Modeling leaf area index from litter collection and tree data in a deciduous broadleaf forest

Leaf area index (LAI) is an important index in ecological and meteorological studies. The litter trap method is commonly used to measure LAI in deciduous forests. To reduce the time consumed in sorting leaf litterfall by species in the litter trap method, we developed four models to predict LAI using litter traps and tree census data. The local dominance model, which estimates the leaf litterfall amount of each species by their local dominance, predicted mean and spatial variability of LAI most accurately compared to the 2 models that did not take into account spatial heterogeneity of species distribution within a forest or the model that estimated litterfall amount from leaf dispersal function. Therefore, this model can be employed instead of sorting leaf litter by species. Furthermore, we found that leaf mass per area (LMA) of at least 10 dominant species are essential for accurate estimation of LAI. Present results suggest that spatial variability of LAI is mainly due to spatial variance of leaf litterfall followed by spatial heterogeneity of species distribution within a forest, and difference in LMA among species.

Research highlights
► Four models to estimate leaf area index (LAI) using litter traps and tree data.
► Local dominancy model accurately predicted the mean and spatial variability of LAI.
► The model can be employed instead of sorting leaf litter by species.
► Leaf mass per area of 10 dominant species were essential for estimating LAI.