Evaluation of soil quality along two revegetation chronosequences on the Loess Hilly Region of China

Science of the Total Environment 633 (2018) 808–815

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Evaluation of soil quality along two revegetation chronosequences on the Loess Hilly Region of China Shujuan Guo a,b, Xinhui Han a,b, Hui Li c, Tao Wang a,b, Xiaogang Tong d, Guangxin Ren a,b, Yongzhong Feng a,b, Gaihe Yang a,b,⁎ a

College of Agronomy, Northwest A&F University, Yangling, 712100, Shaanxi, China The Research Center of Recycle Agricultural Engineering and Technology of Shaanxi Province, Yangling 712100, Shaanxi, China College of Forestry, Northwest A&F University, Yangling, 712100, Shaanxi, China d College of Natural Resources and Environment, Northwest A&F University, Yangling, 712100, Shaanxi, China b c

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• SQI was developed from physicochemical and biological properties. • Key indicators are microbial biomass carbon, silt, clay, total phosphorus and pH. • SQI values for Robinia pseudoacacia L. markedly increased with restoration age. • SQI values for abandoned land reached a steady-state after 27 years of restoration. • SQI values were higher in Robinia pseudoacacia L. than abandoned land.

a r t i c l e

i n f o

Article history: Received 23 January 2018 Received in revised form 18 March 2018 Accepted 18 March 2018 Available online xxxx Editor: Jay Gan Keywords: Vegetation restoration Soil quality index Microbial biomass carbon Restoration pathway

a b s t r a c t Vegetation restoration has been widely implemented to control soil degradation, reduce soil erosion, and improve soil quality. It is vital to understand the mechanisms affecting soil quality in soil restoration processes and to determine an appropriate recover pattern for soil restoration. Thus, a soil quality index was developed using integrated approach to assess soil quality after vegetation restoration in this study. Soil samples were collected from two restoration pathways (afforestation by Robinia pseudoacacia L. and natural recovery of abandoned farmland) with ages sequence of 0, 17,27 and 42 years old at two soil depths (0–10 and 10–20 cm) to measure soil physicochemical and biological properties on the Loess Hilly Region of China, China. The results showed that soil quality index (SQI) was developed based on microbial biomass carbon (MBC), fine particles (FP), and total phosphorus (TP). The MBC, which had the fastest increase rate than TP and FP, had the highest contribution to the final SQI and these contributions increased with recovery age. The MBC values were higher in Robinia pseudoacacia L. than in abandoned land sites at all recovery ages with greater increases along with restoration age. The SQI values significantly increased with increasing restoration age up to 27 years (P b 0.05). After 27 years, SQI values for the AL sites remained stable, while SQI values for RP sites continually improved with increasing restoration age. In addition, SQI values were higher for RP sites than for AL sites for all restoration ages. © 2018 Elsevier B.V. All rights reserved.

⁎ Corresponding author. E-mail addresses: [email protected] (X. Han), [email protected] (G. Ren), [email protected] (G. Yang).

https://doi.org/10.1016/j.scitotenv.2018.03.210 0048-9697/© 2018 Elsevier B.V. All rights reserved.

S. Guo et al. / Science of the Total Environment 633 (2018) 808–815

1. Introduction Vegetation restoration by converting farmland into perennial vegetation occurs globally in various climatic conditions (arid, semi-arid, temperate, and tropical) and ecosystem types (cropland, grassland, and forestland) (Cao et al., 2008; Raiesi, 2011; Templer et al., 2005; Zhang et al., 2011). During vegetation restoration, changes in plant species composition and coverage can alter litter input, root architecture (Schedlbauer and Kavanagh, 2008), and physical (Zhao et al., 2017a), chemical (Lucas-Borja et al., 2012) and biological properties of the soil (Ren et al., 2016a). Changes in soil function and quality may occur as a consequence of these variations (Raiesi, 2017). However, large uncertainties remain concerning the effects of vegetation restoration on soil quality due to differences of revegetation type, restoration age, ecosystems and biomes (An et al., 2009; Zhang et al., 2011). For example, Zhang et al. (2011) found that improvements in soil quality for abandoned land were better than grassland and shrubland after eight years vegetation restoration on the Loess Hilly Region of China. Therefore, it is vital to assess the impacts of vegetation restoration on soil quality during soil restoration processes. Soil quality, defined as the soil capacity to ensure the sustainability of the soil environment and biosphere, can be estimated using various soil quality indicators (Doran et al., 1996; Karlen et al., 2003; Raiesi, 2017). Several soil physical and chemical properties of the soil, such as soil texture, pH, soil water content (SWC), soil organic carbon (SOC) and total nitrogen (TN), reflect soil fertility and structure, and are widely used to indicate soil quality (Raiesi, 2017; Zhang et al., 2011). However, these properties usually change slowly and do not reflect soil quality changes over short time period. Whereas, soil biological properties, such as soil microbial biomass and enzyme activity, are sensitive to soil disturbance and are involved in nutrient cycling and organic matter dynamics (Bastida et al., 2008; Raiesi, 2011). Even though these individual soil properties can be considered as soil quality indicators, the impacts of vegetation restoration on soil quality cannot be assessed using individual soil parameters as they are interdependent and unlikely to thoroughly reflect these complex ecosystems (Raiesi and Kabiri, 2016; Yakovchenko et al., 1996). Therefore, developing a soil quality index based on several different soil characteristics can provide a more effective evaluation of soil quality after vegetation restoration. For example, Mukhopadhyay et al. (2016) developed an SQI evaluate reclaimed coal mine spoil, and recommended two native species for restoration. Using this approach, Zhang et al. (2011) developed an SQI to compare the impacts of different revegetation types on soil quality, and revealed that natural recovery is the best choice for soil restoration on the Loess Plateau. Although SQIs have been showed to be an effective method to reflect soil quality changes in a variety of ecosystems, there is little available information on soil quality evaluation along two chronosequences, especially on the Loess Hilly Region. The Loess Hilly Region of China has a typical semiarid climate and is known for its considerable soil erosion (Li et al., 2016); soil erosion and desertification have resulted in severe land degradation (Bai and Dent, 2009; Li et al., 2016). To change these conditions and restore ecosystems, the Chinese government undertook vegetation restoration programs in the 1950s (Deng et al., 2013; Ren et al., 2016b); to data, N9.27 million ha of farmland have been converted into grassland and forest (Ren et al., 2016a). Vegetation reestablishment on farmland has greatly reduced soil erosion (Fu et al., 2010). The reclaimed land has been stabilized using different vegetation types at various points in time, which provides an opportunity to study the mechanisms affecting soil quality at different stages in the restoration process. Meanwhile, a comparison of the effects of different revegetation types (afforestation by Robinia pseudoacacia. L and natural recovery of abandoned farmland) on soil quality along two chronosequences is essential to select the appropriate vegetation type for restoration in fragile areas. In the present study, we hypothesized that vegetation restoration would improve soil quality, and the stage increase rate in soil quality would decrease

809

in the later recovery age. We also hypothesized that higher soil quality and faster increase rate would be found in afforestation land. Thus, the objectives of this study were to (1) develop a comprehensive soil quality index, (2) evaluate the long-term impacts of vegetation restoration on soil quality, and (3) determine the most suitable revegetation type, which is most capable of restoring soil quality. 2. Materials and methods 2.1. Study sites The study was carried out in the Wuliwan watershed in Ansai County, Shaanxi Province, northern Loess Plateau, China (36°51′41.23′ ′-36°52′50.87′′N, 109°19′49.20′′-109°21′46.46′′E). The region has hilly-gullied loessial landforms with a temperate semiarid climate. The area's mean annual temperature is 8.8 °C and mean annual precipitation is 505 mm (with 70% falling between July and September) (Zhao et al., 2017a). The soil is mainly composed of Calcaric Cambisol, originating from primitive or secondary loess parent materials, which is characterized by weak cohesion and is easily eroded (Fu et al., 2010). The study region has experienced severe soil erosion and degradation. Since the implementation of the vegetation restoration program, farmlands with slopes higher than 25° have gradually been abandoned for natural recovery and afforestation. Robinia pseudoacacia L. is the main species used for vegetation restoration. Our study area has been protected as an experimental site by the Institute of Soil and Water Conservation, Chinese Academy of Science (CAS) since 1973 (Ren et al., 2016b). 2.2. Experimental design, field investigation and sampling In July 2016 (based on the space-for-time substitutions method), we selected sites representing two typical vegetation restoration types, at three recovery ages, with similar environmental conditions: land abandoned for natural recovery for 17 years (AL17), 27 years (AL27), and 42 years (AL42), and land planted with Robinia pseudoacacia L. for 17 years (RP17), 27 years (RP27) and 42 years (RP42). Millet (Setaria italica L.) farmland (FL) was chosen as a reference area (0 years recovery); millet was sown at a depth of 20 cm in May 2016 and the plants were harvested in August 2016. Prior to afforestation, there was little difference in farming practices between the sampling sites. Within each sites, three independent replicate plots (30 × 30 m) were established for sampling. The distance between any two plots was b500 m to ensure that they had similar environmental conditions. Five subplots (1 × 1 m) were established within each plot, at the four corners and the center, to conduct the field investigations. Herb coverage and species presence were determined for all vegetation types (Table 1). After removing the litter layer and debris, soil samples from the 0–10 cm and 10–20 cm soil layers were collected from 10 points in an “S” shape, using a soil auger (5 cm inner diameter). For each soil layer, these ten samples were homogenized to provide a composite sample for each replicate site. The final samples were sieved through a 2-mm screen to remove roots and other debris. Thereafter, these fresh samples were divided into three parts, one of which was used to measure the soil water content (SWC); the second part was air-dried at room temperature, and stored for analysis of the physical and chemical properties; the last part was stored at 4 °C to analyze its biological properties. Root samples were collected from each plot at 10 points in an “S” shape at a soil depth of 0–20 cm using a root auger (9 cm inner diameter). In each plot, five 1 × 1 m random quadrants were established, all the living biomass was removed, and the litter biomass was collected to provide a final litter sample. The soil bulk density (BD) of each soil layer was measured using a soil bulk sampler (5 cm diameter and 5 cm height) with three replicates and then dried in an oven at 105 °C for 48 h.

810

S. Guo et al. / Science of the Total Environment 633 (2018) 808–815

Table 1 The basic data of the land-use types. Vegetation type

Dominant herbaceous species

Aspect

Slope gradient (°)

Elevation (m)

Initial vegetation density (trees/ha)

Herbaceous coverage (%)

Fine root biomass (kg/hm2)

Litter biomass (t/ha)

AL17 AL27 AL42 RP17 RP27 RP42

Lespedeza davurica, Heteropappus altaicus Artemisia vestita, Lespedeza daurica Lespedeza daurica, Artemisia giraldii Pamp Poa sphondylodes, Patrinia heterophylla bunge Poa sphondylodes, Rubia cordifolia Arenariae radix, Poa sphondylodes

N by E 20 N by E 15 N by E 30 N by E 40 N by E 30 N by E 50

40 35 42 35 45 45

1288.1 1285.6 1293.5 1307.8 1291.3 1293.8

– – – 1333 1333 1111

55 ± 3.69c 53 ± 2.7 cd 66 ± 1.5b 43 ± 7.3d 72 ± 3.4ab 77 ± 2.2a

580.8 ± 2.16d 639.8 ± 3.73d 774.7 ± 1.81d 1416.5 ± 6.29c 1686.5 ± 14.98b 1933.9 ± 10.30a

3.57 ± 0.22c 4.96 ± 0.18c 7.21 ± 0.19c 31.37 ± 1.75a 32.87 ± 2.59a 21.44 ± 1.32b

AL: Abandoned land, RP: Robinia pseudoacacia L.

2.3. Sample analysis Soil particle sizes (clay, silt and sand content) were measured using the MasterSizer 2000 method (Malvern MasterSizer 2000, Worcestershire, UK) (Deng et al., 2016). As both clay and silt reflect soil texture, we combine the two into fine particle (FP). The SWC was measured gravimetrically and expressed as a percentage of soil water to dry soil weight. Soil pH was measured using a soil: water ratio of 1:2.5 (PHSJ-4A pH acidometer, Shanghai, China). Soil organic carbon (SOC) was determined by dichromate oxidation (Nelson et al., 1982). Soil total nitrogen (TN) was extracted using the Kjeldahl method (Sparks et al., 1996), and total phosphorus (TP) was assayed using the HClO4-H2SO4 ammonium molybdate ascorbic acid method (Olsen and Le, 1982). Microbial biomass carbon (MBC) and microbial biomass nitrogen (MBN) were determined using the fumigation extraction method (Vance et al., 1987). Enzyme activities were assayed as described by Ren et al. (2016a). For the analysis of enzyme, controls were included without substrate and soil. Catalase activity was determined by the addition of 40 ml distilled water and 5 ml 0.3% H2O2 to 2 g fresh soil, shaken for 20 min (at 150 rpm), and then immediately filtered through Whatman 2 V. The filtrate was titrated with 0.1 mol L−1 KMnO4 under the conditions of sulfuric acid. The final results were expressed as 0.1 mol KMnO4 g−1 20 min−1 (Raiesi and Kabiri, 2016). Saccharase activity was determined using 5 g of fresh soil, 15 ml of 8% glucose solution, 5 ml of 0.2 M phosphate buffer (pH 5.5), and 5 drops of toluene. The samples were incubation 24 h at 37.8 °C, then these solution was filtered (Whatman 2 V) immediately and a 1-ml aliquot was transferred to a volumetric flask with 3 ml 3,5- dinitrylsalicylate, and heated for 5 min. After soil solution reached room temperature, this solution was quantified colorimetrically in an Ultraviolet Spectrometer Subsystem at 508 nm. Finally, results were expressed as mg glucose g−124 h−1. Urease activities were determined using 5 g of fresh soil, 5 ml of citrate solution (pH 6.7) and 5 ml of 10% urea solution. After incubation for 24 h at 37.8 °C, soil solution was diluted to 50 ml with distilled water. The mixtures were filtered and a 1-ml of supernatant was treated with 4 ml of sodium phenol solution and 3 ml of 0.9% sodium hypochlorite solution. The released ammonium was determined colorimetrically in an Ultraviolet Spectrometer Subsystem at 578 nm. Results were expressed as mg −1 NH+ 24 h−1. Alkaline phosphatase activity was determined 4 -N g using 10 g of fresh soil, 2 ml of toluene, 10 ml of disodium phenyl phosphate solution and 10 ml of 0.05 Mborate buffer (pH 9.6). The suspensions were incubated for 2 h at 37.8 °C. After incubation, the samples were filtered, then the filtrate was reacted with 0.5 ml of 2% 4aminoantipyrine and 8% potassium ferrocyanide. The released phenol was quantified colorimetrically in an Ultraviolet Spectrometer Subsystem at 510 nm. Results were expressed as mg phenol g−12 h−1.

the data were considered (Andrews et al., 2002; Brejda et al., 2000). Under a certain PC, only the parameters within 10% of the highest factor loading were selected for indexing (Andrews et al., 2002; Brejda et al., 2000). When more than one parameter was selected in a single PC, correlation analysis was applied to identify their correlations. When these parameters were strongly correlated with each other, only the highest loading factor was retained for the index. If there was no significant correlation between the parameters, each was considered important and, thus, retained in the SQI. Then, these selected soil parameters were transformed using nonlinear scoring functions to obtain unitless (Andrews et al., 2002; Masto et al., 2010). Once transformed, SQI values were calculated using the weighted additive equation (Andrews et al., 2002). High SQI values represent good soil function and a relatively stable soil ecosystem (Andrews et al., 2002; Armenise et al., 2013). 2.5. Data calculation Increase rates of selected soil quality indicators were estimated using the following equations (Zhang et al., 2012). R ¼ ðIx −Iref Þ=Iref  100

ð1Þ

where I refers to soil quality indicators; x is the value in land-use types; and Iref refer to I for farmland. Although the rate of increase in SQI values may not be constant over time, the mean rate of increase could be calculated using the following equation (Deng et al., 2016; Zhang et al., 2012): Stage increase rate of SQI Rstage ¼ △SQI=ð△t  SQIref Þ  100

ð2Þ

where △SQI refer to SQI at start and end of a recovery stage, and △t represents the time interval for each stage. SQIref refers to SQI for farmland. 2.6. Statistical analysis A multivariate analysis of variance (MANOVA) was used to identify the influence of vegetation type, age and soil layer on soil parameters. A one-way analysis of variance (ANOVA) and least significant difference (LSD) were calculated to examine differences between mean values. In addition, the differences were evaluated at the 5% significance level. Correlations among the properties were analyzed using a correlation analysis. All statistical analyses were performed using SPSS version 20.0 (SPSS Inc. Chicago, USA) at P = 0.05. 3. Results 3.1. Vegetation characteristics

2.4. Developing the soil quality index A principal component analysis (PCA) was used to choose appropriate parameters and their weighting factors. Only principal components (PCs) with eigenvalues ≥ 1, that explained at least 5% of the variation in

Vegetation properties differed significantly among the six restored plant areas (Table 1). The lands which were afforested by Robinia pseudoacacia have the similar initial vegetation density. Herbaceous coverage was lowest (43%) in RP17 and was N50% for the other

S. Guo et al. / Science of the Total Environment 633 (2018) 808–815

reclaimed land. Fine-root biomass and litter biomass were considerably higher in the RP areas than in the AL areas and increased significantly with recovery age (P b 0.05). In the same restoration age, fine-root biomasses in RP sites were nearly three times as much as in AL sites, while there were more than three times for litter biomasses. 3.2. Selected soil quality indicators The MANOVA showed that land-use type, soil depth, and their interaction could significantly influence soil parameters as a whole (P b 0.05) (Table S1). SOC, TN, SWC, MBC and MBN, and the activity of four enzymes (catalase, urease, saccharase and alkaline phosphatase) significantly increased, while BD decreased with restoration age in both restoration pathways (Figs. S1, S2 and S3). The SOC, TN, SWC, MBC, MBN and enzyme activities were higher in Robinia pseudoacacia L. (RP) than in abandoned land (AL) sites at each restoration age (Figs. S1, S2 and S3). As all the soil parameters were significantly influenced by land-use type, soil depth or their interaction, they were selected using the PCA for calculating the SQI. The PCA results showed that the first three PCs had eigenvalues N1.0 and explained 81.601% of the total variance (Table 2). The highly weighted parameters in PC-1 were SOC, TN, SWC, urease, saccharase, MBC, and MBN. MBC, which had the highest loading (0.954), was significantly correlated (P b 0.05) with the other six highly weighted parameters parameters (SOC, TN, SWC, urease, saccharase, and MBN) in PC-1 (Table S2). Thus, only MBC was included from PC-1 to estimate SQI. FP and sand were selected from PC-2 (16.261% variance). There was negative correlation between FP and sand (r = −1, P b 0.001). Therefore, only one of them were retained for SQI. PC-3 was characterized by TP and accounted for 8.305% of total variability. The final indicators as determined from PCA for the SQI were MBC, FP, and TP. The results of three-way ANOVAs showed that the selected soil quality indicators (MBC, FP, and TP) differed significant among vegetation types, ages, soil layers or their interaction (Table 3) (P b 0.05). Notably, improvements in MBC corresponded to increased recovery age of both soil layers for the two restoration pathways examined (Fig. 1I and IV). After the implement of vegetation restoration, MBC values were higher in RP sites than in AL sites at the same restoration ages with this difference being greater in the surface soil than in the 10–20 cm soil layer (P b 0.05). FP values ranged from 57.69% to 61.18% for RP sites and from 57.66% to 60.55% for AL sites (Fig. 1 II and V). In both soil layers, FP Table 2 Variable loading coefficients (eigenvectors) of the first two principal components for nine soil variables (0–20 cm layer), and their individual and cumulative percentage of total variance explained by each principal component and eigenvalues. Soil properties

Factor 1

Factor 2

Factor 3

SOC TN TP SWC pH BD FP Sand Catalase Phosphatase Urease Saccharase MBC MBN Eigen value Variance (%) Cumulative variance (%)

0.939 0.895 −0.090 0.870 −0.439 −0.748 0.256 −0.256 0.782 0.828 0.866 0.888 0.954 0.952 7.985 57.035 57.035

0.166 −0.125 −0.371 −0.164 0.315 −0.230 −0.928 0.928 0.273 0.143 −0.180 0.208 0.150 0.004 2.277 16.261 73.296

−0.028 0.039 −0.796 0.184 0.601 0.152 0.216 −0.216 −0.093 −0.019 0.002 0.040 0.072 0.009 1.163 8.305 81.601

Boldface factor loadings are considered highly loaded; Bold-underlined soil factors correspond to the indicators included in the index. SOC, organic carbon; TN, total nitrogen; TP, total phosphorus; SWC, water content; BD, bulk density; FP, Fine particles; MBC, microbial biomass C; MBN, microbial biomass N.

811

values were lower in farmland (0 year of restoration) compared with reclaimed land with no significant differences between these two vegetation types (Table 3). Results also showed that there were no significant differences in RP sites among restoration ages in the 10–20 cm soil layer (P N 0.05). TP values were affected by restoration age only in the 0–10 cm soil layer (Fig. 1 III, P b 0.05). In 0–10 cm layer, the lowest TP value was in the RP42 site and the highest value was in the AL27 site. In the 10–20 cm soil layer, there were no significant differences in TP for either of two restoration pathways among restoration ages, and TP values tended to be higher in RP (0.43–0.51) than in AL (0.36–0.43) sites (Fig. 1 VI). There were no significant differences in TP between soil layers (Table 3). The increase rates of soil quality indicators varied significantly among reclaimed land (Fig. 2). MBC had the fastest increase rate, with a N20% increase in the MBC on reclaimed land. In contrast, the changes of FP and TP with age ranged from −2.39% to 6.04% and −31.04% to 24.13%, respectively. The rate of MBC for AL increased over the first 27 years after restoration, and then remained relatively constant after this point, while MBC continually improved for RP sites with increasing restoration age. 3.3. Soil quality index Soil quality indicators (MBC, FP and TP) were transformed into scores using nonlinear scoring functions, which were shown in Table 4. The weights for MBC, FP and TP were 0.70, 0.20, and 0.10, respectively. The equation for SQI based on normalized equations and weights is given below. SQI ¼ 0:70  S ðMBCÞ þ 0:20  S ðFPÞ þ 0:10  S ðTPÞ

ð3Þ

where S is the score for the indicator. The results of the SQI for each site are shown in Fig. 3. MBC contributed the highest to the final SQI and these contributions increased with recovery age (40.71–60.54% at 0 years, 52.98–69.98% at 17 years, 60.25–73.56% at 27 years, and 64.10–82.12% at 42 years) (Fig. 4). The final SQI values were significantly affected by vegetation type, soil depth and restoration age (P b 0.05; Fig. 3). Improvements in SQI values for RP and AL corresponded to an increase in recovery age, with higher SQI values found in RP sites than in AL sites at the same recovery age. The SQI values were also significantly higher for the 0–10 cm than the 10–20 cm soil layer (P b 0.05). The results also showed that SQI values for AL sites were relatively constant after 27 years of restoration, but continually improved for RP sites with increasing recovery age. The stage increased rates of SQI values varied among restoration age, with 1.28–1.56%, 2.48–2.54%, and 0.25–2.65% increases for stages of 0–17, 17–27 and 27–42 years, respectively (Fig. 4). In the same soil layer, the stage increased rates in RP sites were higher than in AL sites, expect for stage 17–27 in the 10–20 cm layers. For RP sites, there were non-significant differences in the changes in SQI values among different restoration ages (P N 0.05, Fig. 4), although values were higher in the later stage (27–42 year) of vegetation restoration than in the other two earlier stages. However, there were significant differences among restoration stages for AL, with AL27 having the highest rates, followed by AL42, and then with AL17 having the lowest rates (P b 0.05). 4. Discussion 4.1. Influence of vegetation restoration on soil quality index An integrated soil quality index (SQI) based on a combination of soil properties will provide a more effective evaluation of soil quality (Karlen et al., 2003). In the present study, three parameters (MBC, FP, and TP) were chosen to develop the final SQI, consistent with previous studies which also concluded that the above parameters were good indicators of overall soil quality (Raiesi, 2017; Zhang et al., 2011). The

812

S. Guo et al. / Science of the Total Environment 633 (2018) 808–815

Table 3 Results of three-way ANOVA for soil indicators (MBC, fine particles and TP). Soil indicator

MBC FP TP

Type

Type × Layer

Age × Layer

F

Sig.

Age F

Sig.

Layer F

Sig.

Type × Age F

Sig.

F

Sig.

F

Sig.

Type×Age× Layer F

Sig.

103.87 10.26 5.35

0.000 0.003 0.027

231.46 8.01 2.61

0.000 0.000 0.068

451.53 0.29 1.70

0.000 0.597 0.202

61.83 6.98 3.53

0.000 0.001 0.026

8.07 0.06 4.60

0.008 0.803 0.040

26.34 10.75 2.36

0.000 0.000 0.090

19.42 2.45 0.83

0.000 0.081 0.486

MBC, microbial biomass carbon; FP, Fine particles; TP, total phosphorus.

present study showed that this approach can effectively be used to distinguish the influence of vegetation restoration on the soil quality on the Loess Plateau. In the studied ecosystem, microbial biomass carbon (MBC) contributed the highest to the integrated SQI, indicating that the soil microbial biomass played an important role in determining soil quality (Fig. 3). Previous studies demonstrated that microbial biomass was related to

organic matter dynamics and nutrient cycling, and thus it is globally considered as a good soil quality indicator (Anderson, 2003; Raiesi and Kabiri, 2016). In the present study, we also found that significant enhancements to soil quality over the progression of restoration were mainly attributable to higher soil microbial biomass, which was significantly and positively related to SOC and TN (Fig. 3, Table S2). Additionally, MBC showed the quickest increase rate among the soil quality

Fig. 1. Soil quality indicators (MBC, FP, and TP) for different land-use types at the 0–10 and 10–20 cm depths. Values are means (n = 9) ± standard error of mean.

S. Guo et al. / Science of the Total Environment 633 (2018) 808–815

813

that it could be used as an indicator of soil environmental changes (Table 3). Previous studies have shown that vegetation restoration obviously increased soil nutrient contents, but the increase has been observed to be less for phosphorus than for other nutrients (Zhang et al., 2011; J. Deng et al., 2016). The present study also found nonsignificant differences in TP over the course of restoration in two pathways at the 10–20 cm layer, and a lower rate of increase of TP than other indicators (Figs. 1 and 2). This is probably because the phosphorus is rock-derived element, which may not easily respond to changes to the soil ecosystem (Walker et al., 1976). Similar observations were also reported in previous study in the same watershed (Ren et al., 2016a). Phosphorus availability however remains a vital indicator of plant nutrient limitation that can influence processes important to ecosystem functioning (Koerselman and Meuleman, 1996). Thus, phosphorus should be taken into consideration when establishing a comprehensive soil quality index. 4.2. Differential responses of soil quality to restoration pathway

Fig. 2. Restoration ratio of soil quality indicators (MBC, FP, and TP) at the 0–10 and 10–20 cm soil layers.

indicators, which indicated that it was more sensitive to land-use change than the others (FP and TP) and would thus be the most readily detected indicator of variation within the soil environment (Fig. 2). The composition of fine particles (clay and silt) in the soil, as the most important index of soil structure, is also considered an indicator of soil quality. Raiesi (2017) demonstrated that soil texture was strongly correlated with soil quality and determined soil erosion, soil infiltration, and storage capacity of water and nutrients in the soil. In the present study, soil clay content was significantly and positively correlated with SOC and TN, while sand content was negatively correlated with SWC and nutrients (Table S2). Although FP the contribution of FP to the SQI was lower than that of MBC, it still showed similar tendency to that of MBC to improve with restoration age, and thus it could be used to detect impacts of vegetation restoration, particularly within the surface layer. Total phosphorus (TP) is considered the main indicator of soil nutrients contents and could directly reflect soil fertility. Our study showed that TP has been selected for the final soil quality index, illustrating Table 4 Summary of average and range values; normalized equations of the scoring functions.

Average Curve type (scoring function) Weighting factor (standardized)a Slope Normalized equation

MBC

Fine particles

TP

154.28 More is better

59.22 More is better

0.44 More is better

0. 57 (0.70)

0.16(0.20)

0.08(0.10)

−2.5 1 / (1 + (x/154.28)−2.5)

−2.5 1 / (1 + (x/59.22)−2.5)

−2.5 1 / (1 + (x/0.44)−2.5)

MBC, microbial biomass carbon; TP, total phosphorus. a Derived from principal component analysis (Table 2) and values in parenthesis represent weighting factors which are equal to the percentage oftotal variance explained by the factor standardized to unity.

Vegetation restoration modifies plant communities and soil properties (Woods, 2000). It is generally accepted that soil quality increases along with vegetation restoration (An et al., 2009; Yu and Jia, 2014). In the present study, SQI significantly increased along with recovery age at both soil layers (P b 0.05, Fig. 3). This observation is consistent with previous studies, who found improved soil physicochemical and biological properties along revegetation chronosequences (An et al., 2009; Zhao et al., 2017b). This is probably because the older restoration plots can have more extensive root systems, which in turn bring about more exudation and organic matter into the soil, stimulating synthesis of these enzymes, hence facilitating higher microbial biomass (Table 1) (Ren et al., 2016a). Meanwhile, accumulation of organic matter from the input of plant residues can bind microaggregates to form aggregate pores, contributing to higher soil fertility and lower BD, therefore accelerating microial activity and sustaining higher microbial biomass (Table S3) (Ren et al., 2017; Rutigliano et al., 2004). Thus, soil quality increases along with restoration age due to higher microbial biomass. However, previous studies reported that soil quality would not increase over the period of recovery age (An et al., 2009; Zhang et al., 2012). This study also found that stage increase rate of SQI for abandoned land decreased in the later stage of restoration, leading to relatively constant of SQI after 27 years of restoration (Fig. 4). The trends observed for soil quality in abandoned land are in accordance with the report of An et al. (2009), who found that soil nutrients and microbial properties all increased rapidly in earlier succession stages, up to 23 years, but then showed no significant fluctuations in later years. After the farmland has been abandoned, plant communities began to grow with the succession, leading to a large amount of organic matter in the soil and higher root density, which could improve soil quality, ultimately reaching a maximum due to species-specific characteristics after a decade (Cao et al., 2008). In general, the main plants change with the stage in succession. The dominant species are in the Compositae family in the primary succession stage, followed by Leguminosae in the middle stage of succession, and Gramineous in the later succession stage (Table 1) (An et al., 2009). Hence, there was an equilibrium state was exerted for soil quality after 27 years of restoration in the abandoned land. However, continual improvements of soil quality in forestland were observed with restoration age (Fig. 4). This finding is supported by previous studies, which suggested that the plant community tended to be stable in Robinia pseudoacacia L. land after N40 years of restoration (regarded as mature forest), accompanied by a rapid increases in the soil properties due to increasing net primary productivity in the earlier restoration stage on the Loess Hilly Region of China (Ai et al., 2014; Xu and Liu, 2004; Zhang et al., 2010). Thus, different trends in soil quality occurred for the two restoration pathways. Overall, we suggest that vegetation restoration significantly improve

814

S. Guo et al. / Science of the Total Environment 633 (2018) 808–815

Fig. 3. Contributions of selected soil indicators to overall soil quality index (SQI) under different land-use types. Note: Values are means ± SE. Means with different small letters indicate significant differences in abandoned land among the restoration age, whereas capital letters indicate a significant difference in Robinia pseudoacacia L land among the restoration age (p b 0.05). AL: Abandoned land, RP: Robinia pseudoacacia L.

soil quality, but improvements in soil quality would decrease in the later restoration age. 4.3. Differential responses of soil quality to revegetation type Different revegetation types can also have differential impacts on the development of soil quality (Zhang et al., 2011). Previous studies have shown that soil quality in farmland that had been abandoned for natural recovery was lower than that in afforested land (Fazhu et al., 2015; Ren et al., 2016a), although a few studies reported contrasting effects of these two revegetation types on soil quality (Zhang et al., 2011; Zhao

et al., 2017a). In the present study, higher SQI values were found for trees (RP) than for grassland (AL) at each recovery age (Fig. 4). This observation is consistent with previous studies in the same watershed (Ren et al., 2017; Ren et al., 2016b) and may be due to the relatively higher herbaceous coverage and fine root and leaf litter associated with the trees (Table 1), which can protect soil from erosion and enhance soil organic matter content (Deng et al., 2016). Meanwhile, higher organic matter content in the RP was beneficial for improving soil nutrient concentrations and biological properties, and stabilizing the soil structure (Figs. 1 and 2). Furthermore, compared with grass, trees have more developed ventilating tissues, contributing oxygen into the rhizosphere, and they are stimulating microbial population activity and enzyme synthesis, ultimately sustain higher microbial biomass and enzymes activity in the forestland (Caravaca et al., 2005; Zhao et al., 2017a). Therefore, it was not surprising that soil quality was higher in RP than in AL due to the relatively higher soil fertility, microbial biomass, enzyme activity and microstructure of soil aggregates at the same restoration age. In addition, we found that the stage increased rates in RP sites were higher than in AL sites in most cases (Fig. 4), illustrating that soil quality in RP sites had faster increase rate than in AL sites. The faster increase rate of SQI values in RP sites indicates that afforestation by Robinia pseudoacacia is more sensitive to land-use changes and has better capacity for soil recovery. Therefore, we suggested that RP rather than AL would be more suitable for soil restoration on the Loess Hilly Region of China. 5. Conclusion

Fig. 4. Restoration ratio of soil quality index (SQI) under different land-use types at the 0–10 and 10–20 cm soil layers.

We assessed soil quality after farmland conversion along two chronosequence using a soil quality index. Compared with farmland, there were notable differences in the physicochemical and biological properties of the restoration areas, which indicates that vegetation restoration had significant impacts on soil properties. The key indicators selected from soil properties for the SQI were MBC, FP, and TP. Our results showed that both recovery age and restoration pathway significantly influenced the improvement in soil quality (P b 0.05). Increasing recovery age corresponded to significant improvements in soil quality indicating increased resilience and restoration of these farmland areas following vegetation restoration. Stage increase rate of SQI for abandoned land decreased in the later stage of restoration, illustrating that improvements in soil quality decreased along with restoration process. SQI values were higher in RP than in AL at all recovery ages indicating that planting Robinia pseudoacacia L. rather than abandoning farmland for natural recovery is more suitable for soil restoration on the Loess Hilly Region of China.

S. Guo et al. / Science of the Total Environment 633 (2018) 808–815

Acknowledgments The authors especially thank Yadong Xu and Dexin Gao from Northwest A&F University for their help with the fieldwork and this work was supported by the National Natural Science Foundation of China (No. 41571501) and National Natural Key Research and Development Program of China (No. 2017YFC0504601). Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.scitotenv.2018.03.210. References Ai, Z., Chen, Y., Cao, Y., 2014. Storage and allocation of carbon and nitrogen in Robinia pseudoacacia plantation at different ages in the loess hilly region, China. Chin. J. Appl. Ecol. 25, 333–341. An, S., Huang, Y., Zheng, F., 2009. Evaluation of soil microbial indices along a revegetation chronosequence in grassland soils on the Loess Plateau, Northwest China. Appl. Soil Ecol. 41, 286–292. Anderson, T.-H., 2003. Microbial eco-physiological indicators to asses soil quality. Agric. Ecosyst. Environ. 98, 285–293. Andrews, S.S., Karlen, D.L., Mitchell, J.P., 2002. A comparison of soil quality indexing methods for vegetable production systems in northern California. Agric. Ecosyst. Environ. 90, 25–45. Armenise, E., Redmile-Gordon, M.A., Stellacci, A.M., Ciccarese, A., Rubino, P., 2013. Developing a soil quality index to compare soil fitness for agricultural use under different managements in the Mediterranean environment. Soil Tillage Res. 130, 91–98. Bai, Z., Dent, D., 2009. Recent land degradation and improvement in China. AMBIO J. Hum. Environ. 38, 150–156. Bastida, F., Zsolnay, A., Hernández, T., García, C., 2008. Past, present and future of soil quality indices: a biological perspective. Geoderma 147, 159–171. Brejda, J.J., Moorman, T.B., Karlen, D.L., Dao, T.H., 2000. Identification of regional soil quality factors and indicators: I. Central and southern high plains. Soil Sci. Soc. Am. J. 64, 2125–2135. Cao, C.Y., Jiang, D.M., Teng, X.H., Jiang, Y., Liang, W.J., Cui, Z.B., 2008. Soil chemical and microbiological properties along a chronosequence of Caragana microphylla Lam. plantations in the Horqin sandy land of Northeast China. Appl. Soil Ecol. 40, 78–85. Caravaca, F., Alguacil, M.M., Torres, P., Roldán, A., 2005. Plant type mediates rhizospheric microbial activities and soil aggregation in a semiarid Mediterranean salt marsh. Geoderma 124, 375–382. Deng, L., Wang, K.-B., Chen, M.-L., Shangguan, Z.-P., Sweeney, S., 2013. Soil organic carbon storage capacity positively related to forest succession on the Loess Plateau, China. Catena 110, 1–7. Deng, L., Wang, K., Tang, Z., Shangguan, Z., 2016. Soil organic carbon dynamics following natural vegetation restoration: evidence from stable carbon isotopes (δ13C). Agric. Ecosyst. Environ. 221, 235–244. Doran, J.W., Parkin, T.B., Doran, J.W., Jones, A.J., 1996. Quantitative Indicators of Soil Quality: A Minimum Data Set. pp. 25–37. Fazhu, Z., Jiao, S., Chengjie, R., Di, K., Jian, D., Xinhui, H., et al., 2015. Land use change influences soil C, N, and P stoichiometry under 'Grain-to-Green Program' in China. Sci. Rep. 5, 10195. Fu, X., Shao, M., Wei, X., Horton, R., 2010. Soil organic carbon and total nitrogen as affected by vegetation types in Northern Loess Plateau of China. Geoderma 155, 31–35. Karlen, D.L., Ditzler, C.A., Andrews, S.S., 2003. Soil quality: why and how? Geoderma 114, 145–156. Koerselman, W., Meuleman, A.F.M., 1996. The vegetation N:P ratio: a new tool to detect the nature of nutrient limitation. J. Appl. Ecol. 33, 1441–1450. Li, C., Zhao, L., Sun, P., Zhao, F., Kang, D., Yang, G., et al., 2016. Deep soil C, N, and P stocks and stoichiometry in response to land use patterns in the Loess Hilly Region of China. PLoS One 11, e0159075. Lucas-Borja, M.E., Fonseca, T., Linares, J.C., Morote, F.A.G., Serrano, F.R.L., 2012. Does the recruitment pattern of Spanish black pine (Pinus nigra Arn ssp. salzmannii) change the regeneration niche over the early life cycle of individuals? For. Ecol. Manag. 284, 93–99.

815

Masto, R.E., Chhonkar, P.K., Purakayastha, T.J., Patra, A.K., Singh, D., 2010. Soil quality indices for evaluation of long-term land use and soil management practices in semi-arid sub-tropical India. Land Degrad. Dev. 19, 516–529. Mukhopadhyay, S., Masto, R.E., Yadav, A., George, J., Ram, L.C., Shukla, S.P., 2016. Soil quality index for evaluation of reclaimed coal mine spoil. Sci. Total Environ. 542, 540–550. Nelson, D.W., Sommers, L.E., Sparks, D.L., Page, A.L., Helmke, P.A., Loeppert, R.H., et al., 1982. Total carbon, organic carbon, and organic matter. Methods of Soil Analysis Part—Chemical Methods, pp. 961–1010. Olsen, S.R., Le, S., 1982. Phosphorous. In: Page, A.L., Miller, R.H., Keeney, D.R. (Eds.), Methods of Soil Analysis Part 2, Chemical and Microbial Properties. 9. Agronomy Society of America, Agronomy Monograph, Madison, pp. 403–430. Raiesi, F., 2011. Soil properties and C dynamics in abandoned and cultivated farmlands in a semi-arid ecosystem. Plant Soil 351, 161–175. Raiesi, F., 2017. A minimum data set and soil quality index to quantify the effect of land use conversion on soil quality and degradation in native rangelands of upland arid and semiarid regions. Ecol. Indic. 75, 307–320. Raiesi, F., Kabiri, V., 2016. Identification of soil quality indicators for assessing the effect of different tillage practices through a soil quality index in a semi-arid environment. Ecol. Indic. 71, 198–207. Ren, C., Kang, D., Wu, J.P., Zhao, F., Yang, G., Han, X., et al., 2016a. Temporal variation in soil enzyme activities after afforestation in the Loess Plateau, China. Geoderma 282, 103–111. Ren, C., Zhao, F., Kang, D., Yang, G., Han, X., Tong, X., et al., 2016b. Linkages of C:N:P stoichiometry and bacterial community in soil following afforestation of former farmland. For. Ecol. Manag. 376, 59–66. Ren, C., Chen, J., Deng, J., Zhao, F., Han, X., Yang, G., et al., 2017. Response of microbial diversity to C:N:P stoichiometry in fine root and microbial biomass following afforestation. Biol. Fertil. Soils 53, 457–468. Rutigliano, F.A., D'Ascoli, R., Virzo De Santo, A., 2004. Soil microbial metabolism and nutrient status in a Mediterranean area as affected by plant cover. Soil Biol. Biochem. 36, 1719–1729. Schedlbauer, J.L., Kavanagh, K.L., 2008. Soil carbon dynamics in a chronosequence of secondary forests in northeastern Costa Rica. For. Ecol. Manag. 255, 1326–1335. Sparks, D.L., Page, A.L., Helmke, P.A., Loeppert, R.H., Soltanpour, P.N., Tabatabai, M.A., et al., 1996. Methods of Soil Analysis. Part III. Chemical Methods. Templer, P.H., Groffman, P.M., Flecker, A.S., Power, A.G., 2005. Land use change and soil nutrient transformations in the Los Haitises region of the Dominican Republic. Soil Biol. Biochem. 37, 215–225. Vance, E.D., Brooks, P.C., Jenkinson, D.S., 1987. An extraction method for measuring soil microbial biomass. Soil Biol. Biochem. 19, 703–707. Walker, T.W., Syers, J.K., Walker, T.W., Syers, J.K., 1976. The fate of phosphorus during pedogenesis. Geoderma. Geoderma 15, 1–19. Woods, K.D., 2000. Dynamics in late-successional hemlock-forests over three decade. Ecology 81, 110–126. Xu, M., Liu, G., 2004. The characteristics and evolution of soil nutrient in artificial black locust (Robinia pseudoacacia) forest land in the hilly Loess Plateau. Plant Nutrition and Fertilizer Science. 10, pp. 40–46. Yakovchenko, V., Sikora, L.J., Kaufman, D.D., 1996. A biologically based indicator of soil quality. Biol. Fertil. Soils 21, 245–251. Yu, Y., Jia, Q.Z., 2014. Changes in soil organic carbon and nitrogen capacities of Salix cheilophila Schneid along a revegetation chronosequence in semi-arid degraded sandy land of the Gonghe Basin, Tibet Plateau. Solid Earth 6, 2371–2399. Zhang, C., Liu, G., Xue, J., Song, Z., 2010. Evolution of soil enzyme activities of Robinia pseudoacacia plantation at different ages in Loess hilly region. Sci. Silvae Sin. 446, 23–29. Zhang, C., Xue, S., Liu, G.B., Song, Z.L., 2011. A comparison of soil qualities of different revegetation types in the Loess Plateau, China. Plant Soil 347, 163–178. Zhang, Z.-S., Dong, X.-J., Liu, Y.-B., Li, X.-R., Jia, R.-L., Hu, Y.-G., et al., 2012. Soil oxidases recovered faster than hydrolases in a 50-year chronosequence of desert revegetation. Plant Soil 358, 275–287. Zhao, D., Xu, M., Liu, G., Ma, L., Zhang, S., Xiao, T., et al., 2017a. Effect of vegetation type on microstructure of soil aggregates on the Loess Plateau, China. Agric. Ecosyst. Environ. 242, 1–8. Zhao, D., Xu, M., Liu, G., Yao, X., Tuo, D., Zhang, R., et al., 2017b. Quantification of soil aggregate microstructure on abandoned cropland during vegetative succession using synchrotron radiation-based micro-computed tomography. Soil Tillage Res. 165, 239–246.