A fall 2014 study by Cornell University indicates that more than 85 percent of the alfalfa grown in New York state is planted in a mix with perennial grass; 97 percent in the NNY region alone. These mixed forage crops can be harvested as many as four times in one growing season to build feed inventory on dairy farms.
Can Artificial Intelligence Improve Forage Evaluation? NNYADP, Cornell pioneering new field assessment tool
Northern NY. The Northern New York Agricultural Development Program and Cornell University are combining forage science with visual recognition technology to develop a tool to help dairy farmers feed their cattle.
To evaluate the composition of alfalfa-grass crop mixes so popular in the Northeast for feeding dairy cows, Cornell University Animal Science Associate Professor Debbie J. Cherney is pioneering the application of technology similar to facial recognition software used by government and law enforcement agencies.
The farmer-driven Northern New York Agricultural Development Program, a research and technical assistance program for Clinton, Essex, Franklin, Jefferson, Lewis and St. Lawrence counties, is funding the research and development project.
‘Essentially, until now, we have not had a quick or effective way to estimate the alfalfa percentage in a mixed stand prior to harvest,’ says Cherney. ‘Our goal is to develop a user-friendly application that accurately assesses digital images, taken by camera or smartphone; generates crop percentages, and estimates optimal timing for harvesting high quality forage.’
The Northern New York Agricultural Development Program grant also provided funding for Cherney to collaborate with Dairy One testing laboratories in Ithaca, NY, to calibrate the use of near-infrared reflectance spectroscopy, or NIRS, equipment for post-harvest evaluation of forage composition.
Farmers can use post-harvest alfalfa-grass percentages data to plan harvesting schedules the following year, assess stand quality for impact by pests or disease, and enhance nutrient management for optimal fertility, crop production, and environmental stewardship.
This NNYADP research is also of value to beef cattle producers feeding high-forage diets.
Funding for the Northern New York Agricultural Development Program is supported by the New York State Senate and administered through the New York State Department of Agriculture and Markets.
MORE INFO
The research team checked the precision of computer-assisted calculation of pre-harvest alfalfa-grass percentages by collecting, hand-separating, and drying field samples of the photographed crops to determine actual physical percentages.
Photos and forage samples were collected on farms in all six northern NY counties.
Early results in 2014 indicated that the specific grass species in the field mix and seasonal conditions impacted the ability of the software to accurately estimate percentages.
The data generated from analysis of the cell phone photos and field samples collected in 2015 is currently being used to calibrate the pre-harvest evaluation tool Cherney has dubbed AGES: Alfalfa Grass Evaluation System.
‘We want to achieve the highest degree of accuracy; incorrectly estimating crop composition by just 20 percent can result in late harvesting by five or more days, potentially leading to fiber digestibility that is more than 5 percentage units past the desired target for milk production efficiency. That level of inaccuracy can reduce production by 2.5 to 5 pounds of milk per cow per day,’ Cherney says.
‘This tool will efficiently identify the best time to harvest high quality forage crops to help farmers enhance cow health and milk production, reduce the need to buy-in additional forage, and ultimately increase net income,’ Cherney adds.
Cherney notes that ‘calibrations for grass percentage in the NIRS post-harvest samples worked well and, equally as important, one calibration worked just as well for both fresh and ensiled forage samples, great simplifying the process for post-harvest evaluation.’
Cherney is currently completing analysis of forage samples from the 2015 growing season. When finalized the pre-harvest AGES tool will be available on the NNYADP website at www.nnyagdev.org.
The research team includes Dr. Elif Karayilanli, a visiting academic fellow form Turkey, and Dr. Keenan McRoberts, who completed his Ph.D. at Cornell while working on this project.
SIDEBAR
10 Reasons Why Pre- and Post-Harvest Mixed Stand Alfalfa-Grass Percentages are Important — Cornell University Associate Professor Debbie J. Cherney
- Required information, along with alfalfa maximum height, for estimating NDF, a measurement for feed analysis.
- Help identify optimum quality harvest date.
- Allow ranking of fields for harvest, based on alfalfa percentage.
- Help decide when to start treating a field like grass from a fertility standpoint.
- Information for deciding when to rotate a field to a different crop.
- Help assess stand deterioration due to alfalfa pest or disease problems.
- Some nutrient recordkeeping software requires input of alfalfa percentages.
- Required information for some forage quality software, such as the alfalfa-grass version of MILK 2006, that estimates production in pounds per ton and pounds per acre.
- May help with feed ration balancing for optimal milk production.
- Support quality control as a check on how representative forage sampling is, e.g., highly variable alfalfa percentages over time indicates unrepresentative sampling.