Clean water irrigation promotes microbial community recovery in acid mine drainage-contaminated paddy soil: a spatiotemporal analysis based on simulated soil column experiments from Dabaoshan mine, China

Abstract

Microbial communities serve as critical bioindicators and functional drivers of soil restoration processes, particularly in mining-impacted ecosystems undergoing remediation. However, systematic insights into microbial dynamics during clean water restoration of contaminated paddy soils remain limited. This study systematically investigated, by means of column experiments, the temporo-spatial dynamics of microbial community structure and metal speciation in acid mine drainage (AMD)-contaminated paddy soil from the Dabaoshan mining area. The soil was subjected to constant flooding with clean water, including experiments with artificial AMD as a control, for 176 days. The heavy metal fractions present in the soil were determined by sequential extraction. The bacterial community was analyzed at 7 time points and 5 depths using high-throughput 16S rRNA gene amplicon sequencing of the V5–V7 region. Long-term flooding increased the dominance of Firmicutes, Acidobacteria, and Proteobacteria, with limited overlap in significantly enriched taxa during restoration, indicating specialized microbial adaptation or microbial selection. The metal mobility increased as a result of flooding, most strongly in the mobile fractions of Cd at 5 cm depth (FM increased from 62.6% to 68.7%) and Cu at 20 cm depth (FM increased from 16.2% to 21.6%). This was accompanied by a substantial reduction in the residual total reducible-phase Cu (the sum of Fe/Mn oxide-bound fraction F3 and the organic-matter-bound fraction F4) was reduced from 188.4 to 30.8 mg kg⁻¹. Likewise, residual easily migratable Cd (the sum of exchangeable fraction F1 and carbonate-bound fraction F2) was reduced from 5.8 to 0.3 mg kg⁻¹. Such increased mobility might present an increased environmental risk. Canonical correspondence analysis identified the pH, Cu/Cd concentrations, and SO₄²⁻ as primary environmental drivers (cumulative explanation: 72.3%) governing microbial community restructuring. Complementary LEfSe analysis further elucidated potential microbial interaction networks underlying the rehabilitation process. The identified microbial-metal dynamics highlight the importance of integrating biological indicators with geochemical parameters when assessing the rehabilitation efficacy in heavy metal-contaminated agricultural systems.

---

Environmental significance

This study addresses the critical environmental challenge of heavy metal contamination in paddy soils. By quantifying the long-term restoring effect of microbial communities under clean water irrigation, this research provides actionable insights to mitigate adverse effects on soil health. The results highlight that the pH and Cu, Cd, and SO₄²⁻ content are the major factors that shift the distribution of the microbial communities, which directly inform policymakers and stakeholders in designing sustainable management practices. Furthermore, this work bridges the knowledge gap between soil restoration and microbial communities; our findings provide the necessary information.

1 Introduction

Mining industries frequently cause environmental harm due to the release of acid mine drainage (AMD), which poses a serious threat to local soil ecosystems due to its acidity and metal toxicity.^1,2 Contamination of paddy soils with AMD has become a local but serious environmental problem worldwide.³ Soil microbial communities have important ecological functions, including an essential role in nutrient cycling, which directly affects soil ecosystems and plays a key role in maintaining soil fertility; as such, those communities are the main drivers of ecological processes, and their composition is an important indicator of soil tolerance and resilience.^4,5 Environmental remediation of AMD-contaminated soils is a realistic approach to mitigate biodiversity loss and to restore ecosystem functions.^6,7

The spatial gradient characteristics of soil microbial communities provide useful information to determine the restoration stage of contaminated soils, providing valuable feedback to monitor and improve mine restoration practices.^8,9 Changes in ecological quality caused by environmental disturbances or improvements are reflected by rapid shifts in the local microbial soil communities.^10,11 This rapid response of microbes to changes in environmental conditions can provide valuable information and is an early indicator of progress in ecosystem restoration.

The Dabaoshan mine (24°31′37″N; 113°42′49″E) in northern Guangdong Province, China, has been in operation since 1970, with large, long-term AMD discharges into the Hengshi River downstream of a local dam. Due to the lack of alternative irrigation sources, water from the Hengshi River has long been used as irrigation water for the paddy soil that is used for agriculture, causing extreme environmental problems. In 2016, a large-scale sewage treatment plant was built to treat the AMD, which effectively increased the pH of the irrigation water.^12,13 The quality of irrigation water to a large extent affects soil microbial growth and biomass levels, microbial composition and biogeochemical cycles, and is a determinant of soil microbial communities.^14–16 Changes in irrigation water will significantly change the soil properties of AMD-contaminated paddy soil, which in turn promotes the formation of shifted soil microbial communities.^17,18 The microbial communities of AMD-contaminated soil have been investigated in numerous studies by us and other groups.^3,10,13,17 However, these studies focused on AMD-affected water or sediments, whereas systematic insights into microbial dynamics during clean water restoration in paddy soils remain limited. How restoration by means of clean irrigation water affects the soil microbial ecosystem in paddy soil with different contamination gradients remains to be determined.

In this study, we investigated changes in the soil microbial community composition, diversity, and interactions resulting from restoration of AMD-contaminated paddy soil by application of clean water irrigation, which was simulated along a spatial (vertical) gradient by laboratory column experiments. The paddy soil subjected to continuous irrigation with AMD was used as a control. Our aims were to (i) study the spatio-temporal changes in the microbial community abundance, richness and taxonomic composition during soil restoration; (ii) identify the main influencing factors of microbial community composition during this process; (iii) determine how the heavy metal content of the soil affected the microbiome composition during continuous AMD irrigation and during soil restoration. This study not only resulted in a comprehensive understanding of the correlation between water quality and bacterial communities of an irrigation ecosystem but also provided theoretical support for ecological restoration and governmental regulation of contaminated ecosystems.

2 Materials and methods

2.1 Site description and soil collection

This study was conducted on soil collected near the Dabaoshan mine, which is the largest mine in Southern China. It is situated in a humid subtropical monsoon climate zone; this area has been described in detail in previous studies.^19,20 The soil is typical chalky sandy loam that is used for rice production. The soil in the area is contaminated with heavy metals, as a result of multiple years of AMD flooding. In August 2019, soil was collected at a depth of 0–20 cm from Shangba Village (24°27′48″N, 113°48′35″E), which is located in the Hengshi River Basin, where contamination is so serious that the village is locally known as “Cancer Village”. The physicochemical properties of the collected soil have been described before.¹⁹ The prolonged AMD irrigation has led to a seriously acidified soil (pH: 4.30 ± 0.42) with a low cation exchange capacity (8.39 ± 1.37 cmol kg⁻¹). The soil Cu content was 290.25 ± 2.76 mg kg⁻¹, which exceeded the national soil environmental quality standard (GB 15618-2018, pH <5.5) by 6 times, and the Cd content was 9.20 ± 1.87 mg kg⁻¹, 30 times higher than the standard.

2.2 Experimental design

2.2.1 Setup of column experiments. Polyvinyl chloride columns (diameter: 10 cm; height: 28 cm) were used in the experiments. Two layers of filter paper were placed at the bottom of the column, onto which a layer of 3 cm quartz sand was deposited. The collected paddy soil was passed through a 0.425 mm nylon sieve and then used to fill the column (2.0 kg soil; 20 cm height) (Fig. 1). The soil had a pore volume (PV) of 990 mL, and the bulk density in the column was 1.15 g mL⁻¹, which was close to that of the original soil. The saturation water conductivity was (2.08 ± 0.21) × 10⁻⁵ cm s⁻¹. The column was divided into 4 equal segments of 5 cm height, each with sampling outlets to extract pore water (Fig. 1). At the beginning of the experiments, 2 mM CaCl₂ was slowly added from the bottom for 32 h to saturate the soil and remove the residual gas. For experimental flooding, an upwards reverse flow was used to simulate groundwater-driven vertical water migration.

Fig. 1 Schematic diagram of the columns used in the experiments.

2.2.2 Flooding experiments. Deionized water (DI) or artificial AMD was pumped into the soil column at 25 °C at a flow rate of 0.5 mL min⁻¹ using a peristaltic pump (Lange pump, BT100-1L, Rongboheng Constant Flow Pump Company, China). The clean water experiments were labeled S-DI and those with AMD were labeled S-AMD. Artificial AMD consisted of deionized water containing 5005.1 ± 102.2 mg L⁻¹ SO₄²⁻, 1.4 ± 0.1 mg L⁻¹ Cd, 3.5 ± 0.5 mg L⁻¹ Cu, 8.6 ± 0.2 mg L⁻¹ Fe(II), and 180.2 ± 2.1 mg L⁻¹ Fe(III) (pH = 2.41), and it was prepared by diluting stock solutions of H₂SO₄, CuCl₂, CdCl₂, FeCl₃·6H₂O, and FeCl₂. Each experiment was repeated three times. The columns were wrapped with aluminum foil and covered with a shading cloth to prevent algae growth or photooxidation under light. During operation, which lasted for 176 days, 5 mL pore water and 1.0 g of soil were collected at different depths (0, 5, 10, 15, and 20 cm) at various time points (days 0, 7, 21, 51, 84, 128 and 176) by simultaneous extraction with soil moisture samplers (MOM) (Rhizonsphere Research Products, 19.21.22F). These samples were used for subsequent analysis of pH, Fe(II), total Fe (Fe(tot)), SO₄²⁻, Cu and Cd content and for microbial community characterization. Half of each soil sample was immediately stored at −80 °C to be used for microbial sequencing analysis. The other half was freeze-dried for Cu and Cd fraction extraction. Soil fractionation was performed on the original soil and the soil samples collected during operation, using a sequential extraction scheme according to previously described procedures,²¹ which resulted in five fractions: fraction F1 (exchangeable metal species), F2 (carbonate-bound metal species), F3 (Fe/Mn oxide-bound complexes), F4 (organic matter-bound metal species), and F5 (residual metal species).

2.3 Bacterial community analysis of soil

Total soil DNA was extracted from 0.25 g of the soil sample using a MoBio PowerSoil® DNA Isolation extraction Kit (MoBio Laboratory, Carlsbad, CA, USA). DNA quality assessment and quantification were conducted by spectrophotometric analysis using a NanoDrop ND-1000 system (NanoDrop Technologies Inc., Wilmington, DE, USA) and by 1% agarose gel electrophoresis. To characterize the bacterial community, the V5–V7 region of the 16S rRNA gene was amplified from the 31 soil samples (6 time points and 5 depths per treatment plus the original soil) using primers 799F (5-AACMGGATTAGATACCKG-3) and 1193R (5-ACGTCATCCCCACCTTCC-3), followed by Illumina MiSeq amplicon sequencing.²² Multiple sequencing was performed on combined amplicons per sample type, by the use of sample-specific 7 basepair long barcodes. The PCR reaction was performed in 20 µl containing 5 µL reaction buffer, 2 µL of 2.5 mM dNTPs, 10 µM forward and reverse primers, 5 U µL⁻¹ Fast Pfu DNA polymerase, 1 µL DNA template and 14.75 µL ddH₂O. The amplification started with an initial denaturation step at 98 °C for 5 min, followed by 25 cycles of 30 s denaturation at 98 °C, 30 s annealing at 53 °C and 45 s elongation at 72 °C, followed by a final elongation step of 5 min at 72 °C. The amplicons were purified using Vazyme VAHTSTM DNA Clean Beads (Vazyme Company, Nanjing, China) and quantified with a Quant-IT PicoGreen dsDNA detection kit (Invitrogen Carlsbad, CA, USA). Sequencing was performed using an Illumina MiSeq platform (Paiseno Biotechnology Co., Ltd, Shanghai, China), and the MiSeq kit v3 was used for paired-end sequencing. After cleaning and quality-controlling the raw data, QIIME2 (ref. 23) and Mothur software packages were used to attribute validated sequences to their bacterial phylum to genus with a cutoff classification level of 97%.

2.4 Data analysis and statistics

Statistical analyses were performed using SPSS version 19.0. These included the one-way analysis of variance (ANOVA) test and Spearman correlation analysis, followed by the least ANOVA Significant Difference (LSD) test. QIIME2 Scikit-bio software was used to calculate various diversity indices and linear discriminant analysis effect size (LEfSe) analysis. Canonical correspondence analysis (CCA) was performed using normalized sequencing data to correlate microbial community members with soil properties. The relative mobility and bioavailability of the metals in the soil were calculated using the mobility factor (MF) as previously described.¹⁹

3 Results and discussion

3.1 The variations of pore water content and Cu and Cd mobility in paddy soil

The variation of pH, Fe(II), Fe(tot), SO₄²⁻, Cu, and Cd over time was analyzed at different depths of the soil column in the pore water during flooding operation by upwards reverse flow of the columns, as shown in Fig. S1. The dynamic changes in pore water under continuous AMD leaching have been discussed in detail in our previous work.²⁴ Flooding the column containing the contaminated soil with deionized water resulted in a rapid decrease of Fe(II), Fe(tot), and SO₄²⁻content, while the pH value gradually approached neutrality (Fig. S1a). Cu concentrations in the pore water decreased rapidly (within 18 days, at all depths of the column), after which it remained relatively low, with only a slight increase in the top layer as a result of leaching (Fig. S1i). The decrease of Cd concentrations in the pore water was less strong and leaching was stronger compared to Cu (Fig. S1k).

At the end of the flooding experiments, the soil was collected at various depths of the columns and subjected to sequential fractionation. The original soil was likewise analyzed. This produced fractions F1 to F5, as described in the Materials and methods section, whose metal content is shown in Fig. 2. The relative fraction content is shown in Fig. S2. While the original soil contained most Cu in fractions F3 and F4 (collectively containing 188.5 mg kg⁻¹), it was reduced to only 30.8 mg kg⁻¹ after flooding the column with clean water for 178 days (Fig. 2a). The content of Cu in F1 decreased in a depth-dependent manner, reaching as low as 2.9 mg kg⁻¹ (from 26.1 mg kg⁻¹) at 0 cm, but it decreased less strongly to 10 mg kg⁻¹ at 20 cm depth (Fig. 2a). In contrast, the content of F5 had slightly increased, from 34.8 to 44.8 mg kg⁻¹, at all depths. The effects of clean water flooding were even more striking for Cd: it reduced the contents of all fractions combined from 9.2 mg kg⁻¹ to only 0.5 mg kg⁻¹. F1-Cd and F2-Cd in the soil collectively decreased from 5.8 mg kg⁻¹ to 0.3 mg kg⁻¹, indicating that the Cd present in the AMD-contaminated paddy soil had been nearly completely mobilized under clean water irrigation. If the same process would take place in situ, this might potentially lead to the migration of heavy metals towards lower soil layers. As a result, potential risks to groundwater quality or crop uptake would remain, even if the mining soils were restored by irrigation with clean water.

Fig. 2 Vertical distribution of Cu (a and b) and Cd (c and d) in the soil fractions collected at the end of the flooding experiment with deionized water (a and c) and AMD (b and d). The soil was fractionated into following fractions: F1: exchangeable, F2: carbonate-bound, F3: Fe/Mn oxide-bound, F4: organic-bound, and F5: residual metal species.

Flooding the columns with AMD increased the amount of Cu in fraction F5 strongly, with the strongest increase in the top layers. A minor increase was also observed in fraction F1, while the content of fraction F3 slightly decreased (Fig. 2b). The effects on Cd were even more extreme: flooding the soil with AMD strongly increased the Cd content of F1, F2 and F5 to give a total Cd content of 210 mg kg⁻¹ in the top layer and between 100 and 125 mg kg⁻¹ in the deeper layers. The Cd content of fractions F3 and F4 remained very low, and the same amount of Cd was detected at a depth of 20 cm as had been originally present in the soil (Fig. 2d).

The mobility factor (MF) of Cu and Cd as a result of the flooding treatment was calculated for all depths (Table 1). The MF of Cu in the original soil was much lower (16.2%) than that of Cd. In S-DI, the MF of Cd increased to a maximum of 68.7% (at 5 cm depth), and this flooding with clean water increased the MF of Cu to maximal 21.6% (at 20 cm depth). S-AMD treatment likewise increased the MF of Cd, to a maximum of 69.9% (at 5 cm depth), while that of Cu was decreased in most depths, to the lowest value of 8.9% at 5 cm depth. This suggests that Cu may have been fixed in newly formed soil minerals during the AMD flooding. In combination, these results indicate that clean water irrigation of AMD-contaminated paddy soil can create a novel pollution source, leading to heavy metal migration to the lower soil layers, in particular for Cd.

Table 1 Mobility factor (MF) in the soil at various depths before and after flooding^a

		Day 0	Day 176
Depth			0 cm	5 cm	10 cm	15 cm	20 cm
a S-DI: flooding with DI water; S-AMD: flooding with AMD.
Cu	S-AMD	16.2%	12.0%	8.9%	9.8%	8.7%	16.2%
	S-DI	16.2%	12.8%	16.9%	19.5%	20.3%	21.6%
Cd	S-AMD	62.6%	62.9%	69.9%	68.3%	68.7%	69.2%
	S-DI	62.6%	64.4%	68.7%	63.8%	64.1%	62.2%

---

3.2 The impact on microbial diversity and microbial structure during soil rehabilitation

The bacterial community present in the soil was characterized for seven time points during operation, at five depths, by partial 16S amplicon sequencing of the collected soil samples. To compare the diversity of microbial communities under different treatments, the alpha diversity indices Chao1, Shannon, Simpson and others were calculated for a total of 31 sequence datasets per treatment (see Table S1 for S-DI and Table S2 for S-AMD). Among the α-diversity indices of bacterial communities, Chao1, Observed-species, Faith_pd and Shannon indices were slightly lower after clean water flooding compared to AMD flooding (cf. Tables S1 and S2) while no obvious differences were observed in Simpson indices and Pielou_e. Fig. 3a summarizes the data in box-and-whisker plots. This illustrates that the bacterial community displayed a slightly higher microbial richness and diversity following deionized water irrigation than after AMD irrigation, although the difference was not statistically significant (p >0.05).

Fig. 3 Characterization of the bacterial community in the soil before and after the flooding experiment. (a) Alpha diversity of the bacterial communities, as reflected by various indices, in soil of S-DI (Group A) and S-AMD (Group B). A total of 35 samples are shown, representing the communities at 7 time points and 5 depths. (b) Evolutionary branch diagram and bar diagram of LEfSe analysis identifying bacterial taxa with significant differences between S-DI (Group A, in blue) and S-AMD (Group B, in red); only taxa with LDA values >2 are shown.

The impact of AMD pollution on the microbial community structure of the soil was analyzed using LEfSe, to identify any taxonomic units whose presence was significantly different between the two treatments. Fig. 3b shows the evolutionary branch diagram and the LDA value distribution bar plot. A higher LDA value indicates a stronger significance for the differences in abundance between the two treatments. Most taxonomic groups identified by LEfSe were found to be overrepresented in the S-DI group, shown in blue in the figure. Only 8 taxonomic groups were overrepresented in the S-AMD treatment, and they reported lower LDA values than most of the identified S-DI groups (Fig. 3b).

3.3 Spatio-temporal changes in microbial communities

The relative abundance distribution of the ten most abundant phyla is shown in Fig. 4, for seven time points and five depths of two different treatments. As is typical for soil bacterial communities, the dominant phyla were Firmicutes, Acidobacteria and Proteobacteria, which in combination made up 69% in the original soil and increased as a result of flooding to as much as 98% in some samples. Other common phyla were Bacteroidetes, Actinobacteria, Nitrospirae, and Chloroflexi. Roughly speaking, the fraction of Firmicutes showed a decreasing trend over time during both treatments, and a trend of increasing abundance was observed with depth, especially under S-AMD treatment. The fraction of Acidobacteria varied considerably, both with depth and over time; this phylum was generally more abundant following S-AMD treatment, possibly reflecting their adaptability to acidic environments.^25,26 Nevertheless, at the end of both treatments, fewer than 5% Acidobacteria were detected at 5 cm and 10 cm depths, while these depths reported large fractions of Acidobacteria at other time points. This indicates that there were major fluctuations in the phylum distribution over time, in both treatments, to which Acidobacteria seemed to be particularly vulnerable. The relative abundance of Bacteroidetes also varied considerably, with four samples reporting >20% but other samples reporting <2% during S-DI treatment. This phylum varied considerably from one time point to the next, as well as with depth, especially at day 51. Striking differences between samples were also reported for Nitrospirae. In summary, it is clear from this analysis that the microbial communities were highly variable over time and with depth during both treatments.

Fig. 4 Relative abundance of the 10 most abundant phyla identified in S-DI (a) and S-AMD (b), for each time point and depth.

The top 15 most abundant genera are shown in Fig. 5; these made up between approximately 4% and 8% of the identified genera. The relative abundances of the genera collectively reported as ‘others’ in the figure are provided in Tables S3 and S4. The soil initially contained relatively large fractions of Bacillus, Lactobacillus and Staphylococcus, which all decreased with treatment, while the abundance of Candidatus_Koribacter (belonging to the phylum of Acidobacteria) was strongly increased by the treatment. This genus is widely distributed in soils, and its members were found to be tolerant to waterlogging, which may explain their expansion.²⁷

Fig. 5 Relative abundance of the 15 most abundant genera (in %) in S-DI (a) and S-AMD (b).

AMD treatment promoted proliferation of Alicyclobacillus (a genus associated with spoilage) and Fonticella, which is tolerant to metal species,²⁸ especially after 7 days. Sulfobacillus (a Clostridia member) vastly expanded at the later stages of AMD flooding. These are moderately thermophilic acidophilic bacteria typical for sulfur-containing environments where they may promote sulfide oxidation and enhance the leaching of sulfides. Members of this genus are involved in bioleaching of heavy metals.²⁹Anaeromyxobacter is a typical representative heterotrophic metal-reducing bacterium with ability for anaerobic iron reduction. It is often found in paddy soil, and its abundance increased in both treatments, in particular, in S-DI. The relative abundance of Acidiphilium was higher in S-AMD than in S-DI, most likely reflecting the acidic pH of the former. Treatment with clean water also favored Muribaculaceae members, which have been shown to tolerate heavy metals.³⁰

3.4 Major influencing factors of microbial communities

We established correlations through CCA to reveal the interaction between environmental microbial communities and particular soil properties in the soil samples, for which the abundance of the ten most abundant genera was used (Fig. 6). The CCA revealed overlapping trends of pH with Alicyclobacillus and Acidiphilium abundance, whereas sulfate, iron, Cd and Cu content overlapped with abundance of Candidatus-Solibacter and Muribaculaceae. In previous studies, soil pH was considered a key factor driving microbial patterns,^31,32 which was consistent with our findings. The first axis of the CCA explained 74.58% variability at the genus level, with the strongest positive correlation with Cu, Cd, and SO₄²⁻ and the strongest negative correlation with pH. The second axis contributed 3.37%, which revealed positive correlation with all test parameters except for pH.

Fig. 6 Sequencing plot of canonical correspondence analysis (CCA) of bacterial abundance and physicochemical parameters.

3.5 Effects of heavy metal fractions on soil microbial community

The correlations between the concentrations of Cu and Cd in various fractions of soil at the end of the experiment and the top 15 most abundant genera are shown in Fig. 7. Following clean water flooding, the easily migrating fractions F1-Cd, F1-Cu, F2-Cd, F2-Cu, and F3-Cu negatively correlated, in a mostly consistent manner, with Alicyclobacillus, Bacillus, Acidiphilium, and Lactobacillus, and positively correlated with Ruminiclostridium_1, Candidatus_Koribacter, Anaeromyxobacter, Desulfosporosinus, and Fonticella, while for the difficult migrating fractions (F3-Cd, F4-Cd, F4-Cu, and F5-Cu) it was the opposite. However, F5-Cd (which is a non-migrating fraction) grouped with the easily migrating fractions, which was not expected. The pattern was less clear after S-AMD treatment, where F4-Cd grouped together with most of the easily migrating fractions (F1-Cu, F1-Cd, F2-Cu, and F3-Cd), while F2-Cd and F3-Cu grouped with the difficult migrating fractions F4-Cu, F5-Cu and F5-Cd. Nevertheless, also here, the group of Alicyclobacillus, Bacillus, Acidiphilium and Lactobacillus reported consistent trends in relative abundance correlations for nearly all fractions, suggesting that their abundance followed similar patterns. This may suggest that they built a type of ‘core composition’ of the soil microbiome. The same applied to the group of Ruminiclostridium_1, Candidatus_Koribacter, Anaeromyxobacter, Desulfosporosinus, and Fonticella, whose abundance responded in the opposite direction to the analyzed metal species, for both treatments.

Fig. 7 Spearman correlation analysis between the Cd and Cu metal fractions and the relative abundance of microbial components at the genus level in S-DI (a) and S-AMD (b). Significance is indicated with ** for P <0.01 and with * for P <0.05, with positive correlations in blue and negative correlations in red.

4 Conclusions

In this study, we analyzed the changes in heavy metals and the recovery response of microbial communities during restauration of contaminated soil with clean water in column experiments. By comparing the composition and diversity of microbial communities at different time points and depths, the temporal and spatial variation characteristics were revealed. As the flooding time prolonged, the total content of F3-Cu and F4-Cu in combination in the soil decreased from 188.4 mg kg⁻¹ to 30.8 mg kg⁻¹. The total content of F1-Cd and F2-Cd in soil decreased from 5.8 mg kg⁻¹ to 0.3 mg kg⁻¹, and even the less mobile fractions F3-Cd and F4-Cd showed considerable reductions. During restoration, the MF of Cd maximally increased from 62.6% to 68.7% (at a depth of 0–5 cm) and that of Cu most strongly increased from 16.2% to 21.6% (15–20 cm). This suggests that when paddy soil contaminated with AMD is flooded with clean water, it will create new pollution sources, leading to the downward movement of heavy metals to deeper soil layers.

Bacterial community composition changes were apparent at multiple taxonomic levels. The dominance of the three phyla Firmicutes, Acidobacteria, and Proteobacteria strongly increased during flooding. CCA analysis showed that the pH and concentrations of Cu, Cd, and sulfate were the major factors that shifted the distribution of the microbial communities. This study highlights the recovery patterns of microbial communities in AMD-irrigated rice soil remediation and innovatively integrates microbial dynamics and geochemical parameters to provide important scientific evidence for the remediation of AMD-contaminated soils. These experimental findings, conducted using laboratory column simulations, still require confirmation under field conditions, where effects of bioturbation, plant roots, and natural hydrological processes may complicate the process of remediation and may obscure some of the effects on the soil microbial communities, and this will require further research.

Author contributions

Yan Pan: validation, data analysis, sampling, writing – original draft, visualization, supervision, funding acquisition. Zhou Fang, Yuyang Chen & Jinju Zhang: conceptualization, writing and editing. Guining Lu & Zhi Dang: supervision. Chengfang Yang: data analysis, funding acquisition.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The datasets generated and analyzed during the current study are included in this published article and its supplementary information (SI) files. Raw data can be provided if necessary. Supplementary information is available. See DOI: https://doi.org/10.1039/d5em00762c.

Acknowledgements

This study was supported by the National Natural Science Foundation of China (No. 42407023), Natural Science Foundation of the Jiangsu Higher Education Institutions of China (No. 24KJA610006), and Jiangsu Xuzhou Provincial Policy Guidance Project (No. KC23376).

References

1. A. Andalaft, L. Schwarz, P. Pino, R. Fuentes and M. Bórquez, Aybar, Assessment and modeling of nanofiltration of acid mine drainage, Ind. Eng. Chem. Res., 2018, 57(43), 14727–14739.
2. L. Lefticariu, S. R. Sutton, A. Lanzirotti and T. M. Flynn, Enhanced immobilization of arsenic from acid mine drainage by detrital clay minerals, ACS Earth Space Chem., 2019, 3(11), 2525–2538.
3. H. Wang, Y. Zeng, C. Guo, Y. Bao, G. Lu, J. R. Reinfelder and Z. Dang, Bacterial, archaeal, and fungal community responses to acid mine drainage-laden pollution in a rice paddy soil ecosystem, Sci. Total Environ., 2018, 616, 107–116.
4. Z. Zhao, J. Wang, Y. Han, J. Chen, G. Liu, H. Lu and S. Chen, Nutrients, heavy metals and microbial communities co-driven distribution of antibiotic resistance genes in adjacent environment of mariculture, Environ. Pollut., 2017, 220, 909–918.
5. C. D. Wu, J. Wu, X. Liu, Y. Wu, X. Chen and S. Yu, Ecological effects of soil properties and metal concentrations on the composition and diversity of microbial communities associated with land use patterns in an electronic waste recycling region, Sci. Total Environ., 2017, 601, 57–65.
6. P. Thavamani, R. A. Samkumar, V. Satheesh, S. R. Subashchandrabose, K. Ramadass, R. Naidu and M. Megharaj, Microbes from mined sites: Harnessing their potential for reclamation of derelict mine sites, Environ. Pollut., 2017, 230, 495–505.
7. B. Rezaie and A. Anderson, Sustainable resolutions for environmental threat of the acid mine drainage, Sci. Total Environ., 2020, 717, 137211.
8. Á. Hernández, E. C. Arellano, D. Morales-Moraga and M. D. Miranda, Understanding the effect of three decades of land use change on soil quality and biomass productivity in a Mediterranean landscape in Chile, Catena, 2016, 140, 195–204.
9. M. Gastauer, M. P. O. Vera, K. P. De Souza, E. S. Pires, R. Alves, C. F. Caldeira and G. A. Oliveira, metagenomic survey of soil microbial communities along a rehabilitation chronosequence after iron ore mining, Sci. Data, 2019, 6(1), 1–10.
10. Y. Bao, C. Guo, G. Lu, X. Yi, H. Wang and Z. Dang, Role of microbial activity in Fe (III) hydroxysulfate mineral transformations in an acid mine drainage-impacted site from the Dabaoshan Mine, Sci. Total Environ., 2018, 616, 647–657.
11. X. F. Tian, H. W. Hu, Q. Ding, M. H. Song, X. L. Xu, Y. Zheng and L. D. Guo, Influence of nitrogen fertilization on soil ammonia oxidizer and denitrifier abundance, microbial biomass, and enzyme activities in an alpine meadow, Biol. Fertil. Soils, 2014, 50, 703–713.
12. Z. F. Zhou, Y. M. Zheng, J. P. Shen, L. M. Zhang and J. Z. He, Response of denitrification genes nirS, nirK, and nosZ to irrigation water quality in a Chinese agricultural soil, Environ. Sci. Pollut., 2011, 18, 1644–1652.
13. Y. Z. Wang, C. Guo, X. Zheng, C. Ding, G. Lu and Z. Dang, Soil rehabilitation shaped different patterns of bacterial and archaeal community in AMD-irrigated paddy soil, Chemosphere, 2021, 263, 128259.
14. S. Hueso, C. García and T. Hernández, Severe drought conditions modify the microbial community structure, size and activity in amended and unamended soils, Soil Boil. Biochem., 2012, 50, 167–173.
15. A. Meisner, E. Bååth and J. Rousk, Microbial growth responses upon rewetting soil dried for four days or one year, Soil Biol. Biochem., 2013, 66, 188–192.
16. C. Yan, F. Wang, H. Geng, H. Liu, S. Pu, Z. Tian and J. Yao, Integrating high-throughput sequencing and metagenome analysis to reveal the characteristic and resistance mechanism of microbial community in metal contaminated sediments, Sci. Total Environ., 2020, 707, 136116.
17. Y. Pan, H. Ye, X. Li, X. Yi, Z. Wen, H. Wang and Z. Dang, Spatial distribution characteristics of the microbial community and multi-phase distribution of toxic metals in the geochemical gradients caused by acid mine drainage, South China. Sci. Total Environ., 2021, 774, 145660.
18. L. T. Nguyen, Y. Osanai, I. C. Anderson, M. P. Bange, D. T. Tissue and B. K. Singh, Flooding and prolonged drought have differential legacy impacts on soil nitrogen cycling, microbial communities and plant productivity, Plant Soil, 2018, 431, 371–387.
19. Y. Pan, J. Chen, K. Gao, G. Lu, H. Ye, Z. Wen and Z. Dang, Spatial and temporal variations of Cu and Cd mobility and their controlling factors in pore water of contaminated paddy soil under acid mine drainage: A laboratory column study, Sci. Total Environ., 2021, 792, 148523.
20. Y. Pan, H. Ye, Y. Yang, C. Yang, X. Li, T. Ma and G. Lu, Transport and fate of Cu and Cd in contaminated paddy soil under acid mine drainage, J. Environ. Manage., 2023, 334, 117517.
21. A. Tessier, P. G. C. Campbell and M. Bisson, Sequential Extraction Procedure for the Speciation of Particulate Trace Metals, Anal. Chem., 1979, 51, 844–851.
22. B. K. Swan, C. J. Ehrhardt, K. M. Reifel, L. I. Moreno and D. L. Valentine, Archaeal and bacterial communities respond differently to environmental gradients in anoxic sediments of a California hypersaline lake, the Salton Sea, Appl. Environ. Microbiol., 2010, 76(3), 757–768.
23. J. G. Caporaso, J. Kuczynski, J. Stombaugh, K. Bittinger, F. D. Bushman, E. K. Costello and R. Knight, QIIME allows analysis of high-throughput community sequencing data, Nat. Methods, 2010, 7(5), 335–336.
24. Y. Pan, Y. Yang, Q. Liu, M. Fu, G. Lu and Z. Dang, Leaching behavior and mechanism of Cu and Cd in acid paddy soil under acid mine drainage, J. Geochem. Explor., 2025, 107847.
25. W. Sun, X. Sun, B. Li, R. Xu, L. Y. Young, Y. Dong and Q. Wang, Bacterial response to sharp geochemical gradients caused by acid mine drainage intrusion in a terrace: relevance of C, N, and S cycling and metal resistance, Environ. Int., 2020, 138, 105601.
26. B. L. Xu, E. L. Y. Xiao, L. Y. Young and W. Sun, Uncovering microbial responses to sharp geochemical gradients in a terrace contaminated by acid mine drainage, Environ. Pollut., 2020, 261, 114226.
27. T. Lian, L. Cheng, Q. Liu, T. Yu, Z. Cai, H. Nian and M. Hartmann, Potential relevance between soybean nitrogen uptake and rhizosphere prokaryotic communities under waterlogging stress, ISME Commun., 2023, 3(1), 71.
28. Y. Dan, X. Wang, M. Ji, W. Sang, Z. Shen and Y. Zhang, Influence of temperature change on the immobilization of soil Pb and Zn by hydrochar: Roles of soil microbial modulation, Environ. Pollut., 2023, 320, 121109.
29. G. Chen, H. Shi, H. Ding, X. Zhang, T. Gu, M. Zhu and W. Tan, Multi-scale analysis of nickel ion tolerance mechanism for thermophilic Sulfobacillus thermosulfidooxidans in bioleaching, J. Hazard. Mater., 2023, 443, 130245.
30. W. J. Gong, Z. F. Niu, X. R. Wang and H. P. Zhao, How the Soil Microbial Communities and Activities Respond to Long-Term Heavy Metal Contamination in Electroplating Contaminated Site, Microorganisms, 2021, 9(2), 362.
31. K. Zhalnina, R. Dias, P. D. de Quadros, A. Davis-Richardson, F. A. Camargo, I. M. Clark and E. W. Triplett, Soil pH determines microbial diversity and composition in the park grass experiment, Microb. Ecol., 2015, 69, 395–406.
32. D. L. Jones, E. C. Cooledge, F. C. Hoyle, R. I. Griffiths and D. V. Murphy, pH and exchangeable aluminum are major regulators of microbial energy flow and carbon use efficiency in soil microbial communities, Soil Boil. Biochem., 2019, 138, 107584.

---