customary law and developed to inform conservation decisions. The example shows how a genetic survey can answer questions defined by the people responsible for a place, including how bears use a salmon watershed over time. Its results apply to its sampling context. A study in one watershed contributes local knowledge and evidence that can inform broader work without replacing it. Heiltsuk monitoring study
What repeated detections tell researchers
The statistical idea is often called capture and recapture. In a hair study, a capture can mean identifying the animal in a sample. A recapture can mean detecting that individual again on another sampling occasion. Researchers use how often the same individuals reappear to learn about the chances of detecting bears under their survey design. The animals identified are only the ones that left evidence the researchers could recognise.
Imagine two simplified surveys with the same number of distinct bears identified. In one, many individuals appear repeatedly across several sampling occasions. In the other, most appear only once. The two records can imply different detection patterns even though their observed totals are the same. A repeated record provides information about how thoroughly the survey may have sampled the animals. A model still has to allow for differences among sites and animals. A bear that often passes a sampling point is easier to detect than one that rarely visits that part of the study area.
Think also about the edge of the study area. A bear can move across it. If a sample comes from near a boundary, the animal may spend much of its time outside the area the researchers are trying to describe. Spatial methods can use where detections occurred, including how far apart detections of the same individual were. That helps researchers address the relationship between the sampling layout and the space the bears use, with assumptions that should be described in the study. Canadian grizzly assessment
Check where sampling was harder
Imagine sampling sites along accessible valley bottoms and a smaller number in rugged upland terrain. A map of detections may appear concentrated along the valley even if the bears sometimes use both settings. One reason is that the survey had more chances to detect them there. Another is that the animals may genuinely use valley habitat more often. The layout and timing of the survey help researchers work through those possibilities. A reader who sees only a map of points could easily mistake the places with more detections for the only places the animals use.
Researchers can also use different kinds of evidence when the question changes. A remote camera may document that bears passed through an area. A collared animal can provide more information about movement between places. A hair sample can support individual identification when the genetic material is suitable. To combine those records, the team needs to decide what each can establish and how much sampling occurred. It would be misleading to add camera detections, collar locations and hair samples as if they were interchangeable counts of animals.
This also explains why an estimate is often given with an uncertainty interval. The interval reflects what the data and model allow the researchers to infer, under their assumptions. A narrow interval can indicate a more precise estimate within the model. It does not mean every source of bias has disappeared. A broad interval may reflect limited detections, uneven sampling or other uncertainty. If two estimates are close, compare the survey designs and intervals before deciding the difference represents a biological change. The headline number is easier to interpret with that context, especially across survey years.
How a local estimate becomes part of a provincial picture
A local density estimate describes how many bears are estimated per unit of area. Areas differ in habitat, food and human use, so applying one density everywhere would hide important variation. The BC report for the 2018 estimate used a predictive density model informed by North American grizzly inventories. Where suitable field inventories existed, they informed estimates for the areas surveyed and reasonable areas beyond them. The province also considered local expertise where model results did not align with local knowledge.
The report presents results in grizzly bear population units. Those are management and reporting areas. They give readers a geographic frame for interpreting estimates and comparing places. A line between two units marks a boundary in the data. It does not necessarily stop an animal moving across the landscape. If you compare estimates from two publications, check whether the areas being described share the same boundaries and whether the units have changed. BC population unit data
The label “2018 estimate” refers to the year the report is estimating, even if the document appeared later. A newer website may reproduce that same estimate and cite the original report. Publication date and estimate year answer different questions. If you are comparing provincial figures, start with the year represented by the number and then check how the underlying data were gathered.
What to look for when estimates change
A change in the published number can reflect a change in the population, an improvement in data or a change in the estimation method. A new survey may replace an earlier extrapolation. A revised model may handle available data differently. Where a report describes updates for only some population units, compare which parts of the estimate were revised before treating the difference as province-wide growth or decline. The 2018 report explains that it incorporated newer field-based information for some units.
A useful reading routine is to ask four questions.
- Where? Check whether the estimate covers a survey grid, a watershed, a population unit or the province.
- When? Identify the years of fieldwork and the period the estimate represents.
- How? Find out whether the number comes from direct field evidence, modelling or a combination.
- How precisely? Look for an interval or discussion of uncertainty and the conditions that could affect detection.
These questions also help you compare numbers reported for nearby places. A local study may be precise for the area it sampled but have limited relevance beyond it. A provincial estimate can cover a far larger region while relying on different levels of evidence across that region. Both can be useful when you keep their scale and purpose attached.
What a guided sighting can contribute to your understanding
The Campbell River grizzly bear expedition visits mainland Toba Inlet and uses designated viewing platforms with local Klahoose First Nation guides. A platform gives you a clear view of part of the habitat when conditions allow. If a bear returns to the same stretch, ask what could help distinguish a repeat sighting from the arrival of a different individual. You may see markings, size differences or behaviour that help a guide interpret the encounter. A brief or obstructed view can leave the identity uncertain.
A good sighting tells you where a bear was visible at a particular time. A lack of sightings tells you about what people could observe during that visit. Neither observation measures the number of bears using the whole valley. If you are choosing where to see bears in the first place, the guide to seeing grizzlies in Canada covers the travel question. Population studies explain how biologists answer a different question about the animals beyond the viewing area.
Questions about counting grizzlies
Why not simply count photographs?
Photographs can document sightings, but the same animal can appear in multiple frames or during multiple visits. Unless observers can distinguish individuals consistently and account for those they missed, a count of photographs does not become a count of bears.
Does finding more hair mean there are more bears?
It may mean that more samples were collected or that conditions produced more usable DNA. Several samples can come from one bear. Identifying individuals and recording the sampling effort help researchers interpret the difference.
Can a population estimate tell me how many bears I will see?
An estimate describes animals across a defined area and period. Your sighting opportunities depend on where you go, how long you watch and which animals happen to be visible. A guide can explain the habitat in front of you. An estimate has a different use across a larger area, such as helping land managers compare the places bears may depend on over time.
