基于浙江天童20 hm^2常绿阔叶林动态监测样地的群丛划分

Classification of plant associations based on a 20 hm^2 dynamics plot of evergreen broadleaved forest in Mt. Tiantong, Zhejiang, China

为获取能够代表浙江天童山的森林植被典型群丛类型,同时也为植被分类中如何发现过渡类型和确定典型类型提供参考,该研究利用天童20 hm^2森林大样地资料,运用双向指示种分析(TWINSPAN)与除趋势对应分析(DCA),剔除过渡群落,进行群丛划分。结果表明,去除过渡地段后更利于研究区域典型群丛类型的确定。大样地的植被类型可划分为宜昌荚蒾-厚皮香/港柯+云山青冈群丛(Viburnum erosum-Ternstroemia gymnanthera/Lithocarpus harlandii+Cyclobalanopsis sessilifolia Ass.);虎皮楠-柯/木荷+米槠群丛(Daphniphyllum oldhami-Lithocarpus glaber/Schima superba+Castanopsis carlesii Ass.);红毒茴-紫楠/南酸枣+薄叶润楠群丛(Illicium lanceolatum-Phoebe sheareri/Choerospondias axillaries+Machilus leptophylla Ass.)。DCA排序同时能反映各群丛类型分布与环境的相关关系,结果显示,海拔和凹凸度对群丛分布有较大影响,坡度和坡向对群丛分布影响较小。

Aims Our objectives were to obtain typical plant association types in a 20 hm2 forest dynamic monitoring plot in Mr. Tiantong, and provide methodological references for determining transitional types and identifying typical types in vegetation classification. Methods We used two-way indicator species analysis （TWINSPAN） and detrended correspondence analysis （DCA） to remove transitional community and identify typical association types. DCA method was also used to analyze the relationship between the distributions of different plant associations and environmental variables. Important findings The results showed that the classification of typical plant associations was greatly improved by the removal of transitional communities. The forest was divided into three associations： 1） Viburnum erosum- Ternstroemia gymnanthera/Lithocarpus harlandii ＋ Cyclobalanopsis sessilifolia Association, 2） Daphniphyllum oldhami-Lithocarpus glaber/Schima superba ＋ Castanopsis carlesii Association, and 3） Illicium lanceolatum- Phoebe sheareri/Choerospondias axillaries ＋ Machilus leptophylla Association. The results of DCA revealed the importance of environmental variables （elevation, convexity, slope, and aspect） on spatial distribution of plant associations. Elevation and convexity had significant effects on the distribution of associations, while slope and aspect had weak effects. The results confirmed the high accuracy of the association classification we used.