Prediction of Alfalfa Herbage Mass Using Sward Height, Ground Cover, and Disk Technique

Abstract

AbstractRapid, nondestructive methods of estimating herbage mass (HM) are vital to experiments involving frequent collection of data, especially where physical sampling may alter growth. Herbage mass estimation techniques are based on double‐sampling, in which clipped HM and corresponding nondestructive sward measurements are used in regression analysis to obtain prediction equations. Our objective was to evaluate the precision and accuracy of sward height (SH), ground cover (GC), and disk height (DH) as predictors of HM in alfalfa (Medicago sativa L.) at weekly intervals after cutting. Sward height was measured with a sliding crossbar on an upright meter stick. Ground cover was estimated using a 0.18‐ by 0.22‐m grid subdivided into 20 cells. Disk height was measured with 0.96‐ and 1.76‐kg, 0.2‐m2 aluminum disks, allowed to settle on the sward canopy. All nondestructive measures were accompanied by clipping ofthe areas measured. Clipped HM ranged from 12 to 794 g m−2. Herbage mass prediction equations were developed following stepwise regression of sample HM on SH, DH, GC, days of growth, and derived variables. The most precise HM prediction models were based on SH and DH, accounting for 79 to 94% of variation in HM of calibration samples. Applying prediction equations to a validation data set generated predicted HM values. Sward height and DH models accounted for 59 to 86%, and 80 to 97%, respectively, of the variation in HM of validation samples. Herbage mass of validation samples was predicted with lower standard errors of validation by DH models (19.5–35.9 g m−2) than by SH models (60.8–65.6 g m−2) on the same stands in the same year. The disk technique provides a rapid, accurate means of nondestructively monitoring growth in alfalfa plots.