A practical scheme for correcting multiple scattering effects on optical LAI measurements

Accurate and fast non-destructive measurements of leaf area index (LAI) of plant canopies are essential to environmental applications such as water and carbon cycle modelling. A commonly used technique to acquire LAI in situ is based on measurements of radiation transmittance through the canopy with optical instruments. The LAI-2000, that obtains measurements of effective LAI (Le: LAI retrieved assuming random foliage distribution) based on gap fraction at five view angles, is designed to work under diffuse light conditions. The LAI-2000 makes use of blue light to minimise the effect of light scattering in the canopy on LAI measurements. However, actual field LAI measurements are still routinely done under a range of illumination conditions, including direct sunlight. The LAI values measured under conditions of either partial or full direct light are generally smaller than the ones obtained under diffuse conditions. Although this error source is prevailing in many field LAI measurements, hitherto the problem has not been tackled rigorously. To better understand and improve the LAI-2000 measurements taken under non-ideal conditions, measurements were taken in two deciduous and two coniferous forest sites at different times of cloudless days to study how the scattering of the blue light by plant canopies affects LAI measurements. The sites are located in Larose forest near Ottawa, Canada. It is shown through these measurements and modelling with the canopy radiative transfer model Five-scale [Remote Sens. Rev. 19 (2000) 293–305] that the blue light scattering causes underestimation of effective LAI by up to 20% when measured under direct sunlight. A correction for the scattering effect, as a function of solar zenith angle and the effective LAI measured under the sunlit condition, is found through an empirical fit to the measured data in a limited range as well as model-simulated data in the full possible range possible. It is also found that the LAI-2000 fourth ring (47–58° from zenith) gives a more consistent correction than the other rings and that this ring used alone is also suitable for effective LAI retrieval under diffuse conditions. The correction scheme can reduce the error in effective LAI measurements to within 2%. It is therefore suggested that in field programs with logistic constraints, the LAI-2000 time of operation during sunny days can be extended beyond the diffuse illumination conditions near sunrise and sunset since the influence of the direct sunlight on the LAI measurements can be mostly removed using the correction scheme provided in this study.