Sea level rise due to glacier melting caused by climate warming is a major global challenge, but the mechanism of the effect of salinity on soil carbon (C) mineralization in different land types is not clear. The pathways by which salinity indirectly affects soil carbon mineralization rates need to be investigated. Whether or not the response mode is consistent among different land-use types, as well as the intrinsic links and interactions between soil microbial resource limitation, environmental stress, microbial extracellular enzyme activity, and soil carbon mineralization, remain to be demonstrated. In this paper, three typical land-use types (wetland, forest, and agroforestry) were selected, and different salinity levels (0‰, 3‰, 6‰, and 32‰) were designed to conduct a 125-day laboratory incubation experiment to determine the soil CO2 release rate, soil physicochemical properties, and soil enzyme activities, and to correlate C mineralization with biotic and abiotic factors. A correlation analysis of soil physical and chemical properties, extracellular enzyme activities, and carbon mineralization rates was conducted to investigate their intrinsic linkages, and a multiple linear regression of C mineralization at different sites was performed to explore the variability of mineralization among different site types. Structural equation models were established in the pre- and post-incubation stages to study the pathways of soil C mineralization at different incubation times, and the mechanism of mineralization was further verified by enzyme stoichiometry. The results showed that, at the end of 125 days of incubation, the 32‰ salinity addition reduced the cumulative mineralization of forest and agroforestry types by 28.41% and 34.35%, respectively, compared to the 0‰ salinity addition. Soil C mineralization in the three different land-use types was highly correlated with the active C fractions of readily oxidizable C (ROC), dissolved organic C, and microbial biomass C (MBC) in the soil, with the standardized coefficients of multivariate linear regression reaching 0.67 for MBC in the wetland and −0.843 for ROC in the forest. Under long-term salinity additions, increased salinity would reduce the microbial respiratory quotient value by inhibiting β-glucosidase activity, thus indirectly affecting the rate of CO2 release. With added salinity, the mineralization of non-saline soil was more susceptible to the inhibitory effect of salinity, whereas the mineralization of salinized soil was more controlled by soil C pools.
Read full abstract