While studying a large lagoon in Papua New Guinea, meter by meter, Donald Potts decided there had to be a better way to do ecology. A new tuna factory on the lagoon's shores threatened to drastically change the lagoon before Potts could capture a scant fraction of its initial state. Returning to the University of California, Santa Cruz, where he is a professor of ecology and evolutionary biology, Potts was determined to find a faster and more comprehensive way to study ecosystems.

William Pickles, a physicist at Lawrence Livermore National Lab (LLNL), had the answer Potts was looking for: hyperspectral remote sensing. Usually done from aircraft, hyperspectral imaging allows scientists to take pictures of large areas of the Earth from high above. Other imaging techniques are deployed on satellites, but hyperspectral images capture far more information than most satellite images. Potts found he could receive sweeping, detailed data without ever leaving his office, and could potentially detect changes in the environment before they became apparent to the human eye.

That was five years ago. Potts and Pickles, along with UCSC Earth sciences professor Eli Silver, have been working together ever since.

"For the first time, we're able to see the full complexity of these very complicated systems. Up until now, you could only see a small part of the system at one time," Potts said.

The researchers first used hyperspectral imaging to study geothermal activity in Long Valley Caldera, east of the Sierra Nevada, in a project led by Earth sciences graduate student Brigette Martini. "We picked Long Valley as kind of a testing ground because it has been very well mapped," Silver said.

Hyperspectral imaging is revealing new details about this area, including previously unrecognized fault lines. Martini was also able to characterize activity along these fault lines, particularly the movement of hot water and the release of carbon dioxide.

Most recently, Martini discovered a new way to identify areas where high amounts of carbon dioxide are coming directly out of the ground, often causing "tree-kill zones" because of their lethal effect on trees. These areas are potentially dangerous to humans and other animals.

"This result really blew us away," Silver said. "Now we can go to many other areas that aren't well studied and just image the ground to look for high carbon dioxide signatures."

Potts's team is also completing a detailed mapping of a 650-square-kilometer region of Monterey Bay, including the Elkhorn Slough estuary and its watershed. Ocean sciences graduate student Susan Cochran initiated the project, which is being continued by Stacy Jupiter, a graduate student in ecology and evolutionary biology. Imaging the area took just 45 minutes, but interpreting the more than 250 million pixels worth of data has taken two years. Once they have a baseline map, the scientists will re-image the area periodically, and compare the new data with the old data, pixel by pixel.

"This is a very powerful tool for change analysis," Potts said.

Hyperspectral instruments, which are currently carried primarily on airplanes, scan the ground beneath and pick up the characteristic light signatures of the plants, microbes, and minerals below. Sunlight is made up of an array of colors, or wavelengths of light--some visible to the eye and some invisible. Different materials reflect different wavelengths from this array, creating a unique pattern, called a spectrum. Hyperspectral sensors detect at least 100 unique bands of this spectrum, compared to the 7 to 15 bands detected by most current satellite-borne instruments.

Using hyperspectral imaging, scientists can map the boundaries of different kinds of surfaces; they can identify many minerals and some species of vegetation, algae, and microbial films and chart their distributions; they can even learn about the health of distant organisms. For example, healthy plants absorb a lot of red light and reflect strongly in the green and infrared, whereas dying plants reflect light across most wavelengths.

Potts's team is using hyperspectral imaging to assess coastal marine ecosystems and land margins and to track changes caused by natural and anthropogenic effects. "There's a general sense in the scientific community that these shallow coastal margins are where regional and global changes are going to be felt first," Potts said.

Potts and ocean sciences graduate student Daria Siciliano are also part of a multi-agency effort to assess and monitor habitats in the northwest Hawaiian Islands. Called the Northwest Hawaiian Islands Rapid Assessment and Monitoring Program (NOW-RAMP), the project joins UCSC with the U.S. Fish and Wildlife Service, the National Oceanic and Atmospheric Administration, the Hawaii Division of Aquatic Resources, the Bishop Museum of Honolulu, and the University of Hawaii.

The northwest Hawaiian Islands are 10 uninhabited atolls, or low-lying coral reefs, that represent about 69 percent of the total area of coral reefs in the United States. The outermost atoll, Kure, is almost 2,000 kilometers (more than 1,200 miles) from Honolulu and has no infrastructure, making it difficult and expensive to explore from the ground.

"At the moment, they are virtually unexplored. There are no good maps," Potts said.

Potts and Siciliano are using a mixture of remote sensing techniques, including satellite imaging and hyperspectral imaging, to study aspects of the physics, chemistry, biology, and geology of the islands. Siciliano goes out twice a year to "ground-truth," to confirm what they are seeing from the sky and to match unknown spectral signatures to specific species.

Siciliano is using remote sensing to study how fast the Kure atoll is growing, measuring the rate at which polyps and algae deposit calcium carbonate, the skeleton of coral reefs. This calcification process is sensitive to sea temperature and to dissolved carbon dioxide, so changes in calcification rates can signal global climate change.

Siciliano's preliminary results indicate that coral cover and growth rates may be dramatically higher in some of the corals than has been reported in the past 50 years. If this result represents a true change, rather than errors in previous measurements, it may indicate rising sea temperatures, a sign of global warming.

"Worldwide rates of change seem to be accelerating due to human impacts," said Potts, who worries as much about local impacts, like the dumping of sewage and overfishing, as he does about global climate change.

Of 109 countries or major international territories with reef systems, more than 100 are developing countries, where people depend on reef-based economies and are concentrated on the coastlines. Accelerating population growth in these regions means accelerating local impacts, Potts said.

"There is a need to be able to assess and monitor these changes in real time," he said.

Hyperspectral imaging allows almost instantaneous data collection, compared to traditional methods of ecology, where scientists may labor for months or years to get limited data.

"None of the data we are collecting are as detailed as if you had ground crews dedicated to collecting this type of information, but what you lose in the detail, you gain by having equivalent levels of data for every pixel in the image over a large area," Potts said.

"With traditional methods of ecology you never got a sense of the huge scale of an entire estuary or an entire reef," he added. "The ecological phenomena that we're interested in are occurring on the scale of entire systems."

Eventually, Potts hopes to use hyperspectral imaging to monitor rapidly changing areas throughout the world, particularly in developing countries. He may get his first opportunity back in Papau New Guinea, where he is still doing surveys the old-fashioned way, in the water.


Editor's note: Reporters may contact Donald Potts at (831) 459-4417 or potts@biology.ucsc.edu and Eli Silver at (831) 459-2266 or esilver@emerald.ucsc.edu.

Images can be downloaded from the web at http://www.ucsc.edu/news_events/download/.