Greener Greens and Fairer Fairways

Michael Fink
May 01, 1997

Imagine you're a golf-course manager. You drive a special golf cart, laden with the latest soil diagnostic equipment, out to the 16th green. The grass here has been problematic, so you hook up the fiberoptic probe and push it deep into the soil. In just a few minutes the computer prints up a full report including nutrient levels, moisture content, and percentages of several soil components (sand, silt, and clay) at various increments of depth. It turns out that the nitrogen levels are too low in the surface layers, so you make a mental note to add a little to the 16th green during the next fertilization.


Is your green ailing? NIRS might soon make golf courses healthier and greener.

Sounds simple enough – right? It would be if such diagnostic equipment were currently available. And according to André-Ann Couillard, you may not have long to wait. Couillard, a doctoral student in agronomy, believes that by adapting near infrared reflectance spectroscopy (NIRS), soil analysis might soon be this easy.

NIRS works like this: A special lamp emits energy into a grating that separates it into a rainbow of wavelengths between 1100 and 2500 nanometers. The energy is focused onto a sample of material, and detectors read the patterns of reflectance. NIRS is already a valuable tool with many current applications. Large foodstore chains use it to test moisture content in meat. Other companies test animal feed for digestability. One company scans the green portion of turf grass to check nutrient levels.

Until recently, however, one limitation on the use of NIRS has been the need for well-mixed and uniform samples. Here is where Couillard's research differs from previous studies. Instead of mixing the soil layers into a uniform sample, Couillard scans the undisturbed soil cross-section in half-inch portions. Her detailed scrutiny reveals characteristics of a soil sample that most labs would miss. "When you take a soil sample and send it to a lab, they dry, mix, and then test it," explains Couillard. "This averages out the various traits you may wish to examine closely." Scanning small sections of intact soil allows Couillard to identify high and low concentrations of nutrients, peaks in pH, or anomalies in the soil structure, at any soil depth. Problems encountered at a specific layer, she notes, may have a huge effect on plant growth, yet go undetected using conventional testing procedures.

Couillard's most recent work focuses on calibrating the NIRS equipment. She scans soil samples from a given region and feeds the information into a computer program. The samples are then laboratory tested, and the lab values are compared with the scanned data. Couillard then calibrates the program and stores the information in a database for future NIRS testing. "Once you compile a large enough data base," she says, "the computer can recognize most of the samples and you can feel confident with the results. If the computer doesn't recognize a sample, it will predict the characteristics, and I'll verify them with laboratory testing. Again, the new data are used for calibration and then stored in the database." Currently Couillard has data from several hundred samples.

Couillard admits that there is still much to do to perfect the system, and its accuracy is largely dependent on the quality of the lab data used for calibration. "There have been varying degrees of success depending on what characteristics you examine," she explains. "In addition, there are thousands of variations in soil profiles, so you must recalibrate whenever you enter a new region with unusual or untested profiles."

Still, Couillard says that she can often achieve higher than 90 percent correlation with lab-tested values, and the accuracy grows as the database expands. Furthermore, by limiting areas of use (to golf courses in a certain region, for example) she can quickly calibrate for the important characteristics, then use the NIRS scanning techniques to make recommendations for the care of those areas.

"NIRS is a powerful tool," says Couillard. "Its efficiency and accuracy could save huge sums of money for companies that need accurate information for high maintenance applications such as golf courses, parks, and so forth." Such ease in testing could lead to optimum treatment and fertilization, and a substantial reduction in runoff of unnecessary chemicals.

André-Anne Couillard is a doctoral student in agronomy in the College of Agricultural Sciences. Her adviser is A. J. Turgeon, Ph.D., professor of agronomy, 116 Ag Sciences and Industry Bldg., University Park, PA 16802; 814-865-6541; at2@ Her project was supported in part by the Commonwealth of Pennsylvania and the USDA.

Last Updated May 01, 1997