Soil microbial communities shaped by plant diversity and drought
We observed pronounced impacts of plant species richness and soil moisture treatments on community composition of soil bacteria and fungi, indicating an effective conditioning process in Phase I. It has been frequently observed that soil microbial diversity increases with plant diversity, and the hypothesized mechanisms are that root exudates, litter heterogeneity and microhabitats provide more niche opportunities for soil microbes in more diverse plant communities (Youssef & Elshahed 2009; Millard & Singh 2010). However, our results demonstrated that the diversity of microbial communities did not always increase with greater plant diversity. We found that fungal overall diversity and AMF diversity increased and fungal pathogen diversity decreased with plant species richness under ambient conditions, while plant species richness increased AMF diversity under drought conditions. We did not observe any relationship between overall bacterial diversity and plant species richness. Our findings were consistent with an increasing realization that the relationship between microbial and plant diversity can exhibit complicated patterns that might differ for different functional guilds (Rottstock et al. 2014; Prober et al. 2015; Schlatteret al. 2015; Dassen et al. 2017; Mommer et al.2018; Jia et al. 2020).
The positive relationships between overall fungal diversity or AMF diversity with plant diversity could reflect the strong host specificity of these microbes and thus plant mixtures harboured a mix of fungal guilds associated with different plant species (Rottstock et al.2014; Dassen et al. 2017). The neutral relationship between bacterial diversity and plant species richness likely reflected functional redundancy of soil bacteria in our model systems. Thus, high plant species richness did not increase bacterial diversity nor resulted in increased functioning of the whole bacterial community (Allison & Martiny 2008; Louca et al. 2018). Under ambient conditions, the relative abundance of fungal pathogens increased with plant species richness, likely reflecting that increased root biomass and exudates stimulated the growth of more generalist soil pathogens (Eisenhaueret al. 2017) However, high-density root systems likely impeded species-specific soil pathogens in locating their hosts, and thus some specialized pathogen species present in monocultures were absent in plant mixtures, leading to decreased fungal pathogen richness with plant species richness (Mommer et al. 2018).
Biodiversity declined under drought stress for soil bacteria, fungi and AMF. AMF relative abundance decreased and fungal pathogen relative abundance increased under drought conditions. These findings reflect that drought likely limited activities of some guilds that were sensitive to drought stress and selected for particular drought-adapted guilds such as dormant pathogens and drought-resistant AMF (Mariotteet al. 2017; Meisner et al. 2018). Our results provided evidence that plant diversity and drought can condition soil microbial communities, leading to soil legacy effects that can alter subsequent PSFs (De Long et al. 2019).
The more pronounced plant diversity -productivity relationships in Phase II
In Phase I, the BE was not correlated with plant species richness in the ambient treatments even though the values were positive, and the BE was absent in the drought treatments, causing a non-significant plant diversity-productivity relationship (Fig. 3a). In Phase II, however, the BE increased with plant species richness, regardless of soil moisture treatments (Fig. 4d). Our results provided evidence that the positive plant diversity-productivity relationships were more pronounced over time, and soil microbes were one mechanism behind this finding (Schmidet al. 2008; van Ruijven et al. 2020).
These findings were in line with several studies that attributed long-term biodiversity effects to increasing niche complementarity in resource use and facilitation (van Ruijven et al. 2005; Cardinaleet al. 2007; Fargione et al. 2007). However, our results show that these temporal patterns in plant diversity-productivity can be primarily driven by PSFs (Maron et al. 2011; Schnitzer et al. 2011; Kulmatiski et al. 2012; Jing et al. 2015; Wanget al. 2019). We found that the positive CE-plant richness relationships were more pronounced in Phase II compared to Phase I. A possible explanation was that accumulations of species-specific soil pathogens decreased community biomass in the monocultures, while the community biomass of mixtures was promoted due to enhanced diversity of AMF (van Der Heijden et al. 1998; Wagg et al. 2011; van Ruijven et al. 2020). Future studies are necessary to test this hypothesis through identifying species-specific fungal species and recording their shifts in abundance as plant species richness increase.
Interestingly, we observed that the SE became more negative in higher plant richness levels in Phase II (Fig. 4f), and this result was consistent with previous findings that the positive selection effects become negative over time (Fargione et al. 2007; Marquardet al. 2009). In the present study, unproductive plants, D. aegyptium , increased their performance more than productive species in high plant richness levels, leading to a negative selective effect (Fig. S3). In other words, unproductive species might suffer from stronger negative density dependence than productive species, and the possible mechanism was that productive species could produce dense and large root systems in mixtures that declined dilution effects of species-specific soil pathogens (Mommer et al. 2018; van Ruijven et al.2020). Previous studies have suggested that negative PSFs can maintain species richness by preventing productive species from dominating the plant community (Bever 2003; Adler & Muller-Landau 2005; Petermannet al. 2008; Mangan et al. 2010). The implication of our results is that soil conditioning can alter biodiversity effects by increasing complementarity and decreasing selection effects through PSFs.