Estimating the influence of related soil properties on macro- and micro-aggregate stability in ultisols of south-central China

Catena 137 (2016) 545–553

Contents lists available at ScienceDirect

Catena journal homepage: www.elsevier.com/locate/catena

Estimating the influence of related soil properties on macro- and micro-aggregate stability in ultisols of south-central China Jun-guang Wang a, Wei Yang b, Bing Yu a, Zhao-xia Li a, Chong-fa Cai a,⁎, Ren-ming Ma a a Key Laboratory of Arable Land Conservation (Middle and Lower Reaches of Yangtze River) of the Ministry of Agriculture, Soil and Water Conservation Research Centre, Huazhong Agricultural University, Wuhan 430070, People's Republic of China b Hubei Water Resources Research Institute, Wuhan 430070, People's Republic of China

a r t i c l e

i n f o

Article history: Received 31 March 2015 Received in revised form 3 November 2015 Accepted 4 November 2015 Available online xxxx Keywords: Macro- and micro-aggregates Stability Soil properties Ultisols

a b s t r a c t Background: Stable macro- or micro-aggregates are important for preventing soil degradation. The interactions among soil aggregates and stabilizing agents—like clay, soil organic matter (SOM), Fe, and Al oxides—are complex and have not been fully understood. Methods: Eight ultisol samples were collected from the surface (0–10 cm) and subsurface layers (10–20 cm). The macro-aggregate stability was determined by wet sieving, and the micro-aggregate distribution was determined via particle size distribution analysis; however, no chemical dispersant (sodium hydroxide) was applied. Using the PLSR models, the main soil properties that affect macro-aggregate and micro-aggregate stability were estimated. Results: All soils were strongly acidic (pH 4.28–5.56) with low SOM content (b20 g kg−1). The dithionite-citratebicarbonate extractable Fed and Ald were the dominant forms in Fe and Al oxides, much greater than acid ammonium oxalate extractable Feo and Alo. For most soils, the percentage of N5 mm aggregates was the highest, and the percentage of 2–1 mm aggregates was the lowest after wet sieving. Soil parent materials had a significant effect on the particle size distribution of the micro-aggregates. The stability of macro-aggregates and micro-aggregates from Quaternary red clay was stronger than that from Shale (p b 0.05). Regardless of the soil parent materials, the water stability of surface cropland soil macro-aggregates was significantly lower than that of the other land-use types, but the micro-aggregate stability exhibited no trend across different land use types. Conclusion: Ald was the most important binding agent of the macro-aggregates, and clay was the main binding agent of the micro-aggregates, followed by the Fed, Alo, CEC and SOM, while Feo was the weakest agent. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The stability of soil aggregates can affect water movement and storage in soils, and it can also influence soil aeration, erosion, and biological activities as well as crop growth. High and sustainable soil aggregate stability is thus an important characteristic for preserving soil productivity and restraining soil erosion and degradation (Amézketa, 1999). Macro-aggregate stability is often considered as a key index in studies of soil erodibility or soil degradation (Barthès and Roose, 2002; Cammeraat and Imeson, 1998; Idowu, 2003; Yan et al., 2008), and micro-aggregate stability is also used to estimate or predict soil erosion and surface runoff (Darboux and Le Bissonnais, 2007). The weak water stability of micro-aggregates has become a major cause of increasing

⁎ Corresponding author at: College of Resources and Environment, Huazhong Agricultural University, Wuhan 430070, People's Republic of China. E-mail addresses: [email protected] (J. Wang), [email protected] (W. Yang), [email protected] (B. Yu), [email protected] (Z. Li), [email protected], [email protected] (C. Cai), [email protected] (R. Ma).

http://dx.doi.org/10.1016/j.catena.2015.11.001 0341-8162/© 2015 Elsevier B.V. All rights reserved.

water erosion on agricultural lands in some areas of Africa (Opara, 2009). The stability of soil macro- and micro-aggregates can be affected by dozens of different soil intrinsic factors and depends on soil formation processes, biological factors, agricultural management and climate (Amézketa, 1999). Improving our understanding of these processes is fundamental in order to take action to maintain macro- and microaggregate stability in soils. Some researchers have found that soil intrinsic factors—such as the content of clay, organic matter, and oxides—contributed to the cohesive strength of aggregates through binding processes in their selected soils (Levy and Mamedov, 2002; Wuddivira et al., 2006). Soil organic matter is a crucial component for aggregate formation and stabilization (Boix-Fayos et al., 2001; Six et al., 2004; Noellemeyer et al., 2008), and is positively correlated with macro-aggregate distribution (Green et al., 2005; De Gryze et al., 2008; Huang et al., 2010). Boix-Fayos et al. (2001) observed that the macro-aggregate stability depends on the organic matter content when this was greater than 5–6%, and Huang et al. (2010) showed that the percentages of N 0.25 mm water-stable aggregates correlated well with the organic matter contents of eroded ultisols. However,

546

J. Wang et al. / Catena 137 (2016) 545–553

some researchers have shown that soils with higher contents of Fe and Al oxides tend to have more stability for macro- and micro-aggregates (Goldberg et al., 1990; Pinheiro-Dick and Schwertmann, 1996). Igwe et al. (1995, 2009) and Li et al. (2005) showed that in some Inceptisols and ultisols the low organic-matter content did not affect the formation and stabilization of aggregates. They claimed that sesquioxides were ultimately responsible for aggregate formation and stabilization, and their contributions to aggregate formation could vary greatly. There has since been discussion as to which particular element in the sesquioxide group was responsible for aggregate formation. Among sesquioxides, Al was found to be more effective than Fe as an aggregating agent (Keren and Singer, 1990; Mbagwu and Schwertmann, 2006). However, Arca and Weed (1966) indicated that aggregation in oxide-rich soils was more strongly correlated with free Fe oxide content than other soil properties. Other researchers identified oxalate-extracted oxides (Fe, Al) as heavily involved in aggregate formation in some Oxisols, Inceptisols, and Andisols from Brazil, Cameroon, and Chile (Pinheiro-Dick and Schwertmann, 1996; Huygens et al., 2005). Ultisols (more commonly known as “red soils”) cover approximately 1.14 million km2 in southeastern China and are the dominant soil type in South American and Southeastern Asia. Improper land use and soil management, undulating topography, and poor soil properties may have caused severe soil erosion in ultisols areas, which has become one of the most challenging environmental problems in China (Zhao et al., 2000; Zhang et al., 2004). Aggregate water stability is clearly an important characteristic affecting soil erosion (Yan et al., 2008; Wang et al., 2012). Although a number of studies have examined the relationship between chemical properties and macro-aggregate stability in ultisols from subtropical regions, the results have been inconclusive (Huang et al., 2010; Li et al., 2005; Zhang and Horn, 2001). Some studies have indicated that micro-aggregate stability in some ultisols and inceptisols from southeastern Nigeria are affected by iron and aluminum oxides or different land use types (Igwe et al., 2009; Opara, 2009). However, little information exists on the micro-aggregate size distribution and stability in subtropical soils and compares this with the research on macro-aggregates. This study aims to (1) identify the size distribution and stability of the soil macro- and micro-aggregates from subtropical China, (2) estimate the influence of related soil physical and chemical properties on macro- and micro-aggregate stability, and (3) evaluate their roles in the structural stability of these soils. 2. Materials and methods 2.1. Soils Soil samples were collected from Xianning City, Hubei Province in the south-central part of China (Fig. 1.). The sampling sites were in the subtropical zone, which has an annual rainfall of 1572 mm and annual average temperatures of 16.8 °C. The elevation of this area is 32–52 m above sea level. These areas are hilly with different degrees of erosion. Soil erosion in some areas of the ultisol region reaches 7000 t km− 2 per year (He and Sun, 2008). Eight soil samples were collected from the surface layer (0–10 cm) and eight from a subsurface layer (10– 20 cm). Each sampled area was about 0.1 to 0.3 ha. This number of samples is representative for the study area. The eight sample sites were geographically representative for the soils in Hubei Province. They covered the most soil parent materials in subtropical China. Considering the soil parent materials and different land use, we selected the eight soil samples derived from Quaternary red clay and Shale from Xianning City. General characteristics of these sampling sites are shown in Table 1. 2.2. Laboratory methods The wet sieving method of Yoder (1936) was modified to measure macro-aggregate stability. Air-dried soil samples were sieved by hand

Fig. 1. Location of the study area and sampling sites: (a) location of Hubei Province in China, (b) location of Xianning County in Hubei Province, (c) location of Heshengqiao town in Xianning County.

using a column of five sieves at sizes of 5.00, 2.00, 1.00, 0.50 and 0.25 mm. The percentages of different aggregates in the bulk soil as determined with the dry-sieving method were calculated and recorded. Based on these percentages, soil samples of different aggregate sizes were prepared. Approximately 50 g of composed soil samples was placed on the first sieve of a sequential nest and gently moistened to avoid sudden rupture of the aggregates. The sample was then sieved in distilled water at 30 oscillations per minute (along 4 cm amplitude) for 30 min. The resistant soil materials on each sieve were transferred into clean beakers. These soil materials were then oven-dried gently at 40 °C for 48 h and weighed. Macro-aggregate stability was expressed by the content of water-stable aggregates that were greater than 0.25 mm (WSAN0.25), mean weight diameter (MWD), and the aggregate deterioration rate (ADR) by wet sieving. Equations used in this research include: MWD ¼

7 X i¼1

xi wi

J. Wang et al. / Catena 137 (2016) 545–553

547

Table 1 Basic description of selected soils. Sample designation

Parent material

Sample depth (cm)

Altitude (m)

Slope

Orientation

Land use

Qc

Quaternary red clay

0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20

40

10–14%

North

Cropland

48

15–18%

West

Orchard

41

18–20%

West

Prunus × cistena, upland

38

10–13%

East

Tea garden

32

15–18%

North

Cropland

46

17–20%

West

Orchard

41

17–20%

East

Prunus × cistena, upland

52

15–18%

Southwest

Tea garden

Qo Qp Qt Sc

Shale

So Sp St

where xi is the mean diameter of each size class (b0.05 mm; 0.05–0.10 mm; 0.10–0.25 mm; 0.25–0.50 mm; 0.50–1.00 mm; 1.00– 2.00 mm; N2.00 mm), and wi is the proportion of each size class with respect to the total sample. The upper and lower limits of mean intersieve aperture are 3.5 and 0.05 mm, respectively.

ADR ¼

C d −C w  100 Cd

where Cd (%) is the content of dry aggregates, of which the size is greater than 0.25 mm, and Cw (%) is the content of water-stable aggregate, of which the size is greater than 0.25 mm. Particle size distribution was measured by the hydrometer method (Gee and Bauder, 1986). The micro-aggregate distribution was determined as particle size distribution analysis above except that there was no chemical dispersant (sodium hydroxide) applied (only mechanical agitation with an end-over-end shaker). Clay dispersion ratio (CDR) and aggregated silt and clay (ASC) were used to express the microaggregate stability as described by Dong et al. (1983) and Igwe et al. (1999). They were obtained using the following equations:  CDR ¼ W c C c ASC ¼ ð%C c þ %C s Þ−ð%W c þ %W s Þ

where Wc (%) and Ws (%) are the content of clay (b0.002 mm) and silt (0.05–0.002 mm) from the micro-aggregate stability determination, Cc (%) and Cs (%) are the content of clay (b 0.002 mm) and silt (0.05– 0.002 mm) from the particle size distribution measured by the hydrometer method, respectively. The more stable the micro-aggregates, the lower the CDR value or the higher the value of the ASC (%). The basic chemical properties of the soils were determined using the routine methods. Soil pH was measured with a pH meter in a 1.0:2.5 soil:water mixture. Soil organic matter (SOM) was determined using the K2Cr2O7 wet oxidation method (Jackson, 1958). Cation-exchange capacity (CEC) was determined using the ammonium acetate method buffered at pH 7 (Sumner and Miller, 1996). Free Fe and Al oxides (Fed and Ald) were extracted using dithionite-citrate-bicarbonate (DCB) solution (Mehra and Jackson, 1960). The extraction of amorphous Fe and Al oxides (Feo and Alo) was carried out using oxalic acid ammonium oxalate at pH 3.0 (Mckeague and Day, 1966). Pyrophosphate Fe and Al (Fep and Alp) were extracted using sodium pyrophosphate at pH 10 (Bascomb, 1968). The extracted Fe and Al were diluted and determined by an inductively coupled plasma spectrometer (ICP) (VISTA-MPX, Varian, Inc., USA).

2.3. Data analysis Statistical analyses were performed using Excel and SPSS 11.0 software (Statistical Package for the Social Sciences, 2001). Significant differences among the soil samples for macro-aggregate stability were determined using the LSD (Least Significant Difference) procedure for a multiple range test at the 0.05 significance level. Partial least squares regression (PLSR) was used to determine the relationships among related soil properties and macro- and microaggregate stability. In this study, the relationships among the predictors (related soil properties) and the response function (the macro- and micro-aggregate stability) can be inferred from the variable importance for the projection (VIP) and regression coefficients (RCs) of individual predictors in the most explanatory components. The VIP was used to identify the importance of a predictor for both the independent and dependent variables. Terms with large VIP values are the most relevant for explaining the dependent variable. In general, an independent-variable VIP value greater than 1 significantly explains the dependent variable, while a value lower than 0.5 indicates that the independent variable does not significantly explain the dependent variable. The regression coefficients of the PLSR models were used to show the direction of the relationship between the related soil properties and macro- and microaggregate stability. Thus, it is possible to determine which related properties most strongly interact with macro- and micro-aggregate stability. The analyses were performed using the PLSR procedure implemented in SIMCA-P (Umetrics AB. Sweden). The cross-validation was enforced by using the criterion to determine among the number of significant PLSR components. The fraction of the total variation of the dependent variables that could be predicted by a component (Q2) and the cumulative Q2 over all the selected PLSR components (Q2cum) in SIMCA were computed according to the following formula: Q 2 ¼ 1:0−PRESS=SS Q cum 2 ¼ 1:0−∏ðPRESS=SSÞaða¼1;2…mÞ where PRESS is the prediction error sum-of-squares, SS is the residual sum-of-squares, and m is the number of PLSR components. A model is thought to have a good predictive ability when Q2cum is greater than 0.5. In addition, the root mean square error (RMSEE) also provides useful information for calibrating and developing the regression model. 3. Results 3.1. Basic properties of soils Table 2 shows the physical and chemical properties of the soil samples. The sand contents of all the soils from the two parent materials

548

J. Wang et al. / Catena 137 (2016) 545–553

Table 2 Selected physical and chemical properties of the tested soils.a Sample designation

Qc Qo Qp Qt Sc So Sp St

Sample depth (cm)

0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20

Particle size distribution (%) Sand

Silt

Clay

7 7 8 6 9 8 5 5 11 11 8 7 13 8 5 6

47 46 53 55 37 42 53 49 64 62 59 61 63 66 62 60

46 47 39 39 54 50 42 46 25 27 33 32 24 26 33 34

Texture

pH

SOM (g kg−1)

CEC (cmol kg−1)

Fed (g kg−1)

Ald (g kg−1)

Feo (g kg−1)

Alo (g kg−1)

Silty clay

4.28 4.63 4.39 4.32 4.51 4.59 4.39 4.73 5.30 5.56 4.98 5.29 4.80 5.10 4.52 4.81

15.34 13.77 17.15 11.85 21.16 14.46 17.55 13.45 12.53 13.84 12.61 13.37 16.67 12.05 16.35 13.13

15.63 14.73 17.78 15.18 19.38 15.30 15.18 13.00 12.40 13.23 13.25 11.48 16.28 12.35 14.35 13.08

18.00 21.50 17.14 17.61 21.88 23.07 17.00 18.53 13.21 8.75 13.22 11.90 14.96 13.93 14.37 13.73

2.21 2.54 2.39 2.37 2.16 2.65 2.65 2.44 1.41 0.89 1.53 1.41 1.83 1.56 1.58 1.09

1.60 1.74 1.38 1.64 1.58 1.68 1.75 1.95 1.76 1.74 1.39 1.55 1.40 1.62 2.06 2.07

1.04 1.68 1.73 1.96 1.33 1.62 1.94 2.08 1.00 0.94 1.09 0.96 0.95 0.98 1.42 1.32

Silty clay Loam Clay Silty clay Silty loam Silty clay loam Silty loam Silty clay loam

a Sand (2–0.05 mm), silt (0.05–0.002 mm), clay (b0.002 mm). SOM: soil organic matter. CEC: Cation-exchange capacity. Fed and Ald: extracted with dithionite-citrate-bicarbonate. Feo and Alo: extracted with oxalic acid ammonium oxalate.

were the lowest (b 20%) compared with the silt and clay contents. The texture of the soils ranged from silt loam to clay, and the clay content was higher in soils derived from Quaternary red clay than that from the soils derived from Shale with the same land use. All soils were strongly acidic with pH ranging from 4.28 to 5.56. Soil organic matter content in most soil samples was less than 20 g kg−1 except Prunus × cistena upland from Quaternary red clay. Cation exchange capacity values of the studied soils ranged from 11.48 cmol kg− 1 to 19.38 cmol kg−1, showing a minimal difference among them. The contents of SOM from the Prunus × cistena upland (Qp and Sp) surface soils were the highest, followed by tea gardens and orchards with the lowest content appearing in the cropland. The low organic matter of the cropland might be associated with frequent soil tillage for a prolonged period (Zhang and Xu, 2005). The relatively low cation exchange capacity of the soils could be attributed to the leaching process following high intensity rainfall in the region. The dithionite- and oxalate-extracted Fe and Al are shown in Table 2. Dithionite Fe and Al in all soils were higher than were the oxalateextracted forms. Parent materials apparently had a strong influence on the contents of the dithionite extractable Fe and Al in soils: dithionite extractable Fe and Al ranged from 17.00–23.07 g kg− 1 and 2.16– 2.65 g kg−1 from Quaternary red clay whereas they were from 8.75– 14.96 g kg−1 and 0.89–1.83 g kg−1 from Shale. The difference of the

Table 3 The distribution of macro-aggregate of the tested soils by wet sieving method. Sample Sample Fragment size distribution (%) designation depth (cm) N5 mm 5–2 mm 2–1 mm 1–0.5 mm 0.5–0.25 mm Qc Qo Qp Qt Sc So Sp St

0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20

18.16 17.65 41.93 36.71 25.62 32.42 39.67 16.47 5.49 2.51 16.75 22.14 7.24 0.00 10.25 0.74

4.09 9.34 5.19 10.37 12.20 12.76 5.71 4.10 1.44 0.84 1.55 2.19 1.34 0.83 2.82 1.18

4.65 7.05 3.30 6.91 6.52 7.78 2.75 3.26 0.94 0.61 1.07 1.72 1.08 1.08 1.56 0.96

8.19 14.11 5.93 11.63 10.33 11.57 6.03 7.36 1.79 1.46 1.95 4.26 3.00 2.16 3.15 3.92

5.01 14.30 7.22 7.73 14.00 9.25 8.97 8.19 1.64 4.35 3.62 7.03 7.81 4.28 7.74 13.57

oxalate extractable Fe and Al in all soils was not obvious. Oxalate extractable Fe and Al ranged from 1.38–1.95 g kg−1 and 1.04– 2.08 g kg−1 from Quaternary red clay, and they were from 1.39– 2.07 g kg−1 and 0.94–1.42 g kg−1 from Shale, respectively.

3.2. The size and stability of macro- and micro-aggregate Table 3 shows the percentage of N 5 mm aggregates was the highest except for the Sp (10–20 cm) sample, and the percentage of 2–1 mm aggregates was the lowest except for the Qc (0–10 cm) and Sp (10–20 cm) samples. Little difference in percentages was found among macroaggregates of the three sizes (5–2 mm, 1–0.5 mm and 0.5–0.25 mm) for most soils. These results differed from those of Horn (1990), who concluded that the large aggregates (N 3 mm) were broken and small aggregates (0.25–3 mm) were significantly increased. The percentages of all five size fractions of the macro-aggregates from Quaternary red clay were greater than those from Shale, suggesting that the macro-aggregates from Shale suffered more damage after wet sieving. Regardless of the parent materials, the percentages of N5 mm aggregates from the orchard samples (Qo and So) were the highest whereas the percentages from the cropland (Qc and Sc) were the lowest among different land uses. The results of the aggregate stability from the wet sieving method were slightly different due to the choice of the stability indicators (Fig. 2). The value of the ADR was inversely proportional to that of the WSA, and the higher the ADR value, the worse the aggregate stability. For the soils in this study, the values of WSA, MWD and ADR ranged from 8.34% to 73.80%, 0.08 mm to 2.42 mm, and 23.30% to 91.49%, respectively. For similar land uses of the two parent materials, the water stability of the aggregates from Quaternary red clay was stronger than that from the Shale (p b 0.05). In this study, the stability among surface and subsurface soil aggregates showed no significant variation. The particle size distribution of the micro-aggregates is shown in Fig. 3. Because of the different soil parent materials, the particle size distribution of the micro-aggregates was significantly different. The percentage of 0.25–2 mm aggregates from Quaternary red clay (51.85% to 65.18%) was obviously higher than that from Shale (9.35% to 38.00%). The percentage of 0.05–0.25 mm particles from Quaternary red clay was the highest (ranged from 18.35% to 28.20%), followed by 0.05–0.02 mm (5.02% to 12.52%) and b0.002 mm (3.82% to 5.82%). However, the percentage of 0.02–0.002 mm aggregates from Shale was the highest (20.00% to 38.14%), followed by 0.25–0.05 mm (14.00% to 30.45%) and b0.002 mm (4.92% to 10.48%). Regardless of

J. Wang et al. / Catena 137 (2016) 545–553

549

Fig. 2. The content of water stable aggregates N0.25 mm (WSAN0.25), mean weight diameter (MWD) and aggregate deterioration rate (ADR) by wet sieving. Bars with the same letter are not significantly different at the 0.05 probability level by LSD.

the soil parent material, the difference of the aggregates size distribution among land use types or soil depth was insignificant. Macro-aggregates were composed of micro-aggregates, and different size aggregates had different aggregation and stability mechanisms (Oades and Waters, 1991). The micro-aggregate stability indices of the soils are presented in Table 4. The clay dispersion ratio (CDR) was used to estimate the instability of the micro-aggregates whereas aggregated silt and clay (ASC) followed a contrary tendency to that of CDR. The results of CDR and ASC analyses showed that the micro-aggregate stability of the two parent materials was significantly different under the same land use types or soil layers. The values of CDR and ASC from Quaternary red clay ranged from 0.08 to 0.15 and 70.44 to 78.84%%, respectively; by contrast, the values from Shale ranged from 0.16 to 0.38 and 19.90% to 58.74%, respectively. Among the land use types, the micro-aggregate stability was significantly different, but they exhibited no trend under the two parent materials or soil layers.

3.3. Macro- and micro-aggregate stability versus related soil properties The correlation coefficients of the individual soil properties are listed in Table 5. A preliminary analysis indicated high correlation among most soil properties. The results found that the basic soil properties (silt, clay and CEC) were well correlated with different forms of iron and aluminum (Fed, Ald, and Alo). The SOM was well correlated with CEC. However, the Feo seemed to be more independent than the other soil properties. The silt was found to be negatively correlated with the other soil properties (clay, CEC, Fed, Ald, and Alo). Furthermore, the silt and clay were more sensitive to Fed and Ald than to Alo whereas the CEC was more sensitive to Fed and Ald. Among the different forms of iron and aluminum, Fed was more sensitive to Ald than to Alo. Table 6 summarizes the five PLSR models constructed separately for macro- and micro- aggregate stability (WSAN0.25, MWD, ADR, CDI, and ASC). The macro-aggregate characteristic (WSAN0.25, MWD, and ADR)

550

J. Wang et al. / Catena 137 (2016) 545–553

Fig. 3. The particle size distribution of the micro-aggregates.

models extracted one PLSR component that was relevant to 8 predictor variables (related soil properties). The first component accounted for 80.65%, 69.17%, and 82.19% of the variance in the dataset with WSAN0.25, MWD, and ADR, respectively. The addition component cumulatively accounted for 84.87%, 80.41% and 85.93% of the total variance in the WSAN0.25, MWD, and ADR values, respectively. However, the cumulative Q2 (Q2cum) values decreased from 0.75, 0.60 and 0.77 to 0.72, 0.55, and 0.75, respectively (Table 6), which suggested that it did not substantially improve the description of the contributions to the variance. The relative importance of the related soil properties could be obtained by exploring their VIP values and their regression coefficients (RCs) (Table 7). In the case of macro-aggregate stability, the highest VIP value was obtained for Ald from the three indicators, followed by clay, Fed, silt, and Alo for the WASN0.25; clay, silt, Alo, and Fed for the MWD; and clay, silt, Fed, and Alo for the ADR. Macro-aggregate stability appeared to increase with increasing Ald, clay, Fed, and Alo (due to the positive regression coefficients for the WSAN0.25 and MWD, and the negative regression coefficients for the ADR); silt contributed to lower stability. All of the considered soil properties were related to macroaggregate stability to some extent, and some of them had VIP N 1. Others with VIP values below 1 (CEC, SOM, and Feo) were of minor importance for macro-aggregate stability. The first component accounted for 83.91% of the variance in the dataset with CDI of the micro-aggregate stability. The addition component cumulatively accounted for 88.36% of the total variance, and the cumulative Q2 (Q2cum) values increased from 0.83 to 084. However, Table 4 Parameters of micro-aggregate stability of the studied soils.a Sample designation

Sample depth (cm)

CDR

ASC(%)

Qc

0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20 0–10 10–20

0.11f 0.09 fg 0.10 fg 0.15e 0.08 g 0.10 fg 0.11f 0.09 fg 0.38a 0.28c 0.30c 0.33b 0.28c 0.30c 0.20d 0.16e

78.84a 75.48b 70.74 cd 71.34 cd 79.28a 75.50b 70.44d 73.46bc 39.44 h 45.25 g 36.54i 41.26 h 19.90 k 24.60j 52.26e 58.74f

Qo Qp Qt Sc So Sp St a

CDR: Clay dispersion ratio. ASC: Aggregated silt and clay. Values within a column followed by the same letter are not significantly different at the 0.05 probability level by LSD.

the best PLSR prediction was also described by the lowest RMSEE. The addition component did not produce a significant change in Q2cum and RMSEE. For the CDI model, the RMSEE minimum value was achieved with one component, accounting for 83.91% of the CDI. For the ASC model, the RMSEE minimum value was achieved with two components. The addition of the second component significantly improved the model (Table 6). In the micro-aggregate stability models, the highest VIP value was observed for the clay content, followed by the silt, Fed, Ald, and Alo for the CDI, and silt, Fed, Alo, and Ald for the ASC. Other properties were of minor importance for micro-aggregate stability (VIP b 1). Using PLSR modeling, we explored the main soil properties that affect macro- and micro-aggregate stability. The results showed that Ald was superior in binding the macro-aggregates and clay was superior in binding the micro aggregates, followed by Fed, Alo, CEC and SOM; Feo was the weakest contributor to the formation of macro- or micro aggregates. 4. Discussion 4.1. Macro- and micro-aggregate stability of the ultisols After the wetting treatments, the percentage of aggregates N 5 mm from most soils was the highest and the percentage of aggregates around 2–1 mm was the lowest. Zhang and Horn (2001) reported that fragments of 2–1 mm or 0.63–0.2 mm were the dominant fractions in ultisols in China. Huang et al. (2010) found that wet sieving in the tested ultisols resulted in little difference among the different particle size fractions in terms of percentages. For the aggregates from the microaggregate stability determination, the percentage of 0.05–0.25 mm particles from Quaternary red clay was the highest whereas the highest percentage from Shale was found in the fraction of 0.02–0.002 mm particles. Peng et al. (2015) indicated that the size of major proportion of ultisols under long-term fertilization was 0.053–0.25 mm, followed Table 5 Correlation matrix of the independent variables used in the PLSR analysis.

Silt Clay SOMb CEC Fed Ald Feo Alo a

Silt

Clay

SOM

CEC

Fed

Ald

Feo

Alo

1 −0.98** a −0.49 −0.64** −0.89** −0.75** 0.05 −0.52*

1 0.43 0.56* 0.87** 0.76** 0.08 0.62*

1 0.82** 0.40 0.34 0.17 0.12

1 0.62* 0.55* −0.36 0.29

1 0.87** −0.04 0.60*

1 −0.17 0.75**

1 0.28

1

*, p b 0.05. **, p b 0.01. b SOM: soil organic matter. CEC: Cation-exchange capacity. Fed and Ald: extracted with dithionite-citrate-bicarbonate. Feo and Alo: extracted with oxalic acid ammonium oxalate.

J. Wang et al. / Catena 137 (2016) 545–553

551

Table 6 Summary of the PLSR models of macro- and micro-aggregate stability models. Response variable in Y

R2

Q2

Component

% Of explained variability in Y

Cumulative explained variability in Y(%)

Q2cum

RMSEE

WSAN0.25 a

0.81

0.75

MWD

0.69

0.60

ADR

0.82

0.77

CDI

0.84

0.83

ASC

0.85

0.73

1 2 1 2 1 2 1 2 1 2

80.65 4.22 69.17 11.24 82.19 3.74 83.91 4.45 76.92 7.77

80.65 84.87 69.17 80.41 82.19 85.93 83.91 88.36 76.92 84.69

0.75 0.72 0.60 0.55 0.77 0.75 0.83 0.84 0.71 0.73

10.91 10.01 0.49 0.41 10.94 10.09 0.04 0.04 10.00 8.45

a WSA0.25: the content of N0.25 mm water-stable aggregate; MWD: mean weight diameter of wet-sieving; ADR: aggregate deterioration rate. CDR: Clay dispersion ratio. ASC: Aggregated silt and clay.

by the silt-sized fraction (0.002–0.053 mm), and the clay-sized fraction (b0.002 mm) of less than 7% following water extraction. The WSAN0.25, MWD, and ADR were the parameters for estimating the macro-aggregate stability; the CDI and ASC were the parameters for estimating the dispersibility of the micro-aggregate. Our results showed that the stability of both macro- and micro-aggregates from Quaternary red clay was stronger than that from the Shale (Fig. 2. and Table 4). Yan et al. (2008) indicated that the aggregates from the Shale had the least stability compared with other parent materials. Regardless of the soil parent materials, the water stability of surface cropland soil macro-aggregates was significantly lower than that from other land-use types (Fig. 2.), possibly due to the long tradition of farming (Zhang et al., 1997). For the micro-aggregate stability, there was no significant trend among land use types. Lu et al. (2014) also indicated that land use type had no obvious influence on the volumes of macro-, meso-, and micro-pores in ultisols. Compared with other land use types, Opara (2009) concluded that the ultisols from Nigeria under natural forest (NF) exhibited significantly a lower clay dispersion ratio (CDR) and higher aggregated silt and clay as well as higher ASC (%) when compared with other land use types, suggesting greater microaggregate stability. Shrestha et al. (2007) showed that forest soils in Nepal had more macro- than micro-aggregates whereas cultivated soils had a greater proportion of micro- than macro-aggregates. 4.2. Influence of related soil properties in the macro- and micro-aggregate stability of the soils In tropical and subtropical soils, the SOM and sesquioxides are widely regarded as the main organic and inorganic binding agents of soil aggregation; their roles in soil aggregation have been investigated intensively (Denef et al., 2002; Six et al., 2004; Bronick and Lal, 2005; Kögel-Knabner et al., 2008; Pronk et al., 2012; Martins et al., 2013). However, it is difficult to assess their contributions to aggregation of the soil. In the oxide-rich soil, Fe and Al oxides may act as binding agents when organic materials are involved in adsorption on oxide surfaces

(Imhoff et al., 2002; Six et al., 2002). Six et al. (2004) reported that oxides can act as binding agents in three ways: organo-mineral complexes, electrostatic binding between positively charged oxides and negatively charged clay minerals, and a coat of oxides on the surface of minerals. Our results revealed that certain forms of sesquioxides might be responsible for macro- and micro-aggregate stability of the soils. Ald, Alo, and Fed were very important aggregating agents (Table 7). Soil oxides acted as important binding agents in aggregation in ultisols from subtropical China (Zhang and Horn, 2001). Mbagwu and Schwertmann (2006) insisted that Al oxides were better aggregating agents than Fe oxides in some tropical soils. In our study, Ald demonstrated prominent ability in aggregation of macro-aggregates. Li et al. (2005) implied that the stability parameters had significant relationships to Fed and Ald due to the relatively high content of these oxides in these soils. The results by Pinheiro-Dick and Schwertmann (1996) revealed that the oxalate-extractable forms of Fe oxide might be key components for aggregation in some tropical Oxisols and Inceptisols, and Duiker et al. (2003) showed that despite little effect of Fed content on soil aggregation, Feo and organic carbon contents were well correlated with WSA and MWD. However, in our study, the contribution of the Feo to macro- and micro-aggregate stability was among the weakest, determined by the VIP values and regression coefficients (Table 7). We postulate that the Fe and Al alone may not play crucial roles in soil aggregation; however, in association with SOM or clay, they may be indispensable components in soil aggregation. Although the importance and contribution of SOM to aggregate formation and stability has been well reported (Six et al., 2004; Onweremadu et al., 2007; Wagner et al., 2007; Huang et al., 2010), it was not superior in binding the macro- and micro-aggregates in this study (VIP b 1). Mbagwu and Schwertmann (2006) and Igwe et al. (2009) stated that organic matter in some of their soils acted as disaggregating agents or was not correlated with aggregate stability, but Plante and McGill (2002) observed that macro- and micro-aggregation could be improved by increased SOM content. Peng et al. (2015) concluded that the Fe/Al oxides seemed to be major agents of b 0.25 mm

Table 7 VIP values and regression coefficients for macro- and micro-aggregate stability models. WSAN0.25a

Silt Clay SOM CEC Fed Ald Feo Alo

MWD

ADR

CDI

ASC

Regression coefficients

VIP

Regression coefficients

VIP

Regression coefficients

VIP

Regression coefficients

VIP

Regression coefficients

VIP

−0.171 0.174 0.085 0.134 0.172 0.179 −0.038 0.151

1.173 1.189 0.581 0.918 1.176 1.228 0.262 1.035

−0.149 0.152 0.084 0.129 0.148 0.178 −0.064 0.149

1.100 1.118 0.613 0.944 1.084 1.300 0.468 1.091

0.172 −0.175 −0.090 −0.137 −0.172 −0.180 0.037 −0.151

1.171 1.188 0.614 0.929 1.168 1.224 0.254 1.025

0.178 −0.190 −0.100 −0.135 −0.171 −0.163 −0.040 −0.160

1.195 1.272 0.665 0.903 1.142 1.086 0.268 1.069

−0.241 0.287 −0.012 0.036 0.112 0.103 0.221 0.222

1.246 1.335 0.594 0.778 1.080 1.017 0.667 1.030

a WSA0.25: the content of N0.25 mm water-stable aggregate; MWD: mean weight diameter of wet-sieving; ADR: aggregate deterioration rate. CDR: Clay dispersion ratio. ASC: Aggregated silt and clay. SOM: soil organic matter. CEC: Cation-exchange capacity. Fed and Ald: extracted with dithionite-citrate-bicarbonate. Feo and Alo: extracted with oxalic acid ammonium oxalate.

552

J. Wang et al. / Catena 137 (2016) 545–553

aggregates in the ultisols whereas the SOM played a primary role in stabilizing larger aggregates (0.25–2.00 mm). Although Fe/Al oxides were the predominant binding agents in ultisols, the role of SOM in aggregation can be improved by increasing C input (Peng et al., 2015). Igwe et al. (1995) showed that the role of SOM as an aggregating agent in soils may significantly depend on the quantity of organic matter in the soil and the type of soils. In our study, soil organic matter content was mostly less than 20 g kg− 1 except in Prunus × cistena upland from Quaternary red clay (Table 2). Clay can act as a cementing material that holds particles together in the aggregate (Boix-Fayos et al., 2001). Lado et al. (2004) showed that aggregate stability increased with an increase in clay content in the range between 9% and 63% clay, and the increase in aggregate stability was due to the cementing effect of the clay. In our tested soils, the range of the clay content was 24% to 54%, and the results confirmed that clay was an important aggregating agent, especially for microaggregate stability (Table 7). Sumner (1992) found that a large total surface area of Fe and Al oxides might facilitate reactions with clay particles through Coulombic interactions. Wuddivira and Camps-Roach (2007) suggested that the structural stability of soils dominated by kaolinitic clays increased with increasing clay content, but the soils dominated by smectitic clays succumbed easily to slaking forces under fast wetting when clay content increased. Ultisols are known for their high content of 1:1 clay minerals (e.g. kaolinite) and oxides (Lu et al., 2014); Zhou et al. (2013) indicated that electrostatic interactions between oxides and 1:1 clay minerals can lead to aggregate formation through mineral–mineral bonding. Iron oxides, rather than SOM, have been reported to be the dominant cementing agents for aggregation in ultisols (Lu et al., 2014). Our results indicated that clay, Fed, Ald, and Alo were effective in binding aggregates and contributed significantly to macro- and micro-aggregate stability. 5. Conclusion Due to different parent materials, the percentages of all five size fractions (N 5 mm, 5–2 mm, 2–1 mm, 1–0.5 mm, and 0.5–0.25 mm) of the macro-aggregates from Quaternary red clay were higher than those from Shale, and the particle size distribution of the micro-aggregates was significantly different. The stability of the macro- and microaggregates from Quaternary red clay was stronger than that from the Shale. Regardless of the soil parent materials, the water stability of surface cropland soil macro-aggregates was significantly lower than that in other land-use types. For micro-aggregate stability, there was no significant regularity among land-use types. Ald was the most prominent aggregating agent for macroaggregates, and clay was superior in binding the micro-aggregates. The contribution of the Feo to macro- and micro-aggregate stability was the weakest, suggested by VIP values and regression coefficients from PLSR models. The role of SOM in the aggregation of these soils was also not clear, and this could be due to low content in these soils. Our results indicated that the clay, Fed, Ald, and Alo were effective aggregating agents and contributed significantly to macro- and microaggregate stability. Acknowledgments We appreciate anonymous reviewers for improving our paper. Financial support for this research was provided by the National Natural Science Foundation of China (41401303; 41401317) and the Natural Science Foundation of Hubei Province (2014CFA016). References Amézketa, E., 1999. Soil aggregate stability: a review. J. Sustain. Agric. 14 (2), 83–151. Arca, M.N., Weed, S.B., 1966. Soil aggregation and porosity in relation to contents of free iron oxide and clay. Soil Sci. 101, 164–170.

Barthès, B., Roose, E., 2002. Aggregate stability as an indicator of soil susceptibility to runoff and erosion; validation at several levels. Catena 47, 133–149. Bascomb, C.L., 1968. Distribution of pyrophosphate-extractable iron and organic carbon in soils of various groups. J. Soil Sci. 52, 575–581. Boix-Fayos, C., Calvo-Cases, A., Imeson, A.C., Soriano-Soto, M.D., 2001. Influence of soil properties on the aggregation of some Mediterranean soils and the use of aggregate size and stability as land degradation indicators. Catena 44, 47–67. Bronick, C.J., Lal, R., 2005. Soil structure and management: a review. Geoderma 124, 3–22. Cammeraat, L.H., Imeson, A.C., 1998. Deriving indicators of soil degradation from soil aggregation studies in southeastern Spain and southern France. Geomorphology 23 (2–4), 307–321. Darboux, F., Le Bissonnais, Y., 2007. Changes in structural stability with soil surface crusting: consequences for erodibility estimation. Eur. J. Soil Sci. 58, 1107–1114. De Gryze, S., Bossuyt, H., Six, J., Meirvenne, M.V., Govers, G., Merckx, R., 2008. Factors controlling aggregation in a minimum and a conventionally tilled undulating field. Eur. J. Soil Sci. 58, 1017–1026. Denef, K., Six, J., Merckx, R., Paustian, K., 2002. Short-term effects of biological and physical forces on aggregate formation in soils with different clay mineralogy. Plant Soil 246, 185–200. Dong, A., Chester, A., Simsiman, G.V., 1983. Soil dispersibility. Soil Sci. 136, 208–212. Duiker, S.W., Rhoton, F.E., Torrent, J., Smeck, N.E., Lal, R., 2003. Iron (hydr) oxide crystallinity effects on soil aggregation. Soil Sci. Soc. Am. J. 67, 606–611. Gee, G.W., Bauder, J.W., 1986. Particle-Size Analysis. In: Klute, A. (Ed.), Methods of Soil Analysis. Part 1. Physical and Mineralogical Methods, second ed. Agronomy Monographs 9. ASA-SSA, Madison, WI, pp. 383–411. Goldberg, S., Kapoor, B.S., Rhoades, J.D., 1990. Effect of aluminum and iron oxides and organic matter on flocculation and dispersion of arid zone soils. Soil Sci. 150, 588–593. Green, V.S., Cavigelli, M.A., Dao, T.H., Flanagan, D.C., 2005. Soil physical properties and aggregate-associated C, N, and P distributions in organic and conventional cropping systems. Soil Sci. 170, 822–831. He, Y.Q., Sun, B., 2008. Evolvement and Regulation of Red Soil Quality. Science Press, Beijing (in Chinese). Horn, R., 1990. Aggregate characterization as compared to soil bulk properties. Soil Tillage Res. 17 (3–4), 265–289 (5). Huang, L., Wang, C.Y., Tan, W.F., Hu, H.Q., Cai, C.F., Wang, M.K., 2010. Distribution of organic matter in aggregates of eroded ultisols, Central China. Soil Tillage Res. 108, 59–67. Huygens, D., Boeckx, P., Van Cleemput, O., Oyarzun, C., Godoy, R., 2005. Aggregate and soil organic carbons dynamics in South Chilean andisols. Biogeosciences 2, 159–174. Idowu, O.J., 2003. Relationships between aggregate stability and selected soil properties in humid tropical environment. Commun. Soil sci. Plant 34, 695–708. Igwe, C.A., Akamigbo, F.O.R., Mbagwu, J.S.C., 1995. Physical properties of soils of southeastern Nigeria and the role of some aggregating agents in their stability. Soil Sci. 160, 431–441. Igwe, C.A., Akamigbo, F.O.R., Mbagwu, J.S.C., 1999. Chemical and mineralogical properties of soils in southeastern Nigeria in relation to aggregate stability. Geoderma 92, 111–123. Igwe, C.A., Zarei, M., Stahr, K., 2009. Colloidal stability in some tropical soils of southeastern Nigeria as affected by iron and aluminium oxides. Catena 77, 232–237. Imhoff, S., Pires da Silva, A., Dexter, A., 2002. Factors contributing to the tensile strength and friability of oxisols. Soil Sci. Soc. Am. J. 66, 1656–1661. Jackson, M.L., 1958. Soil Chemical Analysis. Prentice-Hall, Englewood Cliffs, NJ, p. 408. Keren, R., Singer, M.J., 1990. Effect of pH on permeability of clay-sand mixture containing hydroxy polymers. Soil Sci. Soc. Am. J. 54, 1310–1315. Kögel-Knabner, I., Guggenberger, G., Kleber, M., Kandeler, E., Kalbitz, K., Scheu, S., Eusterhues, K., Leinweber, P., 2008. Organo-mineral associations in temperate soils: integrating biology, mineralogy, and organic matter chemistry. J. Plant Nutr. Soil Sci. 171, 61–82. Lado, M., Ben-Hur, M., Shainberg, I., 2004. Soil wetting and texture effects on aggregate stability, seal formation and erosion. Soil Sci. Soc. Am. J. 68, 1992–1999. Levy, G.J., Mamedov, A.I., 2002. High-energy-moisture-characteristic aggregate stability as a predictor for a seal formation. Soil Sci. Soc. Am. J. 66, 1603–1609. Li, Z.X., Cai, C.F., Shi, Z.H., Wang, T.W., 2005. Aggregate stability and its relationship with some chemical properties of red soils in subtropical China. Pedosphere 15 (1), 129–136. Lu, S.G., Malik, Z., Chen, D.P., Wu, C.F., 2014. Porosity and pore size distribution of ultisols and correlations to soil iron oxides. Catena 123, 79–87. Martins, M.R., Angers, D.A., Corá, J.E., 2013. Non-labile plant C contributes to long-lasting macroaggregation of an oxisols. Soil Biol. Biochem. 58, 153–158. Mbagwu, J.S.C., Schwertmann, U., 2006. Some factors affecting clay dispersion and aggregate stability in selected soils of Nigeria. Int. Agrophys. 20, 23–30. Mckeague, J.A., Day, J.H., 1966. Dithionite and oxalate Fe and Al as acids in differentiating various classes of soil. Can. J. Soil Sci. 46, 13–22. Mehra, O.P., Jackson, M.L., 1960. Iron oxides removal from soils and clays by a dithionitecitrate system buffered with sodium bicarbonate. Clay Clay Miner. 7, 317–327. Noellemeyer, E., Frank, F., Alvarez, C., Morazzo, G., Quiroga, A., 2008. Carbon contents and aggregation related to soil physical and biological properties under a land-use sequence in the semiarid region of central Argentina. Soil Tillage Res. 99, 179–190. Oades, J.M., Waters, A.J., 1991. Aggregate hierarchy in soils. Aust. J. Soil Res. 29, 815–828. Onweremadu, E.U., Onyia, V.N., Anikwe, M.A.N., 2007. Carbon and nitrogen distribution in water-stable aggregates under two tillage techniques in fluvisols of Owerri area, southeastern Nigeria. Soil Tillage Res. 97, 195–206. Opara, C.C., 2009. Soil microaggregates stability under different land use types in southeastern Nigeria. Catena 79, 103–112. Peng, X., Yan, X., Zhou, H., Zhang, Y.Z., Sun, h., 2015. Assessing the contributions of sesquioxides and soil organic matter to aggregation in an ultisol under long-term fertilization. Soil Tillage Res. 146, 89–98.

J. Wang et al. / Catena 137 (2016) 545–553 Pinheiro-Dick, D., Schwertmann, U., 1996. Microaggregates from oxisols and inceptisols: dispersion through selective dissolutions and physicochemical treatments. Geoderma 74, 49–63. Plante, A.F., McGill, W.B., 2002. Soil aggregate dynamics and the retention of organic matter in a laboratory-incubated soil with differing simulated tillage frequencies. Soil Tillage Res. 66, 79–92. Pronk, G.J., Heister, K., Ding, G.C., Smalla, K., Kögel-Knabner, I., 2012. Development of biogeochemical interfaces in an artificial soil incubation experiment: aggregation and formation of organo-mineral associations. Geoderma 189–190, 585–594. Shrestha, B.M., Singh, B.R., Sitaula, B.K., Lal, R., Bajracharya, R.M., 2007. Soil aggregate- and particle-associated organic carbon under different land uses in Nepal. Soil Sci. Soc. Am. J. 71, 1194–1203. Six, J., Bossuyt, H., De Gryze, S., Denef, K., 2004. A history of research on the link between (micro) aggregates, soil biota, and soil organic matter dynamics. Soil Tillage Res. 79, 7–31. Six, J., Feller, C., Denef, K., Ogle, S.M., Sa, J.C.d.M., Albrecht, A., 2002. Soil organic matter, biota and aggregation in temperate and tropical soils-effects of no-tillage. Agronomie 22, 755–775. Statistical Package for the Social Sciences Inc., 2001. SPSS Advanced Statistics 11.0. SPSS Inc., Chicago. Sumner, M.E., 1992. The electrical double layer and clay dispersion. In: Sumner, M.E., Stewart, B.A. (Eds.), Advaced in Soil Science: Soil Crusting, Physical and Chemical Processes. Lewis Publishers, Boca Raton, pp. 1–31. Sumner, M.E., Miller, W.P., 1996. Cation exchange capacity and exchange coefficients. In: Sparks, D.L. (Ed.), Methods of Soil Analysis. Part 3Chemical Methods 5. ASA and SSSA, Madison, WI, pp. 1201–1229. Wagner, S., Cattle, S.R., Scholten, T., 2007. Soil-aggregate formation as influenced by clay content and organic-matter amendment. J. Plant Nutr. Soil Sci. 170, 173–180.

553

Wang, J.G., Li, Z.X., Cai, C.F., Yang, W., Ma, R.M., Zhang, G.B., 2012. Predicting physical equations of soil detachment by simulated concentrated flow in ultisols (subtropical China). Earth Surf. Process. Landf. 37, 633–641. Wuddivira, M.N., Camps-Roach, G., 2007. Effects of organic matter and calcium on soil structural stability. Eur. J. Soil Sci. 58, 722–727. Wuddivira, M.N., Stone, R.J., Ekwue, E.I., 2006. Soil texture, clay mineralogy and organic matter effects on structural stability and hydraulic characteristics of selected Trinidad soils after rainfall. Trop. Agric. 83, 69–78. Yan, F.L., Shi, Z.H., Li, Z.X., Cai, C.F., 2008. Estimating interrill soil erosion from aggregate stability of ultisols in subtropical China. Soil Tillage Res. 100, 34–41. Yoder, R.E., 1936. A direct method of aggregate analysis of soils and a study of the physical nature of erosion losses. J. Am. Soc. Agron. 28, 337–351. Zhang, B., Horn, R., 2001. Mechanisms of aggregate stabilization in ultisols from subtropical China. Geoderma 99 (1–2), 123–145. Zhang, M.K., Xu, J.M., 2005. Restoration of surface soil fertility of an eroded red soil in southern China. Soil Tillage Res. 80 (1–2), 13–21. Zhang, M.K., He, Z.L., Chen, G.C., Huang, C.Y., 1997. Formation of water-stable aggregates in red soils as affected by land use. Acta Pedol. Sin. 34 (4), 359–366 (in Chinese). Zhang, B., Yang, Y.S., Zepp, H., 2004. Effect of vegetation restoration on soil and water erosion and nutrient losses of a severely eroded clayey plinthudult in southeastern China. Catena 57, 77–90. Zhao, Q.G., Xu, M.J., Wu, Z.D., 2000. Agricultural sustainability of red soil region in southeast China. Acta Pedol. Sin. 37, 433–442 (in Chinese). Zhou, H., Peng, X.H., Perfect, E., Xiao, T.Q., Peng, G.Y., 2013. Effects of organic and inorganic fertilization on soil aggregation in an ultisol as characterized by synchrotron based Xray micro-computed tomography. Geoderma 195–196, 23–30.