Discussion
Our findings show that plant and soil biodiversity have non-substitutable impacts on the temporal stability of community biomass production. In the present study, soil biodiversity loss had a detrimental effect on temporal stability by reducing the community-level mean biomass production, increasing the temporal variability, and weakening compensatory effects (via facilitation of uneven community composition favoring grasses over herbs and legumes). Consistent with previous studies (Tilman et al. 2006; Hector et al. 2010; Hautier et al. 2014; Hautier et al. 2015; Craven et al. 2018; Wanget al. 2019b), an increase in plant diversity in terms of both species richness and functional diversity promoted temporal stability. However, plant and soil biodiversity exerted independent effects on temporal stability. Furthermore, when multitrophic biodiversity was calculated from plant and soil biodiversity, multitrophic biodiversity was positively and linearly associated with the temporal stability of community biomass production. A large number of studies suggest that the biodiversity of a single trophic group, e.g. the diversity of plants, is a major factor stabilizing biomass production (Tilman et al.2006; Hector et al. 2010; Hautier et al. 2014; Hautieret al. 2015; Pennekamp et al. 2018; Craven et al.2018; Wang et al. 2019b). Our study suggests that maintaining ecosystem functions (e.g. biomass production) when faced with environmental variability requires biodiversity at multiple trophic levels.
Compared with plant species richness, the diversity of plant functional traits was a better explanatory variable for temporal stability in diverse plant communities. Exploitative plant species with fast-growing acquisitive traits, e.g. high specific leaf area, leaf nitrogen concentration, specific root length and specific root surface area, could recover rapidly following disturbance, while conservative species with slow-growing traits, e.g. low leaf dry matter concentration, high root average diameter and root tissue density, may have a higher resistance to disturbance (Fischeret al. 2016; Liu et al. 2018; Mahaut et al. 2020; Lozano et al. 2020; Freschet et al. 2020). Plant communities with a higher diversity of functional traits likely imply a stronger asynchronous response of plant species, e.g. a decrease in exploitative plant species compensated by conservative species during disturbance, and vice versa . In the present study, the diversity of plant functional traits increased plant species asynchrony, and subsequently, promoted the temporal stability of biomass production.
Plant species asynchrony is a main underlying mechanism by which plant diversity stabilizes biomass production when faced with disturbance (Loreau & de Mazancourt 2008; Isbell et al. 2009; Hectoret al. 2010; de Mazancourt et al. 2013; Loreau & de Mazancourt 2013; Thibaut & Connolly 2013; Hautier et al. 2014; Craven et al. 2018; Zhang et al. 2019). Supporting past studies, we found that higher plant species richness, soil biodiversity and multitrophic biodiversity promoted the temporal stability of biomass production via enhancing plant species asynchrony. A previous study found that the presence of natural soil biota dramatically increased the growth of legumes and herbs, and thus, promoted plant species asynchrony under simulated environmental variation in precipitation (Pellkofer et al. 2016). Our results suggest that soil biodiversity is crucial for the occurrence of asynchrony of plant species and particularly functional groups. In accordance with past research (Wagg et al. 2014; Wagg et al. 2019; Prudentet al. 2020; Yang et al. 2020), we found that soil biodiversity enhanced the performance of herbs and legumes during wet and dry periods, likely indicating higher resistance during dry periods or faster recovery during wet periods. Specifically, herbs had the opportunity to recover from drought and differed from grasses in response to the simulated variation in precipitation at high soil biodiversity (Fig. 5). Besides, the dramatic reduction in legumes was compensated by the rapid growth of grasses during the first drought disturbance at high and moderate soil biodiversity. However, the compensatory effect was weaker in low soil biodiversity because legumes had already been suppressed by soil biodiversity loss before the first drought disturbance. A recent study shows that soil biodiversity is more important for herbs than grasses, and especially for the persistence of legumes (Yang et al. 2020). Besides, the growth of herbs and legumes is more reliable in the presence of mycorrhizal soil mutualists (Hoeksema et al. 2010; van der Heijden et al. 2016). The absence of these soil mutualists, in addition to the loss of soil biodiversity, could have suppressed the performance of herbs and legumes during and following drought in the present study.
It should be noted that soil biodiversity loss did not affect the growth of herbs and grasses, and even increased the growth of legumes in the monocultures. Therefore, soil biodiversity influenced the performance of plant functional groups by regulating interactions among functional groups in the mixed-species communities. For instance, soil biodiversity increased the mean of plant community evenness by favouring herbs and legumes. Although biodiversity can enhance the temporal stability by promoting community evenness (Thibaut & Connolly 2013), we found that plant community evenness was not related to the temporal stability of community biomass production. Because the proportional abundance of grasses was positively related to temporal stability, soil biodiversity loss should increase temporal stability by promoting the dominance of grasses. However, soil biodiversity loss exerted a negative effect on temporal stability. Thus, the detrimental effect of soil biodiversity loss on temporal stability cannot be directly attributed to its effect on community evenness and the dominance of grasses, but can come from a decrease in asynchrony of plant functional groups.
A decrease in population variance can also contribute to the temporal stability of biomass production, because effects of treatments on temporal stability can be decomposed into their effects on plant species asynchrony and population variance (Thibaut & Connolly 2013). Higher functional richness and soil biodiversity increased population variance, which should decrease temporal stability. However, these detrimental effects were neutralized by an increase in plant species asynchrony. An increase in the temporal stability of biomass production can be caused by both, an increase in the temporal mean of biomass production and a reduction in the temporal standard deviation (Tilman et al. 2006; Hautier et al. 2014). Consistent with most earlier studies (Tilman et al. 2014; Hautier et al. 2014; Craven et al. 2016; Weisser et al. 2017), our study supports the hypothesis that an increase in plant species richness can promote biomass production and decrease the temporal standard deviation of biomass production.
Soil biodiversity loss can decrease or increase biomass production, which depends on soil biodiversity and soil community composition (Wagg et al. 2014). Our results suggest that the temporal mean of biomass production was not altered by soil biodiversity loss when monocultures were included, while the temporal mean was decreased when monocultures were excluded (Fig. 3C and H). This suggests that soil biodiversity loss has a detrimental effect on the temporal mean of biomass production in diverse plant communities. However, soil biodiversity loss did not affect plant diversity-productivity relationships in diverse plant communities. Past research shows soil pathogen suppression has more beneficial effects, while the absence of soil mutualists exerts stronger detrimental effects on biomass production in less diverse plant communities, and consequently, influences plant diversity-productivity relationships (Klironomos et al. 2000; Maron et al. 2011; Schnitzeret al. 2011; Luo et al. 2017; Liang et al. 2019; Wang et al. 2019a). However, soil biodiversity loss did not alter the strength of plant diversity effects on plant biomass production in the present study. These results indicate that the effect of soil biodiversity loss on plant productivity could not be simply due to the loss of single soil pathogens or mutualists. In the present study, soil biodiversity loss reduced plant fungal pathogens in soil inocula, and reduced or even eliminated mycorrhizal soil mutualists in the phylum ofGlomeromycota at the last harvest. Reductions in soil pathogens and mutualists have opposite effects on plant biomass production, and therefore, effects of reductions in soil pathogens and mutualists along soil biodiversity loss might cancel each other out. A decrease in plant productivity could be caused by soil biodiversity loss in general other than the loss of soil mutualists along with soil biodiversity loss (Prudent et al. 2020; Yang et al. 2020).
Our study sheds new light on the consequences of multitrophic biodiversity loss under global anthropogenic change. Anthropogenic influences have been shown to destabilize biomass production via reducing plant diversity (Hautieret al. 2015). Moreover, global anthropogenic change, such as nitrogen deposition, land-use intensification, warming, fertilization and drought, can threaten both soil and plant biodiversity (Tsiafouliet al. 2015; Hautier et al. 2015; Gossner et al.2016; Banerjee et al. 2019; Geisen et al. 2019; Rilliget al. 2019; Zhou et al. 2020). Our study suggests that the loss of multitrophic biodiversity could reduce the temporal stability of biomass production by suppressing the occurrence of plant species asynchrony or decreasing the mean of biomass production and increasing its variation. These results indicate that the prediction of biomass production under global anthropogenic change requires a multitrophic evaluation of biodiversity loss.
In this study, the dilution-to-extinction approach, which has been widely used to investigate the relationships between rare soil microbial species and ecosystem functions (Yan et al. 2015; Hol et al. 2015; Roger et al. 2016; Maron et al. 2018; Domeignoz-Hortaet al. 2020), was employed to create a gradient of soil biodiversity. Our results confirm that less abundant taxa were lost first during dilution, followed by more abundant taxa (Yan et al.2015; Roger et al. 2016; Maron et al. 2018). This approach can simulate a realistic loss of soil biodiversity (Yanet al. 2015; Roger et al. 2016; Maron et al. 2018), because less abundant taxa have a higher risk of extinction than abundant species under anthropogenic changes (Zhou et al. 2020). The present study used an open system, and an increase in bacterial phyla was observed, compared to the initial soil inoculum. This means that propagules from the environment contributed to soil biodiversity at the last harvest. Contamination in a sterile treatment was even detected in closed systems (Wagg et al. 2014) and is generally difficult to avoid. Because soil biodiversity generally had a positive effect on ecosystem functions in previous studies and our present study, this indicates that we have underestimated the effect of soil biodiversity loss on the temporal stability of biomass production.
In summary, our study suggests that plant and soil biodiversity play non-substitutable roles in stabilizing plant community biomass production. Although greater plant diversity can promote the temporal stability of biomass production, the reduction in temporal stability induced by soil biodiversity loss cannot be compensated via increasing plant diversity. This result highlights the significance of multitrophic biodiversity for stabilizing ecosystem functions. Our study has important implications for restoration and conservation management in terrestrial ecosystems. First, it is important to preserve biodiversity at multiple trophic levels to stabilize ecosystem functions, especially in ecosystems suffering from intense disturbance. Second, although biomass production can be stabilized via increasing plant diversity by restoration or conservation measures, we should not ignore the importance of maintaining soil biodiversity, since low levels of soil biodiversity can destabilize biomass production.