INTRODUCTION
Human activities have resulted in substantial increases in the frequency and intensity of extreme climate events worldwide (Meehl et al., 2000; IPCC, 2014). These climate extremes, which are projected to increase in the next decades, may exert large impact on the structure and functioning of plant communities, as observed for various regions following naturally occurring precipitation extremes (Buckland et al., 1997; Breshears et al., 2005; Ciais et al., 2005; PeƱuelas et al., 2007; Knapp et al., 2015) and in studies experimentally imposing precipitation deficits (Heitschmidt et al., 2005; da Costa et al., 2010; De Boeck et al., 2011; Dreesen et al., 2014). Considerable efforts have been directed towards assessing the stability of ecological communities experiencing extreme climate events (e.g., Tilman & Downing, 1994; Wardle et al., 2000; Jasper & Frank, 2010; Hoover et al., 2014). However, until recently research on this topic has mainly focused on the resistance and/or recovery of ecosystem functions (particularly biomass), paying far less attention to the resistance and recovery of the structure of ecological communities. This one-dimensional approach, however, precludes a comprehensive understanding of ecological stability across levels of ecological organization. On the one hand, community functional and structural stability may be positively associated with each other, a scenario that may arise when functional resistance/recovery of an ecosystem is closely linked to the resistance/recovery of community structure (Allison & Martiny, 2008; Baert et al., 2016; Guelzow et al., 2017; Hillebrand et al., 2018; Polazzo & Rico, 2021). On the other hand, functional response of a community to disturbance may not necessarily parallel its structural response. In particular, ecosystem functions are expected to be generally less responsive to disturbance than community structure, as functional redundancy among species may allow communities to mitigate their functional changes despite potentially substantial structural deviation from their pre-disturbance states (Yachi & Loreau, 1999; Allison & Martiny, 2008). Within this context, accumulating theoretical and empirical evidence suggests asynchronous population dynamics among species as a potentially important mechanism stabilizing ecosystem functions (Ives et al., 1999; Yachi & Loreau, 1999; Loreau & de Mazancourt, 2008; Hector et al., 2010; Hautier et al., 2014; Xu et al., 2021). However, greater asynchrony also indicates lower community structural stability, as it corresponds to greater change in species composition (Allan et al., 2011; Hillebrand et al., 2018). Thus it is possible for structural and functional stability to be decoupled or even negatively associated with each other.
Species diversity is often thought as an important factor influencing plant community resistance (Caldeira et al., 2005; Anja et al., 2012; Baert et al., 2016) and recovery (Jasper & Frank, 2010; Kreyling et al., 2017; Wagg et al., 2017). More diverse communities may better withstand disturbance, and recover more quickly from disturbance, due to the greater probability of containing species resistant to disturbance and species resilient after disturbance in more diverse communities, as well as increased asynchronous species responses to environmental changes in more diverse communities (Yachi & Loreau, 1999). On the other hand, theory suggests that more diverse communities may take more time to return to their steady states after disturbance (May, 1973; Ives & Carpenter, 2007), potentially resulting in lower structural and functional recovery with increasing diversity. Moreover, more diverse communities may be more likely to contain species vulnerable to disturbance, making it more difficult to maintain stable structure and function after disturbance. In apparent accordance with these opposing predictions, empirical studies of the diversity-resistance/recovery relationships have reported mixed results. For instance, ecosystem functional resistance has been found to increase (Tilman & Downing, 1994; Kahmen et al., 2005; Isbell et al., 2015), decrease (Pfisterer & Schmid, 2002; Allison, 2004; De Boeck et al., 2008), or remain unchanged (DeClerck et al., 2006; Wang et al., 2007; Van Ruijven & Berendse, 2010; Carter & Blair, 2012) with species diversity. Likewise, positive (Tilman & Downing, 1994; Van Ruijven & Berendse, 2010; Vogel et al., 2012), negative (Pfisterer & Schmid, 2002) and neutral (Carter & Blair, 2012; Xu et al., 2014; Isbell et al., 2015) diversity-ecosystem functional recovery relationships have been reported. By comparison, few studies have addressed the diversity-structural stability relationships, showing that species diversity may increase (Frank & McNaughton, 1991; Baert et al., 2016) or decrease (Van Peer et al., 2004) structural resistance, or shows varied relationship with structural resistance (Shurin et al., 2007); we know of only one study that examined diversity effect on community structural recovery (Baert et al., 2016). Therefore, future studies should consider examining how species diversity relates to both functional and structural resistance/recovery for a more comprehensive understanding of ecological consequences of ongoing biodiversity loss.
Rather than species diversity, the mass ratio hypothesis suggests that the properties of an ecosystem are largely determined by its dominant species (Grime, 1998). Consistent with this hypothesis, functional resistance and/or recovery of grassland communities in response to drought were found to be closely related to the resource-use strategies (Mackie et al., 2019), life history (Ruppert et al., 2015), traits and abundance (Macgillivray et al., 1995; Florence et al., 2014; Stampfli et al., 2018) of dominant plant species. Functional resistance of forest communities to drought has also been reported to depend upon dominant tree species (DeClerck et al., 2006). By comparison, the role of dominant species in modulating community structural resistance and recovery has rarely been assessed. Elucidating the importance of dominant species, relative to species richness and asynchrony, for resistance and recovery at functional and structural levels would provide a more mechanistic understanding of ecosystem dynamics under ongoing anthropogenic changes.
Increase in atmospheric nitrogen deposition has been reported to alter species richness (Vitousek et al., 1997; Suding et al., 2005), species asynchrony (Isbell et al., 2009; Hector et al., 2010), and the abundance of dominant species (Xu et al., 2015a; Bowman et al., 2018), with potential consequences for community resistance and recovery. However, nitrogen enrichment effects on community resistance and recovery remain understudied. The few empirical studies on this topic reported that increased nitrogen availability could increase grassland functional recovery after drought (Kinugasa et al., 2012; Xu et al., 2014), and increase (Hofer et al., 2017) or have little effect (Xu et al., 2014) on grassland functional resistance to drought. No studies, to our knowledge, have explored the linkage between nitrogen input and community structural resistance/recovery. As a common land-use practice in grasslands, mowing may also result in changes in species diversity (Maron & Jefferies, 2001; Williams et al., 2010; Socher et al., 2013) and the abundance of dominant species (Clark & Wilson, 2001; Galvanek et al., 2015). Nevertheless, we also know little about how mowing affects plant community resistance and recovery (but see Vogel et al., 2012), especially from the structural perspective.
In this study, we experimentally simulated nitrogen deposition and mowing in a temperate semiarid grassland in Erguna, Northeast China to assess their effects on grassland resistance and resilience from both functional and structural perspectives. The temperate semiarid grassland constitutes an important component of the Eurasian grassland biome, providing a range of necessary products and services for local human populations. Atmospheric nitrogen deposition is projected to increase in the next decades in this region (Liu et al., 2011), which, combined with mowing as a common grassland management practice (Wang et al., 2011), may impose significant impact on this important ecosystem that is also projected to experience more frequent and intense drought (Lu et al., 2021). During the experiment, our study grassland experienced natural drought in three consecutive growing seasons (2015 to 2017, Fig. S1), reducing aboveground plant biomass by 73.3%, 58.4% and 45.3%, respectively. This natural drought event provides us an excellent opportunity to assess the response of our study grassland to drought under nitrogen deposition and mowing scenarios. We aimed to 1) explore how nitrogen enrichment and mowing influence functional and structural stability (resistance and recovery) of the temperate grassland in response to drought, and 2) elucidate the pathways through which nitrogen enrichment and mowing influence functional and structural stability.