Abstract: High-throughput sequencing technology (16S rRNA) was used to measure OTU of soil bacterial community in plots with different fire intensities (no fire (control), light intensity, moderate intensity, and high intensity) in a secondary Pinus tabuliformis forest in Pingquan City, Hebei Province. OTUs were annotated at different taxonomic levels, which were used to analyze the quantitative changes of dominant soil bacterial groups under different fire intensities. Soil physicochemical properties (soil organic carbon, NH₄⁺-N, NO₃^--N, available phosphorus, water content, pH) and soil enzyme activity (urease, protease, catalase, and dehydrogenase) were used to analyze the relationship between soil microenvironment changes and soil bacterial communities after fire disturbance. The results showed that the composition of bacterial community varied across different fire intensities. At the phylum level, Proteobacteria, Acidobacteria, Actinobacteria, Verrucomicrobia, Chloroflexi dominated the communities. At the genus level, Candidatus_Udaeobacter (Verrucomicrobia, Chthoniobacterales), Bradyrhizobium (Proteobacteria, Rhizobiales), RB41(Acidobacteria, Pyrinomonadales), uncultured_Acidobacteria_bacterium (Acidobacteria, Vicinamibacterales), Mycobacterium (Actinobacteria, Corynebacteriales) dominated the communities. Fire disturbance had a significant effect on the distribution of bacterial communities, with significant differences among different fire intensities (P<0.05). Among the alpha diversity indices, the value of Chao1 was the highest in the moderate fire intensity plots, and Observed_species was the highest in the light fire intensity plots. PD_whole_tree and Shannon index were generally higher in the light, moderate, and high fire intensity plots than in the control plots. Moderateintensity fire maintained a high diversity of soil bacterial community, while high-intensity fire increased community heterogeneity and led to habitat fragmentation, which a legacy effect after many years. Urease and dehydrogenase activities had the greatest degree of interpretation for bacterial community composition in different fire intensity plots. Urease activity was highly positively correlated with Candidatus_Udaeobacter, uncultured_Acidobacteria_Bacteria, RB41. Dehydrogenase activity was highly positively correlated with Candidatus_Solibacter, Bryobacterium, Bradyrhizobia and uncultured_Acidobacteria_Bacteria.

Key words: fire disturbance, high-throughput sequencing, soil bacteria, soil physicochemical property, soil enzyme activity.