Direct and indirect measurements of LAI in millet and fallow vegetation inHAPEX-Sahel

Direct measurements of leaf area index (L) were made in millet and fallow v egetation in the Sahelian zone in Niger, West Africa, Indirect estimates of L were made using two methods. In the first method, transmittance (tau) of the canopy was calculated from measurements of radiation in the photosynth etic waveband (Q) incident upon, and transmitted by the canopy. In the seco nd method, tau was inferred from image analysis of hemispherical photograph s, The canopies were sparse (L < 1) and highly clumped, representing an ext reme test of the indirect methods, which assume that foliage is randomly di stributed. In the first method, the analysis of Lang and Xiang (1986) was used to mini mise the effect of clumping. This involved using leaf dimensions to calcula te an appropriate length scale for linear averaging of t, before a logarith mic transformation was applied. Using this averaging length, indirect estim ates from the transmitted Q method were extremely close to direct estimates in millet. In the fallow vegetation, static sensors were used, so the proc edure of Lang and Xiang (1986) could not be followed exactly, but was appro ximated by setting the temporal averaging period to allow for an appropriat e degree of sun (and shadow) movement. Agreement with direct estimates was reasonable and within the confidence interval of direct estimates. In the second method, L estimates from hemispherical photography were withi n 0.1 of direct estimates at both sites, but there was an apparent underest imation at higher L values. This may be a bias induced by camera positionin g, overexposure of images, or averaging length scale, as the analysis of La ng and Xiang (1986) could not be applied quantitatively to this method. (C) 1999 Elsevier Science B.V. All rights reserved.