Estimation of nadir clumping index (NCI) using airborne laser scanning data over global forest sites

Canopy clumping index (CI) and nadir CI (NCI) describe foliage spatial distribution and are important for canopy radiative transfer and land-surface process simulations. Classical methods, such as the logarithmic gap fraction averaging (LX), the gap size distribution (CC), and the combined gap size and logarithmic averaging (CLX) have been explored to estimate CI in field studies. However, estimating forest NCI from airborne laser scanning (ALS) using classical methods remains challenging, because these methods are not directly applicable to ALS data and the estimation accuracy is greatly hampered by pulse density and footprint size. Here, we adapted the classical methods to estimate NCI from ALS data over 45 global forest sites by accounting for pulse density, footprint size, and other constraints. Results showed that CC yielded the highest mean NCI (0.82 ± 0.20), followed by LX (0.65 ± 0.19) and CLX (0.54 ± 0.18). Validation against field measurements showed that plant area index (PAI) derived by LX and CLX agreed well with field PAI (|bias| ≤ 0.31) obtained from digital hemispherical photography (DHP) and terrestrial laser scanning (TLS), whereas the CC-derived PAI showed systematic underestimation (bias = −1.35). CLX performed best overall, achieving R² = 0.76, RMSE = 1.31, bias = −0.24 for PAI, and R² = 0.22, RMSE = 0.14, bias = −0.07 for NCI. The LX NCI decreased with increasing FVC, CHM, PAI e , and PAI, whereas the CC NCI showed the opposite trend. CLX NCI was overall closer to LX NCI and decreased with increasing structural parameters. The proposed method can be expanded to estimate NCI at a much larger scale. The retrieved NCI values can be used in canopy reflectance and land surface models and directional CI estimation.