Effect of soil coarseness on soil base cations and available micronutrients in a semi-arid sandy grassland

Soil coarseness is the main process decreasing soil organic matter and threatening the productivity of sandy grasslands. Previous studies demonstrated negative effect of soil coarseness on soil carbon storage, but less is known about how soil base cations (exchangeable Ca, Mg, K, and Na) and available micronutrients (available Fe, Mn, Cu, and Zn) response to soil coarseness. In a semi-arid grassland of Northern China, a field experiment was initiated in 2011 to mimic the effect of soil coarseness on soil base cations and available micronutrients by mixing soil with different mass proportions of sand: 0 % coarse elements (C0), 10 % (C10), 30 % (C30), 50 % (C50), and 70 % (C70). Soil coarseness significantly increased soil pH in three soil depths of 0–10, 10–20 and 20–40 cm with the highest pH values detected in C50 and C70 treatments. Soil fine particles (smaller than 0.25 mm) significantly decreased with the degree of soil coarseness. Exchangeable Ca and Mg concentrations significantly decreased with soil coarseness degree by up to 29.8 % (in C70) and 47.5 % (in C70), respectively, across three soil depths. Soil available Fe, Mn, and Cu significantly decreased with soil coarseness degree by 62.5, 45.4, and 44.4 %, respectively. As affected by soil coarseness, the increase of soil pH, decrease of soil fine particles (including clay), and decline in soil organic matter were the main driving factors for the decrease of exchangeable base cations (except K) and available micronutrients (except Zn) through soil profile. Developed under soil coarseness, the loss and redistribution of base cations and available micronutrients along soil depths might pose a threat to ecosystem productivity of this sandy grassland.

Soil base cations are not only essential nutrient cations for both plants and soil microbes, but also serve as one of the main mechanisms of soil acid buffering capacity (Lu et al., 2014) as well as a good indicator of soil fertility (Zhang et al., 2013). Micronutrient availabilities essentially affect terrestrial net primary production, plant quality, and consequently food and forage supply worldwide (Cheng et al., 2010;5 Marques et al., 2015). Current research about desertification and soil coarseness mainly focus on its effects on degradation of forest and grasslands due to logging and overgrazing (Conte et al., 1999;Cao et al., 2008), C and N depletion in soils and plant components (Zhou et al., 2008;Bisaro et al., 2014), soil compaction and erosion risk (Allington and Valone, 2010), and soil physical properties of particle size distributions 10 (Su et al., 2004;Huang et al., 2007). However, less is known about the changes in soil base cations and availabilities of micronutrient during dryland desertification and soil coarseness.
Soil coarseness is suggested to cause decrease of soil silt and clay contents (Zhou et al., 2008), decline in soil C and nutrient (such as N and P) concentrations (Xie et al.,15 2015), and losses in species diversity and productivity (Zhao et al., 2006;Huang et al., 2007). As biogeochemical cyclings of base cations and micronutrients are largely controlled by soil organic matter (SOM) (complexation and chelation) (Sharma et al., 2004) and properties of soil mineral (reversible sorption and desorption processes) (Jobbá gy et al., 2004), decrease of SOM and soil fine particles would potentially 20 decrease soil base cations and micronutrient availability. Also, the changes in SOM along soil depth could shape the vertical distribution of base cations and available micronutrients (Sharma et al., 2004).

The Horqin Sandy Land, or Horqin Sandy Grassland is an important part of Inner
Mongolia grassland and one of the main sandy areas in northern China covering 25 approximately 43,000 km 2 (Li et al., 2004). In Horqin region, the soils are prone to aeolian soil erosion and soil coarseness especially when natural sandy grassland is converted into farmland (Li et al., 2004). To examine the effect of soil coarseness during desertification on the concentrations of soil base cation (exchangeable Ca, Mg, K, and Na) and available micronutrient (Fe, Mn, Cu, and Zn) of this region, we set up 30 Solid Earth Discuss., doi:10.5194/se-2016-18, 2016 Manuscript under review for journal Solid Earth Published: 27 January 2016 c Author(s) 2016. CC-BY 3.0 License. a field experiment in Zhanggutai by mixing the soil with different mass proportions of sand: 10% (light soil coarseness), 30% (moderate soil coarseness), 50% (heavy soil coarseness), and 70% (severe soil coarseness). We hypothesized that both soil base cations and available micronutrients would decrease with the increasing degree of soil coarseness due to the decrease of SOM and soil fine particles. We also expect that soil 5 base cations and available micronutrients would decrease with soil depth.  (Li et al., 2000;Chen et al., 2005). The mean annual temperature is 6.2 º C and mean annual precipitation is about 450 mm being a semi-arid region (Chen et al., 2005). Soil texture of the experiment site is sandy soil with 99.32 ± 0.13 % sand, 0.45 ± 0.14 % silt, and 0.23 ± 0.02 % clay (means ± standard deviation, data measured from control soil). The soil type is classified as a Aeolic Eutric Arenosol 20 according to the FAO classification (IUSS Working Group WRB, 2014). This area constitutes an agro-pastoral ecotone which is severely degraded due to excessive cultivation and grazing (Chen et al., 2005). 25 In May 2011, a complete randomized design was applied to the site. Thirty 4 m×4 m plots were established for five treatments with six replicates per treatment. Adjacent plots were separated by 1 m buffer zone and PVC plates to prevent water and nutrient exchanges. A certain mass proportion of 2 mm-sieved river sand (siliceous, pH 7.5 ± 0.2) was mixed with native soil for each of three depths (0-20 cm, 20-40 cm, and  Table 1.

Soil sampling and chemical analysis
In October 2015 (i.e. after 2 years of plant community settled), a composite soil sample was taken from three randomly selected locations within each plot from three 20 soil layers of 0-10 cm, 10-20 cm, and 20-40 cm, respectively. Fresh soil samples were sieved through 2 mm screen and visible plant roots were taken out. After transportation to laboratory, the soils were air-dried and a subsample of the soil was ground for C and N analysis.

Soil pH and particle size distribution
Soil pH was determined in a 1:2.5 (w/v) soil-to-water extract of soil samples from all treatments with a PHS-3G digital pH meter (Precision and Scientific Corp., Shanghai, China). Soil particle size distribution was determined by the pipette method in a sedimentation cylinder, using Na-hexamethaphosphate as the dispersing agent (Zhao

Statistical analyses
The normality of data was tested using the Kolmogorov-Smirnov test, and 20 homogeneity of variances using Leven's test. Effects of soil coarseness on soil pH, fine particles, base cations and available micronutrients were determined by one-way ANOVA. Multiple comparisons with Duncan design were performed to determined difference in soil parameters among soil coarseness degrees. Pearson correlation analysis was used to examine the relationship among soil parameters. Multivariate 25 linear regression analyses (stepwise removal) were conducted to determine variables that made significant contributions to variance of soil base cations and available micronutrients. All statistical analyses were performed in SPSS 16.0 (SPSS, Inc., Chicago, IL, U.S.A) and statistical significance was accepted at P < 0.05.

Soil pH
Soil coarseness significantly increased soil pH by up to 8.8% across three soil depths ( Fig. 1a; Table 2). For both 0-10 cm and 10-20 cm soils, the highest soil pH was detected in C70 (7.3 and 7.4, respectively) and C50 (7.2 and 7.3, respectively) soils, 5 which were followed by C30 and C10 soils (Fig 1a). Significant and positive overall effect of soil depth was detected on soil pH ( Fig. 1a; Table 2). For both C0 and C50 treatments, soil pH of 10-20 cm and 20-40 cm was significantly higher as compared to that of 0-10 cm soil (Fig. 1a). Soil pH in10-20 cm of C10 and C30 was significantly higher than that in 0-10 cm of C10 and C30, respectively (Fig. 1a). Significant 10 interactive effect of soil coarseness and soil depth was found on soil pH ( Table 2).
Proportions of soil fine particles (< 0.25 mm) were determined in 0-10 cm soil.

Soil base cations
Across three soil depths, soil coarseness significantly decreased both exchangeable Ca 20 and Mg concentrations by up to 29.8% and 47.5%, respectively, as compared to C0 ( Fig. 2a,b). Both exchangeable Ca and Mg concentrations were the lowest in the C70 and followed by C50 as compared to C0 in all soil depths (Fig. 2a,b). Soil depth significantly decreased soil exchangeable Mg, while showed no effect on exchangeable Ca (Table 2). Both soil coarseness and soil depth had no impact on soil 25 exchangeable K (Fig. 2c). At 0-10 cm, C50 and C70 significantly decreased soil exchangeable Na by 22.3% and 24.2%, respectively, as compared to C0 (Fig. 2d). Soil exchangeable Na did not change with soil depth (Fig. 2d, Table 2). Soil available Fe significantly decreased with soil coarseness degree by as much as 17.1% in C10, 22.0% in C30, 36.6% in C50 and 62.5% in C70 across three soil depths (Fig. 3a). Soil coarseness significantly decreased soil available Mn for 0-10 cm (by up to 17.3% in C70) and 10-20 cm (by up to 45.4% in C70) soils (Fig. 3b). Both soil available Fe and Mn significantly decreased with soil depth (Fig. 3a,b; Table 2). 5 Significant negative desertification effect was detected on soil available Cu by 14.7% -44.4% as compared to C0 in 0-10 cm soil (Fig. 3c). For both C30 and C50 treatments, soil available Cu concentration of 10-20 cm soil was significantly higher than that in 0-10 cm and 20-40 cm soils. Soil available Zn concentration was not affected by soil coarseness but it decreased with soil depth (Fig. 3d; Table 2).

Regression analyses between soil parameters
All regression analysis were conducted for 0-10 cm soil as the data of fine particles (< 0.25 mm) were only available for 0-10 cm soil. At 0-10 cm soil, soil pH significantly and negatively correlated with exchangeable Ca, Mg and Na, and with available Fe, 15 Mn and Cu (Table 3). Soil fine particles (< 0.25 mm) significantly and positively correlated with exchangeable Ca, exchangeable Mg, exchangeable Na, available Fe, available Mn, and available Cu (Table 3). The SOC significantly and positively correlated with exchangeable Ca, Mg, Na, available Fe, Mn, and Cu (Table 3).
According to multiple regression models, change of soil fine particles explained 20 65.5%, 75.7%, 31.4%, 24.0% of variations in exchangeable Ca, Mg, Na, and available Mn (Table 3). Soil pH explained 75.7% of variation in available Fe ( Table 2). The SOC explained 59.3% of variation in available Cu (Table 2). correlation between soil pH and exchangeable Ca, Mg and Na (Table 3). Indeed, with the increase of soil pH, soil base cations (such as Ca 2+ and Mg 2+ ) and available micronutrients (Fe 2+ , Mn 2+ and Cu 2+ ) would precipitate with OH - (McLean, 1982) resulting in the decrease of soil base cations and available micronutrients under soil coarseness.

Effect of soil depth on base cations and available micronutrients
The hypothesized decrease of exchangeable base cations and available micronutrients with soil depth was partially supported as only exchangeable Mg (Fig. 1b), available Fe (Fig. 2a), Mn (Fig 2b) and Zn (Fig. 2d) decreased with soil depth. Vertical 5 distribution of soil nutrients can be influenced by two opposite processes, leaching and biological cycling (such as plant absorption) (Truggill, 1988). Being a ubiquitous process in ecosystems, plant absorption of nutrients can transport soil elements aboveground and return the litterfall to soil surface (Stark, 1994). Especially in this sandy land or desertified grassland, plants tend to accumulate SOM or nutrients to 10 form 'island of fertility ' (Cao et al., 2008). In these sandy soils, leaching is also an essential process in shaping the vertical distribution of soil nutrients (Truggill, 1988).
As leaching moves nutrients downward while biological cycling moves them upward (Jobbá gy and Jackson, 2001), the unchanged Ca, K and Na concentrations might be the combining effects of leaching and biological cycling. Our results are in contrast 15 with previous studies suggesting that ecosystem were more capable to retain K than other base cations (Nowak et al., 1991;Jobbá gy and Jackson, 2001). In this case, it is obvious that many environmental factors, like soil types and plant community composition can be drivers for the vertical distribution of base cations and micronutrients. Stronger effect of plant absorption than leaching might contribute to 20 the shallower distribution of exchangeable Mg (Fig. 1b), available Fe (Fig. 2a), Mn ( Fig. 2b) and Zn (Fig. 2d). The dominant role of plant cycling in determining the vertical distribution of Mg, Fe, Mn and Zn might illustrate that these elements were scarcer and more limiting nutrients for plant growth in this semi-arid sandy ecosystem (Jobbá gy and Jackson, 2001).

Conclusions
The results showed that grassland soil coarseness decreased soil base cations of exchangeable Ca, Mg and Na as well as available micronutrients of Fe, Mn and Cu.
The loss of SOM, decrease of soil fine particles, and increase of soil pH were the Wang, T., Xue, X., Zhou, L., Guo, J.: Combating aeolian desertification in northern China, Land Degrad. Dev., 26, 118-132, 2015a. 30 Solid Earth Discuss., doi:10.5194/se-2016-18, 2016 Manuscript under review for journal Solid Earth     Letters indicate significant differences among treatments (lowercase letters) and differences among soil depths when averaging across all treatments (capital letters).