Abstract: To examine the response of soil microbial community to pruning of encroaching shrub species and explore the relationships between soil microbial community composition/structure and soil physicochemical properties and plant community, we set up treatment of pruning Caragana microphylla plots and control plots without pruning in the C. microphyllaencroached grassland of Inner Mongolia. Three months later, soil physicochemical properties were measured, composition and structure of plant communities were investigated, and soil microbial community composition and structure at the depths of 0-5 cm and 15-20 cm were analyzed by phospholipid fatty acids (PLFA) method. Compared with control, pruning treatment decreased the total aboveground biomass, with a significant decrease of Artemisia intramongolica biomass, while significantly increased species richness and Shannon index. Soil pH was significantly lower at 0-5 cm layer, and soil total carbon content, total nitrogen content, and C/N were significantly higher at 15-20 cm soil layer. The PLFAs contents of total microbes, Gram-positive bacteria (G⁺), Gram-negative bacteria (G^-), bacteria (B), fungi (F) and actinomyces (Act) were significantly higher, and G⁺/G^- was lower at 0-5 cm layer. Results of redundancy analysis (RDA) showed that soil pH was positively correlated with the PLFAs content of G⁺, and negatively correlated with that of G^-, F and Act. Soil total nitrogen content was positively correlated with PLFA contents of G⁺, G^-, B, F and Act, while Shannon index was negatively correlated with that of G⁺ and positively correlated with that of G^-, F and Act. Total aboveground biomass was negatively correlated with the PLFAs contents of G^-, F and Act. These results indicated that soil microbial community composition and structure would respond rapidly to the changes of plant community and soil physicochemical properties caused by pruning, which could be used as an indicator for soil health and provide guidance for the restoration of shrubencroached grassland.

Key words: site index model, DEM, potential distribution area, spruce forest.