About Us | Help Videos | Contact Us | Subscriptions
 

CSA News Magazine - Features

New applications of near-infrared spectroscopy in crop and soil analysis

 

doi:10.2134/csa2017.62.0203
  1. Melissae Fellet

Above: A mixed stand of alfalfa and grass in New York. Image courtesy of Ken Paddock. Inset: NIRS forage analysis. Image courtesy of Dairyland Laboratories, Inc. (www.dairylandlabs.net/).

 
 

This article in CSA NEWS

  1. Vol. 62 No. 2, p. 10-11
    unlockOPEN ACCESS
     
    Published: February 13, 2017


 View
 Download
 Alerts
 Permissions
Request Permissions
 Share

Farmers and ranchers send soil and crop samples to commercial laboratories for chemical tests that measure nutrient concentrations as well as protein, fiber, and carbohydrates that influence the nutritional quality of feed. These tests, however, can be time consuming, and some require special equipment and extremely clean conditions.

Near-infrared spectroscopy (NIRS), an analytical technique used for plant, animal, and soil analysis, provides faster and cheaper results, reducing laboratory costs as much as 80%. This method correlates light reflected from a sample with laboratory measurements of carbon, nitrogen, phosphorus, protein, fiber, or carbohydrate concentrations, among other tests. If the target parameter of a new sample falls within the set used for calibration, NIRS correlations can be applied to different spectrometers as well as samples collected from different regions, states, and countries.

Two recent journal articles extend near-infrared spectroscopy to new applications in crop and soil analysis: one in Crop Science estimates the ratio of alfalfa and grass in forage gathered in the northeastern United States, and the other, in the Journal of Environmental Quality, estimates plant-available phosphorus in soil collected around the Chinese countryside.

Alfalfa–Grass Mixture

Forage fields in the northeastern U.S. contain mixtures of alfalfa and grass. Farmers need to know the amount of grass in the fields because the composition of the feed affects dairy cow nutrition and milk production. “Grass has more digestible fiber than alfalfa, but cows digest the fiber in alfalfa faster than that in grass,” says ASA and CSSA Fellow Jerry Cherney, a Cornell University professor and forage specialist. Digestible fiber in an alfalfa–grass mixture is one factor in determining the nutritional quality of feed. “It’s important to harvest a forage field at the right time to get good quality alfalfa and grass for the animals.”

Some commercial forage analysis laboratories in the U.S. estimate species composition based on the amount of fiber or nitrogen-containing compounds in a plant sample, but none of the labs can provide an accurate estimate of the ratio of alfalfa and grass. Attempts to use NIRS to estimate species composition yielded models that were not successful with samples collected in later years.

To gather enough data to develop a broadly applicable NIRS calibration of alfalfa–grass composition, Cherney and his colleagues collected samples from 91 sites in eight New York counties during 2011, 2012, and 2014. Most of the sites were farmer’s fields although some were experimental forage plots. The researchers separated the alfalfa and grass by hand; then they dried and ground each type of plant. Finally, the researchers recombined the alfalfa and grass in mixtures of 0, 20, 40, 60, 80, or 100% alfalfa. Portions of each mixture were also ensiled for 30 days.

Scientists at Dairy One Cooperative, Inc., an agriculture analysis non-profit, measured a NIR spectrum for each dried and ensiled sample. The spectral signals at wavelengths between 1108 and 2492 best correlated to the known mixture percentages from both the freshly dried and ensiled samples, providing a way to precisely and accurately estimate mixture composition from a NIR spectrum.

The researchers validated their model using 98 samples collected in 2015 that were separated, dried, ground, and recombined in random known mixtures. The mixture percentages estimated from the NIR signals still accurately reflected the known composition. And because all three NIR instruments used at Dairy One provided similar results, this calibration could be used at other labs, Cherney says.

The researchers are now testing forage samples from fields in the Midwest to see if the same model can accurately predict the composition of those fields.

Soil Phosphorus

Farmers want to know the amount of phosphorus in their soils to manage their phosphorus applications wisely. Excess phosphorus can seep into runoff and pollute waterways, and unnecessarily adding the nutrient to fields depletes raw phosphorus reserves. In China, more than half the soils are deficient in phosphorus, and at the current rate of consumption, China is expected to run out of clean phosphorus reserves in 20 years.

Laboratory tests of soil samples can provide estimates of phosphorus concentrations dissolved in the soil or bound to sediments. Different tests are best for particular soil conditions. For example, the Mehlich-3P test uses acidic conditions to remove phosphorus from acidic or neutral soil while the Olsen-P test uses slightly alkaline conditions to analyze phosphorus in calcium-rich or alkaline soil.


Rice plants in Hunan province, China. Image courtesy of Marie van Maarschalkerweerd.

 

Near-infrared spectroscopy could provide faster, cheaper, and less laborious phosphorous measurements, but researchers have found that the NIR spectra only weakly correlate with Mehlich-3 P and Olsen-P laboratory measurements.

Cao Weidong from the Chinese Academy of Agricultural Sciences, Peter Holm from the University of Copenhagen, and their colleagues wondered if NIR soil spectra would correlate with a different laboratory analysis: diffusive gradient in thin films (DGT). In this technique, phosphorus passes from moist soil through a filter into an absorbing gel, creating a gradient in the soil that mimics the depletion zone around a plant root. Researchers extract the phosphorus absorbed in the gel, the amount of which reflects the amount of phosphorus available to a plant more accurately than other tests.

For this study, the researchers collected NIR spectra of paddy soils from six locations in rural China. They also performed both Olsen-P and DGT-P tests on each sample. Finally, the researchers analyzed the NIR spectra for regions that statistically correlated to the laboratory phosphorus measurements. Because NIRS does not measure phosphorus directly, any correlations result from secondary interactions between phosphorus and organic components of the soil.

Using DGT-P as a reference provided a stronger correlation than Olsen-P although both correlations were weak (R2 = 0.52 for DGT-P vs. 0.04 for Olsen-P). Performing the analysis with more similar soils could improve the strength of the correlation, but the results would be less generally applicable, says lead author Marie van Maarschalkerweerd, University of Copenhagen and Sino-Danish Center for Education and Research. With a larger and more focused dataset, using DGT-P as a reference for NIR soil spectra might be developed into a reliable method that helps farmers determine if their fields contain enough phosphorus, she says.

 

Footnotes


Comments
Be the first to comment.



Please log in to post a comment.
*Society members, certified professionals, and authors are permitted to comment.