The soil tested in this experimental study was collected from the Jiulongquan gully reclamation area of Nanniwan Town, Baota District, Yan’an City in Loess Plateau (Figure 1). The average temperature in the gully watershed is 8.1°C, the annual rainfall is 530–600mm, and the frost-free period is 120–140 days. The thickness of the soil layer is 0.4–5.0 m in gully. The soil is developed on the loess and shale parent material, with obvious profile layers. The soil upper part (0–60 cm) is dense and the lower part (60–80 cm) has obvious viscosification (Ma et al., 2020a). The mean soil grain size distribution from 18 samples were described as follows: jiulongquan gully: clay (<0.002 mm; 6.3%), silt (0.002–0.05 mm; 66.3%) and sand (>0.05 mm; 27.4%), mengjia gully: clay (<0.002 mm; 5.7%), silt (0.002–0.05 mm; 59.1%) and sand (>0.05 mm; 35.2%) (Ma et al., 2020b). Both are silty loam soil. The two gullies belong to Nanniwan watershed. The main crops are maize, there are abandoned land, barren grass land and damaged forest land in Mengjia gully. Jiulongquan gully is the main gully, which is 9.8 km long, and the valley width is generally between 250 and 500 m. The gully erosion control and land construction project has been implemented in stages. The total construction scale was 662.96 hm². The completion time was 2014 at the earliest and 2020 at the latest. Mengjia gully is a branch gully, which is 8.3 km long and 120–150 m wide (Lei et al., 2024). The project has not been implemented. The both gully approximately 1 km apart have similar climates, channel shapes, and soil conditions.

The sampling times were April 2023. Using the method of spatial replaced time, the control soil that was collected in Mengjia gully (N36°15′48″∼36°19′13″, E109°40′22″–109°42′41″). The soil in the tilled layer (0–20 cm) was collected 3 years after construction (completed in 2020), and 9 years after construction (completed in 2014) in Jiulongquan gully (N36°14′40″∼36°19′25″, E109°35′50″∼109°39′50″). All samples were transported back to the laboratory placed in incubators. Ice plates and ice cubes sealed with ziplock bags are placed at the bottom, around and top of the incubator. Part of the sample (1 kg) was immediately passed through a nylon screen with a aperture of 2 mm, mixed, sealed and stored in a refrigerator at 0°C–4°C away from light for DOC and MBC, the other part was naturally air-dried for SOC and other soil physicochemical properties determination, and the rest of the soil sample was used for the layout of cultivate test.

The experiment was conducted at the Fuping pilot base of Shaanxi Institute of Land Construction and Land Engineering Technology. The size of the cultivation pots was 10 cm × 20 cm of PVC material pipe. Five kilograms of soil were placed in each cultivation pot. Thirty days after planting in the maize field, maize plants showing the same growth pattern were selected and transplanted into cultivation pots. Nine pots were planted, three of which were the control group (soil collected from Mengjia gully), three pots were from the 3-year construction group and three pots were from the 9-year construction group (soil collected from jiulongquan guly). After the maize was transplanted, all treatments were watered 500 mL every 3 days. The cultivation pots were placed in an artificial climate room in order to label the cultures in a closed system with a labelling test time of 100 days. Light was provided for 12 h per day (8:00–20:00), and the plants were maintained at a relative humidity of 80%–90%. Continuous labelling of carbon isotopes (¹⁴C-CO₂) was self-generated by a reaction of NaHCO₃ (1 mol•L^-1) and HCL (1 mol•L^-1). The CO₂ concentration in the greenhouse was maintained at about 300–350 μmol·L^-1 by controlling the addition amount (Ge et al., 2012).

After 100 days of marked culture, destructive sampling was performed, the maizewas cut along the base, and shoots and roots were collected. The soil attached to the root system was put into a CaCl₂ (0.01 mol•L^-1) solution, shaken gently, and then washed with distilled water. After the crowns and roots greened, they were dried to a constant weight. SOC and DOC were determined using a German TOC/TN analyser (MultiC/N 3100). Soil microbial carbon was determined using the chloroform fumigation-K₂SO₄ extraction method, and ¹⁴C in the soil was determined using the method proposed by Wu et al. (Wu and O'Donnell, 1997).

Microsoft Excel software was used for data processing, and SPSS 22 software was used for statistical analysis. The significance of differences among different treatments was tested by one-way ANOVA.

The ¹⁴C-DOC/DOC ratios were 6.72%, 7.73%, and 14.64% before, 3 years after, and 9 years after reclamation, respectively. Compared with the control, ¹⁴C-DOC/DOC increased by 15.03% and 89.39% after 3 and 9 years of reclamation, respectively. Before, 3 years after, and 9 years after gully reclamation, the ¹⁴C-MBC/MBC ratios were 1.70%, 7.19%, and 7.67%, respectively. ¹⁴C-MBC/MBC increased by 322.94% and ¹⁴C-MBC/MBC by 6.67% at 3 and 9 years following gully reclamation. The soil microbial content increased significantly after gully reclamation, and the rate of increase showed a decreasing trend. After 9 years of project implementation, maize root input contributed the most to soil soluble organic carbon. In the evaluated period, the values of ¹⁴C-DOC/DOC of maize were all greater than ¹⁴C-MBC/MBC (Table 3).

The ratios of ¹⁴C-DOC/¹⁴C-SOC were 3.54%, 2.42%, and 2.48% before, 3 years after, and 9 years after gully reclamation, respectively. ¹⁴C-MBC/¹⁴C-SOC were 10.92%, 9.72%, and 17.97% after 3 and 9 years of gully reclamation. The proportions of ¹⁴C-DOC/¹⁴C-SOC and ¹⁴C-MBC/¹⁴C-SOC in the soil decreased after 3 years but began to increase again after 9 years.

¹⁴C-SOC/SOC represents the renewal rate of SOC, and the ¹⁴C-SOC/SOC ratios were 0.73%, 1.28%, and 1.99% before, 3 years after, and 9 years after gully reclamation. Compared with the control, the ¹⁴C-SOC/SOC increased by 75.34% after 3 years and 55.49% after 9 years, respectively. As the project implementation time increased, the renewal rate of organic carbon increased, but the renewal rate slowed.

In this study, the proportion of photosynthetic carbon fixed in the plants, roots and soil after the end of the growing period of maize is similar to that of previous studies on rice, ryegrass, wheat and forage (Lu et al., 2004; Kuzyakov and Schneckenberger, 2004; Lu et al., 2005; Hütsch et al., 2002; He et al., 2008). The photosynthetic production capacity of maize was not the same in different construction years, but the root-shoot ratio was very close (0.11–0.13), and the rhizosphere deposition efficiency was similar. This may be because during the process of carbon input from maize photosynthetic carbon into the ground, a part of carbon was returned to the atmosphere through plant respiration and rhizosphere microbial respiration, and the construction mainly affected this part of carbon, while the effect on root-shoot ratio was negligible. After a certain number of years of construction, with the increase of planting years, the soil gradually ripened (organic matter content>10 g·kg^-1, soil salt content<1 g·kg^-1, soil erosion<1,000 t·(km.a)⁻¹), creating a good condition and foundation for the conversion of photosynthetic carbon. The distribution of photosynthetic carbon in the SOC pools in the reconstructed soils increased, but the increase amplitude decreased with the soil ripening. In the long run, Compared to barren gullies, soils under gully management had higher levels of photosynthetic carbon. This may be due to the implementation of organic land reconstruction and irrigation and drainage projects to regulate farmland groundwater level, ensured reasonable water content of farmland layer, prevented natural disasters such as drought and flood and soil erosion, improved soil quality, provided high-quality microenvironmental conditions for maize plant growth and root respiration, and enhanced the photosynthetic production capacity of maize. These changes conducived to the accumulation of SOC (Feng et al., 2022). In the loess Plateau, the mixing of sand and soft rock can be reconstructed to form a “soil” with good water and fertility preservation, which can reduce soil erosion and improve the soil carbon sequestration capacity of sandy land (Guo et al., 2023). In the area of bare rock and gravel land, soil reconstruction can be carried out with sand, shale and other materials, which can solve the problems of barren land, water leakage and fertilizer leakage in the region, and was conducive to soil carbon sequestration (Lei et al., 2022). All these proved that soil reconstruction to improved soil quality and ecological environment was conducive to soil carbon sequestration. The gully erosion control and land construction project reduced the sediment transport of the basin by about 10%, and increase the soil layer thickness and soil fertility of the cultivated land (Song, 2020b). In addition, various fertilization measures also increased the biomass of maize to varying degrees, thus increasing the input of SOC (Wan et al., 2021).

DOC and MBC were the main destinations for new carbon in the soil activated carbon pool and were closely related to the release of CO₂ and CH₄ via soil respiration. This also explained that DOC and MBC were often used as indicators to characterize soil active organic carbon, and were also considered to be sensitive indicators to evaluate the response of land use type change to SOC (Kuzyakov and Jones, 2006). MBC only accounts for 1%–5% of SOC, but the size of MBC reflects the activity of microorganisms and the rate of turnover of SOM and nutrients. MBC can control the accumulation and loss of SOC through the decomposition and transformation of input SOM (Gil-Sotres et al., 2005). During the growth period of maize, the renewal rates of DOC and MBC were 14.64% and 7.67% respectively, while SOC was 1.9%, indicating that the input of photosynthetic carbon into maize rhizosphere had a greater impact on the change of DOC and MBC contents than SOC, which was basically consistent with previous studies (Wang et al., 2004). This may be because the rhizosphere secretion was an important source of DOC during the growth of maize, and it was also a raw material for the growth of soil microorganisms, thus affecting the dynamic changes of DOC and MBC. The ¹⁴C-DOC/DOC of maize were all greater than ¹⁴C-MBC/MBC, indicating that the effect of construction on DOC was greater than that of MBC. This may be due to the fact that water resources were controlled mainly through irrigation and drainage in gully erosion control and land construction project, soil water content changed greatly before and after reconstruction, and DOC was more affected by water content than MBC. In addition, it was found that maize growth would significantly affect ¹⁴C-DOC and ¹⁴C-MBC, which may be due to the higher photosynthetic productivity and root biomass of maizein the reconstructed soils system, and the increase of available carbon sources to soil microorganisms, thus changing the content of rhizosphere deposited carbon as well as the population and quantity of soil microorganisms (He et al., 2008).

It was generally believed that the depolymerisation and dissolution of SOC are prerequisites for its mineralisation. Some studies indicate that the mineralisation of SOC was carried out by microorganisms and was affected by many factors, such as temperature, water, soil properties and plant communities (Karhu et al., 2022). Previous studies have shown that increases in atmospheric CO₂ concentration and temperature do not increase the accumulation of SOC, but accelerated the turnover rate of SOC (Fang et al., 2022). Soil carbon mineralization rate was negatively correlated with soil clay content. Due to different litter and root exudates of different plant communities, the SOC in soil also change (Yang et al., 2022). During the incubation period of this study, the cumulative mineralization amount of ¹⁴C-CO₂ continued to increase, and the mineralization rate gradually decreased and stabilized. This was because the engineering disturbance in the early stage of construction was large. With the increase of construction years, the physical and chemical properties of soil changed, the soil quality gradually improved, and the mineralization rate gradually stabilized. In the early stage of experiment, the mineralization rate of the soil tested was relatively fast, mainly because the SOC and DOC in the soil in the early stage were decomposed rapidly, and a large amount of nutrients were released rapidly, which promoted the microbial activity. However, with the extension of cultivate time, the easily decomposed components in the soil began to turn to the components that were difficult to decompose after being used up by microorganisms, and the mineralization rate slowed down. This was similar to the research conclusion of Zhang P. et al. (Zhang et al., 2011).

During the experimental period, the variation of ¹⁴C-DOC fluctuated frequently, with no apparent pattern. The difference of ¹⁴C-DOC was obvious among different construction years. This may be due to the low content of ¹⁴C-DOC and the instability of the detected data, or it may be related to the soil structure and soil water content, which was one of the main environmental factors affecting the decomposition, transformation and accumulation of SOM (Schurer et al., 2022). At the initial stage of cultivation, the mineralized amount of ¹⁴C-MBC in soil of different construction years was relatively large, and with the progress of cultivation time, it tended to stabilize after 40 days of culture. This was because, water, O₂ and nutrients were abundant, and the transformation process was active during the early stages of cultivation. Later, these substances were gradually consumed and reduced, and the transformation process slowed or even stopped. Another reason was that the microbial substrate was gradually utilised in the later stage, the nutrients in the cultivation system were limited, microbial activity was stabilised, and the transformation process tended to be stable (Six et al., 2006).