Revegetation pattern affecting accumulation of organic carbon and total nitrogen in reclaimed mine soils

Study area

The study was conducted in Heidaigou surface mine within the Junggar Banner in Erdos on the Loess Plateau, China (39°43′–39°49′N and 111°13′–111°20′E; Fig. 1), and was approved by the land environmental protection office of Shenhua Group Zhungeer Energy Co. (project number: KZCX2-XB3-13-02). The mine encompasses a total area of 52.11 km², and the mine elevation is between 1025 m and 1302 m above sea level. The site has a temperate continental arid climate with a mean annual temperature of 7.2 °C. The mean annual precipitation is 404.1 mm (ranging from 213.0 to 459.5 mm), with 60–70% received during the growing season between July and September. The average annual evaporation is 1943.6 mm and the relative humidity is 58%. The wind is mostly calm, blowing in a north–northwest direction at an average speed of 2.2 m s⁻¹. The predominant soil series prior to mining was mainly loess soil (Calcaric Regosol, FAO/UNESCO, 1988).

The mine has six overburden dumps, and the northern dump was selected for this study (Fig. 1). Since 1993, large-scale reclamation and revegetation activities have been implemented at this dump for the purpose of soil ripening, and by 2005, the reclamation area had reached 1.48 km². In the process of reclamation, at least 1 m of top subsoil (the deep loess parent materials stripped during mining) was applied on the top of the dump and appropriate leveling was conducted to create a flat surface. Four predominant artificial revegetation patterns were selected and compared to the natural recovery pattern to evaluate their effects on SOC and TN accumulation: the planting of arbors (Ar), bushes (Bu), arbor–bush mixtures (AB), and grasslands (Gr). This study was not a replicated field plot experiment or an established design. Therefore, the vegetation types comprising Ar, Bu, AB, and Gr were established at four, four, two, and three typical vegetation collocations, respectively, as pseudo-replications due to the lack of true replicates, with a total of 13 artificially revegetated sites (ARSs). Revegetation at these selected ARSs was implemented between 2002 and 2005. The vegetation was sparse and it only included scattered Leymus chinensis (Trin.) Tzvel. and Leymus secalinus (Georgi) Tzvel., so only one experimental site was established for the natural recovery approach and it was designated as the natural recovery site (NRS). The NRS was not planted and it had been unmanaged since 2004. All of the selected RMSs had similar soils and terrain conditions, which provided an excellent opportunity to compare the different revegetation patterns. The UNSs located within 2 km of the reclaimed dump and without significant disturbance due to mining activities were included in the study as reference sites. Historically, the UNSs were mainly sloping farmlands, which were later converted into forests or grasslands due to the implementation of the state-funded “Grain-for-Green project” in 1999. Three Bu sites, two AB sites, and three Gr sites with vegetation type changes since approximately 2001 were selected for this study. In total, 22 experimental sites were established in this study (Fig. 1). Global Positioning System coordinates (with a resolution of 3 m) were recorded for each sampling location, and the details of these sites are shown in Table 1.

Soil sampling and analysis

At each RMS and UNS, disturbed soil samples were collected at depths of 0–10, 10–20, 20–40, 40–60, and 60–100 cm from five randomly selected locations using a 20 cm by 5 cm soil auger in July 2015. The selected replicate locations were at least 20 m from the boundary of the site (from the area where the vegetation type changed) and they were separated by a distance of approximately 10 m to account for spatial variability. Soil samples were collected from Ar, Bu, and AB sites near the center of the inter-tree space. In total, 550 soil samples were collected. The soil samples were air-dried, crushed, and passed through 0.25–mm sieves before performing SOC and TN measurements. The SOC concentration (g kg⁻¹) was measured using the Walkley–Black method (Nelson & Sommers, 1982), and the TN concentration (g kg⁻¹) was determined using the semi-macro Kjeldahl method (Bremner & Mulvaney, 1982).

One 1 m deep pit was dug at one typical site of each vegetation type in the RMS and UNS, and core samples of undisturbed soil (100 cm³) were collected at depths of 0–10, 10–20, 20–40, 40–60, and 60–100 cm. Three replicate measurements were conducted in each layer. The soils were transported to the laboratory and dried to a constant weight at 105 °C to calculate the bulk density (g cm⁻³). The SOC stock (SOC_i) and TN stock (TN_i) (kg m⁻²) were then calculated using the following formula: ${\displaystyle {\mathrm{SOC}}_s=\sum \frac{{\mathrm{SOC}}_i\times {\mathrm{BD}}_i\times {\mathrm{d}}_i}{100}}$ ${\displaystyle {\mathrm{TN}}_s=\sum \frac{{\mathrm{TN}}_i\times {\mathrm{BD}}_i\times {\mathrm{d}}_i}{100}}$

where SOC_i, TN_i, BD_i, and d_i represent the SOC concentration (g kg⁻¹), TN concentration (g kg⁻¹), bulk density (g cm⁻³), and thickness (cm) of the i th layer, respectively.

At each RMS and UNS, aboveground leaf litter was also collected from five randomly selected quadrants measuring 1 m × 1 m. The leaf litter was placed in mesh bags and transported to the laboratory to determine the dry mass (g m⁻²) by weighing the litter after oven drying at 60 °C for 48 h.

Statistical analysis

All data were tested for normality and homogeneity of variance. One-way analysis of variance (ANOVA) and least significant difference (LSD) tests (p < 0.05) were used to detect differences in the concentrations and stocks of SOC and TN associated with different revegetation patterns and soil depths. Few typical vegetation collocations were available for each land and the NRS was not field replicated, so the five sampling locations within each site were also used as pseudo-replicates for the statistical analysis. The independent sample t-test was applied to compare the concentrations and stocks of SOC and TN between RMSs and UNSs, and only the soil values obtained at the Bu, AB, and Gr sites were used. All statistical analyses were performed using SPSS 16.0 software.

SOC and TN concentrations under different revegetation patterns

The concentrations of SOC and TN in the soil profile are shown in Table 2. The SOC concentration was strongly stratified based on soil depth for all of the revegetation patterns. The SOC concentrations were highest in the top depth (0–10 cm) and decreased as the soil depth increased, although a slight increase was observed at a depth of 60–100 cm for Bu and AB. Significant decreases occurred in the 0–40 cm depth for Gr and the 0–20 cm depth for the other four revegetation patterns. The SOC concentration did not differ significantly below 40 cm or 20 cm. As shown in Table 2, the revegetation pattern had a significant effect on the SOC concentration (p < 0.05). At a depth of 0–40 cm, the average SOC concentration ranged from 3.54 g kg⁻¹ for Gr to 1.50 g kg⁻¹ for NRS and it was ranked in the following order: Gr > Bu > AB > Ar > NRS. The average SOC concentrations (0–40 cm) for Gr, Bu, AB and Ar were approximately 136.5%, 74.4%, 55.9%, and 21.5% greater, respectively, than that for the NRS. Except for AB at depths of 10–20 and 20–40 cm and Ar, the SOC concentrations in the ARSs were significantly higher than those in the NRS. Among the ARSs, Gr contained significantly higher SOC concentrations than all the other revegetation patterns at all depths. The SOC concentration in Ar was significantly lower than that in Bu at all depths but lower than that in AB at a depth of 0–10 cm. No significant differences were observed between Bu and AB at any depth. Below 40 cm, the differences in the SOC concentration were smaller among the revegetation patterns. In addition, no significant differences were found between the ARSs and the NRS, except for Gr. Gr had the highest SOC concentration among the revegetation sites, it did not differ significantly from those in AB and Bu. In addition, no differences were observed between Ar, Bu, and AB.

The soil TN concentration also varied significantly according to the soil depth and revegetation pattern (p < 0.05, Table 2). The TN concentration exhibited a significant decreasing trend from 0–40 cm for Ar, Bu, and Gr and from 0–20 cm for AB and the NRS. The soil TN concentration in the 0–40 cm interval exhibited a similar trend to that of SOC and it was ranked in the following order: Gr > Bu > AB > Ar > NRS. The TN concentrations in the ARSs with Bu and Gr at all depths and with AB at a depth of 0–10 cm were significantly greater than those in the NRS, which was consistent with the SOC results. However, unlike SOC, the highest soil TN concentration in Gr was observed in the 0–20 cm depth. The differences among the revegetation patterns decreased below 40 cm and no differences were observed below 60 cm.

SOC and TN stocks under different revegetation patterns

The average SOC stock varied from 0.52–1.17 kg m⁻² in the top 0–20 cm of the soil and from 1.80 to 3.78 kg m⁻² in the entire soil profile (0–100 cm) with the different revegetation patterns (Fig. 2). The SOC stock in the 0–20 cm depth accounted for 26.4–32.8% of the total SOC stock in the 0–100 cm depth. Except for those in Ar, the SOC stocks in the ARSs were significantly greater than those in the NRS in both the 0–20 cm and 0–100 cm depths. The total SOC stocks at 100 cm with Bu, AB, and Gr were 51.4%, 59.9%, and 109.9% higher, respectively, compared with that in the NRS. Gr had the highest SOC stocks, which were 38.7% and 31.3% higher than those for Bu and AB, respectively, at a depth of 0–100 cm. However, the SOC stocks in Bu and AB were not significantly different.

The average soil TN stocks varied between 0.06 and 0.11 kg m⁻² in the 0–20 cm interval and between 0.22 and 0.34 kg m⁻² in the 0–100 cm interval (Fig. 2). The TN stock in the 0–20 cm depth accounted for 26.5–31.7% of the total TN stock in the 0–100 cm depth. Similar to the SOC stocks, the TN stocks were higher in the ARSs than those in the NRS, excluding those in Ar. Gr had significantly higher TN stocks compared with Bu and AB, but the TN stocks did not differ between Bu and AB.

Comparison of SOC and TN concentrations and stocks in RMSs and UNSs

As shown in Figs. 3 and 4, there were significant differences in the SOC and TN concentrations and stocks between the RMSs and UNSs (p < 0.05). The average SOC concentration in the 0–100 cm interval was lower in the RMSs than that in the UNSs , i.e., 24.8% lower for Bu, 30.5% lower for AB, and 31.0% lower for Gr. However, the differences in the SOC concentrations between the RMSs and UNSs varied among the vegetation types and soil depths, but a significant difference was only observed for AB and Gr, and it was mainly in the 0–20 cm soil interval. For 0–20 cm, the RMSs contained significantly lower SOC stocks than the UNSs for AB and Gr, but the SOC stocks were similar to those in the UNSs for Bu. However, in the 0–100 cm interval, there were no significant differences in the SOC stocks between the RMSs and UNSs for all vegetation types.

The mean TN concentration in the 0–100 cm soil interval was lower in the RMSs than the UNSs, i.e., 29.4% lower for Bu, 27.8% lower for AB, and 43.1% lower for Gr (Fig. 3). In particular, the RMSs had significantly lower TN concentrations than the UNSs up to a depth of 100 cm for Bu and Gr and up to a depth of 40 cm for AB. In contrast to the SOC, the TN stocks in the RMSs were significantly lower than those in the UNSs in both the 0–20 cm and 0–100 cm intervals (Fig. 4).

Comparison of aboveground plant litter in RMSs and UNSs

There were significant differences in the amounts of aboveground plant litter between the RMSs and UNSs (p < 0.05). The amounts in the ARSs were significantly higher than those in the UNSs for most cases, whereas the amounts in the NRS were significantly lower than those in the UNSs. Among the ARSs, the woodlands, i.e., A, B, and AB, had significantly higher amounts of aboveground plant litter than the grassland, i.e., G. However, no significant differences were found among the A, B, and AB.