Assessing Factors That Affect the Estimation of a Canopy's Gap Fraction and Extinction Coefficient Using Discrete Airborne LiDAR Data.

Gap fraction ( p(?)) and extinction coefficient (k) are critical variables to predict canopy structure parameters, such as the leaf area index (LAI). The discrete airborne LiDAR scanner (ALS) provides an entirely new method to estimate p(?) and k of a canopy. However, it is widely recognized that estimating p(?), derived from the laser penetration index (LPI), and estimating k, are strongly influenced by the characteristics of the LiDAR sensor, observation geometry, and canopy properties. This study conducted a global sensitivity analysis to comprehensively assess the influences of factors deriving from the above three sources on p(?) and k, using the point cloud data simulated by the computer simulation model. Also, the applicability of point cloud for estimating p(?) and for estimating k using the normal vector reconstruction (NVR) method under different conditions was investigated. Results showed that the influence of explored factors on different LPIs of the canopy is dominated by some single factor, but the specific factors are LPI-and LAI-dependent. In contrast, the influences from explored factors on k showed an identical pattern across different support region sizes (SSRs) and leaf angular distribution types. The retrieved k was influenced by pulse density-related factors. The estimated p(?) accuracy varied with the form of LPI and the range of LAI. The capability of successfully estimating k relied on the leaf angular distribution. The estimated k was influenced by the inherent target selection effect of LiDAR return, which requires further investigation.