A National Collaboration Releases Their First Publication

The following piece was written by OAS Communications Coordinator Ryan Bower for the Snapshot Wisconsin newsletter. To subscribe to the newsletter, visit this link.

For the last three years, Snapshot Wisconsin has been contributing to a similar citizen science program called Snapshot USA, and Snapshot USA recently reached an important milestone worth celebrating. They released their first publication! Congratulations, Snapshot USA!

In honor of Snapshot USA reaching this milestone, the Snapshot Wisconsin team wanted to highlight this fellow citizen science project and share with our volunteers a lesser-known way that Snapshot Wisconsin data is being used.

What is Snapshot USA?

Snapshot USA is a national effort to bring together trail camera data from across the country and learn about what drives the distribution of mammal species within the United States. Snapshot USA takes a similar approach to Snapshot Wisconsin, having people classify trail camera photos to generate usable data for science. The main differences are that Snapshot USA is a nationwide effort and is focused entirely on mammals.

Snapshot USA was organized in 2019 by scientists from the North Carolina Museum of Natural Sciences and the Smithsonian Conservation Biology Institute. They asked fellow researchers, citizen science programs (including Snapshot Wisconsin) and private citizens to upload and classify their trail camera photos. Much to everyone’s excitement, over 150 people and programs participated in the effort.

Better yet, people contributed photos from 110 locations across all 50 states, proving that there are people all over this country who value efforts like Snapshot USA and Snapshot Wisconsin.

Snapshot Wisconsin’s Contributions

For our part, the Snapshot Wisconsin program was thrilled to support a fellow citizen science project. We submitted data from 2019 and 2020, and we are working on submitting data from this last year as well. It is important to us to support other programs like Snapshot USA and build up science together.

Every year, Snapshot Wisconsin has contributed data from around 10 of our trail cameras in the Clam Lake elk camera grid. Snapshot USA required at least ten cameras clustered within a 5km area, so only a few areas of our camera grid like the elk camera grids met the requirement.

Despite limitations in which cameras we could include, the Snapshot Wisconsin team is glad that we were able to contribute to this effort at all. Jennifer Stenglein, one of Snapshot Wisconsin’s lead scientists, said, “Snapshot Wisconsin is not set up to have areas with clustered cameras. We made a special exception for the elk grid because the data are used to monitor the growing elk herds across the state. Fortunately, the elk grid matched the minimum requirements to participate.”

Stenglein also mentioned how important it was to her personally that Snapshot Wisconsin contributed to this nationwide effort. “As a scientist, having open data is huge. There is so much trail camera data out there, but it’s [isolated] to specific programs or people. Snapshot USA created a place for trail camera data to come together and be available. That allows scientists to ask questions we couldn’t before, like how climate change is impacting species at a national level.” Stenglein is excited to see what other researchers do with the compiled data.

How Snapshot USA Operates

In addition to sampling populations from across a wider scale than Snapshot Wisconsin, Snapshot USA samples from all major habitats and development zones found within the United States. When a new collaborator joins the program, they select the combination of setting (Urban, Suburban, Rural, Wild, Other) and habitat (Forest, Grassland, Desert, Alpine, Beach, Anthropogenic, Other) that matches their camera site.

Our volunteers may notice that some of these site combinations differ from what Snapshot Wisconsin uses. For example, urban deployment does not fit Snapshot Wisconsin’s criteria for setting up a camera. Stenglein thought that the addition of urban areas adds an interesting element to the dataset, but it shows a fundamental difference in what Snapshot Wisconsin and Snapshot USA are trying to capture.

Next, collaborators upload their photos from a specified time window. In 2019, Snapshot USA collected photos from the 14-week period from August to November. Once uploaded, collaborators could start classifying photos, similarly to how our camera hosts do it.

One important difference is that Snapshot USA puts all their photos through a second round of classification – this time by an expert. Expert review happens within Snapshot Wisconsin as well, but only for the species we’ve learned are classified with lower accuracy. Our accuracy analyses have shown that volunteers do a great job of accurately classifying most species, especially the most common species, so Snapshot Wisconsin only expertly classifies the photos of the hard-to-classify species and rare species. Besides, Snapshot Wisconsin would not be able to expertly classify its 60+ million photos. However, this extra step is possible for a program like Snapshot USA.

“Limiting the time window for data collection is really common in trail camera studies,” said Stenglein. “I don’t know if there is any perfect time window for Snapshot USA to choose, since you will always miss something. However, it does make sense for them to select a window of time. It would be too challenging to collect a whole year’s worth of data, let alone have an expert review.”

Once both rounds of classifications are done, the data are assembled into a package and prepped for release in the form of a new publication. This type of publication is called a “data paper” because its main purpose is to release a new dataset for others to work with.

“It’s a cool, new trend in science for data papers to come out,” said Stenglein. “I’ve seen more effort being put towards proper archiving of data. Researchers can use these datasets to test their own hypotheses and come up with new and exciting insights into wildlife distributions in the USA. I think this is where research needs to be, so it’s encouraging to see this trend.”

2019 Data Is Released

In April 2021, Snapshot USA officially published their 2019 dataset. The paper was published in the scientific journal Ecology and had around 100 different authors.

In total, the dataset included photos from 1,509 cameras across 110 locations, and all 50 states and the District of Columbia contributed data. The dataset had 166,036 observations (photos) and found 83 unique mammal species. Seventeen bird species were also detected, which impressed Stenglein, but the project wasn’t looking for birds, only mammals.

“All together, that’s an impressive number of species detected,” said Stenglein. “Trail cameras aren’t set up to see all species equally. Birds, for example, often spend most of their time above the line of sight of cameras, so capturing 17 species of birds is pretty cool.”

Snapshot Wisconsin’s contributions included sightings of just over 20 of the 83 mammal species found by Snapshot USA. Given the small area that the photos came from, seeing 20 species is a healthy number. If we were able to use more of the grid, that number would have been much higher.

The paper reported that the three most detected species nationwide were white-tailed deer, squirrels and raccoons, in that order. Snapshot Wisconsin’s own data visualization tool, the Data Dashboard, also shows a similar trend, with white-tailed deer and squirrels being the top two species detected in Wisconsin. Racoons weren’t third, but they are high on the list.

Coyotes were the most widespread species detected across the nation, which surprised some of the Snapshot Wisconsin team. However, Stenglein explained, “It may be because there is only one major species of coyote. Deer and other common animals change species as you go across the country. Mule deer, white-tailed deer and black-tailed deer each have different ranges across the U.S.”

What’s Next?

Stenglein was proud of the Snapshot USA team for pulling this effort together. As one of the main researchers for Snapshot Wisconsin, Stenglein knows how much work it is to collect photos from hundreds of sources and extract usable data from them. Stenglein mentioned that it is great to see another citizen science project release their first publication. “Our Snapshot Wisconsin team only has so much capacity to work on decision-support tools, so it is cool to know that these data will be used in more ways and by more people.”

Stenglein also mentioned that there is a second publication in the works already. This publication will release the 2020 dataset. It’s nice to see such a quick turn around time for the second publication.

“The peer review process can easily take months to years,” explained Stenglein, “so there will always be a lag. However, I expect that this first lag will be the biggest. I’ve already seen process improvements on the data uploading side. They’ve moved to a more efficient process, which really helps.”

Stenglein believes Snapshot USA has expanded its data collection to Europe as well for the 2021 season, which could offer some interesting comparisons for researchers.

Stenglein’s final thoughts for the Snapshot USA program were:

“I’m so impressed that they pulled this off. We know from Snapshot Wisconsin how difficult it can be to keep things running smoothly, especially when it comes to IT infrastructure and solutions. I wish Snapshot USA all the luck as they continue to expand their program, and I look forward to working with them each year. What you’ve accomplished is impressive. Remember that.”