Plants in natural wetland ecosystems often experience hydrological disturbances (McGowan et al., 2011), and water level regulation is considered to be a useful tool in vegetation restoration (Leira and Cantonati, 2008; Yuan et al., 2017). Climate change causes strong hydrological disturbances, such as tidal variation, flooding, and extreme drought, which inevitably impose stress on plant individuals and communities, and numerous studies have examined the effects of such hydrological disturbances on the performance of plants in various transitional terrestrial and amphibious ecosystems (Wright et al., 2015; Wang and Li, 2016).

Abiotic disturbances and biotic interactions are important factors affecting plant community structure (Tilman, 1988; Grant et al., 2014). One classic approach for determining the spatial distribution of a plant community is identifying how ecological processes differ across environmental gradients (Keddy et al., 1997; Guo and Pennings, 2012). In recent decades, numerous studies have focused on plant-plant interactions along gradients of environmental disturbance (Bertness, 1991; Weigelt et al., 2002; Wang and Li, 2016) and have put forward the stress-gradient hypothesis (SGH), which predicts that facilitation and competition simultaneously affect neighboring plants and that the net outcome of these interactions shifts from negative to positive with increases in environmental stress (Bertness and Callaway, 1994; Maestre et al., 2009; Smit et al., 2009; He et al., 2013). This is a well-supported hypothesis (Liancourt et al., 2005; Maestre et al., 2005; Lortie and Callaway, 2006), but few studies have focused on the effects of hydrological disturbances, such as changes in water depth and hydrological fluctuations, on interspecific interactions among emergent wetland species (Luo et al., 2015).

In aquatic ecosystems, water depths fluctuate rather than being maintained at constant levels (Deegan et al., 2007; Wang P. et al., 2016). Increasing fluctuation and water depth can increase nutrient loss and tissue damage and decrease photosynthesis (Sasikala et al., 2008), which may cause species interactions to shift between facilitation and competition. There are also some opposing views, e.g., Bornette et al. (1994) found that aquatic plants performed best when flooding events occurred at an intermediate frequency, and Nielsen et al. (2013) found that increased hydrologic stability led to a loss of diversity in wetland communities. Because wetland species may differ in their responses and tolerance to hydrological changes, we predicted that water depth and hydrological fluctuations may greatly affect plant growth and interspecific interactions.

To better understand the importance of water depth and its fluctuations in the interactions among emergent plants in wetlands, we addressed the following questions: (1) Do water depth and hydrological fluctuations affect the growth of wetland plants? (2) Are the interspecific interactions among emergent species affected by hydrological fluctuations and water depth? To answer these questions, we conducted a greenhouse experiment testing the effects of three treatment categories (interspecific interactions, water depth, and hydrological fluctuations) on two common emergent wetland plants collected from the banks of the Chaobai River in Beijing.

The growth of both species was significantly decreased at the 30 cm water depth (Table 1), agreeing with the results of many previous studies (Voesenek et al., 2006; Panda et al., 2008). The 30 cm water depth was much higher than the initial height of both S. planiculumis and P. australis, which may cause reductions in the oxygen supply and carbohydrate use efficiency and increases in anaerobic respiration and nutrient consumption (Gibbs and Greenway, 2003), leading to an overall decrease in biomass accumulation. In nature, adult P. australis can survive in relatively deep water, over 1 m, because the mother ramets can provide nutrients to water-stressed seedlings through clonal integration. In our study, all the plants of both species were single ramets at the beginning, and the initial height was markedly lower than the static water level of 30 cm and higher than 10 cm. P. australis gained relatively more biomass in the 10 cm treatments than in the so cm treatments at the beginning of growth, and we predicted than the 30 cm water level will benefit the P. australis in later stages. Furthermore, many emergent species can morphologically adapt to submergence stress by elongating their shoots or leaves (Wang et al., 2014). In the case of S. planiculumis, the average height increased markedly with increasing water depth (Table 2), demonstrating morphological adaptation. Thus, morphological adaptation may explain the response of emergent species to water depth.

Some studies have found that hydrological fluctuations cause plants to become re-exposed to the ambient oxygen environment, resulting in the production of acetaldehyde and reactive oxygen species (Boamfa et al., 2005; Luo et al., 2012; Steffens et al., 2013), which are harmful to plant growth. However, the hydrological fluctuations did not obviously alter the growth of the two study species (Table 2). Some previous studies have found that plants are able to tolerate fluctuations of up to 30 cm without a significant drop in final biomass (Deegan et al., 2007; Wang P. et al., 2016). Luo et al. (2015) found that a small amount of hydrological fluctuation did not alter the growth of wetland communities because small changes may allow parts of the plant shoots to remain in the water, which causes the production of fewer oxidative toxins that can injure the plant tissue. In our study, the 20 cm range of fluctuation may have been too small to affect the growth of these two emergent species. Another possible explanation for this result may be that both S. planiculumis and P. australis have a strong ability to tolerate such water fluctuations and exhibit compensatory growth. Previous studies have also found that some other sources of stress or resources, such as low soil nutrients, nitrogen pulses, and altered water content (Geng et al., 2006; Sun et al., 2009; Wang et al., 2015), could not change the growth of target plants because the plants were insensitive to such stresses.

The inhibition or promotion of growth may be caused not only by submergence stress but also by interspecific interactions with neighbors (Wang Y. J. et al., 2016), and both interspecific and intraspecific interactions can be affected by the hydrological conditions. There is a trade-off between facilitation and competition in the interactions among plants, which means that plants exhibit greater competition in a suitable environment and more facilitation in a stressful environment (Maestre et al., 2009; Zhou et al., 2017).

Both the hydrological fluctuation and water depth treatments significantly affected the interspecific interactions between S. planiculumis and P. australis (Table 2; Figures 1, 2). Specifically, the mean RII values for P. australis increased with increasing water depth and fluctuation, while the values for S. planiculumis decreased (Figure 4). P. australis benefited more from the interspecific interaction than S. planiculumis when disturbance caused by fluctuating hydrological conditions increased; the plants performed much better in mixed culture than in monoculture under the fluctuating water conditions at 30 cm. This result mainly occurred because different plant species exhibit different levels of environmental tolerance, and P. australis appears to be better able to adapt to greater water depths and increasing fluctuation, such as those evaluated here, than S. planiculumis. For instance, Edwards et al. (2003) found that the amount of time necessary for morphological adaptation will depend on many factors, but, as an indication, it took the macrophyte Eleocharis cellulosa 3 weeks to adjust its morphological characters to new water level conditions and 9 weeks to shift its biomass partitioning. In addition, Deegan et al. (2007) established hydrological fluctuation treatments for emergent wetland species; Triglochin procerum did not respond to the treatments due to its ability to photosynthesise under water, while Typha domingensis responded negatively to increasing amplitude, and other species responded positively.

In nature, S. planiculumis and P. australis coexist during the early growing season (April to June). P. australis dominates the relatively deep-water areas through its clonal growth and morphological characters. Mauchamp and Mésleard (2001) found that the tolerance of P. australis to submergence stress increased with maturity, which indicates that time affects the interspecific interactions. Unlike P. australis, S. planiculumis accumulates more biomass under competition with its neighbors in a relatively static environment, but these effects disappear when hydrological disturbance is increased (Figure 2B). The results also indicate that P. australis is better able to adapt to disturbance than S. planiculumis. Some studies have found that the adaptive strategies of plants under high levels of disturbance caused by changing hydrological conditions mainly include (1) morphological adaptations via stem elongation and the formation of adventitious roots and aeration tissues, (2) alterations to metabolism, and (3) the regulation of physiological activity by changing the concentrations of hormones such as ethylene, gibberellin, and abscisic acid (Vasellati et al., 2001; Zhang et al., 2007).

Our study showed that (1) both S. planiculumis and P. australis accumulated more biomass at a relatively low water level (10 cm) than at a high water level, which indicates that water depth and wetland emergent plant growth were negatively correlated during the early growing season and that hydrological fluctuations did not obviously alter the growth of either target species; and (2) P. australis benefited from its interspecific interactions with S. planiculumis, which shifted from competition to facilitation with increasing water depth and hydrological fluctuations. These results partly explain the distribution patterns of both S. planiculumis and P. australis in relation to different hydrological conditions. P. australis showed an advantage in using varied hydrological conditions in competition with other emergent species, such as S. planiculumis. Moreover, these results may also contribute to our understanding of the organization and assembly of estuarine wetland plant communities and may have important implications for understanding how emergent wetland plant communities respond to the natural variation and human modification of hydrological conditions. However, the lack of O₂ content data and ORP data may limit the understanding and the results will vary in relation to different biotic and abiotic environmental stressors and their complex combinations in natural habitats. Further studies should evaluate the shifts between competition and facilitation in response to multiple environmental stressors in wetland plant communities.

JZ and L-JC conceived and designed the experiments; JZ, L-DZ, XP, and YN executed the experiment and measured the data; JZ, L-DZ, and JL analyzed the data and made the figures; L-JC, WL, and X-MK contributed to writing and editing the manuscript; JZ wrote the paper; JZ, L-DZ and M-XZ revised the paper.