Soil Aggregates Stability and Carbon Density Under Three Plantations on Loess Plateau, China

[Background] This paper aimed to research the inuence of types of plantations on soil aggregate stability and carbon density. The soils of Robinia pseudoacacia, Pinus tabulaeformis and M. pumila plantations were selected as the research objects in the south of Loess Plateau. The mass percentages of soil water stable aggregates with different sizes were measured by fast wetting method, and then, the mean weight diameter (MWD) was calculated to asses soil aggregate stability. Soil organic carbon, inorganic carbon and bulk density were analyzed using traditional analysis methods. The roles played by soil organic and inorganic carbon in the formation of water stable aggregates were discussed. [Results] The content of soil organic carbon and MWD was signicantly higher in P. tabuliformis plantation compared with R. pseudoacacia and M. pumila plantations. Conversely, the content of soil calcium carbonate and total carbon density was the lowest in P. tabuliformis plantation. Soil organic carbon was the highest at the surface soils (0-10 cm), which was 2.21-3.96, 2.10-3.00 and 1.65-2.53 times compared with subsurface soils in Robinia pseudoacacia, Pinus tabulaeformis and M. pumila plantations respectively. While soil carbonate calcium and inorganic carbon density uctuated with the increasing of soil layers. Aggregate organic carbon increased while carbonate calcium content decreased with the increasing of aggregate size respectively. [Conclusion] We concluded that: (1) P. tabuliformis played a more important role in soil organic carbon sequestration compared with R. pseudoacacia and M. pumila; (2) there was a threshold value of carbonate calcium content, when it was lower or higher than this threshold, it had positive or negative correlation with MWD; (3) bigger water stable aggregates had higher soil organic matter and lower carbonate calcium. R. pseudoacacia plantation. The MWD did not differ signicantly at all layers under M.pumila plantation.


Introduction
The area of plantation of trees in China has reached to 61.68 million hectares, coming the rst in the world. In the study area, in order to accommodate degradation of soils, arid and cold climate, exotic species Robinia pseudoacacia, which is fast growing and aid N-xation, has been introduced for a long time; and native species Pinus tabulaeformis has also been chosen as predominant reproducing tree species. Meanwhile, Malus pumila was also planted widely by local farmers in order to seek higher economic pro ts than others (such as Triticum aestivum and Zea mays). Each plant species provides a different quality and quantity of organic material into soils, causing changes of soil properties [9]. Chen et al. found the age-related ne root biomass [13] and C: N: P stoichiometry [12] were signi cantly different between P. tabulaeformis and R. pseudoacacia in the Loess Plateau. Bhattacharya et al. [23] showed tree species signi cantly changed soil properties through roots and litters. Zhang et al. [26] concluded that P. tabulaeformis plantation greatly increased plant residues and vegetation coverage, thereby improving soil micro-environment and ultimately improving physical-chemical properties and biological activities of soil.
Mataix-Solera et al. [9] concluded that compared with other species, Pinus halepensis should be planted in afforestation projects in semi-arid areas, because this species could produce more hydrophobic substances into the soil than other species. However, Liu et al. [35] found that broadleaved forests can most obviously improve soil quality, followed by broadleaved coniferous mixture and coniferous plantations. In addition, Nie et al. [27] found plant species had no signi cant impact on soil organic carbon concentration. Cao et al. [32] also found that the differences of soil nutrients between black locust and Chinese pine were not obvious.
Afforestation has been seen as an important role to reduce soil erosion and improve soil quality, which is usually expressed by the content of soil organic carbon (SOC) or soil organic matter (SOM). Many parameters including bulk density and aggregate stability correlated well with SOC or SOM. Aggregate stability, indicating soil structural stability, is often considered as an important soil quality [8]. It was believed that the SOC content played an important role in the soil aggregation [27]. In addition to, aggregate size distribution and stability can also be used to indicating of soil degradation [1]. Mean weight diameter (MWD) of water stable aggregates has been recognized as an important indicator on the capability of the soil to resist against water erosional forces. Meanwhile, stable aggregate played an important role in the stabilization of SOC [33]. Micro-aggregates are bound together by young organic matter into larger macro-aggregates and a breakdown of macro-aggregates results in a release of labile SOM [4]. It is generally accepted that the stability of SOC in soil aggregates is closely related to the protection of aggregate structure from microbial decomposition [33].
The focus of afforestation changed to carbon stocks in the last decades [15]. Among the numerous sources of greenhouse gases, emissions of CO 2 are affected by changes of land use [23]. Soil carbon consists of soil organic carbon (SOC) and soil inorganic carbon (SIC) [6]. Determining changes in soil organic carbon (SOC) and inorganic carbon (SIC) caused by plantation is important for estimating the regional carbon budget and evaluating ecological effects [30]. The choice of tree species plays an important role in SOC accumulation. The Populus plantation was main carbon sink with a carbon sequestration rate of 9.50 t C/ (ha·yr) in Beijing [36]. Li and Liu [25] believed that soil was a great reservoir for C storage in black locust plantation. The quantitative contribution of plantation types to C stocks is much debated. For example, Cao et al. [31] found that the SOC densities of the N-xing black locust plantations were signi cantly lower than those of the Chinese pine plantations and secondary oak forests. However, Wang et al. [5] found that after 23 years' growth, N-xing species performed better in restoring soil carbon. Tong et al. [28] concluded that the total soil C stock in a layer of 100 cm was in the order: Robinia pseudoacacia > Populus tomentosa > Caragana korshinskii in the Loess hill region of China.
Therefore, the objective of the study was to quantifying aggregate stability, organic and inorganic carbon density in soil pro les of R. pseudoacacia, P. tabuliformis and M. pumila plantations. We hypothesized that: (1) plantation type and soil layers would affect soil properties such as soil aggregate stability, soil organic and inorganic carbon content and (2) content of organic carbon and inorganic carbon was different in water stable aggregates with different sizes. The study will contribute to tree species selection in afforestation as carbon sink increase as well as water and soil conservation measures.

Objects And Methods
Study area and soil sampling.
The study area is located in the north of Liquan County situating the south region of Loess Plateau, China (Fig. 1). The mean annual temperature is 12.6℃ with the highest monthly temperature in July (34℃) and the lowest in January (-4℃) and with 214 frost-free days. The mean precipitation is 517.6 mm. After level-terrace site preparation on the same slope, Robinia pseudoacacia and Pinus tabuliformis were planted fty years ago, Pinus tabuliformis plantation is above Robinia pseudoacacia plantation. And the Malus pumila plantation was planted thirty years ago at the level ground on bottom of the slope. The studied soil is named cinnamon soil in the Chinese genetic classi cation [37], which has been developing from loess parent material. The mean diameter at breast height (DBH) of Malus pumila, Robinia pseudoacacia and Pinus tabuliformi was 12.98 cm, 12.73 cm and 15.42 cm respectively, the average pH of surface soil under the site of them is 8.42, 8.33 and 7.97 respectively. During September 2018, a plot was chosen in the central area of each plantation, in which three soil pro les were digged into 50 cm (in Malus pumila plantation) or 80 cm (in Robinia pseudoacacia and Pinus tabuliformis plantations) with intervals of 10 cm using a spade for sampling. Soil samples were air-dried for one-week and stored at room temperature.
Experimental design. 5-10 g dry aggregates with 2-5 mm diameter were gently submerged into distilled water for 30 minutes and then washed onto a 0.05 mm sieve using ethyl alcohol. After shaking with hands about 15 times all soil particles with different diameters were washed into a 200 mL beaker using ethyl alcohol, and separated by moving a cascade of sieves with openings of 5, 2, 1, 0.5, 0.2, 0.1 mm after drying at 40℃.The water stable aggregates with the diameters of < 0.1 mm, 0.1-0.2 mm, 0.2-0.5 mm, 0.5-1.0 mm,1.0-2.0 mm and > 2.0 mm were weighed and mass percentages were calculated [34]. The treatment was replicated three times.
SOM concentrations of the aggregate fractions and the bulk samples were determined using the oil bath-K 2 Cr 2 O 7 titration method. Bulk density was measured with a standard technique using a cutting ring driven vertically downward into the midrange of each horizon. The content of inorganic carbon was determined by volume of carbon dioxide from reaction with hydrochloric acid [14].

Data analysis.
Mean weight diameter (MWD) was calculated using the following formula. Where (in millimetre) is a mean diameter of two consecutive sieves, and w i is corresponding mass percent.
Soil organic carbon density (hereafter: SOCD) and soil inorganic carbon density (hereafter: SIOCD) at different soil layers were computed using the following equations: Where SOC (i) and SIOC (i) is the content of SOC(%) and 0.12* CaCO 3 (%) in the layer i respectively, BD(i) is the soil bulk density (g/cm 3 ) in the layer i, H(i) is the soil layer's thickness(10 cm).
The diagrams in the paper were drawn with ggplot2 packages of R [20] and Excel 2013. SPSS software (version 16.0) (SPSS Inc., Chicago, IL, USA) was used to perform the statistical and variance analyses. One-way ANOVA followed by LSD's test was used to compare the signi cance of difference of MWDs among the different soil layers under the same plantation and among the three plantations at the same soil layer. An independent t-test was used to test the differences between R.pseudoacacia and P. tabuliformis plantations in the 50-80cm soil layers.

Results
Aggregates stability characteristics.
Plantation types in uenced signi cantly the MWDs of soil water stable aggregates (Fig. 2). MWD was 0.18-0.27 mm, 0.35-0.57 mm and 0.11-0.15 mm under R.pseudoacacia, P. tabuliformis, and M.pumila plantations respectively. The MWD among three plantations differed signi cantly at all soil layers. The MWD of P. tabuliformis was 22-67% greater than that of R. pseudoacacia across the 0-80 cm layers with 10 cm intervals. MWD was 39% and 70% higher in R. pseudoacacia and P. tabuliformis than in M.pumila across 0-50cm layers. The variation characteristics of MWD in soil pro le under three types of plantation were different. The MWD of 20-80cm is higher than that of 0-20cm under the P. tabuliformis plantation. The MWD of 0-10cm was the highest, followed by 10-30cm, and that of 30-80cm was the lowest under R. pseudoacacia plantation. The MWD did not differ signi cantly at all layers under M.pumila plantation.
The content of SOC and calcium carbonate (CaCO 3 ).
Our ndings showed that 26%, 31% and 31% of SOC were distributed in the top soil (0-10 cm) in P. tabuliformis, R. pseudoacacia and M.pumila plantations respectively. Meaningwhile, the differences of SOC content were smaller in the subsurface layers (10-60 cm) than in the surface layer (0-10 cm) between plantations of R.pseudoacacia and M.pumila. However, in all soil layers, the SOC content was signi cantly the highest in the P. tabuliformis plantation than in the R. pseudoacacia and M.pumila plantations (Fig. 3A). Bulk density characteristics.
The bulk density varied between 1.41 g/cm 3 and 1.56 g/cm 3 in the M.pumila plantation, between 0.91 g/cm 3 and 1.31 g/cm 3 in P. tabuliformis plantation, and from 1.16 g/cm 3 to 1.48 g/cm 3 in R. pseudoacacia plantation (Fig. 3C). Bulk density was always higher in M.pumila plantation than in the two others in the whole soil layers.
Soil organic carbon density and soil inorganic carbon density.
The content of soil organic matter and inorganic carbon in water stable aggregate with different size.
The CaCO 3 content declined successively with increasing aggregate size and was the lowest in > 2 mm fraction (Fig. 5). Inversely, the SOM content increased linearly with the increase in aggregate size, and the highest was in > 2 mm fraction (3.93%) (Fig. 5). It was obvious that there was a tradeoff between SOM and CaCO 3 content.

Discussion
An evaluation of the changes in soil properties as a consequence of forest planting is important, especially for fragile ecological areas [2]. Our results supported the hypothesis that bulk density, aggregates stability and the content of carbon were affected by tree species, and they also changed obviously among different soil layers. Additionally, water stable aggregates with bigger sizes had more soil organic matter content, but less calcium carbonate content. According to the Kyoto protocol, forest planting is an important action to build a soil carbon pool in terrestrial ecosystems [21]. The total carbon density was in the order of R. pseudoacacia > M.pumila > P. tabuliformis in all soil layers. Compared with M. pumila and P. tabuliformis plantations, P.tabuliformis plantation was more conducive to accumulation of soil organic carbon, decrease of soil inorganic carbon and bulk density, and then could improve the water stability of soil aggregates, which will facilitate the improvement of soil quality and protection of soil and water from loss.
Changes in soil properties among three plantations.
Soil aggregate stability is considered to be a property that provides information on soil quality [11]. In addition, many studied soil parameters aligned with SOC [8], especially such as aggregate water stability. It was observed that water stability of soil aggregates and SOC content both were in the order of P. tabuliformis > R. pseudoacacia > M.pumila, indicating that soil quality and related soil functions varied greatly among three plantations. Similarly, Chen et al. [13] found that soil organic carbon was higher in 10-year-old P. tabulaeformis than in 10-year-old R.pseudoacacia stands. In our study, it was demonstrated that P. tabuliformis would be more conducive to increase of SOC and improvement of aggregate stability compared with R. pseudoacacia and M.pumila. This is in agreement with results found by previous studies, which observed Pinus plantation could supply soil with considerably more organic material and improve water repellency of soil [3,22]. Increase in SOC was mostly associated with an increase in soil hydrophobicity, especially under the wax/aromatic oil rich litter of the Pinus halepensis trees [9]. Moreover, the presence of water repellency can play an important role in the formation and stabilization of aggregates and can avoid high levels of soil degradation [11]. From this, it seemed logical that soil quality indicators including SOC and aggregate stability have been improved by P. tabuliformis plantation in the present study. Additionally, the ne root biomass (FRB) and ne root production (FRP) of P.tabulaeformis were greater than those of R. pseudoacacia [13]. Small and ne roots produced optimal conditions to form and stabilize aggregates due to the polysaccharides being produced by the microorganisms [1]. Meanwhile, BD was signi cantly related to most other soil parameters and it could be used as an indicator of soil structure [2]. The soil BD was in the order of P. tabuliformis < R. pseudoacacia < M.pumila, which indicated there was negative variation trend between BD and MWD or SOC. Our study demonstrated that the increase of SOC corresponded to the improvement of aggregate stability and decrease of BD, which was most signi cant in P. tabulaeformis plantation, and R. pseudoacacia plantation followed. At M. pumila plantation, MWD of water stable aggregates was the smallest (Fig. 2), which was similar with the lowest SOC (Fig. 3A). It was visible that M.pumila plantation had the lowest values of SOC and MWD, but the highest bulk density. This could be explained by long term cultivation management such as pruning and weed control resulting in the lower input and higher decomposition of organic substances. Lal [21] also believed that most soil under the managed ecosystems contained a lower SOC pool than their counterparts under natural ecosystems due to the depletion of the SOC pool in cultivated soil. Our ndings also showed that soil organic matter was largest signi cantly in the surface layer (0-10 cm) compared to the other soil layers. This was in agreement with previous studies conducted in other forest ecosystems [24], and was a logical result as the surface layer is the main place of soil organic matter sources such as dry branches and fallen leaves.
Effects of tree species on soil carbon sequestration.
It is also becoming increasingly clear that carbon accumulation in soil represents an important carbon stocks [24]. The quantitative relationship between the changes of SOC and SIC stocks in deep pro les following vegetation restoration should be further determined [30]. Our results showed SOCD ranged from 0.64-2.63 kg/m 2 and SIOCD ranged from 0.37-4.08 kg/m 2 over the 0-80 cm soil pro les. Compared with P. tabulaeformis, The STCD increased by 59%-100% and 3% -46% in R. pseudoacacia and M. pumila plantations respectively. SIOCD to STCD ratios in R. pseudoacacia and M. pumila plantations were 56-85% and 53-76%. Zethof et al. [8] and Wang et al. [29] also found inorganic carbon content was often much larger than that of organic carbon in semi-arid regions. SIOCD to STCD ratios were lower (12%-55%) under P. tabulaeformis plantation, which was due to coniferous trees having more acid exudate compared with broadleaf trees. And then, more carbonate would be dissolved [29]. In addition, higher organic matter accumulation under P. tabulaeformis plantation increased saturated hydraulic conductivity [2], which also increased carbonates leaching. The content of CaCO 3 and SIOCD uctuated with deepening soil layers, showing there was more inorganic carbon content in subsoils. Han et al. [30] also con rmed that the maximum soil in-organic carbon (SIC) values was at 60-100 cm soil layer. SIC content at subsoil increased signi cantly due to the dissolution and leaching of carbonates from topsoil and the subsequent precipitation in subsoils [6]. There was a lot more carbon in deep soil than we once thought, and the underlying processes inhibiting its turnover are still largely unknown [17].
Aggregate-CaCO 3 content decreased with increasing aggregate size.
In many semi-arid regions, where the presence of carbonates in soil is frequent, it is necessary to study the correlation between carbonates and aggregate stability. For example, Fernández-Ugalde et al. [19] thought carbonates must be considered when modelling soil structure formation. Calcium bridging is the dominant factor for the long-term positive effect of calcium addition on the structural stability of soil [10]. Chrenková et al. [11] found carbonate content had a positive in uence in MWD for sandy soils. In semiarid calcareous soils, Fernández-Ugalde et al. [19] found that the interaction of maize straw and carbonates resulted in a higher stability of macro-aggregates (> 250um) in carbonated soil than noncarbonated soil, then concluded the formation of secondary carbonates within and/or around macroaggregate could explain this stability. In our case, the effect of carbonates in stabilization of aggregates showed two-sidedness. On the one hand, aggregate-inorganic carbon decreased with increasing aggregate size, except for the smallest size which had lower concentration than the next-bigger size. This occurred because the CaCO 3 could make the soil particles consolidated in the dry state. But when soil was wetted by water, the CaCO 3 could dissolve in water and make the soil particles separated and become dispersed. Therefore, the aggregates of calcareous soil were easily broken under the fast wetting condition. Our study also found that P. tabulaeformis could effectively decrease the content of CaCO 3 , which lead to the improvement of soil stability and reducing of soil erosion. On the other hand, there was a linear positive correlation between inorganic carbon content and MWD of water stable aggregates in the soil pro le of P. tabulaeformis plantation, while a linear negative correlation was found in the soil pro le of R. pseudoacacia plantation (Fig. 6). Therefore, this paper assumed that there was a threshold value of inorganic carbon content in soil. When it was higher than the threshold value, the aggregate structure would be destroyed with the dissolution of calcium carbonate in the process of rapid wetting; when it was lower than the critical value, calcium carbonate could be used as aggregate cement. Uddivira et al. [16] found that adding Ca 2+ to the soil with Ca 2+ concentration of 2.7 and 3.1cmol/kg increased the percentage content of water stable aggregates, on the contrary, adding Ca 2+ in the soil with Ca 2 + concentration of 13.9 cmol / kg would disperse more aggregates. Virto et al. [7] also believed that compared with the soil with 30% carbonate content, the carbonate in the soil with 15% carbonate content played a major cementation role in soil aggregates.
Aggregate-SOM content increased with increasing aggregate size.
Qiao et al. [33] suggested that micro-aggregates played key roles in protecting SOC based on more recalcitrant SOC stored in micro-aggregates. However, higher organic carbon content in the macroaggregates (> 2 mm fraction) compared to the micro-aggregates ( 0.2 mm fraction) in our results, that was similar to previous studies [19], which observed that when organic inputs were increased in a calcareous soil, a greater proportion of stable macro-aggregate (> 250um) would be formed in comparison to a non-calcareous soil of similar characteristics. Gao et al. [18] also found the largest amount of organic carbon was stored in > 0.25 mm size aggregates. Elliott [4] observed more organic matter associated with macro-aggregates than with micro-aggregates in a temperature grassland soil.
The conclusion also was line with the theory of aggregate hierarchy, which believes an increase in carbon concentration with increasing aggregates size class because that large aggregate-size classes are composed of small aggregate-size classes plus organic binding agents [4]. In general, there was a tradeoff between SOC and CaCO 3 , which was similar to the conclusions of Yang et al. [6] and Han et al. [30].

Conclusions
Organic materials are the major cementing agents in uencing aggregation formation and stabilization. Calcium carbonate may be a dispersing agent when it was higher than a threshold value. On the contrary when it was lower than the threshold value, it could be used as aggregate cement. Soil organic carbon and inorganic carbon were affected by tree species. The economic forest M. pumila decreased the content of SOM and MWD of water stable aggregates due to effect of management. M. pumila and R. pseudoacacia were not as bene cial as P. tabulaeformis for the sequestration of SOC and improving water stable aggregates.

Declarations
Ethics approval and consent to participate The location of sample area Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.  Soil total carbon density under several soil layers in three arti cial forest types The content of calcium carbonate (CaCO3) and soil organic matter (SOM) in water stable aggregates with different sizes Figure 6 Relationship between MWD of water stable aggregates and CaCO3 content in Pinus tabulaeformis (A) and Robinia pseudoacacia (B)