Not all wood is created equal.
We know this by looking at the grain on a piece of cut timber. But what’s going on inside the wood to make it that way—and how do you measure that at such a scale that captures the variability of what’s being grown?
This is what Joe Dahlen’s lab at the University of Georgia Warnell School of Forestry and Natural Resources is working to determine.
The variability of wood is key to how he conducts his research. As a natural product, no piece is ever the same. This means Dahlen, an associate professor, has had to get creative in developing new and rapid methods for measuring wood properties. In many cases, off-the-shelf equipment is modified or combined with other equipment to suit his needs. In other cases, he’s used circuitry knowledge to design, construct and program his own automated imaging systems.
“We have tools that we’ve modified; other tools that we built ourselves, all of which help us to achieve higher quality research at UGA,” he said. He frequently works with the UGA Instrument Design and Fabrication Shop to modify or fabricate the customized parts he uses in his research.
Before computers, most analysis of wood was done with a microscope or by measuring the density of “whole” samples.
“Wood density is pretty easy to measure, and we now have off-the-shelf equipment that measures density at high resolutions across tree rings,” said Dahlen. “But with microscopes, large sample sizes are hard to achieve because throughput is so slow.”
Other wood properties are much more challenging to measure and can require very expensive equipment. So, Dahlen searches for alternative methods that capture similar information at lower cost.
In one lab, he worked with a graduate student and a staff member to build a system to prepare, scan and analyze tree rings collected from wood disks. They were curious about the link between qualitative stem assessments done on standing trees and how accurately it applied to the wood found within the tree. They developed equipment to scan hundreds of tree cookies, or cross sections, and analyzed the rings.
While research remains to be done in this area, the project was an important step in scaling tree ring data to the log and tree level.
This is just one example of Dahlen’s research to better understand the wood that trees produce. Tree growth is a function of the environment and genetics, some of which can be manipulated through forest management such as species, planting density, thinning intensity, site preparation and fertilization. Factors such as the climate cannot be controlled—but are equally important.
All of these factors influence the best use of wood, whether it’s quality building products, pulp or energy products. A major thread through his work is creating models that identify quality issues before trees are cut and transported to a mill. An emerging goal is to link his work to carbon sequestered in products and in standing trees.
He has two labs, one located within the Forest Resources buildings on South Campus and the other at the Pete Phillips Wood Utilization Plant Sciences Building at Whitehall Forest. The lab on South Campus serves as the main measurement lab, with equipment to measure wood using sound waves, X-rays and visible and non-visible light. The lab is also used to pulp wood and measure the fibers, with the information used by the pulp and paper industry and to link his work with tree physiology. The Whitehall lab is used for sample preparation and includes saws, chippers, sorters and scales, along with equipment for testing large pieces of lumber.
His work feeds into a number of projects. He leads the Wood Quality Consortium, which is a cooperative effort with the forest industry and the U.S. Forest Service to better understand wood coming from managed loblolly pine stands. He’s wrapping up a project with the U.S. Forest Service looking at the wood characteristics of naturally regenerated and planted longleaf pine over a variety of sites. A project with the Department of Energy is advancing the use of machine learning and highly detailed spectral imaging. He’s also starting two projects with the National Science Foundation’s Center for Advanced Forestry Systems looking at silvicultural treatment effects on lumber quality and methods of measuring wood density on trees before they are cut.
As his lab has started to generate more and more data, he is turning to artificial intelligence and machine and deep learning to help analyze and better understand the data. By exploring different analysis methods, it allows them to record properties not visible to the naked eye, and computer programs allow them to process more data.
“You hear ‘big data,’ but somebody needs to generate big data—in most cases you can’t just beam it from the cloud, right?” he said. “The subject area is interesting, but a big challenge I see is building up the scientific infrastructure to be able to answer a lot of the questions.”
Now, with an array of instruments that take measurements from the macro to the molecular level, he’s on the cutting edge of exploring these answers. “We need to be able to get good data to answer what’s happening in the trees, but we can’t do that unless we have automated equipment and we’re pretty efficient in the field and the lab.”