World of Knowledge
Researchers At the Center for Advanced Spatial Technologies Work to Bring Geographic Information to Everyone.
Imagine trying to gather information from a library where the information is stored in different buildings miles away from one another. After driving from one location to another, you discover that the information is also in different languages, and translating it all into the same language could take months–or years.
Welcome to the world of digital spatial data–a wealth of information collected by various local, state and national agencies: Voting districts and census blocks; dams, harbors and springs; ditches and ferry crossings; boat ramps and power plants; churches, factories and post offices; precipitation levels, crops and soil types–lots of data from vastly different sources, written in different languages for different programs and housed at various agencies across the United States.
What many people saw as a massive morass of data to be gathered and slowly sorted through meant something different to Fred Limp and his colleagues of the Center for Advanced Spatial Technologies (CAST): They saw an opportunity. He and a team of researchers at the University of Arkansas created GeoStor, a database that integrates large amounts of information from satellite imagery; local, state and federal agencies; and remote sensing technologies into a useable, readable format that anyone with a Web browser can access.
Working with several geospatial technology companies and many government, private and non-profit agencies, researchers at CAST specializing in database management, geospatial interoperability, remote sensing applications and computer programming brought together a diverse range of information to create GeoStor.
“We built the glue that plugs everything together,” Limp said. “It’s a research effort that empowers other research efforts.”
GeoStor has changed a researcher’s relationship to data gathering. Previously, scientists followed what was called the 80/20 rule: Data collection took up 80 percent of the time, and research using the data 20 percent of the time. GeoStor enables researchers to reverse that ratio and increase their research productivity.
“You used to have to be a specialist just to get started,” Limp said. “But we’ve lowered the complexity of getting there.”
CAST uses Global Positioning Systems (GPS), remote sensing technology and a variety of mapping techniques in its day-to-day research. GPS relies on 24 satellites stationed around the globe about 12,500 miles up. These satellites broadcast signals that a GPS receiver on earth can pick up. The receiver triangulates its position using the satellite signals and allows a person to find his or her exact location on Earth.
As the remote sensing satellites revolve around Earth, they collect a wealth of data about its surface. Instruments aboard the satellites collect detailed information that can be used by scientists to determine soil types, crop types, slopes, water quality and chlorophyll content of forest canopies–among other things.
To develop maps containing specific types of information, CAST researchers collect “ground truth” data. They visit different sites in Arkansas, determining their exact location using GPS and collecting visual information on the landscape. They take this information back to the office and incorporate it into a computer representation of the image taken by the satellite. They can then tell the computer to use the “ground truth” information to seek other similar areas on the satellite image. In this way, they can identify all the cotton or rice fields or all the oak forests in Arkansas.
The researchers also use databases of information from other sources–Army Corps of Engineers, U.S. Geological Survey, Arkansas Highway and Transportation Depart-ment–to create comprehensive maps. They compile all the information into one easy-to-use format that has become GeoStor.
The result is that when people point their browsers at GeoStor, they can find out the location of all drive-in theaters in the state; the types of crops being grown near Conway; flood zones or hospitals; cliffs and mines; schools and churches. They can find city streets and county roads, dams, rivers and ditches. People can view the information in map form or as tables. GeoStor also links to other information sources, such as data on specific schools.
GeoStor represents the tip of the iceberg in terms of the work done by CAST researchers. At any given time, CAST juggles numerous research, education and outreach projects, ranging from mapping the land-use and land-cover in Arkansas to helping the state government redraw political districts in the state.
Researchers working with CAST have worked with the Nature Conservancy to create maps of endangered birds in Latin America, to determine conservation priorities. They have worked with the National Park Service to create easy access to archaeological information online for federal agencies.
They have worked on projects with other federal agencies, including NASA and the U.S. Geological Survey, and with state agencies, including the Arkansas Department of Information Services and the Arkansas Soil and Water Commission.
CAST employs between 35 and 45 people, including graduate students, high school students and professional staff. The Center works with people across the University of Arkansas campus in various disciplines, offering researchers the latest in innovative technologies. This has led to interdisciplinary cross-collaboration in fields as diverse as engineering, agriculture, anthropology and sociology. By doing this, researchers at CAST have opened up new avenues of research for University of Arkansas scientists and scholars, creating opportunities to look at whole fields in a different light.
CAST Gets All Wet
Indrajeet Chaubey, assistant professor of biological and agricultural engineering, studies water quality issues that affect the drinking supply for almost 300,000 people–Beaver Lake. Traditional water quality methods require researchers to sample water at several points on the lake, bring it back to a laboratory, and spend days analyzing its contents. Because of the time, money and complexity involved, finding sources of pollution in the watershed or determining the effects of a new subdivision or farm on water quality seem nearly unattainable.
But Chaubey intends to attain these goals by using remote sensing data. Working with CAST, Chaubey and his graduate students plan to map the water quality of Beaver Lake using satellite imagery and ground truth data.
“The satellite can give me information on the whole lake,” Chaubey said.
The researchers know when the satellite will snap pictures of Beaver Lake, measuring its reflectivity. They will visit the lake within 24 hours of the satellite passage, together with researchers from the U.S. Geological Survey, and collect water samples at strategic points along the lake for analysis. They will analyze the water for turbidity, sedimentation, chorophyll, nutrient content and nitrogen and phosphorous levels. They will sample at different times of year because water quality varies seasonally.
They can then correlate these “ground truth” data with the reflectivity data from the satellite to create a detailed model of the lake’s water quality.
This model could be used to predict the water quality of the lake given a set of circumstances such as drought or flooding.
Chaubey is also working with CAST to develop a decision support system for government officials and land owners so they can have detailed information about how land use in the Beaver Lake watershed affects water quality in the lake.
“Right now we don’t have a single model that tells us all the things we want to know,” Chaubey said.
Chaubey will use data from CAST to integrate three water quality models–one that looks at runoff from land to stream, one that examines runoff from stream to lake, and a third that looks at the lake’s response to runoff.
Geographic Information Systems (GIS) maps offer detailed information on soil, land use and topography. Chaubey and his students will use these data in the integrated models and re-convert it into GIS so that decision makers can use the information to make decisions about the watershed.
The Environmental Protection Agency, the Arkansas Department of Environmental Quality and the Arkansas Soil and Water Commission, among others, would use such technology to determine where they need to put their resources. And farmers, city planners and other citizens could use the system to minimize impact on water quality when changing farming practices, building new subdivisions or siting landfills.
Chaubey and his colleagues Marty Matlock, assistant professor of biological and agricultural engineering, and Brian Haggard, adjunct assistant professor and hydrologist for the U.S. Department of Agriculture, have also received funding for a project in the Eucha Watershed that spans Arkansas and Oklahoma. The new project, funded by the USDA Nutrient Science for Improved Watershed Management program, will again make use of the expertise at CAST.
High-Tech Data May Help Low-Income Households
Arkansas scores low in terms of technology: The state has the second lowest percentage of households with telephones (88.7 percent), the second lowest percentage of households with a computer (29.8 percent) and the second lowest percentage of households with Internet access (14.7 percent) in the United States.
Despite this technological lag, Joe Schriver, director of the School of Social Work, believes that technology actually holds some of the answers to Arkansas’ problems with poverty. Working with CAST, he hopes to use geospatial information to understand poverty in Arkansas and develop policies and practices that will help alleviate adverse economic conditions.
“Everyone assumes that the high poverty rates in Arkansas are a given,” said Schriver. “But we see a loss of human and social capital that could be of benefit to the state.”
Schriver wants to take technology where it has not gone before: To make a difference to populations that can’t even access it.
CAST has a wealth of information that social workers, legislators and other decision makers can use to help guide people out of poverty.
“The data show there are really two states of Arkansas. There’s an affluent Arkansas and a large group of people who survive in the context of tremendous poverty,” Schriver said. “The social, health, economic and education indicators reflect a picture that has to change.”
CAST researchers can analyze data to show the highest concentrations of households without telephones. They can show the locations of social workers, the locations and characteristics of schools, the percentage of the population living in poverty, immigration and emigration statistics and ethnic diversity.
Understanding where certain populations or conditions in the state are located will allow social workers and others to see what currently exists and determine what needs to exist, Schriver said.
School of Social Work researchers will also work with CAST to access global data to study other approaches to social and economic development that might inform change in Arkansas.
“We want to take this technology and make it make a difference to populations that don’t have access to this kind of information and have not benefited from it,” Schriver said.
A Boring Dilemma
A big brown beetle has begun to make its presence known in the Ozark National Forest by eating oaks from the inside out and slowly killing them.
The red oak borer, a little-studied insect that lives most of its life cycle inside of oak trees, has in recent years proliferated, causing the death of thousands of trees in the Ozark forest. Researchers with the U.S. Forest Service first began noticing the tree deaths in 1999 near Clarksville, Ark., and the problem has increased since then. Scientists estimate that the inch-long insect has done about $1 billion worth of damage to commercial timber in Arkansas and Missouri in the past three years.
Fred Stephen, University Professor of entomology, is working with CAST to track the borer’s past, present and future path as it chews its way through forests. Eventually the researchers plan to produce maps that show high-risk areas, so land owners can get information on the risk of mortality their oaks face.
The borers pose a problem researchers rarely see: native to the Ozark mountains, the insects have never before caused the kind of devastation seen in recent years, killing tens of thousands of trees.
“All of a sudden, for unknown reasons, it has reached this horrendous density,” Stephen said.
The insects have a two-year life cycle, most of which takes place inside the tree. The adult females lay their eggs under the lichens on the bark. The eggs hatch into larvae that bore into the bark and carve galleries in the phloem, a soft layer of woody tissue that lies beneath the bark. They later migrate into the heart of the tree, chewing holes in the middle.
Normally the trees can defend themselves against a few invaders, but the borer’s mysterious population explosion has made resistance futile. Before the oak decline began in earnest, a tree infested with red oak borers might have three or four adults emerging from its interior; researchers recorded the highest number of adults at 71. Now researchers have seen 700 to 800 adults emerging from a single tree.
“It turns the trees into Swiss cheese,” Stephen said.
In spring, the trees appear to be normal, sporting green foliage through the wet, cool weather. By August, however, the diseased trees begin to show their true colors–the top leaves turn brown, crumble and fall off, indicating the trees have become stressed.
Stephen and his colleagues believe that the advanced age and density of the oaks, the drought stress of the past several years, combined with diseases and secondary insects, have created an “oak decline” event that has weakened trees to the point that red oak borer can easily kill them.
Satellite imagery available through CAST can detect the amount of chlorophyll present in a given area–mapping its presence and absence. It can also give the researchers detailed data on the slope, aspect, vegetation type, soil type and ridge top and valley locations in the landscape. This information, coupled with information on the age, density, diameter and vegetation composition of the forest, collected in collaboration with the U.S. Forest Service, will be used to determine which forests have the heaviest infestations of this insect and to see if it’s possible to predict its spread.
Stephen and his graduate students will also dissect the trees to look at the beetles’ life stages, hoping to get a better idea of why an insect that once lived relatively peaceably in the Ozarks ecosystem has suddenly become a killer.