Deng et al. Carbon Research (2023) 2:19 https://doi.org/10.1007/s44246-023-00051-7 ORIGINAL ARTICLE Open Access Carbon in Chinese grasslands: meta‑analysis and theory of grazing effects Lei Deng1,2, Zhouping Shangguan1,2*, Stephen Mackenzie Bell3, Andrey V. Soromotin4, Changhui Peng5, Shaoshan An1,2, Xing Wu6, Xingliang Xu7, Kaibo Wang8, Jianping Li9, Zhuangsheng Tang10, Weiming Yan1,2, Fengbao Zhang1,2, Jiwei Li2, Jianzhao Wu1 and Yakov Kuzyakov11,12 Abstract Globally, livestock grazing is an important management factor influencing soil degradation, soil health and carbon (C) stocks of grassland ecosystems. However, the effects of grassland types, grazing intensity and grazing duration on C stocks are unclear across large geographic scales. To provide a more comprehensive assessment of how grazing drives ecosystem C stocks in grasslands, we compiled and analyzed data from 306 studies featuring four grassland types across China: desert steppes, typical steppes, meadow steppes and alpine steppes. Light grazing was the best management practice for desert steppes (< 2 sheep ­ha−1) and typical steppes (3 to 4 sheep ­ha−1), whereas medium grazing pressure was optimal for meadow steppes (5 to 6 sheep ­ha−1) and alpine steppes (7 to 8 sheep h ­ a−1) lead- ing to the highest ecosystem C stocks under grazing. Plant biomass (desert steppes) and soil C stocks (meadow steppes) increased under light or medium grazing, confirming the ‘intermediate disturbance hypothesis’. Heavy grazing decreased all C stocks regardless of grassland ecosystem types, approximately 1.4 Mg ­ha−1 per year for the whole ecosystem. The regrowth and regeneration of grasslands in response to grazing intensity (i.e., grazing optimization) depended on grassland types and grazing duration. In conclusion, grassland grazing is a double-edged sword. On the one hand, proper management (light or medium grazing) can maintain and even increase C stocks above- and belowground, and increase the harvested livestock products from grasslands. On the other hand, human-induced overgrazing can lead to rapid degradation of vegetation and soils, resulting in significant carbon loss and requiring long-term recovery. Grazing regimes (i.e., intensity and duration applied) must consider specific grassland characteris- tics to ensure stable productivity rates and optimal impacts on ecosystem C stocks. Highlights (1) Light grazing can maintain and even increase C stocks in grassland biomasses. (2) Human-induced overgrazing leads to significant carbon loss that requires long-term recovery. (3) Heavy grazing decreases C stocks regardless of grassland ecosystem types. (4) Grassland state before grazing is critical to the response of grassland to grazing duration. (5) The ‘intermediate disturbance hypothesis’ fits to non-degraded grasslands before grazing. Keywords Grazing duration, Grazing intensity, Grassland, Overgrazing, Plant and soil, Carbon dynamics, Land use change Handling editor: Hailong Wang. *Correspondence: Zhouping Shangguan shangguan@ms.iswc.ac.cn Full list of author information is available at the end of the article © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. Deng et al. Carbon Research (2023) 2:19 Page 2 of 17 Graphical Abstract 1 Introduction light and medium grazing can actually be beneficial for Global grasslands cover approximately 40% of terres- soil C accumulation (Hafner et al. 2012; Jiang et al. 2020), trial area, store 10%–30% of soil organic carbon (C) the absence of grazing can lead to soil C loss because with a sequestration rate of 0.5 Pg ­yr−1, and harbor litter accumulation decreases with little stimulation of more than 10% of terrestrial biomass C (Booker et al. grass growth (Gao 2007; Hassan et al. 2021). Thus, there 2013; Dlamini et al. 2016). Due to their large area and are complex interactions between grazing and grassland potential for sequestering atmospheric C ­ O2, grass- C stocks (Milchunas and Laurenroth 1993). The ecosys- lands are critical to the global C cycle (Deng et al. 2017; tem C stocks in various grassland types may have spe- Abdalla et al. 2018; Bai and Cotrufo 2022). However, cific response patterns to increasing grazing intensities about 49% of global grassland area has been degraded because of the differences in: 1) climate, 2) net primary to some extent by overgrazing (Bardgett et al. 2021). In production, 3) composition of plant communities, and, China, nearly 61% of grasslands suffer from degradation 4) soil properties. Accurate estimations of the thresholds due to intensive grazing (Zhou et al. 2014). Degraded of grazing intensity (i.e., stocking rates) is particularly grasslands not only fail to provide subsistence for the important for promoting C sequestration in different local people (Zhao et al. 2017; Bardgett et al. 2021), but grassland types. While it is known that grassland ecosys- also fail to mitigate climate change owing to the neg- tem C stocks are affected by grazing intensities and graz- ative effect on ecosystem C stocks (Wang et al. 2016; ing animals (Eldridge et al. 2017), grazing duration must Deng and Shangguan, 2021). Therefore, there is a new also be considered in any robust synthesis (Zhang et al. challenge to increase or at least maintain C sequestra- 2017; Deng et al. 2017). Even under light grazing condi- tion in grassland ecosystems and protect existing ter- tions, grassland ecosystems can degrade through long- restrial C stocks in the context of climate change. term grazing (Jiang et al. 2020), sometimes leading to Grazing intensity is the main factor affecting the woody encroachment (Gao et al. 2021) which decreases C source/sink function of grasslands directly (Zhao productivity and biodiversity (Bai et al. 2012; Zhang et al. 2017). Most studies have shown that overgrazing et al. 2021a). Additionally, heavy grazing during the few strongly declined C stocks in grassland (Dlamini et al. first years may have no noticeable effects on grassland C 2016; Jiang et al. 2020; Liu et al. 2018). Grassland pri- stocks (Milchunas et al. 1998; Elmore and Asner 2006). mary productivity dramatically drops because of over- Soil C content has also been shown to linearly increase grazing, thereby reducing the C input from vegetation under light grazing for more than 40 years (Conant et al. to soil (Derner et al. 2006; Zuo et al. 2018). Overgrazing 2001). This indicates that the effects of grazing duration accelerates ­CO2 emission from the soil due to organic on ecosystem C stocks remain largely unclear, and the matter decomposition, thus speeding up C transfers into optimal grazing duration and intensities for C sequestra- the atmosphere and promoting soil erosion (Chen et al. tion need to be quantified based on a broad meta-anal- 2016; Zhao et al. 2017; Moinet et al. 2016). Although ysis taking into account the specifics of grassland types. Deng et al. Carbon Research (2023) 2:19 Page 3 of 17 Grazing effects on C stocks depend on plant biomass, types, grazing intensities, grazing durations, and ecosys- vegetation diversity, and soil properties (Deng et al., tem components. To our knowledge, no study has been 2014a; McSherry & Ritchie 2013; Abdalla et al. 2018), all conducted to test each of these theories based on a large of which are affected by local climate and environmental dataset including all these factors. conditions, e.g., precipitation, temperature, light intensity The grassland area in China covers 265 million ha, and duration of the vegetation restoration (Altesor et al. accounting for 27.6% of the country’s total land area 2005; Yan et al. 2013; Deng et al. 2017). While the above- (http://​w ww.​gov.​cn/​xinwen/​2021-​08/​26/​conte​nt_​56334​ ground biomass decreases with grazing pressure (Zhu 97.​htm). Based on the temperature, precipitation and et al. 2016, 2018), soil nutrient turnover is accelerated altitude ranges, the grassland ecosystems in China are (Pineiro et al. 2009; Dong et al. 2021). This is attributed usually classified into four major types: desert steppes, to the addition of animal dung and urine, which in turn typical steppes, meadow steppes, and alpine steppes affects soil C dynamics and microbial activities (Vandan- (Kang et al. 2007; Yan et al. 2013). Nearly all unculti- dorj et al. 2017). Grazing also affects soil properties such vated Chinese grasslands are grazed by large mammals as nitrogen (N) content, pH, bulk density, infiltration such as cattle, sheep, goats and yak (Yan et al. 2013), and rate, microporosity, and moisture, all of which are closely thus, grazing is a critical factor affecting ecosystem func- related to ecosystem C stocks (Han et al. 2008). Although tion (Deng et al. 2014a, 2017; Eldridge et al. 2017; Bai many individual studies have clarified the effects of graz- and Cotrufo 2022). Despite this, only two meta-analyses ing on soil C stocks and the underlying mechanisms reported the effects of grazing on C stocks in China’s (Conant et al. 2001; Hafner et al. 2012; Jiang et al. 2020; grasslands (Yan et al. 2013; Jiang et al. 2020), and they Bai and Cotrufo 2022), generalized relationships are still mainly evaluated grazing intensity based on limited data- unclear due to regional differences in climate, soil prop- bases without considering the effects of grassland types, erties, and grassland types, as well as disturbance factors grazing duration and stocking rate. Here, we conducted such as grazing intensities/stocking rates and grazing a more in-depth investigation into the wider effects of duration. Therefore, a more systematic analysis as well as grazing on C stock in grassland ecosystems. We hypoth- a theoretical review is required. esized that grassland carbon stocks’ response to grazing Four hypotheses have been proposed for the struc- are diversified among different grassland ecosystems and ture and functions of grassland ecosystems in responses their responses to grazing intensity are affected by graz- to grazing: the ‘intermediate disturbance hypothesis’ ing duration. This synthesis compiled and analyzed data (Tilman and Downing 1994), the ‘grazing optimization from 306 studies carried out on 216 sites in China, in hypothesis’ (Edelstein-Keshet 1986), the ‘dynamic bal- an effort to answer two research questions: (a) How do ance hypothesis’ and the ‘dynamic imbalance hypoth- ecosystem C stocks (above- and belowground) respond esis’ (Palmer et al. 2016). The ‘intermediate disturbance to increasing grazing intensities in four dominant grass- hypothesis’ suggests that light or medium grazing land types: desert, typical, meadow, and alpine steppes? changes the existing vegetation stock and stimulates (b) What are the temporal patterns of ecosystem C stocks regrowth leading to increased photosynthesis, more during grazing duration under each grazing intensity? By biomass production per vegetation season, and more C using these data to verify theories of grassland response input belowground (Tilman and Downing 1994). This to grazing and to explore pathways towards sustainable regrowth was termed as a “compensatory effect”, which grassland management, this study provides new insights is linked to the ‘grazing optimization hypothesis’ (Edel- and a foundation for management decisions regarding stein-Keshet 1986). However, the increasing grazing global grasslands. reduces the recovery potential of the vegetation, and con- sequently, this means that further grazing intensity leads 2 Materials and methods to the collapse of the grassland ecosystem (McNaughton 2.1 Data compilation 1979). The ‘dynamic balance hypothesis’ refers to the To develop a comprehensive database in China, peer- coordinated and stable-state achieved in an ecosystem reviewed journal articles published before 2021 were within a certain period, including the stability of eco- searched using Web of Science (http://​apps.​webof​ system structure, functions, and energy input and out- knowl​edge.​com/), Google Scholar (http://​schol​ar.​ put (Palmer et al. 2016). Finally, the ‘dynamic imbalance google.​com/), and China Knowledge Resource Inte- hypothesis’ proposes that the structure and functions of grated Database (http://​w ww.​cnki.​net/) with the fol- an ecosystem are increased or decreased under grazing, lowing search term combinations: ‘grazing’ or ‘grazing implying difficulties maintaining the ecosystem steady- intensity’ and ‘biomass’ or ‘soil carbon’ and ‘grass- state (Palmer et al. 2016). Unfortunately, these theories land’ or ‘pasture’ or ‘meadow’ or ‘rangeland’ and have been individually developed for specific grassland ‘China’. We also searched for articles that were cited Deng et al. Carbon Research (2023) 2:19 Page 4 of 17 in the publications we found in the previous step. To The raw data were either obtained from tables or avoid bias in the selection and increase relevance, we extracted by digitizing graphs using the GetData Graph extracted papers based on the following criteria: Digitizer (version 2.24, ver. 2.24, http://​getda​ta-​graph-​ digit​izer.​com, Russian Federation). For each study, the (a) Only studies using paired-site and/or chronose- following information was compiled: sources, location quence approaches, with similar soil and climatic (i.e., longitude and latitude), climatic data (i.e., mean conditions for both the grazed and ungrazed annual precipitation, and mean annual temperature), ele- sites were selected for the database; studies were vation, land use types (including grazing exclusion sites excluded if the experiments were not adequately and grazing sites), grassland types (i.e., desert, typical, replicated or if the paired sites were confounded by meadow and alpine steppes), dominant species, grazing different soil types. intensity, stocking rates (i.e., yak, cattle, sheep or horse (b) Grazing intensity was clearly defined in quantitative per hectare), grazing duration (i.e., years since grazing), or qualitative terms for stocking rate, grazing dura- grassland utilization rate (i.e., aboveground biomass utili- tion, and livestock type (i.e., cattle, sheep, or others), zation rate), grazing period (i.e., growth period, ungrowth (c) Experiments with durations of less than one grow- period or whole year), above- and belowground biomasses ing season were excluded to avoid short term noise, or C stocks, litters biomasses or stocks, total biomasses or (d) Soil C stocks were provided or could be calculated C stocks, plant coverage, plant height, plant diversity (i.e., based on SOC or SOM content, bulk density, and species richness, Shannon–wiener Index, Margalef rich- soil depth, ness Index, Pielou Evenness Index), soil depth from soil (e) Sampling depths for belowground biomass varied surface, bulk density, soil pH, soil organic C (SOC) con- in studies. Considering that most root biomass is tent and total N content (TN) in each layer of 0–100 cm distributed in the top 50 cm and most studies sam- soil depths. Stocking rates and grassland utilization rates pled roots to this depth, data of root biomass in (Eq. 1) corresponding to different grazing intensities in 0–50 cm were used to investigate the C stocks in the four grassland types of the dataset are shown in Fig. belowground biomass, S1. The grazing intensities were categorized based on the (f ) Only the first rotation of grazing was considered text of each study, that is ungrazed, light, medium, and and data for 0–100 cm soil layers were extracted. heavy. To make different units of stocking rate compara- The synthesis only estimated soil C stocks for the ble among different sites, we normalized the raw values of top 20 cm because more than 90% of the studies stocking rates to standardized animal units by using the were conducted at this soil depth. criteria: one yak equivalent for three sheep/goat, and one cow-calf or horse equivalent for four sheep/goat (Val- In total, the final dataset comprised of 306 studies lentine 1990). If climatic factors were not reported, we (Appendix Dataset S1), including 216 sites in 16 prov- extracted these data from the WorldClim database (http://​ inces of China (Fig. 1), which represented most of the www.​world​clim.​org/) based on site location information grassland area of China. This dataset is the largest com- (latitude and longitude). Moreover, to depict more appar- piled thus far with a focus on the effects of grazing on ent trends of the C pools, the ages of grazing duration grassland C stocks in China. The latitude of the study were divided into six groups: 1, 2–3, 4–5, 6–10, 11–20, sites ranged from 25.21° to 49.34° N and the longitude and > 20 years. We divided the stocking rate into six stock- ranged from 81.00° to 148.16° E. Altitude ranged from ing rate categories to explore the effects on the grassland 140 to 4730 m. Grazing intensity provided in the col- C stock changes: < 2, 2–4, 5–6, 7–8, 9–10, and > 10 sheep lected studies ranged from low to medium and high graz- units per hectare. ing. Grasslands in China range from the high-altitudes of the Qinghai-Tibetan plateau to the low-altitude steppes of Inner Mongolia. The grassland ecosystems in China 2.2 Data calculation are classified into four major types (Kang et al. 2007): 2.2.1 Grassland utilization rate desert steppes, typical steppes, meadow steppes, and As animal stocks only feed on the aboveground biomass alpine steppes. The classifications were mostly provided in grasslands, we used the following equation to calculate in the individual studies, but when they were absent, we the grassland utilization rate: defined the grassland types based on the temperature, AGBCK − AGBG precipitation and altitude ranges (Yan et al. 2013). The Ru = × 100% (1) climatic conditions for each grassland type are presented AGBCK in Fig. 1. In which, R­ u is the grassland utilization rate (%), A ­ GBCK is the aboveground biomass in the ungrazed sites (g m ­ −2), Deng et al. Carbon Research (2023) 2:19 Page 5 of 17 Fig. 1 Locations of the sampling sites reported in the individual studies collected in this synthesis. The grassland ecosystems are classified into four major types (Kang et al. 2007): typical steppes, meadow steppes, desert steppes and alpine steppes. The above figures indicate the livestock equivalents as follows: one goat equivalent for one sheep, one yak equivalent for three sheep/goats, and one cow-calf or horse equivalent for four sheep/goat, which is estimated by the herbivorous value of the livestock (Vallentine 1990). The left bottom inset figure indicates the range of mean annual temperature (MAT) and mean annual precipitation (MAP) for the sites of the four grassland steppes in this meta-analysis and ­AGBG is the aboveground biomass in the grazing only had aboveground biomass (AGB) or belowground sites (g ­m−2). biomass (BGB), the study used the root/shoot ratio (R/S) of each grassland type and grazing intensities to calculate their belowground biomass (BGB) or aboveground bio- 2.2.2 Vegetation C stock mass (AGB) (Fig. S2). For vegetation C stock, we used the following equation (Deng et al. 2017): 2.2.3 Soil C stocks CB = B × Cf (2) If the samples reported only had SOM, their SOC was calculated by the relation between SOM and SOC as fol- In which, C­ B is the vegetation C stock (g ­m−2), B is the lows (Deng et al. 2017): vegetation biomass (g ­m−2) or litter biomass (g ­m−2), and ­Cf is the plant or litter biomass C content. The plant bio- SOC = SOM × 0.58 (3) mass C coefficient for estimating the herbaceous C stock where SOC is the soil organic C concentration (g k­ g ) −1 was set at 0.45 (Deng et al. 2017). If the samples reported and SOM is the soil organic matter (g ­kg−1). Deng et al. Carbon Research (2023) 2:19 Page 6 of 17 The SOC stocks were calculated using the following global average in terms of soil C distribution with depth. equation (Deng et al. 2017): Previous meta-analyses have also used the same method, and concluded that depth corrections did not alter the SOC × BD × D Cs = (4) overall pattern of soil C stock dynamics during vegeta- 10 tion development (Li et al. 2012; Deng et al. 2014b, 2017; in which, ­Cs is soil organic C stocks (Mg ­ha−1); SOC is Yu et al. 2023). soil organic C content (g ­kg−1); BD is soil bulk density (g ­cm−3), and D is soil thickness (cm). 2.3 Data analysis Soil BD estimates are critical for calculations of ­Cs, but Analysis of variance (ANOVA) was conducted to evalu- many studies did not measure this. We established an ate whether the changes in plant C stocks, soil C stocks, empirical relationship between SOC content and soil BD plant height, plant coverage, R/S, soil pH, and soil C/N with the reported values for ungrazed sites and grazed were significantly different depending on grazing inten- sites from the Appendix Dataset S1. Then, the missing sity, grazing duration, grassland types, or grazing peri- values of soil BD were interpolated using the predicted ods. Differences were evaluated at the 0.05 significance values from the empirical functions (Exponential Decay, level (p < 0.05). When the homogeneity of variance was Double, 4 Parameter, Fig. S3, Eqs. 5 and 6): confirmed and significance was observed at the p < 0.05 level, a least significant difference (LSD) test was used for BDUnG = 1.53e0.01SOC − 0.13, r 2 = 0.47, p < 0.0001, n = 113 multiple comparisons. Pearson correlations were used to (5) analyze ecosystem C stocks with climate (MAP, MAT), BDGG = 1.27e 0.01SOC 2 + 0.20, r = 0.42, p < 0.0001, n = 333 elevation, grassland utilization rate, grazing duration, (6) R/S, and plant community characteristics in the differ- To increase the comparability of data derived from dif- ent grazing grasslands (Table S1). Regression analysis was ferent studies, the original soil C data were converted to conducted to analyze the relationships among plant bio- soil C stocks in the top 20 cm using the depth functions mass C stocks as well as the relationships between soil C developed by Jobbágy and Jackson (2000) according to stocks and above- or belowground biomass C stocks, the the following equations: relationships of R/S, soil C/N ratios with stocking rates or grazing duration, and the relationships of soil N stocks Y = 1 − βd (7) and soil C stocks, TN content and SOC content (Table S2). In addition, a multivariable linear regression analy- 1 − β 20 sis was conducted to quantify the relationships between X20 = × Xd0 (8) ecosystem C stocks and five driving factors of MAP, 1 − β d0 MAT, elevation, stocking rate, and grazing duration in where Y represents the cumulative proportion of the the four grassland types. All statistical analyses were per- soil C stock from the soil surface to depth (d, cm); β is formed using the SPSS software program, ver. 17.0 (SPSS the relative rate of decrease in the soil C stock with soil Inc., Chicago, IL, USA). depth; ­X20 denotes the soil C stock in the upper 20 cm; d0 denotes the original soil depth (cm); ­Xd0 is the original 3 Results soil C stock. Although Jobbágy and Jackson (2000) pro- 3.1 Grassland C stocks change with grazing intensity vided the depth distribution of soil C for 11 biome types Grassland type and grazing intensity affected ecosys- globally, there was no significant difference in the depth tem C stocks (Fig. 2). Heavy grazing decreased plant distribution among biome types or between individual biomass and soil C stocks in all grassland types, but biomes and the global average. Therefore, in the present light grazing slightly increased the above- and below- study, the global average depth distributions for C were ground biomass stocks only in the desert steppes adopted to calculate β (i.e., 0.9786) in the equations (Li (Fig. 2). Light grazing increased belowground plant et al. 2012). biomass (roots) in typical steppes, but had no effects It should be noted that potential uncertainties may be on roots in alpine and meadow steppes (Fig. 2). Both introduced by this dataset standardization, mainly due light grazing and medium grazing increased the soil C to the difference in C distribution through the soil pro- stocks in all grassland types, except in alpine steppes file between ungrazed and grazing sites, and among the (Fig. 2). In terms of ecosystem C accumulation, light different stages following grazing exclusion. However, grazing is good for desert steppes and typical steppes, as it has been stated, there was no significant difference whereas medium grazing is the best option for among the 11 biome types included in Jobbágy and Jack- meadow steppes and alpine steppes (Fig. 2). son’s study (2000) or between individual biomes and the Deng et al. Carbon Research (2023) 2:19 Page 7 of 17 Fig. 2 Carbon stocks depending on grazing intensity in the four grassland types of desert, typical, meadow and alpine steppes. UG, Ungrazed; LG, Low grazing; MG, Medium grazed; HG, High grazed. Shoot C, aboveground biomass carbon stocks; Root C, belowground biomass carbon stocks; Litter C, litter biomass carbon stocks; Total biomass C, total biomass carbon stocks including both shoot C and root C; Soil C, soil carbon stocks; Ecosystem C, ecosystem carbon stocks including all plant biomass C, litter C and soil C. p < 0.05 indicates significant differences among the grassland types or grazing intensities or their interactions at 0.05 level. All the data are presented as Mean ± SE Grazing intensities depend on stocking rates. Over- 3.2 Grassland C stocks change with grazing duration all, higher stocking rates reduced plant and soil C Grazing duration is an important factor of ecosystems stocks. Along with the increasing grazing intensi- C stocks. Of the 72 potential carbon, intensity and ties, aboveground biomass carbon stocks decreased grassland type effects, two-thirds showed significant from 0.6 Mg ­ha−1 to 0.1 Mg ­ha−1, from 1.1 Mg ­ha−1 to linear declines, but 16 showed a unimodal response, 0.5 Mg ­ha−1, from 2.0 Mg ­ha−1 to 0.4 Mg ­ha−1 and generally increasing up to 2 or 4 years, then declined from 1.3 Mg ­ha−1 to 0.8 Mg ­ha−1 in the desert steppes, (Fig. 4). Grassland ecosystem C stocks increased in typical steppes, meadow steppes and alpine steppes, the early grazing duration (< 3 yr) under light grazing respectively (Fig. 3). Compared with ungrazed grass- (Figs. 4 and S5), especially in the desert, typical, and lands, grazed grassland C stocks did not change at meadow steppes. For the short-term grazing duration the stocking rates of about 1 ~ 2, 3 ~ 4, 5 ~ 6, and 7 ~ 8 (< 6 years), light grazing increased plant biomass C sheep units per hectare for desert steppes, typical stocks in the desert steppes, but they declined during steppes, meadow steppes, and alpine steppes, respec- long-term grazing (Fig. 4). Light grazing (2 to 4 sheep tively (Fig. 3, Fig.S4). ­ha−1) increased belowground biomass in the typi- cal steppes (Fig. 4). The belowground biomass in the meadow steppes under light grazing (1 to 2 sheep ­ha−1) Deng et al. Carbon Research (2023) 2:19 Page 8 of 17 Fig. 3 Grassland carbon stocks depending on stocking rates in the four grassland types of desert, typical, meadow and alpine steppes. Note: UG, Ungrazed. Shoot C, aboveground biomass carbon stocks; Root C, belowground biomass carbon stocks; Litter C, litter biomass carbon stocks; Total biomass C, total biomass carbon stocks including both shoot C and root C; Soil C, soil carbon stocks; Ecosystem C, ecosystem carbon stocks including all plant biomass C, litter C and soil C. All the data are presented as Mean ± SE and medium grazing (2 to 4 sheep ­ha−1) were stable 3.3 Factors affecting grassland C stocks depending even under long-term grazing (Fig. 4). Plant biomass C on grazing stocks in the alpine steppes were also stable under light Climate is an important factor affecting grassland ecosys- grazing (2 to 4 sheep ­ha−1), though C stocks decreased tems (Fig. 5, Tables S1 and S2). Plant C stocks in grass- for litter, soil and the whole ecosystem (Fig. 4). When lands usually increased with precipitation but they were the grazing duration is longer than 2 years, soil C stocks not influenced by temperature (Fig. 5). This phenomenon decrease faster than aboveground or belowground bio- was only found in the ungrazed grasslands (Fig. 5). How- mass (Fig. S5). Overall, soil C stocks decreased from ever, grazing offset the correlations between plants C 59.6 Mg ­ha−1 to 26.4 Mg ­ha−1, from 64.3 Mg ­ha−1 to stocks and climate conditions (Table S1). Plant C stocks 14.8 Mg ­ha−1, from 53.2 Mg ­ha−1 to 18.5 Mg ­ha−1 in the remained stable independent of MAP or MAT under all light, medium, and heavy grazing, respectively (Fig. S5). the grazing intensities (Table S1). Soil C stock increased All litter C stocks decreased linearly with grazing dura- with precipitation, but it was reduced with temperature tion, regardless of grassland types (Fig. 4). It is worth (Fig. 5). In the light and medium grazed grasslands, soil noting that heavy grazing decreased C stocks linearly C stocks decreased with stocking rates, but there were no with grazing duration, at about 1.4 Mg ­ha−1 per year in correlations in the heavy grazing grasslands (Table S1). the whole grassland ecosystems (Fig. S6), mainly linked Ecosystem C stocks in the light grazing grasslands to the decline in soil C stocks. decreased with increasing Pielou index, indicating that light grazing increased the evenness of the grasslands. Higher plant coverage increased soil C stocks (Table Deng et al. Carbon Research (2023) 2:19 Page 9 of 17 Fig. 4 Response of grassland C stocks to grazing duration (maximum ~ 30 years) under each grazing intensity in the four grassland types. The values below each grazing intensity indicate the corresponding stocking rate (sheep h ­ a−1) in the dataset of this synthesis. The numbers under the bell shaped curves show the period when the maxima will be reached in years. Shoot C, aboveground biomass carbon stocks; Root C, belowground biomass carbon stocks; Litter C, litter biomass carbon stocks; Total biomass C, total biomass carbon stocks include both shoot C and root C; Soil C, soil carbon stocks; Ecosyst. C, ecosystem carbon stocks include all plant biomass C, litter C and soil C S1). The multivariable linear regression model analysis biomass of the desert steppes (Fig. 2) because: 1) the veg- showed that soil C stocks were mainly affected by the etation productivity index was the highest under light climate factors, stocking rate, and grazing duration in all grazing by stimulating the compensatory plant growth grassland types, but factors affecting plant C stocks were (Evju and Mysterud 2010; Zhu et al. 2018); and, 2) live- determined by the grassland types (Table S2). In the typi- stock increases the evenness of the grasslands (Table S1) cal meadow steppes, plant C stocks were mainly affected and reduces the competition of dominant species for by grazing intensity and duration. In the desert and resources (water, nutrients, light), which provides space alpine steppes, the MAP and MAT were the main factors for the survival of alien species and the fast recovery of influencing plant C stocks (Table S2). native species (Yan et al. 2013; Zhu et al. 2018; Gao et al. 2021). This conforms to the ‘intermediate disturbance 4 Discussions hypothesis’ (Tilman and Downing 1994). Therefore, rota- 4.1 Effects of grazing on plant biomass C stocks tional grazing intensity maintains or increases grassland Grazing disturbance is crucial for the growth and productivity and is an important grassland management development of plant species, and thus can change the strategy to increase C stocks. community composition, structure, diversity, and pro- Light grazing, however, had no effects on shoot biomass ductivity (Fig. S7; Zhao et al. 2017; Mou et al. 2018). Our in the alpine steppes (Fig. 2), because alpine meadows meta-analysis showed that grazing decreased the shoot have higher compensatory growth or over-compensatory biomass in the desert, typical and meadow steppes, espe- growth under grazing (Kuzyakov et al. 2002; Lu et al. cially under heavy grazing (Fig. 2), because shoots were 2017) and thus increase the primary productivity (Sun removed and the remaining stubbles were damaged by et al. 2019). Heavy grazing strongly decreased the shoot livestock trampling (Ferraro and Oesterheld 2002). Light and root C stocks (Figs. 2, 3 and 4), as overgrazing (i.e., grazing, however, slightly increased the shoot and root grazing leading to degradation of grassland) with higher Deng et al. Carbon Research (2023) 2:19 Page 10 of 17 Fig. 5 Effects of the MAP (mean annual precipitation) on plant biomass C stocks in the ungrazed sites, and of the MAP and MAT (mean annual temperature) on soil C stocks in ungrazed and grazing sites, as well as the relationships between soil C stocks and plant biomass C stocks in ungrazed sites. Shoot C, aboveground biomass carbon stocks; Root C, belowground biomass carbon stocks; Litter C, litter biomass carbon stocks; Total biomass C, total biomass carbon stocks including both shoot C and root C; Soil C, soil carbon stocks. All regression lines are significant at p < 0.01 (except Litter C, which is at p < 0.05) stocking rates reduced the growth and regeneration of Zhu et al. 2018); 2) selective feeding by livestock reduces grassland plants, and reduced C allocation belowground the litter amounts and litter inputs (Fig. 3), thus reduc- (Kuzyakov et al. 2002; Unteregelsbacher et al. 2012; Liu ing the soil C content (Sun et al. 2011; Qu et al. 2022); et al. 2015). Overgrazing altered the distribution of mat- 3) trampling, particularly in the heavy grazing, destroys ter and energy among the plant organs, leading to a small soil structure and increases soil erosion which reduces leaf area index (LAI), short grass height, sparse plant cov- C in topsoils (Zhao et al. 2017); and, 4) soil C ­ O2 emis- erage (Fig. S7), and low light interception. Consequently, sions increase due to soil acidification (decreased soil photosynthesis decreased, especially in the high grazing pH) under intensive grazing (Fig. 7; Zamanian et al. 2018; intensity. If the grassland is overgrazed during the seed- Raza et al. 2020). Light grazing can increase the soil C bearing period, the livestock decreases the tiller number, stocks in the desert steppes, typical steppes, and meadow leaf, plant height, growth rate, and total biomass (Zheng steppes (Fig. 2) because grazing has a stimulatory effect et al. 2010; Eldridge et al. 2016), and thus, it affects the on root growth (Fig. 4). More assimilation products are productivity of the entire grassland (Milcu et al. 2016). allocated to the roots as storage for regrowth following Therefore, the compensatory regeneration of herbage grazing (Kuzyakov et al. 2002; Pucheta et al. 2004; Hafner is weak, and present C stocks of plant biomass are low et al. 2012), thereby increasing root/shoot ratios (Fig. under overgrazing (Patty et al. 2010). S2 and S8). Additionally, soil organic matter increases with animal manure brought outside (Liu et al. 2020) 4.2 Effects of grazing on soil C stocks and declines in organic matter decomposition due to Soil C stocks depend on the balance between the C low microbial activities under grazing (Zhao et al. 2017; inputs by plants and soil organic matter decomposition Zheng et al. 2021), leading to an increase in soil C stocks. (Deng et al. 2014b, 2017). According to our meta-analy- However, light grazing and medium did not affect soil C sis, grazing decreased the soil C stocks in typical steppes stocks in the alpine steppes (Fig. 3) because: 1) appropri- (Fig. 3) because: 1) grazing decreases shoot biomass ate (light – medium) grazing does not reduce C inputs and photosynthesis, which reduces the belowground C from the plants into soils due to higher compensa- input (Zheng et al. 2010; Unteregelsbacher et al. 2012; tory growth or over-compensatory growth of the alpine Deng et al. Carbon Research (2023) 2:19 Page 11 of 17 steppes (Fig. 3); 2) animal manure inputs can offset the Deng et al. 2021) and a decrease in ecosystem C stocks. C losses due to trampling (Milchunas and Laurenroth In summary, this decrease in ecosystem C stocks under 1993); and, 3) grazing changes the composition of plant long term grazing is due to: 1) low plant productivity and species in the alpine steppes, and increased plant diver- carbon inputs (Gamoun 2014; Deng et al. 2021); 2) soil sity promotes soil organic matter formation (Zhao et al. N deficiencies due to higher N losses, which limit plant 2009; Gao et al. 2021). The stability of soil C stock in growth (McNaughton et al. 1997); and, 3) high propor- alpine grasslands indicates good adaptability and resist- tions of heterotrophic microorganisms that accelerate ance to grazing (Wang et al. 2016). soil organic matter decomposition (Patra et al. 2005; Raza C-N interactions are very important factors in deter- et al. 2020). mining whether C sinks in ecosystems can be sustained over the long term (Luo et al. 2006; Deng et al. 2017). In 4.3 Effects of grazing on ecosystem C stocks the grazed grasslands, soil C/N decreased during the first Grazing affects grassland ecosystems through trampling, 4–5 years (Fig. 6) due to a fast decline in the soil C pools feeding and the excrements of livestock (McSherry and (Fig. S5). Under long-term grazing (> 10 yr), the declin- Ritchie 2013; Deng et al. 2017; Wilson et al. 2018). Graz- ing soil C pool reached a stable level (Fig. S5). However, ing duration is an important factor influencing C stocks. long-term grazing also leads to higher N mineraliza- The short-term (1 yr) grazing, light grazing and medium tion and N ­ 2O emissions (McNaughton et al. 1997) due grazing slightly increased the above- or belowground bio- to the activities of heterotrophs, nitrifiers, and deni- mass (Fig. 4 and S5), but heavy grazing did not (Fig. S6). trifiers (Patra et al. 2005), leading to more soil N losses Compared with ungrazed conditions, light or medium than in the earlier stages. Long-term intensive grazing grazed grasslands, heavy grazing reduced plant cover- reduces the biological nitrogen fixation capacity (Zhang age and density, and species diversity (Fig. S7), and thus et al. 2021b). So, soil C/N ultimately increased follow- decreased plant biomass (Gamoun 2014). ing long-term grazing (Fig. 6). More soil N loss also The short-term grazing effects also confirm the ‘inter- means a lack of nutrients for plant growth, which in mediate disturbance hypothesis’ (Tilman and Downing turn reduces grassland productivity (Hao and He 2019; 1994): medium grazing maintained and even increased Fig. 6 Soil pH and soil C/N ratios depending on grazing intensities in the acidic and alkaline soils, and the C/N ratios depending on grazing duration. UG, Ungrazed; LG, Low grazing; MG, Medium grazed; HG, High grazed. Small letters above the error bars indicate significant differences among the grazing intensities or grazing duration at p < 0.05. Values in parenthesis are the numbers of observations. All the data are presented as Mean ± SE Deng et al. Carbon Research (2023) 2:19 Page 12 of 17 vegetation productivity and diversified the community S6, Zhu et al. 2018). Plant productivity is the main driving structure. However, regardless of the grazing intensity, force for ecosystem C sequestration because soil and eco- all ecosystem C pools increased with short-term grazing system C stocks increase with the above- and belowground (2 ~ 3 yr) to various degrees (Fig. 4 and S5). This could be plant biomass (Fig. 7, Deng et al. 2021). Therefore, optimal attributed to the rapid turnover of nutrients, accelera- grazing intensities must always be considered to maintain tion of plant growth, and the transformation of organic original production rates and ensure that the grassland matter, indicating that grasslands have a compensatory ecosystems remain stable and sequester C (Fig. 4). response to grazing. These findings, as impacted by the grazing duration, also confirm the ‘grazing optimization 4.4 Theory of grazing effects on ecosystem C stocks hypothesis’ (Edelstein-Keshet 1986) because grass growth The structure and functions of grassland ecosystem is stimulated during short-term grazing (< 3 yr), after responses to grazing fit four theories, i.e., the ‘intermedi- which the compensatory effect disappears (Fig. S5). ate disturbance hypothesis’ (Tilman and Downing 1994), The best grazing intensity maintaining the highest eco- the ‘grazing optimization hypothesis’ (Edelstein-Keshet system C stocks for desert and typical steppes is light graz- 1986), the ‘dynamic balance hypothesis’ and the ‘dynamic ing (Fig. 2). Medium grazing is the best grazing intensity imbalance hypothesis’ (Palmer et al. 2016). Based on the for ecosystem C stocks in meadow steppes and alpine results of this synthesis, we summarize these conceptual steppes (Fig. 2). When the grazing intensity exceeds the models to reflect the response of grassland C stocks to tipping-point, the ecosystem C stocks will drop. These grazing intensity and duration (Fig. 8). results also confirm the ‘grazing optimization hypothesis’ The ‘intermediate disturbance hypothesis’ assumes that (Edelstein-Keshet 1986). Before the stocking rate increases moderate disturbance is beneficial to maintain higher to the optimal level, the ecosystem C stocks increase, and community productivity and diversity, and medium graz- then decrease with the increase of grazing intensity. Light ing accelerates vegetation growth, reduces canopy shield- or medium grazing increases the ecosystem C stocks ing to solar radiation, increases forage photosynthetic because plants undergo over-compensatory growth (Zhu capacity, and thus enables compensatory effects on forage et al. 2018). Under heavy grazing, the rate of plant regen- growth (Tilman and Downing 1994). The compensatory eration cannot compensate for the biomass removal rate, effect on forage growth is further developed in the ‘grazing and thus the total net primary productivity drops (Fig. optimization hypothesis’ (Edelstein-Keshet 1986), which Fig. 7 The relationships among plant biomass carbon (C) stocks (n = 286) as well as the relationships between soil C stocks (n = 253) and shoot or root C stocks. Shoot C, aboveground biomass carbon stocks; Root C, belowground biomass carbon stocks; Litter C, litter biomass carbon stocks; Total biomass C, total biomass carbon stocks including both shoot C and root C; Soil C, soil carbon stocks. R.2 is the coefficient of determination. All regression lines are significant at p < 0.001 Deng et al. Carbon Research (2023) 2:19 Page 13 of 17 Fig. 8 Conceptual models of the response of grassland C stocks to grazing intensity from ungrazed through low grazing and medium grazing to high grazing. The conceptual models reflect the response of grassland C stocks to grazing duration elucidated in this synthesis. The ‘Unimodal curve’ in panel A indicates the first increase of grassland C stocks to a maximum, and then the subsequent decrease. Panel A fits the ‘Intermediate disturbance hypothesis’ (IntermD hypothesis) (Tilman and Downing 1994). Panel B fits the ‘dynamic imbalance hypothesis’ (DynamIB hypothesis) (Palmer et al. 2016). The grassland response cycle to grazing intensity can be divided into two stages of C stock changes, that is CompS, compensatory growth stage, and DegradS, degradation stage (A). The green “cross circle” in panel A indicates the optimal grazing intensity, where grazing has a positive effect on grassland C stocks. The model only reflects the general C stock response to grazing intensity and it does not cover all cases (e.g., some grasslands can be grazed with low intensity for a very long time without reaching degradation). Grassland state before grazing is critical to the response of grassland to grazing duration. The ModerD hypothesis fits to non-degraded grasslands before grazing (A). And the DynamIB hypothesis fits to degraded grasslands before grazing (B). It is worth noting that dry steppes (Dry St) have a lower optimal grazing intensity than wet steppes (Wet St) (A), and the grassland C stocks decreased faster in dry steppes than in wet steppes (B). Dry St are steppes located in arid and semi-arid areas, and Wet St are steppes located in humid areas. The grey boxes in panels A and B reflect the grassland ecosystem C components of each of the grassland types that fit the corresponding models. Overgrazing indicates grazing that leads to the degradation of a grassland assumes that the primary production increases first and (Figs. 4 and 8). Soil C stocks increased in desert, typical and then decreases with rising grazing intensity (Deng et al. meadow steppes during the first 3 years of grazing (Fig. 4 2014a). The primary productivity reached the peak when and 8). The compensation effect will gradually disappear the grazing intensity was medium, corresponding to the with the increase in grazing duration, until they reach a optimal stocking rates (McNaughton 1979). The ‘interme- new equilibrium (Fig. 8). The turning point after the com- diate disturbance hypothesis’ and the ‘grazing optimization pensatory growth stage could represent the optimal graz- hypothesis’ can have the same outcome when grasslands ing duration (Fig. 8). Therefore, the compensatory effect is do not suffer from extreme grazing. In our meta-analysis, related to grazing duration and intensity. the plant biomass and soil C stocks in the desert steppes The ‘dynamic balance hypothesis’ refers to the stable confirm the ‘intermediate disturbance hypothesis’ respond- state achieved by interactions between organisms and ing to grazing intensity (Fig. 8), indicating that appropri- environment within a certain period, including the stabil- ate grazing intensity of less than 2 sheep h ­ a−1 is better ity of ecosystem structure, functions and energy (Palmer for increasing ecosystem C stocks (Fig. 8) in the desert et al. 2016). The shoot C stocks in the typical steppes steppes. and root C stocks in alpine steppes under light grazing The belowground biomass C stocks in the typical steppes remained independent of grazing duration (Figs. 4 and 8). increased after light grazing (Fig. S6), which confirms the They confirmed the ‘dynamic balance hypothesis’ respond- ‘grazing optimization hypothesis’ (Edelstein-Keshet 1986). ing to grazing. The dynamic balance of the C stocks is The specific performance of grazing optimization is the determined by grazing intensities. For example, the root compensatory growth of plants (Lu et al. 2017). Compensa- C stocks in the meadow steppes was unchanged at or tory growth is the result of coadaptation and coevolution below medium grazing but reduced significantly under among grasslands, livestock and environment respond- heavy grazing (Fig. 2). So, heavy grazing disturbs the bal- ing to disturbance, occurring at various organizational anced C budget. Soil C stocks in the alpine steppes (Figs. 2 levels of plant organs, individual plants, populations and and 3) decreased with grazing (Fig. 8), indicating that communities (McNaughton 1979). Along with increas- grazing imbalanced C inputs and outputs. It confirmed ing grazing duration, the belowground biomass C stocks the ‘dynamic imbalance hypothesis’ (Palmer et al. 2016). of the typical steppes linearly increase under light grazing, Therefore, plant and soil C stocks’ responses to grazing but a “unimodal” trend is common under medium grazing Deng et al. Carbon Research (2023) 2:19 Page 14 of 17 intensity are related to the grassland type and grazing (Fig. 5). So, reasonable grazing may reduce the sen- duration. sitivity of grasslands to climate change. 4.5 Implications for grazing management in grasslands Grazing is the primary use for natural grassland eco- 5 Conclusions systems (Yan et al. 2013; Zhou et al. 2014; Dlamini et al. The ‘intermediate disturbance hypothesis’ and the ‘grazing 2016). The C stocks in grassland ecosystems of China optimization hypothesis’ are most relevant for the early stages could increase by up to 0.21 Pg ­yr−1 (i.e., 1.7 Mg ­ha−1 ­yr−1) of light grazing, such as in desert steppes. The ‘dynamic bal- if the comprehensive ‘Returning Grazing Land to Grass- ance hypothesis’ is most relevant for light grazing conditions land’ Project is implemented (Deng et al. 2017). However, such as in typical steppes and alpine steppes and the over- a total cessation of grazing is in conflict with the policy degradation stage due to overgrazing. The ‘dynamic imbal- goal of increasing food production, so grazing exclusion is ance hypothesis’ is most relevant for the intermediate stage therefore not feasible. Based on this meta-analysis, there of light grazing and medium grazing, or the stages prior to are four options for grazing land management strategies: degradation under heavy grazing. However, it should also be noted that the grassland state before grazing is critical for the (1) Reasonable grazing intensity is conducive to main- response of the grassland to the grazing duration. If the grass- taining grassland ecosystem C pools. The appro- land is not degraded before grazing, it fits the ‘intermediate priate intensity of grazing utilization, whether light disturbance hypothesis’; if the grassland is in different degra- grazing or medium grazing, has positive effects dation states, it is more in line with the ‘dynamic imbalance on ecosystem C stocks (Fig. 2). The best stocking hypothesis’. The model of grassland response to grazing is reg- rates in the desert steppes, typical steppes, meadow ulated by the grassland’s type, grazing intensity, and grazing steppes and alpine steppes are about 1 ~ 2, 3 ~ 4, duration. In conclusion, grassland grazing is a double-edged 5 ~ 6 and 7 ~ 8 sheep units per hectare, respectively sword. On the one hand, proper management can maintain (Fig. S5). At these grazing intensities, livestock will and even increase the grassland C stock and the harvest live- not decrease C stocks, and they may even increase C stock products. On the other hand, grassland degradation can stocks of grassland ecosystems. be accelerated by human induced overgrazing, resulting in (2) Rotational grazing cycles are conducive to the healthy serious carbon losses. So, more appropriate and customized development of grasslands. All the components of management strategies should be considered whenever pos- ecosystem C in the short-term grazing (< 3 yr) do sible, such as light or medium grazing, rotational grazing or not indicate degradation, regardless of the grazing non-growing season grazing, to promote C sequestration on intensity (Fig. S4). Although grazing exclusion has no grasslands and climate change mitigation. effects on shoot biomass after 3 years, it can increase shoot biomass to some extent in the first 3 years Supplementary Information (i.e., synthesis of grazing exclusion effects; Deng et al. The online version contains supplementary material available at https://​doi.​ org/​10.​1007/​s44246-​023-​00051-7. 2017). Accordingly, every 3 years (i.e., 3 years of graz- ing followed by 3 years of rest) may be the optimal Additional file 1: TableS1. Pearson correlations of ecosystem carbon (C) grazing rotation cycle to maintain stable pastures. stocks with climate (MAP, mean annual precipitation; MAT, mean annual (3) Grazing in the non-growing season helps to main- temperature), elevation and grassland utilization rate, grazing duration, root/shoot ratio (R/S) and plant community characteristics in the grass- tain C sequestration. Non-growing season grazing lands with four grazing intensities. TableS2. Multivariable linear regression strongly increased ecosystem C stocks due to the model analysis between ecosystem carbon (C) stocks and five driving higher productivity in alpine steppes, compared with factors: MAP (x1), MAT (x2), elevation (x3), stocking rate (x4) andgrazing duration (x5) in four grassland types under grazing. FigureS1. Stock- grazing during the growing season or the whole year. ingrates and grassland utilization rates (see below) corresponding to However, whether non-growing season grazing is three grazing intensities in the four grassland types of the dataset, as beneficial for improving grassland C stocks in other well as the relationships between stocking rate and grassland utilization rates. FigureS2. Root/shoot ratios (R/S) depending on grazing intensi- grassland types requires further study. ties in the four grassland types of typical, desert, meadow and alpine (4) Reasonable grazing improves the ability of eco- steppes. FigureS3. Empirical functions to estimate the missing soil bulk systems to cope with climate change. Grazing can density based ondata from studies reporting organic carbon content and bulk density in the ungrazed (n=133) and grazed (n=333) sites. Figure. change the correlations between plants C stocks S4 Plant biomass, litter and soil carbon (C) stocks depending on stocking and climate conditions (Table S1). Plant C stocks rates in the four grassland types. FigureS5. Plant biomass, litter and soil remained unchanged with climate change under carbon (C) stocks depending on grazing duration underthe three grazing intensities of light, moderate and heavy grazing. Figure S6. Grassland all grazing intensities (Fig. 5), and soil C stocks carbon stocks change with grazing duration under the heavy grazing. Fig- responded slower to climate change under graz- ureS7. Plant community height and coverage change with grazing inten- ing than when they were under ungrazed condition sities in the four grassland types in China. FigureS8. There lationships Deng et al. Carbon Research (2023) 2:19 Page 15 of 17 Altesor A, Oesterheld M, Leoni E, Lezama F, Rodriguez C (2005) Effect of graz- of root/shoot ratios (R/S) with stocking rates (A) and grazing duration ing on community structure and productivity of a Uruguayan grassland. (B). FigureS9. Grassland carbon (C) stocks depending on grazing periods Plant Ecol 179:83–91. https://​doi.​org/​10.​1007/​s11258-​004-​5800-5 of alpine steppes. Bai Y, Cotrufo MF (2022) Grassland soil carbon sequestration: current under- Additional file 2: Appendix Dataset S1. standing, challenges, and solutions. Science 377(6606):603–608. https://​ doi.​org/​10.​1126/​scien​ce.​abo23​80 Bai Y, Wu J, Clark CM, Pan Q, Zhang L, Chen S, Wang Q, Han X (2012) Grazing Acknowledgements alters ecosystem functioning and C:N: P stoichiometry of grasslands This study was supported by the National Natural Science Foundation of along a regional precipitation gradient. J Appl Ecol 49:1204–1215. https://​ China (42277471), the Second Tibetan Plateau Scientific Expedition and doi.​org/​10.​1111/j.​1365-​2664.​2012.​02205.x Research Program of the Chinese Academy of Sciences (2019QZKK0603), Bardgett RD, Bullock JM, Lavorel SL et al (2021) Combatting global grassland the Strategic Priority Research Program of the Chinese Academy of Sciences degradation. Nat Rev Earth Envi 2:720–735. https://​doi.​org/​10.​1038/​ (XDA23070201), the Scientific and Technological Innovation Project of Shaanxi s43017-​021-​00207-2 Forestry Academy of Sciences (SXLK2021-0206), the RUDN University Strategic Booker K, Huntsinger L, Bartolome JW, Sayre NF, Stewart W (2013) What can Academic Leadership Program, and the Tyumen Oblast Government project ecological science tell us about opportunities for carbon sequestration No. 89-DON on arid rangelands in the United States? Glob Envir Chang 23:240–251. https://​doi.​org/​10.​1016/j.​gloen​vcha.​2012.​10.​001 Authors’ contributions Chen J, Zhou X, Wang J et al (2016) Grazing exclusion reduced soil respiration L.D., Z.S. and Y.K. designed the research, L.D., J.L., and J.W. collected the data, but increased its temperature sensitivity in a meadow grassland on the L.D. performed the research, K.W., J.L., Z.T., W.Y. and F.Z. contributed new Tibetan Plateau. Ecol Evol 6(3):1–13. https://​doi.​org/​10.​1002/​ece3.​1867 analytical tools, L.D. and X.W. analysed the data, and L.D., Z.S., S.B., A.S., C.H., Conant RT, Paustian K, Elliott ET (2001) Grassland management and conversion S.A., X.X. and Y.K. wrote the paper. The author(s) read and approved the final into grassland effects on soil carbon. Ecol Appl 11:343–355. https://​doi.​ manuscript. org/​10.​1890/​1051-​0761(2001)​011[0343:​GMACIG]​2.0.​CO;2 Deng L, Shangguan ZP (2021) High quality developmental approach for soil and Availability of data and materials water conservation and ecological protection on the Loess Plateau. Front All research data of the current study are accessible from the corresponding Agr Sci Eng 8(4):501–511. https://​doi.​org/​10.​15302/J-​FASE-​20214​25 author on reasonable request. Deng L, Sweeney S, Shangguan ZP (2014a) Grassland responses to grazing disturbance: Plant diversity changes with grazing intensity in a desert steppe. Grass for Sci 69:524–533. https://​doi.​org/​10.​1111/​gfs.​12065 Declarations Deng L, Liu GB, Shangguan ZP (2014b) Land use conversion and changing soil carbon stocks in China’s ‘Grain-for-Green’ Program: a synthesis. Glob Competing interests Chang Biol 20:3544–3556. https://​doi.​org/​10.​1111/​gcb.​12508 All the authors declare no competing interests. Deng L, Shangguan ZP, Wu GL, Chang XF (2017) Effects of grazing exclusion on carbon sequestration in China’s grassland. Earth-Sci Rev 173:84–95. Author details https://​doi.​org/​10.​1016/j.​earsc​irev.​2017.​08.​008 1 State Key Laboratory of Soil Erosion and Dryland Farming On the Loess Pla- Deng L, Peng CH, Li JW, Liu YL, Hai XY, Liu QY, Huang CB, Shangguan ZP, Kuzya- teau, Institute of Soil and Water Conservation, Northwest A&F University, No. kov Y (2021) Drought effects on soil carbon and nitrogen dynamics in 26 Xinong Road, Yangling 712100, Shaanxi, China. 2 Institute of Soil and Water natural ecosystems. Earth Sci Rev 214:103501. https://​doi.​org/​10.​1016/j.​ Conservation, Chinese Academy of Sciences and Ministry of Water Resources, earsc​irev.​2020.​103501 Yangling 712100, Shaanxi, China. 3 Institute of Environmental Science Derner JD, Boutton TW, Briske DD (2006) Grazing and ecosystem carbon stor- and Technology (ICTA), Universitat Autònoma de Barcelona (UAB), Bellaterra, age in the North American great plains. Plant Soil 280(1/2):77–90. https://​ Spain. 4 Research Institute of Ecology and Natural Resources Management, doi.​org/​10.​1007/​s11104-​005-​2554-3 Tyumen State University, 6 Volodarskogo Street, 625003 Tyumen, Russia. Dlamini P, Chivenge P, Chaplot V (2016) Overgrazing decreases soil organic 5 Center of CEF/ESCER, Department of Biological Science, University of Quebec carbon stocks the most under dry climates and low soil pH: a meta- at Montreal, Montreal H3C 3P8, Canada. 6 State Key Laboratory of Urban analysis shows. Agr Ecosyst Envir 221:258–269. https://​doi.​org/​10.​1016/j.​ and Regional Ecology, Research Center for Eco‑Environmental Sciences, agee.​2016.​01.​026 Chinese Academy of Sciences, Beijing 100085, China. 7 Institute of Geographic Dong L, Martinsen V, Wu YT, Zheng Y, Liang CZ, Liu ZL, Mulder J (2021) Effect of Sciences and Natural Resources Research, Chinese Academy of Sciences, grazing exclusion and rotational grazing on labile soil organic carbon in Beijing 100101, China. 8 State Key Laboratory of Loess and Quaternary Geology, north China. Eur Soil Sci 72:372–384. https://​doi.​org/​10.​1111/​ejss.​12952 Institute of Earth Environment, Chinese Academy of Sciences, Xi’an 710075, Edelstein-Keshet L (1986) Mathematical theory for plant-herbivore systems. J Shaanxi, China. 9 School of Agriculture, Ningxia University, Yinchuan 750021, Math Biol 24:25–58. https://​doi.​org/​10.​1007/​BF002​75719 Ningxia, China. 10 Key Laboratory of Grassland Ecosystem of the Ministry Eldridge DJ, Poore AGB, Ruiz-Colmenero M, Letnic M, Soliveres S (2016) Eco- of Education, College of Grassland Science, Gansu Agricultural University, system structure, function, and composition in rangelands are negatively Lanzhou 730070, Gansu, China. 11 Department of Soil Science of Temperate affected by livestock grazing. Ecol Appl 26:1273–1283. https://​doi.​org/​10.​ Ecosystems, Department of Agricultural Soil Science, University of Göettingen, 1890/​15-​1234 Göttingen, Germany. 12 Peoples Friendship University of Russia (RUDN Univer- Eldridge DJ, Delgado-Baquerizo M, Travers SK, Val J, Oliver I (2017) Do grazing sity), 117198 Moscow, Russia. intensity and herbivore type affect soil health? Insights from a semi-arid productivity gradient. J Appl Ecol 54:976–985. https://​doi.​org/​10.​1111/​ Received: 6 December 2022 Revised: 2 April 2023 Accepted: 10 April 1365-​2664.​12834 2023 Elmore AJ, Asner GP (2006) Effects of grazing intensity on soil carbon stocks following deforestation of a hawaiian dry tropical forest. Glob Chang Biol 12(9):1761–1772. https://​doi.​org/​10.​1111/j.​1365-​2486.​2006.​01198.x Evju M, Mysterud A (2010) Interaction between local climate and grazing determine the population dynamics of the small herb Viola biflora. Oeco- References logia 163(4):921. https://​doi.​org/​10.​1007/​s00442-​010-​1637-x Abdalla M, Hastings A, Chadwick DR et al (2018) Critical review of the impacts Ferraro DO, Oesterheld M (2002) Effect of defoliation on grass growth: a quan- of grazing intensity on soil organic carbon storage and other soil titative review. Oikos 98(1):125–133. https://​doi.​org/​10.​1034/j.​1600-​0706.​ quality indicators in extensively managed grasslands. Agr Ecosyst Envir 2002.​980113.x 253:62–81. https://​doi.​org/​10.​1016/j.​agee.​2017.​10.​023 Gamoun M (2014) Grazing intensity effects on the vegetation in desert range- lands of Southern Tunisia. J Arid Land 6:324–333. https://​doi.​org/​10.​1007/​ s40333-​013-​0202-y Deng et al. Carbon Research (2023) 2:19 Page 16 of 17 Gao XL, Li X, Zhao L, Kuzyakov Y (2021) Shrubs magnify soil phosphorus deple- Milcu A, Eugster W, Bachmann D et al (2016) Plant functional diversity tion in Tibetan meadows: conclusions from C:N: P stoichiometry and increases grassland productivity-related water vapor fluxes: an ecotron deep soil profiles. Sci Total Envir 785:147320. https://​doi.​org/​10.​1016/j.​ and modeling approach. Ecology 97:2044–2054. https://​doi.​org/​10.​1890/​ scito​tenv.​2021.​147320 15-​1110.1 Gao YH (2007) Study on carbon and nitrogen distribution pattern and cycling Moinet GYK, Cieraad E, Hunt JE, Fraser A, Turnbull MH, Whitehead D (2016) Soil process in an alpine meadow ecosystem under different grazing inten- heterotrophic respiration is insensitive to changes in soil water content sity. Chinese Academy of Science, Beijing but related to microbial access to organic matter. Geoderma 274:68–78. Hafner S, Unteregelsbacher S, Seeber E, Xu X, Li X, Guggenberger G, Miehe G, https://​doi.​org/​10.​1016/j.​geode​rma.​2016.​03.​027 Kuzyakov Y (2012) Effect of grazing on carbon stocks and assimilate par- Mou XM, Li XG, Zhao NN, Yu YW, Kuzyakov Y (2018) Tibetan sedges sequester titioning in Tibetan montane pasture revealed by 13CO2 pulse labeling. more carbon belowground than grasses: a 13C labeling study. Plant Soil Glob Chang Biol 18:528–538. https://​doi.​org/​10.​1111/j.​1365-​2486.​2011.​ 426:287298. https://​doi.​org/​10.​1007/​s11104-​018-​3634-5 02557.x Palmer MA, Zedler JB, Falk DA (2016) Ecological Theory and Restoration Ecol- Han GD, Hao XY, Zhao ML et al (2008) Effect of grazing intensity on carbon and ogy. In: Palmer MA, Zedler JB, Falk DA (eds) Foundations of Restoration nitrogen in soil and vegetation in a meadow steppe in Inner Mongolia. Ecology. Island Press, Washington, DC Agr Ecosyst Envir 125:21–32. https://​doi.​org/​10.​1016/j.​agee.​2007.​11.​009 Patra AK, Abbadie L, Clays-Josserand A et al (2005) Effects of grazing on Hao Y, He Z (2019) Effects of grazing patterns on grassland biomass and microbial –functional groups involved in soil N dynamics. Ecol Monogr soil environments in China: a meta-analysis. PLoS One 14(4):e0215223. 75:65–80. https://​doi.​org/​10.​1890/​03-​0837 https://​doi.​org/​10.​1371/​journ​al.​pone.​02152​23 Patty L, Halloy SRP, Hiltbrunner E et al (2010) Biomass allocation in herba- Hassan N, Sher K, Rab A, Abdullah I, Zeb U, Naeem I, Shuaib M, Khan H, Kham ceous plants under grazing impact in the high simi-arid Andes. Flora W, Khan A (2021) Effects and mechanism of plant litter on grassland 205:695–703. https://​doi.​org/​10.​1016/j.​flora.​2009.​12.​039 ecosystem: a review. Acta Ecol Sin 41:341–345. https://​doi.​org/​10.​1016/j.​ Pineiro G, Paruelo JM, Jobbagy EG, Jackson RB, Oesterheld M (2009) Grazing chnaes.​2021.​02.​006 effects on belowground C and N stocks along a network of cattle exclo- Jiang ZY, Hu ZM, Lai DYF, Han DR, Wang M, Liu M, Zhang M, Guo MY (2020) sures in temperate and subtropical grasslands of South America. Glob Light grazing facilitates carbon accumulation in subsoil in Chinese Biogeoch Cycl 23:GB2003. https://​doi.​org/​10.​1029/​2007G​B0031​68 grassland: a meta-analysis. Glob Chang Biol 26(12):7186–7197. https://​doi.​ Pucheta E, Bonamici I, Cabido M, Díaz S (2004) Belowground biomass and org/​10.​1111/​gcb.​15326 productivity of a grazed site and a neighbouring ungrazed exclosure in a Jobbágy E, Jackson R (2000) The vertical distribution of soil organic carbon grassland in central Argentina. Austr Ecol 29:201–208. https://​doi.​org/​10.​ and its relation to climate and vegetation. Ecol Appl 10:423–436. https://​ 1111/j.​1442-​9993.​2004.​01337.x doi.​org/​10.​1890/​1051-​0761(2000)​010[0423:​T VDOSO]​2.0.​CO;2 Qu Q, Zhang J, Hai XY, Wu JZ, Fan JW, Wang DF, Li JW, Shangguan ZP, Deng L Kang L, Han X, Zhang Z, Sun OJ (2007) Grassland ecosystems in China: review (2022) Long-term fencing alters the vertical distribution of soil 13C and of current knowledge and research advancement. Phil Trans Roy Soc B SOC turnover rate: revealed by MBC-13C. Agr Ecosyst Envir 339:108119. Biol Sci 362:997–1008. https://​doi.​org/​10.​1098/​rstb.​2007.​2029 https://​doi.​org/​10.​1016/j.​agee.​2022.​108119 Kuzyakov Y, Biryukova OV, Kuznetzova TV, Mölter K, Kandeler E, Stahr K (2002) Raza S, Na M, Wang P, Ju X, Chen Z, Zhou J, Kuzyakov Y (2020) Dramatic loss of Carbon partitioning in plant and soil, carbon dioxide fluxes and enzyme inorganic carbon by nitrogen-induced soil acidification: the past, present activities as affected by cutting ryegrass. Biol Fert Soils 35(5):348–358. and future of Chinese croplands. Glob Chang Biol 26(6):3738–3751. https://​doi.​org/​10.​1007/​s00374-​002-​0480-6 https://​doi.​org/​10.​1111/​gcb.​15101 Li DJ, Niu SL, Luo YQ (2012) Global patterns of the dynamics of soil carbon Sun DS, Wesche K, Chen DD et al (2011) Grazing depresses soil carbon stor- and nitrogen stocks following afforestation: a meta-analysis. New Phytol age through changing plant biomass and composition in a Tibetan 195:172–181. https://​doi.​org/​10.​1111/j.​1469-​8137.​2012.​04150.x alpine meadow. Plant Soil Envir 57:271–278. https://​doi.​org/​10.​17221/7/​ Liu N, Kan HM, Yang GW et al (2015) Changes in plant, soil, and microbes in a 2011-​PSE typical steppe from simulated grazing: explaining potential change in soil Sun J, Zhou TC, Liu M et al (2019) Water and heat availability are drivers of the C. Ecol Monogr 85:269–286. https://​doi.​org/​10.​1890/​14-​1368.1 aboveground plant carbon accumulation rate in alpine grasslands on the Liu S, Zamanian K, Schleuss P, Zarebanadkouki M, Kuzyakov Y (2018) Degrada- Tibetan Plateau. Glob Ecol Biogeogr 29(1):50–64. https://​doi.​org/​10.​1111/​ tion of Tibetan grasslands: consequences for carbon and nutrient cycles. geb.​13006 Agr Ecosyst Envir 252:93–104. https://​doi.​org/​10.​1016/j.​agee.​2017.​10.​011 Tilman D, Downing JA (1994) Biodiversity and stability in grasslands. Nature Liu S, Wang J, Pu S, Blagodatskaya E, Kuzyakov Y, Razavi B (2020) Impact of 367(6461):363–365. https://​doi.​org/​10.​1038/​36736​3a0 manure on soil biochemical properties: a global synthesis. Sci Total Envir Unteregelsbacher S, Hafner S, Guggenberger G, Miehe G, Xu X, Liu J, Kuzyakov 745:141003. https://​doi.​org/​10.​1016/j.​scito​tenv.​2020.​141003 Y (2012) Response of long-, medium- and short-term turnover processes Lu XY, Kelsey KC, Yan Y, Sun J, Wang XD, Cheng GW, Neff JC (2017) Effects of of the carbon budget to overgrazing on the Tibetan Plateau. Biogeo- grazing on ecosystem structure and function of alpine grasslands in chemistry 111:187–201. https://​doi.​org/​10.​1007/​s10533-​011-​9632-9 Qinghai-Tibetan Plateau: a synthesis. Ecosphere 8:e01656. https://​doi.​org/​ Vallentine JF (1990) Grazing Management. Academic Press, California, USA 10.​1002/​ecs2.​1656 Vandandorj S, Eldridge DJ, Travers SK, Delgado-Baquerizo M (2017) Contrasting Luo YQ, Field CB, Jackson RB (2006) Does nitrogen constrain carbon cycling, or effects of aridity and grazing intensity on multiple ecosystem functions does carbon input stimulate nitrogen cycling? Ecology 87:3–4. https://​ and services in Australian woodlands. Land Degrad Devel 28(7):2098– doi.​org/​10.​1890/​05-​0923 2108. https://​doi.​org/​10.​1002/​ldr.​2736 McNaughton SJ (1979) Grazing as an optimization process: grass ungulate Wang XY, Yi SH, Wu QB, Yang K, Ding YJ (2016) The role of permafrost and soil relationships in the Serengeti. Am Nat 113:691–703. https://​doi.​org/​10.​ water in distribution of alpine grassland and its NDVI dynamics on the 1086/​283426 Qinghai-Tibetan Plateau. Glob Planet Chang 147:40–53. https://​doi.​org/​ McNaughton SJ, Banyikwa FF, McNaughton MM (1997) Promotion of the 10.​1016/j.​glopl​acha.​2016.​10.​014 cycling of diet-enhancing nutrients by African grazer. Science 278:1798– Wilson CH, Strickland MS, Hutchings JA, Bianchi TS, Flory SL (2018) Grazing 1800. https://​doi.​org/​10.​1126/​scien​ce.​278.​5344.​1798 enhances belowground carbon allocation, microbial biomass, and soil McSherry ME, Ritchie ME (2013) Effects of grazing on grassland soil carbon: a carbon in a subtropical grassland. Glob Chang Biol 24:2997–3009. https://​ global review. Glob Chang Biol 19:1347–1357. https://​doi.​org/​10.​1111/​ doi.​org/​10.​1111/​gcb.​14070 gcb.​12144 Yan L, Zhou G, Zhang F (2013) Effects of different grazing intensities on Milchunas DG, Laurenroth WK (1993) Quantitative effects of grazing on grassland production in China: a meta-analysis. PLoS One 8(12):e81466. vegetation and soils over a global range of environments. Ecol Monogr https://​doi.​org/​10.​1371/​journ​al.​pone.​00814​66 63(4):327–366. https://​doi.​org/​10.​6084/​m9.​figsh​are.c.​33089​07 Yu ZJ, Wang KB, Li JW, Shangguan ZP, Deng L (2023) Mixed plantations have Milchunas DG, Laurenroth WK, Burke IC (1998) Livestock grazing: animal and more soil carbon sequestration benefits than pure plantations in China. plant biodiversity of shortgrass steppe and the relationship to ecosystem For Ecol Manage 529:120654. https://​doi.​org/​10.​1016/j.​foreco.​2022.​ function. Oikos 83:65–74. https://​doi.​org/​10.​2307/​35465​47 120654 Deng et al. Carbon Research (2023) 2:19 Page 17 of 17 Zamanian K, Zarebanadkouki M, Kuzyakov Y (2018) Nitrogen fertilization raises ­CO2 efflux from soil inorganic carbon: a global assessment. Glob Chang Biol 24:2810–2817. https://​doi.​org/​10.​1111/​gcb.​14148 Zhang J, Zuo X, Zhou X, Lv P, Lian J, Yue X (2017) Long-term grazing effects on vegetation characteristics and soil properties in a semiarid grassland, northern China. Envir Monit Assess 189:216. https://​doi.​org/​10.​1007/​ s10661-​017-​5947-x Zhang RY, Tian DS, Chen HYH, Seabloom EW, Han GD, Wang SP, Yu GR, Li ZL, Niu SL (2021a) Biodiversity alleviates the decrease of grassland multi- functionality under grazing disturbance: a global meta-analysis. Glob Eco Biogeogr 31:155–167. https://​doi.​org/​10.​1111/​geb.​13408 Zhang L, Wang XT, Wang J, Wan Q, Liao LR, Liu GB, Zhang C (2021b) Grazing exclusion reduces soil N ­ 2O emissions by regulating nirK- and nosZ-type denitrifiers in alpine meadows. J Soils Sed 21:3753–3769. https://​doi.​org/​ 10.​1007/​s11368-​021-​03035-5 Zhao BB, Niu KC, Du GZ (2009) The effect of grazing on above-ground biomass allocation of 27 plants species in an alpine meadow plant community in Qinghai-Tibetan Plateau. Acta Ecol Sin 29:1596–1606. In Chinese. Zhao FZ, Ren CJ, Shelton S, Wang ZT, Pang GW, Chen J, Wang J (2017) Grazing intensity influence soil microbial communities and their implications for soil respiration. Agr Ecosyst Envir 249:50–56. https://​doi.​org/​10.​1016/j.​ agee.​2017.​08.​007 Zheng SX, Ren HY, Lan ZC, Li WH, Wang KB, Bai YF (2010) Effects of grazing on leaf traits and ecosystem functioning in Inner Mongolia grasslands: scal- ing from species to community. Biogeosciences 7:1117–1132. https://​doi.​ org/​10.​5194/​bg-7-​1117-​2010 Zheng ZZ, Li LF, Makhalanyane TP et al (2021) The composition of antibiotic resistance genes is not affected by grazing but is determined by microor- ganisms in grassland soils. Sci Total Envir 761:143205. https://​doi.​org/​10.​ 1016/j.​scito​tenv.​2020.​143205 Zhou W, Gang CC, Zhou L, Chen Y, Li JL, Ju WM, Odeh IO (2014) Dynamic of grassland vegetation degradation and its quantitative assessment in the northwest China. Acta Oecologica 55:86–96. https://​doi.​org/​10.​1016/j.​ actao.​2013.​12.​006 Zhu GY, Deng L, Zhang XB, Shangguan ZP (2016) Effects of grazing exclusion on the plant community and soil physicochemical properties in a desert steppe on the Loess Plateau, China. Ecol Engin 90:372–381. https://​doi.​ org/​10.​1016/j.​ecole​ng.​2016.​02.​001 Zhu GY, Tang ZS, Shangguan ZP, Deng L (2018) Overgrazing depresses soil carbon stock through changing plant diversity in temperate grassland of the Loess Plateau. Plant Soil Envir 64:1–6. https://​doi.​org/​10.​17221/​610/​ 2017-​PSE Zuo XA, Zhang J, Lv P et al (2018) Effects of plant functional diversity induced by grazing and soil properties on above- and belowground biomass in a semiarid grassland. Ecol Ind 93:555–561. https://​doi.​org/​10.​1016/j.​ecoli​ nd.​2018.​05.​032 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations.