Study On The Relationship Between Surface Pollen Assemblage And Vegetation In Qaidam Basin

Pollen analysis is one of the important measures to reconstruct paleoenvironment. More and more people are deducing the paleovegetation based on the pollens in the Quaternary sediments. The study on surface pollen assemblages and their relationships with the modern vegetation and climate is the foundation for recovering paleoenvironment and paleovegetation.How to understand the relationship between pollens and vegetation is an important question for ecologist and Palynologist . The promising trends are to link pollen analysis more closely with the topics of ecology and numberical analysis. In this study, the author summarize the characteristics of surface pollen assemblages in the different communities ,the distribution of typical pollens and the relationship between pollens and vegetation in the southern slope of the Eastern Qilian Mountain and Qaidam Basin.There are some arboreal pollens in the surface pollen assemblage of the shrub, steppe communities. Vegetation investigations showed that there are not arboreal plants in the study area, so these arboreal pollens are spreaded from the outside regions.The surface pollen assemblages have obvious distinctions in the different ecological regions. In the forest communities, the percentage of woody pollen is more than the herbaceous pollen's,however it is contrary in the deserta ,shrubs and steppe communities.In the forest communities of this study area,the percentage of woody pollen is 45.35%-78.20%,the herbaceous pollen's is 21.8%-54.64%.But in the shrubs, deserta and steppe communities, the percentage of shrubs and herbaceous pollens is more than 81.51%, the woody pollens'is less than 18.48%. Artemisia and Chenopodiaceae are the major components in the desert and steppe communities pollen combination. Artemisia pollen is the major component of the steppe communities whose constructive species are Gramineae plants.The results of the surface pollen analysis show that pollen assemblage matches the corresponding plants,but it can't accurately reflect the percentage of their matrix vegetation.For example, the taxa of Picea, Betula, Ephedra,Artemisia and Chenopodiacea pollens are over-representative; Compositae is very-representive;Populus,Gramineae, Leguminosea, Ranunculaceae and Roseceae are under-representative. The percentage of Picea pollen is 10% and the Betula pollen's is 27.55%-52.13%,where the Picea and Betula trees are dominant in the communities. The percentage of Ephedra pollen is 60.20% where the Ephedra shrubs are dominant in the communities. The highest percentage of Gramineae pollen is 20.85% where the Gramineae plants are dominant in the communities.Gramineae pollen is common in the other communities,but the content is very low,and most of them is lower than 5%.The percentage of Rosaceae pollen is only 16.32% in the Potentilla fruticosa-shrub.The characteristics of pollen commbination in the different communities shows that the Chenopodiaceae plants are more drought-resistant than Artemisia plants. Chenopodiaceae pollens are dominant in the arid area ,and in the semi-arid environment, it is mainly Artemisia.From west to east, Ephedra, Cyperaceae, Gramineae pollens showed an increased trend,with the increasing precipitation. The pollen types of needle trees, Cyperaceae,Potamogeton and Thalictrum represent wetter environmental condition.The pollen types of Artemisia, Chenopodiaceae,Caragana,Tamarix and Nitraria are present drier environmental condition, and pollen types of Gramineae, Compositae, Ranunculaceae, Leguminosae, Labiatae, Rosaceae represent tran-sitional environmental condition between the two.Cluster Analysis shows close relationship between the surface pollen assemblages and vegetation communities, except for individual samples.The results of the Clustering Analysis divide the brushes and steppe communities samples into 6 groups.The results of the Principal Components Analysis divide them into 8 groups. The second axis shows the relationship between the samples and humidity based on the ecological environment of each group; the first axis shows the relationship between the samples and temprature based on the growth habit of each plant. In this study, the result of Clustering Analysis is roughly the same as the Principal Component Analysis,but the Principal Component Analysis is superior to Clustering Analysis.