Discussion

To the best of our knowledge, our meta-analysis has provided the first global evidence that belowground fine-root attributes could be modified to meet elevated resource demands in species-rich plant communities. We demonstrated that species mixtures increased fine root biomass, with more pronounced increases observed in older stands and deeper soil layers. Although on average, plant mixtures did not alter other fine root traits, the effects of species mixtures on specific root length shifted from negative to positive with stand age, positive in two-species mixtures to negative in more species-rich mixtures, and positive to negative with soil depth. The effects of plant mixtures on root length density shifted from positive to negative, mostly in croplands, with increasing mean annual temperature. Plant mixtures had no effects on weighted rooting depth in grasslands or planted forests but had a positive effect under cold and moist climates, and a negative effect under warm and dry climates in natural forests.
Unsurprisingly, we found that species mixtures, on average, increased fine root biomass, and the positive mixture effects increased with species richness and soil depth with more pronounced species richness effects in both older stands and deeper soil layers. Our results extended the aboveground overyielding to belowground (Cardinale et al. 2007; Zhang et al. 2012; Liang et al. 2016), particularly, the pronounced diversity effects in more species-rich and older stands and in deep soils (Ma & Chen 2016). This result supported the notion that complementary effects increase with species richness in mixtures (Barryet al. 2019), and functional redundancy decreases with stand age, while interspecific facilitation increases with soil depths (Jobbágy & Jackson 2001; Makita et al. 2010; Reich 2012; Forrester & Bauhus 2016). This finding suggests elevated water and nutrient demands in species-rich and old mixtures, leading to deeper soil exploration. Moreover, the mixture effects on root biomass were significantly different between ecosystem types, which might have resulted from differences in the average species richness of mixtures between ecosystem types (Table S3).
However, contrary to the prediction of the stress gradient hypothesis, we found that the mixture effects on root biomass increased with mean annual precipitation, particularly in species-rich mixtures. Although interspecific facilitation may be enhanced through resource limitations (Maestre et al. 2009; Forrester & Bauhus 2016), it is possible that increased water and nutrient availability augmented niche differentiation, which consequently leads to stronger diversity effects on productivity (Searle & Chen 2019). Further, in dry climates, species mixtures can increase soil moisture content, which alleviates heightened water requirements for roots (Lange et al. 2014). The enhanced mixture effects on root biomass with water availability in more diverse communities might be attributable to stronger positive resource partitioning in species-rich mixtures (Barry et al. 2019).
On average, root functional traits, including the root:shoot ratio, community weighted-mean rooting depth, root length density, specific root length, mean root diameter, and root nitrogen content, did not differ between species mixtures and the mean of corresponding monocultures. The lack of a mixture effect on the root:shoot ratio suggested that fine root overyielding was of the same magnitude as its aboveground counterpart on a global scale. Furthermore, the variations of mixture effects on root biomass were synchronous with those of root length density and specific root length (Fig. S2). The neutral mixture effects may have been attributable to the fact that the majority of the original studies consisted of two species mixtures with short experimental durations, in which mixture effects are expected to be minimal due to limited interspecific interactions between individual plants (Lei et al. 2012; Beyer et al. 2013; Siebenkäs & Roscher 2016). Nevertheless, the mixture effects on several functional traits were highly dependant on the species richness in mixtures, stand age, soil depth, or environmental stress.
We found that the effects of species mixtures on root length density shifted from negative to positive from young to old stands, topsoil to deep soils, and warm to cold climates. Firstly, the increasing plant mixture effects on root length density with stand age were anticipated, since diversity effects should facilitate fine roots to have a high resource uptake capacity to satisfy the elevated water and nutrient demands (Cardinale et al. 2007; Zhang et al. 2012), which could be achieved by the higher horizontal soil volume utilization of fine roots in older stands (Brassard et al.2013; Ma & Chen 2017). Secondly, increased root length density with soil depth implied that fine roots penetrated deeper into the soil to uptake additional soil resources to compensate for the overyielding of plant mixtures (Zhang et al. 2012; Ma & Chen 2017; Oram et al. 2018). Lastly, in colder climates where fine roots are likely to face lower resource availability due to the slower fine root decay rate (See et al. 2019), the increased interspecific facilitation in mixtures (Forrester & Bauhus 2016) might increase root length density for an improved resource capacity. The shifted mixture effects on root length density suggested that the intense competition for resources overridden interspecific facilitation in young stands, shallow soil, and colder sites.
We also found that the mixture effects on specific root length shifted from positive to negative from two to higher numbers of species in mixtures, topsoil to deep soils, and from negative to positive with increasing stand age. Firstly, the decreased specific root length with the species richness in mixtures implied that fine roots reduced resource uptake efficiency in more diverse communities (Ostonen et al. 2007), which might have resulted from more intense resource competition. A recent meta-analysis demonstrated that the soil organic carbon content exhibited small variations with increasing species richness globally (Chen et al. 2019), which suggested a stable soil nutrient pool regardless of species richness. Therefore, increased soil nutrient competition leads to a decreased specific root length for a more conservative strategy with lower carbon costs (Reich 2014). Secondly, in alignment with the notion that interspecific facilitation increases with soil depth (Forrester & Bauhus 2016), we found that the mixture effect on specific root length decreased in deeper soils for lower resource availability. To support aboveground progressive overyielding with species richness (Liang et al. 2016), the high diversity effects on specific root length in surface soil might be a compensating strategy of resource uptake in this root-rich soil layer (Yuan & Chen 2010). Lastly, the increased mixture effects on specific root length with stand age were attributable to elevated water and nutrient demands in older stands, which resulted in a high resource uptake efficiency. Moreover, we found that the negative diversity effects on specific root length shifted to neutral in older stands (> 5 years, Fig. 3b). Due to the positive mixture effects on soil organic carbon content with stand age (> 5 years) (Chen et al. 2019), the intense competition for nutrients in more diverse communities might be counteracted in older stands.
We found that plant mixture effects on root attributes were highly dependant on the species richness in mixtures, stand age, soil depth, or environmental stress. To address the high water and nutrient demands in the support of greater aboveground productivity in plant species mixtures, fine roots increased the root biomass and/or root length density, but decreased the specific root length, in relation to both the species richness in mixtures and soil depth. We also found that the plant mixture effects on root biomass, root length density, and specific root length increased with stand development. Across global climatic variations, the mixture effects on root biomass increased with the mean annual precipitation, and the increased trends were more pronounced in more diverse plant communities, while the mixture effects on root length density decreased with the mean annual temperature. Our analysis highlights the need to incorporate the number of species in mixtures, stand age, and soil depth profiles, toward examining mixture effects on root attributes. Because of the dominant role of fine roots in soil resource exploration, our results suggest that increased fine root biomass with shifts in fine root traits enhanced soil resource uptake to support high primary production in mixtures.

Table 1 Fine-root traits and resource uptake strategies.