IntroductionEcological restoration has received worldwide attention due to its vital roles in carbon (C) sequestration, biodiversity conservation, desertification prevention and soil fertility improvement of degraded lands. A target to restore at least 15% of the degraded lands globally has been proposed by United Nations Convention on Biological Diversity. However, there is evidence showing that many restoration projects are failed or get limited success. In this sense, it is crucial to monitor the dynamics of ecosystem structures, functions and processes during ecological restoration, including soil fertility indices such as soil organic C (soil C hereafter), soil total nitrogen (TN), soil total phosphorus (TP), and soil microbial communities. Soil C is a perfect proxy for judging whether degraded lands have improved, as increase in soil C benefits the improvement of water holding capacity, nutrient retention capacity, soil structure and soil biological quality. Soil TN and TP are the major limiting nutrients of ecosystem productivity and play key roles in ecosystem restoration or succession. Soil microbial communities play key roles in soil biogeochemical cycles and exerts direct and indirect effects on the development of plant communities. Additionally, soil microbial communities can be used as an integrated measure of soil health or fertility due to their important roles in soil nutrient cycling. Many factors change following agricultural abandonment or during ecological restoration. Different restoration strategies, e.g., restoration with different plant species with or without management, may be considered when restoring a given degraded area. As different tree species or their combinations may have differential effects on substrate quantity and quality, rhizo/ mycorrhizospheric chemistry, and other soil properties, they may induce distinct profiles of soil microbial communities. SOM with high cellulose and lignin content and high C:N ratio are likely decomposed via fungal-dominated “slow” pathways, while moist, N-rich tissues are mainly decomposed via bacterial-dominated “fast” pathway. Karst ecosystems are widely distributed over the Earth’s land surface, including southwest China. In the past, a large portion of the karst lands in southwest China has been degraded due to intensive human disturbances especially agricultural activities. In the present study, soil nutrients and microbial communities (characterized by phospholipid fatty acid (PLFA) method) under three commonly adopted restoration strategies in the karst area of southwest China were compared, i.e., i) restoration with an economic tree species Toona sinensis (TS), ii) restoration with Guimu-1 hybrid elephant grass (GG), and iii) restoration with a combination of Zenia insignis and Guimu-1 hybrid elephant grass (ZG). Cropland was used as reference. We hypothesized that the patterns of soil C and TN in the present study (observed in 2017) would be similar to those observed in 2014 (Hypothesis I). On the other hand, fungi are sensitive to disturbance, e.g., tillage, but bacteria can rapidly adapt to frequently changing soil environment. Subsequently, the abundance of fungal community may increase more pronouncedly relative to that of bacterial community when the disturbance ceases. We therefore hypothesized that the fungal to bacteria PLFA (F:B) ratio would increase following agricultural abandonment (Hypothesis II).Materials and methodsThis study was conducted at Guzhou catchment in Guangxi Zhuang Autonomous Region, southwest China. This region is located in the subtropical humid forest life zone with a monsoon climate. The experiment site used to be a cropland under maize-soybean rotation before 2002 and was distributed over the bottom area of a slope of about 20. The major objective of the experiment was to investigate the temporal dynamics of soil properties and plant community following agricultural abandonment. The experiment adopted a completely randomized block design with five treatments (four restoration strategies and a cropland as reference) and three replications. The three restoration strategies included in the current study are described below:Restoration with Toona sinensis (TS), an economic tree species. The tender shoot in spring is a popular vegetable in China.Restoration with Guimu-1 hybrid elephant grass (GG), ahybrid of elephant grass and Pennisetum alopecuroides.This grass is perennial and has high biomass production, harvested for 3 to 4 times per year and used as fodder for beef.Restoration with a combination of Zenia insignis and Guimu1 hybrid elephant grass (ZG). Zenia insignis is a native tree species with the leaves used as additive to the fodder for beef. The Guimu-1 hybrid elephant grass is usually harvested for 3e4 times per year and used as fodder for beef.For most of the sampling sites, obvious organic layer was absent, so samples from organic layer were not collected. The ten soil samples in a plot were mixed as a composite sample. A subsample of fresh soil samples was freeze-dried and used for PLFA analysis. A subsample of fresh soil was used for the analyses of microbial biomass C (MBC) and N (MBN) using chloroform-fumigation extraction method. A subsample of soil was air-dried and passed through 0.15 mm mesh sieve for measurements of soil C, TN, TP, pH and exchangeable calcium (Ca) and magnesium (Mg). Subsamples were further dried at 105 C for 24 h to determine gravimetric soil water content (SWC). Phospholipid fatty acids (PLFAs) were extracted from 8 g of freeze-dried soil and were analyzed. Data were tested for normality of variance, and were log transformed when necessary. One-way ANOVA was used to examine the significant difference in the biotic and abiotic variables. Pearson’s correlation analysis was used to identify the relationships among the biotic and abiotic variables. Stepwise multiple linear regression approach was used to identify and evaluate the contributions of the strongest explanatory variables to PLFA biomarkers.