LUKE MARZEN

Marzen researches wetlands using cutting-edge techniques

A multidisciplinary team of scientists from across the Auburn campus, including Luke Marzen, professor of geography, are researching the Alabama wetlands through a $363,800 grant from the National Institutes for Water Resources/U.S. Geological Survey National Competitive Grant Program. The three-year project is titled Inventory, Classification,

and Assessment of Alabama’s Geographically Isolated Wetlands. Marzen is a co-primary investigator on the project and is working with others from Auburn including Sam Fowler, co-primary investigator and director of the Water Resources Center; Al Schotz, an ecologist and biologist with the Environmental Institute; and Michael Barbour, GIS database manager for the Alabama Heritage Program. The statewide project is based

on a partnership agreement between the Environmental Protection Agency Region 4,
the Alabama Department of Environmental Management, and Auburn University. The agreement designates Auburn as a Watershed Center of Excellence. Information collected by the team will support the Alabama Department of Environmental Management’s Wetland Management Program, which strives to strengthen the state’s ability to implement a comprehensive wetlands program by providing baseline data to fill information gaps. The project will inventory, classify, and assess isolated wetlands in Alabama to enhance protection of these biologically diverse ecosystems.

“The wetlands are unique, because as of now, they do
not fall under the protection of the federal government, and a lot of folks are interested in trying to protect them,” said Marzen, whose specialty is in remote sensing and geographic information systems.

Marzen is a physical geographer who has spent much of his career investigating the relationships between human activity and natural environments using remote sensing and GIS. For the wetland project, he has been exploring new methods in geographic object-based image analysis where he fuses aerial satellite imagery with airborne and terrestrial LiDAR.

“Traditionally, the image analysis would have been done by manually digitizing aerial photography. Then, we would have looked at the spectral reflectance of individual pixels and grouped them into things that might be the same, like water. The problem with that method is water looks like other things when you are looking at pixels. Water could easily be confused with pavement or rooftops,” Marzen said. “What we have done is utilize new geospatial technology to try and automate the process. We group the pixels into objects, which is called segmentation, based on several factors and image characteristics such as tone, texture, shape, size, height, and spectral reflectance to extract the wetlands from the imagery. The traditional method is pixel-based, and we are moving from looking at individual pixels to looking at objects. This represents a paradigm shift in remote sensing. It will help us to more accurately map the area and it speeds up the process as well.”

Marzen said the more accurate and faster system of mapping will result in a more conclusive assessment of the state’s natural resources.

“It’s important to have an inventory of our natural resources. We have actually documented all the wetlands in the state, even though the grant is only for the isolated wetlands,” said Marzen. “There is a national wetlands inventory that was produced in the early 80s and is not very accurate. We will share our data with the national inventory as well.”

About Marzen