凋落叶空间扩散模型在常绿阔叶林的适用性分析

Applicability analysis of spatially explicit model of leaf litter in evergreen broad-leaved forests

利用凋落叶空间扩散模型研究单株植物的凋落叶扩散过程，这对预测凋落叶在地表的分布格局有重要意义,本文依据浙江天童20hm^2动态监测样地植被调查数据和叶凋落量数据，分别对20种目标树种进行凋落叶空间扩散模型的拟合，以及模型适用性分析,模型假设叶凋落量和植株胸径之间服从异速生长关系，并且叶凋落量随距离呈指数降低，通过极大似然法估计模型参数,结果表明：所有树种实际叶凋落量和理论叶凋落量相关性显著；但树种间的模型预测精度相差较大，各树种理论叶凋落量解释实际叶凋落量变异的百分比为16．0％～74．0％，平均为49．3％,模型预测精度与叶凋落量数据的标准差、树种平均胸径、树种平均叶片干质量呈显著正相关．根据各树种的分布格局，使凋落物筐覆盖到不同胸径母树周围不同距离处，确定各树种的最优扩散函数，以及不断改进已有的扩散函数可以提高模型的预测精度。

The spatially explicit model of leaf litter can help to understand its dispersal process, which is very important to predict the distribution pattern of leaves on the surface of the earth. In this paper, the spatially explicit model of leaf litter was developed for 20 tree species using litter trap data from the mapped forest plot in an evergreen broad-leaved forest in Tiantong, Zhejiang Province, eastern China. Applicability of the model was analyzed. The model assumed an allometric equation between diameter at breast height （DBH） and leaf litter amount, and the leaf litter declined exponentially with the distance. Model parameters were estimated by the maximum likelihood method. Results showed that the predicted and measured leaf litter amounts were significantly correlated, but the prediction accuracies varied widely for the different tree species, averaging at 49.3 % and ranging from 16.0% and 74.0%. Model qualities of tree species significantly correlated with the standard deviations of the leaf litter amount per trap, DBH of the tree species and the average leaf dry mass of tree species. There were several ways to improve the forecast precision of the model, such as installing the litterfall traps according to the distribution of the tree to cover the different classes of the DBH and distance apart from the parent trees, determining the optimal dispersal function of each tree species, and optimizing the existing dispersal function.