Are Iron Tailings Suitable for Constructing the Soil Profile Configuration of Reclaimed Farmland? A Soil Quality Evaluation Based on Chronosequences

Iron tailings used as soil substitute materials to construct reclaimed farmland soil can effectively realize the large-scale resource utilization of iron tailings and reduce environmental risks. It is vital to understand the mechanisms affecting reclaimed soil quality and determine the appropriate pattern for reclamation with iron tailings. Thus, a soil quality index (SQI) was developed to evaluate the soil quality of reclaimed farmland with iron tailings in a semi-arid region. Soil samples were collected from two reclamation measures (20 cm subsoil + 20 cm iron tailings + 30 cm topsoil and 20 cm subsoil + 20 cm iron tailings + 50 cm topsoil) with reclamation years of 3 (R3), 5 (R5), and 10 (R10) at three soil depths (0–10, 10–20, and 20–30 cm) to measure 13 soil physicochemical properties in western Liaoning, China. Adjacent normal farmland (NF) acted as a reference. Results indicated that iron tailings were suitable for constructing the soil profile configuration of reclaimed farmland. SQI of reclaimed soil increased with the reclamation year, but it has not reached the NF level after 3 years, while it was better than NF after 5 years. The nutrient content of reclaimed soil increased with the reclamation year, but it still did not reach the NF level after 10 years. SQI of R10 (with 50 cm topsoil) was also better than NF but slightly lower than R5 (with 30 cm topsoil). For the semi-arid region with sticky soil texture, the topsoil thickness of reclamation was not the thicker the better, and 30 cm topsoil covered on iron tailings in western Liaoning could achieve a better reclamation effect than 50 cm.


Introduction
The exploitation of mineral resources supports rapid economic development while also causing substantial ecological and environmental problems, which has become one of the key challenges facing the sustainable development of the contemporary world [1][2][3]. Globally, surface mining destroys the regional farmland, forest, and landscape by creating huge overburden dumps and wastelands [4]. These overburdened materials mostly consist of large boulders, loose rock fragments and tailings, devoid of organic matter and nutrients, which are left unmanaged and create environmental pollution, commonly known as "mine spoil" [5,6]. Solid wastes such as mine spoil, if not properly treated and reused, would cause serious pollution to the environment and health risks [7,8]. Solid waste was a misplaced resource, and the level of recycling of resources is one of the important signs of social progress and a pathway toward green and sustainable development [9][10][11]. The report of the 19th National Congress of the Communist Party of China pointed out that we should 'establish and improve the economic system for the development of a green low-carbon cycle', and the circular economy is also called the 4R economy: Reduce, Reuse, Recycle, and quality, it is necessary to select the most representative indicators according to the research objectives, considering the factors such as cost and difficulty of test methods. Minimum Dataset (MDS) can use the least indicators to monitor and reflect changes in soil quality caused by changes in soil management measures, which has been widely used to evaluate soil quality [31]. Therefore, our study intends to evaluate the soil quality of the farmland reclamation with iron tailings in western Liaoning by using the SQI based on the MDS.
This research contributes to confirming the feasibility of solid wastes such as iron tailings can be recycled for constructing the soil profile configuration of reclaimed farmland in the existing literature through a soil quality evaluation based on chronosequence and revealing the reconstruction mechanisms of farmland reclaimed with iron tailings and the optimal reconstruction profile configuration. The specific objectives of this study were to (1) develop an SQI evaluation process and analyze the reclaimed soil quality indicators characteristics at different profile configurations and reclamation years; (2) evaluate the SQI of reconstructed farmland based on the MDS and determine the changing mechanisms and key impact indicators of the reclaimed soil quality; and (3) explore the optimal soil profile configuration of reclaimed farmland in western Liaoning and enhance our understanding of the farmland reclamation process with iron tailings to guide the reclamation technology improvement and the management of soil after reclamation.

Study Area
Our study was conducted on reclaimed farmland with iron tailings at a surface iron mining (Jianping Shengde Rixin Mining Co., Ltd.) in Jianping County, Chaoyang City, western Liaoning Province, China, at 41 • 45 N, 119 • 37 E (Figure 1), which is a region rich in iron ore resources, and mining wastes, such as iron tailings and waste rocks, occupy large amounts of land. This region is characterized by a semi-arid monsoonal climate with a mean annual temperature of 7.6 • C, mean annual precipitation of 467 mm, and a mean annual effective evaporation of approximately 1853 mm. According to the soil classification system of China, the soil in this region belongs to Hapli-Ustic Argosols [32]. The original farmland has a thin tillage layer (about 15 cm), sticky soil texture, and poor soil moisture regimes, which leads to the farmland being mostly medium and low yield fields. Aridity is the primary limiting factor of regional farmland quality.
Mining companies must reclaim mining wasteland in accordance with Chinese regulations, and due to the lack of reclamation soil sources in western Liaoning, Jianping sheng Rixin Mining Co., Ltd. combined waste iron tailings and mining stripping soil to reclaim mining wastelands as farmland. In the early stage of reclamation, due to the lack of systematic and scientific theoretical guidance, the specific reclamation schemes and soil profile configuration are different in different periods. At the beginning of reclamation, according to the correlation standard and reclamation practices, it was considered that the thicker the reclamation soil was, the better the effect was. The subsoil with a thickness of 20 cm was filled in the lowest layer, the iron tailings with 20 cm were filled in the middle layer as the soil moisture retention layer, and the tillage soil stripping from mining was covered with 50 cm as the topsoil to form a reconstructed soil profile. In recent years, due to the lack of soil sources, combined with the characteristics of regional farmland and crop growth conditions, the topsoil thickness was changed to 30 cm. On the one hand, the main purpose of filling reclamation with iron tailings is to construct a water retention layer to solve the limitation of water shortage in agricultural development in semi-arid areas. On the other hand, the iron tailings and topsoil are mixed by rotary tillage to improve the regional sticky soil texture.

Soil Sampling and Analysis
Through field investigation, the soil profile configuration of two typical iron tailings reclamation farmland was (1) 20 cm subsoil + 20 cm iron tailings + 30 cm topsoil and (2) 20 cm subsoil + 20 cm iron tailings + 50 cm topsoil. Soil samples were collected at Figure 1. Location of the study area. Note: the color pentagrams represent the distribution of soil sample plots of reclaimed farmland. NF represents the adjacent normal farmland as a reference; R3 represents farmland reclaimed for 3 years with a soil profile configuration of 20 cm subsoil + 20 cm iron tailings + 30 cm topsoil; R5 represents farmland reclaimed for 5 years with a soil profile configuration of 20 cm subsoil + 20 cm iron tailings + 30 cm topsoil; R10 represents farmland reclaimed for 10 years with a soil profile configuration of 20 cm subsoil + 20 cm iron tailings + 50 cm topsoil. The photos of the map were adopted from www.google.com/maps (accessed on 27 May 2022).
Mining companies must reclaim mining wasteland in accordance with Chinese regulations, and due to the lack of reclamation soil sources in western Liaoning, Jianping sheng Rixin Mining Co., Ltd. combined waste iron tailings and mining stripping soil to reclaim mining wastelands as farmland. In the early stage of reclamation, due to the lack of systematic and scientific theoretical guidance, the specific reclamation schemes and soil profile configuration are different in different periods. At the beginning of reclamation, according to the correlation standard and reclamation practices, it was considered that the thicker the reclamation soil was, the better the effect was. The subsoil with a thickness of 20 cm was filled in the lowest layer, the iron tailings with 20 cm were filled in the middle layer as the soil moisture retention layer, and the tillage soil stripping from mining was covered with 50 cm as the topsoil to form a reconstructed soil profile. In recent years, due to the lack of soil sources, combined with the characteristics of regional farmland and crop growth conditions, the topsoil thickness was changed to 30 cm. On the one hand, the main purpose of filling reclamation with iron tailings is to construct a water retention layer to solve the limitation of water shortage in agricultural development in semi-arid areas. On the other hand, the iron tailings and topsoil are mixed by rotary tillage to improve the regional sticky soil texture. Location of the study area. Note: the color pentagrams represent the distribution of soil sample plots of reclaimed farmland. NF represents the adjacent normal farmland as a reference; R3 represents farmland reclaimed for 3 years with a soil profile configuration of 20 cm subsoil + 20 cm iron tailings + 30 cm topsoil; R5 represents farmland reclaimed for 5 years with a soil profile configuration of 20 cm subsoil + 20 cm iron tailings + 30 cm topsoil; R10 represents farmland reclaimed for 10 years with a soil profile configuration of 20 cm subsoil + 20 cm iron tailings + 50 cm topsoil. The photos of the map were adopted from www.google.com/maps (accessed on 27 May 2022). According to the research results of Bünemann et al. [37], combined with the experience and the specific situation of the study area, the total dataset (TDS) of soil quality was established, including 13 indicators (Table 1). In our study, principal component analysis (PCA) combined with Norm value and Pearson correlation analysis was used to select the soil indicators that can best reflect the soil quality characteristics and have significant indigenous effects on the evaluation results from the TDS as the minimum dataset (MDS) [38]. Norm value is the length of the vector norm of the indicator in the multidimensional space composed of components; the longer the length, indicating that the greater the comprehensive load of the indicator in all principal components, the stronger the ability to explain comprehensive information [39]. The formula is as follows (Equation (1)): where N ik is the comprehensive load of the indicator i on the first k principal components with eigenvalue ≥1; µ ik is the load of the indicator i on the principal component k; λ k is the eigenvalue of the principal component k.
The factors with high eigenvalues and soil variables with high factor loading were assumed to be indicators that can foremost represent farmland soil [40]; hence, the retained principal components are selected according to the eigenvalue >1, and the loading values of the indicator was within 10% of the maximum loading value [41,42]. If a single component contains more than one soil attribute, the multivariate correlation coefficient is used to determine whether the variable is redundant. For variables with significant correlation, a variable with a high Norm value was selected for soil quality evaluation, and the rest were excluded. If the highly weighted variables are not correlated, each variable can be used for soil quality evaluation [26].

Evaluation Model of Soil Quality Index (SQI)
SQI is a comprehensive reflection of soil function by calculating the weight and score of each soil quality evaluation index. The greater the value, the better the soil quality [37]. According to the positive and negative correlation between each soil quality evaluation index and soil quality, the membership function between the evaluation index and soil quality was established, and the membership degree of the index was calculated by Equations (2) and (3) [43][44][45]. Then, the role of each factor is calculated by using the factor load in principal component analysis, and their weights are determined by Equation (4). Finally, the comprehensive evaluation index of soil quality is calculated by Equation (5) through the weighted comprehensive method and addition multiplication [46].
where S i is the standard value of soil variable, x ij is the measured value of soil quality index i in the year j, x i−max is the maximum value of index i, and x i−min is the minimum value of index i.
where w i is the weight of the soil quality index i, C i is the common factor variance, C is the sum of the common factor variance.

Data Analysis
The SPSS (Statistical Program for the Social Sciences, release 25.0) was used to perform a correlated statistical analysis of the data. All variables follow the normal distributions (tested with the Kolmogorov-Smirnov test at the p-value of 0.05). One-way ANOVA (analysis of variance) was carried out to compare the means of soil characteristics of normal farmland and reclaimed soil chronosequence sites. Differences between individual means were tested using DMRT (Duncan's multiple range test) at p < 0.05 significance level. PCA-loaded variables were subject to Pearson correlation analysis. Origin 2021 was used for drawing.

Feature of Soil Quality Evaluation Indicators
Results of soil physical and chemical analyses indicated that the characteristics of reclaimed soil quality improved with the increase in reclamation years (Figures 2-5). BD is a sensitive indicator of soil compaction, and the soil with low BD is loose, which is beneficial to water storage, and vice versa. BD of reclaimed farmland was higher than that of NF in different reclamation years. At the depths of 0-10 cm, 10-20 cm, and 20-30 cm, BD of R3 was significantly increased by 26.37%, 18.26%, and 20.13% (p < 0.05), R5 was significantly increased by 19.71%, 16.48%, and 17.13% (p < 0.05), and R10 was significantly increased by 7.86%, 11.33%, and 11.54% (p < 0.05), respectively, compared with NF ( Figure 2). In addition, BD increased with soil depth. At different soil depths, the differences in R3 and R5 were not significant (p > 0.05), but in R10, the difference was significant (p < 0.05). With the increase in reclamation years, BD at all levels decreased significantly. SWC of NF was the highest, reclaimed farmland was significantly lower than that of NF, and the differences between different reclamation years were significant (p > 0.05). With the reclamation year increase, SWC showed a continuous growth trend in each soil depth. The SWC of R10 at 0-10 cm, 10-20 cm, and 20-30 cm increased by 57.53%, 21.37%, and 20.75%, respectively, compared with R3. There was no significant difference in SWC of different soil depths within the same reclamation year.
Clay of NF was the highest in all sample plots, and clay of reclaimed farmland in R3, R5, and R10 was significantly different from that of NF (p < 0.05) ( Figure 3). There was no significant difference in clay among different reclamation years (p > 0.05), and with the increase in reclamation year, clay increased slightly. However, there were significant differences among different soil depths in the same reclamation year (p < 0.05). The silt was the highest in NF and the lowest in R5, and it was significantly different among each sample plot (p < 0.05), while it has no significant difference among different soil depths in the same reclamation year (p > 0.05). The sand was the lowest in NF and the highest in R5, and it was significantly different among each sample plot (p < 0.05). It has no significant difference among different soil depths in the same reclamation year (p > 0.05). 0.05). With the increase in reclamation years, BD at all levels decreased significantly. SWC of NF was the highest, reclaimed farmland was significantly lower than that of NF, and the differences between different reclamation years were significant (p > 0.05). With the reclamation year increase, SWC showed a continuous growth trend in each soil depth. The SWC of R10 at 0-10 cm, 10-20 cm, and 20-30 cm increased by 57.53%, 21.37%, and 20.75%, respectively, compared with R3. There was no significant difference in SWC of different soil depths within the same reclamation year.  Table 1 for abbreviations and units.
Clay of NF was the highest in all sample plots, and clay of reclaimed farmland in R3, R5, and R10 was significantly different from that of NF (p < 0.05) ( Figure 3). There was no significant difference in clay among different reclamation years (p > 0.05), and with the increase in reclamation year, clay increased slightly. However, there were significant differences among different soil depths in the same reclamation year (p < 0.05). The silt was the highest in NF and the lowest in R5, and it was significantly different among each sample plot (p < 0.05), while it has no significant difference among different soil depths in the same reclamation year (p > 0.05). The sand was the lowest in NF and the highest in R5, and it was significantly different among each sample plot (p < 0.05). It has no significant difference among different soil depths in the same reclamation year (p > 0.05).  Table  1 for abbreviations and units.  Table 1 for abbreviations and units. SWC. Different capital letters indicate significant differences (p < 0.05) among different reclamation years in the same soil layer, and different lowercase letters indicate significant differences (p < 0.05) among different soil layers in the same reclamation year. See Table 1 for abbreviations and units.
Clay of NF was the highest in all sample plots, and clay of reclaimed farmland in R3, R5, and R10 was significantly different from that of NF (p < 0.05) ( Figure 3). There was no significant difference in clay among different reclamation years (p > 0.05), and with the increase in reclamation year, clay increased slightly. However, there were significant differences among different soil depths in the same reclamation year (p < 0.05). The silt was the highest in NF and the lowest in R5, and it was significantly different among each sample plot (p < 0.05), while it has no significant difference among different soil depths in the same reclamation year (p > 0.05). The sand was the lowest in NF and the highest in R5, and it was significantly different among each sample plot (p < 0.05). It has no significant difference among different soil depths in the same reclamation year (p > 0.05).  Table  1 for abbreviations and units.  Table 1 for abbreviations and units. There were significant differences in pH between different sample plots and different soil depths (p < 0.05) (Figure 4). pH decreased with the increase in reclamation year at 0-10 cm and 10-20 cm depths, from 7.62 and 7.61 in R3 to 7.41 and 7.55 in R10, respectively. OM of NF was the highest, which was greater than 10.0 g/kg at 0-10 cm and 10-20 cm depths. With the increase in the reclamation year, OM increased significantly (p < 0.05), but OM was still less than 8.0 g/kg. In the same reclamation year, OM decreased significantly with the increase in soil depth (p < 0.05). OM. Different capital letters indicate significant differences (p < 0.05) among different reclamation years in the same soil layer, and different lowercase letters indicate significant differences (p < 0.05) among different soil layers in the same reclamation year. See Table 1 for abbreviations and units.
TN of NF was significantly higher than the other three reclaimed farmlands ( Figure  5). At 0-10 cm, TN of R3, R5, and R10 were decreased by 73.39%, 56.88%, and 46.79%, respectively, compared with NF. TN increased significantly with the reclamation year increase (p < 0.05), but there was no significant difference at different soil depths in the same reclamation year (p > 0.05). TP was highest in NF and lowest in R3; it was 0.37 g/kg and 0.27 g/kg at 0-10 cm, respectively. TK of each reclaimed farmland was higher than NF,  Table 1 for abbreviations and units.  Table  1 for abbreviations and units.

Construction of MDS for Soil Quality Evaluation
In the results of PCA, the eigenvalues of the first three components were greater than 1, and their cumulative contribution rate reached 90.96%, indicating that the minimum dataset can replace the whole dataset for soil quality evaluation ( Table 2). The first principal component variance was 60.25%, in which TN had the maximum loading value. The loading values of TN, BD, OM, and AN were within 10% of the maximum loading value, while TN had a high correlation with the other three variables ( Figure 6), respectively, 0.914, 0.993, and 0.975 (p < 0.01); therefore, only TN in PC-1 was selected as the MDS. The variance of PC-2 was 19.83%. pH had the maximum loading value, and TP, AP, and clay loading values were within 10% of the maximum loading value. According to Figure 6, the correlation coefficients of pH and TP, AP were respectively −0.85 (p < 0.01) and −0.60 (p < 0.05), while the correlation between pH and clay is very low, and the correlation between clay and TP, AP were all very low, while TP and AP had a high correlation with 0.74 (p < 0.01), according to the Norm value, AP and clay were selected in MDS. The variance of PC-3 was 10.88%, silt and sand were within 10% of the maximum loading value, while the correlation coefficient of silt and sand was very high at −0.98 (p < 0.01), sand was selected in the MDS depending on the Norm value. The MDS of the soil quality evaluation of farmland soil constructed using iron tailings in the semi-arid region consists of TN, AP, clay, and sand.   Table 1 for abbreviations and units.
There were significant differences in pH between different sample plots and different soil depths (p < 0.05) (Figure 4). pH decreased with the increase in reclamation year at 0-10 cm and 10-20 cm depths, from 7.62 and 7.61 in R3 to 7.41 and 7.55 in R10, respectively. OM of NF was the highest, which was greater than 10.0 g/kg at 0-10 cm and 10-20 cm depths. With the increase in the reclamation year, OM increased significantly (p < 0.05), but OM was still less than 8.0 g/kg. In the same reclamation year, OM decreased significantly with the increase in soil depth (p < 0.05).
TN of NF was significantly higher than the other three reclaimed farmlands ( Figure 5). At 0-10 cm, TN of R3, R5, and R10 were decreased by 73.39%, 56.88%, and 46.79%, respectively, compared with NF. TN increased significantly with the reclamation year increase (p < 0.05), but there was no significant difference at different soil depths in the same reclamation year (p > 0.05). TP was highest in NF and lowest in R3; it was 0.37 g/kg and 0.27 g/kg at 0-10 cm, respectively. TK of each reclaimed farmland was higher than NF, and it showed a significant downward trend with the increase in the reclamation year. AN of NF was significantly higher than that of each reclaimed farmland (p < 0.05). At 0-10 cm, 10-20 cm, and 20-30 cm, AN was 66.03, 48.56 mg/kg, and 43.21 mg/kg, respectively, but in reclaimed farmland, it was all less than 40 mg/kg, while with the increase in reclamation year, AN showed a significant upward trend. AP of NF was significantly higher than that of reclaimed farmland, and there was no significant difference between different reclamation years, but it was significantly different at 0-10 cm from 10-20 cm and 20-30 cm (p < 0.05). AK of NF was significantly lower than that of reclaimed farmland, and it was the highest in R3, which was 138.75 mg/kg, 94.22 mg/kg, and 108.98 mg/kg at 0-10 cm, 10-20 cm, and 20-30 cm, respectively. With the increase in reclamation year, AK decreased significantly. In addition, AK among different soil depths in the same reclamation year was significantly different (p < 0.05). In general, nitrogen and phosphorus increased with the reclamation year but were all lower than NF; potassium was significantly higher than NF after reclamation but decreased with the reclamation year.

Construction of MDS for Soil Quality Evaluation
In the results of PCA, the eigenvalues of the first three components were greater than 1, and their cumulative contribution rate reached 90.96%, indicating that the minimum dataset can replace the whole dataset for soil quality evaluation ( Table 2). The first principal component variance was 60.25%, in which TN had the maximum loading value. The loading values of TN, BD, OM, and AN were within 10% of the maximum loading value, while TN had a high correlation with the other three variables (Figure 6), respectively, 0.914, 0.993, and 0.975 (p < 0.01); therefore, only TN in PC-1 was selected as the MDS. The variance of PC-2 was 19.83%. pH had the maximum loading value, and TP, AP, and clay loading values were within 10% of the maximum loading value. According to Figure 6, the correlation coefficients of pH and TP, AP were respectively −0.85 (p < 0.01) and −0.60 (p < 0.05), while the correlation between pH and clay is very low, and the correlation between clay and TP, AP were all very low, while TP and AP had a high correlation with 0.74 (p < 0.01), according to the Norm value, AP and clay were selected in MDS. The variance of PC-3 was 10.88%, silt and sand were within 10% of the maximum loading value, while the correlation coefficient of silt and sand was very high at −0.98 (p < 0.01), sand was selected in the MDS depending on the Norm value. The MDS of the soil quality evaluation of farmland soil constructed using iron tailings in the semi-arid region consists of TN, AP, clay, and sand. Note: The variable corresponding to the bold value is selective further due to its relatively high scores. Variable loading coefficients (eigenvalues) of the first three factors were extracted using 13 soil attributes, their eigenvalues, and individual and cumulative percentage of total variance explained by each factor. Factor loadings are considered highly weighted when within 10% of the variation of the absolute values of the highest factor loading in each factor. Bold-underlined soil attributes correspond to the indicators included in the MDS. See Table 1 for abbreviations.
Empirically, the main purpose of using iron tailings as the matrix to fill reclaimed farmland in areas with sticky soil texture was to improve soil texture, and it has the downside of weak capacity in holding fertilizer. Clay and sand are basic elements used to reflect soil texture. TN and AP are basic elements used to maintain crop growth. Accordingly, TN, AP, clay, and sand are suitable and essential for evaluating the soil quality of reclaimed farmland with iron tailings.
Note: The variable corresponding to the bold value is selective further due to its rela scores. Variable loading coefficients (eigenvalues) of the first three factors were extracte soil attributes, their eigenvalues, and individual and cumulative percentage of total v plained by each factor. Factor loadings are considered highly weighted when within variation of the absolute values of the highest factor loading in each factor. Bold-unde attributes correspond to the indicators included in the MDS. See Table 1 for abbreviation Empirically, the main purpose of using iron tailings as the matrix to fill farmland in areas with sticky soil texture was to improve soil texture, and it has t side of weak capacity in holding fertilizer. Clay and sand are basic elements used soil texture. TN and AP are basic elements used to maintain crop growth. Acc TN, AP, clay, and sand are suitable and essential for evaluating the soil qua claimed farmland with iron tailings.

Soil Quality Evaluation Based on MDS
When performing PCA again for the four selected SQI evaluation indicators in MDS, each PC explained a certain amount (%) of the variation in the dataset (Table 3). TN has the highest contribution value in SQI, with a commonality of 0.833, followed by sand, with a commonality of 0.824, and the commonality of clay and AP is 0.559 and 0.393, respectively. The weight of TN, AP, clay, and sand was 0.319, 0.151, 0.214, and 0.316, respectively. We calculated the scores of each index and sum the weighted scores of each variable to obtain the SQI value of each sample plot in 0-10, 10-20 cm, and 20-30 cm depths (Figure 7). SQI was significantly higher (p < 0.05) at 0-10 cm (0.454-0.636) than 10-20 cm (0.383-0.528) and 20-30 cm (0.262-0.504) of each sample plot. SQI of NF was 0.542, 0.528, and 0.262, respectively, at 0-10 cm, 10-20 cm, and 20-30 cm. Among three kinds of reclamation farmlands in different years, the SQI of R5 was the highest at 0.636 at 0-10 cm, followed by R10 (0.597). SQI of R3 was both the lowest at 0-10 cm (0.454) and 10-20 cm (0.383); it decreased by 16.24% and 27.46%, respectively, compared with NF, but increased by 24.05% at 20-30 cm. SQI of R5 was significantly improved (p < 0.05). At 0-10 cm and 20-30 cm, it was significantly higher than that in the normal farmland, increasing by 17.34% and 92.37%, respectively, and restored to the normal farmland level at 10-20 cm. SQI at 0-10 cm and 20-30 cm of R10 was also significantly better than that of NF, increasing by 10.15% and 86.26%, respectively, but was slightly lower than that of R5. There were significant differences between R5 and R3 but no significant differences between R5 and R10.

Applicability Verification of Soil Quality Evaluation Method Based on MDS
Generally, the soil quality can be evaluated with high accuracy through the TDS of the soil quality evaluation indicators. However, due to the numerous indicators, experimental analysis is complicated and time-consuming. The indicator dataset can be simplified through a series of statistical analyses, but it will lead to a decrease in evaluation accuracy. Therefore, it is necessary to verify the applicability of MDS of the evaluation indicator in a specific region or a specific soil. The common factor variance of each indicator of TDS was obtained by PCA, and then, the weight of each indicator of TDS was obtained ( Table 3). The above method was used to analyze the soil quality of TDS.
SQI based on the MDS (MDS-SQI) and SQI based on the TDS (TDS-SQI) were used for regression analysis to verify the accuracy of the comprehensive value of soil quality based on the MDS (Figure 8). MDS-SQI and TDS-SQI met the linear regression relationship (p < 0.01), and the correlation coefficient was 0.840. The regression equation is y = 0.840x+ 0.147 (n = 36, R 2 = 0.712, p < 0.01), where y represents TDS-SQI, x represents MDS-SQI. The above analysis shows that MDS can replace TDS, and the quality evaluation of farmland soil reclaimed by iron tailings through the MDS indicator system has high accuracy.

Applicability Verification of Soil Quality Evaluation Method Based on MDS
Generally, the soil quality can be evaluated with high accuracy through the TDS of the soil quality evaluation indicators. However, due to the numerous indicators, experimental analysis is complicated and time-consuming. The indicator dataset can be simplified through a series of statistical analyses, but it will lead to a decrease in evaluation accuracy. Therefore, it is necessary to verify the applicability of MDS of the evaluation indicator in a specific region or a specific soil. The common factor variance of each indicator of TDS was obtained by PCA, and then, the weight of each indicator of TDS was obtained ( Table 3). The above method was used to analyze the soil quality of TDS.
SQI based on the MDS (MDS-SQI) and SQI based on the TDS (TDS-SQI) were used for regression analysis to verify the accuracy of the comprehensive value of soil quality based on the MDS (Figure 8). MDS-SQI and TDS-SQI met the linear regression relationship (p < 0.01), and the correlation coefficient was 0.840. The regression equation is y = 0.840x + 0.147 (n = 36, R 2 = 0.712, p < 0.01), where y represents TDS-SQI, x represents MDS-SQI. The above analysis shows that MDS can replace TDS, and the quality evaluation of farmland soil reclaimed by iron tailings through the MDS indicator system has high accuracy.
for regression analysis to verify the accuracy of the comprehensive value of soil quality based on the MDS (Figure 8). MDS-SQI and TDS-SQI met the linear regression relationship (p < 0.01), and the correlation coefficient was 0.840. The regression equation is y = 0.840x+ 0.147 (n = 36, R 2 = 0.712, p < 0.01), where y represents TDS-SQI, x represents MDS-SQI. The above analysis shows that MDS can replace TDS, and the quality evaluation of farmland soil reclaimed by iron tailings through the MDS indicator system has high accuracy.

Chronosequence Evolution of Soil Quality in Reclaimed Farmland
Soil is an interconnected system, and the reconstruction process contains a series of chain reactions that take a certain amount of time [1]. Soil texture was improved after reclamation with iron tailings. Compared with NF, the most significant features of the reclaimed soil in each reclamation year were high sand and low clay. Sand had the largest value in each reclaimed soil, and clay was significantly lower than NF (Figure 7). This was because the soil in the study area is cinnamon soil with high clay content. The iron tailings were filled under the topsoil and would be mixed into topsoil by tillage during the crop planting. Iron tailings were mostly irregular granular [47], and their specific surface area was larger than that of clay particles, increasing the sand content. Meanwhile, lots of iron tailings reduce the clay content and regulate the soil mechanical composition, which fully demonstrates that reclaiming farmland with iron tailings can effectively improve the soil texture in western Liaoning, and the results were consistent with those of Yang [48].
The reclaimed soil quality was poor in the early stage of reclamation; SQI of R3 was significantly lower than NF because TN and AP of R3 were significantly lower than NF ( Figure 5). Soil nitrogen and phosphorus are essential elements for plant growth, and the phosphorus content affects soil fertility and physical and chemical characteristics such as SWC, pH, and OM [49]. Although TN of reclaimed soil showed an overall upward trend with the reclamation year, it was always smaller than NF. TN of R10 was only 53% of NF, and the nitrogen supply capacity was significantly poor. The variation trend of AP was consistent with TN and showed an upward trend with the reclamation year, but it was also less than NF. AP of R10 was only 70% of NF, and the phosphorus supply capacity was significantly poor, which was consistent with Li et al. [50], Duo and Hu [51], and Li et al. [52], that TN and AP showed an upward trend with the reclamation year, but it was always lower than that of normal farmland. The texture of iron tailings is sandy, and the water and fertilizer conservation abilities are poor. The nutrient of reclaimed soil was low, and the recovery time was long. The accumulation of soil nutrients is a long-term process, and it needs to be improved by changing the irrigation mode, rational planting, and following the principle of small amounts and multiple applications of fertilizer.
After 5 years of reclamation, SQI was significantly higher than NF, indicating that the reclamation with iron tailings has a great influence on the comprehensive quality of soil, and the reclaimed soil after 5 years can reach or even better than the comprehensive quality of normal farmland in the region. The research results of Mukhopadhyay et al. [6] also showed that the quality of reclaimed soil improved with the increase in reclamation year. Cao et al. [53] reported that the reclaimed soil was largely restored after 12 years of reconstruction; however, the recovery was not completed.

Effect of Profile Configuration on Reclaimed Soil Quality
Reclaimed soil quality will recover better with the increase in reclamation year [6], but the SQI of R10 was slightly less than R5. By comparing their profile configurations, R5 covered 30 cm topsoil on iron tailings, while R10 covered 50 cm topsoil. Through field investigation, local farmers plowed farmland with a depth of 30-40 cm by rotary tillage machine in spring. Therefore, the topsoil with 30 cm would mix iron tailings in the plowing process, which effectively increases the sand content of the topsoil and improves the sticky soil texture, and the number of iron tailings mixed in the topsoil increases with the years of cultivation and the soil texture was getting better and better (Figure 9). However, iron tailings were difficult to be mixed into the topsoil with 50 cm through plowing, so the effect of using iron tailings to improve the sticky soil texture was poor. In addition, the sticky soil texture in the region leads to poor soil water retention capacity, and aridity is the primary limiting factor of regional farmland quality [48]. In the reclamation process, the role of filling iron tailings in the middle layer is to use the pore structure of iron tailings to build a water retention layer and improve the water holding capacity of the reclaimed soil. The research results of Jia [54] showed that in farmlands similar to the soil texture of our study area when the rainfall is less than 20 mm, the infiltration depth is 20 cm; when the rainfall is 20-50 mm, the infiltration depth is 40 cm, and when the rainfall is more than 50 mm, the infiltration depth can reach the soil layer below 40 cm. The mean annual precipitation in Jianping County is 467 mm, with the highest rainfall in July, about 137 mm. Generally, the maximum individual rainfall is less than 50 mm, so it is estimated that the infiltration depth of an individual rainfall in the study area will not exceed 40 cm. Therefore, the soil moisture migration might make it difficult to In addition, the sticky soil texture in the region leads to poor soil water retention capacity, and aridity is the primary limiting factor of regional farmland quality [48]. In the reclamation process, the role of filling iron tailings in the middle layer is to use the pore structure of iron tailings to build a water retention layer and improve the water holding capacity of the reclaimed soil. The research results of Jia [54] showed that in farmlands similar to the soil texture of our study area when the rainfall is less than 20 mm, the infiltration depth is 20 cm; when the rainfall is 20-50 mm, the infiltration depth is 40 cm, and when the rainfall is more than 50 mm, the infiltration depth can reach the soil layer below 40 cm. The mean annual precipitation in Jianping County is 467 mm, with the highest rainfall in July, about 137 mm. Generally, the maximum individual rainfall is less than 50 mm, so it is estimated that the infiltration depth of an individual rainfall in the study area will not exceed 40 cm. Therefore, the soil moisture migration might make it difficult to reach the iron tailings layer with 50 cm topsoil, which makes the soil water retention effect of the iron tailings layer poor, while the soil moisture could reach the iron tailings layer with 30 cm topsoil and effectively maintain the soil moisture, thereby promoting the nutrient cycling in soil (Figure 9). The amount of iron tailings mixed in the topsoil increases with the years of cultivation, which made the soil moisture retention ability stronger and the nutrient cycling ability better. As shown in Figure 2b, SWC after reclamation for 5 years was significantly higher than that after 3 years.
Our result is different from some previous cognition, that was, the thicker the reclaimed soil in land reclamation is, the better it is [55]. For semi-arid regions, the topsoil thickness when iron tailings are used for reclamation is not the thicker, the better, but the thickness suitable for regional conditions is the best. In western Liaoning, based on regional soil characteristics and tillage practice, when iron tailings are used as mine soil to reconstruct farmland, 30 cm topsoil covered on iron tailings can achieve a good reclamation effect, and with the increase in reclamation year, it is conducive to improving the quality of reclaimed soil. This reclamation pattern consumes large amounts of iron tailings and forms a soil water-retaining layer, which meets the requirements of regional crop cultivation and forms a good reclamation effect. Moreover, reducing the coverage thickness of the topsoil can effectively save soil resources and reclamation costs [56].

The Variation Regulation of Soil Quality in Vertical Profiles with Different Reclamation Years
The MDS-SQI was significantly higher (p < 0.05) at 0-10 cm (0.454-0.636) than 10-20 cm (0.383-0.514) and 20-30 cm (0.325-0.504) (Figure 10) of the reclaimed farmland. The MDS-SQI observed for NF was 0.542, 0.528, and 0.262, respectively, at 0-10 cm, 10-20 cm, and 20-30 cm. It showed that the quality of topsoil in reclaimed soil was higher than that of subsoil. In a similar study on reclaimed sites, the quality of topsoil was higher than that of subsoil [57]. For both the soil layers, the overall trend of SQI has followed the order: R3 < NF < R10 < R5. After 5 years of reclamation with iron tailings, the soil comprehensive quality at each soil depth could reach or even better than the level of regional normal farmland. Our result is different from some previous cognition, that was, the thicker the reclaimed soil in land reclamation is, the better it is [55]. For semi-arid regions, the topsoil thickness when iron tailings are used for reclamation is not the thicker, the better, but the thickness suitable for regional conditions is the best. In western Liaoning, based on regional soil characteristics and tillage practice, when iron tailings are used as mine soil to reconstruct farmland, 30 cm topsoil covered on iron tailings can achieve a good reclamation effect, and with the increase in reclamation year, it is conducive to improving the quality of reclaimed soil. This reclamation pattern consumes large amounts of iron tailings and forms a soil water-retaining layer, which meets the requirements of regional crop cultivation and forms a good reclamation effect. Moreover, reducing the coverage thickness of the topsoil can effectively save soil resources and reclamation costs [56].

The Variation Regulation of Soil Quality in Vertical Profiles with Different Reclamation Years
The MDS-SQI was significantly higher (p < 0.05) at 0-10 cm (0.454-0.636) than 10-20 cm (0.383-0.514) and 20-30 cm (0.325-0.504) (Figure 10) of the reclaimed farmland. The MDS-SQI observed for NF was 0.542, 0.528, and 0.262, respectively, at 0-10 cm, 10-20 cm, and 20-30 cm. It showed that the quality of topsoil in reclaimed soil was higher than that of subsoil. In a similar study on reclaimed sites, the quality of topsoil was higher than that of subsoil [57]. For both the soil layers, the overall trend of SQI has followed the order: R3 < NF < R10 < R5. After 5 years of reclamation with iron tailings, the soil comprehensive quality at each soil depth could reach or even better than the level of regional normal farmland.

Key Indicator Identification and Policy Implications
By using a series of soil quality indicators, the effect of filling reclamation with iron tailings on soil quality was studied through the SQI method based on MDS, and the applicability of this method in the region was verified. The MDS was screened by PCA combined with the Norm value, and the Norm value was introduced to consider the load of the indicator on all principal components to avoid the loss of information on other principal components [58]. The summary of the relevant research results of soil quality evaluation based on MDS by Bünemann et al. [37] showed that bulk density, pH, organic matter,

Key Indicator Identification and Policy Implications
By using a series of soil quality indicators, the effect of filling reclamation with iron tailings on soil quality was studied through the SQI method based on MDS, and the applicability of this method in the region was verified. The MDS was screened by PCA combined with the Norm value, and the Norm value was introduced to consider the load of the indicator on all principal components to avoid the loss of information on other principal components [58]. The summary of the relevant research results of soil quality evaluation based on MDS by Bünemann et al. [37] showed that bulk density, pH, organic matter, sand percentage, silt percentage, total nitrogen, available phosphorus, and soil water content have a high frequency of use. The three indicators (sand, TN, AP) determined in our study were consistent with most of the results. In addition, clay was selected as MDS in our study, indicating that in addition to sand, TN, and AP, the effect of clay content on soil quality in the study area was also significant, which was determined by the regional natural conditions and soil characteristics. The study area is a typical semi-arid region, the soil texture was sticky, and filling reclamation with iron tailings would change the soil texture. Therefore, the four MDS indicators selected in our study are the key indicators affecting the quality of reclaimed soil in western Liaoning.
The traditional filling reclamation materials such as fly ash and gangue usually contain harmful heavy metal elements such as Cd, Pb, and Hg, and the reclaimed soil has potential ecological harm [26]. In addition, if these heavy metals diffuse with runoff, they will also pollute a wider range of soil [51]. The results of our previous study found that the content of heavy metals such as Cd, Cr, Cu, Zn, Pb, Ni, Hg, and As did not exceed the risk intervention value for pollution of agricultural land in China [59], even lower than the heavy metal content in the regional native soil, with no toxicity and are not a source of pollution for the soil and crops [48]. In addition, our previous column leaching test results also showed that the heavy metal content in the leachate was very low, which met the standard [59], especially the content of Cr, Pb, and Cd was very low, which was almost impossible to be detected. This once again proves that it is feasible to use iron tailings as reclamation materials of mine wasteland, which can increase the farmland area and reduce the tailings accumulation, and save the cost of tailings pond management.
In general, our study evaluated the reclaimed soil quality after the implementation of different iron tailings reconstruction farmland techniques to identify the most effective reclamation techniques. The results of this study can also provide valuable policy implications on improving the treatment of waste iron tailings, guiding the formulation of land reclamation technical standards, and promoting ecological restoration of the mining area. For example, in the future formulation of relevant technical standards for land reclamation in China, the appropriate thickness of topsoil cover should be reasonably determined according to the actual situation of reclamation areas. In addition, the research results have a positive guiding role in the formulation of technical policies for "harmless", "reduction", and "resource utilization" of solid waste in China. Furthermore, through the wide promotion and application of this technology, the use of iron tailings to reclaim historical legacy mines in the region can effectively solve the problems of limited reclamation resources and shortage of repair funds faced by local governments in the ecological restoration of historical legacy mines and improve the comprehensive utilization value of abandoned lands. Our research results are of great significance to promoting the ecological restoration of mines, realizing the sustainable development of ecological civilization construction, and supporting the "UN decade on Ecosystem Restoration" action.

Conclusions
Iron tailings were confirmed to be suitable as soil substitutes for constructing the soil profile configuration of reclaimed farmland. The comprehensive quality of reclaimed soil improved with the reclamation year, but it has not reached the level of regional normal farmland after 3 years of reclamation. The soil quality after 5 years of reclamation was better than that of normal farmland. SQI of R10 was also better than NF but slightly lower than R5. The quality of topsoil was better than that of subsoil in the same reclaimed farmland. The thickness of topsoil would affect the reclaimed soil quality. The soil quality of 30 cm topsoil covered in 5 years of reclamation was better than that of 50 cm topsoil covered in 10 years of reclamation. For the semi-arid region with sticky soil texture, the thickness of reclaimed topsoil is not the thicker, the better. The topsoil covering 30 cm after iron tailings filling in western Liaoning could achieve a better reclamation effect; the topsoil texture was improved, and the reclamation cost was effectively saved.
Our study mainly analyzes the effect of the measures that have been completed by regional mines to reclaim farmland with iron tailings. The study was not carried out based on the strict and systematic experimental scheme design but conducted targeted research according to the actual reclamation process with iron tailings. According to the results of our study, it is expected to further establish a test site for systematic research in the study area in the future, to provide a basis for the improvement of ecological restoration theory and waste resource utilization technology in mining areas.