Soil Organic Carbon Mineralization and its Temperature Sensitivity Along a Vegetation Restoration Gradient in Subtropical China

Background and aims Soil organic carbon (SOC) mineralization produces important CO 2 ux from terrestrial ecosystems which can provide feedbacks to climates. Vegetation restoration can affect SOC mineralization and its temperature sensitivity (Q 10 ), but how this effect is related to soil moisture remains uncertain. Methods We performed a laboratory incubation using soils of different vegetation restoration stages (i.e., degraded vegetation [DS], plantation [PS], and secondary natural forest [SFS]) maintained under different moisture and temperature conditions to explore the combined effects of vegetation restoration and soil moisture on SOC mineralization and Q 10 . Results We found that cumulative SOC mineralization in PS and SFS were about 11.7 times higher than that in the DS, associated with higher SOC content and microbial biomass. Increased soil moisture and temperature led to higher SOC mineralization in the SFS and PS. However, in the DS, soil moisture did not affect SOC mineralization, but temperature enhancement solely increased (158.7%) SOC mineralization at the 60%MWHC treatment. Furthermore, signicant interactive effect of vegetation restoration and soil moisture on Q 10 was detected. At the 60%MWHC treatment, Q 10 declined with vegetation restoration age. Nevertheless, at the 30%MWHC treatment, Q 10 was lower in the DS than that in the PS. Higher soil moisture did not affect Q 10 in the PS and SFS, but enhanced Q 10 in the DS. Conclusions Our results highlight that the responses of SOC mineralization and Q 10 to vegetation restoration were highly dependent on soil moisture and substrate availability, and vegetation restoration reduced the inuence of soil moisture on Q 10 .


Introduction
Vegetation restoration has been expanding worldwide since the middle of the twentieth century due to environment conservation programs and policy incentives ( Previous studies reported that soil properties, such as SOC content, nutrients, C: N ratio, C quality, and soil microbes, will be changed along the vegetation restoration chronosequence (Berthrong et  It was shown that the ratio of actinomycetes to bacteria (Liu et al. 2017) and fungal phospholipid fatty acid (Qin et al. 2019) were positively correlated with Q 10 , whereas the abundance of gram-negative bacteria decreased with the increasing Q 10  . But, other studies have shown that changes in soil physicochemical properties rather than microbes control C mineralization (Zhang et al. 2019; Li et al. 2020). For instance, soil C: N ratio or C quality is recognized as the main factor affecting Q 10 values ). The large biochemical complexity of soils from a variety of ecosystems has led to great discussions about the sensitivity of SOC to temperature (Qi and Li 2017; Wang et al. 2019). However, when environmental factors were considered, the responses of SOC mineralization and Q 10 become more confusing (Suseela et al. 2012).
Soil moisture strongly affects SOC mineralization through soil aeration, substrate supply, and microbial activity (Suseela et al. 2012;Li et al. 2018; Moyano et al. 2013). In general, the optimum soil moisture for SOC decomposition is frequently found at intermediate levels, above or below which SOC decomposition rate decreases (Craine and Gelderman 2011; Suseela et al. 2012). Studies on the effects of soil moisture on Q 10 have produced inconsistent results. For example, Jiang et al (2013) observed a reduction in Q 10 of soil respiration in a mixed broad-leaved forest when simulated precipitation was doubled from ambient levels. Moonis et al (2021) found an opposite result that wetting treatment (50% water-lled pore space) increased Q 10 by 25.0%. A laboratory study found that the highest Q 10 of R h occurred at intermediate soil moisture levels (45%WHC), but the nature of this interaction varied between two different soils (Craine and Gelderman 2011). The inconsistency of soil moisture effects on SOC decomposition and Q 10 is possible due to the confounded effects of different environmental factors and soil properties (Pregitzer et al. 1999 [MWHC]) and temperature (18˚C vs. 28˚C). Soil physical, chemical and microbial properties were measured before the incubation. Soil microbial biomass C (MBC), dissolved organic C (DOC), microbial community composition and evolved CO 2 -C were measured throughout the incubation. On the basis of our knowledge, we predicted that (1) greater SOC mineralization would be found with vegetation restoration age; (2) increased soil moisture and temperature would enhance SOC mineralization in all vegetation restoration stages; (3) lower Q 10 would be found with vegetation restoration age, but soil moisture would have a less effect on Q 10 in the SFS than that in the DS, due to soils with longer vegetation restoration years always have a greater biodiversity and might be more adaptive to changes in environmental conditions such as soil moisture (Yang et al. 2018;Xu et al. 2021).

Study site
Soil samples used for incubation were collected from a typical vegetation restoration area (116°18′-116°31′E, 25°33′-25°48′N) of Changting county, Fujian province in southeastern China. The area is characterized by a subtropical monsoon climate, with mean annual temperature of 18.3℃ and mean annual rainfall of 1730 mm, of which approximately 75% occurred from March to August between 1981 and 2010. The soil of the study site is classi ed as red soil, equivalent to ultisol according to the USDA soil classi cation system (Buol et al. 2003).
Due to historical reasons, forest degradation of this area was very serious in the middle of the last century. Since the 1970s, the government began to organize vegetation restoration projects which led to the vegetation coverage rate increased notably (Lu et al. 2018).

Soil sampling and preparation
Soil samples were collected from three vegetation restoration stages, namely vegetation degraded soil (DS), plantation soil (PS) and secondary natural forest soil (SFS). Currently, there are almost no serious vegetation degradation areas in Changting County. Therefore, the DS was collected from a small vegetation degradation site where was protected for science, education and visit by the local government. At the vegetation degraded site, vegetation is quite sparse, soil erosion is severe and dwarf Pinus massoniana Lamb. is sporadically distributed. The PS was collected from plantation forest which was restored from 1998. The dominant species of the plantation forest are Pinus massoniana Lamb.. We collected the SFS from a secondary natural forest (with age > 70 years) which was protected as soil samples were passed through a 2-mm sieve and homogenized. Three soil samples of each vegetation restoration site were used for determining chemical and physical properties and phospholipid fatty acids (PLFAs). The remaining soil samples were stored at 4°C for weeks prior to the incubation experiment.

Soil properties and PLFA analyses
Soil pH was measured using a soil: water ratio of 1: 2.5. Soil water content was determined using the methods of oven-dried and weighted. The contents of SOC and total nitrogen (TN) measured by an elemental analyzer (EA3000, EuroVector, Italy). Total phosphorous (TP) was measured photometrically after samples were digested with perchloric acid and hydro uoric acid.
Soil MBC was determined by the chloroform fumigation-extraction method following Brookes et al (1985). The extractable C from the unfumigated samples was considered as DOC. Particulate organic C (POC) was determined using a method that was described by Fang et al (2020). Brie y 15 g of air-dried soil was placed in a 100 mL of 5 g L − 1 sodium hexametaphosphate solution by handshaking the mixture for 5 min prior to being placed on a reciprocal shaker (90 r• min − 1 ) for 18 h. The dispersed soil sample was passed through a 53 µm sieve and rinsed with deionized water. The material that remained on the 53 µm sieve was considered the POC fraction. All of the samples were dried at 60°C, weighed, then nely ground for the determination of organic C.
The organic C content was converted to bulk soil POC content according to the fraction mass ratios.
The measurements of PLFAs pro les were determined on 8 g of freeze-dried soil after extraction using a mixture containing chloroform, methanol and citrate buffer (Bossio and Scow 1998). And then the extracted fatty acid methyl esters were analyzed following the procedures described by Fang

Soil incubation experiment
Soil samples were incubated under two soil moisture (60%MWHC vs. 30%MWHC) and two temperature (18℃, close to the annual mean temperature at the sampling site, and 28℃ for obtaining Q 10 )  regimes for 120 days. Such a design produced 12 treatment combinations with 9 replicates per combination (three vegetation stages × two moistures × two temperatures × 9 replicates). Each incubated soil sample (50 g dry) was placed in a 250 mL Erlenmeyer asks covered by a sealing which has small holes to maintain gas exchange and is used to reduce evaporation. Soil moisture of each sample was adjusted to the designed contents by injecting distilled water slowly on the sample surface to ensure uniform moisture penetration. And soil samples were rst pre-incubated at 23°C for 1 week to minimize the "plus effect" and then put into incubators set at 18 and 28℃, respectively. To maintain constant soil moisture levels, soil water was checked and adjusted every 4-5 days by weighing it. In the incubators, a certain amount of air was introduced through the air compressor to avoid anaerobic conditions (Whitaker et al. 2014).
During the incubation, three replicates of each combination were used for the determination of soil CO 2 evolution at days of 1, 3, 5, 7, 10, 15, 20, 25, 30, 45, 70, 95 and 120. Compressed air was used to ush the headspace for ca. 60 s to standardize the starting CO 2 concentration of each incubation ask (Whitaker et al. 2014) during each measurement. Gas samples were collected twice from the headspace, immediately after closing the ask and 30 min later, with a needle cylinder and stored in a gas sampling bag. The CO 2 concentration of gas samples was determined by gas chromatography (7890B, Agilent, USA) within 24 h. The standard gases used to calibrate the gas chromatograph included four different CO 2 concentrations in the N 2 makeup gas. SOC mineralization rate (R) was calculated according to Huang et al 2019 using the following equation: where v is head space volume of incubation ask (total volume of ask minus soil volume), m is dry soil weight, △c/△t is the average CO 2 concentration difference per hour, T is the incubation temperature, and C M is the molar mass of C. The ideal gas law was used to determine the molar volume of CO 2 at the incubation temperature.
The cumulative amount of SOC mineralized (C m ) was calculated using the following equation: Where R is daily SOC mineralization rate, p is incubation period, and D is incubation time (day). We calculated the temperature sensitivity (Q 10 ) of SOC mineralization during the incubation as follows (Conant et  Where t c and t w are the times required to respire a given amount of soil C at relatively cold (T c , 18°C) and warm (T w , 28°C) temperatures during incubation.
Three replicates of soil samples were collected for each treatment combination at day of 15, 45, and 120 after the start of the experiment. After soil samples were passed through a 2-mm sieve, each soil sample was divided into two parts. One part was stored at 4°C and used for determination of soil water content, MBC, and DOC content. The other part was freeze-dried and used for determination PLFAs as mentioned above. The metabolic quotient (qCO 2 ) and the decomposability of DOC (R h : DOC) were estimated by dividing the cumulative mineralization by the corresponding MBC and DOC, respectively.

Statistical analysis
Data were logarithmically transformed to meet the assumptions of normality and homogeneity of variances when necessary. The differences of soil properties among three vegetation restoration stages were tested using one-way analysis of variance (ANOVA). Three-way ANOVA was used to test the effects of vegetation restoration, moisture, temperature and their interaction on cumulative SOC mineralization, MBC, DOC, qCO 2 , R h : DOC, and microbial composition. Two-way ANOVA was used to test the effects of soil moisture, temperature and their interaction on cumulative SOC mineralization, MBC, DOC, qCO 2 , R h : DOC, and microbial composition

SOC mineralization
After the 120-day incubation, cumulative SOC mineralization was signi cantly affected by vegetation restoration, moisture, temperature and their interactions (Fig. 1, Table 1). Cumulative SOC mineralization was signi cantly higher in the PS (1522.7 µg CO 2 -C g − 1 soil) and SFS (1521.3 µg CO 2 -C g − 1 soil) than that in the DS (129.5 µg CO 2 -C g − 1 soil). Both increased soil moisture and temperature led to signi cantly higher SOC mineralization in the SFS and PS (Fig. 1). However, in the DS, increased soil moisture has little effects on SOC mineralization, and increased soil temperature only increased (P < 0.001) SOC mineralization by 158.7% at the 60%MWHC treatment.  (Fig. 2), but we detected signi cant interaction between the two factors (P < 0.01). Then, we compared the Q 10 among different vegetation restoration stages at the same soil moisture level or between two soil moisture levels at the same vegetation restoration stage. We found that Q 10 was signi cantly higher in the DS than that in the PS and SFS at the 60%MWHC treatment, but Q 10 was signi cantly lower in the DS than that in the PS and SFS at the 30%MWHC treatment. Increased soil moisture did not markedly in uence Q 10 Fig. 3). Soil DOC content was signi cantly affected by vegetation restoration stage, soil moisture, temperature, and the interaction of vegetation restoration and soil moisture. Soil MBC and DOC contents were signi cantly higher in the SFS and PS than that in the DS (Fig. 3

Soil microbial composition
The relative abundance of GP and GN was solely signi cantly affected by vegetation restoration, and the relative abundance of fungi was signi cantly in uenced by vegetation restoration, soil moisture, temperature and the interaction of vegetation restoration and soil moisture. (Table 1, Fig. 5).The relative abundance of GN was signi cantly lower in the DS than that in the PS and SFS. On the contrary, the relative abundance of GP and the ratio of GP: GN was signi cantly higher in the DS than that in the PS and NS. The relative abundance of fungi was signi cantly higher in the PS than that in the DS and SFS. The relative abundance of fungi was higher at 30%MWHC than that at the 60%MWHC in the PS. At the 60%MWHC treatment in the SFS and at the 30%MWHC treatment in the DS, increased temperature decreased the relative abundance of fungi.

Soil properties among vegetation restoration stages and correlations
Vegetation restoration has signi cant in uences on the soil pH, SOC, TN, TP contents, and microbial composition ( Table 2). Soil pH was higher (P < 0.05) in the DS than that in the PS and SFS. On the contrary, the SOC, TN, TP, MBC, DOC, and POC were showed the order: DS < PS < SFS. The total PLFAs content and the relative abundance of GN were also showed the order: DS < PS < SFS, but the relative abundance of GP was signi cantly higher (P < 0.05) in the DS than that in the PS and SFS. The relative abundance of fungi was signi cantly higher in the PS than that in the DS and SFS. The ratio of GP: GN and F: B followed the order: DS > PS > SFS. SOC is soil organic carbon, TN is total nitrogen, TP is total phosphorus, MBC is microbial biomass carbon, DOC is dissolved organic carbon, POC is particulate organic carbon, GP is gram-positive bacteria, and GN is gram-negative bacteria.
At the end of the incubation, soil cumulative SOC mineralization was positively correlated (P < 0.001) with the content of MBC, and increased exponentially with the increase (P < 0.0001) of DOC content (Fig. 6) in the DS than that in the PS and SFS (Fig. 1, Table 1) was found. Lower SOC mineralization in the DS would be attributed to lower total SOC or lower labile SOC, mainly POC, DOC and MBC contents ( Table 2) Despite very different soil properties between the PS and SFS, the Cumulative SOC mineralization did not differ. The possible reason is that, although total SOC was signi cantly lower in the PS than that in the SFS, the amount of POC and MBC was slightly smaller ( Table 2, P > 0.05) in the PS than in the SFS, which likely resulted in similar amount of SOC mineralization in the PS and SFS. Likewise, Huang et al (2019) showed that although total SOC was higher in the natural forest soil, the amount of labile SOC was 27-28% greater in the plantation soil leading to greater SOC mineralization in the forest plantation soil than that in the natural forest soil. There are also some incubation experiments revealed that the uxes of CO 2 were positively correlated with DOC during the several months of incubation suggested that soil CO 2 emission was mainly from DOC (Conant et al. 2008;Gershenson et al. 2009). The DOC content did not differ between the PS and SFS which could partly explain the similar amount of SOC mineralization in the PS and SFS (Fig. 3, 6).  (Fig. 4). Such changes in qCO 2 suggesting that increased temperature stimulated microbial respiration in the in the PS, SFS, and 60%MWHC level in the DS in our study. Although microbial biomass and the abundance of major microbial groups were not remarkably altered by temperature (Fig. 3, 5), Schindlbacher et al (2011) also reported that ve years of soil warming did not affect microbial biomass or the abundance of major microbial groups, but signi cantly increased qCO 2 , thus leading to an increase in soil respiration.
However, increased temperature did not enhance SOC mineralization under 30%MWHC level in the DS which may be relate to the interactive effect of soil water de cit and substrate limitation (Suseela et  In the PS and SFS, we found that SOC mineralization was signi cant higher at 60%MWHC level (almost at an optimum soil water content for microbial respiration) than that at 30%MWHC level. As mentioned above, suitable soil moisture can facilitate the diffusion of soluble C substrates ) and improve microbial substrate availability (Balogh et al. 2011), which can stimulate the microbial decomposition of SOC.
Compared to 30%MWHC, signi cantly increased MBC and DOC content at 60%MWHC level in the PS and SFS (Fig. 3, 6) could support this point. But increased soil moisture did not stimulate SOC mineralization in the DS, which may due to the limited soluble C substrates in the DS (Moyano et al. 2012). In the DS, 60%MWHC treatment did not increase DOC content under both two incubation temperatures also indicated the substrate limitation (Fig. 3). In addition, the inherent lower labile SOC content (Table 1) in the DS also should be responsible for the substrate limitation for microbes. These ndings suggested that, to some extent, rising moisture and temperature would stimulate SOC decomposition, but these responses are highly dependent on soil properties or the quantity and quality of the substrates. that the response of Q 10 to vegetation restoration was highly dependent on soil moisture. In detail, at the 60%MWHC treatment, Q 10 was decreased with vegetation restoration age (Fig. 2), this partly supports our expectation. Inversely, at the 30%MWHC treatment, Q 10 was signi cantly lower in the DS than in the PS and SFS (Fig. 2). Likewise, previous studies presented soil property may interact with numerous climatic variables, such as temperature and precipitation, to in uence Q 10 (Table 2), and the incubation experiment done little to reverse this trend (Fig. 5).

Responses of Q 10 to vegetation restoration and soil moisture
We found Q 10 was signi cantly lower in the DS than in the PS and SFS at the 30%MWHC treatment (Fig. 2). result that increased soil moisture did not affect Q 10 in the PS and SFS, but enhanced Q 10 in the DS, which is consistent with our expectations that soil moisture had a less effect on Q 10 in the SFS than in the DS. The potential mechanism could also explained by the substrate availability. In our study, water de cit limited microbial biomass at the 30%MWHC in the DS which could lead to a lower Q 10 (Fig. 3, 6). Previous studies also have shown that drying can decrease Q 10 of soil respiration, and which was attributed to substrate limitation

Conclusion
In this study, we found that cumulative SOC mineralization in PS and SFS was about 11.7 times higher than that in the DS, possibly due to the quite lower SOC content and microbial biomass in the DS. Increased soil moisture and temperature led to signi cantly higher SOC mineralization in the SFS and PS. However, increased soil moisture did not affect SOC mineralization in the DS, and increased temperature solely increased SOC mineralization at the 60%MWHC treatment in the DS. The discrepancy responses of SOC mineralization to moisture and temperature indicated that, to some extent, rising temperature and moisture would stimulate SOC decomposition, but these responses are highly in uenced by soil inherent substrate availability. Higher soil moisture did not affect Q 10 in the PS and SFS, but enhanced Q 10 in the DS. The Q 10 value declined (P < 0.05) with vegetation restoration age at the 60%MWHC treatment, but it was signi cantly lower in DS than that in the PS at the 30%MWHC treatment. Our results suggested that the response of Q 10 to vegetation restoration was highly dependent on soil moisture and substrate availability.  Regression plots linking cumulative CO2-C emission to soil microbial carbon (MBC) and dissolved organic carbon (DOC) content. The dotted lines represent 95% con dence intervals.