Mike Hamilton, director of the James San Jacinto Mountains Reserve in Riverside County, refers to himself as a "wirehead." Others have called him the first "digital naturalist." However he is classified, this son of an electrical engineer is equally comfortable studying a rare plant species or troubleshooting a balky computer. And now, thanks to a 10-year, $40-million grant from the National Science Foundation, he and a host of colleagues will have the opportunity to explore new areas where technology and field research overlap.
Their NSF-funded project is called CENS - the Center for Embedded Networked Sensing - and will involve scientists and engineers from UCLA, UC Riverside, University of Southern California (USC), California Institute of Technology (Caltech), and California State University, Los Angeles. It will focus on designing networks of embedded sensors for use in a range of different environments. Embedded sensors are ubiquitous and have transformed our lives. They're the inexpensive, miniaturized microprocessors that control everything from the antilock braking system in your car (along with dozens of other systems), to the beeper in your waffle iron that lets you know when breakfast is ready.
The CENS Project will seek to extend the use of embedded sensors into four very different scientific endeavors. One CENS team will design sensors to monitor plankton in the ocean. Another will use them to create smart buildings that detect and respond to the first tremors of an earthquake. A third CENS group will look at integrating sensors into buildings or wells to detect and prevent groundwater contamination. And at the James Reserve, the sensors will be used to provide an unprecedented look at the forest ecosystem. As Hamilton noted in announcing the grant: "These devices are so small and cheap we can put hundreds of them throughout a study area. Some will be small mobile robots, while others will have fixed locations, and all will communicate directly to the Internet via wireless Ethernet. The applications for these devices are extremely wide, and the potential benefits are enormous."
Those who know Hamilton shouldn't be too surprised at this latest venture. Throughout his 20-year career with the NRS he has always been at the forefront in exploring how new technologies might be used to advance research. Back in the eighties, he was using an Apple II and a laserdisc to create one of the first virtual tours (his subject, of course, was the James Reserve). Over the last decade, his innovative work with GIS technology has helped local firefighters reduce the destruction caused by periodic wildfires by focusing their mitigation efforts and improving their containment strategies. The reserve itself is powered by a solar energy system he designed, installed, and maintains.
"The science paradigm is completely affected by technology," Hamilton says. "Nobody will deny that. There aren't any Luddite scientists anymore. When I started using GIS in the early eighties, people said, GI what? But by the early nineties, half of our field biology students were taking a GIS course to learn how to use it. Today it's used by anybody who deals with anything that relates to the landscape. The same will be true with these new technologies. They're going to change the way we work."
To see an early prototype of what Hamilton envisions, log on to the James Reserve's website: <http://www.jamesreserve.edu/>. In addition to the usual text and photos, the site incorporates a "Wildlife Observatory" that consists of feeds from a number of other unique sources. Strategically placed video cameras provide close-ups of nest boxes, meal worm feeders, hummingbird feeders, and a "bat apartment." A remote weather station provides real-time feeds on the current temperature, precipitation, and wind patterns at the reserve. There's even a robotic camera that web surfers can control - panning, tilting, and zooming in on objects they want to observe more closely.
Originally installed as part of an NSF education grant, these cameras were designed to help elementary students study field biology from their classrooms. Over time, however, they've served a much larger process: introducing a whole new audience to the reserve system. During the last year, the James website recorded more than a million visits, compared with the 1,500 students, teachers, and scientists who were actually able to visit the reserve during the same period. And though these virtual visitors don't get to smell the fresh mountain air or hike through the woods, in many ways their experiences can be just as rewarding. For example, a researcher interested in the nesting behavior of western bluebirds can learn more about their behavior by reviewing a season of archived "nest-cam" images than from a month of on-site observation.
Hamilton believes a network of embedded sensors will make virtual visits to the reserve even more important. "I see this as really enhancing the time-scale side of doing science," he explains. "In typical field biology, you try to capture significant observations in those periods of time when you have live humans out there, and you have no data in between, except maybe a few expensive satellite images. We tend to ignore some questions because they can't be done in a cost-effective manner. This technology opens up the possibilities for potential new discoveries. If you had the right sensors, you could listen to all the birds singing in the James Reserve every hour for the entire breeding season. That data set is extremely different from anything we've ever collected because it's continuous and it doesn't involve a human, so there isn't the disturbance factor."
To illustrate the potential impact of doing remote fieldwork, Hamilton takes reserve visitors outside and shows them a wooden box nailed to a tree. Though the box is empty right now, he's already working on the equipment that will be installed there. "We call it the MossCam," he says, opening the door to the box and indicating where a camera and processor will be mounted. "It's designed for the remote sensing of external physiological change in this patch of Tortula princeps."
Hamilton points out a smudge of dried moss on a granite rock a few feet from the camera box. "[UC Berkeley professor] Brent Mishler and his students are studying the physiology of this arid moss species. This station will give them the ability to continuously monitor the moss from their lab. They can watch it through all seasons to see how it responds to changes in the environment - primarily precipitation and temperature. Our camera system will be a multispectral camera that will capture visible light and infrared, and it will be a web cam so the data will be collected every thirtieth of a second and uploaded to a server where it can be sent via FTP or viewed as a web page via the Internet." [Editor's note: MossCam is now a reality at - <http://www.jamesreserve.edu/mosscam>.]
He goes on to explain that the remote-sensing characteristics will be directly comparable to weather station data that will be available right next to it. The students will be able to correlate visible infrared light changes in the plant as the moss dries up and rehydrates after being dry. How rapidly that change occurs and how often will be visible to the camera.
The MossCam will serve its purpose well, allowing a team of scientists hundreds of miles away to track the subject. But it's still just a first step in realizing Hamilton's vision. "If this was software," he says with a laugh, "I'd call it Version 0.0. The next step is to cut the line, at least the data line to make the unit wireless ... and it's not a smart sensor yet because it's just going to stream data constantly to a server. A human will still have to make decisions about what data we collect, when we store it and when we don't. There isn't any intelligent processing going on at the sensor itself. That will be the next step."
Hamilton's goal is to have 100 nodes at James within a few years. Some will have specialized uses; others will be general-purpose nodes collecting a wide range of parameters. "One of our collaborators is John Rotenberry [director of the UC Riverside-administered NRS reserves]," he notes. "He's an ornithologist interested in the acoustical data that could be processed for bird identification, location, and interaction. We might have special nodes programmed to identify unique individual songs. By triangulating this reading with other nodes hearing the same sound, we can pinpoint the location of that singing bird, all in real time. Over time we can track the locations of those singing territories as a 'cloud' of presence and absence activities. We can use this data to create a cloud that indicates where the singing male bird spends most of its time. The cloud could be darker towards areas of higher frequency and lighter where frequency is lower. Then we can have these adjacent clouds of other birds so we can see whether individuals are spending a lot of time at the boundaries interacting and defending the edges, or whether they're spending more time in the center defending the higher quality habitat where the female is."
Embedded sensors will also change the scale at which researchers can conduct their studies. GIS resolution is fine for doing landscape-level analysis, but the new systems will let scientists visualize phenomena on a much smaller scale. "The frontier is the ever-shrinking scale of observation, both in time and space. We're taking our scale of observation down from communities of vegetation to what's happening in a single tree or what's happening within a square meter surrounding that vegetation. You need very different tools to do that. And on a time scale, we're going from seasonality - four times a year - to once every minute or once every second."
Hamilton acknowledges that the visualization of such data "is a real challenge because you're dealing with orders of magnitude more data than you were in the past." But he isn't too concerned about being overwhelmed by so much data: "We're developing a small microcomputer that will have the ability to look at the data that comes from the sensors, process it, and then decide whether to store it, send it, or notify a human. That's called 'intelligent processing.' Once we figure out how to classify and determine what it is we're looking for, a lot of these processes can be automated to where the software will tell us when something important is happening."
Hamilton foresees a day when thousands of tiny, inexpensive sensors scattered throughout the forest are constantly monitoring the movement of animals and changes in the environment. Given current technology trends, he thinks it's realistic that soon some of these devices will be the size of beetles. "Each sensor should actually be designed so it is part of the niche we are trying to monitor," he explains. "And sometimes the niche will change over time, so the node really ought to be able to track that as well. I could see the devices going up a tree and coming back down, or going underground, or going under water and coming out again like an amphibian. Ideally, our designs should almost mimic the life forms they're designed to monitor."
Just as the website's Wildlife Observatory has opened up the James Reserve to a global audience much larger than it can hold physically, Hamilton foresees a day in the near future when embedded technology will do the same thing for the entire Natural Reserve System: "We serve a fundamental need in terms of preserving and protecting these areas and making them open to scientists, but these are not undisturbed ecosystems. People are a constant element here, and the success of a reserve is largely judged upon how much it is used. We need to have the ability for science to go on here, but there is a finite carrying capacity at any site before [use of the reserve] begins to change the ecosystem and impact the experiments of other scientists. We're hitting our heads against that already at many of our sites. How can you allow new people to study in areas that are already at capacity? I think the virtual reserve is one way to do that. Developing a passive remote-sensing network that allows researchers to have virtual access is going to be crucial."
Personally, Hamilton sees this latest focus as one he will maintain for the rest of his career. "This will be my last decade, my swan song. I'm planning to see the system used at all of our field stations and to facilitate its implementation at other field stations around the world. After the first three years of the CENS Project, I see the rest of the reserves becoming part of the experiment, and I certainly will see the NRS fully wired in less than ten years. We're at the point where I think we all better fasten our seat belts, 'cause this is going to happen fast." - JB
For more information, contact:
Michael Hamilton
Reserve Director
UC James Reserve
Box 1775
Idyllwild, CA 92549-1775
Phone: 909-659-3811
Email: director@jamesreserve.edu |
