Asymmetric effects of grazing intensity on macroelements and microelements in grassland soil and plants in Inner Mongolia Grazing alters nutrient dynamics of grasslands

Abstract Grazing is a traditional grassland management technique and greatly alters ecosystem nutrient cycling. The effects of grazing intensity on the nutrient dynamics of soil and plants in grassland ecosystems remain uncertain, especially among microelements. A 2‐year field grazing experiment was conducted in a typical grassland with four grazing intensities (ungrazed control, light, moderate, and heavy grazing) in Inner Mongolia, China. Nutrient concentration was assessed in soil and three dominant plant species (Stipa krylovii, Leymus chinensis, and Cleistogenes squarrosa). Assessed quantities included four macroelements (carbon (C), nitrogen (N), phosphorus (P), and magnesium (Mg)) and four microelements (copper (Cu), iron (Fe), manganese (Mn), and zinc (Zn)). Soil total C, total N, total P, available N, and available P concentrations significantly increased with grazing intensity but soil Mg, Cu, Fe, Mn, Zn concentrations had no significant response. Plant C concentration decreased but plant N, P, Mg, Cu, Fe, Mn, and Zn concentrations significantly increased with grazing intensity. In soil, macroelement dynamics (i.e., C, N, and P) exhibited higher sensitivity with grazing intensity, conversely in plants, microelements were more sensitive. This result indicates macroelements and microelements in soil and plants had asymmetric responses with grazing intensity. The slopes of nutrient linear regression in C. squarrosa were higher than that of S. krylovii and L. chinensis, indicating that C. squarrosa had higher nutrient acquisition capacity and responded more rapidly to heavy grazing. These findings indicate that short‐term heavy grazing accelerates nutrient cycling of the soil–plant system in grassland ecosystems, elucidate the multiple nutrient dynamics of soil and plants with grazing intensity, and emphasize the important function of microelements in plant adaptation in grazing management.


| INTRODUC TI ON
Grassland ecosystems, covering 40% of the global land area, are one of the most important components of the terrestrial ecosystem, and function as a critical carbon (C) sink (White, Murray, & Rohweder, 2000). Grazing is a traditional grassland management, which has important economic and ecological functions, like providing animal products for human, preventing species invasion in some regions (Kang, Han, Zhang, & Sun, 2007;Kooijman & Smit, 2001).
Soil nutrient availability is important to plant growth and also regarded as a driving force in the alteration of species composition, community structure and function, and even community succession (Koerselman & Meuleman, 1996). However, there are conflicting results on the effect of grazing intensity on soil C, N, and P cyclings in grassland ecosystems. For example, many studies have supported that soil C, N, and P concentrations decreased with grazing intensity, reducing soil nutrient cycling (Han et al., 2008;He et al., 2011;Oliveira Filho et al., 2019). Other studies have shown that soil C, N, and P concentrations increased with grazing intensity, accelerating soil nutrient cycling (Liu et al., 2018;Liu, Zhang, Chang, Kan, & Lin, 2012;Reeder & Schuman, 2002;Reeder, Schuman, Morgan, & Lecain, 2004), while some have suggested that soil nutrient concentrations had no significant response with grazing intensity (Cui et al., 2005;Milchunas & Lauenroth, 1993). Compared with C, N, and P cyclings, there is poorly understanding of microelement dynamic in soil with grazing intensity, which affects plant community structure and function. Moreover, previous studies about soil nutrient cycling under different grazing intensities are examined by the spatial sequence instead of temporal sequence, which is lack of accurately control to background information (e.g., soil texture, vegetation heterogeneity, and grazing intensity) (Cui et al., 2005;Han et al., 2008;Steffens et al., 2008).
The impact of grazing on plant growth is not only limited by the changes in soil nutrient cycling but also extends to direct physical disturbance, like trampling and foraging (Liang, Gornish, Mariotte, Chen, & Liang, 2019;Ritchie, Tilman, & Knops, 1998). Plants must rapidly respond to these disturbances by regulating nutrient strategies, such as C, nitrogen (N), and phosphorus (P) concentrations and their stoichiometry (Liang et al., 2019;Yin et al., 2010).
However, plant nutrient concentrations under different grazing intensities remain controversial (He et al. 2019;Han et al., 2008;Ma et al., 2019). For example, some studies have shown plant N and P concentrations increased and plant N/P ratio decreased with grazing intensity (Li et al., 2016), while others reported the opposite result (He et al., 2019). Microelements are essential elements of some functional compounds in plants. For example, Copper (Cu) and iron (Fe) are important components in chlorophyll, some enzymes, and proteins (DalCorso, Manara, Piasentin, & Furini, 2014;Hänsch & Mendel, 2009). Many microelements are also involved in the primary and secondary metabolic processes of plants (Hänsch &Mendel, 2009), especially in N andP metabolisms (DalCorso et al., 2014). Thus, more attention needs to be paid to these microelements, which regulate plant growth and physiological activity (e.g., photosynthesis and respiration) (DalCorso et al., 2014).
Understanding the multi-nutrient dynamics in plants is helpful to elucidate the adaptation of plants to grazing and reveals the underlying mechanisms of changes in community structure and function.
The grassland in Inner Mongolia is an important piece of the Eurasian grassland ecosystem and is representative of grasslands in northern China. (Kang et al., 2007;Li, Jäschke, Guo, & Wesche, 2019). In the past three decades, overgrazing and change in grazing management practices (i.e., seminomadic farming systems to intensified settled livestock farming) have degraded approximately 90% of the grassland . Previous studies have shown the nutrient cycling of the soil-plant system remained uncertain in grassland ecosystems in Inner Mongolia (Cui et al., 2005;Han et al., 2008;He et al., 2011). Moreover, plant species have various nutrient strategies to grazing due to selective foraging (Ritchie et al., 1998).
Unfortunately, the responses of multi-nutrients in different plant species to grazing are poorly understood.
This study aims to clarify the changes in multiple elements of soil and plants with grazing intensity in grassland ecosystems in Inner Mongolia. A grazing experiment with four grazing intensities was con-

| Study sites
The study site (44°40.66′N, 116°28.32′E, a.s.l. 1,100 m) was located in a typical grassland in Inner Mongolia, China. The mean annual temperature of the study site is 0.5-1.0°C, with mean monthly temperatures ranging from −19.0°C (January) to 21.4°C (July). The mean annual precipitation is 280.5 mm, approximately 80% of which occurs during the growing season (May to September). The mean annual potential evaporation is about 1,600 to 1,800 mm. According to Chinese soil classification, the soil is chestnut. The dominant species is Stipa krylovii, and other species include Leymus chinensis, Cleistogenes squarrosa, Carex duriuscula, Potentilla tanacetifolia. S. krylovii, Leymus chinensis, and Cleistogenes squarrosa are perennial grasses and the aboveground biomass of these three species accounts for 70% of the total community aboveground biomass and greatly affects community structure and function. Grazing exclusion occurred from 2007 to 2012 in this region, and the aboveground biomass was harvested at the termination of the growing season once a year.

| Experimental design
We selected a flat area with uniform soil and vegetation and estab- where W is the weight of experimental sheep (31 kg), N is the number of experimental sheep (0, 4, 8, and 16 experimental sheep in one plot, respectively), D is the grazing time (91 days), W SSU is the weight of standard sheep (50 kg), and S is the area of grazing plot (1.33 hm 2 ). Grazing intensity included CK-ungrazed control (0 SSU day/year hm 2 ), LG-light grazing (170 SSU day/year hm 2 ), MG-moderate grazing (340 SSU day/year hm 2 ), and HG-heavy grazing (680 SSU day/year hm 2 ). Four grazing intensities were randomly assigned to the grazing plots with three replicates.

| Sample collection and measurements
Soil and plant samples were collected at the time of peak above- Plant and soil samples were ground with a mill (MM400; Retsch).
Total C and N concentrations were measured by a Vario EL (vario EL III CHNOS Elemental Analyzer, Elementar Analysensysteme GmbH).
Total P, Mg, Cu, Fe, Zn, and manganese (Mn) concentrations were determined by using ICP-OES (iCAP 6300 ICP-OES Spectrometer, Thermo Scientific). Soil available P and N concentrations were measured by an ultraviolet spectrophotometer (UV-2550, UV-Visible Spectrophotometer, Shimadzu) and alkaline hydrolysis diffusion, respectively. In this study, soil available N mainly includes ammonia nitrogen, nitrate nitrogen, amino acids, amides, and easily hydrolyzed proteins. Because of short-term grazing (2012-2015), we paid more attention to the effect of grazing intensity on the nutrient concentrations of soil and plants compared with time effect. Levene's test was used to examine the homogeneity of variances for all data. One-way ANOVA with liner polynomial contrast was selected to compare the differences in nutrient concentrations of soil and plants among different grazing intensities. One-variable linear regression was selected to quantify the relationships between plant N and P and other nutrient elements (i.e., Mg, Cu, Fe, Mn, and Zn). Data were indicated by mean ± 1SE (n = 3), and the significance level of data analysis was 0.05. All data analyses were conducted by SPSS 21.0 (SPSS Inc.).

| Response of soil nutrients to different grazing intensities
Macroelements and microelements in soil had inconsistent responses to grazing intensity. Specifically, the means of soil total C, total N, total P, available N, and available P concentrations significantly increased with grazing intensity (Figure 1; p < .05), especially at the depth of 0-10 cm. However, the means of soil Mg, Cu, Fe, Mn, and Zn concentrations had no significant response with grazing intensity ( Figure S1).  (Table 1).

| Response of plant nutrients to different grazing intensities
Plant C/N ratio in the three species remarkably decreased with grazing intensity (Figure 4a-c; p < .05), while plant N/P ratio had no consistent response (Figure 4d-f).   1). More importantly, the slopes of nutrient linear regression in C. squarrosa were higher than those in S. krylovii and L. chinensis ( Figure 5).  (Figure 3a-i), which aligned with findings from previous studies (Li et al., 2016;Zheng, Ren, Li, & Lan, 2012). Because mature and senescent plant tissue was foraged and plants need amounts of N and P to restructure stem and leaf (Ma et al., 2019). The new growth, comprised of young leaves and tillers, contained higher N and P concentrations, particularly in heavy grazing. Moreover, the increase in soil nutrient concentrations ( Figure 1) might indirectly accelerate plant growth, which also increased plant N and P concentrations.

| D ISCUSS I ON
We found that plant C/N ratio significantly decreased with grazing intensity (Figure 4a-c; p < .05), which was consistent with previous studies (Bai et al., 2012;Zheng et al., 2012). It confirmed that grazing accelerated N cycling and plant C/N ratio could indicate plant nutrient strategies to grazing disturbance. However, plant N/P ratio had no consistent response with grazing intensity (Figure 4d-f). This result indicated LG

| Effect of grazing on plant species
Plants had various nutrient strategies to grazing in grassland ecosystems (Liang et al., 2019). Previous studies reported that long-term heavy grazing decreased the dominance of L. chinensis and S. grandis, resulting in a surplus of community space and resources. C. squarrosa used these resources to reproduce rapidly, thereby degrading the grasslands (Wang, Liang, Liu, & Hao, 2000). We found that the slopes of nutrient linear regression in C. squarrosa were higher than that in S. krylovii and L. chinensis ( Figure 5). Higher nutrient concentrations in C. squarrosa might have a higher photosynthetic rate, protein and nucleic acid synthesis capacity, nutrient use efficiency, and drought resistance, which renders them more tolerant to heavy grazing. This result indirectly suggested that C. squarrosa had a higher capacity of nutrient acquisition under heavy grazing. Nutrient dynamics of different species might explain the changes in species composition and community structure during grassland degradation from the perspective of plant nutrient utilization.

| Effect of grazing on soil nutrients
In soil, grazing had more notable effects on macroelement concentrations than those of microelements. Short-term heavy grazing could accelerate the decomposition of plant litter, feces, and F I G U R E 5 Relationships between plant N and P concentrations and other element concentrations under different grazing intensities urine via livestock trample, which directly increased the input of organic matter and stimulated the mineralization of C, N, and P (Liu et al., 2018;Steffens et al., 2008). In addition, Hamilton and Frank (2001) reported that grazing increased the root exudation of carbon and increased the number of soil microbes. These reasons resulted that soil total C, total N, total P, available N, and available P concentrations increased with grazing intensity (Figure 1), supporting part of studies (Liu et al., 2018;Reeder & Schuman, 2002;Reeder et al., 2004). In contrast, soil microelement concentrations had no significant response to grazing intensity ( Figure S1), further showing that soil macroelements had sensitive indicators for grazing. This result supported previous studies (Mathews, Sollenberger, Nair, & Staples, 1994).
Interestingly, He et al. (2011) and Steffens et al. (2008) found soil nutrients decreased with grazing intensity. In this study, 6-year grazing exclusion before the experiment provided enough organic matter (e.g., plant litter) for soil nutrient cycling. In addition, grassland ecosystems possess a certain elasticity and threshold to grazing. Short-term heavy grazing had positive effects on soil nutrients , while extended grazing time could have negative effects on the nutrient balance of the soil-plant system (He et al., 2011;McSherry & Ritchie, 2013;Steffens et al., 2008). Thus, compared with plant nutrients, the impacts of grazing intensity on soil nutrients were complex, depending on land use pattern, community type, and grazing intensity and time (Reeder & Schuman, 2002).
In the future, the effect of grazing on soil nutrients should be longterm and continuous observation.

ACK N OWLED G M ENTS
We appreciated the reviews for their warm-hearted work and valuable suggestions. We would like to thank Courtney Buoncore at Princeton University for his assistance with English language and grammatical editing. This study was supported by The National Basic Research Program (2014CB138802).

CO N FLI C T O F I NTE R E S T
No conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data associated with this paper has been uploaded in Dryad: https:// doi.org/10.5061/dryad.37pvm cvgz.