We can use cameras to estimate the abundance of various species in a given area (which we refer to as density). The important components of this calculation are the length of time the camera is operating, the total duration (time) over which each species is present in the camera field of view, and the size of area that is sampled by each camera.

This figure shows the density estimates at individual camera deployments. On the y-axis (vertical axis) of each graph, you'll see the number of cameras. On the x-axis (horizontal axis), you'll see the estimated density of each species for which data were collected. The pattern, repeated for each species, shows that low densities are associated with the majority of cameras, because most cameras are triggered infrequently by an animal passing by (it's one camera in a big area). Occasionally we see a high density for a given species at an individual camera deployment that has been triggered a lot e.g. Canada Lynx. 

Over half of the deployments did not detect Canada Lynx, and the density was less than 0.1 individuals per km² in many other deployments (which translates to a handful of images over the course of the deployment duration). However, one deployment detected lynx nine times, yielding nearly 50 photos, which resulted in an estimated density of 0.8 individuals per km².

We can also calculate animal densities on each trapline. We use a simulation technique to estimate the uncertainty associated with the density estimates, reflected by the error bars around each point. In this context, ‘uncertainty’ reflects our confidence (or lack of) in each estimate. The wider the bar around a point, the less certain an estimate is. Each point can be interpreted as an estimated density for that species within that trapline, with 90% confidence that the true density value of the species lies somewhere within the range represented by the bar. 
 

When the error bars of two estimates do not overlap, we can be reasonably confident that the estimates are different. For example, we can see the error bars do not overlap in the density of Moose at Traplines 2 and 3; we can interpret this to mean that Trapline 3 likely has more moose. However, the estimate for Moose density on Trapline 1 is highly uncertain so we cannot be confident that Moose are more or less common on Trapline 1 than on either of the other two traplines. Conversely, it appears that Black Bears are most abundant on Trapline 1, less so on Trapline 2, and are the least abundant on Trapline 3, since the error bars do not overlap.