Effects of rubber-based agroforestry systems on soil aggregation and associated soil organic carbon: Implications for land use

Geoderma 299 (2017) 13–24

Contents lists available at ScienceDirect

Geoderma journal homepage: www.elsevier.com/locate/geoderma

Effects of rubber-based agroforestry systems on soil aggregation and associated soil organic carbon: Implications for land use Chunfeng Chen a,b, Wenjie Liu a,⁎, Xiaojin Jiang a,⁎, Junen Wu a,b a b

Key Laboratory of Tropical Forest Ecology, Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, Menglun, Yunnan, 666303, China University of Chinese Academy of Sciences, Beijing 100049, China

a r t i c l e

i n f o

Article history: Received 14 January 2017 Received in revised form 16 March 2017 Accepted 19 March 2017 Available online xxxx Keywords: Agroforestry systems Aggregate stability Soil organic matter Erosion Aggregate-associated carbon

a b s t r a c t Rubber-based agroforestry (Hevea brasiliensis) systems are considered the best way to improve soil properties and the overall environmental quality of rubber monoculture, but few reports have examined soil aggregate stability in such systems. The objective of this study was to examine the management and landscape effects on water stable soil aggregates, soil aggregate-associated carbon, nitrogen content and soil carbon, and nitrogen accumulation in Xishuangbanna, southwestern China. Treatments were rubber monoculture (Rm) and four rubberbased agroforestry systems: H. brasiliensis–C. arabica (CAAs), H. brasiliensis–T. cacao (TCAs), H. brasiliensis– F. macrophylla (FMAs) and H. brasiliensis–D. cochinchinensis (DCAs). The results showed that, with the exception of CAAs, the rubber-based agroforestry treatments significantly increased total soil organic carbon (SOC) and N contents and enhanced the formation of macroaggregates compared to the rubber monoculture treatment. SOC and N contents in all water-stable aggregate fractions were significantly higher in rubber-based agroforestry systems (except CAAs) compared to rubber monoculture. The macroaggregate fractions contained more organic carbon and nitrogen than the microaggregate fractions. The proportions of C and N loss from slaking and sieving were shown to have significantly negative correlations with the mean weight diameter and the SOC and N concentrations in bulk soil. The results suggest that soil surface cover with constant leaf litter fall and extensive root systems in the rubber-based agroforestry systems increased soil organic carbon and nitrogen, helped improve soil aggregation, reduced soil erosion, decreased carbon and nitrogen loss, and ultimately improved the carbon and nitrogen accumulation rates. Given that the soil physical-chemical properties improvement and the patterns of the intercropping system played key roles in managing artificial forests, we recommend that local governments and farmers should prefer T. cacao, F. macrophylla and D. cochinchinensis and not C. arabica as the alternative interplanted tree species within rubber plantations. © 2017 Elsevier B.V. All rights reserved.

1. Introduction The spread of monoculture rubber plantations has occurred throughout the Xishuangbanna Region, resulting in 22.14% of the landscape covered by rubber (Xu et al., 2014) and barely 3.6% of that occupied by important tropical seasonal rainforest (Li et al., 2007). The transformation from both primary and secondary forests to rubber (plantations) and its continued intensification has resulted in numerous negative environmental consequences, particularly increased soil erosion (Mann, 2009), reduced water infiltration (Ziegler et al., 2009), soil nutrient loss and environmental degradation (Chaudhary et al., 2009; Qiu, 2009). The concentrations of organic carbon and nitrogen have also been reported to decline especially when native ecosystems ⁎ Corresponding authors at: Key Laboratory of Tropical Forest Ecology, Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, Menglun, Mengla, Yunnan 666303, China. E-mail addresses: [email protected] (W. Liu), [email protected] (X. Jiang).

http://dx.doi.org/10.1016/j.geoderma.2017.03.021 0016-7061/© 2017 Elsevier B.V. All rights reserved.

are converted to rubber (Li et al., 2012). Thus, a combined planting pattern of rubber and interplanting or a rubber-based agroforestry system could improve biodiversity, ecosystem services and the use of natural resources, which are important ways to promote the sustainable development of agriculture and the environment (Nath et al., 2005; Viswanathan and Shivakoti, 2008; van Noordwijk et al., 2012). However, over the past decade, the types of agroforestry systems and their soil properties have varied extensively. Although several studies concerning the temporal and spatial variability of soil properties have been conducted in this region (Zhang et al., 2007; Li et al., 2012), soil organic carbon (SOC) and the impacts on soil aggregates under different rubber-based agroforestry systems has received little attention. Soil aggregates are the basic units of the soil structure that control the dynamics of soil organic matter (SOM) and nutrient cycling (Jastrow et al., 1996; Chevallier et al., 2004). SOM is known to have a strong relationship with aggregate formation and stabilization. Soil aggregation is described using a hierarchical model (HM) and is generally divided into macroaggregates (N0.25 mm) and microaggregates

14

C. Chen et al. / Geoderma 299 (2017) 13–24

(0.25–0.053 mm) with differing binding agents. Soil microaggregates are typically formed by binding microbial polysaccharides with smaller soil particles such as silt and clay, whereas macroaggregates are typically formed by more transient factors such as enmeshment by fugal hyphae and fine roots (Rochester, 2011). This concept is supported by the observation that slaking-resistant macroaggregates (N0.25 mm) contain more organic matter than microaggregates (b0.25 mm) and more labile organic matter is abundant in macroaggregates than in microaggregates (Jastrow et al., 1996; Six et al., 2000). Therefore, macroaggregates are thought to be sensitive to changes in soil management such as cultivation practice and organic inputs, whereas microaggregates are less sensitive. Management practices, such as agroforestry systems and interplanting, which promote the maintenance and accumulation of soil C, have been increasingly accepted by farmers because of the growing interest in the conservation of SOM (Jose, 2009; Ramachandran Nair et al., 2009). For example, agroforestry systems that leave more plant residues on the soil surface generally allow for improvements in soil aggregation and aggregate stability. SOM can increase the amount of aggregates, especially macroaggregates, and promote the stability of aggregates (Elliott, 1986; Six et al., 2000). Meanwhile, soil aggregation can increase SOC storage by reducing loss from microbial mineralization and by water erosion. For the former, soil organic matter can be physically protected from microbial decomposition through sorption to clay minerals (Oades, 1984; Hassink et al., 1993) or other organic molecules and through isolation in micropores (Adu and Oades, 1978; Foster, 1981) and enclosure within soil aggregates (Tisdall and Oades, 1982), thus reducing the risk of being decomposed to CO2 into the atmosphere. Mineralization studies using intact versus crushed aggregates revealed the existence of a physically protected SOC pool in soil macroaggregates. For the latter, erosion reduced the amount of soil C by causing the degradation of the soil structure and removing C from one site and depositing it elsewhere (Gregorich et al., 1998). Water erosion tends to redistribute the smallest and least dense particles (small aggregate or clay), and organic C losses can be sensitive and extensive compared to bigger aggregates where organic C accumulate (Woods and Schuman, 1988). Organic C loss from soil occurs mainly through the mineralization of soil organic matter to CO2, whereas plenty of losses can also occur by the leaching of soluble organic C and by the flowing away of C bonded in clay. Although numerous studies have examined the mechanism and influence of aggregates in protecting SOC from mineralization (Woods and Schuman, 1988; Gregorich et al., 1998), few studies focus on the efficiency and mechanism of aggregates in protecting SOC against the destructibility and loss from erosion and leaching in that study area. We aimed to evaluate the influence of rubber-based agroforestry system management on soil aggregate stability, soil fertility and SOC and N loss in tropical hillside rubber plantations. Specifically, the objectives of this study were (1) to compare the differences in water stable aggregates, soil carbon, soil nitrogen and aggregate-associated carbon and nitrogen concentrations among rubber monoculture and four rubber-based agroforestry systems and (2) to prove the hypothesis of whether increased organic matter inputs, which varied in rubberbased agroforestry systems, could help improve soil aggregation, C storage and N availability relative to rubber monoculture management. 2. Materials and methods 2.1. Study sites The studied areas are located in the Xishuangbanna Tropical Botanical Garden (XTBG; 21°55′39″N, 101°15′55″E) in Yunnan Province, SW China. The climate is characterized by annual average temperatures ranging from 24 to 29 °C, high annual average atmospheric humidity (86%), and an average annual rainfall of 1557 mm with three seasons (fog-cool season: from November to February; hot-dry season: from

March to April; and rainy season: from May to October) (Vogel et al., 1995). The tropical southern monsoon dominates the climate and contributes 80–90% of the annual rainfall during the rainy season, whereas the subtropical jet streams prevail and deliver dry and cold air during the dry season. The research plots have slopes between 27 and 31° and are sandy loam in texture. The mean elevation of the plot is 760 m, ranging from 710 to 860 m (Fig. 1). The soils are classified as laterites (Oxisols) developed from arenaceous shale sediments approximately 2 m deep (Vogel et al., 1995). The parent material consisted of a 30–40 cm thick layer of gravel deposited by a distributary of the Mekong River. Studies were conducted in a typical catchment (19.3 ha) covered with rubber monoculture (clone PB86) arranged in double rows and planted at a density of 2 m × 4.5 m; there were 16-m-wide gaps between the rows. Rubber trees were tapped every other day from the end of March to mid-November (approximately 120 times per year), and the annual mean latex yield was approximately 250 kg ha−1. The experiment included four rubber-based agroforestry ecosystems that represented different land uses and management (CAAs, H. brasiliensis–C. arabica; TCAs, H. brasiliensis–T. cacao; FMAs, H. brasiliensis–F. macrophylla; and DCAs, H. brasiliensis–D. cochinchinensis). The four associated intercrops (approximately 10 years old) were planted in the 16 m interrows between the double rows of rubber monoculture. In CAAs, C. arabica trees were planted in five rows, each 1 m apart and containing plants spaced 1.6 m apart. C. arabica trees reached approximately 2.2 m high and were 4 m apart from the rubber trees. In TCAs, T. cacao trees were planted in five rows, with rows and plants within row spaced 2 m apart. T. cacao trees reached approximately 3.6 m and were approximately 3.5 m apart from the rubber trees. In FMAs, F. macrophylla trees were planted in eight rows, each spaced 1 m apart and 0.8 m between each plant in each row. F. macrophylla trees reached 4.2 m and were 3 m apart from the rubber trees. In DCAs, D. cochinchinensis trees were planted in five rows, with 1.5 m apart and 2.5 m between each plant in each row. D. cochinchinensis trees reached approximately 2.3 m and were 3.5 m apart from the rubber trees (Table 1). Morphological characteristics of the understory plant species and the rubber tree in the different types of the rubber-based agroforestry systems were shown in Table 1. The crops' leaf area index (LAI) and canopy closure rate were determined by using a plant canopy analyzer (LAI-2200; Li-Cor Inc., USA). Litterfall was collected from 1 m2 areas on the soil surface. The samples were oven-dried at 65 °C and then weighed. The sites were selected in this study were based on similarities in soil parent material, rubber age (the sites had approximately 25-year-old rubber, and rubber tree diameters were approximately 20–25 cm), and similar geographical position with a common north-facing slope (ranging from 85° to 94°). A commercial fertilizer containing N, P and K was point-applied in March and August at a dose of approximately 0.1 kg N per tree hole per year in each study site (Li et al., 2012). 2.2. Sampling and measuring methods Using the S-shaped sampling strategy, 32 undisturbed soil samples for soil structure determination in each soil depth (0–5, 5–15 and 15–30 cm) were collected in different ecosystems or agro-ecosystems in November 2014, and eight soil samples were mixed into 2 kg soil samples (Bissonnais, 1996) resulting in four mixed soil samples. Prior to the determination of water-stable aggregation, the soils were ovendried at 45 °C. The analysis of water-stable soil aggregates was carried out using a modified Yoder type apparatus (Yoder, 1936). Briefly, a bank of sieves 200 mm in diameter with mesh aperture of 0.053, 0.25, 0.5, 1, 2 and 5 mm, containing 100 g air-dried soil on the top of the sieve, were submerged in deionized water for 10 min at room temperature. Soils were then sieved under water by moving the sieves up and down for a period of 5 min. It should to be noted that floating organic materials or floatable plant materials, which may increase a slight

C. Chen et al. / Geoderma 299 (2017) 13–24

15

Fig. 1. Location of the study site (21°55′39″N, 101°15′55″E) in Yunnan Province, southwest China.

over-estimation of SOC and N contents of each aggregate fraction, were decanted and removed. The material retained on each sieve was removed and oven-dried at 60 °C for 24 h and weighed to calculate the proportion of water-stable aggregate fraction for every particle size. Subsamples of the intact sieved fractions and whole soil (i.e., the starting material) were finely ground and then analysed so that C and N recovery could be calculated. The elemental C and N concentrations were measured by dry combustion with a Carlo Erba NC2500 elemental analyzer (Milan, Italy). It was found that the soil in this area was free of inorganic C, as there was no reaction when HCl was added to the soil; hence, all C measured was considered to be equivalent to the SOC (Chivenge et al., 2011). The pH of the soil was potentiometrically measured in a supernatant suspension of a 1:2.5 soil:liquid mixture. The particle size distributions (silt, sand and clay fractions) were determined after dissolution of CaCO3 with 2 mol dm−3 HCl and decomposition of the organic matter with 30% H2O2. Silt, sand and clay fractions were determined according to pipette method (van Reeuwijk, 2002) after the dispersion of samples using Na(PO3)6.

2.3. Calculations and statistical analyses The water stable aggregate index (WSA), which accounts for aggregate stability, was calculated using the following equation: WSA ¼ Mr =Mt  100

ð1Þ

where Mr is mass of resistant aggregates (g), and Mt is the total mass of wet sieved soil (g). The aggregate indices of soil mean weight diameter (MWD) for each treatment were calculated according to the following formulas: n

ð2Þ

MWD ¼ ∑ xi yi i¼1

where xi is the mean diameter of each particular size of aggregates separated by sieving (mm), yi is the percent of the weight of aggregates in that size range to the total dry weight of soil, and n is the number of separated aggregate classes. To determine the total mass of C and N associated with the whole mass of each individual aggregate size class recovered from 100 g of soil, aggregate C and N were calculated as the mass of C and N per whole mass of sand-free WSA. Subsamples of bulk soil from each plot were also ground, and total C and N were determined for the whole soil as described above. C loss (SOC external to aggregates) is equal to the SOC content in 100 g of whole soil minus the total SOC associated with the whole mass of each individual aggregate size class recovered from 100 g of soil. In addition, the proportion of C loss is equal to the C loss (SOC external to aggregates) divided by the SOC content in 100 g of whole soil. The data for each soil layer (0–5, 5–15, and 15–30 cm) were subjected to one-way analysis of variance (ANOVA) using SPSS 13.0, and all comparisons among the different forest types for the soil physicalchemical properties were conducted using least significant difference

Table 1 Morphological characteristics of rubber trees and understory species in the different types of rubber-based agroforestry systems. Species

Hevea brasiliensis

C. arabica

T. cacao

F. macrophylla

D. cochinchinensis

Plant height (m) Height of 1st branch (m) Stem diameter (cm) Leaf size (cm2) Crown breadth (m) Leaf area index (m2 m−2) Canopy closure rate (%) Litterfall (kg ha−1)

20.14 ± 1.48 2.03 ± 0.36 24.45 ± 3.7 80.13 ± 12.2 12.84 ± 1.6 1.31 ± 0.18 71.67 ± 4.36 907.12 ± 22.97

2.16 ± 0.23c 0.27 ± 0.15b 5.02 ± 0.42bc 95.73 ± 12.19b 1.55 ± 0.27c 2.42 ± 0.31c 39 ± 4.41c 1209.7 ± 16.6c

3.62 ± 0.76b 0.8 ± 0.15a 6.83 ± 1.08b 243.54 ± 82.24a 3.34 ± 0.88a 3.65 ± 0.62a 37.33 ± 4.18c 2712.04 ± 32.77a

4.2 ± 0.48a 0.21 ± 0.03b 3.09 ± 0.49c 227.74 ± 60.94a 2.14 ± 0.5b 2.27 ± 0.29c 69.33 ± 2.92a 1548.22 ± 24.18bc

2.23 ± 0.29c 0.26 ± 0.04b 16.81 ± 4.73a 102.58 ± 18.3b 2.19 ± 0.42b 3.11 ± 0.54b 48.22 ± 4.68b 1905.4 ± 28.66b

Means and standard error (n = 9). Treatments indicated by the same letter are not significantly different at P b 0.05 (one-way ANOVA).

16

C. Chen et al. / Geoderma 299 (2017) 13–24

tests (LSD). Significant differences among means were evaluated by Tukey's honest significant difference (HSD) at P b 0.05. The effects of agroforestry type, soil depth and other soil parameters were analysed using the general linear model procedure (PROC GLM). The differences among land uses for concentrations of recovery of C (SOC within macroand microaggregates, grams per kilogram of whole aggregates), C loss (SOC external to aggregates), and the proportion of C loss were determined using ANOVA. The relationships between the proportion of C and N loss and the C and N concentrations of whole soil and the relationships between the proportion of C and N loss and MWD were determined by linear regression. 3. Results 3.1. Soil properties After 10 years of rubber-based agroforestry ecosystem restoration, many soil properties were considerably changed (Table 2). Soils beneath Rm, CAAs, and FMAs were loamy clay, with a particle-size distribution of approximately 30% sand, 30% silt, and 40% clay in the 0–5 cm layer. The soils in TCAs and DCAs were clay with a particle-size distribution of 20% sand, 30% silt, and 50% clay. Patterns were similar at the 5–15 and 15–30 cm layers, although the soils at those depths contained slightly more silt and clay relative to the 0–5 cm depth. SOC and N contents were significantly greater in rubber-based agroforestry ecosystems (except CAAs) than those in Rm systems (P b 0.01). The C:N ratio values ranged from 7.99 to 9.40, with significantly lower values in CAAs sites at a 0–5 cm depth and significantly higher values in TCAs sites at a 5–30 cm depth. The soil pH values ranged from 4.7 to 5.4, with slightly lower values in Rm systems compared with agroforestry systems across each soil depth. 3.2. Aggregates distribution and stability The results showed that rubber-based agroforestry ecosystems had a significant main treatment effect on the proportion of soil in the 6 fractions and on MWD among the 0–5, 5–15, and 15–30 cm sampling depths (Table 3, Fig. 2). Compared with Rm systems, rubber-based agroforestry ecosystems increased the proportion in the N5 and 5–2 mm fractions and reduced the proportion in the 1–0.5, 0.5–0.25, and 0.25–0.053 mm fractions at a soil depth of 0–5 cm. The proportions of the mass of N5 mm significantly increased by 107.33%, 350.77%, 253.32% and 350.59% for CAAs, TCAs, FMAs and DCAs, respectively (P b 0.01). Meanwhile, the proportions

of the mass of N 0.25 mm significantly increased by 19.47%, 12.74% and 15.27% for TCAs, FMAs and DCAs, respectively (P b 0.01). In the 5–15 and 15–30 cm soil depths, TCAs and DCAs increased the proportions in the N5, 5–2 and N 0.25 mm fractions, whereas CAAs and FMAs decreased those fractions when compared with Rm. The proportion of bigger macroaggregates declined with soil depth, resulting in an increased proportion of smaller macroaggregates or microaggregates (Table 3). The proportions in the N5 and N 0.25 mm fractions in each rubber-based agroforestry ecosystem declined in the sub-soil layers (5–15 and 15–30 cm depths) when compared with the 0–5 cm depth, but those increased in Rm. However, CAAs had a negative effect on soil aggregation, as it had smaller N 0.25 mm fractions than Rm in both 5–15 and 15–30 cm soil layers, i.e., lower water-stable aggregate contents. Particularly, the N 0.25 mm fractions (macroaggregate size) and the N 2 mm fractions (large macroaggregate size) under TCAs and DCAs were both significantly greater than those under the other sites. The effects of agroforestry systems on the MWD are shown in Fig. 2. The MWD was significantly greater under rubber-based agroforestry systems than that under the Rm system at a 0–5 cm depth (P b 0.05). The MWD values with CAAs, TCAs, FMAs and DCAs were 22.52%, 120.07%, 70.9%, and 102.62% higher than that with Rm, respectively. In the depths of 5–15 and 15–30 cm, the MWD was significantly higher under TCAs and DCAs than that under Rm, however, no significant differences were found with CAAs and FMAs. The MWD increased gradually with the soil layer depth with four treatments, and the differences with rubber-based agroforestry systems were obvious, especially in the 0–5 cm layer. 3.3. Soil carbon and nitrogen concentrations within aggregates Carbon and nitrogen concentrations in the N 2, 2–1, 1–0.25, 0.25–0.053, and b0.053 mm fractions generated by wet-sieving were estimated. The effects of agroforestry treatments or soil depth treatments (main treatment effects) on carbon, nitrogen and the C:N ratio within aggregates were presented in Fig. 3. The effect of agroforestry systems on soil aggregate organic carbon and nitrogen concentrations was significant at all depths (P b 0.05). All of the rubber-based agroforestry systems produced higher soil aggregate organic carbon content at each depth compared with rubber monoculture (except CAAs). At the depth of 0–5 cm, the aggregate SOC and N content under the five land uses followed a deceasing sequence FMAs N DCAs N TCAs N CAAs N Rm, whereas sequence DCAs N FMAs N TCAs N CAAs N Rm occurred at depths 5–15 and 15–30 cm. In the topsoil, compared with Rm, the proportions of N2 mm aggregate

Table 2 Soil physical and chemical characteristics of the landscape elements at the study areas. Land use

C (g kg−1)

N (g kg−1)

C:N

Sand (%)

Silt (%)

Clay (%)

pH

0–5 cm Rm CAAs TCAs FMAs DCAs

17.00 ± 0.20b 16.81 ± 2.20b 23.57 ± 1.73a 26.90 ± 1.40a 24.33 ± 0.27a

1.86 ± 0.01c 1.97 ± 0.24c 2.57 ± 0.20b 2.98 ± 0.10a 2.68 ± 0.01ab

9.12 ± 0.05a 8.51 ± 0.17b 9.16 ± 0.20a 9.01 ± 0.18a 9.09 ± 0.11a

35.25 ± 0.84a 30.60 ± 1.18b 24.37 ± 0.52c 24.32 ± 2.01c 27.05 ± 1.11bc

34.61 ± 0.89b 35.91 ± 1.05b 26.76 ± 1.12c 38.72 ± 0.12a 26.68 ± 0.15c

30.15 ± 0.49c 33.48 ± 0.13bc 48.88 ± 0.99a 36.96 ± 2.32b 46.28 ± 1.16a

4.8b 5b 4.8b 5.2a 5.2a

5–15 cm Rm CAAs TCAs FMAs DCAs

13.71 ± 0.81c 15.59 ± 0.24bc 19.67 ± 1.91a 18.30 ± 2.14ab 20.62 ± 0.05a

1.63 ± 0.06c 1.89 ± 0.03bc 2.10 ± 0.20ab 2.20 ± 0.24ab 2.31 ± 0.01a

8.41 ± 0.17c 8.23 ± 0.050c 9.40 ± 0.07a 8.30 ± 0.06c 8.91 ± 0.01b

27.60 ± 0.36a 25.19 ± 0.83b 25.54 ± 0.87ab 22.87 ± 0.61c 21.52 ± 0.82c

35.32 ± 0.70b 36.86 ± 0.41ab 26.41 ± 0.96d 38.7 ± 0.36a 30.55 ± 0.57c

37.08 ± 0.36b 37.94 ± 0.89b 48.05 ± 1.52a 38.43 ± 0.35b 47.93 ± 0.72a

4.8b 4.9b 4.8b 5.1ab 5.3a

15–30 cm Rm CAAs TCAs FMAs DCAs

10.36 ± 0.64c 13.14 ± 0.68b 17.32 ± 1.83a 13.62 ± 0.30b 16.02 ± 1.16a

1.29 ± 0.04c 1.64 ± 0.07b 1.93 ± 0.18a 1.64 ± 0.04b 1.92 ± 0.13a

8.04 ± 0.29b 7.99 ± 0.11b 8.96 ± 0.14a 8.29 ± 0.36b 8.35 ± 0.05b

26.87 ± 0.84a 21.84 ± 0.25b 22.87 ± 0.60b 21.51 ± 1.08b 16.82 ± 0.54c

34.74 ± 1.06b 39.28 ± 0.41a 26.12 ± 0.24c 38.10 ± 1.02a 33.76 ± 0.48b

38.39 ± 0.26d 38.88 ± 0.23 cd 51.01 ± 0.84a 40.39 ± 0.15c 49.42 ± 0.63b

4.9b 4.7b 4.8b 5.2a 5.4a

Rm: rubber monoculture, CAAs: H. brasiliensis–C. arabica, TCAs: H. brasiliensis–T. cacao, FMAs: H. brasiliensis–F. macrophylla, DCAs: H. brasiliensis–D. cochinchinensis; Means and standard error (n = 3). Treatments indicated by the same letter are not significantly different at P b 0.05 (one-way ANOVA).

C. Chen et al. / Geoderma 299 (2017) 13–24

17

Table 3 Proportional aggregate weight (%) of different soil fractions acquired from wet sieving at the study areas. Land use

Water stable aggregate (%)

WSA (%)

N5 mm

5–2 mm

2–1 mm

1–0.5 mm

0.5–0.25 mm

0.25–0.053 mm

0–5 cm Rm CAAs TCAs FMAs DCAs

5.87 ± 0.33d 12.17 ± 1.22c 26.46 ± 0.54a 20.74 ± 1.49b 26.42 ± 0.26a

12.97 ± 1.11d 12.69 ± 1.46d 30.20 ± 0.35a 18.26 ± 0.79c 22.75 ± 0.98b

14.46 ± 0.66abc 13.01 ± 1.06bc 12.82 ± 0.26c 14.87 ± 0.29ab 15.05 ± 0.23a

28.11 ± 0.41a 26.52 ± 1.92a 12.61 ± 0.13c 21.52 ± 0.93b 14.76 ± 0.7c

11.36 ± 0.53a 9.21 ± 0.82b 4.83 ± 0.13d 6.64 ± 0.64c 4.89 ± 0.37d

8.15 ± 0.44a 6.14 ± 0.89b 2.92 ± 0.23c 3.73 ± 0.46c 3.16 ± 0.17c

72.76 ± 0.56c 73.59 ± 1.12c 86.93 ± 0.29a 82.03 ± 0.61b 83.87 ± 0.45b

5–15 cm Rm CAAs TCAs FMAs DCAs

11.27 ± 1.75b 4.39 ± 0.96d 13.21 ± 0.37b 7.52 ± 0.17c 20.22 ± 0.43a

13.06 ± 1.46c 8.64 ± 1.85d 23.01 ± 1.39b 13.03 ± 0.97c 27.85 ± 0.63a

18.8 ± 2.04ab 15.17 ± 1.09c 17.49 ± 0.75bc 21.57 ± 0.64a 14.86 ± 0.56c

24.62 ± 1.16c 31.92 ± 0.21a 22.02 ± 1.02d 27.75 ± 0.7b 15.76 ± 0.73e

9.04 ± 0.43b 11.53 ± 0.87a 6.34 ± 0.55c 8.41 ± 0.22b 5.32 ± 0.24d

5.08 ± 0.45b 6.55 ± 0.23a 3.98 ± 0.37 cd 4.62 ± 0.38bc 3.23 ± 0.19d

76.79 ± 0.61b 70.66 ± 1.54c 82.07 ± 0.21a 78.28 ± 0.51b 84.02 ± 0.60a

15–30 cm Rm CAAs TCAs FMAs DCAs

4.8 ± 1.12b 3.41 ± 0.08bc 12.27 ± 0.67a 2.89 ± 0.41c 10.9 ± 0.71a

6.92 ± 1.42b 7.45 ± 1.77b 17.36 ± 1.73a 7.9 ± 0.75b 20.9 ± 1.5a

19.76 ± 1.44ab 19.48 ± 2.52ab 17.61 ± 0.55b 22.31 ± 1.89ab 23.3 ± 0.64a

33.40 ± 1.53a 33.22 ± 1.74a 23.78 ± 1.41b 32.14 ± 0.89a 20.06 ± 1.23b

11.08 ± 0.96a 9.26 ± 0.99ab 7.58 ± 0.38bc 10.15 ± 0.95a 6.47 ± 0.33c

6.89 ± 0.62a 6.19 ± 0.67a 4.71 ± 0.37bc 5.51 ± 0.32ab 3.63 ± 0.21c

73.56 ± 2.12c 72.83 ± 1.72c 78.60 ± 0.93ab 75.38 ± 0.47bc 81.63 ± 0.61a

Rm: rubber monoculture, CAAs: H. brasiliensis–C. arabica, TCAs: H. brasiliensis–T. cacao, FMAs: H. brasiliensis–F. macrophylla, DCAs: H. brasiliensis–D. cochinchinensis; Means and standard error (n = 5). Treatments indicated by the same letter are not significantly different at P b 0.05 (one-way ANOVA).

SOC significantly increased by 11.77%, 35.60%, 56.15%, and 38.26% for CAAs, TCAs, FMAs and DCAs, respectively (P b 0.05); meanwhile, the proportions of 0.25–0.053 mm aggregate SOC significantly increased 10.76%, 30.38%, 55.73%, and 48.63%, respectively (P b 0.05). There were no significant differences in the proportions of 2–1 and 1–0.25 mm at the depth of 0–5 cm between Rm and CAAs, but these proportions differed significantly among the other agroforestry systems (P b 0.01). In addition, at the depths of 5–15 and 15–30 cm, the aggregate SOC and N content had the same trend as the depth of 0–5 cm. The proportion of b0.053 mm aggregate SOC was significantly higher under TCAs, FMAs and DCAs than under Rm and CAAs. However, the proportions of b 0.053 mm aggregate SOC were much higher than most of the other aggregate fractions. For N0.053 mm aggregate

fractions, aggregate SOC and N contents increased with the increasing aggregate fraction sizes. The C:N ratios of the soil aggregate fractions were significantly different among rubber monoculture and rubberbased agroforestry systems at the 0–15 cm soil depth; however, they were not significantly different at the 15–30 cm soil depth. Meanwhile, for N 0.053 mm aggregate fractions, the values of C:N increased with increasing aggregate fraction sizes, whereas the C:N values of b 0.053 mm clay fractions were high as well. 3.4. Recovery and loss of C and N from slaking and sieving The recovery of SOC had significant differences among the five sites at the three soil depths (0–5 cm: F = 51.11, P b 0.01; 5–15 cm: F =

Fig. 2. Mean weight of the water stable aggregate (MWD) values with wet sieving under the rubber monoculture system and rubber-based agroforestry systems. Means and standard error (n = 5). Bars with different lower case letters indicate significant differences at P b 0.05. See Table 1 for abbreviation.

18

C. Chen et al. / Geoderma 299 (2017) 13–24

Fig. 3. The characteristics of soil aggregate organic carbon content, soil aggregate nitrogen content and C:N values for the size fractions under the monoculture rubber system and rubberbased agroforestry systems. Means and standard error (n = 5). Bars with different lower case letters indicate significant differences at P b 0.05. See Table 1 for abbreviation.

89.95, P b 0.01; 15–30 cm: F = 29.64, P b 0.01), with high values observed in agroforestry systems and lower values observed in Rm (Table 4). However, the recovery of SOC in samples from the three soil depths for Rm and CAAs did not show a significant difference. In addition, the recovery of SOC decreased with increasing layers. Partitioning by fraction (Fig. 4), macroaggregates accounted for approximately 70% of the total amount of recovery of C across the landscape elements, which was much higher than microaggregates (10%).

Soil C loss (SOC external to aggregates) values showed significant differences among sites and depths (P b 0.05) (Table 4). However, C loss values were lower in Rm than in most agroforestry systems, which was apparently due to the differences of initial soil C concentrations and inherent soil variability. Therefore, we calculated the relative proportions of C loss in total soil C. The results clearly showed that the proportion of C loss was significantly different among the five sites at the three soil depths (0–5 cm: F = 3.66, P b 0.05; 5–15 cm: F = 5.14,

Table 4 Soil organic carbon (SOC) and nitrogen balance under the monoculture rubber system and rubber-based agroforestry systems. Land use

Total mass of aggregate C (g kg−1)

C loss (g kg−1)

Proportion of C loss (%)

Total mass of aggregate C (g kg−1)

N loss (g kg−1)

Proportion of N loss (%)

0–5 cm Rm CAAs TCAs FMAs DCAs

13.40 ± 0.05c 14.46 ± 0.37c 20.50 ± 0.18b 22.39 ± 1.07a 20.36 ± 0.52b

3.60 ± 0.24a 2.35 ± 0.51a 3.06 ± 0.32a 4.50 ± 0.74a 3.97 ± 0.30a

21.19 ± 0.05a 12.59 ± 1.98b 12.99 ± 0.71b 16.79 ± 0.77ab 16.35 ± 2.35b

1.52 ± 0.01d 1.69 ± 0.04c 2.27 ± 0.04b 2.54 ± 0.12a 2.33 ± 0.05b

0.35 ± 0.01ab 0.28 ± 0.04b 0.31 ± 0.04b 0.44 ± 0.03a 0.35 ± 0.05ab

18.53 ± 0.20a 14.13 ± 0.69b 11.91 ± 1.33b 14.79 ± 0.36b 12.94 ± 1.94b

5–15 cm Rm CAAs TCAs FMAs DCAs

10.77 ± 0.37b 11.60 ± 0.25b 15.38 ± 0.28a 16.28 ± 1.54a 17.60 ± 0.07a

2.94 ± 0.56a 3.99 ± 0.49a 4.29 ± 0.40a 2.02 ± 1.24a 3.02 ± 0.05a

21.91 ± 4.03ab 25.58 ± 1.25a 21.80 ± 1.32ab 11.63 ± 3.28c 14.65 ± 0.34bc

1.31 ± 0.03c 1.44 ± 0.03c 1.70 ± 0.02b 1.96 ± 0.15a 2.07 ± 0.01a

0.32 ± 0.03ab 0.45 ± 0.03a 0.39 ± 0.03a 0.23 ± 0.12b 0.24 ± 0.01b

19.76 ± 2.89ab 23.66 ± 1.20a 18.79 ± 1.26ab 15.88 ± 3.04bc 10.52 ± 0.49c

2.61 ± 0.26ab 3.49 ± 0.81ab 4.59 ± 0.92a 2.69 ± 0.34ab 2.13 ± 0.49b

26.11 ± 2.07a 27.00 ± 3.54a 26.87 ± 0.80a 19.78 ± 0.38b 14.24 ± 1.02b

0.98 ± 0.06d 1.26 ± 0.08c 1.55 ± 0.07b 1.36 ± 0.01c 1.70 ± 0.10a

0.30 ± 0.06a 0.38 ± 0.06a 0.38 ± 0.03a 0.28 ± 0.01a 0.22 ± 0.03a

23.95 ± 5.06a 23.34 ± 3.20a 19.85 ± 1.42ab 17.35 ± 0.61ab 11.50 ± 0.76b

15–30 cm Rm 7.75 ± 0.64d CAAs 9.65 ± 0.72 cd TCAs 12.73 ± 0.92ab FMAs 10.92 ± 0.05bc DCAs 13.89 ± 1.03a

Rm: rubber monoculture, CAAs: H. brasiliensis–C. arabica, TCAs: H. brasiliensis–T. cacao, FMAs: H. brasiliensis–F. macrophylla, DCAs: H. brasiliensis–D. cochinchinensis; Means and standard error (n = 3). Treatments indicated by the same letter are not significantly different at P b 0.05 (one-way ANOVA).

C. Chen et al. / Geoderma 299 (2017) 13–24

19

Fig. 4. Carbon (a–c) and nitrogen (d–f) distribution among the different aggregate fraction sizes in the monoculture rubber system and rubber-based agroforestry systems (n = 5, error bars are SE). See Table 1 for abbreviation.

P b 0.05; 15–30 cm: F = 5.478, P b 0.05). In addition, the proportion of C loss from the 0–5 cm depth was significantly greater in Rm compared to the agroforestry systems (P b 0.05). Meanwhile, the proportion of C loss from the 5–15 and 15–30 cm depths was significantly greater in Rm than FMAs and DCAs; however, there were no significant differences among Rm, CAAs and TCAs. The recovery of C increased with increased C content in bulk soil (F = 435.509; P b 0.001) and increased with increased MWD values (F = 35.898; P b 0.001). The slope of this relationship is similar to that between the recovery of N, bulk soil N content, and MWD values. The proportion of C loss was reduced with increased C content in bulk soil (F = 5.915; P b 0.05) and with increased MWD values (F = 7.514; P b 0.05). 4. Discussion 4.1. Aggregate size distribution and stability as affected by different intercropped treatments In this study, the greatest impacts on aggregate mass distribution and aggregate stability were associated with the conversion from monoculture rubber forest to rubber-based agroforestry systems. Our results showed that the amount of aggregates and MWD in rubberbased agroforestry systems were significantly higher than Rm (except for CAAs). These findings agreed with previous studies by Gupta et al. (2009) and Gama-Rodrigues et al. (2010). They found that the amount of macroaggregates and MWD were higher when the sole crop was converted to agroforestry systems. Although changes in soil aggregation following the conversion of monoculture rubber forest to rubberbased agroforestry systems may be contributed to the mechanical disturbance associated with cultivation patterns, the agroforestry systems

in this study practically localize the effects of such disturbances, as farmers typically sow rubber trees as well as inter-planting trees, apply fertilizer into same size holes (approximately 30 cm), and utilize the same operation of tapping to ensure that no other form of soil displacement occurs. Thus, most impacts of forest conversion on soil aggregates are likely attributable to other mechanisms, such as altered community composition and biomass, changes in organic matter inputs, and shifts in microclimate or soil construction. The data obtained in this study showed that MWD in all the four rubber-based agroforestry systems was significantly higher than in rubber monoculture in topsoil, and the WSA in all but one system (CAAs) was significantly greater than in rubber monoculture in that soil layer (Figs. 2 and 7). This phenomenon was based on at least four types of evidence. First, agroforestry systems generally enhance organic matter accumulation in soils through the inclusion of permanent vegetation and cover plants, which would be expected to increase soil microbial populations and earthworms and consequently increase the amount of macroaggregates (Udawatta et al., 2008). We obtained a similar result, namely, that higher contents of SOC and N were found in agroforestry systems (except for CAAs in 0–5 cm soil depth) than in Rm; the content of C increased 13.94–65.93%, and the content of N increased 17.79–57.24% using the same fertilization methods and fertilizer (Table 2). Furthermore, the relationship between SOC stock, the mass of the N 0.25 mm aggregate size class and the MWD indicated a strong interdependence at all systems and emphasized that the increase in SOC stock of rubber-based agroforestry systems also increases macroaggregates and aggregation indices. Second, the distribution and penetration of roots (especially fine roots) into the soil could also contribute to the variation in formation and the stability of soil aggregates (Pohl et al., 2009; Erktan et al., 2016). In general, fine roots could induce the

20

C. Chen et al. / Geoderma 299 (2017) 13–24

formation of macroaggregate due to the compressing action of growing roots and the bending of the organic matter, which originated from the dead roots (Erktan et al., 2016). Thus, agroforestry systems, which contain abundant fine roots in the soils, may improve the amount of macroaggregate compared to rubber monoculture. Third, the protective effect of leaf litter and the canopy of the agroforestry system acted as mulch against the beating action of raindrops and erosion from runoff, which would break down bigger aggregates into smaller ones (Bruce et al., 1992; Gupta et al., 2009). We also quantified the leaf fall in these systems (Table 1), and the results showed that the amount of litter in agroforestry systems was larger than in Rm and was closely linked with macroaggregates and MWD in treatments and depths. Fourth, clay was also considered to be the major binding agent for aggregate formation and the stabilization in these soils (Frenkel et al., 1978; Shainberg et al., 1992; Wick et al., 2009). It has been reported that soil clay content can influence the degree of soil aggregation because the higher reactive surface of clay promotes binding with organic materials (Shainberg et al., 1992). Significantly higher clay content in rubber-based agroforestry systems (24%–56%) contributed to greater aggregation through the reintegration of clay particles into the aggregates, whereas Rm had low clay content (11%–49%). Similar research was reported by Gupta et al. (2009), who found more richness of clay in poplar-based agroforestry systems, resulting in more water stable aggregates and increased MWD compared with a monoculture poplar system. Important differences among intercropped treatments were found in the size and distribution of aggregates. Our results showed that distribution in macroaggregates, in particular those N 0.25 mm, were improved when rubber monoculture was intercropped with T. cacao (TCAs) and with D. cochinchinensis (DCAs) at all soil depths. However, there were no significant changes in those fractions when rubber monoculture was intercropped with C. arabica (CAAs) and with F. macrophylla (FMAs) (except for 0–5 cm soil depth). The diversity of rubber-based agroforestry systems could account for this phenomenon. As described above, fine roots play an important role in soil aggregation. However, the root systems among rubber trees and intercropped species were complex and changeable in this study area, especially in the systems that contained different combinations of tree species. Indeed, previous studies indicated that almost all fine roots of cocoa trees were concentrated in the surface soil layer (N90% were in the upper 10 cm of the soil) (Carr and Lockwood, 2011). Only 33% of the fine roots of coffee trees were located in the upper 10 cm of the soil, and 73% were located in the upper 30 cm of the soil (Cuenca et al., 1983); meanwhile, 60.3% of the fine roots of Flemingia macrophylla were concentrated in first 10 cm of soil, 24.4% were in 10–20 cm, and 11.2% were in 20–30 cm (Budelman, 1990). The distribution of fine roots in different interplantings could explain the reason that the WSA and MWD in CAAs and FMAs (except in 0–5 cm soil depth) were significant lower than TCAs and DCAs at all soil depths. Another reason is clay content. Specifically, we found a higher content of clay in TCAs (48.05%– 51.01%) and in DCAs (46.28%–49.42%) at all soil depths, whereas a lower content of clay was found in CAAs (33.48%–38.88%) and FMAs (36.96%–40.39%). A positive correlation was also observed between soil clay and WSA, as well as MWD; this indicated that soil clay played important roles in the formation and stability of soil aggregates in these study regions. 4.2. Soil aggregate SOC, N and C:N ratio as affected by agroforestry treatments A significant effect of land use change was observed with SOC and N concentrations in aggregate fractions. The result of this study showed that the aggregate SOC and N contents were higher in TCAs, FMAs and DCAs than in CAAs and Rm, which largely corresponded to the changes demonstrated in bulk soil SOC and N. This was due to the fact that the inputs of additional plant material, such as litter and fine roots, into the soils were very different under different rubber-based agroforestry

treatments, and the resulting aggregate SOC and N were also significantly different. In addition, the treatment differences in SOC and N storage that varied with aggregate fractions essentially reflected the treatment effects on the distribution of soil among the different aggregate fraction sizes (Fig. 3). This suggested that SOC and N concentrations within aggregate fractions were not only the result of SOC and N storage in bulk soil, which was largely affected by land use but also due to the redistribution of soil among the difference aggregate fractions. Tisdall and Oades (1982) and Hassink (1997) proposed the concept of the model of aggregate hierarchy (HM), which stated that the addition of organic matter to soils first results in the formation of SOM associations with clay and silt particles. The formation of microaggregates (b0.25 mm) and macroaggregates (N 0.25 mm) begins if the SOM binding capacity of the clay and silt fractions are saturated. Thus, in accordance to the concept of aggregate hierarchy, microaggregates are bound together to form macroaggregates by transient binding agents (i.e., microbial- and plant-derived polysaccharides) and temporary binding agents (i.e., roots and fungal hyphae) (Tisdall and Oades, 1982; Six et al., 2000). The aggregate hierarchy shows the tendency of an increase in C concentration with increasing aggregate-size classes because larger aggregate-size classes are composed of smaller aggregate-size classes incorporated with organic binding agents (Elliott, 1986). We obtained the same result, which showed that the SOC and N concentrations within aggregates were increased with increasing aggregates both rubber-based agroforestry ecosystems and rubber monoculture ecosystems soils (Fig. 3). Namely, SOC content within the N 0.25 mm aggregate fractions (macroaggregates) was higher than that within the 0.25–0.053 mm aggregate fraction (microaggregates). However, in all soils, the higher SOC concentration was found in the aggregate fraction b0.053 mm when compared to the other aggregate fractions. This is contradictory to the results of John et al. (2005), who observed that lowest SOC concentration was included in that aggregate fraction. The underlying mechanism for such a trend is not clear. Further investigation into the speciation of SOC in the soils, including particulate organic C, may be needed to elucidate the process involved. Low values of the soil C:N ratio in these ecosystems were obtained because of the warm ecosystems in the study area, indicating a fast decomposition of organic residues and the prevalence of a medium humified humus across the sample plots (Callesen et al., 2007; Vodnik et al., 2008); the values of C:N ratio ranged from 7.5 to 9.5 (Table 1). Oades (1984) suggested that organic matter in macroaggregates was a result of mineral particles and particulate organic matter, primary plant materials with a wide C:N ratio. Buyanovsky et al. (1994) indicated that C:N ratios within macroaggregates were higher than within microaggregates and was attributable to the incomplete humified organic plant residues, indicating that active binding agents were root hairs and fungi. We had similar results; that is, the C:N ratios of aggregate fractions decreased from coarser to finer-sized fractions in all treatments and reflected the mineralization and humification status of the SOM (Fig. 3). The high C:N ratio of the organic matter in macroaggregates was a result of mineral grains and particulate organic matter, primary plant residues with a wide C:N ratios. SOM in macroaggregates is poorly protected against decomposition of microorganisms, releasing less-available organic matter and microbial materials with lower C:N ratios (Callesen et al., 2007). 4.3. Soil C and N recovery and loss from slaking and sieving as affected by agroforestry treatments The resistance of soil aggregates to the slaking and dispersive action of water is important for maintaining a porous soil structure. The lower recovery rate of C and N in the soils might be an effect of its poor resistance against the applied slaking force. However, soils under this region may experience high soil degradation rate due to (a) the lack of multilayer vegetal cover, especially with the monoculture planting patterns and (b) special climate conditions, which cause abundant rainfall. Under these conditions, soil surfaces are periodically removed by soil

C. Chen et al. / Geoderma 299 (2017) 13–24

erosion, and the soil structures are degraded from bigger particles to the smallest least dense particles (small aggregate or clay), consequently resulting in C and N loss by leaching of soluble organic C and by the flowing away of C bonded in small aggregate/clay. While previous studies have reported the risks of carbon losses by runoff and erosion (Palis et al., 1990; Bissonnais et al., 2007), no study to date has observed a loss or degradation of SOC and N in response to the resistance of soil aggregates to the slaking and dispersive action of water using the sieving method. For this study, we found that the more SOC content in the whole soils, the higher recovery of SOC (total mass of aggregate C) observed through the procedure of the bigger aggregates breaking down into smaller ones after shaking and sieving (Figs. 4 and 5). We also observed that the MWD and SOC content in whole soils both showed significantly negative correlations with the proportion of C and N loss, i.e., the more SOC content in the whole soils and the more stable aggregates, the less proportion of C loss occurred during the slaking and sieving procedure (Table 3, Figs. 4 and 5). This could indicate that the C and N in rubberbased agroforestry systems (except for CAAs) were much more stable

21

than in rubber monoculture. The result could be explained by the process of soil physical and chemical properties degradation, which is closely linked with the reduction of its organic matter concentration that is essential to aggregation (Zeytin and Baran, 2003), and the latter is critical to the stabilization of the carbon pool through physical protection within aggregates (Balabane and Plante, 2004). We hypothesized that in the long run, higher residue inputs and enhanced protection of the soil surface would lead to improved aggregation under rubberbased agroforestry systems, meanwhile, physical protection of the aggregates could reduce C loss during the slaking and sieving, thus improving carbon and nitrogen accumulation rates. Specifically, following 10 years of rubber-based agroforestry systems, the addition of interplantings was highly correlated with an increase in soil organic C storage, which is presumably a reflection of an increased C sink in the soil. The persistent binding agents (i.e., root and humic acid moieties) and temporary binding agents (i.e., mycorrhizal hyphae and glycoprotein) were probably significantly increased in the soils, and this presumably led to macroaggregate formation and/or the increase in further production of macroaggregates; meanwhile, it reduced the

Fig. 5. The correlations among the recovery of C, proportion of C loss and C content in bulk soil (a, c, e), and the relationships among the recovery of N, proportion of N loss and N content in bulk soil at all soil depths (b, d, f). Symbols represent plot level means (n = 3), and error bars are SE.

22

C. Chen et al. / Geoderma 299 (2017) 13–24

Fig. 6. The correlations among the recovery of C, proportion of C loss and MWD in bulk soil (a, c), and the relationships among the recovery of N, proportion of N loss and MWD in bulk soil at all soil depths (b, d). Symbols represent plot level means (n = 3), and error bars are SE.

Fig. 7. Schematic diagram illustrating the effect of rubber-based agroforestry on soil aggregation and organic carbon (SOC) dynamics. (a) Conversion of rubber monoculture to rubberbased agroforestry improves soil organic matter due to the inputs of litterfall and roots. (b) Formation of aggregates by diverse biomass-C inputs under rubber-based agroforestry, and (c) processes involved accumulation and redistribution of SOC among aggregate fractions. Plant roots and litterfall inputs make a large contribution to SOC, which play an important role in aggregation. Cone represents the positive correlation between the SOC contents of different aggregate size fractions; dashed lines emphasize the possible inclusion of the microaggregate size fraction (0.053–0.25 mm) within any macroaggregates.

C. Chen et al. / Geoderma 299 (2017) 13–24

breakdown from macroaggregates to microaggregates or to smaller particles (Tisdall and Oades, 1982; Oades, 1984). An increase in the proportion of macroaggregates resulted in an improvement of the physical protection of C provided by these aggregates, followed by a reduction in nutrient turnover rates (increase of C and N) from the soils (Fig. 6). 5. Conclusion The purpose of the study was to examine water stable soil aggregates (WSA), the distribution of SOC and N within the aggregates and soil carbon and nitrogen accumulation in the soils between rubber monoculture and rubber-based agroforestry systems. In this study, N10 years of agroforestry practices have significantly improved the quality of soil from rubber monoculture and particularly increased soil aggregation, enhanced soil carbon and nitrogen accumulation and improved SOC and N distribution within the aggregates. The order of soil physical-chemical properties improvement and the continuity degree of the rubber-based agroforestry systems from highest to lowest were TCAs, DCAs, FMAs and CAAs. Because aggregate formation was associated with increased carbon storage, an increase in the SOC and N content could correspond to increasing aggregate fraction size. Accordingly, the organic carbon and N fractions in different aggregate sizes varied among the five ecosystems. The macroaggregate fraction (N 0.25 mm, especially in the 1–0.25 mm fraction) contained more organic SOC and N than the microaggregate did. The study showed that the proportion of SOC and N loss during wetting sieving and slaking was negatively correlated with WMD, suggesting that the aggregate could physical protect SOC and N from degradation. The study also showed that the proportion of SOC and N loss was negatively correlated with the content of SOC and N in whole soil, suggesting that rubber-based agroforestry systems have a greater potential for C and N sequestration compared with the adjacent rubber monoculture sites. Given that the soil physical-chemical properties improvement and continuity played key roles in managing artificial forests, we recommend that the local government would prefer T. cacao, F. macrophylla and D. cochinchinensis and not C. arabica as the alternative tree species interplanted within rubber plantations (Fig. 7). Acknowledgements We thank Mr. Liu MN, Miss Fu Y, Miss Liu JQ and the Central Laboratory of XTBG for their help. This study was supported by the National Natural Science Foundation of China (31570622 and 41271051), the Natural Science Foundation of Yunnan Province (2013FA022 and 2014HB042), and the Chinese Academy of Sciences 135-Project (KFJEW-STS-084). References Adu, J.K., Oades, J.M., 1978. Physical factors influencing decomposition of organic materials in soil aggregates. Soil Biol. Biochem. 10 (2), 109–115. Balabane, M., Plante, A.F., 2004. Aggregation and carbon storage in silty soil using physical fractionation techniques. Eur. J. Soil Sci. 55 (2), 415–427. Bissonnais, Y.L., 1996. Aggregate stability and assessment of soil crustability and erodibility: I. Theory and methodology. Eur. J. Soil Sci. 47 (4), 425–437. Bissonnais, L.Y., Blavet, D., Noni, D.G., Laurent, J.Y., Asseline, J., Chenu, C., 2007. Erodibility of Mediterranean vineyard soils: relevant aggregate stability methods and significant soil variables. Eur. J. Soil Sci. 58 (1), 188–195. Bruce, R.R., Langdale, G.W., West, L.T., Miller, W.P., 1992. Soil surface modification by biomass inputs affecting rainfall infiltration. Soil Sci. Soc. Am. J. 56 (5), 1614–1620. Budelman, A., 1990. Woody legumes as live support systems in yam cultivation. Agrofor. Syst. 10 (1), 47–59. Buyanovsky, G.A., Aslam, M., Wagner, G.H., 1994. Carbon turnover in soil physical fractions. Soil Sci. Soc. Am. J. 58 (4), 1167–1173. Callesen, I., Raulund-Rasmussen, K., Westman, C.J., Tau-Strand, L., 2007. Nitrogen pools and C:N ratios in well-drained Nordic forest soils related to climate change and soil texture. Boreal Environ. Res. 12, 681–692. Carr, M.K.V., Lockwood, G., 2011. The water relations and irrigation requirements of cocoa (Theobroma cacao L.): a review. Exp. Agric. 47 (04), 653–676. Chaudhary, V.B., Bowker, M.A., O'Dell, T.E., Grace, J.B., Redman, A.E., Rillig, M.C., Johnson, N.C., 2009. Untangling the biological contributions to soil stability in semiarid shrublands. Ecol. Appl. 19 (1), 110–122.

23

Chevallier, T., Blanchart, E., Albrecht, A., Feller, C., 2004. The physical protection of soil organic carbon in aggregates: a mechanism of carbon storage in a vertisol under pasture and market gardening (Martinique, West Indies). Agric. Ecosyst. Environ. 103 (2), 375–387. Chivenge, P., Vanlauwe, B., Gentile, R., Six, J., 2011. Organic resource quality influences short-term aggregate dynamics and soil organic carbon and nitrogen accumulation. Soil Biol. Biochem. 43 (3), 657–666. Cuenca, G., Aranguren, J., Herrera, R., 1983. Root growth and litter decomposition in a coffee plantation under shade trees. Developments in Plant and Soil Sciences. Vol. 71. Springer Netherlands, pp. 477–486. Elliott, E.T., 1986. Aggregate structure and carbon, nitrogen, and phosphorus in native and cultivated soils. Soil Sci. Soc. Am. J. 50 (3), 627–633. Erktan, A., Cécillon, L., Graf, F., Roumet, C., Legout, C., Rey, F., 2016. Increase in soil aggregate stability along a Mediterranean successional gradient in severely eroded gully bed ecosystems: combined effects of soil, root traits and plant community characteristics. Plant Soil 398 (1–2), 121–137. Foster, R.C., 1981. Polysaccharides in soil fabrics. Science 214 (4521), 665–667. Frenkel, H., Goertzen, J.O., Rhoades, J.D., 1978. Effects of clay type and content, exchangeable sodium percentage, and electrolyte concentration on clay dispersion and soil hydraulic conductivity. Soil Sci. Soc. Am. J. 42 (1), 32–39. Gama-Rodrigues, E.F., Nair, P.R., Nair, V.D., Gama-Rodrigues, A.C., Baligar, V.C., Machado, R.C., 2010. Carbon storage in soil size fractions under two cacao agroforestry systems in Bahia, Brazil. Environ. Manag. 45 (2), 274–283. Gregorich, E.G., Greer, K.J., Anderson, D.W., Liang, B.C., 1998. Carbon distribution and losses: erosion and deposition effects. Soil Tillage Res. 47 (3), 291–302. Gupta, N., Kukal, S.S., Bawa, S.S., Dhaliwal, G.S., 2009. Soil organic carbon and aggregation under poplar based agroforestry system in relation to tree age and soil type. Agrofor. Syst. 76 (1), 27–35. Hassink, J., 1997. The capacity of soils to preserve organic C and N by their association with clay and silt particles. Plant Soil 191 (1), 77–87. Hassink, J., Bouwman, L.A., Zwart, K.B., Bloem, J., Brussaard, L., 1993. Relationships between soil texture, physical protection of organic matter, soil biota, and C and N mineralization in grassland soils. Geoderma 57 (1), 105–128. Jastrow, J.D., Miller, R.M., Boutton, T.W., 1996. Carbon dynamics of aggregate-associated organic matter estimated by carbon-13 natural abundance. Soil Sci. Soc. Am. J. 60 (3), 801–807. John, B., Yamashita, T., Ludwig, B., Flessa, H., 2005. Storage of organic carbon in aggregate and density fractions of silty soils under different types of land use. Geoderma 128 (1), 63–79. Jose, S., 2009. Agroforestry for ecosystem services and environmental benefits: an overview. Agrofor. Syst. 76 (1), 1–10. Li, H., Aide, T.M., Ma, Y., Liu, W., Cao, M., 2007. Demand for rubber is causing the loss of high diversity rain forest in SW China. Biodivers. Conserv. 16 (6), 1731–1745. Li, H., Ma, Y., Liu, W., Liu, W., 2012. Soil changes induced by rubber and tea plantation establishment: comparison with tropical rain forest soil in Xishuangbanna, SW China. Environ. Manag. 50 (5), 837–848. Mann, C.C., 2009. Addicted to rubber. Science 325 (5940), 564–566. Nath, T.K., Inoue, M., Myant, H.L.A., 2005. Small-scale agroforestry for upland community development: a case study from Chittagong Hill Tracts, Bangladesh. J. For. Res. 10 (6), 443–452. van Noordwijk, M., Tata, H.L., Xu, J., Dewi, S., Minang, P.A., 2012. Segregate or integrate for multifunctionality and sustained change through rubber-based agroforestry in Indonesia and China. Agroforestry–The Future of Global Land Use, pp. 69–104. Oades, J.M., 1984. Soil organic matter and structural stability: mechanisms and implications for management. Biological Processes and Soil Fertility. Springer Netherlands, pp. 319–337. Palis, R.G., Okwach, G., Rose, C.W., Saffigna, P.G., 1990. Soil erosion processes and nutrient loss. II. The effect of surface contact cover and erosion processes on enrichment ratio and nitrogen loss in eroded sediment. Soil Res. 28 (4), 641–658. Pohl, M., Alig, D., Korner, C., Rixen, C., 2009. Higher plant diversity enhances soil stability in disturbed alpine ecosystems. Plant Soil 324 (1–2), 91–102. Qiu, J., 2009. Where the rubber meets the garden. Nature 457 (7227), 246–247. Ramachandran Nair, P.K., Mohan Kumar, B., Nair, V.D., 2009. Agroforestry as a strategy for carbon sequestration. J. Plant Nutr. Soil Sci. 172 (1), 10–23. van Reeuwijk, L.P., 2002. Procedures for Soil Analysis. sixth ed. International Soil Reference and Information Centre (ISRIC), Wageningen, The Netherlands. Rochester, I.J., 2011. Sequestering carbon in minimum-tilled clay soils used for irrigated cotton and grain production. Soil Tillage Res. 112 (1), 1–7. Shainberg, I., Levy, G.J., Rengasamy, P., Frenkel, H., 1992. Aggregate stability and seal formation as affected by drops' impact energy and soil amendments. Soil Sci. 154 (2), 113–119. Six, J., Elliott, E.T., Paustian, K., 2000. Soil macroaggregate turnover and microaggregate formation: a mechanism for C sequestration under no-tillage agriculture. Soil Biol. Biochem. 32 (14), 2099–2103. Tisdall, J.M., Oades, J., 1982. Organic matter and water-stable aggregates in soils. J. Soil Sci. 33 (2), 141–163. Udawatta, R.P., Kremer, R.J., Adamson, B.W., Anderson, S.H., 2008. Variations in soil aggregate stability and enzyme activities in a temperate agroforestry practice. Appl. Soil Ecol. 39 (2), 153–160. Viswanathan, P.K., Shivakoti, G.P., 2008. Adoption of rubber-integrated farm-livelihood systems: contrasting empirical evidence from the Indian context. J. For. Res. 13 (1), 1–14. Vodnik, D., Grčman, H., Maček, I., Van Elteren, J.T., Kovačevič, M., 2008. The contribution of glomalin-related soil protein to Pb and Zn sequestration in polluted soil. Sci. Total Environ. 392 (1), 130–136. Vogel, A.W., Wang, M.Z., Huang, X.Q., 1995. People's Republic of China: Reference Soil (Lotosal) of Tropical Southern Yunnan Province. Institute of Soil Science–Academica

24

C. Chen et al. / Geoderma 299 (2017) 13–24

Sinica, Nanjing, and International Soil Reference and Information Centre, Wageningen. Wick, A.F., Stahl, P.D., Ingram, L.J., 2009. Aggregate-associated carbon and nitrogen in reclaimed sandy loam soils. Soil Sci. Soc. Am. J. 73 (6), 1852–1860. Woods, L.E., Schuman, G.E., 1988. Cultivation and slope position effects on soil organic matter. Soil Sci. Soc. Am. J. 52 (5), 1371–1376. Xu, J., Grumbine, R.E., Beckschäfer, P., 2014. Landscape transformation through the use of ecological and socioeconomic indicators in Xishuangbanna, Southwest China, Mekong Region. Ecol. Indic. 36, 749–756.

Yoder, R.E., 1936. A direct method of aggregate analysis of soils and a study of the physical nature of erosion losses. Agron. J. 28 (5), 337–351. Zeytin, S., Baran, A., 2003. Influences of composted hazelnut husk on some physical properties of soils. Bioresour. Technol. 88 (3), 241–244. Zhang, M., Fu, X.H., Fang, W.T., Zou, X., 2007. Soil organic carbon in pure rubber and tea-rubber plantations in southwestern China. Trop. Ecol. 48 (2), 201. Ziegler, A.D., Fox, J.M., Xu, J., 2009. The rubber juggernaut. Science 324 (5930), 1024–1025.