The study was conducted at Chishan Forestry Center of Dongshan County (23°38°N, 117°24°E) in southeastern Fujian Province, China (Fig. 1). The climate of this region is of the subtropical maritime monsoon type with a mean annual temperature ranging from 13.1 ℃ in January to 27.3 ℃ in July. Mean annual precipitation and evapotranspiration are 1104 mm and 2028 mm, respectively. The Rainy season is from May to September when 61% of the rainfall occurs, while the dry season is usually from November to the next February (Fig. 2). The soil is a typical udipsamment with low nutrient content and a thickness of 80-100 cm.

Five forest stands, including a secondary forest and plantations of Casuarina, Pine, Acacia, and Eucalyptus, were selected as experimental groups in December 2014. These stands have similar elevations and soil types, each with an average slope of less than 10°. The detailed site characteristics and topsoil properties of the five forest stands are shown in Table 1.

The secondary forest is composed of native vegetation resulting from secondary succession with distinct tree, shrub, and herbaceous layers, which regenerates naturally and has been afforded long term protection by local people. The age of the secondary forest is more than 50 years, and the dominant species in the canopy are L. glutinosa and Celtis sinensis Willd. The shrub layer is predominantly Glochidion eriocarpum, Litsea rotundifolia, Bridelia tomentosa, and Anodendron affine (Hook. et Arn.) Druce, and the main herb species are Ophiopogon bodinieri and Arthraxon hispidus var. hispidus. Both Casuarinas and Acacia plantations were established in 1992 with main tree species of C. equisetifolia and Acacia crassicarpa, respectively. The pine plantation was planted with Pinus elliottii Engelm. and slash pine in 1993. The Eucalyptus plantation was established in 2004 with Eucalyptus urophylla × E.grandis crosses; the previous species on this site was P. elliottii, which was planted in 1976. The canopy densities of the different plantations are greater than 0.7 with few shrubs and grasses in the understory (Gao et al., 2018).

In each of the five forest stands, four 20 m × 20 m plots were established with a 10 m buffer between each plot. In July 2015, ten soil cores (1.0 m length × 5 cm diameter) were collected along the diagonal of each plot from the soil surface to a 100-cm depth. Fine roots (< 2 mm diameter) were washed out and the remaining fine-root biomass was estimated following Yang et al. (2004).

Five litter traps (1 m × 1 m) with 1 mm nylon mesh were randomly arranged 0.3 m above the forest floor in each plot for litter collection. From March 2015 to February 2016, litter was collected monthly for determining annual litterfall.

Ten soil pits along the diagonal of each plot were used to collect soil at both 0-10 cm (topsoil) and 10-20 cm (subsoil) depths in April (spring), July (summer), September (autumn), and November (winter) 2015. A 500 g sample was collected from each layer of every plot, and 300 g portions of each were air-dried, ground, and sieved through a 2-mm sieve for microbial biomass analysis. The remaining soil was ground and sieved through a 0.149-mm sieve for total soil carbon (TC) and total N (TN) analysis. Soil bulk density in each layer was calculated using intact soil cores (100 cm³). Soil temperature was obtained using an instantaneous digital thermometer (AM-11T, Avalon company, USA), and soil moisture was determined using an oven-drying method.

Soil microbial biomass C (MBC) and N (MBN) were analyzed following the chloroform fumigation-extraction method (Brookes et al., 1985). Briefly, soils were fumigated for 24 h with CHCl₃, after which the fumigant was removed, the soils were extracted with 0.5 M K₂SO₄, and their C and N concentrations were estimated based on the differences between the chloroform fumigated and unfumigated soil divided by correction factors of 0.45 (Joergensen and Müller, 1996a) and 0.54 (Joergensen and Müller, 1996b), respectively. Specifically, the organic C concentration in the extracted solution was measured using a TOC-VCPH/CPN Analyzer (Shimadzu, Japan) and the organic N content was measured using a continuous flow analyzer (SAN+ +, Skalar Analytical B.V.). Soil TC and TN were determined using an elemental analyzer (Vario EL Ⅲ, Elementar Analysensysteme GmbH, Hanau, Germany). Soil pH was measured with a pH meter.

Data from each of the four plots were averaged for further analysis. To test the significances of differences in soil C and N contents, soil pH, soil bulk density, as well as the soil MBC and MBN contents between different forest types, we used one-way analysis of variance (ANOVA) with Duncan’s multiple comparison method at a significance level of 0.05. We also used Pearson correlation analysis to determine how annual litterfall, litter C and N content, fine root biomass, root C and N content, soil temperature, and soil moisture correlated with the soil MBC and MBN contents. All statistical analyses were performed using the PASW Statistics 18.0 for Windows.

The soil pH, soil C and N content, and soil C: N ratio under different forests all decreased with the increase of soil depth (Table 2), whereas the soil bulk density increased with soil depth. In the topsoil, the soil pH values of plantations were significantly lower than that of the secondary forest, and no significant difference was found in soil pH values among Pine, Acacia, and Casuarinas, which were all significantly lower than that of Eucalyptus. The soil bulk density of Casuarinas was significantly higher than those of the other forests, followed by Eucalyptus and Acacia. However, the soil bulk density of Pine and secondary forest were much lower than those of the other forests. On average, the soil C and N contents of secondary forest were significantly higher than those of the plantations, the soil N content of Eucalyptus was significantly higher than that of Casuarinas, and the soil C: N ratios of Eucalyptus and Casuarinas were significantly higher than those of the other forests.

In the subsoil, the soil pH and soil C and N concentrations in the secondary forest were all higher than those of the plantations, whereas the soil bulk density was lower than those of the plantations. No significant differences were found in soil pH, soil C content, or soil bulk density among the different plantations. In terms of soil N content, there was no significant difference among Eucalyptus, Pine, and Acacia plantations, which were each significantly higher than that of Casuarinas, and no significant difference was observed in the soil C: N ratio among the different forests.

The temperatures of topsoil and subsoil were both higher in summer and autumn and lower in spring and winter, which is consistent with the general rule (Fig. 3a and 3b). The water contents in both topsoil and subsoil under different forests showed apparent seasonal variations, and were always higher for a given forest type in winter and spring but lower in summer and autumn (Fig. 3c and 3d). In the topsoil, the soil water contents of the five forests in different seasons were all higher in secondary forest and lower in Casuarinas. In the subsoil, the soil water content in winter was higher in Eucalyptus, but lower in Pine and Casuarinas, and no significant differences were found in spring, summer and autumn among the different forests.

No significant difference was found in the seasonal variations of MBC content in the topsoil of Pine and Casuarinas plantations or in the subsoil of Acacia plantation, except that the MBC contents under different forests varied greatly among the different seasons (Fig. 4), being higher in spring or winter and lower in summer or autumn. The peak values of MBC content vary with forest type, which were 189.8, 101.6, 132.6, 58.5, and 536.1 mg kg^-1 in the topsoil and 128.3, 122.3, 55.4, 34, and 299 mg kg^-1 in the subsoil of Eucalyptus, Pine, Acacia, Casuarinas, and secondary forest, respectively.

The MBN content of different forests showed obvious seasonal variations, except for Eucalyptus (Fig. 4), which were higher in spring or winter and lower in summer or autumn. The peak values of MBN content were 12.7, 12.2, 19.9, 15.3, and 60.7 mg kg^-1 in the topsoil and 17.6, 9.6, 13.9, 11.8, and 60.5 mg kg^-1 in the subsoil of Eucalyptus, Pine, Acacia, Casuarinas, and secondary forest, respectively.

The annual mean MBC and MBN contents under different forests decreased with the increase of soil depth (Fig. 5). The mean MBC contents in Eucalyptus, Acacia, and Casuarinas differed significantly between topsoil and subsoil, but no significant differences were found in Pine and secondary forest between the two soil layers. The mean MBC content of secondary forest was significantly higher than those of the plantations in the same soil layer, and the mean MBC contents both in topsoil and subsoil of Eucalyptus, and in the subsoil of Pine were all significantly higher than the corresponding values of Casuarinas. No significant difference was found among the other plantations. The mean MBN contents differed significantly between topsoil and subsoil in Acacia, and no significant differences were found between the two soil layers under the other forests. The mean soil MBN content of secondary forest was significantly higher than those of the plantations in the same soil layer, and no significant difference was found in MBN content among the plantations in the same layer.

The soil MBC: MBN ratio in the topsoil of Eucalyptus was significantly higher than that of the subsoil. However, no significant difference was observed between topsoil and subsoil in the other forests. In the topsoil, the soil MBC: MBN ratio of Eucalyptus was significantly higher than those of the other forests, and no significant difference was observed among other forests. In the subsoil, the corresponding ratio was highest in Pine, followed by Eucalyptus and Acacia plantations, and lowest in secondary forest and Casuarinas (Fig. 6).

The percentages of soil MBC and MBN within TC and TN, respectively, in topsoil were all lower than those in subsoil under each of the different forests. The percentages of soil MBC within TC under Eucalyptus, Pine, Casuarinas, and secondary forest differed significantly between topsoil and subsoil. The percentages of soil MBN within TN under Eucalyptus, Acacia and secondary forest differed significantly between topsoil and subsoil. Specifically, the percentage of soil MBC within TC in the subsoil of Eucalyptus was significantly higher than that of Casuarinas. In addition, no significant differences were found among different forests either in topsoil and subsoil. The percentage of soil MBN within TN in the topsoil of Casuarinas was significantly higher than those of Eucalyptus, Acacia and Pine, and the corresponding value in secondary forest was also higher than that of Pine. However, no significant differences were found among different forests either in topsoil and subsoil (Fig. 7).

The effects of soil temperature and water content on the seasonal changes of soil MBC and MBN vary with forest types. However, the seasonal variations of soil MBC and MBN under different forests were all negatively correlated with soil temperature and positively correlated with soil water content (Table 3).

The MBC and MBN contents under different forests were significantly and positively correlated with TC, pH and litter N, and negatively correlated with soil bulk density and litter C. In addition, the MBN content was also significantly and positively correlated with root N both in topsoil and subsoil, and the MBC content in the subsoil was significantly and negatively correlated with root C (Table 4).

The qualities and quantities of soil C and soil N can be affected by forest types due to the variations in tree species composition, litter quantity and quality, root turnover rate, as well as root exudates in different forests (Paul et al., 2002; Pérez-Cruzado et al., 2012; Wang et al., 2013; Hoogmoed et al., 2014; Deng and Shangguan, 2017). In our study, the soil C and N contents of the secondary forest were significantly higher than those of plantations, and these differences were mainly attributed to the former having a longer stand age and higher litter quantity and quality. The development time of the secondary forest is longer than the plantations, so it takes longer to incorporate the C and N inputs into the soil through litter and fine roots. In addition, the secondary forest has higher annual litterfall and root biomass, as well as higher C and N contents with a lower C: N ratio (Table 1). Therefore, the secondary forest can input more C and N into the soil because of its more rapid litter decomposition and root turnover rates. No significant difference was found in soil C content between N-fixing trees and non-N-fixing trees in the plantations, the soil N content was lowest under Casuarinas, and no noticeable difference was observed in soil N content between Acacia and non-N-fixing trees. Generally, the N-fixing trees have higher soil C and N contents than non-N-fixing trees under the same management practices and site conditions (Resh et al., 2002; Ussiri et al., 2006; Wang et al., 2010), because of the higher N contents in their litter and root exudates. Decomposition of these components can increase soil N content (Forrester et al., 2007), which will accelerate the tree growth and increase the C input from roots to soil, and consequently, increase the soil C content (Resh et al., 2002; Jandl et al., 2007). However, some studies have suggested that phosphorus is the main factor that limits the growth and N-fixation of nodular plants (Pearson and Vitousek, 2001). So the lower available phosphorus content in soil may limit the growth of N-fixing bacteria, thus affecting the N-fixation capacity of N-fixing trees (especially non-legume trees), and the N-fixing trees without N-fixing capacity may input less C and N into the soil compared with non-N-fixing trees (Hoogmoed et al., 2014). Therefore, the lower soil N content under N-fixing trees in our study may be related to the lower soil available phosphorus content in coastal sandy dunes (Table 1).

The soil microbial biomass varies greatly among different forests, which is the combined result of tree species and environmental factors (Liu and Wang, 2010; Li et al., 2014). In our study, the MBC and MBN contents under secondary forest were significantly higher than those of plantations in the different soil layers, and the MBC contents under Casuarinas were the lowest, in both topsoil and subsoil. No significant difference in MBN content was observed among plantations. First of all, compared with the plantations, the diversity of tree species of secondary forest provided a variety of litter and root exudates with various quality and quantity characteristics, which would provide rich substrates for the soil microorganisms, and accelerate their growth and reproduction (Wardle, 2010). Secondly, the litter C: N ratio and lignin: N ratio of secondary forest were much lower and consistent with easier litter decomposition; however, the higher tannin content and higher C:N ratio in the litter of Casuarinas may limit its decomposition rate (Ye et al., 2012). In addition, the withered branches of Casuarinas were often harvested for fuel by local residents, which can greatly reduce the carbon input from its litter to soil, and ultimately affect the growth and reproduction of soil microorganisms.

The mean MBC and MBN contents decreased with the increase of soil depth under the different forests, which is consistent with previous findings (Hu et al., 1997; Li et al., 2014; Wang et al., 2016b). The MBC contents differed significantly between topsoil and subsoil under Eucalyptus and Casuarinas, and the MBN and MBC contents all differed significantly between topsoil and subsoil under Acacia. According to the previous studies, most of the soil microorganisms are heterotrophic types. The topsoil has a good water permeability and high pH due to the full contact with air; the quantity of litter input is higher and the root exudates are rich in nutrients in the topsoil, which is conducive to the growth and reproduction of soil microorganisms, and leads to the increased microbial activity (Yang et al., 2009; Luan et al., 2011; Wang et al., 2016b). With the increase of soil depth, however, the litter decomposition and root exudates are gradually reduced, which slows down the synthetic metabolism of microorganisms in the subsoil, and leads to decreases in MBC and MBN contents with the increase of soil depth (Hu et al., 1997; Li et al., 2014).

The MBC: MBN ratio reflects the relative ratio between fungi and bacteria in the soil (Qiu et al., 2010), and the fungi are dominant in a microbial community when the MBC: MBN ratio is between 4 and 15, while the bacteria are dominant when the MBC:MBN ratio is between 3 and 5 (Sarathchandra et al., 1988). In this study, the MBC: MBN ratio under Eucalyptus was significantly higher than those under other forests in topsoil, and the MBC: MBN ratio under Pine was significantly higher than those of other forests in subsoil, which means the advantages of bacteria in the soils of Eucalyptus and Pine plantations were higher than those of other forests in topsoil and subsoil, respectively. Studies have shown that under both Eucalyptus and Pine are eutrophic species of mycorrhizal fungi, and three types of ectomycorrhizal, endogenous mycorrhizal and mixed mycorrhizal are found under Eucalyptus (Zhu et al., 2001). The Pine is a mainly ectomycorrhizal species (Tam, 1994), while Acacia and Casuarinas are leguminous and actinomycetes plants, respectively. Therefore, there are large proportions of rhizobium and actinomycetes among the symbiotic bacteria under Acacia and Casuarinas plantations. As a result, the MBC: MBN ratios under Acacia and Casuarinas were lower than those of Eucalyptus and Pine. Furthermore, the MBC: MBN ratio under coniferous forests is generally higher than that of broad-leaved forests (Liu and Wang, 2010), which means that fungi are more dominant in coniferous forests than in broad-leaved forests. This may also be one of the reasons why the MBC: MBN ratio under Pine was higher than under the other forests in this study.

The percentages of soil MBC and MBN within TC and TN in topsoil, respectively, were lower than those in subsoil under the same forests, which indicate that the carbon and nitrogen accumulations in the subsoil are higher than in the topsoil, and the soil organic matter is transferred from topsoil to subsoil (Li et al., 2014). These percentages represent the conversion efficiency of soil organic carbon by microorganisms and reflect the effectiveness of the utilization of soil substances (Chang et al., 2011). Previous studies have shown that these percentages in broad-leaved forest were higher than in coniferous forest, which indicates that the ability to maintain soil microbial biomass and the accumulations of soil carbon and nitrogen under broad-leaved forest are higher than under coniferous forest (Li et al., 2014). In our study, the percentage of MBC within TC under Pine plantation was the lowest in topsoil (no significant difference), while the percentage of MBN within TN under Pine plantation was the lowest in both topsoil and subsoil, which is consistent with the existing results.

In this study, the MBC and MBN contents of different forests showed similar seasonal variations, which were single peak curves with higher values in spring or winter and lower values in summer or autumn. This agrees with the results from the coastal sandy dunes of Taiwan (Chen et al., 2005). Although May to September is the rainy season in the study area (Fig. 2), the soils in this sandy area have high temperatures and lack sufficient water-holding capacity during this period, resulting in the high evaporation dispersion. Therefore, the effective water content of the soil can be extremely diminished, leading to a significantly lower soil water content in summer and autumn than in winter and spring (Fig. 2). The lower soil water content in summer and autumn may limit the effectiveness of soil substrate activity; the soil microorganisms may die because of the lack of “food”, leading to the sharp decline in soil microbial biomass. In contrast, the suitable soil temperature and moisture in winter and spring provide a good metabolic environment for soil microorganisms, which promotes their growth and reproduction. In addition, the growth rate of plants tends to be higher in summer, which means a larger demand for soil nutrients. This may also be one of the important reasons for the lower soil microbial biomass in summer, because it would limit the availability of nutrients to soil microorganisms.

The soil pH is considered to be the main factor driving the spatial distribution of a soil microbial community, and pH is closely related to the microbial activity; when the pH value is lower than 7, the soil microbial activity increases gradually with an increase of soil pH (Shen et al., 2013; Feng et al., 2014; Landesman et al., 2014). As an important source of soil C and N, the quantity and quality of root exudates and litter can affect the input of organic carbon to the soil by providing carbon, nitrogen, and energy sources for microbial synthesis and metabolism (Wang et al., 2016b), and consequently, affect the soil microbial biomass. In this study, the differences of MBC and MBN contents under different forest types are closely related to the litter quality and soil physical and chemical properties. The MBC and MBN contents were significantly positively correlated with TC, pH, and litter N, and negatively correlated with soil bulk density and litter C. In addition, the MBN contents are also significantly positively correlated with root N in both topsoil and subsoil, whereas the MBC content in the subsoil is significantly negatively correlated with root C, which is consistent with the existing results (Xu et al., 2009; Li et al., 2014).