Comparison of three methods for estimating leaf area index of marandu palisadegrass swards under continuous stocking

The objective of this research was to compare three methods for estimating LAI in marandu palisadegrass (Brachiaria brizantha Hochst. ex A. Rich cv. Marandu) swards subjected to four grazing heights (10, 20, 30 and 40 cm) under continuous stocking by beef cattle. Evaluated methods were: (1) Direct, destructive (LAI(DM)); (2) Indirect, destructive, based on the use of tiller population density and mean leaf area per tiller (LAI(TPD)); and (3) Indirect, non-destructive, based on the use of a LAI-2000 canopy analyser equipment (LAI(LICOR)). LAI(DM) was estimated on herbage samples harvested using four 0.30 x 0.37 m metallic frames per paddock. Measurements were carried out in a monthly basis and results grouped into five seasons of the year: summer, autumn, winter, early and late spring. The experiment was carried out according to a complete randomised block design, with four replications and a split-plot arrangement, where sward heights were allocated to plots and LAI estimation methods to sub-plots. The results showed that the canopy analyser overestimated LAI values in relation to the other two methods. This overestimation is mainly due to the excessive accumulation of dead material at the base of swards, particularly on those kept taller (30 and 40 cm). The LAI(DM) was the method that better detected variations in sward height and seasons of the year, and the LAI(TPD) showed intermediate values, with a pattern of variation in absolute values similar to that obtained for the LAI(DM). The canopy analyser should not be used when LAI estimates are being generated for simulation and mathematical modelling purposes and to estimate the efficiency of incident radiation use.