Grassland covers a large area of the planet, accounting for 24% of the global vegetative land area (Ren et al., 2008). In China, the total area is approximately 400 million hectare, contributing more than 40% of the total land area (Zhao and Li, 2018). Despite their importance, grasslands face a high threat, as they have been rapidly declining in the past decades due to biodiversity loss (Carbutt et al., 2017; Wang et al., 2020a). Both the dynamics and diversity of grassland ecosystems are easily affected by natural and artificial disturbances such as fire, grazing, and mowing (Loram-Lourenco et al., 2020; Vermeire et al., 2020; Slette et al., 2021; Qin et al., 2022). These disturbances alter the plant community structure and ecosystem function of grasslands as they are managed at inappropriate scales, intensities, and frequency levels, thereby leading to their degradation (Lu et al., 2012; Zhang et al., 2020). In the past decade, prescribed burning and mowing have become important tools for managing and restoring grassland ecosystems in China (Lu et al., 2012; Zhu et al., 2021), even though they significantly impact grasslands’ soil properties, plant growth, and nutrient cycles (Dowhower et al., 2021), further disturbing various ecological processes, such as herbivory and litter decomposition (Lu et al., 2012; Bai and Cotrufo, 2022). Few studies have investigated the effects of fire and mowing on the microbial community structure in grassland ecosystems, despite the fact that microorganisms play key roles in regulating the nutrient cycle and carbon fixation (Zhao et al., 2020; Bai and Cotrufo, 2022).

Prescribed burning is a major strategy used worldwide to manage grassland ecosystems, as it directly drives the use of natural resources, such as water, light, and nutrients, by native plants (Fidelis et al., 2012; Gordijn and O'Connor, 2021). Plant physiological processes that require water, energy, and nutrients are altered by prescribed burning (Lu et al., 2012; Boughton et al., 2013), and plant nutrition is a key feature affected by the fire in grasslands (Fultz et al., 2016). Reinhart et al. (2016) reported that nutrient concentrations (N and P) were much higher in burned soils than in the control soils. Variations in the nutrient cycle in response to prescribed burning may result in changes in the primary productivity of grasslands (Fidelis et al., 2012). Furthermore, previous studies have focused on the effect of fire on plant diversity and soil nutrients, whereas its effects on microbial community structure in grasslands are still not well investigated (Lu et al., 2012; Wragg et al., 2018; Vermeire et al., 2020). Soil microorganisms have greater potential to affect nutrient mobility. They may also compete with plants for the same limiting nutrients in grasslands (Bing et al., 2016; Yang et al., 2020a; Bonanomi et al., 2022). Predicting the complex response of biodiversity to fire disturbance are also dependent on the microbial and functional groups in soils (Farnsworth et al., 2014). Thus, a better understanding of the relationships among plant growth, soil nutrients, and microbial communities under the prescribed burning treatment is needed to provide a context for the sustainable development of grassland ecosystems (Gordijn and O'Connor, 2021).

Mowing, the mechanical removal of vegetation, is considered an effective management practice in grasslands, which has characteristics similar to those of prescribed burning (Vermeire et al., 2020; Török et al., 2021). Compared with fire, mowing requires less planning and presents low ecological risk (Fidelis et al., 2012). Another benefit is that mowing increases plant diversity in grassland ecosystems by promoting aboveground conditions, besides preserving the high conservation value of grassland ecosystems (Fidelis et al., 2012; Vermeire et al., 2020). Therefore, mowing may be an attractive alternative to prescribed burning for vegetation management. However, plant productivity and soil nutrients are key factors that potentially limit the response of ecosystems to mowing (Chen et al., 2021). Yang et al. (2020b) found that mowing had no significant influence on soil nutrient concentrations; however, plant N and P content were increased by mowing. Soil microorganisms can participate in the biogeochemical cycles of various nutrients (N and P) in grassland ecosystems (Bing et al., 2016; Wang et al., 2020b; Wilcox et al., 2022). However, the potential influence of mowing on the microbial community in grasslands is not well understood.

Miscanthus, a perennial herb widely distributed in western China, is the main forage in summer grasslands and forest steppes. It plays a crucial role in sustaining the ecosystem functions of grasslands. The goal of this study was to evaluate the effects of prescribed burning and mowing on plant growth, soil properties, and microbial communities in grassland ecosystems. The specific aims were to 1) assess their effects on plant growth and soil properties, 2) further indicate their effects on microbial communities (bacteria and fungi), and 3) analyze the internal relations of plant biomass, soil properties, and microbial community.

Our experimental results positively highlight that both mowing and fire significantly (p < 0.01) reduced Miscanthus growth and dry material weight. This finding is consistent with results of previous studies on induced fire (Toma et al., 2010) and mowing (Clifton-Brown et al., 2017; Magenau et al., 2021). However, this effect was larger in the deeply mowed plots than in the burned plots. In particular, deep mowing induced a 5.2-fold decrease in dry plant biomass relative to the control, which is attributed to the significant alteration of plant growth traits, soil nutrient availability, and the community and structure of soil microorganisms.

First, after harvest, the tiller number per cluster in SH5 was the lowest among all treatments, as it showed a 2.2-fold decrease relative to that of the control (125 vs. 265); however, fire only induced a 1.1-fold decrease in the tiller number per cluster relative to that of the control (231 vs. 265). The decrease in tiller number in Miscanthus growth generally induced a decrease in aboveground dry weight, which is indicated by the significantly positive relationship between the dry plant weight per cluster and the tiller number per cluster ( Figure 6 ). This finding is consistent with those of previous studies (Lee et al., 2017; Malinská et al., 2020). In addition, plant height in the deep-mowed plot was the lowest among all treatment plots. This could be directly affected by soil nutrient availability (Cadoux et al., 2012; Ivanyshyn et al., 2018) and will be discussed later.

Second, during Miscanthus growth, the soil nutrient availability (K and P) was much lower in both the mowed and burned plots than in the control, and this effect was much larger in the deep-mowed plot. This can be attributed to the removal of the aboveground straw by mowing, which directly leads to the loss of nutrients over their decomposition into soils (Kahle et al., 2001), similar to grasslands (Bakker et al., 2002; Hejcman et al., 2010; Magenau et al., 2021) and croplands (Buyanovsky and Wagner, 1986; DuPont et al., 2010; Cade-Menun et al., 2013). Fire also induces nutrient loss, as the burned soil is easily washed away by rain (Andreu et al., 1996; Thomas et al., 1999; Gimeno-García et al., 2000). As documented in many studies (Cadoux et al., 2012), soil nutrient availability is generally considered a key factor affecting Miscanthus growth (Himken et al., 1997). For example, in an alkaline grassland soil, both P and K deficiencies generally induced a decrease in Miscanthus (Roncucci et al., 2015; He et al., 2020). In particular, both mowing and fire during Miscanthus growth significantly induced a decrease in soil N availability in the 20–40 cm layer ( Figure 2 ), which is the area of plant root activity for N uptake. Indeed, the role of N limitation as a factor determining successional processes in grass biomass production is significant, as the most striking result of previous studies to date is the significant response of plant growth to N fertilization (Van der Woude et al., 1994; Craine and Jackson, 2010; Toma et al., 2010; Whitaker et al., 2015; Zhao et al., 2016; Lee et al., 2017). If sustained, our current experimental results indicate that nutrient limitation (N, P, and K) is an important factor determining the rates of Miscanthus accumulation in repeated fires and mowed grasslands where there is a loss of soil nutrients.

Our third hypothesis is that variations in bacterial and fungal composition are altered by fire and mowing, which indirectly alters other soil factors, such as nutrient and micronutrient availability. Soil microbial communities play a crucial role in regulating soil nutrient availability (Cheplick et al., 1989; McCaig et al., 1999; Unger et al., 2013), suggesting that their interaction determines soil sustainability and the development and productivity of plants (Song et al., 2021). Mowing and fires alter the source of energy and nutrients for soil microbial growth, as illustrated in this study ( Table 1 ; Figures 1 – 4 ), in which both fire and mowing redistributed the community composition and diversity of soil bacteria and fungi. This redistribution is much greater in the deep mowing than in the BG, which could be a crucial factor resulting in the diffidence of Miscanthus aboveground biomass. The coverage, richness, and diversity of the soil microbial community were well isolated in soils with four grassland managements, according to three alpha diversity indices ( Figure 3 ; Figure S1 ). Thus, our experimental findings indicate that deep mowing slightly altered the composition of soil bacteria, but largely altered the composition of soil fungal communities relative to other treatments ( Figure 3 ). This finding is consistent with previous research showing the responses of soil bacterial and fungal communities to various grassland mowing and management practices (Cui et al., 2020). This effect can be attributed to two factors. Soil microbial communities are sensitive to variability in mowing dryland ecosystems on a global scale owing to changes in nutrient sources (Delgado-Baquerizo et al., 2017). Deep mowing exacerbates the negative impacts of soil water loss on soil microbial communities (Blagodatsky and Smith, 2012). Moreover, this practice exacerbates the loss of soil nutrients ( Figure 2 ) due to the greater loss of the litter biomass (i.e., a much lower stubble heights-induced by the deep mowing) relative to other treatments. Therefore, the bacteria and fungi were significantly altered by deep mowing compared to other treatments because of soil environmental conditions and nutrients. First, deep mowing over Miscanthus growth significantly decreased soil OC and N ( Table 1 ), which can be used to provide a direct energy source and nutrients to soil microbes, affecting their growth and abundance. This study also observed that in SH5, soil microbial abundance (e.g., Firmicutes, and especially Proteobacteria) increased relative to that of the control ( Figure 3 ). This is consistent with previous results, demonstrating that Proteobacteria was abundant in less healthy soils (Li et al., 2014), as its abundance is sensitive to soil nutrient status (Esperschütz et al., 2007; Unger et al., 2013). This induced alteration in deep mowing significantly led to a decrease in the abundance of Basidiomycota and an increase in Ascomycota among the fungal communities. In particular, it further illustrated a much greater effect on fungal community diversity at the level of alpha diversity index (considering the Simpson diversity, ACE, and Chao 1 richness indexes) than other treatments ( Figures 4 , 5 ). Second, due to the loss of key energy and nutrients induced by plant growth, the soil microbial community diversity was altered. This study highlights that the sources of N, K, and P in the deep mowing over Miscanthus growth significantly decreased ( Table 1 ); thus, the soil microbial community diversity was significantly altered ( Figures 3 , 4 ). However, the reason for this remains unclear and needs to be further investigated to determine the differences between bacteria and fungi in soils with poor fertility. More interestingly, the factors of soil pH, available P and K, total N (TN), alkali-hydrolyzed N (AhN), and total OC explained 43% of soil bacterial community and 66% of soil fungal community ( Figure 7 ), highlighting their importance for soil microbial communities. In particular, for SH5, Proteobacteria and Acidobacteria at the phylum level were significantly related to the sources of energy and nutrients, including SOM, total N and P, and available N and P ( Figure 7A ). Similarly, the soil factors in SH5 were significantly related to those in most fungal communities. As discussed above, our experimental results clearly indicated that deep mowing resulted in a higher redistribution of the soil microbial community among all treatments. Although the stability of microbial communities in grasslands significantly affects soil quality and immunity (McCaig et al., 1999; Bonkowski and Roy, 2005; Otto et al., 2005; Cui et al., 2020), the structure and function of microbial communities in dry grasslands are influenced by numerous physicochemical soil factors (Jangid et al., 2011; Griffiths and Philippot, 2013; Zheng et al., 2019). Microbial communities are, therefore, regarded as a crucial mechanism affecting soil quality, which supports our third hypothesis. The use of various grassland managements has already been reported to alter the species abundance and composition of the soil microbial community during grass growth by altering soil nutrient availability, thereby affecting the aboveground biomass, as illustrated in Table S2 , which shows that biomass best explained the interaction between soil quality and microbial community. Here, our results demonstrate that the composition of the soil microbial community is significantly affected by deep mowing by altering soil chemical properties (especially soil N and P availability, as well as SOC), indirectly driving Miscanthus productivity.

Both fire and mowing significantly decreased the soil available N, P, and K content, tiller number per cluster, and plant height during Miscanthus growth relative to the control. However, these effects were greater in the deep-mowing plot than in the other plots. After harvest, both fire and mowing significantly decreased the dry weight of Miscanthus, whereas deep mowing induced a 5.2-fold decrease in dry plant biomass relative to the control. This could be attributed to the removal of aboveground straw by mowing, which directly leads to the loss of nutrients in soils. If sustained, our current experimental results indicate that nutrient limitation (N, P, and K) was an important factor determining the rates of Miscanthus accumulation in the repeated fires and the deep-mowing grasslands due to the loss of soil nutrients. In addition, both fire and mowing redistributed the community composition and diversity of the soil bacteria and fungi. This redistribution was much greater in deep mowing than in fire and light mowing. Soil bacterial abundance (e.g., Firmicutes, and especially Proteobacteria) in the deep-mowing plot increased relative to that of the control; this is consistence with previous results, indicating that Proteobacteria are abundant in less healthy soils as they are sensitive to the soil nutrient status. In contrast, deep mowing led to a significant decrease in the abundance of Basidiomycota and an increase in Ascomycota among the fungal communities, whereas it had an even greater effect on fungal community diversity at the level of alpha diversity indexes (considering the Simpson diversity, ACE, and Chao 1 richness indices) than other treatments. Thus, the variation in bacterial and fungal composition was altered by deep mowing, which indirectly altered other soil factors such as nutrient and micronutrient availability, resulting in a decrease in plant productivity. This finding indicates that deep mowing, rather than fire, induces soil processes and health development, thus restraining grass sustainability in a dry-alkaline grassland.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material .

ZM: Experiment; data collection, writing the original draft, and analysis. YW: Editing, formal analysis, and made suggestions of the manuscript. QL and WL: Experiment and editing. HW and YL: Experiment, editing, and analysis. MY: Conceptualization, editing, made suggestions of the manuscript, and corresponding author. All authors contributed to the article and approved the submitted version.