Introduction
It is clear that environmental changes could alter plant community composition and diversity, and competition (plant – plant interaction) is often considered as one of the main drivers of the consequence of environmental changes (Grime, 1979; Tilman, 1982). Therefore, plant ecologists have been investing tremendous efforts in interpreting and predicting how the intensity (amount of inhibition or facilitation in the target’s performance as a consequence of a neighbor’s presence) and importance (the impact of a neighbor on the target expresses as a proportion of the total environmental impact on the target) of competition varies along the environmental gradients especially soil nutrition gradient, and have concluded three theory models with empirical evidences (Bertness et al. 1994; Brooker et al. 2005; Gaucherand et al. 2006; Grime, 1979; Laurent et al. 2017; Tilman, 1982). First, it is assumed that competition inhibition would decrease from benign (fertile) habitats to harsh (unfertile) habitats, which is called “stress gradient hypothesis” (Bertness and Callaway 1994; Grime, 1979). Second, it predicts that total competition is equally important throughout soil resource gradients, and belowground competition is most intense for soil water and/or nutrition in unfertile sites and aboveground competition is most intense for light and space (aerial competition) in fertile sites, which is called Tilman’s theory (Tilman, 1987). Third, it argues that competition importance increases with the increase of resource availability and it is higher in fertile environments, which is called Grimes’s theory (Grime, 1977).
Although plenty of experiments have successfully supported these models under different soil resource conditions, there is a growing number of experiments that have emphasized the importance of species identities (both target and neighbor) in influencing competition for resources and consequently influencing the community composition and structure (Read, 2016; Suding et al. 2004; Yelenik et al. 2017). The responses of a single target species varying based on the identities of neighbors have been found in several studies (Kong et al. 2018; Stoll et al. 2000). For example, Keddy et al. (2002) found that the response of a single target species to competition was equally sensitive under two different environmental conditions but varied with the identities of 63 experimental competitors. These experiments have also shown that the effect of neighbor identity to competitive intensity may even surpass that of environmental differences. Although there were not consistent conclusions, two components of species, the effects of neighbors on resource availability (the growth potential) and the ability of individuals to tolerate these effects, were usually considered when researchers tried to understand how species identities influenced the outcomes of plant-plant interaction (Goldberg, 1996; Suding et al. 2004). In addition, ecologists found that the co-occurring species may compete or facilitate one another by modifying the microbial environment to change nutrition availability (Brooker et al.2008). Several studies have found out that a species could influence the outcomes of plant-plant interaction by immediate effect when another species co-occurred with it, legacy effect or plant-soil feedback when another species grew in the soil that it has been conditioned (Zhao et al. 2018).
Suding and Goldberg (2001) addressed the processes responsible for the patterns of species composition change after the environmental changes, and suggested that both competitive reduction and competitive change were two alternative hypotheses to explain the relative competitive abilities of different target species affected by different neighbor species in changing environmental conditions. The competitive reduction hypothesis predicts no shift of competitive hierarchies of the target species, suggesting environmental changes may modify species associations simply; while the competitive change hypothesis predicts competitive hierarchy of target species may shift with the change of environmental conditions, thus influencing the species composition, not just the intensity of competition (Suding et al. 2001). Therefore, considering the species identities is very important for understanding the community processes and intrinsic mechanisms in the context of global changes which have altered the soil resource conditions and community composition and structure dramatically.
Arid and semi-arid grasslands are very sensitive to the environmental changes. In this region, nature ecosystems are being degrading to more stress conditions because of over-grazing and drought (Huang et al. 2019; Ren et al. 2018). Simultaneously, the observed acceleration in N cycle caused by fossil fuel combustion, fertilizer use and legume shrub invading into grassland substantially have contributed to soil nutrition (Chen et al. 2019; Zhou et al. 2019). The different responses (growth potential or tolerance) of different species to such changes would influence the outcome of plant-plant interaction, alter plant community composition and diversity. Similar to the worldwide terrestrial communities, the semi-arid steppe of China is suffering from global changes (Li et al. 2012). In this region, Stipa grandis community is the most normal, stable and representative community, and it is very sensitive to environmental changes; and compared with S. grandis ,S. krylovii occupies dryer and more infertile habitats. The previous experiments demonstrated that the lower tolerance of S. grandis to drought, disturbance and infertile soil condition might be the reason for the replacement of S. grandis by S. krylovii (Chen et al. 2013). Recent study showed that soil conditioned by S. grandis with high N treatment could increase S. krylovii ’s competition superiority and N-addition effect in a low-density controlled experiment (Zhao et al. 2018). Therefore, exploring how the identities of neighbors influenced the competition hierarchies of S. grandis and S. krylovii under different soil nutrition conditions is very important for us to understand the mechanisms of community dynamics and predict community processes.
In this study, we conducted a microcosm experiment to explore how soil nutrition condition, neighbor species influenced the intensity and importance of target species (S. grandis and S. krylovii ), how plantation condition, species identity influenced nutrition-addition effect, and how the variation of the competitive hierarchy altered with the change of soil nutrition when the target species were grown with different neighbors. Especially, we conjectured that the factor of soil nutrition condition might not be the main driver that influenced the intensity and importance of competition on S. grandis and S. krylovii . What’s more, we hypothesized that S. krylovii would show a higher competitive hierarchy than S. grandis in the low fertility condition. Such results would be very important for us to understand and predict the community dynamic in the semi-arid steppe of China when facing environmental changes.