4 | Discussion

Many studies have explored the response of flowering phenology to global change. Three aspects of our study, however, distinguish it from these previous studies. First, while a few studies of the effects of global changes on phenology have considered one factors, our study examines the effects of N addition, plant diversity loss, and their interactions on plant flowering phenology in a single experiment. Second, our study explores the effects of N addition and plant diversity loss on multiple flowering phenology events. Third and most importantly, through three different analysis methods, we confirmed that N addition and plant diversity loss altered flowering phenology through its effects on functional traits rather than abiotic factors.

4.1 | The effects of N addition and plant diversity loss on flowering phenology

The flowering phenology is one of the more important factors determining the reproductive success of plants, because the timing of flowering means when plants expose it reproductive organs to the changing biotic and abiotic environments (Rathcke & Lacey, 1985; Kudo & Hirao, 2006). In our study, N addition delayed the first flowering day (FFD) but advanced the last flowering day (LFD), and then resulted in a shortened flowering duration (FD), which consistent with those of previous manipulative experiments in natural grasslands (Smith et al., 2012; Xi et al., 2015). Nevertheless, the 1.92 d delay in FFD at the N addition plots is shorter than that reported from the alpine grasslands in North America (Smith et al., 2012); the 1.2 d advance in LFD and 3.2 d shortened in FD is longer than that reported from the alpine grasslands in China (Xi et al., 2015). Which suggesting that the effects of N addition on flowering phenology varies among ecosystems and phenological stages.
Different from the effects of N addition, plant diversity loss advanced the FFD but delayed LFD, which resulted in an extension in FD. The 0.31 d advance per species lost in FFD is similar to the results (ranging from an advancement of 1.8 d per species lost to a delay of 0.7 d per species lost, average is 0.6 d advance) that reported from the serpentine grasslands in North America (Wolf et al., 2017), suggesting that the effects of plant diversity loss on FFD varies among species. Moreover, we noticed that the amplitude of changes (0.64 d) in LFD is more than that in FFD (0.31 d) after per species lost, which revealing that LFD may be more sensitive to plant diversity loss.

4.2 | The effects of N addition and plant diversity loss on functional traits

We found evidence to support our second hypothesis that increased N inputs promote light acquisition traits. Specifically, leaf mass and area of M. sativa was promoted after N addition. These results were inconsistent with that reported from the alpine grasslands in China (Liu et al., 2017; Zhang et al., 2019), which may be induced by the differences in species. M. sativa, a legumes, can fix N by means of symbionts to relieve their N limitation (Bordeleau & Prévost, 1994), and thus the nutrient acquisition traits of M. sativa had no changes under N addition. Meanwhile, the relief of N limitation accompany with the increasing demand of light (Hautier et al., 2009). Hence, the changes in light demand of M. sativa under N addition could lead to an increase in leaf mass and area.
Our findings for plant diversity loss also support our second hypothesis that both light (specific leaf area) and nutrient (leaf nitrogen content and relative abundance) acquisition traits was promoted under plant diversity loss. These results were inconsistent with the effects of plant diversity loss on grass species that reported from the grasslands in Germany (Gubsch et al., 2011), which may be caused by the differences in species. Closed canopies are characterized by pronounced gradients in spectral light quality and quantity (Jones, 1992). In the present study, plant diversity loss impacts on morphological traits associated with light acquisition were mainly attributable to the changes in relative height of M. sativa. Relative height decreased with plant diversity loss, indicating higher efforts for light acquisition should be allocated by plants (Spehn et al., 2005; Lorentzen et al., 2008). Hence, M. sativa exhibited high specific leaf area in adjustment to increasing competition for light. Moreover, nitrogen is not the limitation resource of M. sativa. Higher light acquisition traits are accompanied with higher nutrient acquisition traits (Hautier et al., 2009; Zhang et al., 2019), which resulted in an increase in leaf nitrogen content and relative abundance under plant diversity loss.

4.3 | Functional traits rather than abiotic factors determine flowering phenology

Our findings not support for our third hypothesis that functional traits and abiotic factors co-diver the response of flowering phenology to N addition and plant diversity loss, the contributions of functional traits for changes in flowering phenology was far more larger than abiotic factors (Fig. 5, 6, and 7). Specifically, the advance in FFD under biodiversity loss was mainly induced by the decrease in available soil N, which was consistent with the result from Wolf et al. (2017) in American serpentine grassland. Lower available soil N advanced the plant switches from the growth to reproduction stage (Wang & Tang 2019), and then indirectly caused the advancement in FFD at the monocultures.
However, the response of LFD and FD to biodiversity loss and N inputs was mainly driven by leaf nitrogen content and relative abundance. It has been proved that plant morphological and physiological traits could elucidate the changes in plant phenology (Jia et al., 2011; König et al., 2018; Pérez-Ramos et al., 2019) and the alteration of living strategies of plants (Wilson & MacArthur, 1967; He et al., 2008; Pérez-Ramos et al., 2019). At the monocultures, higher efforts for nutrients (leaf nitrogen content and relative abundance) acquisition means increase in intraspecific competition for resource among individuals of M. sativa (Lorentzen et al., 2008; Liu et al., 2017), and K-strategy with higher survive rate was chosen by M. sativa at the monocultures, which resulted in delaying LFD, extending FD, and decreasing FN (Fig. 2). As species richness increase, the intraspecific competition for resource decrease, but the interspecific competition among different species increase (Hector et al., 1999; Tilman et al., 2001). However, because the predominant competition advantage of M. sativa (high relative abundance and plant height, Fig. 4) at the mixed-cultures, r-strategy was chosen by M. sativa with more flower numbers and lower survival rate (Guiz et al., 2018), which resulted in advancing LFD, shortening FD, and increasing FN (Fig. 5b, c, and d). Therefore, the gradually shifts from r-strategy to K-strategy of M. sativa led to the changes in flowering phenology under plant diversity loss.