Improved algorithm for estimating canopy indices from gap fraction data in forest canopies

An algorithm has been proposed to estimate the leaf area index (LAI) and canopy closure of forest canopies from the distribution of gap fraction measured by an LAI-2000 plant canopy analyser by inverting a forest gap fraction theoretical formula. The algorithm is based on the idea that the canopy closure is estimated from the reading of canopy analyser in the near-zenith ring. However, the measured gap fraction, especially at the near-zenith view directions, is subject to random variations due to insufficient spatial sampling. For that reason a regularisation of the inverse problem is proposed, based on the idea that the random fluctuations of the gap fraction can be described by means of their expansion by eigenvectors and eigenvalues of the covariance matrix. An algorithm to simulate the eigenvectors and eigenvalues and to evaluate the random coefficients of the expansion has been presented. Another novel feature of the algorithm is that it provides an opportunity to correct for the shadowing effect of tree trunks in the gap fraction data. To apply the algorithm, estimates for certain inventory data on the stand under study are needed, such as the mean tree height, crown depth, breast-height trunk diameter and stem number. The application of the algorithm appears to be efficient when the respective stand data can be obtained from a forestry database or at least their expert estimates are available. A few examples demonstrating fairly good performance of the algorithm, especially in relatively open boreal and sub-boreal forests, have been given.