中国东北样带 (NECT)东部森林区的植被与表土花粉的定量关系

Study of Quantitative Relationships Between Vegetation and Pollen in Surface Samples in the Eastern Forest Area of Northeast China Transect

基于中国东北样带东部森林区 8个植被样方和表土花粉资料 ,定量分析了每种植物与表土花粉类型的关系。结果表明 :花粉类型与植物关系密切 ,在α =0 .0 5的显著性水平上 ,相关系数 (correlationcoefficient)C值大都在0 .5以上 ;表土花粉组合与植物群落间有较大的相似性 ,相似系数 (CC)大于 5 0 %。采用定量描述花粉与植物关系的参数 :联合指数 (associationindex)A、超代表性指数 (over_representationindex)O、低代表性指数 (under_representationindex)U、相关系数C和代表性系数 (representationcoefficient)R等值 ,应用TWINSPAN分类、PCA排序和回归分析 ,把东北样带的表土花粉类型划分为 4类 :具代表性类群组Group 1,它们能正确反映植被 ;超代表性类群组Group 2 ,它们常具有高花粉值 ,与植被不成比例 ;低代表性类群组Group 3,它们的花粉很难分离提取 ;以及低代表性类群组Group 4,其花粉在土壤地层中较常见。在此基础上 ,分析了花粉的R值与A值的关系 ,指出A值和R值一样也是校正地层花粉数据的良好参数 ,并建立了二者间的回归方程 :A =- 0 .0 42 1R2 +0 .2 42 5R +0 .392 6 (r2 =0 .6 0 2 1)。该研究为用古花粉资料准确复原古植被奠定了基础。

The eastern forest stands of Northeast China Transect (NECT) were chosen to study the quantitative relationships between vegetation and pollen in surface samples. The indices of A (association index), O (over_representation index), U (under_representation index), C (correlation coefficient) and R (representation coefficient) for each pollen type were calculated. The results indicated that the relationships between vegetation and pollen type in surface samples were significant and the correlation coefficients of 70% pollen types were more than 0.5 (α=0.05); the similarity between pollen assemblage and plant community was good and coefficient of similarity was more than 50%. 69 pollen types found in the surface samples could be divided into four groups with TWINSPAN classification and PCA ordination according to A,O,U and C. The four groups reflected the pollination characteristic of plants and the state of pollen conserved in soil. Group 1 was the associative group which could accurately reflect the local vegetation; Group 2 was over_representative group which had high pollen percentage outproportional to vegetation; Group 3 was the under_representative group in which the pollen were hardly obtainable from the soil, and Group 4 was also an under_representative group, in which pollen were easily obtainable in soil where plants directly grow from. The study also showed that A was a parameter to rectify pollen data and it was easier to obtain than parameter R . The parameters A and R have close relation and their regression equation was: A=-0.042?1R 2+0.242?5R+0.392?6(r 2=0.602?1). These groups and indices provide a solid foundation for using pollen data in accurately reinstating vegetation.