Does phytoextraction with Sedum plumbizincicola increase cadmium leaching from polluted agricultural soil?

Abstract Sedum plumbizincicola is a cadmium (Cd) and zinc hyperaccumulator that can activate Cd by rhizosphere acidification. However, there is little understanding of the Cd leaching risk from polluted soil during phytoextraction process. Here, pot and column experiments were conducted to monitor soil Cd leaching characteristics under different rainfall simulation conditions during S. plumbizincicola phytoextraction. Soil Cd leaching increased significantly with increasing simulated rainfall intensity. Compared with normal rainfall (NR), weak rainfall (WR) resulted in a 34.3% decrease in Cd uptake by S. plumbizincicola and also led to a 68.7% decline in Cd leaching. In contrast, Cd leaching under heavy rainfall (HR) was 2.12 times that of NR in the presence of S. plumbizincicola. After two successive growing periods, phytoextraction resulted in a 53.5–66.4% decline in the amount of soil Cd leached compared with controls in which S. plumbizincicola was absent. Even compared with maize cropping as a control, S. plumbizincicola did not instigate a significant increase in Cd leaching. The contribution of Cd leaching loss to the decline in soil total Cd concentration was negligible after phytoextraction in the pot experiment. Overall, the results contribute to our understanding of soil Cd leaching risk by phytoextraction with S. plumbizincicola. NOVELTY STATEMENT Repeated phytoextraction by hyperaccumulator Sedum plumbizincicola is an important remediation technology to remove Cd from contaminated soils. At the same time, Sedum plumbizincicola can also activate soil Cd by rhizosphere acidification. However, studies on the leaching risk of soil activated Cd during the phytoextraction process are very few. This study looked at the effects of Sedum plumbizincicola growth on soil Cd leaching with the changes in rainfall simulation and plant type. Results showed that repeated phytoextraction with Sedum plumbizincicola did not increase Cd leaching from contaminated soil.


Introduction
Potentially toxic metal pollution of agricultural land is an environmental issue of considerable public concern in many countries, especially cadmium (Cd) pollution in south China (Bolan et al. 2013;MEP and MLR 2014;Huang et al. 2019).This represents a serious threat to food safety and human health and also has negative effects on soil chemical and biological properties (Zhou et al. 2016;Mu et al. 2019;Qin et al. 2021;Tang et al. 2022).Metals accumulate particularly in surface soils and their bioavailability and mobility may be inhibited due to the changes in soil properties such as increased soil organic matter, clay contents and pH (Tang et al. 2016).However, considerable quantities of available metals remain in polluted soils which can be readily taken up by the roots of crops and hyperaccumulator plant species (Wu et al. 2018;Mu et al. 2019) or translocated to lower soil depths and potentially contaminate the groundwater by leaching (Fan et al. 2022).The mobility and bioavailability of Cd are higher than those of Zn, Cu, As and Pb, suggesting a considerable risk of Cd leaching from polluted soils (Kang et al. 2021).Ye et al. (2022) found the migration of Cd was more pronounced than that of Cu by soil columns, representing a greater potential risk of groundwater pollution than Cu.At elevated concentrations, finding effective strategies to lower soil Cd leaching has become an important method of alleviating further pollution of food and water resources.
Many physical, chemical and biological techniques have been introduced to ensure food safety by lowering and eliminating Cd that is already present in agricultural soils (Qin et al. 2021).The Chinese government also implemented a policy of remediation during fallow periods to control metal(loid) pollution of agricultural land in 2016, and important progress in pilot studies on fallow policy has been made in Hunan province, China (Zhao et al. 2017).Although managed fallow with lime addition can effectively decrease soil Cd availability, there were no significant differences in Cd availability between general fallow and normal rice cultivation (Gu et al. 2021).The availability of Cd is therefore high in polluted soils without any management (i.e., general fallow), and may leach from polluted soils in downward soil water movement.Several techniques have been used to slow down or prevent Cd losses from polluted soils, including the application of organic and inorganic amendments (Ouyang et al. 2017;Hamid et al. 2022) and plant uptake (Luo et al. 2019).Based on the hyperaccumulator plant species that are extremely tolerant to potentially toxic metals, phytoextraction is considered to be a viable and environmentally benign biological technique for Cd removal from polluted soils (Bhat et al. 2022).S. plumbizincicola is a Cd/Zn hyperaccumulator that has been widely used to remediate Cd-polluted Chinese soils (Deng et al. 2016;Wu et al. 2018;Hu et al. 2019).In these studies, S. plumbizincicola usually had high Cd phytoextraction efficiency, but the residual Cd of polluted soil was still high after the first crop.Thus, S. plumbizincicola should be planted repeatedly to continuous remediate the highly Cdpolluted soil.However, the hyperaccumulator shows rootinduced increases in soil acidification that are clearly related to higher soil Cd bioavailability (Sun et al. 2019).If the activated Cd in in the rhizosphere cannot be fully taken up by S. plumbizincicola, the excessive available Cd may increase the Cd leaching risk from polluted soils during phytoextraction process.
Attributing to the monsoon climate there are significant differences in precipitation between the dry and wet seasons across the Yunnan Plateau, and the cumulative rainfall in the wet season accounts for > 75% of the annual total (Xiao et al. 2016).Furthermore, the large interdecadal oscillation features of precipitation can also lead to extreme weather events such as heavy rainfall and long-term drought events.Previous studies indicate that the migration of metals is directly controlled by the intensity of rainfall, and heavy rainfall leads to clear Cd leaching from polluted soils (Ouyang et al. 2017;Ye et al. 2022).S. plumbizincicola has been grown in Yunnan province and has been found to have high Cd phytoextraction efficiency in polluted soils (Fan et al. 2019).However, the growth period of S. plumbizincicola coincides mainly with the wet season.Gul et al. (2019) found that amendments of EDTA, ammonium nitrate and citric acid increased the Cd concentrations of soil leachates during phytoextraction with Pelargonium hortroum.At present, few studies have been carried out to investigate the leaching characteristics of Cd in polluted soils under the interactions between rainfall simulation and phytoextraction.
In order to clarify whether the repeated phytoextraction by S. plumbizincicola can increase the Cd leaching from polluted soils.Here, Cd-polluted agricultural soils were collected from Yunnan province and repeatedly phytoextracted with S. plumbizincicola in pot experiments.The objective was to investigate the effects of S. plumbizincicola on Cd leaching from polluted soils under different simulated rainfall intensities.A low Cd accumulating plant species of maize was also selected as a control plant-soil system for further evaluation of the Cd leaching characteristics of polluted soils.The study was designed to provide reliable information on the Cd leaching risk from polluted soils phytoextracted with S. plumbizincicola.

Site description and sampling
Soil samples were collected from agricultural land in Lanping county, Yunnan province, southwest China (26 21 0 56 00 N, 99 23 0 23 00 E), an area with a low latitude mountain-plateau monsoon climate with a distinct wet season from May to October and a dry season from November to April, and mean annual temperature and rainfall values of 13.7 C and 1008 mm, respectively.The soil is classified by the Chinese Soil Taxonomy as Orthic Anthrosols developed from purple sandstone, and maize is the dominated crop of the study area.Due to long-term mining activities, local agricultural soils have been highly polluted with metals, especially Cd (Zhou et al. 2021).On April 2019 the topsoil (0-20 cm) and subsoil (20-40 cm) were collected from a polluted field.After air-drying, the soil samples were sieved through a 5-mm mesh and stored prior to the pot and column experiments.

Experiment I: pot experiment
A glasshouse pot experiment was conducted at the Institute of Soil Science, Chinese Academy of Sciences, Nanjing.The pH, total Cd concentration and 0.01 M CaCl 2 -extractable Cd concentration of the topsoil were 5.54 ± 0.05, 1.72 ± 0.07 mg kg À1 and 0.127 ± 0.001 mg kg À1 , respectively.In the present study, soil total Cd concentration was 5.7 times the Cd risk screening value (0.3 mg kg À1 ) according to Soil Environmental Quality Risk Control Standard for Soil Contamination of Agricultural Land (Trail) (GB15618-2018).Based on the annual rainfall (1008 mm) at the study site, the normal annual rainfall that falling on 1 m 2 surface of soil was about 1008 L, i.e., approximately 4 L water per kilogram of topsoil (0-20 cm depth).According to local meteorological monitoring data, over 70% of the rainfall occurs from May through October (Xiao et al. 2016), i.e., about 3.0 liters of rainwater per kilogram of soil.During the normal growing season of local crops (May to October) regular exposure to seasonal drought or heavy short-term rainfall usually induces soil water shortage or waterlogging at the experimental area.Therefore, three rainfall treatments were simulated in the pot experiment, namely weak rainfall (WR, 1.5 L kg À1 ), normal rainfall (NR, 3.0 L kg À1 ) and heavy rainfall (HR, 4.5 L kg À1 ).Sedum plumbizincicola was grown for two periods in the pot experiment, namely "SP" treatment.Soils without S. plumbizincicola were set up as controls (CK1).There were total of six treatments and each treatment had four replicates.Each pot (inner diameter-Â height: 20 cm Â 25 cm) contained 2.5 kg topsoil with a layer of quartz sand and a nylon net laid on the bottom to hold the soil and avoid extra soil loss with leaching.At the bottom of each pot there was also a small drainage hole ($ 1 cm) to permit leaching of soil solution.At the end of the growing period the total leachate from each pot was collected and pooled to obtain a homogeneous sample for further analysis.
The first (1 st ) and second (2 nd ) growing periods of S. plumbizincicola were 4 July to 12 December 2019 and 10 January to 29 May 2020, respectively.Before the rainfall simulation treatments, four S. plumbizincicola seedlings had grown for two months in each pot to obtain robust and uniform plants under the same water management regime.Inorganic fertilizers KH 2 PO 4 and CO(NH 2 ) 2 were added at 0.34 and 0.21 g kg À1 soil to each pot on 4 July, 17 September and 22 October 2019 at the 1 st growing period.Then 0.34 g KH 2 PO 4 kg À1 soil and 0.21 g CO(NH 2 ) 2 kg À1 soil were added to each pot on 30 March 2020 at the 2 nd growing period.In the following growth period until the S. plumbizincicola harvest, deionized water was sprayed on the soil surface to simulate the WR, NR and HR treatments.The rainfall simulation treatments lasted from 28 September to 10 December in the 1 st growing season and from 13 March to 27 May in the 2 nd growing season.The amount of water and spraying time in each treatment depended on the soil moisture status in the controls (CK1).The amount of water sprayed was 100-200 ml on each occasion and the spraying time was usually at an interval of 1-6 days in the controls (CK1), and the details were showed in Table S1.

Experiment II: soil column experiment
A soil column experiment was conducted in a plastic greenhouse at Lanping county, Yunnan province.The pH, total Cd concentration and 0.01 M CaCl 2 -extractable Cd concentration were 5.64 ± 0.02, 1.70 ± 0.12 mg kg À1 and 0.10 ± 0.02 mg kg À1 in the topsoil, and 6.30 ± 0.06, 0.65 ± 0.20 mg kg À1 and 0.014 ± 0.003 mg kg À1 in the subsoil.There were three treatments, namely "unplanted" control soil (CK2), soil planted with the Cd hyperaccumulator S. plumbizincicola (SP) and soil planted with maize (Zea mays L., M) with low Cd accumulation.Three S. plumbizincicola seedlings or one maize seedling were transplanted in a soil column.There were four replicate columns of each treatment.The 1 st and 2 nd growth periods were 12 April to 14 September 2019 and 20 April to 28 September 2020, respectively.Basal compound fertilizer (1.16 g kg À1 ) was mixed with the soil in each column on 12 April 2019 (1 st period) and 20 April 2020 (2 nd period), and then 0.99 g KH 2 PO 4 and 0.62 g CO(NH 2 ) 2 were applied in solution to each column on 5 August 2019 (1 st period) and 25 June 2020 (2 nd period).All treatments received the same amounts of purified water which were equivalent to the amount of precipitation throughout the plant growth period.Totals of 12.9 and 13.2 L of water were added to the soil columns in the 1 st and 2 nd growing periods, respectively.
The inner diameter and height of the soil columns were 15 cm and 50 cm, respectively.At the bottom of each column, quartz sand and nylon netting were laid successively to hold the soil and avoid extra soil loss during leaching.In the present study, the bulk density of subsoil was higher than that of topsoil.Therefore, 3.3 kg air-dried subsoil (20-40 cm depth) was placed on the nylon netting and then 2.9 kg air-dried topsoil (0-20 cm depth) on top.There was a small hole ($1-cm-diameter) at the bottom of each column to permit leaching.Throughout the experimental period, leachate was collected on 4 June, 1 July, 5 August, 15 September and 3 December 2019, and on 25 June, 7 September and 29 September 2020.The leachate collected each sampling time was retained and pooled to obtain a homogeneous sample for subsequent analysis.Each column was equipped with soil water samplers at 0-20 cm soil depth comprising a 5-cm porous membrane with a pore size of 0.15 mm (Rhizon Flex, Eijkelkamp Agrisearch Equipment, Netherlands).One day before soil solution sampling, soil water content in the columns was adjusted with deionized to produce the same water content across the treatments.

Sampling and chemical analysis
S. plumbizincicola and maize shoots were harvested at the end of each growth period.Shoot samples were washed with deionized water, oven-dried at 80 C to constant weight and weighed.Sub-samples of oven-dry shoots were finely ground prior to chemical analysis.After harvest, soil samples were collected, air-dried, and passed through 2-mm and 0.15-mm mesh sieves for chemical analysis.
Soil pH was determined using a glass electrode at a 1:2.5 (w/v) soil:water ratio.Soil available Cd was extracted with 0.01 M CaCl 2 (Houba et al. 2000).Soil samples ($ 0.20 g) were digested with a mixture of 5 ml HNO 3 and 5 ml HCl and the plant samples ($ 0.50 g) were digested with a mixture of 6 ml HNO 3 and H 2 O 2 in 50-ml sealed digestion polytetrafluoroethylene (PTFE) containers (Zhou et al. 2018).The leachates were analyzed by accurately transferring 15 ml solution to 50-ml PTFE containers and evaporating to near dryness.Each residue was digested with an acid mixture of 5 ml HNO 3 and 5 ml HCl (Zhou et al. 2021).Certified soil reference material GSS-5 purchased from the Institute of Geochemical and Geophysical Exploration, Langfang, Hebei, China, was used and the concentration of Cd was determined within 100 ± 10% of the certified value.Cadmium concentrations in the extract, digest and soil solution were determined by inductively coupled plasmamass spectrometry (ICP-MS, Optima 8000, Perkin Elmer, Waltham, MA).

Statistical analysis
Data are expressed as mean ± standard deviation (SD).Oneway analysis of variance (ANOVA) followed by LSD (least significant difference) post hoc test were conducted to test the treatment effects of simulated rainfall intensities and plant types.Results were considered significantly different at p < 0.05.If data were not normally distributed or heteroscedasticity was detected by Levene's test, values were log-transformed prior to statistical analysis.Statistical analysis was conducted using the IBM Corporation SPSS version 20 software package.

Sedum plumbizincicola biomass and cadmium content
In experiment I the shoot biomass and Cd concentration and uptake of S. plumbizincicola at the 1 st growth period was higher than at the 2 nd growth period (Table 1).Compared with NR, WR resulted in significant decreases in shoot biomass of 44.8 and 42.0% at the 1 st and 2 nd growth periods, respectively.In contrast, shoot Cd concentrations were higher in WR than in NR.After two successive growing periods, the total uptake values of Cd in S. plumbizincicola decreased significantly by 34.3% in WR compared with NR, especially in the case of the plants harvested after the 1 st growing period (by 37.8%).However, there were insignificant differences in shoot biomass, Cd concentration or Cd uptake between NR and HR across the two successive growing periods.In the experiment II the shoot concentrations and uptake of Cd in both maize and S. plumbizincicola at the 1 st growth period was higher than at the 2 nd growth period (Table 1).Although S. plumbizincicola had a lower shoot biomass than maize the concentration and uptake of Cd were both much higher in S. plumbizincicola than in maize.

Leaching losses of Cd from polluted soil
In experiment I, S. plumbizincicola phytoextraction decreased the quantity of leachate and Cd leached in polluted soil at both 1 st and 2 nd growing periods (Table 2).Compared with CK1, repeated phytoextraction with S. plumbizincicola (SP) significantly decreased the total amounts of Cd leached by 66.4, 61.1 and 53.5% in WR, NR and HR treatments, respectively.With increasing simulated rainfall, the quantity of leachate and Cd leached from soil also increased significantly.In comparison with NR, WR decreased the total amounts of Cd leached by 68.7 and 75.6%, respectively, in soil with and without phytoextraction by S. plumbizincicola.In contrast, the total amounts of Cd leached from HR were 2.12 and 0.95 times that from NR in polluted soil with and without S. plumbizincicola phytoextraction, respectively.
In experiment II, plant growth decreased the amount of soil leachate, especially maize growth (Table 2).Compared with CK2, the total amounts of soil Cd leached decreased significantly by 58.2 and 39.9% with maize (M) and S. plumbizincicola (SP) growth, respectively.Due to the large standard deviation, there were no significant differences in the amounts of Cd leached between maize (M) and S. plumbizincicola (SP) growth.

Changes in soil pH and total/available Cd concentrations
In experiment I, WR had the lowest soil pH but the highest CaCl 2 -extractable Cd concentrations of the three rainfall simulation treatments (Figure 1).In CK1 (without S. plumbizincicola) there were no significant changes in soil total Cd concentrations with increasing simulated rainfall intensity.In contrast, significant declines in soil total Cd concentration were observed with increasing rainfall intensity in polluted soil following S. plumbizincicola repeated phytoextraction.There were no significant differences between NR and HR treatments in soil pH, total Cd concentration or CaCl 2 -extractable Cd concentration.Compared with CK1, S. plumbizincicola phytoextraction resulted in significant decreases in soil total Cd and CaCl 2 -extractable Cd concentrations by 21.0-32.6 and 42.7-61.2%at the 1 st growing period, and by 29.2-51.9and 44.6-71.0%at the 2 nd growing period.S. plumbizincicola phytoextraction had little effect on soil pH across the two successive growing periods.
In column experiment II the 20-40 cm depth soil had a higher pH but lower total and CaCl 2 -extractable Cd concentrations than the 0-20 cm depth soil.Of the three treatments the "unplanted" control soil (CK2) had the lowest pH but the highest CaCl 2 -extractable Cd concentration in the 0-20 cm depth soil.Compared with maize (M), the pH and total and CaCl 2 -extractable Cd concentrations in 0-20 cm depth polluted soil with S. plumbizincicola (SP) decreased significantly by 0.45 unit, 41.3 and 34.0% at the 1 st growing period, and by 0.34 unit, 58.4 and 56.4% at the 2 nd growing period.However, in the 20-40 cm depth soil there were no significant changes in soil total or CaCl 2 -extractable Cd concentrations among the CK2, M and SP treatments.

The pH and cadmium concentrations in the soil solution
As shown in Figure 2 the Cd concentration of soil solution decreased with increasing length of experimental time.Of the three treatments, maize planting (M) caused the highest pH value and lowest Cd concentration of the soil solution in experiment II.There were no significant differences in soil solution pH between CK2 and SP treatments at either the 1 st or the 2 nd growing period, but S. plumbizincicola (SP) still resulted in lower Cd concentrations in the soil solution in comparison with CK2 after July 1, 2019.Attributed to the lower pH, a higher Cd concentration was observed in soil solution following S. plumbizincicola growth (SP) than maize growth (M) at the 1 st growing period from May 1 to September 1, 2019.However, there were no significant differences in Cd concentration of the soil solution between SP and M treatments at the 2 nd growing period from 25 June to 8 September, 2020.

Effects of S. plumbizincicola growth on Cd leaching
Here, the amounts of Cd leached from polluted soil increased with increasing simulated rainfall intensity (Table 2).Similar results were obtained by Ouyang et al. (2017) and Ye et al. (2022) who found that the cumulative losses of Cd from soils increased significantly as the rainfall intensities increased in a column leaching experiment.Estimated as the difference in Cd stocks in two adjacent soil depths, Cd leaching also increased with increasing frequency of irrigation in a long-term field experiment (McDowell 2022).Even in the presence of S. plumbizincicola growth the cumulative Cd leaching still increased with increasing rainfall intensity (Table 2).Thus, Cd leaching from soil was directly enhanced by increasing rainfall intensity, and then represented a higher potential risk of groundwater pollution (Fan et al. 2022;Ye et al. 2022).
Compared with CK2 (absence of S. plumbizincicola) the presence of S. plumbizincicola significantly decreased soil Cd leaching at each growing period (Table 2).A previous study reports that rhizosphere acidification induced by S. plumbizincicola played an important role in soil Cd mobilization (Sun et al. 2019).It seems that the roots of S. plumbizincicola had a strong ability to activate Cd from soil solid particles and then made the Cd in the rhizosphere more readily leached.However, the Cd concentrations of soil solution did not increase in the presence of S. plumbizincicola compared with CK2 (Figure 2).S. plumbizincicola is a typical Cd hyperaccumulator plant that has been shown to have a substantial capability of taking up solution Cd quickly by roots and transporting Cd from the roots to the shoots efficiently (Hu et al. 2013).Consequently, the activated Cd in rhizosphere was more accessible to take up by the roots of S. plumbizincicola immediately and a minimum amount of Cd would have leached down.In the bulk soil repeated phytoextraction resulted in a significant decline in soil CaCl 2extractable Cd concentration, but this had no acidification effect on the soil (Figure 1).In addition, the amount of leachate derived from soil also decreased significantly due to continuous metabolism and utilization of water by hyperaccumulator (Table 2).Chen et al. (2004) report that Vetiveria zizanioides growth decreased leaching by 37% compared with unplanted control soil.The continual uptake of Cd and water by S. plumbizincicola would therefore significantly lower the soil available Cd pool and thus decrease the opportunity of soil Cd leaching risk.At the 1 st growing period there were significant interaction effects on soil Cd leaching between rainfall simulation and S. plumbizincicola phytoextraction (Table S2).It is concluded that the continual uptake of Cd by S. plumbizincicola would cause further decline in Cd leaching as result of increasing rainfall intensity.At the 2 nd growing period the amount of Cd leaching was not significantly affected by an interaction between rainfall simulation and S. plumbizincicola phytoextraction.
In comparison with CK2, the significant decline in Cd leaching from polluted soil planted with maize may be largely attributable to the continual Cd uptake by maize (Table 1) and the lower soil Cd availability resulting from the higher pH (Figures 1 and 2).Due to the increments of soil solution Cd concentration (Figure 2) and quantity of leachate (Table 2), it appears that S. plumbizincicola growth might lead to more soil Cd leaching compared with maize growth, especially at the 1 st growing period.However, there were no significant differences in soil Cd leaching between S. plumbizincicola and maize growth (Table 2).This is mainly attributable to the higher uptake capacity for soil solution bioavailable Cd of S. plumbizincicola than of maize, which might then alleviate the soil Cd leaching risk.Compared with the non-hyperaccumulator Thlaspi arvense, Luo et al. (2019) found the hyperaccumulator Thlaspi caerulescens significantly decreased the leaching losses of Cd from polluted soil in a pot experiment.Here, S. plumbizincicola also showed greater uptake of Cd than of maize (Table 1).Overall, S. plumbizincicola phytoextraction significantly decreased the amounts of Cd leached from polluted soil, consequently mitigating the potential risk from Cd to groundwater pollution.Even compared with the maize crop, S. plumbizincicola did not significantly increase Cd leaching from polluted soil.

Contribution of Cd leaching to soil phytoextraction efficiency
Previous field studies found that the declines in total amounts of soil Cd were higher than the total amounts of Cd taken up by S. plumbizincicola after repeated phytoextraction (Deng et al. 2016;Zhou et al. 2018).However, in pot experiments there were no significant differences between Cd taken up by S. plumbizincicola and the decreasing Cd amount of polluted soils (Figure 3).Taking into consideration the Cd taken up by S. plumbizincicola at the field scale, Deng et al. (2016) considered that the differences between Cd uptake and soil Cd decline may be related to the inhomogeneity of the soil samples collected, losses of Cd through surface runoff or leaching down the soil profile, and atmospheric Cd deposition.In general, inhomogeneous soil sampling is attributable to the variation in root distribution which causes spatial heterogeneity of soil Cd concentrations at 0-20 cm depth after phytoextraction (Hu et al. 2019;Luo et al. 2019).In pot experiments the phytoextracted soils are usually uniformly mixed to overcome the inheterogeneity of soil sampling.In addition, the annual atmospheric Cd deposition was 25.6 g ha À1 at the site where soil sampling was conducted (Zhou et al. 2021), i.e., a contribution of only $ 0.011 mg Cd per kilogram of soil (0-20 cm depth).Even if all atmospheric Cd deposition entered the soil, it would account for only a very small proportion of the total decline in soil Cd concentration during repeated phytoextraction.V azquez et al. (2011) considered that rainfallenhanced leaching made the largest contribution to the decline in total and available Cd concentrations in pyritic sludge-polluted soils of the Vicario area in southwest Spain.As shown in Table 2, the total amounts of Cd leached were 0.17-3.66mg pot À1 , values significantly lower than the decrease in soil total Cd of 1.62-3.24mg pot À1 (Table 1) after S. plumbizincicola phytoextraction.According to one sample t-test, soil total Cd decreases were not significantly different from the test value of zero duo to the large standard deviation (Figure 3).These results indicate that Cd leaching was not the critical process influencing soil Cd decline during repeated phytoextraction in the pot experiment.Finally, surface runoff would not occur in the pot experiment.Excluding the processes of soil sampling, atmospheric deposition, leaching and surface runoff, the Cd uptake of S. plumbizincicola was the key process determining the decrease in soil Cd during repeated phytoextraction in the pot experiment of present study.

Conclusions
In the pot experiment, increases and decreases in the intensity of simulated rainfall both resulted in lower Cd uptake by S. plumbizincicola.Soil Cd leaching increased significantly with increasing simulated rainfall intensity.Due to the strong uptake capacity for Cd of the roots, S. plumbizincicola growth resulted in a significant decline in soil available Cd pool and then decreased the amount of Cd leached from polluted soil.Even compared with the conventional crop maize, S. plumbizincicola growth did not cause a significant increase in soil Cd leaching.Moreover, the contribution of Cd leaching losses to decreasing soil Cd concentration was almost negligible after repeated phytoextraction.Repeated phytoextraction with S. plumbizincicola can therefore be recommended as a remediation technique which removes Cd efficiently from polluted soil and also prevents soil Cd leaching losses.Measurement of soil Cd leaching may be used as a basis for risk assessment of the phytoextraction technique with S. plumbizincicola.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This work was supported by the National Natural Science Foundation of China (42007132, U2002210).

ORCID
Longhua Wu http://orcid.org/0000-0002-0287-8963 The significant decreases in shoot biomass by WR treatment indicate that extended water deficiency significantly inhibited the growth of S. plumbizincicola.Dole zal et al. (2022) observed similar results in which extended drought stress reduced the growth and establishment of temperate grassland plants, especially shallow-rooted species.The lack of significant differences in shoot biomass between HR and NR treatments indicates that the growth of S. plumbizincicola could adapt to a certain intensity of heavy rainfall.However, water stress (including flooding) would also result in a significant decrease in the shoot biomass of S. plumbizincicola due to a decline in photosynthetic rate(Cui et al. 2009).At the end of each growth period we observed a very low soil water content in the WR treatment compared with the NR and HR treatments (FigureS1).A previous study demonstrated that water-limited conditions (35% WHC) increased soil Cd bioavailability compared with normal water conditions (70% WHC), and then increased the Cd concentrations in wheat (Triticum aestivum)(Khan et al. 2019).Comparing with water-limited condition, the reduction in soil Cd bioavailability might be attributed to the precipitation of Cd with sulfide as insoluble form of CdS at low redox potential under water-saturated condition(Wu et al. 2019).In comparison with NR and HR treatments, the lowest pH and highest CaCl 2 -extractable Cd concentration of soil in the WR treatment may well explain the highest Cd concentration in S. plumbizincicola shoots.The NR treatment produced the largest Cd uptake and phytoextraction efficiency by S. plumbizincicola compared with WR and HR

Figure 1 .
Figure 1.Changes in pH and total and CaCl 2 -extractable Cd concentrations of polluted soil at the 1st and 2nd growing periods in experiments I (a, b, c) and II (d, e, f), respectively.Different lowercase characters indicate significant differences among the treatments within each index (p < 0.05).

Figure 2 .
Figure 2. Changes in pH and Cd concentrations of the soil solution at the 1st and 2nd growing periods in experiment II (No pH data available for the sampling date September 8, 2020).

Figure 3 .
Figure 3.Total Cd uptakes by S. plumbizincicola (SP) and maize (M) and soil total Cd decline in each pot after repeated phytoextraction across both growing periods.

Table 1 .
Biomass, Cd concentrations and Cd uptake of S. plumbizincicola (SP) and maize (M) at the 1 st and 2 nd growing periods in experiments I and II, respectively.
Different lowercase characters in the same column indicate significant differences among the treatments within each experiment (p < 0.05).

Table 2 .
Amounts of leachate and Cd leached from soil at the 1 st and 2 nd growing periods in experiments I and II, respectively.