Species richness both impedes and promotes alien plant invasions in the Brazilian Cerrado

In a field study conducted at 38 sites in two regions, we measured the abundance of alien invasive species, species richness of the plant community, total and soil extractable pools of P and N, soil phosphatase activity and the root phosphatase activity of nine common plant species. An analysis of the data using structural equation modelling (SEM) revealed no significant relationship between soil extractable-P concentrations and the abundance of alien plants (Fig. 1), despite the fact that previous studies in Cerrado have found P fertilization to promote the invasion of alien species³¹. However, the SEM did find abundance of alien invasive plants to be influenced by native species richness in two contrasting ways. One way was a direct negative relationship between species richness and the abundance of invasive species, which is consistent with the stochastic niche hypothesis and with results of some previous studies^14,15,16. This pattern was also observed in a direct regression between the two variables (Suppl. Figure 2). The other way was an indirect and positive effect of species richness that was mediated via phosphatase, suggesting that invasive plants may benefit from organic P released through phosphatase produced by soil microbes and/or plant roots.

A positive relationship between plant species richness and soil phosphatase activity (Figs. 1 and 2a) is consistent with the positive correlation between soil phosphatase activity and manipulated plant species richness observed in the Jena Biodiversity Experiment²⁹. The authors attributed the relationship to a positive effect of plant diversity on soil microbial activity and a tight coupling between soil C and P cycling. Because their SEM results did not explain one third of the variation in the data, the authors suggest that there may be a direct link, involving some as yet unknown path, between plant species diversity and soil phosphatase activity²⁹. Our finding of a significant positive correlation between root phosphatase activity and species diversity in three of our nine Cerrado species tested (Fig. 2b–j and Suppl. Table 1) not only supports this idea, but suggests a possible mechanism. We conclude, therefore, that under unfertilized conditions alien plants benefit more from the release of organic-P than from a relatively high availability of mineral-P, and that this release is higher in species-rich communities.

To further explore the relationships among plant species richness, plant and soil phosphatase activity, plant P and N uptake, and the abundance of invasive plants, we conducted a mesocosm experiment. We used Cerrado soil in which we planted simple communities composed of one of the two commonest invasive grasses, either Melinis minutiflora or Urochloa decumbens²², combined with the native grasses Saccharum asperum and Setaria poiretiana and/or the leguminous forb Stylosanthes guianensis (3 plants per mesocosm; see Suppl. Figure 3 for the design). To half of the mesocosms, we added a relatively high dose of inorganic P fertilizer (equivalent to 3 g P m⁻² in the form of dissolved Na₂HPO₄). This increased total soil P, but had no effect upon extractable P (Suppl. Figure 4a,b) or on plant P uptake (Suppl. Table 2), indicating that the added phosphate had been immobilized³⁴. This rapid immobilization of P in Cerrado soil explains why the addition of inorganic P had almost no effect on traits or on competition of these plants (Suppl. Table 2). It also underlines the importance for Cerrado plants of other P forms, perhaps including mono- or diester-bound organic P and phytic acid. It may be that Cerrado plants have a competitive advantage over soil microbes in accessing these forms of P, whereas microbes are more effective in competing for inorganic phosphate.

The results show that species richness had a significant positive effect upon the root phosphatase activity of all plant species, upon P uptake of four out of five species (Fig. 3a,b, Suppl. Table 2), and upon community P-uptake (Suppl. Figure 6g). These were large effects, with the average phosphatase activity and P uptake in the three-species-mixtures being 2.5 times higher than in monocultures (‘All’ in Fig. 3a,b). Unlike the field study, however, there was no correlation between plant species richness and soil phosphatase activity (Suppl. Figure 4e). Hence, for our mesocosm study it is clear that the species richness effect on phosphatase activity was not due to the enzyme activity of soil microbes but to that of the plants themselves, whereas the plant richness effect on soil phosphatase activity in field study could be due to both microbial and plant root activity. Indeed, for at least three plant species we observed a plant species richness effect on root phosphatase under field conditions (Fig. 2b–d). There was also no relationship between species richness and soil extractable-P (Suppl. Figure 4d), supporting the conclusion that the enhanced P uptake in the species mixtures was derived from non-labile P forms such as organic compounds. For two or three of the species studied the differences in root phosphatase activity between treatments were associated with differences in root morphology (specific root branching and surface area) as measured 13 weeks later (Suppl. Figure 5). This association could be causal, since phosphatase activity is known to be highest in the surface cells and apical meristems of roots³⁶, but it may also reflect a morphological next to the physiological phosphatase activity response to a greater P demand in the species mixtures.

We also found interspecific differences in foliar δ¹⁵N in the mesocosm experiment (Fig. 4a). The legume Stylosanthes guianensis had the lowest foliar δ¹⁵N in both monocultures and mixtures, presumably because it obtained most of its N through N₂ fixation. The alien grass Urochloa decumbens had lower foliar δ¹⁵N than the other three grass species in the monocultures (Fig. 4a), which could be due to the uptake of a different N form³⁷, although this difference was no longer detectable in the species mixtures (Fig. 4b,c). Plant N uptake of all four grass species and of the plant communities increased on average with increasing species richness (Fig. 3c, Suppl. Figure 6d). None of the grasses had a significantly lower foliar δ¹⁵N when growing with Stylosanthes guianensis compared to the monocultures (Fig. 4a,b), suggesting that this enhanced N uptake was not due to transfer from the N-fixing species. Instead, the species richness effect can be explained, at least in this short-term experiment, through the complementary use of different N sources (in particular, soil-N vs. atmospheric-N)^17,24,26. The enhanced N uptake in more diverse communities may also have created a greater demand of other potentially growth-limiting resources, such as P. Hence, species complementarity for different N sources, and enhanced plant N uptake, may have been a driving force for the observed richness effect on plant phosphatase activity (Suppl. Figure 6f,h,i). Similarly, enhanced N uptake in more diverse plant communities has been demonstrated for a broad range of European grasslands³⁸. Whether this is the case in the Cerrado plant communities, and whether it is a (co)-driving factor for the observed species richness effect on soil phosphatase activity, and hence on alien plant invasion (Fig. 1), remain to be tested.

Only one species in the mesocosms, the alien grass Melinis minutiflora, produced more biomass in mixtures than in monocultures (Fig. 3d). This is the most prominent alien invasive plant in the region of our study^22,31 (i.e., the region where also the mesocosm soil was collected), and appears to be the species best able to benefit from the additional P and N available in more diverse communities. Its success was not due to it producing a greater root length than other species (data not shown), and presumably reflects greater nutrient use efficiency although we lack the data (on nutrient residence time) to support this. The expansion of this alien species in our mesocosms did not lead to a higher community biomass (Suppl. Figure 6a), but to a shift in species abundances mainly at the cost of one of the native grasses, Setaria poiretiana (Fig. 3c). Our results support the conclusion, convincingly demonstrated in a meta-analysis, that native plants often promote the performance of alien plants relative to other co-inhabiting native species³⁹. Possibly, the most facilitative native species in our study was the leguminous forb Stylosanthes guianensis, for instance because it used a distinct different N source, but species interaction effects on root phosphatase activity, foliar δ¹⁵N and P or N uptake were also observed in mixtures of only grass species. Facilitation among species may also have been influenced by other soil microbes than N₂-fixing bacteria, such as mycorrhizal fungi. All species of our study are associated with arbuscular mycorrhiza. We did not record abundances of these fungi, or investigate their role in nutrient transfer among species in our study, but this is worth further investigation.

We note that the variation in species richness in our mesocosm experiment was only in the range of one to three species, whereas species richness in natural plant communities generally go far beyond this number of species. Results of long-term biodiversity experiments such as in Cedar Creek or Jena, as well as model predictions, have shown that already between monocultures and 2, 3 or 4 species mixtures very strong species richness effects are found on ecosystem properties and processes mediated by organism interactions^40,41. Moreover, the observed patterns in our mesocosm experiment and our field study—where species richness ranged between 1 and 55 per 4-m² species—were very consistent with each other. We therefore presume that the species richness effects on resource facilitation, as we observed in the mesocosm experiment, are very likely to occur in the field as well, and probably even stronger.

Source: Ecology - nature.com