Introduction

Soil organic matter (SOM) affects soil physical, chemical, and biological properties and hence is regarded as a key attribute of soil fertility (Paustian et al. 1997; Verma et al. 2010). Meanwhile, SOM plays an important role in mitigating elevated atmospheric CO2 level (Lal 2004). In agricultural soils, land use change or management practices can disrupt SOM dynamic equilibrium and produce a marked effect on soil organic C stock (Lal 2004; Leifeld and Kögel-Knabner 2005). Paustian et al. (1997) estimated that global cultivated soils released about 50 Pg C through the mineralization of SOM. Depletion of the SOM caused a decline in soil fertility and crop productivity and increased atmospheric CO2 concentration (Lal 2004; Banerjee et al. 2006). Therefore, appropriate management practices on agricultural soils should be adopted for increasing SOM and, thus, sustaining soil productivity and mitigating global climate change.

Agroforestry, in which trees are grown in association with crops and/or pasture and livestock on a land use system, is believed to increase SOM and improve soil fertility (Lal 2004). Compared with monocropping systems, trees in agroforestry systems can enhance the SOM level by adding the quantity of aboveground and belowground organic matter inputs to soils (Jose 2009; Nair et al. 2009). Moreover, the effect of agroforestry practices on SOM varies with soil texture (e.g., clay or clay plus silt; Six et al. 2002; Gupta et al. 2009). However, at least 10 years after the establishment of agroforestry systems is necessary to detect a significant change in SOM in tropical regions (Young 1997), and more time is required in temperate regions due to the lower organic matter inputs and slower SOM turnover rate (Oelbermann et al. 2004). Among the different pools of SOM (labile, slow, and recalcitrant), labile fractions of SOM, such as microbial biomass and particulate organic matter (POM), have a shorter turnover time than total SOM and are more sensitive to changes in land management practices (Cambardella and Elliott 1992; Leifeld and Kögel-Knabner 2005; Mungai et al. 2006; Zhang et al. 2006). Previous studies found that changes in soil microbial biomass (Wang et al. 2005) and POM (Marquez et al. 1999) occurred in a relatively short time after the establishment of agroforestry systems, although there was no significant change in total SOM. Therefore, identifying the changes in these labile SOM fractions may provide early information about the effects of agroforestry practices and soil texture on SOM dynamics.

Since 2000, the Grain-for-Green Program has been performed to control soil erosion and arrest soil degradation in arid and semiarid areas of northern China (Li 2004). Consequently, agroforestry systems have been widely adopted by smallholder farmers due to their dual roles in generating income and maintaining environmental services. However, there is little information on the changes in labile SOM fractions following land use change from monocropping to agroforestry systems in these regions. Knowledge about the effects of agroforestry systems on labile SOM fractions is needed to better understand SOM dynamics in these regions.

In this study, we investigated soil microbial biomass C (MBC), microbial biomass N (MBN), particulate organic matter C (POMC), particulate organic matter N (POMN), as well as soil total organic carbon (TOC) and total N (TN) in 4-year-old poplar (Populus euramericana cv. ‘N3016’)-based agroforestry systems and monocropping systems with two different soil textures (sandy loam and sandy clay loam) in a semiarid region of Northeast China. The main objective of our study was to assess the effect of agroforestry practices on labile SOM fractions in both sandy loam and sandy clay loam soils. We hypothesized that, irrespective of soil texture, soil microbial biomass and particulate organic matter would increase following land use change from monocropping to poplar-based agroforestry systems due to enhanced organic matter input to soils.

Materials and methods

Study site and experimental design

The study site was located in Jianping County (40°17′–42°20′ N, 119°10′–120°02′ E), west of Liaoning Province, Northeast China. The study site has a semiarid temperate monsoon climate with a mean annual temperature of 6.5°C, precipitation of 467 mm (more than 60% falling between June and August), and the frost-free period of about 148 days per year. The soil is classified as Entisols of suborder Fluvent according to the US Soil Taxonomy, and it mainly includes two types of texture: sandy loam and sandy clay loam. In the study site, agroforestry systems have been established on marginal agricultural lands since 2000 due to the Grain-for-Green Program.

In order to examine the effect of agroforestry practices on labile SOM fractions, poplar-based agroforestry systems were established on marginal agricultural lands in both sandy loam (sand, silt, and clay fractions of 80%, 11%, and 9%, respectively) and sandy clay loam (sand, silt, and clay fractions of 64%, 13%, and 23%, respectively) soils in April 2005. Agricultural lands were flat and shared similar land management history. Based on the information gathered through conversation with local inhabitants, agricultural lands had been cultivated for more than 60 years before establishing agroforestry systems. For this study, three pairs of agroforestry and monocropping systems were arranged in each soil texture, and the agroforestry system and the corresponding monocropping system were in a neighborhood. In agroforestry systems, nursery-raised 2-year-old poplar seedlings were planted in pits of 40 × 40-cm size and 40-cm depth at a spacing of 3 × 4 m, and the stand density was 833 trees per hectare. Six parallel rows of crops were planted within the tree rows at spacing of 50 cm; the first row was 75 cm from the tree row. Since 2007, pruning of tree branches in agroforestry systems was conducted every year after the leaves fell, and the pruning materials were removed from the systems. Other management practices including cover crops, irrigation, fertilization, and weed control were conducted according to the conventional methods used by local farmers. In 2005–2008, the cropping sequences in the agroforestry and monocropping systems were sorghum (Sorghum bicolor), millet (Setaria italica), and 2 years of maize (Zea mays), respectively. All the sites were tilled to about 15-cm depth before seeding by using a tractor pulling a semi-mounted plow. Crops were usually sowed in late April to early May and harvested in late September to early October. Before sowing, compound fertilizers containing 60 kg N ha−1, 60 kg P2O5 ha−1, and 30 kg K2O ha−1 had been applied every year. During crop growth periods, 60, 60, and 90 kg N ha−1 had been applied for sorghum, millet, and maize, respectively.

Sample collection and analyses

In order to avoid direct effects of current year’s management practices (e.g., tillage, irrigation, and fertilization) on labile SOM fractions, soil sampling was conducted in April 2009 when fields were fallow. For each pair of treatments, a plot of 12 × 21 m was identified for agroforestry system and a plot of 15 × 15 m for monocropping system. Each pair of sampling plots was within a distance of 10–20 m. In agroforestry systems, poplar diameter at breast height (DBH, 1.3 m) and height were measured for each plot during soil sampling. Based on DBH and tree height, tree aboveground biomass (including stem and branch) was calculated using species-specific allometric equations (Liang et al. 2006). Detailed information about the allometric equations has been given in Mao and Zeng (2010). In each plot, poplar litter mass was collected from five randomly selected quadrats (50 × 50 cm), placed in paper bags, returned to the laboratory, oven-dried to a constant mass at approximately 65°C for 48 h, and then weighed. Because crop residues were generally removed from the field during soil tillage, we did not measure the mass of crop residues in agroforestry and monocropping systems. Detailed information about poplar stand characteristics in agroforestry systems is shown in Table 1.

Table 1 Stand characteristics of 4-year-old poplar-based agroforestry systems in a semiarid region of Northeast China
Full size table

In the preliminary study, we found that selected soil properties showed no significant spatial variations with distances from the tree rows (data not shown). Therefore, in the present study, six soil cores (4 cm in diameter) were randomly collected at the 0- to 15-cm and the 15- to 30-cm layers in each plot following removal of surface litter and thoroughly mixed by layer to form a composite sample. After removing plant roots, macrofauna, and visible plant debris, the soil was sieved through a 2-mm mesh screen and divided into two subsamples. One subsample was stored at 4°C until analysis for soil MBC and MBN and the second subsample air-dried and stored at room temperature for the determination of TOC, TN, POMC, and POMN. Air-dried soil samples were ground to pass through a 0.25-mm sieve for the measurement of TOC and TN. The TOC concentration was determined by the wet oxidation method with K2Cr2O7 and concentrated H2SO4 (Walkley and Black 1934). To determine TN, the samples were first digested using the Kjeldahl method (Bremner 1996). The TN concentration in the digested solution was measured using a continuous-flow autoanalyzer (AutoAnalyzer III, Bran+Luebbe GmbH, Germany).

Soil MBC (Vance et al. 1987) and MBN (Cabrera and Beare 1993) were measured by the chloroform fumigation–extraction method. Fumigated and non-fumigated soils (equivalent to 20 g dry weight) were extracted with 50 mL 0.5 mol L−1 K2SO4 for 30 min and filtered. Organic C and TN in fumigated and non-fumigated extracts were measured by the dichromate oxidation method (Vance et al. 1987) and alkaline persulfate oxidation method (Cabrera and Beare 1993), respectively. Soil MBC was calculated by dividing the difference in extractable organic C concentration between fumigated and non-fumigated soils with a conversion factor of 0.38 (Vance et al. 1987), and MBN was calculated by dividing the difference in TN concentration between fumigated and non-fumigated soils with a conversion factor of 0.45 (Cabrera and Beare 1993).

Soil POMC and POMN were determined by the method of Cambardella and Elliott (1992). A 20-g air-dried soil sample was dispersed in 50 mL 5 g L−1 sodium hexametaphosphate solution on a reciprocal shaker, and then the soil suspension was passed through a 0.053-mm sieve and rinsed several times with distilled water. The material remaining on the sieve was defined as POM. Soil POM was transferred to a glass beaker, oven-dried at 50°C for 24 h, weighed, ground, and passed through a 0.25-mm sieve for the measurement of organic C and TN. Organic C and TN were determined by the methods described above.

Statistical analyses

Data were tested for normality using the Kolmogorov–Smirnov test, and all data followed a normal distribution (data not shown). At each soil layer, a paired t test was used to determine significant differences in soil properties between land use types for the same soil texture, and an independent samples t test was used to examine significant differences in soil properties between soil textures for the same land use type. Data were statistically analyzed using SPSS 13.0 for Windows software package; the accepted significance level was α = 0.05.

Results

In poplar-based agroforestry systems, there were no significant differences in tree aboveground biomass and litter mass between sandy loam and sandy clay loam (Table 1). Land use change had no significant effects on TOC, TN, and C/N ratio (Table 2). For each soil texture, no significant differences in TOC, TN, and C/N ratio existed between agroforestry and monocropping systems. For each land use type, sandy loam had lower TOC and TN than sandy clay loam at each 15- to 30-cm layer (Table 2).

Table 2 Soil TOC and TN under monocropping and 4-year-old agroforestry systems in a semiarid region of Northeast China
Full size table

Land use change from monocropping systems to poplar-based agroforestry systems increased soil MBC both in the 0- to 15-cm and 15- to 30-cm layers, irrespective of soil texture (Fig. 1). In sandy loam soils, agroforestry systems had a higher percentage of MBC to TOC (MBC/TOC) than monocropping systems at each soil layer (Table 3), but only had greater soil MBN at the 15- to 30-cm layer (Fig. 1). For each land use type, soil MBC in sandy clay loam was significantly greater than that in sandy loam (Fig. 1). In addition, at the 15- to 30-cm layer, monocropping systems in sandy clay loam had higher MBN than that in sandy loam (Fig. 1), whereas agroforestry systems in sandy clay loam had lower MBC/TOC and percentage of MBN to TN (MBN/TN) than that in sandy loam (Table 3).

Fig. 1
figure 1

Soil microbial biomass carbon (MBC) and nitrogen (MBN) in monocropping and 4-year-old poplar-based agroforestry systems. Data are the mean values; vertical bars are standard errors (n = 3). For each soil layer, means with different capital and lowercase letters are significantly different between land use types for each soil texture and between soil textures for each land use type, respectively, at P < 0.05

Full size image
Table 3 Soil MBC/TOC, MBN/TN, POMC/TOC, and POMN/TN under monocropping and 4-year-old agroforestry systems in a semiarid region of Northeast China
Full size table

Agroforestry systems had lower soil POMC, POMN, and the percentage of POMC to TOC (POMC/TOC) than monocropping systems only at the 0- to 15-cm layer in sandy loam soils (Fig. 2 and Table 3). In sandy clay loam soils, agroforestry systems generally had higher soil POMC, POMN, POMC/TOC, and the percentage of POMN to TN (POMN/TN) than monocropping systems at the 0- to 15-cm layer and had greater POMN and POMN/TN at the 15- to 30-cm layer (Fig. 2 and Table 3). For the monocropping systems, sandy loam had greater POMC/TOC and POMN/TN than sandy clay loam in both the 0- to 15-cm and the 15- to 30-cm layers (Table 3) and only had higher POMC and POMN in the 0- to 15-cm layer (Fig. 2). However, agroforestry systems established on sandy loam only had greater POMC/POMN than those established on sandy clay loam at the 15- to 30-cm layer (Fig. 2).

Fig. 2
figure 2

Soil particulate organic carbon (POMC) and nitrogen (POMN) in monocropping and 4-year-old poplar-based agroforestry systems. Data are the mean values; vertical bars are standard errors (n = 3). For each soil layer, means with different capital and lowercase letters are significantly different between land use types for each soil texture and between soil textures for each land use type, respectively, at P < 0.05

Full size image

Discussion

Our results partly supported the hypothesis that soil microbial biomass increased following land use change from monocropping to agroforestry systems. Poplar-based agroforestry practices increased MBC both in sandy loam and sandy clay loam soils and generally had no effect on soil MBN. In poplar-based agroforestry systems, the enhanced organic matter inputs to soils and improved understory microclimate (Amatya et al. 2002; Lee and Jose 2003) may account for the increased soil MBC. In addition, no significant difference in soil MBN between agroforestry and monocropping systems is probably attributed to the high amounts of N fertilizer application (120–150 kg N ha−1 year−1; Kaur et al. 2000) or shift in soil microbial community structure (Chang and Trofymow 1996).

Inconsistent with our hypothesis, poplar-based agroforestry practices only increased soil POMC and POMN in sandy clay loam soils, but did not enhance soil POMC and POMN in sandy loam soils and even led to declines in soil POMC and POMN at the 0- to 15-cm layer. This may result from the difference in soil clay or silt plus clay content (Six et al. 2002). Soil POM represents undecomposed plant residues and partly decomposed plant material at an early stage of decomposition (Leifeld and Kögel-Knabner 2005). Soil texture exerts an influence on the protection of POM by directly incorporating POM into microaggregates and indirectly affecting aggregate dynamics (Six et al. 2002). Soils with high clay or silt plus clay content can effectively protect POM from decomposition and enhance POMC and POMN accumulation in soils (Cote et al. 2000). Moreover, Bauhus et al. (1998) found that the specific microbial respiration (the amount of CO2–C respired per unit of microbial biomass C) in clay soils (80% clay content) was half that in tilled soils (46% clay content). Therefore, following a land use change from monocropping to agroforestry systems, higher clay content in sandy clay loam soils might protect POM against decomposition and reduce the amount of organic C and N mineralized, thus resulting in a greater soil POMC and POMN. In sandy loam soils, the declines in soil POMC and POMN in agroforestry systems may be explained by the relative low protection of POM due to the coarse soil texture (Bauhus et al. 1998; Cote et al. 2000) as well as the priming of soil labile C induced by fresh organic matter inputs (Fontaine et al. 2004) and/or N fertilizer application (Wander et al. 2007).

In the present study, no significant differences in TOC and TN were found between 4-year-old poplar agroforestry systems and monocropping systems for both sandy loam and sandy clay loam soils in a semiarid region of Northeast China, although there was increased organic matter input to soils in agroforestry systems. Oelbermann et al. (2006) and Peichl et al. (2006) also found that TOC did not significantly increase in 13-year-old hybrid poplar (Populus deltoids × nigra DN-177) and Norway spruce (Picea abies) agroforestry systems in southern Canada, respectively. In poplar-based agroforestry systems, increased soil organic matter inputs through tree leaf and root litters may be offset by the accelerated decomposition of organic matter induced by site preparation (Mao et al. 2010), the priming effect of fresh organic carbon supply (Fontaine et al. 2004), and annual tillage practices (Lal 2004). Considering that the effect of agroforestry system on TOC and TN depends on the age of the system (Oelbermann et al. 2004), further studies are needed to understand the temporal changes in TOC and to quantify the effects of stand age on TOC dynamics in semiarid temperate agroforestry systems.

Soil MBC and POMC play an important role in the formation and stabilization of soil aggregates (Liao and Boutton 2008; Gartzia-Bengoetxea et al. 2009), which are critical in the protection and accumulation of soil organic carbon (Six et al. 2002). Soil POM not only acts as a nucleus for macroaggregate formation with material accumulating around the POM but also is an important agent in binding microaggregates to form macroaggregates (Bronick and Lal 2005). Meanwhile, soil microbial secretions and residues are critical in the maintenance of soil aggregate stability and soil physical structure (Liao and Boutton 2008). Increased MBC and POMC may help the formation and maintenance of soil aggregate structure. Moreover, soils with high clay content may promote aggregate formation (Six et al. 2002). Therefore, compared with sandy loam soils, poplar-based agroforestry systems established on sandy clay loam soils can protect SOM against decomposition and may favor organic matter accumulation in soils in semiarid regions of Northeast China. Actually, Gupta et al. (2009) observed that poplar-based agroforestry systems sequestered more C in soils with greater clay content in a semiarid region of northwest India. Our results suggest that agroforestry systems established on soils with higher clay content may promote the sequestration of soil organic carbon in semiarid regions of Northeast China.

In our study, poplar aboveground biomass and litter mass in agroforestry systems did not vary with soil texture (Table 1), albeit sandy clay loam had greater TOC and TN than sandy loam (Table 2). In poplar-based agroforestry systems, the amount of nutrient (N, P, and K) provided by fertilizer application during the growing season may be greater than that supplied by mineralizing of organic matter or weathering of parent material in soils. Therefore, poplar stand characteristic showed no difference between sandy loam and sandy clay loam.

In conclusion, land use change from monocropping to poplar-based agroforestry systems influenced labile SOM fractions in semiarid regions of Northeast China, although there was no significant change in SOM. Moreover, the effects of agroforestry practices on labile SOM fractions varied with soil texture. Our results suggest that labile SOM fractions are more sensitive than total SOM to land use change and can provide early information about the changes in SOM in semiarid regions of Northeast China. Further studies should be conducted to assess the effects of stand age and soil texture on SOM dynamics in semiarid temperate agroforestry systems.