Physicist Knut Stamnes sits at his computer, staring at the riotous swirl of glowing green, blue, orange, and yellow on his screen. This striking image could be an abstract painting, with bright hues sitting amidst a sea of black, but it’s not. It is actually a data map of the Santa Barbara Channel, with the solid black background depicting land and the sea flowing with color.

This data quantifies the chlorophyll and dissolved or floating inorganic matter such as silt, clay or sand in these waters. High chlorophyll levels may indicate an overgrowth of the algae that sits at the base of the ocean food chain. Stimulated by phosphorus from fertilizers, the algae proliferate, choking waters of oxygen and sunlight and killing or contaminating sea life. In 2004, the UN Environment Programme identified 146 oxygen-depleted “dead zones” in the world’s oceans.

But Dr. Stamnes’ research does not focus on algae blooms. He and his colleagues have spent ten years developing methods to accurately interpret information from space—satellite data—on the environment. Their goal has been to develop “virtual modeling” assessment tools to study water quality. “Our model could be used,” says Stamnes, “to develop diagnostic tools that analyze water health and monitor coastal regions where many people live and where most of the world’s fisheries are located.”

To create a model that can reliably “see” what’s in the sea, Stamnes used NASA satellite data that measures the sunlight reflected off the water and back into space. The tricky part, Stamnes explains, is “atmospheric correction” for dust, pollutants from power plants and other sources, and other atmospheric particles. “Up to 90 percent of the signal is generated from stuff in the atmosphere,” he notes.

In order to determine how the atmosphere and the sea scatter and absorb radiation—and create a working computer model—the team decided to crunch data from one specific region. The Santa Barbara Channel proved ideal. NASA satellites have monitored this strip of Pacific Ocean for decades. And University of California Santa Barbara researchers have regularly sampled Channel waters by boat, giving Stamnes’ team water quality information to ground-truth their computer models. Their modeling results were published in October in the International Journal of Remote Sensing.

The last step in making this a fully-functional model is speeding up computer analysis of satellite data. It took three days to analyze one image. But using a multi-processing “neural network” computing system, similar to those used in artificial intelligence, the team amped the process to just 10 minutes per image, 1,000 times faster.

There are other important roles for this work as well. Stamnes’ modeling method could help calculate how much of atmospheric greenhouse gases are absorbed by the oceans. NASA is currently pointing its satellites around the globe, gathering data to help us better understand Earth’s carbon cycle.

This current research grew out of prior studies, also with climate change implications, conducted in a very different location: Barrow, Alaska, 320 miles north of the Arctic Circle. There, Stamnes helped establish a field station as part of the Department of Energy’s Atmospheric Radiation Measurement Program to study the role of clouds in the climate system.

Other research (a collaboration between the University of Alaska where Stamnes taught at the time and a Japanese consortium) used satellite data on reflected radiation to map changes in snow and sea ice and to calculate how radiation from sunlight penetrates ice—very important pieces of the climate puzzle. “The only way we can look at the whole polar region is with satellites—which is why I was drawn to remote sensing,” says Stamnes.

Researchers went out on the tundra to study the types and quality of snow and ice, information needed to develop accurate models from satellite data. An important factor for sea ice stability seems to be how much it melts in the summer. “Old,” re-frozen ice is twice as reflective as new ice, which absorbs radiation much easier. “New” sea ice is therefore much more likely to completely melt again the following summer.

The amount of time snow sits on the ground also affects its ability to reflect or absorb sunlight. Satellite data details the grain size in snow, helping to estimate its “age.” Researchers corroborated that information by sampling snow, and placing it under a microscope to look at its structure: newly-fallen snow has smaller grain size than older snowpack.

Stamnes hopes to return to his snow and ice studies in the Arctic later this year, while continuing water quality modeling research. Remote sensing has the potential to help us answer important environmental questions, he says, and the ability to reliably analyze data from space is extremely important in monitoring the poles and the oceans and helping us understand and quantify climate change.

By Sharon Guynup

sharonguynup@mac.com

For more information, please contact:

Knut Stamnes
Professor
Burchard
Room 712
Phone: 201.216.8194
Fax: 201.216.5638
kstamnes@stevens.edu