Fraser Shilling has developed a unique way to analyze something many of us would rather not see at all: roadkill.

Shilling is a University of California, Davis, professor who studies transportation and landscape ecology, among other things. Along with website developer Dr. David Waetjen, he created the California Roadkill Observation System, which draws on volunteers who submit observances of animals killed on roadways. The system allows Shilling to gather data that gives great insight into how wildlife species are impacted not just by roads and traffic, but also by California’s ongoing drought.

Unlike most projects that track one species or region, the project provides an opportunity to track many different species across the state of California. In partnership with other systems around the world, there is potential to build a valuable tool to understanding how drought and climate change affect all living things.

Water Deeply: How did you end up tracking roadkill?

Fraser Shilling: I would drive I-5 when I was living in Oregon 20 years ago, when I would come to California to visit my family, and I would see lots of roadkill. Barn owls, actually, were the main thing that I would notice, and I’d wonder what they were until I finally stopped and checked them out. That got my interest going.

Over the last eight years I’ve been associated with the Road Ecology Center here at U.C. Davis, and one of the areas I focus on is transportation ecology, which studies the effects of transportation on natural and human communities. I also look at connecting projects on water and water sustainability and projects on mercury in fish.

A few years ago, in 2009, I asked my programmer (Waetjen) to create the website, the California Roadkill Observation System, so that I could more formally record roadkill observations and provide a way for anyone to do it in California.

Water Deeply: How is data collected?

Shilling: The main way it’s collected is through the website. We also created in the past an app for smartphones and didn’t find there was a lot of penetration with that. It just has to do with who our contributors are. The youngest are in their 20s and the oldest in their 70s; it’s a wide range. A lot of our people are happy to not use an app.

This is a volunteer effort for us as well. We don’t have funding for this. Dr. David Waetjen, who is also a volunteer, created this website for fun, which is kind of neat in a way. Until this year it was the largest system like it in the world. It’s probably got the least funding of any system in the world because it has zero funding. But it’s a really neat volunteer-run system and it’s just he and I, basically, at the center of it with this big cloud of observers helping us with the data collection.

We’d like to get funding because we are trying to go to the next level, and we’re going to need more funding and stability. In the meantime, it’s really an ad hoc effort.

Water Deeply: How do people volunteer information for the database?

Shilling: Primarily they are pulling over and identifying the animal and providing some information of where the animal is, and a picture is optional. About 15 percent of our observations have a picture with them. We use that information to verify species identity. It allows us to make generalizations of how accurate people are being in identifying species, and they tend to be pretty accurate. Anyone can do it, and most of our enthusiastic observers are natural history buffs or biologists.

Water Deeply: Were there any surprises as the data began to amass?

Shilling: The A to Z species that we would get. From acorn woodpecker to zebra-tailed lizard, we get a wide, wide range of California’s biodiversity showing up. In California there are approximately 630 full-time or part-time native land-dwelling vertebrates. Out of those we have over 400 showing up in our database as roadkill. So about two-thirds of California’s biodiversity end up on roads. That was pretty surprising. I think it’s one of the first studies to really find something like that. That such a wide range of species can be impacted in this way.

Water Deeply: Is it possible to tell from your data how drought has affected animal movement patterns?

Shilling: At the beginning of the drought, there was actually an increase in the rate of roadkill. Between 2010 and 2014 in all of these categories, there’s been change, and the primary change has been an increase in the rate of roadkill. The exception is medium-size mammals, which for some reason have been reducing. The interesting thing is between last year and this year, there’s been a decline for mammals and amphibians, and I think this is possibly because we are actually starting to see a decline in wildlife populations in California.

In those five years, the main thing that has been going on that could affect them is the drought. We always have changes in land use and traffic gradually increasing and new road development and so forth. But the biggest change over the last five years is the drought. That being said, all these animals are affected continuously by human activity. There are all these other stressors that are always going on, and the drought is sort of the straw that breaks the camel’s back, because it has such a big effect on vegetation and running water availability.

There’s been an increase up until the last year in the number of mule deer getting hit, and then recently there’s been a decrease. Between 2010 and 2014 there was a doubling in the rate of mule deer getting hit. In 2014 and 2015 it’s gone down to half. What it basically means to me is that it’s not that they are now staying away from roads, but there are fewer of them to be hit.

Amphibians are responding in the same way. I think it was for the same reasons. They were moving around more looking for water, and now there’s just fewer of them.

They move around for different reasons. There’s a lot of daily movement and that’s really the important part here. Animals are moving around to meet their daily needs, and because we have so many roads and so much traffic, quite a lot of them encounter roads as they’re moving around to meet those needs.

Water Deeply: Why do you think medium-sized animals were being hit less when the other classifications were increasing?

Shilling: Most of the medium-sized animals we see are the scavengers of the animals being hit on the side of the road: raccoons, striped skunk, possums. I suspect that since they depend on being able to scavenge, if there’s potentially less stuff to eat, we’re going to get less of them coming out there and getting killed.

Water Deeply: Do you have any ideas about how wildlife can be diverted from roadways?

Shilling: The simplest thing that’s done fairly regularly, but not so much in California, is to build fencing along the roads and highways and provide crossing structures to go underneath roads. Probably the states that are best at this are Rocky Mountain states. Typically you’ll have miles of fencing leading to an earth crossing structure to allow animals to cross from one side of the road to the other.

The most common way is to build an undercrossing, sort of a long tunnel, and there are animals that will use those and we can measure that, but very few species use these and a lot of species will use the overcrossings. So the difference is … the undercrossing will cost maybe half a million and the overcrossing will cost you $5 million. So because we prioritize concrete or humans over wildlife, we just don’t build them. It’s not that we don’t have the money. We have plenty of money. We spend $11 to 12 billion dollars a year on transportation projects in California, but we spend only a couple of million a year on these crossing structures. That’s a 1 in 10,000 ratio, so if we were to up that ratio to say, 10 percent for wildlife, then you would have enough to build these structures and actually do something.

In terms of moving across a road then tunnels, undercrossings, overcrossings — they’re not getting built fast enough.

Water Deeply: What have you discovered about how animals use wildlife corridors?

Shilling: It’s sort of a conservation planning gimmick, in a way. Wildlife don’t really use corridors. There are a few examples of wildlife following the same path year after year, like caribou, or grizzly bears will follow the same trail from one year to the next. But in terms of the way animals move across landscapes, they don’t use them. There’s no such thing as a wildlife corridor.

The complicated answer is that very few mammals are tolerant of coming close enough to the highways to cross it. That doesn’t mean they don’t actually wander through, and that’s why we see them showing up as roadkill, but in terms of regular movement only a few species are willing to do that, so we do have ways we can theoretically funnel the animals from one side to the other, but in practice it’s a lot more challenging than that. What it comes down to is the most effective structures are big wide overcrossings where the animal can barely tell they are going across a road. It’s vegetated and continuous and looks like regular habitat.

Water Deeply: Who is looking at this data, and how can it be applied to other fields of study?

Shilling: We get data requests all the time, and most of them have to do with the roadkill aspect of it. A newer kind of data request we’re getting is that they are not just animals that are coming to the road: They are evidence of a particular species occurring in a particular place on a particular date. That’s really important because in California we don’t have any general wildlife tracking system except for this one. This is California’s biggest wildlife tracking system for a broad range of species. The Department of Fish and Wildlife doesn’t have a system like this. The birders have their own system and have many more observations, but that’s just for birds.

If someone was interested in bobcats, bar owls, mule deer, etc., then our system has the most occurrence data for those species, and we’re starting to use the data in that way where we say, let’s look at how the distribution of these animals is occurring right now, and is that going to change over time. So if you were interested in climate-change effects, this would be one type of data that would be useful to understand how wildlife is impacted by climate change.

We finished our sixth year, and if we go for another 10 years, then you’re starting to talk about a long-term data set, and when you combine that with historical data, you can start to better understand climate-change effects on wildlife.

Water Deeply: Are there other uses, such as tracking invasive species?

Shilling: Yes. There’s a couple of squirrels we’re starting to see, the eastern fox squirrel and eastern grey squirrel, that have been invading different areas of California over the last 20 years. They were first spotted in Southern California, but now we’re seeing them in many areas of the state. We had our first observation of an eastern fox squirrel in the Sierra Nevada from a roadkill observation.

So when we use our roadkill data and compare it to what you might call the official data for these squirrels, we have a little bit of a different picture. We cover more areas in the state than the so-called official database, so what that means is if you really want to track invasion, you’d want to combine the data from these different sources. And for us it pointed to our roadkill data as an important source in understanding ecological processes such as invasion.

Water Deeply: Do you have partnerships with others who are collecting data on roadkill?

Shilling: We tend to chat because we know each other around the world, so we can exchange methods on collection and analysis. We can share how we manage these data sets as they grow. It becomes really important to have good web access to the data and systems for things like smartphone data collection.

There’s a system in Idaho, a couple of smaller systems, and then a lot of state highway departments retrieve carcasses from roadsides and they sometimes record where they retrieved it and what kind of animal it was. Usually it’s deer, moose or elk. So there are a variety of systems around the country, but the ones like this one, there are probably a dozen in the world of this scale. There’s a few in Australia, a few in Europe, one in South Africa, a couple in South America. We actually inspired a lot of those other systems because we were one of the first to operate at this scale. It makes it easier to collaborate because we know each other, and their data systems are similar to ours.

Water Deeply: What are some of these other systems learning?

Shilling: Brazil just did a really nice thing that I’m going to copy. I feel free to do that because they copied our database structure by our encouragement, so now I’m going to copy how they’re doing this one study. They are estimating total roadkill for the country. It numbers in the hundreds of millions per year, so I ‘m going to figure out how to do that for California.

Other people have done this. The Human Society of the U.S. estimates that 1 million vertebrates a day are killed on U.S. roads and highways. That number is very close to Brazil’s number, and Brazil is almost the same size, so it suggests that we are able to make these order of magnitude comparisons on how big of an impact we’re having. That’s important, because if we want to conserve wildlife, then we need to know how we’re killing them, and roads and traffic are one of the primary causes of mortality for almost all species in the country.

Water Deeply: How can people change their own behavior to minimize roadkill?

Shilling: The primary cause of roadkill is traffic speed. You can drive 75 and think you’re aware, but it doesn’t mean you’re going to miss hitting animals. So it really is a question of speed.

In a lot of parks around the world, including in the U.S., speed limits are used as the main way to reduce roadkills and it’s very, very effective. So you could start with enforced speed limits in areas where you have wildlife. If driving the speed limit is going to take you an extra hour, leave earlier.

If we’re not willing to slow down, then it really is a question of building these overpasses. Fifty percent of accidents caused by animals are from swerving to avoid the animal.

One conversation we really won’t hear about much is not only can we stop new road development in certain areas like the Amazon, where it’s incredibly destructive, but can we think about removing roads and restoring roadlessness that really makes habitats.

The front edge of our learning is that we are capable of potentially killing a wide range of species on roads, we are capable of getting a few species safely across the roads, but we are still having an extensive unmitigated impact on wildlife. One thing we’re learning is we’re really not solving the problem, and that’s an important lesson. We really need to know when we’re not doing enough.

Water Deeply: Do people ever give you strange looks when you are photographing roadkill on the side of the road?

Shilling: Surprisingly not. I do get people wanting to know what I’m doing. They know that roadkill is there. They see it. For most people, the primary way they see wildlife is as roadkill. If you think about the last time you saw a fox, it was probably as roadkill. So when people see it and they see someone else looking at it, I don’t think they see it as surprising.

Top image: A volunteer’s photo of a roadkill rabbit, one of many uploaded to the wildlifecrossing.net website. (California Roadkill Observation System)