Changes in Soil Organic Carbon and Total Nitrogen at a Small Watershed Scale as the Result of Land Use Conversion on the Loess Plateau

Soil organic carbon (SOC) and total nitrogen (total N) are important soil components for agricultural production. Soil quality is related to the total amount of SOC and total N sequestered in the soil. Land use plays a major role in the distribution and amount of SOC and total N. This study analyses the amount of SOC and total N under various land cover types in 1987, 2005 and 2010, and evaluated their storage in land use conversions in a comprehensively managed watershed on the Loess Plateau, China. Results show that concentrations of SOC and total N in shrub land and natural grassland areas were significantly higher than for other land uses (farmland, orchard, abandoned farmland, manmade grassland) while cropland had the lowest concentration. Storage of SOC and total N increased along the revegetation chronosequence. As the storage of SOC in 2005 and 2010, they were 3461.86 × 108 and 4504.04 × 108 g respectively. Soil organic carbon storage were enhanced one third just during 5 years. The effects of land use on SOC and total N were the most significant in the upper soil layers. The correlation between SOC, total N, and the C/N ratio indicated that the best combination of land uses were natural grassland and shrub land. They efficiently influenced the distribution and storage of SOC and total N, and benefited vegetation restoration.


Introduction
Soil organic carbon is an important element of the global carbon stock and contains approximately two times more carbon than the atmosphere or vegetation [1].Soil carbon pool estimations (both organic carbon and inorganic carbon) in China are essential for appraising global terrestrial carbon inventories; they also optimize the mitigation of carbon dioxide (CO 2 ) accumulation in the atmosphere [2].Soil organic carbon and total nitrogen are the key indicators for estimating soil quality and act as important carbon and nitrogen reservoirs [3], and understanding the distribution of SOC and total N stocks are essential in achieving improvements in soil quality [4].Scholars have performed a lot of studies on the SOC distribution and stock on global [5], country [6] and regional [7] scales.However, these studies had a high degree of uncertainty due to the dates came from different sources, so the inherent and spatial variability were just simulated pattern.It is different with inventory data (measured values) after the processing of spatial distribution technology that would reflect real spatial variation.Soil organic carbon were influenced by climate, hydrology, soil, land use, et al. abiotic factors and the other biotic factors, while land uses were the most sensitive to display human disturbance.Thus, monitoring the SOC in different land uses is essential for estimating the SOC distribution and stock.
The equilibrium of carbon and nitrogen stocks is the result of the inputs and outputs [8,9] of the carbon and nitrogen cycle.Some studies have reviewed the effects of land use conversion on soil carbon stocks, such as forest clearing [10,11] changes in tropical forest cover [12,13] disturbance and recovery [14] cultivation [15] deforestation for pasture, land-use management of crop, pasture and forest, and cultivation of native vegetation to grasslands [16].Conversion from cultivation land into perennial vegetation land is found to accumulate SOC and total nitrogen by increasing carbon and nitrogen source from litter decomposition and soil mineralization [3,[17][18][19].Organic carbon losses due to land-use change from grasslands and forests to croplands are estimated 20% to 25% in the zone of cultivation within the first 40-50 yrs.[20].These losses show a rapid decline over the first 20 years after which soil organic carbon levels gradually stabilize at a new steady state over about the next 30 years [21][22][23].During the past two centuries, land uses practices have modified decomposition dynamics by changing soil aeration, water dynamics and storage, as well as the biochemistry and quantity of crop residues [24,25].Cultivation of pasture soils has resulted in a 25 to 50 % decrease in soil organic carbon [26,27].
During the last century, China implemented the most damaging type of outmoded agricultural management, which led to habitat fragmentation and soil erosion.Increasing population pressures and sloped terrain, subjected to intensive cultivation, were the major causes that led to environment fragmentation and serious soil erosion.In order to withstand further deterioration of natural ecosystems, the Chinese government has launched a serious of nation-wide conservation projects focusing on vegetation rehabilitation and land recovery [12].One of the most promising programs in achieving sustainable ecological restoration was the Grain for Green project.The project, implemented in 1999, focused on the recovery of damaged ecosystems.The crucial outcomes were sustainable agricultural development, controlled soil erosion, improved land quality, and widespread conversion of sloping cropland (the loess hilly area) to sustainable land uses.Croplands with slopes of greater than 15 • were converted to green space [28].
This manuscript examines quantity and distribution of SOC and total N under various developmental stages of different land uses.We hypothesized that: (1) Land uses conversion would affect SOC and total N during the different developmental stages (1987, 2005 and 2010), and SOC and total N would increase after vegetation restoration; (2) shrub land and natural grassland related changes in SOC and total N significantly, natural grassland is the overwhelmingly dominant land use in SOC storage increasing and soil nitrogen mineralization promoting.Therefore, the objectives are: (1) Judge the effects of Grain for Green project implemented on the Loess Plateau under various developmental stages; (2) to analyze the influence of land uses in relation to increase SOC storage and promotion of soil nitrogen mineralization.

Study Area
The comprehensive management watershed (Shanghuang) (106 • 26 106 • 30 E, 35 • 59 36 • 02 N) is in the hilly-gully region on the Loess Plateau, Guyuan, Ningxia Province.The watershed has a semi-arid climate with an average annual temperature of 6.9 • C and an average annual rainfall of 419 mm .The rainy season starts in July and continues until October, with the average rainfall in August accounting for 24% of the total annual precipitation.The dominant soil type in the study area is loessal Cambisol [29].The elevation of the study area ranges from 1530 to 1822 m above sea level, and the area covers 8.19 km 2 .According to field investigations, satellites and remote sensing images, the land use in the study area are: Platform farmland, sloping farmland, cropland, terraced farmland, orchard, forest land, shrub land, manmade grassland, natural grassland, residential land, water area, flood land, and fallow land (Figure 1a-d).In 1982, the study area was characterized by broken topography, serious soil erosion, deteriorated ecological environment and lagging agricultural productivity.It is the result of long time over-exploitation and lower vegetation coverage.The primary land uses of sloping farmland, natural ArcGIS was used to perform a land use change analysis.Remotely sensed data was interpreted using digitized 1982 1:10000 scale topographic maps, color infrared aerial photographs from 1990, 2005 orthophotos and 2010 land use field studies (Table 1).In order to introduce uniformity and comparability, we used a five-class land use system: (I) Shrub land; (II) manmade grassland; (III) natural grassland; (IV) farmland; (V) abandoned farmland.The principle form of manmade shrub land is Caragana Korshinskii with two associated species, Armeniaca sibirica and Amygdalus davidiana.The dominant manmade grassland plant is Medicago sativa L. The dominant species in natural grassland are Artemisia gmelinii, Stipa bungeana, Artemisia scoparia, Artemisaia stelleriana, Stipa grandis, Thyme and Potentilla chinensis.In 1982, the study area was characterized by broken topography, serious soil erosion, deteriorated ecological environment and lagging agricultural productivity.It is the result of long time overexploitation and lower vegetation coverage.The primary land uses of sloping farmland, natural grassland, shrub land, and manmade grassland accounted for 1.3% of the total area (5.3 and 5.0 ha; Figure 1a: 1982).After the establishment of the research station, the land cover underwent a fundamental change.An optimization model for land use, based on the theory of aggro-ecological systems, was implemented to effectively improve the ecological environment and conserve soil and water.After grass was cultivated between 1982 and 1990, the areas of manmade shrub land and grassland rose to 27.2% (Figure 1b: 1990).A significant change in land use occurred between 2002 and 2010 during the "Grain for Green" restoration project.The Grain for Green program was implemented in Shanghuang to solve environmental and economic development issues; with regard to the former, the objective was to control soil erosion and improve land quality.As the result of the special terrain designation (90% of the area is hilly, with 51% of the land at an elevation of 1534.33-1822.30m with closely dissected, sharp-edged, steep slopes [12], the program relocated high-slope farmland (>25 • slope) to lower-slope areas and established shrub and grass land.During the years of Gain for Green, the study area was changed into a demonstration zone for successful land reclamation and comprehensive management, the area of shrub land occupied 66.7% of total area and farmland occupied 16.16% (Figure 1c,d).

Experimental Design and Sample Collection
SOC and total N data were collected by the same research group (1987, 2005, and 2010).But sample numbers and soil depths for each time period with different objectives were non-uniform.In 1987, 35 samples were collected from natural grassland at a sampling depth of just 0-10 cm.In July 2005, 45 samples were collected at sampling depths of 0-20 and 20-40 cm by the foundation of Nation Sciences, China (40701095).In June 2010, 42 samples were collected at sampling depths of 0-10, 10-30, and 30-60 cm by the foundation of Nation Sciences, China (4097117) (Tables 1 and 2).Land use information was collected for soil samples in 1987, 2005, and 2010 (Figure 1 and Table 1), while soil sample numbers, depths and other basic management details were collected in 2005 and 2010 (e.g., vegetation types, dominant species, site coordinate fertilization).A non-equidistant irregular grid (0.8 × 0.8 km) method was used for sample collection in 1987, 2005, and 2010.Six or seven replicated points were collected for each site.The dichromate oxidation method was used to measure SOC and the micro Kjeldahl method was used to determine total nitrogen [36].The core ring method was used to determine bulk density.

Statistical Analysis
How to compare SOC and TN with non-uniform data directly?It is too difficult, so we use the indexes of the storage of soil organic carbon and total nitrogen, those indicators are the total quantity and spatial value of SOC and nitrogen.It can eliminate the influence that caused by different sample numbers.And the same time, two parameters H i and n in equations eliminate the influence the thickness of soil and soil layers.
The density of soil organic carbon total nitrogen can indicate the contributions of global changes that form soil quality and terrestrial ecosystem.Generally, the mass of soil organic carbon and total nitrogen contained in soil per square meter, that represent soil organic carbon density and soil total nitrogen density.
D SOC -Soil organic carbon density C i -The gravel content on layer i (particle size > 2 mm, volume %); B i -Soil bulk density on layer i (g/cm 3 ); SOM i -The content of soil organic carbon on soil layer i (g/kg); H i -The thickness of the soil (cm); n-The numbers of total soil layer (cm); 0.58-The conversion coefficient of Bemmenlen (SOC); D TN -Soil total nitrogen density; C i -The gravel content on layer i (particle size > 2 mm, volume %); B i -Soil bulk density on layer i (g/cm 3 ); TN i -The content of total nitrogen on soil layer i (g/kg); H i -The thickness of the soil (cm); n-The numbers of total soil layer (cm); S SOC , S TN -Soil organic carbon storage, soil total nitrogen storage; S i -The area of land uses; D SOC , D TN -Soil organic carbon density, soil total nitrogen density; A variety of statistical analyses were used to identify the effects of land use and soil depth on SOC and total N concentration.One-way ANOVA analysis with land use type (shrub land, manmade grassland, natural grassland, farmland and abandoned farmland) as the main factor was employed to test the significance of mean differences.The level of significant differences was p < 0.05.A univariate procedure was used to examine the normality of distribution of the observed data before conducting ANOVA.After the ANOVA test indicated a significant effect at p < 0.05, the means of SOC and total nitrogen for different land use types were classified using the Duncan multiple-range procedure.All statistical analyses were performed with SAS Enterprise Guide, Version 4.1.

Land Uses Conversion Between 1982 and 2010
Based on the agricultural history of the area, the developmental stages of land use conversion were divided into three parts (Table 1).In the first stage (1982)(1983)(1984)(1985)(1986)(1987)(1988)(1989)(1990), the primary land uses were natural grassland (46%) and farmland (40%) (Figure 1a).Due to increasing population pressure, natural grassland was converted into farmland.In the second stage, between 1990 and 2005, the area of natural grassland decreased by 21.04% and farmland became the major land use.Farmland increased by 10.85% from 316 ha in 1987 to 408 ha in 2005.As ecological restoration began to solve the problem of fragmentation, farmland and manmade grassland gradually decreased.In the third period (2005-2010), the area of orchard, shrub land and natural grassland increased.Farmland decreased dramatically from 407 ha in 2005 to 130 ha in 2010 at the percent of 34.39%, shrub land became the dominate land use.One-way ANOVA analysis indicated that land use had a significant effect on SOC and total N.In Table 3, natural grassland had the lowest concentration of SOC and total N in 1987, 5.90, and 0.70 g•kg −1 , respectively.Whereas, after more than 10 years, they reached a high level in 2005 and in 2010.Among the different land uses, the concentration of SOC and total N declined in the following order: Natural grassland > shrub land, abandoned farmland > manmade, and grassland > farmland.Farmland had the lowest values of SOC and total N.As depth increased, the concentration of SOC and total N decreased gradually.The downward trend in concentration in natural grassland were drastic, SOC decreased 5.7 g•kg −1 from 0-20 to 20-40 cm in 2005, 5.46 g•kg −1 from 0-10 to 10-30, and 5.09 g•kg −1 from 10-30 to 30-60 cm in 2010.However, in farmland the reduction was not obvious, it decreased 1.67 g•kg −1 in 2005, and 1.26 g•kg −1 from 0-10 to 10-30, and 1.14 g•kg −1 from 10-30 to 30-60 cm in 2010.For natural grassland, the ratio of soil organic carbon/ total nitrogen (C/N) was 8.43 in 1987 in Table 4, it was much lower than in 2005 and 2010.Shrub land showed the highest C/N ratios in 2005.However, natural grassland had the greatest C/N ratio in 2010 The lowest values were shown in farmland and abandoned farmland, whether in 2005 or in 2010; the C/N ratios of natural grassland and shrub land were higher than manmade grassland, abandoned farmland, and farmland in 2005 and 2010.
Based on the correlation analysis, the regression lines in Figure 2 describe the relationship between SOC and total N among the different land uses.The correlation between SOC and total N among the different land uses were positive.Shrub land (0.8189) and natural grassland (0.9318) had the highest correlation coefficients, and the variation tendencies were proximity (Figure 2a,c).The results indicate that SOC and total nitrogen of shrub land and natural grassland have the better couplings.Based on the correlation analysis, the regression lines in Figure 2 describe the relationship between SOC and total N among the different land uses.The correlation between SOC and total N among the different land uses were positive.Shrub land (0.8189) and natural grassland (0.9318) had the highest correlation coefficients, and the variation tendencies were proximity (Figure 2a, c).The results indicate that SOC and total nitrogen of shrub land and natural grassland have the better couplings.

Distribution of BD, SOCD, and SND in 2005 and 2010
The distributions of soil bulk density (BD), soil organic carbon density (SOCD), and soil nitrogen density (SND) among the different land uses in 2005 and 2010 are shown in Figure 3.In 2005 and 2010, the BD values in manmade grassland and farmland were higher than natural grassland and shrub land.The differences among the values for each land use in 2005 were considerable (Figure 3a), manmade grassland, had the highest BD (1.26 g•m −3 ), which was 1.11 times higher than the minimum value (natural grassland).The comparison of SOCD and SND among the different land uses in 2005 and 2010 are shown in Figure 3b, c.The difference between SOCD values for the various land uses in 2005 was greater than in 2010, especially for natural grassland.SOCD was 3.72 g•m −3 in 2005 (0-20 cm), which was two times more than abandoned farmland (2.08 g•m −3 ).However, abandoned y = 0.1076x + 0.1464 R² = 0.7912 Total N (gꞏkg

Distribution of BD, SOCD, and SND in 2005 and 2010
The distributions of soil bulk density (BD), soil organic carbon density (SOCD), and soil nitrogen density (SND) among the different land uses in 2005 and 2010 are shown in Figure 3.In 2005 and 2010, the BD values in manmade grassland and farmland were higher than natural grassland and shrub land.The differences among the values for each land use in 2005 were considerable (Figure 3a), manmade grassland, had the highest BD (1.26 g•m −3 ), which was 1.11 times higher than the minimum value (natural grassland).The comparison of SOCD and SND among the different land uses in 2005 and 2010 are shown in Figure 3b,c.The difference between SOCD values for the various land uses in 2005 was greater than in 2010, especially for natural grassland.SOCD was 3.72 g•m −3 in 2005 (0-20 cm), which was two times more than abandoned farmland (2.08 g•m −3 ).However, abandoned farmland has the lowest SOCD and SND of the different land uses; the order was: Natural grassland > farmland > shrub land > manmade grassland > farmland.

Distribution of BD, SOCD, and SND in 2005 and 2010
The distributions of soil bulk density (BD), soil organic carbon density (SOCD), and soil nitrogen density (SND) among the different land uses in 2005 and 2010 are shown in Figure 3.In 2005 and 2010, the BD values in manmade grassland and farmland were higher than natural grassland and shrub land.The differences among the values for each land use in 2005 were considerable (Figure 3a), manmade grassland, had the highest BD (1.26 g•m −3 ), which was 1.11 times higher than the minimum value (natural grassland).The comparison of SOCD and SND among the different land uses in 2005 and 2010 are shown in Figure 3b, c.The difference between SOCD values for the various land uses in 2005 was greater than in 2010, especially for natural grassland.SOCD was 3.72 g•m −3 in 2005 (0-20 cm), which was two times more than abandoned farmland (2.08 g•m −3 ).However, abandoned farmland has the lowest

Discussion
Land use conversion significantly influences the terrestrial ecosystem structure, which makes the contribution of SOC and STN vary [3].Microclimate, soil variation, and organic matter input are affected by land uses that impact SOC decomposition and N mineralization [2,[37][38][39].[40] indicated that cultivation does not increase accumulation of SOC or support perennial vegetation.SOC storage

Discussion
Land use conversion significantly influences the terrestrial ecosystem structure, which makes the contribution of S OC and S TN vary [3].Microclimate, soil variation, and organic matter input are affected by land uses that impact SOC decomposition and N mineralization [2,[37][38][39].[40] indicated that cultivation does not increase accumulation of SOC or support perennial vegetation.SOC storage decreases by 42% and 59% when forest and grassland were converted to crop, respectively [18].Soil nitrogen storage in black soil under grassland has increased in the past 20 years of vegetation rehabilitation.Additionally, there was a sustained decrease in SOC and soil nitrogen when natural ecosystems were subjected to human disturbance or were converted into agriculture land [41,42].
In the present study, the concentration of SOC and total N declined in the following order: Natural grassland > shrub land, abandoned farmland > manmade, grassland > farmland.The lowest concentration of SOC and total N in farmland would result in: (I) Lower organic matter that decreased with aerial growth, biomass production, and crop harvest; (II) long-term cultivation and uptake of nitrogen by crops would enhance mineralization and utilization of soil organic matter [35].Natural grassland and shrub land are the most suitable land uses in Shanghuang watershed for ecological restoration and have greater SOC and total N than other land uses.This is because: (I) A sufficient carbon source is offered by higher litter inputs [15,[43][44][45]; (II) lower herbaceous plants and a closed canopy is created by a moist environment; (III) a vertical distribution of root systems is created [33].
The root system distribution directly influences the vertical distribution of SOC and total N. Abundant carbon resources are produced by ageing, decay, and decomposition of roots [46].Grass roots are always concentrated at a depth of 0-40 cm, indicating that a dramatic reduction of SOC in grasslands at greater soil depths is caused by fewer roots [47].Although, some researchers show that manmade grassland is an effective means of grassland system restoration [48,49].The present study shows that SOC and total N of manmade grasslands were lower than in other land uses (just a little higher than in farmland).Manmade grasslands containing Medicago sativa is the main forage vegetation for the animal husbandry industry in the Shanghuang watershed.In the early years of cultivation, Medicago sativa has a high nutrient consumption demand.As the years of cultivation increase, root nodule bacteria and litter decomposition generate and secrete nitrogen and organic carbon, which adds carbon and nitrogen to the soil [50].However, this increase in SOC and total N does not satisfy the requirements of plant growth.Years of cultivation break the equilibrium of supply and demand, which decreases the amount of nutrients in the soil.
Soil bulk density (BD), the unit volume of the undistributed soil, is influenced by soil density and pore space and decreases with greater pore space and fluffier textures.The purpose of the Grain for Green project was to convert farmland (low productivity, unsustainable, and unreasonable land management conditions) into shrub land and grassland with small soil bulk density.[51] reported that the vegetation on farmland improved soil quality after 8 years of secondary succession.As succession proceeded, vegetation litter and humus accumulated on the surface and the root systems influenced soil bulk density.SOC and total N was significantly higher at the soil surface than at other soil profile depths.[14] indicated that SOC and total N declined consistently with project depth in sandy loam soils (Typic Udivitrands) of New Zealand.[21] also reported that SOM in cultivated lands at a depth of 0-20 cm were significantly reduced by 49% relative to SOM content of pasture in the highlands of southern Turkey when pasture was cultivated.Differences between soil profile depths were dependent on plant residue that was attributed to vegetative cover, years under development, land use type, and human activity.The incorporation of plant residue into the surface soil layer is easily distracted by external factors, especially in the 5 cm soil layer.
The soil C/N ratio is an important indicator of the soil microorganism community structure and changes in soil quality in terrestrial ecosystems [52,53].The C/N ratio is closely related to soil nitrogen mineralization, fixation, and nitration [8].The differences in soil C/N ratios among land uses result from: (I) The accumulation rate of SOC and total N; (II) the chemical properties of organic matter [54].Tong, C.L, et.al.[55] reported that the SOC and total N accumulation are not synchronous.
The accumulation rate of SOC is faster than that of total N. Litter decomposition is the major reason why the C/N ratio in farmland was lower than in natural grassland and shrub land.The lignin, waxiness, and organic matter in shrub land and grassland are higher than in farmland where it has difficulty decomposing [54,56].The lignin content of grass roots was 15-19% [57], and 9-12% in crop (wheat and corn) roots [58].Therefore, shrub land with the highest C/N ratio is related to a greater amount of material in a slowly decomposing organic matter fraction.
The advantages of ecological restoration in Shanghuang watershed were not only seen in vegetation cover and production increases, but also in soil microbial parameters and nutrient enhancements.It also improved land productivity, strengthened soil carbon sinks, and decreased greenhouse gas emission.

Conclusions
Through the process of ecological restoration, natural grassland gradually recovered and large proportions of farmland were converted into shrub land and abandoned grassland.Soil carbon and nitrogen storage in Shanghuang increased along the revegetation chronosequence.The positive coefficient of C/N ratios indicated that the best combination of land uses were natural grassland and shrub land.The contribution of land use conversion on SOC and total N concentration was obvious in the Shanghuang watershed.The projects on the Chinese Loess Plateau promoted soil organic carbon storage and soil nitrogen mineralization.
. In order to introduce uniformity and comparability, we used a five-class land use system: (I) Shrub land; (II) manmade grassland; (III) natural grassland; (IV) farmland; (V) abandoned farmland.The principle form of manmade shrub land is Caragana Korshinskii with two associated species, Armeniaca sibirica and Amygdalus davidiana.The dominant manmade grassland plant is Medicago Sativa L. The dominant species in natural grassland are Artemisia gmelinii, Stipa bungeana, Artemisia scoparia, Artemisaia stelleriana, Stipa grandis, Thyme and Potentilla chinensis.

Figure 2 .
Figure 2. Correlations between SOC and total N during 2010 in the different land uses.(a) Shrub land; (b) Manmade grassland; (c) Natural grassland; (d) Farmland; (e) Abandoned farmland

Figure 2 .
Figure 2. Correlations between SOC and total N during 2010 in the different land uses.(a) Shrub land; (b) Manmade grassland; (c) Natural grassland; (d) Farmland; (e) Abandoned farmland

Figure 3 .
Figure 3. Distributions of soil bulk density (BD), soil organic carbon density (SOCD), and nitrogen density (SND) among the different land uses in 2005 and 2010.(a) Soil bulk density; (b) Soil organic carbon density; (c) Soil nitrogen density.
(a) SSOC and STN of natural grassland in the 0-10 cm profile depth; (b) SSOC and STN for the different land use types in 2005 and 2010.

Figure 3 .
Figure 3. Distributions of soil bulk density (BD), soil organic carbon density (SOCD), and nitrogen density (SND) among the different land uses in 2005 and 2010.(a) Soil bulk density; (b) Soil organic carbon density; (c) Soil nitrogen density.

3. 5 .
Soil Organic Carbon and Total Nitrogen Storage (S SOC and S TN )Soil carbon and nitrogen storage (S SOC and S TN ) in the surface soil of natural grassland was compared for 1987 and 2010.

Figure 4 .
Figure 4. Soil carbon and nitrogen storage of different land use types in 2005 and 2010.(a) Soil carbon storage; (b) Soil nitrogen storage.

Figure 4 .
Figure 4. Soil carbon and nitrogen storage of different land use types in 2005 and 2010.(a) Soil carbon storage; (b) Soil nitrogen storage.

Table 1 .
Land use conversion between 1982 to 2010 among various land uses.

Table 1 .
Land use conversion between 1982 to 2010 among various land uses.

Table 3 .
Concentration of SOC and total N in 1987, 2005, and 2010.
Values are mean ± standard.

Table 5 .
Distributions of soil carbon and nitrogen storage.

Table 5 .
Distributions of soil carbon and nitrogen storage.(a) SSOC and STN of natural grassland in the 0-10 cm profile depth; (b) SSOC and STN for the different land use types in 2005 and 2010.