ARS | Publication request: Estimation of bermudagrass forage intake from canopy spectral absorbance measurements using hyperspectral radiometry

Submitted to: International Grasslands Congress
Publication Type: Proceedings
Publication Acceptance Date: April 12, 2008
Publication Date: June 14, 2008
Citation: Gao, F.Q., Brown, M.A., Wang, X.Z., Starks, P.J., Wu, J.P. 2008. Estimation of bermudagrass forage intake from canopy spectral absorbance measurements using hyperspectral radiometry. In: Proceedings of the International Grasslands Congress, Multifunctional Grasslands in a Changing World, June 27 - July 5, 2008, Hohhot, Inner Mongolla, China. 1:407. Available on-line: http://www.rangelandcongress.com/ChinaCongress.htm

Interpretive Summary: Grazing animal gains on forage have a lower cost of gain than animals fed in feedlot on mixed rations. However, often nutrient levels on warm-season forages can fall below animal requirements and protein or energy supplementation is desirable. Conventional methods of estimation of forage quality are slow and labor intensive processes. Recent work has suggested that it is possible to assess forage quality in real-time on the pasture, which allows stocker operators the technology to quickly recognize the need for supplementation. However, uncertainty exists in actual nutrient deficiencies because of the difficulty estimating forage intake in grazing animals. Research at the USDA-ARS Grazinglands Research Laboratory was done to develop equations the potential for estimating forage intake on bermudagrass pasture using hyperspectral forage canopy light absorbance measured by hand-held spectroradiometers. Results from the study indicated strong relationships between bermudagrass canopy light reflectance measured from 410 nm to 1010 nm and dry matter intake in test lambs. Consequently, it was concluded that prediction of forage dry matter intake using hyperspectral forage canopy reflectance is feasible and that, combined with allied forage quality estimates, real-time precision supplementation of grazing livestock is possible.

Technical Abstract: Hyperspectral forage canopy absorbance was estimated on eight random plots in each of three 1.2 ha common bermudagrass pastures weekly over a period of 9 weeks from June through early August, 2005 using spectroradiometers measuring light reflectance from 410 nm to 1010 nm. Forage in each plot was harvested using a sickle mower each week and weighed amounts were fed to one of 12 individually penned lambs to evaluate the potential for estimating forage intake on the pasture using hyperspectral forage canopy light absorbance. Dry matter intakes were expressed as a percent of body weight and corrected for fixed effects and time before regression on forage canopy wavelengths from 410 nm to 1010 nm using stepwise regression procedures. Dry matter intakes as a percent of body weight were also regressed on forage canopy light absorbance using partial least squares procedures. The best 10-, 20-, 30-, 40-, 50-, and 60-variable prediction models yield R2 values of 0.27, 0.49, 0.61, 0.73, 0.81, and 0.92, respectively (all P < 0.01). Partial least squares regressions yielded R2 similar to the 60-variable model. Forage dry matter intake as a percentage of body weight can be predicted on pasture using forage canopy absorbance estimates from hyperspectral radiometers. In conjunction with forage quality estimates, it appears possible to precisely estimate nutrient balance of grazing animals and provide precision supplementation to meet production targets.