Study On Vegetation Restoration And Plant Species Selection Of The Landslide Areas Induced By The Wenchuan Earthquake In Beichuan County

In order to shorten the process of vegetation restoration in the Wenchuan Earthquake zone, this paper aimed to provide scientific advice on species and combination selection by 29 typical sample surveys in landslide blanks and undamaged woodland of Beichuan County. Using x2 test, Pearson and Spearman correlation analysis, principal component-cluster analysis, TWINSPAN classification and CCA, DCA, DCCA sorting methods, the paper found out the dominant species in study area and their interspecific association, divided site types and plant types, and made clear the relationship between plant community and environment. Results showed that:(1) There were 250 kinds of species in the study area, which belonged to 197 genera and 78 families. The majority of them were Asteraceae, Poaceae, Leguminosae and Rosaceae, among which 97 were dominant species. (2) Based on the result of principal component-cluster analysis, the research area was divided into seven categories, and the plant communities were divided into nine groups according to TWINSPAN analysis. The results clarified that, the types of survey samples and plant communities were not completely consistent in the study area, and there was a large difference in plant diversity between different site types. (3) Vegetation restoration were largely influenced by soil pH value, altitude, soil texture, slope and soil organic matter content, indicating that the environmental factors were important for vegetation restoration in the research area. (4) By comparing plant and soil comprehensively, it was found that the measures of "dry masonry" and "ecological bag and fence" were the optimum engineering measures to restore vegetation in the study area, however, the "bamboo fence" measure was so poor that it was not suitable for the process of vegetation restoration in Beichuan County on a large scale separately. (5) The soil conditions of artificial restoration plots, such as the contents of soil organic matter and available soil phosphorus, were needed to be improved. Finally, it is concluded that vegetation restoration in the damaged area is still in the early stage. Vegetation hierarchy before landslide, landslide damage or man-made factors are likely to be the main factors influencing the landslide blanks vegetation distribution. The article suggests that closely correlated plant-pairs should be used reasonably in the study area, in order to form a good structure of plant community. According to ecological recovery goals, this paper approves 14 kinds of suitable species for artificial restoration plots in study area.