Skip to main content Skip to main content

The Habitable Planet: A Systems Approach to Environmental Science 

Ecosystems Interview with Robert Crabtree

Interviewer: How did you first get involved in the ecosystem in science?

BOB: When I was seven years old, and then again when I was fourteen years old, I came to Yellowstone on a vacation and really fell in love with what I perceived as a scientist as God’s ecosystem. I mean this is a vestige left on the earth of what it probably was like before we came in high numbers and impacted things. So as a kid, I really fell in love with the area and always kept it in the back of my head that this would be a great place to do research. Since sixth grade I knew I wanted a PhD in Ecology and I did so. I guess I was lucky in knowing what I wanted to do since I was a kid and in having an opportunity to do so.

Interviewer: Tell me a little about the Center.

BOB: I’m the Chief Scientist and founder of Yellowstone Ecological Research Center where we specialize in long-term studies in the Yellowstone ecosystem. We look at things at large scales and we try to do collaborative research. We bring together a lot of scientists in different disciplines to attack and focus on these large scale ecosystem questions.

Interviewer: What is special about the Yellowstone ecosystem?

BOB: Yellowstone is probably one of the last great wildernesses left on earth. It’s a place where you can go and see what it was like before European man came to this continent. Besides being the world’s first national park, it’s the last temperate ecosystem that’s nearly pristine left on Earth. It also contains some of the last vestiges of what Lewis and Clark saw in their famous expedition two hundred years ago. We have a full complement of all the original species and we have intact patterns and processes over large scales — for example, animal migrations, and large scale fire patterns. So it’s a perfect outdoor laboratory to take advantage of natural policy experiments.

Interviewer: What are some of the other things about Yellowstone that make it special for a study for scientific purposes?

BOB: The Yellowstone ecosystem is a very large area, twenty million acres. And in the center is Yellowstone and Grand Teton National Park, the more pristine systems. If you want to understand anything about science, you have to have a control without impact by humans. Then those national parks are surrounded by human impacts, a beautiful framework, to understand how ecosystems function and how to better protect them by taking advantage of this unique world setting.

Interviewer: Is there any parallel between Yellowstone and tropical forest preserves that are around the world?

BOB: Yes. I think all these great last ecosystems on earth serve as study sites or what we like to call large postage stamps on the surface of the earth. In order to protect our biosphere that’s distributed across our Earth, we have to have these ecosystems that we study and understand in order to understand the whole Earth’s system.

Interviewer: At a very basic level, what is the ecosystem? What are the different levels in Yellowstone?

BOB: The term ecosystems can be very vague and diffuse. But you can can understand ecosystems by thinking about them as complex. They are made up of parts and pieces like species, including us. And they’re also made up of pattern and process like a fire or the flow of a stream. In order to start understanding them and research them, we need to have basic subdivisions of ecosystems. The most common is thinking of an ecosystem as a food web, again, complex, no less complex than the ecosystem itself. We think of things as consumers at the top of that food pyramid, for example, the wolf. And the wolf eats the elk; the next trophic level down, elk eats grass. The next trophic level down, grass lives off of energy from the sun and the nutrients from the soil. So you can think of ecosystems as more of a top down example that I’ve just described, or a bottom up through the trophic levels. That’s a very typical way of how we view ecosystems in order to study them.

Interviewer: Tell us about the Yellowstone food web top and explain what a tophic level is.

BOB: If we understand what a food web is, a trophic level is essentially the organizational division of levels of the food web. You can go from top down or bottom up, the bottom being plants and nutrients and sunlight to create food for essential predators that eat those grasses and predators that eat those primary consumers..

Interviewer: Where does the fox live in the food web?

BOB: An essential missing part of Yellowstone was the wolf that was last taken out or extirpated or locally made extinct in the 1920’s. Sixty years went by and a lot of effort went into bringing back this last missing piece, the wolf. Based on studies done before wolves were reintroduced and after and continuing, we can really take advantage of this ecosystem scale experiment. In 1995 and 1996, after twenty years of effort by a lot of organizations, wolves captured in social groups in Canada were transported down, acclimated in pens, and released. Teams of researchers were ready to find out what was going to happen.

Interviewer: What did you hope to learn?

BOB: A question that scientists often try to answer to understand about ecosystems is this. Are the things at the top of the food chain more important in structuring communities? Or is it things at the bottom of the food chain, or bottom-up effect? This allowed us an experiment from the top down. That’s much easier to do than, for example, reintroducing ants or soil microbes everywhere. The only way to understand food web effects at very large spatial scales is to bring back an easy experiment to conduct at the top of the food chain. The wolves came and the researchers and educators took full advantage of that.

Interviewer: Let’s go back historically. Tell us a little bit about the history of the wolves in this part of the world.

BOB: There had been rumors of the great wonders of Yellowstone from trapping journals and Native American stories. So Congress funded a famous expedition known as the Hayden Expedition in 1871. I think forty thousand dollars worth of funding went toward planning this incredible scientific expedition to document whether or not these stories about wonderful geothermal features and wildlife were actually true. In 1872, a few months after Ferdinand Hayden and his expedition came back, Congress did something that’s really truly American – created the world’s first national park — something that is perpetuating its popularity worldwide ever since. In managing and protecting this park, many things had to happen including tryng to make it more like what the pristine system was. Actually when it was made a park, there were already some big impacts to the park. Fur trapping had occurred. Ungulates like elk and deer were being poached. And the park was actually formed by bringing in the military to help protect it. Soon after that, they moved into a phase of trying to manipulate the park to get it back to a healthy state, which included the removal and eradication of predators. Predators were thought to be bad, taking down and reducing populations of antelope, deer and elk. So, the park as well as organizations across the United States had a big predator eradication campaign going on. And it ended up in extincting wolves, mountain lions — they tried with the coyote, but were unsuccessful. Now we have a chance to reverse that and take advantage of it by understanding the reintroduction of the wolf after management itself removed that species.

Interviewer: What were some of the effects immediately after the last wolf disappeared?

BOB: There wasn’t a lot of research back in the 1920’s and 1930’s; but what’s wonderful about studying in a place like Yellowstone National Park is that we have probably the best long-term historical records of any place in the West. We do have an idea just from sighting records of the rangers who were trained as great natural history biologists and ecologists to be able to document the changes that are perceived to have happened. For example, coyotes were not that commonly seen. Wolves were seen very commonly; fox were seen very commonly and coyotes were seen occasionally. If you look at the sighting records in the 1930’s and 1940’s, obviously there were no wolves, a decrease in the number of fox, but a huge increase in the coyote that took over, at least partially, the ecological niche of the wolf.

Interviewer: What were other effects of the wolf’s disappearance. What about small rodents?

BOB: One thing we’re fortunate in doing in working with other researchers and scientists both in the park and at colleges and universities is to be able to look at some of the research projects that took place before the wolf reintroduction, like the scavenger complex, those species that consume winter coat carcasses and how they were affected by a big increase in the number of elk and deer killed by wolves, a new kind of scavenger or carcass to feed off of. And how an increased population of coyotes released from the absence of wolves were now going to be knocked down in numbers, (and they have been by around fifty percent), and what effects that could have on the food web. For example, we’ve noted a fairly substantial increase in the number of small mammals or rodents like ground squirrels and pocket gophers and voles in those areas where wolves are very dense and they have pretty much eliminated the coyote.

Interviewer: What about the elk?

BOB: The elk is a fairly complicated story in that the park has been managing the elk, and the elk are also managed by hunting practices outside the park in and we do have big impacts from the wolf. But we also have impacts from the great fires of 1988 and from climate change that has certainly lessened the impact of severe winters. In the last decade, we’ve had a series of mild winters that have also affected the elk.

Interviewer: What else have you learned about the effects of the wolf introduction?

BOB: You have the system that was intact with wolves the turn of the century. Now we have wolves back and rediscovering and redefining their niche in the ecosystem. So one of the fundamental questions that scientists have about the wolf reintroduction is whether the wolf is going to fit right back in to the niche it once had or have things changed substantially enough to where their new niche is actually different than what it was at the turn of the century.

Interviewer: Why would you want to answer that question?

BOB: Because ecosystems worldwide have been impacted by things like the removal of the wolf at the turn of the century. If we are to succeed in restoring ecosystems and protecting our terrestrial biosphere and so on, we have to fully understand the effects of human activities that try to restore those ecosystems.

Interviewer: Would you say that all this effort is just to help restoration efforts or is there a pure science involvement too?

BOB: I think the pure science and our conservation goals to restore need to go hand and hand. This is a perfect example by answering that question — is the restoration of a missing species as simple as getting them to re-colonize the area and expecting them they’ll fit right back into that niche? Or is it naïve for us to imagine that things are that simple? The initial results are it’s more complicated that we thought.

We are seeing effects from impacts outside the park, in areas that are more human managed and dominated with agricultural practices and urbanization, for example, that are affecting the Yellowstone ecosystem in such a way that wolves probably aren’t going to fit right back into the niche they once had. We’re seeing effects from, for example, invasive species, that is those plants or pathogens that have come from another continent that have their impact in reshuffling the food web or the trophic system to where wolves won’t have a chance to fit back into where they once were at the turn of the century.

Interviewer: What are some of the measurements that you’re taking with elk?

BOB: One of the main things we need to do as scientists in order to understand the impact of the wolf reintroduction is to take a simple inventory of the species that they directly impact. For example, the main prey of wolves are elk, which constitutes about ninety percent of their diet, but they also prey on big-horn sheep or an occasional mountain goat or a moose or a mule deer or a white-tail deer. And recently they’ve been learning how to prey upon bison. You have one of the most diverse set of ungulates or herbivores ianywhere in the west or the United States. Herbivores include not only the big deer-like creatures like moose and deer and elk, but also the smaller herbivores like ground squirrels and voles. What we’re seeing are increases in some and decreases in others. We need to factor out the reasons for that and see much is from the wolves versus some of the other factors we’re also investigating.

Interviewer: Is it true that right after the wolves were exterminated, the number of elk went way up?

BOB: It’s so complicated. The park rangers actually removed sometimes hundreds and thousands in the ‘50’s and ‘60’s both by shooting and also by trans-locating them to other areas. I don’t know if you know, but we came close to extirpating the elk alongside the near extinction of bison. We greatly reduced elk across North America. Yellowstone was used as a site to capture and trans-locate elk. So we never allowed the system to respond to see what the impact of the loss of the wolf would be. Cow populations went up. Certainly the predation rates on elk were greatly decreased. In fact, the only other predator that was an obligate carnivore on elk was the mountain lion, whose numbers were also greatly reduced. So we probably did have an increase in elk numbers; but we weren’t really monitoring as closely as we should have. And there were many other impacts to that elk population. So, again, a complex system.

Interviewer: You eliminated the top predator at Yellowstone. They were gone for sixty years. Suddenly they come back. What a perfect experiment. How come it isn’t giving you clear answers that you’re looking for?

BOB: What we’re finding out is that it’s very complex. Ecosystems are complex. Many things happen at many levels and many scales. We are seeing some very clear impacts due to the wolf reintroduction; but it’s not clean or as clear as if we were doing an experiment in a laboratory setting where we can clearly control certain factors and not others. Right now, we’re having to take advantage of one treatment in somewhat of a control situation; but there are many other treatments and impacts going on at the same time that are interacting with the reintroduction of the wolf. The greatest charge sent to us as scientists is to try to unravel that and tease apart that complexity and rightly assign the big signal that is coming from the wolf reintroduction on the entire ecosystem and food web of Yellowstone.

Interviewer: Can you talk about the effect on riparian ecosystems.

BOB: The term ‘riparian habitat’ has to do with the vegetation that lives alongside streams, rivers and lakes. So it’s a vegetation that has adapted to the wetter, aquatic systems. Riparian ecosystems often are areas where you see the highest concentration of species or bio-diversity. We call them hot spots. The riparian habitats of Yellowstone probably comprise somewhere around one percent of the total land area but affect over seventy percent of the species in the park.

Interviewer: What is the current health of the riparian systems?

BOB: Riparian ecosystems have changed dramatically starting a few years after the wolf reintroduction. One of the confounding factors has been that in 1995 and 1996 and 1997 we had huge floods right at the time when wolves were reintroduced. What the evidence is initially showing is that both the wolves and the floods have improved the health of the riparian habitats in the park. You know, flooding is a factor that leads to germination and reproduction and growth of cottonwoods, for example. And the evidence also suggests that ungulates or herbivores like elk are spending less time browsing upon willow and so the willow is increasing in its growth and productivity in these riparian habitats.

Interviewer: Describe a willow stand that has been eaten by an elk.

BOB: Early in the winter you’ll see elk herds come in and they’ll be grazing upon the grasses. And then they’ll run into a little willow bush and perhaps nip the new year’s growth. Later in the winter, when everything is snow-covered, the elk will come in and the only thing they really have access to is willow. They’ll often mow an area clean; anything that sticks up above the snow is removed. In other areas, in particular where the wolves have created fear in the elk, where wolves can go in and effectively kill elk, elk might not even be going down into those areas; and willow is allowed to grow up and get high enough to where it escapes browse height from the elk. And we’re seeing this in several areas of the northern part of the park. We call it release height willow.

Interviewer: What would be the evidence you’d look for in order to determine the different effects of the elk?

BOB: When a willow plant is hedged or cropped down by browsing from multiple species, not just elk but moose, for example, that will stimulate a lot of compensatory growth. So you’ll have up to a meter or more of new growth in summer. When you go out in the fall and early winter you’ll see these big, lush willows that are trying to grow, escape browse height, but also gather enough energy from the sun to germinate and send out seeds and grow their populations, become more fit. And then if you go there later in the winter, you’ll see differential browsing by elk and other herbivores. And it’s that complex pattern that scientists are trying to understand for what could be the future for Yellowstone Park and other ecosystems that are trying to be healthy. How we better manage those species is going to determine the fate of those communities.

Interviewer: What are some of the problems that you have in trying to study the riparian systems?

BOB: Riparian areas in habitats are extended over large areas through the entire ecosystem. And we simply don’t have the money or the labor force to go out and examine all these. But we can intensively look at representative samples of those riparian habitats and then turn to our remote sensing data that covers the whole ecosystem and make great inference or extrapolate our results to the whole ecosystem accurately.

Interviewer: Would you define remote sensing.

BOB: Remote sensing is essentially measuring things or sensing things from a remote set of instruments. It could be a camera on the ground or as in our case, sophisticated sensors or advanced cameras being flown on airplanes, but it can also mean images from satellites orbiting the earth looking down on our ecosystem taking data.

Interviewer: Why is Yellowstone one of the places of interest to the remote imaging community?

BOB: The value of the Yellowstone ecosystem is as a reference or a gauge of what the earth should look like. A lot of scientists, including NASA and other federal agencies, like to see their sensors flown over Yellowstone because they know that our group has this rich amount of ground data. And without ground data or what’s called ground truth and validation we really can’t make the most out of these wonderful remote sensing data sets. So a lot of people like to study what’s going on in Yellowstone; plus, it’s just a fun place to be.

Interviewer: What is hyperspectral remote sensing?

BOB: A lot of remote sensing uses really sophisticated cameras that have been our visible view into four to seven spectral channels — often red, green, blue and near red or infrared. Wonderful products have come out of that, in particular, the work horse of all remote sensing instruments called LANDSAT that ecologists have used for decades now to look at land use change and things changing over time. Recently, there has been some fabulous technology developed primarily by NASA called hyperspectral or many spectral channels. You can think of one of these hyperspectral cameras as almost like a biological DNA fingerprint of the landscape. You can examine and look at many features from soil to plant chemistry to even things in the atmosphere by using one of these sophisticated hyperspectral cameras that spread out information into hundreds of spectral channels of which the visible spectra that we’re used to is but a minority of those spectral bands.

Interviewer: What do you gain when you have that many channels available?

BOB: What we learn in the world of science is the more information you have, the more you can learn. We often deal with something called bias — bias in our research. And the more information you have, the more ammunition you have to correct that bias.

For example, one thing extremely important to stream and aquatic repair in ecosystems is the influence of large woody debris. That’s pretty much what it sounds like, trees that fall over, branches that raft down, beaver cut willow and it’s laying around on the gravel beds, floods often distribute a lot of these nutrient-rich dead trees. Remote sensing in the form of hyperspectral cameras are very sensitive to, what’s called the lignan and cellulose. The constituents of dead wood show up almost like a bright spotlight in what’s called a shortwave infrared well out of the visible wavelengths that our human eyes can sense. So if you have a hyperspectral camera, large, woody debris stands out like a spotlight. And we can map them precisely and understand their influence on the dynamic nutrient flow within our repairing stream system.

Interviewer: Can you give me a specific example of how you’re using hyperspectral remote sensing?

BOB: One of the primary ways we’re using a hyperspectral remote sensing is in our studies of the effects of the wolf reintroduction on the riparian habitats. Those species that occupy the riparian habitats are using all those spectral channels. The high fidelity quality of the instrument to not just map willow, but to record its height and the changing amount of large, woody debris, and the decrease in the wet grasses or the increase of the marsh. All of these things interact, for example, they may account for more voles or more squirrels because the habitat has increased. Hyperspectral can even be used to identify individual species of plants as well as community types, whereas a camera with a reduced set of spectral bands can’t. In order to understand unexplained impacts to the food web, plants, mammals, insects, birds — we have to understand fully all the different habitat types they inhabit. And hyperspectral is a wonderful measurement instrument to identify all these different habitat features in a complicated riparian habitat.

Interviewer: What is the relationship to the teams you send out on the ground?

BOB: Right now in our geospatial analysis lab in Bozeman, we’re processing the data from the airborne sensors or the satellite sensors. What’s most important is to get out in the fields in the mountains of Yellowstone like Sarah and other technicians and researchers we have to understand what it is we’re seeing. In order to look at the impacts of the wolves, we have to understand every object and feature in the image in order to bring in our analysis. And we also have to get out there and do things that remote sensing cameras can’t do, for example, count animals and trap small mammals that are the food items of both wolves and coyotes, but also respond to the vegetation that we’re trying to measure with remote sensing. We have to get out there and measure the heights of willow and a variety of other things or else we really can’t use this remote sensing data to understand the impacts of the wolf reintroduction.
Without the research team to go out and make sense of the high tech images, we will not have a handle on what’s going on in the real world. Our research teams have to get out and validate what we’re seeing in the images. And without extensive measurements taken, we just can’t tell the true story.

Interviewer: How do you correlate the actual locations where people are and different places on the image?

BOB: The approach that we take with making the most of our remote sensing data sets comes in two parts. One is to establish long-term data plots where we specifically look at the spectral information on those plots. But we also allow the spectral information, the mapped images, to dictate where we go in the field. We have to specifically look at certain color patterns or certain regions of interest that have something very unique there, find those coordinates, send them to our research teams and then go out and often make discoveries like an unknown released amount of willow or a patch of invasive species. So it comes in two forms, both looking at the spectral information of the remote sensing data in our long-term established plots to relate it to the time stamp of the data sets we have, but also to allow the imagery to go out and discover new sites where things are actually changing.

Interviewer: So how do you examine the images? Do you just look at them?

BOB: There are established geospatial algorithms to identify certain spectral features based on either spectral libraries or known spectral signatures that we go out and collect in the field. It’s kind of like throwing the information into the mix and out pops similar things that you’ve put into it and then maps those similar features.

Interviewer: How do you get the data from a flight?

BOB: We work very closely with the scientists and the engineers on board these airborne flights. A lot of flight planning goes into it, a lot of careful watching of the weather, because if clouds get in the way, we don’t get information at all. So weather permitting, excellent data usually comes out of these types of sensors. Once they get off the plane, it goes into a system where the imagery is linked to real coordinates or what’s called co-registration and geospatial registration. We have to put real coordinates of known surveyed sites on the ground on those maps so we can tell what’s where. And without that, change detection later or bringing in another data set like a historic aerial photo, we just lose out on all that information. Then that data moves into software programs on our computers here at the lab; and we start processing that for information of interest.

Interviewer: What do you do with all this data when it comes back?

BOB: It gets entered into spreadsheets on our databases. They’re summarized and we apply statistical analysis techniques in order to piece together all the players of the story. And then we do more and more sophisticated analysis to try to factor out the basic question that we ecologists face — what are the major causes amongst many known causes that have a bigger impact than others. Ecosystems are multi-causal interactive systems. Rarely is it a single, simple cause that is what we’re looking for. It’s usually complicated. And those factors that are important vary over space and time. We’re looking at patterns over space and time and all that variability and complexity and we’ve assigned it to specific packages or parameters, for example, the impact of floods, the impact of climate change, the impact of wolves, the impact of drought on soil microbes that affect plants, that affect rodents. That’s the basic charge we have as ecologists: trying to look at patterns over space and time and assign them to factors we’re measuring in a field.

Interviewer: Why should the average person care about this data?

BOB: Ecosystems like Yellowstone provide us with a basic life support services that we need as humans – clean air, clean water, and all those beautiful areas that we’ve come to love that have a lot of species that live there like wolves and grizzly bears and wonderful lush willow plug plants. If we don’t have complicated sensors and measurement tools, we can’t protect those systems. A lot of the questions we have as scientists that we use to help inform decision-makers with good scientific information is only going to come from sophisticated and complex research instruments like hyperspectral remote sensing or the AVIRIS sensor. In addition to that, we need big views of those ecosystems that we’re trying to protect and these kinds of airborne and satellite-based measurement systems are really the only tools we have to protect these large areas we call ecosystems.

Series Directory

The Habitable Planet: A Systems Approach to Environmental Science 

Units