Canopy transmittance models for estimating forest leaf area index

Abstract
Leaf area index was estimated in old-growth and young post-fire coniferous forests in northwestern Colorado. A line quantum sensor was used to measure canopy transmittance at different solar zenith angles. Leaf area indices were estimated from canopy transmittance data according to three different models and were subsequently compared with leaf area indices derived from existing allometric equations. Of the three canopy transmittance methods evaluated, a Beer–Lambert model adjusted for diffuse light and solar zenith angle was in closest agreement with allometric leaf area index estimates (11.5% average difference), followed closely by the Beer–Lambert model (14.4% average difference). Leaf area index predicted by a one-dimensional inversion model did not agree well with allometric estimates (30.6% average difference). Differences in methods of data processing were found to have significant effects on final results. Subtraction of diffuse photosynthetically active radiation increased the leaf area indices. Calculation of leaf area index at each sampled point and determination of a final mean leaf area index approximated the allometrically derived values more closely than did derivation of leaf area index only once from an averaged gap-fraction value. Leaf area index estimates varied with sun angle.