A meta-analysis of the canopy light extinction coefficient in terrestrial ecosystems

The canopy light extinction coefficient （K） is a key factor in affecting ecosystem carbon, water, and energy processes. However, K is assumed as a constant in most biogeochemical models owing to lack of in-site measurements at diverse terrestrial ecosystems. In this study, by compiling data of K measured at 88 terrestrial ecosystems, we investigated the spatiotemporal variations of this index across main ecosystem types, including grassland, cropland, shrubland, broadleaf forest, and needleleaf forest. Our results indicated that the average K of all biome types during whole growing season was 0.56. However, this value in the peak growing season was 0.49, indicating a certain degree of seasonal variation. In addition, large variations in K exist within and among the plant functional types. Cropland had the highest value of K （0.62）, followed by broadleaf forest （0.59）, shrubland （0.56）, grassland （0.50）, and needleleaf forest （0.45）. No significant spatial correlation was found between K and the major environmental factors, i.e., mean annual precipitation, mean annual temperature, and leaf area index （LAI）. Intra-annually, significant negative correlations between K and seasonal changes in LAI were found in the natural {K=2/π[cosαcosθsina^-1（tanθtanα）＋（1＋cos^2a-cos^2θ^1/2）],a＋θ〉π/2 K=cosαcosθ,α＋θ≤π/2 k K is usually calculated with the Beer Lambert Law （Monsi and Sakei, 1953）： K = - In （Ii/Io） cosθ/（LAIΩ）,（2）ecosystems. In cropland, however, the temporal relation- ship was site-specific. The ecosystem type specific values of K and its temporal relationship with LAI observed in this study may contribute to improved modeling of global biogeochemical cycles.