Effect of ecological restoration on carbon storage of damaged mountain slope in China’s East Ussuri River Basin

The survey site

The survey site is located in the east of Ussuri River Basin (48°22′N, 134°28′E, 40–60 m altitude) (Fig. 1). The region is affected by monsoons, with cold and dry winters, and warm and humid summers. The annual average temperature is 3.4 °C and the annual average precipitation is 954.7 mm. The natural ecosystem consists of dark brown soil and a secondary forest dominated by Ulmus pumila and Betula platyphylla (Liang et al., 2022; Liu, Liao & Chang, 2024).

In the study region, roads construction has resulted in severely degraded mountain slope (Fig. 2). To address this environmental challenge, two restoration strategies were implemented in 2015: (1) Aggregate spray-seeding restoration technology (ASS) Developed by Qingdao Greensum Ecology Co., Ltd., China (Chen et al., 2024), this technique involved spraying a 7–10 cm growth medium composed of clay, organic additives, soil amendments, and necessary slow-release fertilizers. To facilitate natural succession, the seed bank incorporated native plants (Ulmus pumila, Betula platyphylla), and pioneer species (Lespedeza davurica, Caragana microphylla, Hippophae rhamnoides Linn, and Amorpha fruticosa). This multispecies approach targets both immediate erosion control (pioneers) and long-term carbon sequestration (native trees); (2) planted forests. This approach focused on manually planting native saplings (Ulmus pumila, Betula platyphylla) with pit planting technique (Jo & Park, 2017). Both restoration projects were supplied with regular irrigation and weeding during the first 3 years to ensure establishment.

Site design

Field surveys were conducted in August 2023 during the peak vegetation growth season. We selected three slope types: aggregate-sprayed restored slope (S1), planted forest slope (S2), and undamaged natural slope (S3). We established three 20 × 20 m² survey plots in each slope type (nine plots total). Plant community assessments targeted the tree layer, shrub layer, herbaceous layer, dead wood layer, and litter layer. A nested quadrat design was employed to assess plant community: a central 10 × 10 m² quadrat was surveyed tree layer for species composition, height, and diameter at breast height (DBH). A nested 2 × 2 m² quadrat within the tree quadrat was used to record shrub composition and diameter at breast height (DBH). A 1 × 1 m² subplot within the shrub quadrat quantified herbaceous herbaceous composition. Dead wood volume (decay stages) and litter thickness/biomass were measured within the 10 × 10 m² quadrat. In our survey, trees and treelets were identified as DBH ≥2 cm, others were treated as shrub (Deb, Roy & Wahedunnabi, 2015).

Aboveground biomass samples (trees, shrubs, herbs) were collected, placed in labeled paper bags, and oven-dried at 65 °C for 48 h. Dried samples were ground (<0.2 mm) for carbon (C) concentration analysis. Given the rocky substrates and shallow soil profiles (≤10 cm depth), composite soil samples were collected from five systematically randomized cores (0–10 cm depth) per plot. Samples were homogenized, air-dried, sieved (2 mm), and analyzed for biochemical properties. Plant and soil C content was determined using via oxidation with potassium dichromate under acidic conditions, followed by titration measurement, soil N content was determined using the Kjeldahl method, and soil P content using the phosphorus molybdenum blue photometry method after digested in sulfuric acid-hydrogen peroxide (Nelson & Sommers, 1973).

Carbon storage calculation

The carbon storage of plant tree layer is calculated using Eqs. (1)–(6) (Eggleston et al., 2006; Zeng, 2017). The aboveground biomass of trees is calculated using Eq. (1) (Zeng & Tang, 2012). Use Eqs. (2)–(4) to derive Eq. (5) (Zeng & Tang, 2011), and the belowground biomass is calculated using Eq. (5).

(1) $$M_a=a\times D^b$$where $M_a$ is aboveground biomass (kg), $\mathrm{a}$ is parameter (a = 0.3ρ), ρ is basic wood density (g/cm³), D is diameter at breast height (cm), b is parameter (b = 7/3).

(2) $$M_a=a_1\times D^{a_2}+{\epsilon }_{Ma}$$

(3) $$M_b=b_1\times D^{b_2}\times {\hat{M}}_a+{\epsilon }_{Mb}$$

(4) $$R=b_1\times D^{b_2}+{\epsilon }_R$$where $M_a$ is aboveground biomass (kg), $M_b$ is belowground biomass (kg), R is the root to shoot ratio (the ratio of belowground to aboveground biomass), D is diameter at breast height (cm), ${\hat{M}}_a$ is the predicted value of aboveground biomass, $a_i$ and $b_i$ are model parameters, ${\epsilon }_i$ is the error term.

(5) $${\hat{M}}_b=c_1\times D^{c_2}$$where c₁ = a₁ × b₁ = 0.0599, c₂ = a₂ + b₂ = 2.16; ${\hat{M}}_b$ is the predicted value of belowground biomass.

(6) $$C_{tree}=(M_a+{\hat{M}}_b)\times C{F}_{tree}$$where C_tree is the carbon storage per unit area (kg/m²), $M_a$ is aboveground biomass (kg), ${\hat{M}}_b$ is the predicted value of belowground biomass, CF_tree is the organic carbon content (g/kg).

The carbon storage of shrub layer, herb layer, dead wood layer, and litter layer are calculated using Eq. (7) (Eggleston et al., 2006):

(7) $$C_x=B_x\times C{F}_x$$where C is the carbon storage per unit area (kg/m²), B is the biomass (kg/m²), CF is the organic carbon content (g/kg), and x is shrub, herb, dead wood, and litter.

The soil carbon storage is calculated using Eq. (8) (Eggleston et al., 2006):

(8) $$C_s=\rho b\times C{F}_s\times 0.1$$where C_s is the soil carbon storage per unit area (kg/m²), ρb is the soil bulk density (g/cm³), and CF_s is the soil organic carbon content (g/kg), focusing on a 0.1 m soil thickness in this survey.

Data statistics

The carbon storage of aggregate spray seeding restored slopes, planted forests restored slopes, and naturally undamaged slopes is represented by the average values of three sample plots. To analyze the relative contributions of plant density and soil physicochemical properties to ecosystem carbon storage, the R language relaimpo package is utilized. Differences in measured variables across slope types were assessed using one-way ANOVA with Tukey’s post-hoc test applied where ANOVA results were significant (P < 0.05). Linear regression analysis via the Origin analysis tool was performed to evaluate relationships between ecosystem carbon storage and soil bulk density, soil total nitrogen, and plant density, thereby elucidating key drivers of carbon sequestration. All statistical analyzes were performed in R version 3.6.1 (R Core Team, 2019).

Current status of vegetation and soil carbon storage

The total carbon storage differed significantly (P < 0.05) across restoration types, with S3 (undamaged slope) exhibiting the highest storage (10.64 kg/m²), followed by aggregate spray seeding restored slope S1 (5.95 kg/m²) and planted forest slope (1.14 kg/m²) (Fig. 3, Table 1).

Carbon partitioning revealed distinct patterns: At S1, aboveground storage was dominated by trees (1.53 kg/m²) with minimal shrub (0.02 kg/m²), herbaceous (0.0033 kg/m²) and litter (0.0057 kg/m²) contributions, while belowground pools were soil-dominated (3.93 kg/m² vs. roots: 0.47 kg/m²). S2 showed negligible aboveground storage (shrubs: 0.0073 kg/m²; herbs: 0.0017 kg/m²; litter: 0.0014 kg/m²) with soil accounting for virtually all carbon (1.13 kg/m²). In contrast, S3 displayed strong tree carbon accumulation (1.59 kg/m²) and exceptional belowground soil storage (8.58 kg/m²), dwarfing root contributions (0.44 kg/m²) (Table 2).

Analysis of biological and soil physical and chemical properties in slope ecosystems

Vegetation density and structural attributes differed significantly among slope ecosystems (P < 0.05). The Shannon-Wiener index reached its maximum at S1, significantly exceeding values at S2 and S3 (which showed no statistical difference). Conversely, the Simpson index peaked at S2, followed sequentially by S1 and S3 (Fig. 4, Table 1). The total plant density varied markedly across sites: aggregate spray-seeded slope (S1) recorded 5.16 plants/m², planted forest slope (S2) 2.35 plants/m², and undamaged reference slope (S3) 3.47 plants/m². At S1, tree, shrub, and herbaceous layer densities significantly surpassed those at S3, though tree diameter at breast height (DBH) was significantly lower than in the reference ecosystem (Tables 1, 3).

Comparative soil analysis revealed distinct physicochemical gradients (Tables 1, 4). The aggregate spray-seeded slope (S1) exhibited significantly higher soil water content (SWC), total nitrogen (TN), total phosphorus (TP), and total potassium (TK) than both planted (S2) and undamaged (S3) slopes. Conversely, soil bulk density at S1 was significantly lower than at S2 and S3.

Analysis of influencing factors on ecosystem carbon storage

Variance partitioning analysis revealed that plant and soil factors collectively explained 99.8% of carbon storage variation in aggregate-sprayed restored slope (S1), with vegetation contributing 19% to this model (Fig. 5). Soil physicochemical properties dominated the explanatory power (80.8%), where soil bulk density emerged as the primary predictor (23.7%), followed sequentially by soil total nitrogen (19.2%), soil total potassium (14.8%), soil total phosphorus (11.7%), and soil water content (11.4%). The correlation analysis further demonstrated significant associations: ecosystem carbon storage showed strong negative correlation with soil bulk density (P < 0.01), while exhibiting a significant positive correlation with soil total nitrogen (P < 0.01) and plant density (P < 0.001) (Fig. 6).

Impact of ecological restoration on slope carbon storage

This study quantifies carbon storage patterns in damaged slopes of the Ussuri River Basin after 8 years of restoration. Our results reveal that ASS successfully reestablishes a carbon distribution profile analogous to undamaged reference slopes (Table 2), confirming its efficacy in restoring ecosystem carbon functions. The technique’s “tree, shrub, grass” plant configuration used in ASS can rapidly establish a stable plant community, (>5% vegetation coverage (Li et al., 2011)), with vegetation carbon dominated by trees (98%) consistent with natural forest ecosystems (Wang et al., 2011). Notably, tree layer carbon in ASS restored slopes matched undamaged slopes (1.53 vs. 1.59 kg/m², Table 2), confirming recovery of vegetative sequestration function. However, this remains about 50% below Heilongjiang’s forest average 3.3 kg/m² (Liu, Wang & Luan, 2011), attributable to the younger age and smaller DBH (Ren & Xia, 2017). Carbon accumulation is projected to increase with restoration duration (Shi et al., 2015; Ahirwal & Maiti, 2018), indicating significant future potential in plant carbon sequestration but require long-term monitoring.

This study found that the soil carbon storage of ASS restored slopes is 3.5 times that of planted forest restored slopes, confirming the effectiveness of ASS ecological restoration in increasing soil carbon storage (Xu, Qu & Wang, 2020). The reasons for this are related to the ASS restoration technology: firstly, the technology uses a substrate rich in organic matter to construct an artificial soil layer (Li et al., 2011), significantly increasing the organic carbon content in the soil; secondly, by establishing a high-density seed bank, ASS promotes the input of organic matter from seed residues, plant litter, and roots into the soil, further enhancing soil carbon storage (Wu et al., 2023); finally, soil erosion is a major factor in the loss of soil organic carbon (Liu et al., 2023), and the artificial soil structure constructed by ASS is stable and has strong erosion resistance (Li et al., 2011), effectively promoting the fixation of soil carbon. However, compared with undamaged slopes, the soil carbon storage of ASS restored slopes is still lower. This may be related to the low amount of forest litter return. The results of this study show that the litter carbon storage in the sprayed restoration area is only 41% of the carbon storage of undamaged slopes, consistent with the results of a study where natural forest land was converted to artificial forest (Ran, Xu & Wan, 2022). This indicates that the reduction in the amount of litter is a key factor leading to low soil organic carbon storage. In addition, another reason for the low soil carbon storage in aggregate spray seeding areas may be related to the thin artificial soil layer in the damaged slope area. Relevant studies have pointed out that in rocky desertification ecosystems, due to high rock exposure rates and thin soil, the soil carbon storage after restoration is far below the national average (Lu et al., 2019). A study by Lei et al. (2023) also found that the soil organic carbon pool in the spoil dump slopes of northern mining areas still has a lower content than the soil in natural areas after restoration.

Impact of ecological restoration on plants and soil

Vegetation restoration constitutes the primary objective in rehabilitating damaged slopes. This study revealed significantly higher plant density and Shannon-Wiener index at aggregate-spray restored slopes (S1) compared to planted forest restored slopes (S2) after 8 years and undamaged slopes (S3) (Table 3), indicating superior vegetation recovery of ASS method (Li et al., 2011). Soil water content and nutrient availability serve as key drivers of plant community composition, profoundly influencing species richness and diversity (Fehmi & Kong, 2012; Wu et al., 2014). Consistent with Bennie et al. (2008), we observed significantly elevated soil water content (SWC) and nutrient concentrations (TN, TP, TK) at S1 than that of the S2 and S3. These enhanced soil conditions likely underpin the observed differences in vegetation structure. The high-density seeding methodology intrinsic to aggregate spray seeding further contributed to these outcomes (Tamura et al., 2017). This approach facilitated rapid establishment of a multi-strata vegetation structure (trees, shrubs, and grasses) on bare slopes within 2 years. Plant density subsequently increased progressively during community succession from pioneer to late-stage species (Li et al., 2011).

Damaged mountain slopes, characterized by challenging site conditions such as shallow soils, water scarcity, and low fertility, present significant obstacles to ecological recovery (Hüttl & Frielinghaus, 1994; Xu et al., 2005; Egli & Poulenard, 2016; Moos et al., 2018; Hu et al., 2021; Wang et al., 2022; Shen et al., 2023), necessitates prioritizing soil quality improvement. Ecological restoration demonstrably counters soil degradation, enhances nutrient cycling, and sustains soil quality to facilitate ecosystem recovery (Wang et al., 2014; Li et al., 2016). Consistent with Morrison & Foster (2001), our findings indicate significantly higher levels of soil total nitrogen (TN), total phosphorus (TP), and total potassium (TK) at site S1 compared to sites S2 and S3. This nutrient enrichment primarily stems from enhanced litter inputs and root system dynamics. Critically, the deliberate incorporation of leguminous species at S1—noted for nitrogen-rich litter and roots—promoted soil aggregate formation and stability (Kumar et al., 2020). Enhanced aggregate stability significantly improves key soil physicochemical properties (Regelink et al., 2015), including reduced bulk density and increased water retention capacity (see Table 4). These improved soil physical properties subsequently enhance the stability of carbon and nitrogen pools (Ma et al., 2022). Coupled with rapid nutrient cycling rates characteristic of broad-leaved species, this significantly contributes to soil fertility enhancement (Shao et al., 2020). Furthermore, the artificial soil matrix employed in aggregate spray seeding technology, typically amended with organic matter, soil conditioners, and necessary slow-release fertilizers, provides an initial plant growth foundation. Subsequent vegetation recovery, driven by continuous accumulation of litter, root network development, and root exudation, further facilitates progressive improvement in soil physicochemical properties (Yang et al., 2022; Dan, 2023).

Analysis of influencing factors on ecosystem carbon storage

Ecological restoration effectively enhances ecosystem carbon storage through dual pathways: accelerated vegetation growth and promoted soil carbon sequestration (Li, Niu & Luo, 2012; Lange et al., 2015; Zheng et al., 2024). Plant density is an important variable influencing the forestry biomass accumulation (Marquard et al., 2009) by regulating resource competition and crown development (Postma et al., 2021). Contrary to self-thinning theory, which typically predicts negative density-biomass relationships due to competitive exclusion (Yoda, 1963; Ouyang et al., 2019; Hao et al., 2022), our study revealed a positive correlation between plant density and carbon storage (P < 0.001), explain 19% of the variation after 8-year restoration (Figs. 5 and 6). This apparent contradiction resolves when considering stand density context: at low densities (0.5 plants/m²), wider spacing reduces competition, enabling greater branch expansion and individual tree growth (Fanin et al., 2025). This maximizes per-unit biomass in late-successional stages. In addition, a relative high initial densities (e.t., via hydroseeding) enhances resource acquisition efficiency per unit area by promoting rapid canopy closure to capture more solar radiation, improving interception of water and nutrients before loss, suppressing weeds, and accelerating early biomass accumulation (Westgate et al., 1997; Tamura et al., 2017).

Our study found that carbon storage in aggregate-spray restored slopes is significantly negatively correlated with soil bulk density (P < 0.01), indicating that reducing soil bulk density promotes carbon accumulation on such slopes. This relationship arises because bulk density directly influences key soil properties governing carbon dynamics, including aeration, infiltration, water-holding capacity, and solute transport (Chai & He, 2016; Xu, He & Yu, 2016). Elevated bulk density can impede root development and plant growth, thereby reducing photosynthetic rates and aboveground productivity (Li & Wang, 2000; Liu & Shan, 2003). Consistent with this mechanism, site S1 exhibited a 24% lower bulk density than site S3 (Table 4), corresponding to more favorable vegetation conditions. This finding aligns with Chen et al. (2024), who demonstrated that aggregate-spray stabilization (ASS) effectively restructures soil on mining tailings ponds to achieve lower bulk density.