---

作 者 ：尹锴,崔胜辉*,赵千钧,花利忠,石龙宇,吝涛

期 刊 ：生态学报 2009年 29卷 11期 页码：6085~6094

关键词：城市森林；植物多样性；冗余分析；双重筛选逐步回归；林下叶层；

Keywords:urban forest, plant diversity, redundancy analysis, duplicato bolting stepwise regression, understory,

摘 要 ：以4类多样性指数作为衡量城市森林林下层植物多样性差异的定量指标，并同时记录对应样方尺度的环境变量特征。以期应用冗余分析(RDA)手段提取城市森林林下层植物多样性的干扰控制因子，从而进一步揭示主要环境变量对城市森林林下层植物多样性变异的贡献率。RDA分析结果表明枯落物盖度、距林缘距离、小路面积和垃圾数4变量为能够显著解释林下叶层植物多样性变化的最小变量组合，解释信息量百分比达61%，10变量共同解释的信息量为72.1%。在RDA分析的基础上，对10环境变量组合与植物多样性特征进行双重筛选逐步回归，发现草本密度与海拔、坡度、距林缘距离、枯落物盖度呈极显著相关；草本层Pielou均匀度指数与海拔、坡度、郁闭度、枯落物盖度、岩石盖度呈极显著相关；草本层Simpson多样性指数与坡度、枯落物盖度、距林缘距离、岩石盖度、伐桩数呈极显著相关；灌木层Pielou均匀度指数则与海拔、郁闭度、小路面积、岩石盖度呈极显著相关。

Abstract:In this paper, four diversity indexes were considered as quantitative indicators of understory for urban forest, while the features of environment variables were recorded at the corresponding quadrat scale. By applying the redundancy analysis (RDA) the control factors that affect understory diversity in urban forest were extracted and the contributions to understory diversity were further disclosed. RDA results showed that litter cover, distance to the forest edge, road area and garbage four\|variables were the minimum variables combination which can significantly explain the understory plant diversity, with 61% information explaination. The 10 variables could explain 72.1% of all information. Based on the RDA, 10 environmental variables and understory features were performed duplicato bolting stepwise regression. The result showed that herb density and altitude, slope, distance to forest edge, litter cover were significantly related; Pielou evenness of herb significantly related with the altitude, slope, canopy cover, litter cover and rock cover; Simpson diversity of herb significantly related with slope, litter cover, distance to forest edge, rock cover and stake; Pielou evenness of shrub significantly related with altitude, canopy cover, road area and rock cover.

---