Microbial biomass, enzyme and mineralization activity in relation to soil organic C, N and P turnover influenced by acid metal stress

International audience

English. This study focused on the potential of using soil microbial biomass, enzyme and mineralization activities involved in organic C, N and P turnover, to evaluate the quality of a subtropical agricultural soil affected by long-term acid metal stress. Fractions of C, N and P involved in soil organic matter, microbial biomass and mineralization processes were estimated. Total enzyme activity (FDA) and eight hydrolase activities (xylanase, amylase, b-glucosidase, invertase, N-acteyl-glucosaminidase, urease, alkaline and acid phosphatases) in different decomposition stages of organic C, N and P were selected to characterize the soil functional diversity. These biological datasets were compared with soil metal variables (total contents and free and ligand-complexed ions of Cu, Pb, Zn, Cd, Al and Mn), using principal component analyses, co-inertia and discriminant analyses. The multiple statistics indicate that the metal variables were significantly related with not only general biological factors, but also respective datasets of biomass, enzyme activities and mineralization rates (all P< 0.001). In general, metal variables were inversely related to parameters and indices of microbial biomass C, N and P, FDA and C-related polysaccharidase and heterosidase activities, and P mineralization. As comparison, metal variables exhibited positive relationships with parameters and indices of N-related N-acteyl-glucosaminidase, urease, ammonification, total N mineralization and metabolic quotient, compared with inhibited nitrification. Specifically, free and complexed metal cations showed higher bioavailability than total contents in most cases. Cu, Pb, Al and Mn had different ecotoxicological impacts than Cd and Zn did. Stepwise regression models demonstrated that metal variables are key stress factors, but most of them excluded soil pH. Furthermore, spatial distribution in land uses and of sampling sites clearly separated the soil samples in these models (P< 0.001). We conclude that such a statistical analysis of microbiological and biochemical indices can provide a reliable and comprehensive indication of changes in soil quality and organic nutrient cycling, after exposure to long-term acid metal stress.