Introduction

Both in China and worldwide, rapid economic development is closely linked to the utilization of mineral resources. As a traditional secondary industry, mineral resource extraction is an important pillar of economic development. However, the expansion of mineral resource extraction can lead to a range of negative consequences, including severe environmental pollution and ecological damage1,2. Therefore, it is important to conduct ecological restoration in mining areas.

When conducting mine restoration, it is essential to identify the causes of ecosystem destruction, for example, mining activities, domestic pollution of workers, transportation, develop corresponding restorative measures, and monitor land disturbance and damage3. When carrying out mine restoration, exploring the restoration role of plants is to establish a stable plant community, utilize the role of dominant species, and then utilize the functions of plant communities4. Mine rehabilitation is a long-term process, and the duration of reclamation is the main factor controlling the changes in soil and vegetation properties. It is therefore essential to monitor mines at different times post-restoration to assess the effectiveness of restoration efforts and track the progress of ecological recovery5. Analyzed the factors affecting the rehabilitation of mines and assessed the capacity and value of ecological rehabilitation6. Compared the differences between restored mine areas and undisturbed areas and made a series of recommendations for mine restoration7.

Seed dormancy is an excellent biological characteristic, which is the adaptability of plants to resist adverse environmental conditions formed in the long-term phylogenetic process, which is conducive to the survival and reproduction of the race. Consider the changes in soil seed bank (SSB) and its relationship with aboveground vegetation (AVC) in the ecological restoration of mining areas, in order to facilitate a three-dimensional analysis of the ecological restoration effect. The seed reserves in soil play a vital role in the earth’s future biodiversity, serving as a potential source for vegetation regeneration and contributing to the establishment of new plant communities8,9,10. Grazing, in particular, can have significant impacts on the soil seed reserve, which can subsequently influence the characteristics of both plants and the soil itself11. –12.

The SSB has a role in maintaining biodiversity and can be used to study past vegetation conditions in arid desert ecosystems. It can be used to analyze how aboveground vegetation is conducive to desert management and exploring the changes of the SSB in these specific environments is conducive to the in-depth study of desert ecosystems13. Exploring ways to break the dormancy cycle of the desert SSB would significantly enhance the existing research on the SSB, such as the amount of precipitation, the level of temperature14.

Long- and short-term changes in grassland plant communities and the migration of forest vegetation can have an impact on the stability of the SSB and ecosystems, and the SSB is critical for the subsequent restoration of disturbed land15. There are two types of soil seed banks, persistent SSB and transient SSB. Changes in the relationship between SSB and AVC were correlated with changes in species similarity and importance16,17.

Correlations between the SSB and AVC can be determined through a range of indicators, including SSB species composition, density, and richness18. A comprehensive analysis of SSBs, comparing natural ecosystems with artificially reconstructed ecosystems, will allow for a better understanding of the dynamics of SSBs by integrating human and natural impacts, ultimately contributing to long-term ecosystem stability.

In the ecological restoration of mines, the focus is not only on changes in vegetation and soil properties, but also the potential vegetation resource in the SSB, because seed germination can replenish above-ground vegetation. Combining the instantaneous and permanent SSB of mines with vegetation and soil physicochemical properties for research is beneficial for long-term ecological restoration. The main purpose of above research is to remediate the mining damage to the environment and ensure that ecological restoration reaches a stable state. Such as the SSB of specific populations within mining waste areas can accelerate ecological restoration and its study will facilitate insights into mine population dynamics and biodiversity changes19. And reported the effects of seedling germination in the SSB after mining in arid regions, and found that seed longevity increased under arid conditions, while warmer and wetter conditions decreased seed viability20.

Xilinhot Shengli West No. 2 open-pit coal mine, located in the typical grassland area of China, possesses a unique vegetation belt that is very important to the natural environment and economy of northern China. Therefore, the objectives of this study were to: (1) investigate the changes in the SSB and AVC in a long-term reclaimed coal mine dump; (2) determine the relationship between the SSB and AVC in a coal mine dump at different times post-restoration; and (3) determine the impact pathways of soil properties and AVC on the SSB.

Methods

Study site

The study area was the Shengli West No. 2 open-pit coal mine in Xilinhot City, Inner Mongolia Autonomous Region (Fig. S1). The study area is located in the eastern part of Inner Mongolia Autonomous Region, with geographic coordinates of 43°57′N latitude and 115°57′E longitude. The average annual precipitation of this area is 210.4 mm and the average annual temperature is 7–8 °C, with a freezing period of up to 5 months and an annual cumulative temperature of 2,682 °C. Xilinhot City is located in the eastern Inner Mongolia Plateau, with an arid and semi-arid continental climate. The main species in the grassland area are Leymus chinensis and Stipa grandis. The soil types are black calcium and light chestnut calcium soils under a natural grassland, and open-pit coal mine dumping sites can provide valuable nutrient indicators for ecological assessments (Table S1). The economy of Xilinhot City is dominated by secondary industry, and its mineral resources contribute significantly to the economic development and employment of the region.

Field surveys

For ecological restoration, the Shengli West No. 2 open-pit coal mine dumping sites in Xilinhot City were covered with soil that was seeded with a mixture of Gramineae and Leguminosae species. The main species are Medicago sativa、Astragalus laxmannii、Caragana microphylla、Elymus nutans、Bromus inermis、Melilotus officinalis. The growth and reproduction characteristics of the species mixture included seed and asexual reproduction. The “space-for-time substitution” method was applied21, and four dumping sites were selected that had received the same restoration measures with different post-restoration time intervals (2, 4, 7, and 17 years). The surrounding undamaged natural grassland was used as the control (CK), and a total of five sample plots were selected. One sampling point was established 50 m from each sample site, and 5 sampling points were established at each sample site; hence, a total of 25 sample points were selected (Table S2) .

The SSB and AVC surveys and SSB sampling began in July 2023. The AVC survey recorded the full range of plant species, cover, density, and height in the 1 × 1 m sample plots, a total of 25 in five plots (Table S2). The SSB was sampled close to the location of the AVC survey to ensure that the vegetation corresponds to the seed bank, sample size as above. In the vegetation sample plots, 20 × 20 cm plots were selected for soil sampling using a soil auger. Surface soil (0–10 cm) was extracted using an earth auger, the soil samples were sieved through a 2 mm sieve to remove stones and plant roots, and stored in labeled self-sealing bags in the dark.

Soil samples

The soil samples were tested for various soil indicators. Soil organic matter (SOM) was determined using a K2Cr2O7 oxidation and FeSO4 titration22. Available nitrogen (AN) was determined by NaOH hydrolysis23. Available phosphorus (AP) was determined by NaHCO3 extraction22. Soil salt content (SSC) was determined by evaporation of the filtrate after shaking (soil/water ratio = 1:5;)24. Available potassium (AK) was determined by NaHCO3 extraction22. Bulk density (BD) and soil water content (SWC) were determined by drying at 10 °C for 24 h.

Seedling emergence experiment

In this experiment, germination test was used to identify the density and species of soil seed bank, which can shorten the determination time and make the work simple. Storing the finished soil samples away from light for a period of time helped to seed dormancy. The the seedling emergence experiment was undertaken in located on an open terrace within the Inner Mongolia University. The field site was a sunny and stable location with flat ground, good drainage, and natural light. The soil samples from the 25 sample points were spread out in a sprouting tray (28 × 28 × 5 cm). The thickness of the soil was 0.75–1 cm, and the soil was watered regularly every day to keep it moist, and to ensure that the germination conditions (i.e., temperature, water, light) in the SSB were the same. Cover with insulating film to prevent seed contamination. The experiment was conducted for eight weeks, after which few new seedlings emerged. The soil samples were placed in sunlight with sufficient air circulation, and the species and number of non-sprouting plants was recorded. Every three days, the types and numbers of sprouted plants were recorded (unrecognizable plants were transplanted to other germination trays or grown in the original sprouting tray until they could be recognized).

Statistical analysis

To estimate the seed density, the number of sprouted plants in each germination tray was scaled up in equal proportions.

SSB density: the amount of seedlings sprouted on the sprouting tray was expanded in equal proportions according to the size of the sampling area to obtain the SSB density.

A one-way analysis of variance (ANOVA) was conducted, with Duncan and least significant difference (LSD) tests (P < 0.05) used to make comparisons among the different post-restoration time intervals and determine the significance of the changes in SSB density and α-diversity25. A normality test was conducted to test the normality of the relevant data, and an outlier test was conducted to facilitate the selection of subsequent analysis methods. Shapiro-Wilk test was selected for normality test and descriptive statistics retrieval was selected for outlier test.

A Pearson’s correlation analysis was conducted to explore the relationship between the variables, including the number of species, density, cover, height, and α-diversity of the AVC; and the number of species, density, and α-diversity of the SSB. The relationship between the vegetation and SSB characteristics was analyzed in combination with soil properties (SOM, AN, AP, SSC, AK, SWC, and BD). The use of structural equations can better reflect the relationship between soil physicochemistry and above-ground vegetation and soil seed bank after restoration, and can reflect the positive and negative relationship between the corresponding indicators, which is meaningful for the adjustment of the later ecological restoration. Based on26, the number and density of SSB and AVC species were selected for analysis, and then the AN, SSC, and SWC of soil were determined and the predictive variables of SSB vegetation in the soil were determined. Finally, a structural equation model (SEM) was constructed. Path prediction for indicators for which structural equation analysis was performed (Fig. S2). The IBM SPSS Amos 26 Graphics package was used to fit the path model and its structural relationship with the data.

A Bray–Curtis-based non-metric multidimensional scaling (NMDS) was used to analyze the similarities and differences in the species composition and density of the SSB and AVC as the time post-restoration increased. The data were standardized prior to the analysis27,28. A factor analysis was performed on the indicators related to the AVC and SSB using a principal component analysis (PCA), and the composite scores of the SSB and AVC were derived by dimensionality reduction as the number of years post-restoration increased29.

With IBM SPSS Statistics 25 used for data analysis, Origin 2021 used for data visualization, and ArcGIS 10.6 used for the graphical plotting of the study area profile.

Results

Changes in AVC and SSB characteristics with increasing time post-restoration

In the vegetation of the mine and natural grassland, there were 11 families and 33 plant species documented (Table 1, S3). The density of the AVC was 94 − 192.8 plant m− 2 (Fig. 1a). In terms of changes in species’ importance values, the restored 17-year old drainfield was similar to the natural grassland (Table S4; Fig. 1c). The highest density and cover of vegetation in the 7 years of restoration was 192.8 plant m− 2 and 64.66%, respectively, and the highest overall height of vegetation was 199.24 cm m− 2 in natural grassland. (Table S5).

Table 1 Species composition and quantity of different families in the AVC.
Full size table
Fig. 1
Fig. 1
Full size image

(a) AVC density. (b) SSB density. (c) Importance values for different life types of vegetation communities. The density of the AVC and SSB in the open-pit coal mine dumping sites in different years post-restoration and the density of the natural grassland AVC and SSB in plants m− 2. The pink at the top represents shrubs, the dark blue in the center represents perennials, and the light blue at the bottom represents annuals and biennials. Different letters in the graph represent significant differences for the same index, P < 0.05. 2a represents restoration of a 2-year-old dump, 4a represents restoration of a 4-year-old dump, 7a represents restoration of a 7-year-old dump, 17a represents restoration of a 17-year-old dump, and CK represents natural grassland.

In both the mine site’s SSB and the natural grassland, 10 families and 34 plant species were identified (Table 2, S6). The SSB density was in the range of 475.00 − 705.00 (plant m− 2), with the maximum density in the natural grassland and the lowest density in the restored area 2 years post-restoration (Fig. 1b). Leymus chinensis and S. grandis were the most abundant species in natural grassland, while Lespedeza daurica, Heteropappus altaicus, and Thalictrum petaloideum were present in the natural grassland, but had not yet appeared in the SSB of the open-pit coal mine dumping sites.

Table 2 Species composition and quantity of different families in the SSB.
Full size table

As the restoration period increased, the density of Gramineae plants in the AVC and SSB gradually increased and approached that of natural grassland. The density of Leguminosae plants in the AVC and SSB initially increased and then decreased. The density of S. collina in the SSB gradually decreased. The density of S. collina in the AVC initially increased and then decreased (Fig. 2a, b). In the AVC, the proportion of perennial plants in the area restored for 17 years and natural grassland was significantly higher than that in the area restored for 2–4 years. The proportion of annual and biannual plants in the area restored for 2–4 years was significantly higher than that in the area restored for 17 years and natural grassland (Fig. 3a, b).

Fig. 2
Fig. 2
Full size image

(a) The trends of plant changes in different vegetation communities. (b) The changing trends of different plants in the SSB. Gramineae, Leguminosae and S. collina were selected for a comparison of trends in the different years post-restoration. Uppercase letters denote differences between groups and lowercase letters indicate differences within a group, P < 0.05. 2a represents restoration of a 2-year-old dump, 4a represents restoration of a 4-year-old dump, 7a represents restoration of a 7-year-old dump, 17a represents restoration of a 17-year-old dump, and CK represents natural grassland.

Fig. 3
Fig. 3
Full size image

(a) The proportions of the perennial, annual, and biennial plant density in the AVC. (b) The proportions of the perennial, annual, and biennial plant density in the SSB. Different letters in the graph represent significant differences for the same index, P < 0.05. 2a represents restoration of a 2-year-old dump, 4a represents restoration of a 4-year-old dump, 7a represents restoration of a 7-year-old dump, 17a represents restoration of a 17-year-old dump, and CK represents natural grassland.

The α diversity index of the AVC was higher in natural grassland than in the restored areas, and the Margalef index of natural grassland was significantly larger than that of the dumping sites (Fig. 4a). The Margalef index results revealed that the SSB was significantly larger in natural grassland than in the areas 2, 4, and 17 years post-restoration (Fig. 4b).

Fig. 4
Fig. 4
Full size image

(a) AVC α-diversity. (b) SSB α-diversity. (c) Similarity of SSB and AVC at different years of restoration. The α-diversity of the AVC and SSB in the open-pit coal mine dumping sites in different years post-restoration and the natural grassland AVC and SSB. The Jaccard similarity score of the SSB and AVC in the open-pit coal mine dumping sites in different years post-restoration. Different letters in the graph represent significant differences for the same index, P < 0.05. 2a represents restoration of a 2-year-old dump, 4a represents restoration of a 4-year-old dump, 7a represents restoration of a 7-year-old dump, 17a represents restoration of a 17-year-old dump, and CK represents natural grassland.

The relationship between the AVC and SSB

As the restoration period lengthened, the resemblance between the SSB and AVC slowly diminished. The SSB of the area 17 years post-restoration was less similar to that of the aboveground vegetation than that of the natural grassland. The overall similarity of the SSB and AVC in the restored area shifted from medium to low over the 17 year period post-restoration (Fig. 4c).

The correlations between soil indicators, AVC, and SSB

The similarity coefficients, indicating the extent of correlation between the indicators of the SSB and AVC, were generally low, with only moderate and low correlations observed. The SOM and AP were significantly positively correlated with the number of species in the SSB. The BD had a negative correlation with vegetation height, while SWC had a negative correlation with vegetation density (Fig. 5a). The SEM indicated that the number of species in the SSB increased with the increase in AVC species, and the density of the SSB increased with the increase in the number of SSB species. The number of species in the AVC decreased with the increase in SSC and increased with the increase in AN. The density of the AVC increased with the growth of SSC and decreased with the growth of SWC (Fig. 5b).

Fig. 5
Fig. 5
Full size image

(a) Correlation between the vegetation and SSB indicators. (b) The SEM of SSB, AVC and soil properties. S is the SSB and V is the AVC. Red symbolizes positive correlations, while blue denotes negative correlations, with the change from lighter to darker colors representing a gradual strengthening. The indicators are the number of species, density, and α-diversity for the SSB; and number of species, density, cover, height, and α-diversity for the AVC; and SOM, AN, AP, SSC, AK, SWC, and BD. The variables were selected as follows: SSB select species number and density, AVC select species number and density, soil physicochemical properties select AN, SSC, and SWC. The thickness of the line represents the strength of the relationship, while the dashed line indicates no significant difference (* P < 0.05, * * P < 0.01, * * * P < 0.001).

The similarity between the SSB and AVC in different years post-restoration

There was a high degree of similarity between the areas 2, 4, and 7 years post-restoration. Among the different time periods post-restoration, the AVC 17 years post-restoration had the highest degree of similarity with the AVC in natural grassland (Fig. 6a). The SSB in the natural grassland was found to be unlike the SSB 2, 4, and 7 years post-restoration. The SSB in the natural grassland displayed some similarity with the SSB 17 years post-restoration. (Fig. 6b).

Fig. 6
Fig. 6
Full size image

(a) The Bray − Curtis based NMDS for the AVC of the restored area and natural grassland. (b) The Bray − Curtis based NMDS for the SSB of the restored area and natural grassland. The analysis of similarity was based on species composition and density. The shading from light to dark indicated the time post-restoration (from 2 to 17 years). The orange color is the natural grassland, and points that are close together on the visualization are more similar to each other, while points that are far apart are more dissimilar. 2a represents restoration of a 2-year-old dump, 4a represents restoration of a 4-year-old dump, 7a represents restoration of a 7-year-old dump, 17a represents restoration of a 17-year-old dump, and CK represents natural grassland.

Comprehensive evaluation of the SSB and AVC

The PCA of the correlation indexes of vegetation communities yielded results with a cumulative contribution of 83.3% (Table 3; Fig. S3a, b). The PCA of the correlation indexes of the SSB yielded results with a cumulative contribution of 91.9% (Table 3; Fig. S3c), i.e., the principal components extracted from the data accounted for a substantial portion of the total variation observed in the correlation indexes.

Table 3 The SSB and AVC composition matrix.
Full size table

The comprehensive scores of each index of the SSB followed the order of natural > 4 a > 17 a > 7 a > 2 a, indicating a fluctuating pattern (Table 4). The comprehensive scores of each index of the AVC followed the order of natural > 7 a > 4 a > 17 a > 2 a, indicating an increasing and then decreasing trend (Table 5).

Table 4 Comprehensive scores for the SSB.
Full size table
Table 5 Comprehensive scores for the AVC.
Full size table

Discussion

The SSB bias may affect the species in the AVC in different locations or time periods18. In this study the trend of change for character such as overall density, coverage, and average height was also an initial increase followed by a decrease, which may be related to the gradual disappearance of replanted species as the time post-restoration increased, together with the entry of natural grassland species into the mine site, although not yet on a large scale. Previous studies have found that artificial restoration can temporarily increase vegetation density, while environmental improvements (such as vegetation growth, soil nutrients) increase the proportional cover of the dominant species30. Perennial species were more abundant in native grasslands, and annual and biannual species were more abundant in restored areas after disturbance, because the disturbance reduced the number of perennial species31. The density of perennial herbs increased after artificial restoration and gradually resembled that of natural grassland32. This would explain the fact that weeds have become fewer 17 years after the restoration in this study, although the vegetation density is lower.

Previous studies have shown that SSB species number and density increase as environmental conditions become more suitable33. Some studies have shown that persistent SSB species and density increase and then decrease with increasing disturbance16. This is similar to the results of this study. Weed species are abundant on disturbed lands, but their populations are unstable and tend to decline over time34. Often, some dominant herbaceous species will increase in abundance over time post-restoration35. This fits with this study’s decrease in weeds and increase in dominant perennial species as the number of years of restoration increases. It was previously found that a plantation forest did not have many species among the natural community a short time after planting36. In the SSB and AVC, the density of important species such as Ajuga laxmanii and M. sativa, which were sown during restoration, displayed an increasing and then decreasing trend. However, there were also some species, such as L. chinensis, whose density displayed an increasing trend, which may be because A. laxmanii and M. sativa are generally not present in natural grasslands. Species such as Leymus chinensis is dominant in natural grasslands and will gradually enter the restored grassland. It relies on asexual reproduction, resulting in an increasing number of individuals over time. Ajuga laxmanii and M. sativa primarily rely on seed reproduction, which can lead to a decrease in density in the later stages of the SSB.

Vegetation recovery after disturbance has been reported to follow a directional pattern, with the same or similar trends in species composition, density, and diversity37. This is similar to the trends of SSB and AVC diversity observed in the present study. Previous studies have shown that restoration sites contain many annual and biannual plants, and diversity declines approximately 7 years post-restoration38. Other studies have shown that the species diversity decreases as the time post-restoration increases27. The diversity of the native SSB decreases as invasive species enter the area39. The high index values for species diversity in natural grasslands in this study were similar to the results of previous studies that found significant differences between natural and restored areas40. Found that the species richness of natural wetlands is significantly higher than that of agricultural land, supporting the notion that natural areas tend to exhibit higher species diversity compared to agricultural land41. Earlier studies found that the SSB has a temporal variability that may affect its diversity and persistence42. In the present study, the diversity index was higher in areas 4 − 7 years post-restoration, which may be related to their initial artificial maintenance, watering, and fertilization, while areas that had been restored for a longer period of time did not receive this maintenance, resulting in a trend of decreasing species diversity.

Factors such as topography and soil can affect the species composition of the SSB and AVC43. This could explain the low similarity found in this study between the SSB and AVC, i.e., a moderately similar or low similarity relationship between the SSB and AVC. Some studies have shown that the high spatial and temporal heterogeneity of the SSB in semi-arid regions can result in low similarity values44. The influence of the SSB in the early stages of vegetation recovery makes the similarity higher at this initial stage45. The similarity between the AVC and SSB is often considered low, which may be related to species accumulation rates and sampling time46. Long-term studies have shown that the species composition of the SSB and vegetation in the study area differ significantly over time47,48. The reasons for the decrease in similarity in this study may be related to the accumulation of seeds over time, heterogeneity of soil conditions, burial depth, and weakened connection with the AVC.

The correlations between the number of species, density, diversity, and other indicators of the SSB and AVC were weak, and the indicators with significant correlations were generally of medium or low magnitude, which was consistent with the SSB and AVC having a low similarity and weak correlation8. Reports indicate notable links between the physical and chemical characteristics of soil and the prevalence of the AVC49, but no study has identified a significant relationship between soil nutrients and the SSB50. The significant differences in the vegetation between undisturbed and damaged areas also confirmed the results of this study51. The combined PCA scores for the SSB and aboveground vegetation were highest for natural grassland. Recent studies have found that over time easily germinated seeds tend to become depleted in the SSB, while dormant seeds become more prevalent compared to the aboveground vegetation47. Seeds in the SSB may have a dormancy period, and declines in soil physicochemical properties and a prolongation of the dormancy period can affect the adaptability and dynamics of the AVC52. The prolonged seed dormancy and failure to germinate in a timely manner resulted in a slight decline in the number, density, and diversity of species in the area 17 years after restoration, which could also have had an effect on the decline in the AVC index scores.

Conclusion

As the time post-restoration increased, the number of dominant species in the SSB and AVC gradually increased. The dominant species of natural grasslands are increasing year by year in the mining area’s dumping ground in the SSB, indicating that the SSB has great potential in the subsequent restoration, and can serve as a potential resource for the AVC, as well as maintaining the stability of ecosystem restoration. The number of annual and biannual species decreased, and the proportion of perennial biological species gradually increased. The number of Gramineae species in SSB and AVC increased with the increase in time post-restoration. As the number of natural grassland species increased in the restored area, it was predicted that species diversity would gradually stabilize and display an upward trend. The relationship between soil physicochemical properties and SSB was relatively weak, while the relationship with AVC was relatively strong. As the restoration period increased, the similarity between the SSB and the AVC gradually decreased, and the degree of influence of the AVC on the SSB decreased. Therefore, in land restoration practices, replanting should occur after a specific restoration period to ensure the stability and diversity of species in the interval between the artificial restoration of open-pit coal mine dumping sites and natural grassland succession.