甘肃兴隆山国家自然保护区表土花粉组合及数量分析

POLLEN ASSEMBLAGES AND NUMERICAL ANALYSIS OF SURFACE SAMPLES FROM XINGLONG MOUNTAIN NATIONAL RESERVE,GANSU

通过对兴隆山国家自然保护区不同植物群落40个样点的表土花粉组合特征、数量分析[分层聚类分析(HCA)和降趋势对应分析(DCA)],研究了花粉组合与植物群落的关系。结果显示:1)大部分森林和部分灌丛的花粉组合与植物群落之间对应关系很好,如青杆林(云杉属组合)、辽东栎林(栎属-桦属-蒿属组合)、桦林(桦属-蒿属-沙棘属组合)和沙棘灌丛(沙棘属组合);2)蔷薇灌丛(蒿属-蔷薇科组合)和草原(蒿属-菊科组合)的花粉组合基本上能反映群落特征;3)山杨林、糙皮桦灌丛和高山灌丛草甸的花粉组合与植物群落的对应性较差;4)云杉属花粉的长距离传播能力有限;5)栎属花粉仅在辽东栎林中含量较高;6)桦属花粉含量较高时(>30%)能够反映桦林的存在,含量较低时(<10%)可能当地无桦树生长;7)沙棘属和蒿属花粉具超代表性;8)豆科、禾本科和蔷薇科花粉具低代表性;9)HCA和DCA分析对于区分主要花粉类型和揭示环境梯度是有效的,但反映植被细节特征的能力有限。

In this paper 40 surface pollen samples from Xing- long Mountain Nature Reserve in Gansu Province were analysed for relationship between pollen and vegetation. The data were classified using Hierarchical Cluster Analysis （HCA） and were ordinated by using Detrended Correspondence Analysis （DCA）. The relationship between surface pollen assemblage and plant community shows that. 1）the pollen assemblages from most of the forests and part of the shrubs seem to correspond well with the distribution of the pollen producers, such as Picea wilsonii forest（Picea assemblage）, Quercus liaotungensis forest （Quercus-Betula-Artemisia assemblage）, Betula platyphylla forest （Betula-Arte- misia-Hippophae assemblage）, and Hippophae rhamnoides ssp. sinensis shrub（Hippophae assemblage）; 2）the pollen assemblages from Rosaceae shrub （Arte- misia-Rosaceae assemblage） and steppe （ Artemisia- Compositae assemblage） may represent characteristics of plant community moderatelye 3） the pollen assem- blages from Populus davidiana forest, Betula utilis shrub, and alpine/subalpine shrub and meadow are not well associated with their plant communities. This study also sheds light on several important issues in pollen dispersal and representation. The higher Picea percentages correspond closely to the distribution of spruce trees in the spruce forest and mixed forest, and the occurrence of Quercus pollen is consistent with the distribution of oak trees. The higher Betula pollen value （〉30%） probably indicates the existence of birch forest, and birch trees may not be locally present at sites where Betula pollen occurs at values 〈 10%. Hippophae and Artemisia are over-represented, whereas Leguminosae, Gramineae and Rosaceae are under-represented pollen taxa. In addition, HCA and DCA are effective to indicate the distinction between different pollen assemblages and environmental gradient, but limited in detailed explanation of vegetation from pollen assemblages.