Some Season 1 Results…

As we get ready to launch season 2 of Snapshot Wisconsin, we are excited to share some initial results from season 1.  But first, thanks to all participants for making season 1 a success!

At least 29 animal species were detected in season 1.  It will likely come as no surprise to anybody that deer, elk, and squirrels and chipmunks were the most frequently observed organisms.

One thing users are often curious about is the level of agreement among Zooniverse classifiers.  There are many metrics that can be used to measure agreement, one of which is shown below: the proportion of votes cast for the ‘consensus’ species.  A value of 1 means that 100% of people who looked at a photo event agreed upon the animal.  A value of less than 1 indicates some level of disagreement. For example, if the value = 0.5, then half of all classifiers thought the image was one species, and the other half thought it was something else.  The chart to the right shows that for most species, more than half of all classifiers agreed. The tougher species tend to be those that are small and move quickly (think mink or weasels), and tend to take poorer quality photos for these reasons. (Please note that images can be  viewed in larger format at the bottom!)


The tough species tend to be small and move quickly (think mink or weasels), and often the picture quality for these organisms is not that great.  

Anyway, elk were a big focus of season 1, and there are some interesting things that popped out of the classifications. One metric that is of general management interest is the cow-calf ratio, which is an index of recruitment (how many calves are being produced). The classified images suggest a ratio of roughly 0.5, so a very rough guess is that on average, a female elk gave birth to about 0.5 calves (obviously, never exactly 0.5, though!).  

Most of the time, cow elk appeared to move in small groups, while bull elk were relatively solitary:


Elk appeared to be primarily crepuscular (most active in morning and evening). Changes in elk activity across time (or from July into early February) provide some of the more interesting results.  Bull elk were markedly more active in autumn; this activity peak corresponds with the rutting period, during which bulls compete for harems of females.  We suspect that many bull elk move off of the camera grid outside of the rutting season.  In contrast, cow elk appear to become less active during the rut.


The spatial distribution of bull and cows across the entire time period suggests moderate overlap, consistent with bulls and  cows using space differently during some times of the year, and coming together during the rut. There also appears to be some partitioning of space between elk and canid species.  Below, we have lumped the frequency of coyote and wolf events across the camera grid, and also present some of the patterns associated with some other common species (bears, deer, turkey).


The patterns we observe in elk activity suggest that, beyond improving existing monitoring,  cameras can teach us quite a bit about how elk use space and behave over different times of year. Although what we present here are simply summary patterns, we are beginning to analyze some of this information more thoroughly. In the long run, things like elk responses to land-cover, changes in plant productivity over the course of a year, and predators appear to be fruitful topics to investigate.


Tags: , , ,

2 responses to “Some Season 1 Results…”

  1. Daniel Nogueira says :

    Awesome results, thanks for the hard work.


  2. Stuart Rapley says :

    So nice to see the results 🙂
    About to start Season Two right now


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: