Effects of different scenarios of temperature rise and biological control agents on interactions between two noxious invasive plants

An increasing number of exotic plants and their biological control agents are being introduced into new regions. Therefore, it is necessary to study their interactions and to manage the future directions of plant invasions under different scenarios of temperature rise.


| INTRODUC TI ON
When exotic plants are introduced into new ranges, they develop novel interactions with native plants. Therefore, many researchers have proposed hypotheses to explain the invasion mechanisms of exotic plants and to predict and manage exotic plants according to the interactions between exotic plants and natives (Jeschke, 2014;Mitchell et al., 2006). After an increasing number of exotic species have invaded the same region (Seebens et al., 2018), one exotic species encountering another exotic species within a habitat becomes unavoidable. Most habitats have been invaded multiple times, and studies of the interactions between multiple invasive species may be as important as studies of native-invasive interactions for understanding the invasion process and for future management of invasive species. However, few studies have analysed invader interactions (Kuebbing & Nuñez, 2015;Kuebbing et al., 2013).
The invasion meltdown hypothesis suggests that some exotic species can facilitate one another's invasions even if they did not undergo co-evolution (Simberloff, 2006;Simberloff & Von Holle, 1999).
Herbivory is an important factor affecting interactions among exotic plants (Clay et al., 1993;Hartley & Amos, 1999;Mulder & Ruess, 1998; Van et al., 1998). In their introduced ranges, native herbivores tend to accumulate on invasive plants over time due to exotic plants lacking effective defences against them  or due to the rapid adaptation of predators to invasive species (Siemann et al., 2006). Moreover, the co-evolved specialist enemies of some invasive plants may be introduced into the same area as biological control agents for these invasive plants (Bellows, 2001). Herbivores cause different levels of damage to exotic plants (Carpenter & Cappuccino, 2005;Jogesh et al., 2008), which may alter the interactions among exotic plants. The invasive plant Eichhornia crassipes (Mart.) Solms (Pontederiaceae) is highly competitive and can replace other invasive plants, but it was found that Pistia stratiotes L. could replace E. crassipes after specialist insects were introduced to control the latter species (Center et al., 2005).
Climate warming is the major component of global change (Walther et al., 2009) and can affect the distribution, performance and competitive displacement of invasive plants (Cornelissen et al., 2019;Lehmann et al., 2020;Lu et al., 2013). Many studies have confirmed that climate warming can enhance the competitiveness of exotic invasive plants towards native plants (He et al., 2012;Liu et al., 2017;Wu et al., 2017). Occasionally, it has also reduced an invader's dominance in areas where invasive plants and native plants grow together (Verlinden et al., 2014). Climate warming can also indirectly alter interactions among plants by affecting the interactions between plants and herbivores. Specialist enemies of invaders can expand their distribution ranges under climate warming, and native plants can be released from competition via disproportionate increases in herbivory on the invader. Therefore, the plant community shifts from being dominated by the invading plant to being dominated by native plants (Lu et al., 2016). In ecosystems that have undergone multiple invasions, climate warming may affect exoticexotic interactions in the same way that it affects exotic-native interactions. However, there have not been reports on the effects of herbivores and climate warming on the interactions among invasive plants. In addition to the increase in climate warming, extreme high-temperature events are predicted to become more frequent, more intense and longer in duration (Easterling et al., 2000;Meehl & Tebaldi, 2004 (Francis, 2012).
A. philoxeroides and E. crassipes were introduced to China in the 1930s and 1950s, respectively. Currently, A. philoxeroides has invaded large warm temperate zones, though extremely high-temperature events will slow this replacement process.

K E Y W O R D S
Alternanthera philoxeroides, competition between invasive plants, Eichhornia crassipes, herbivory, temperature rise areas south of the Yellow River Basin, and E. crassipes is prevalent in south of the Yangtze River (Li & Xie, 2002). Both species have invaded various freshwater bodies in China, where they have changed hydrological environments by destroying native animals and plant communities, reducing species diversity, altering ecosystem structure and function and disrupting food chains and nutrient cycling (Lolis et al., 2020;Portela et al., 2020;Tanveer et al., 2018). To control these two noxious weeds, the specialist beetles A. hygrophila (for A. philoxeroides) and N. eichhorniae (for E. crassipes) were introduced to China from the original ranges of the two plants in 1986and 1995, respectively (Ding et al., 2001Ma, 2001

| Studied species
Eichhornia crassipes is a free-floating aquatic plant. It reproduces both sexually and asexually. In invaded regions, it increases in population size mainly through vegetative reproduction, forming new ramets from axillary buds on stolons that are produced through the elongation of internodes. Under suitable conditions, the population of E. crassipes can double between 1 and 3 weeks through the spread of these daughter plants and will form thick, extensive mats that affect their ecosystems (Gopal, 1987). Sexual reproduction rarely occurs, owing to the lack of suitable pollinators and of appropriate sites for germination and seedling establishment in invaded regions (Barrett, 1980). In tropical China, E. crassipes can grow throughout the year. In subtropical regions, the plant dies in the winter and sprouts new plants from axillary buds on the stem base in the following year.
Alternanthera philoxeroides is a perennial herbaceous plant that is both stoloniferous and amphibious. It can grow prostrate along the ground or across the water surface, rooting at the nodes, anchoring to the shore and forming tangled mats that affect ecosystems.
A. philoxeroides expands its population by producing new shoots from each stem node; these shoots can become new plants if they are disconnected. Sexual reproduction in A. philoxeroides is absent in China.
Neochetina eichhorniae originated from South America and is one of the most widely distributed biological control agents of E. crassipes. The adult weevils are nocturnal, and they feed externally (foliage and petiole) within E. crassipes, causing a decline in photosynthesis in E. crassipes. Their eggs are often laid in holes that are chewed into the petiole of E. crassipes by females. N. eichhorniae experiences three larval instar stages. Newly hatched larvae tunnel towards the bases of petioles and into the crowns of the plants, where they excavate small burrows. Water flows into these tunnels and burrows, causing tissue decay and host death. Fully grown larvae exit the crown and move down to the roots to pupate underwater.
The minimum generation time is 10 weeks, and the average lifespan of adults is over 90 days (Julien, 2001).
Agasicles hygrophila is a monophagous flea beetle used as a biological control agent for A. philoxeroides. It is native to the Paraná River basin, Río de la Plata and the maritime borders of Uruguay and southern Brazil. Females deposit an average of 32 eggs in a mass of two parallel rows on the undersides of the leaves of A. philoxeroides (Telesnicki et al., 2011). The three larval instars and the adults feed on the leaves and stems of A. philoxeroides, producing feeding holes and trenches on the leaves. Prepupae chew a circular hole into the internode of the plant and then pupate in the hollow stems. Generally, the period from egg laying to adults hatching is three weeks, and the average lifespan of adults in South China is approximately 6 weeks (Fu, 2007). Generally, 9-20 quadrats (1 m × 1 m) were placed along three to six transects, and more quadrats and transects were placed in communities with high species richness and complex community structures.

| Field surveys
In some communities with low species richness, the distances between the two quadrats and transects were long (30 m, 50 m, 100 m, 200 m); in some communities with high species richness and complex community structures, the distances were shorter (5 m, 10 m, 20 m). Briefly, the quadrats in each community were selected to fully reflect the local community structure as much as possible. For each quadrat, the coverages of A. philoxeroides and E. crassipes were determined by visual estimation on a 22-degree scale: 0.5%, 1%, 5%, 10%, 15%, 20% …100%. Then, all the plants of A. philoxeroides and E. crassipes were collected and dried at 70°C for more than 48 hr to determine the total biomass of each species. Next, the mean coverage and biomass of A. philoxeroides and E. crassipes in each location were calculated.
The defoliation (% of leaf area removed) of A. philoxeroides and E. crassipes by herbivores was also assessed. First, the existence of specialist enemies (N. eichhorniae and A. hygrophila) was determined directly by, searching for the specialist enemies in the communities, or indirectly, by looking for traces of foliage feeding. Then, 10-20 plants of A. philoxeroides or E. crassipes in each quadrat were randomly selected. The damaged leaves on each of the selected plants were checked, and the damage was scored on a seven-degree scale: 1 (the whole leaf was damaged), 3/4 (most parts of the leaf were damaged), 1/2 (half of the leaf was damaged), 1/4, 1/8, 1/16 and 1/32.

The defoliation of each plant (D) was calculated as follows:
where N t is the total number of individual plants and N r is calculated as the sum of the leaf damage scores.

| Common garden experiment
In order to examine the effects of different scenarios of temperature rise on the interactions between A. philoxeroides and E. crassipes under the pressure of herbivory, we adopted the temperature difference caused by the difference in latitudes to examine these effects. Therefore, common garden experiments were established at two sites (Ezhou City, 30.271713°N, 114.568791°E; Xinmi City, 34.463588°N, 113.355276°E) in China. Meanwhile, the same experimental set-up that was established in the natural environments at Ezhou and Xinmi was established in a greenhouse in Ezhou city.
The air temperature during our experiment in the Ezhou greenhouse could reach 50°C at midday on certain days; hence, the Ezhou greenhouse was considered to represent extreme high-temperature events. Ezhou and Xinmi experience a monsoon climate, with hot, wet summers and cold winters. Ezhou is located in a subtropical zone.
Over the past 30 years, the mean annual precipitation was 1402 mm, the minimum mean air temperature was 1.9°C (in January), and the maximum mean temperature was 33.3°C (in July). In contrast, Xinmi is located in a temperate zone. Over the past 30 years, the mean annual precipitation was 640.9 mm, the minimum mean temperature was −3.7°C (in January), and the maximum mean temperature was 31.7°C (in July) (http://data.cma.cn/). During our experiment, the daily average temperature and accumulated light intensity in Ezhou were both higher than those in Xinmi (Figure 1a,c). The average temperature in the greenhouse was 1.86°C higher than that in the natural environment ( Figure 1b). During the experiment, the average temperature in the greenhouse was 33.13°C, and the instantaneous maximum temperature reached 52°C. Therefore, the Ezhou greenhouse experienced extremely high temperatures.
The common garden experiment had a split-plot design, with the experimental site (Ezhou, Ezhou greenhouse and Xinmi) as the whole-plot factor and the culture type (monoculture or mixed culture) and herbivory treatment (no herbivory or herbivory by specialist enemies) as subplot factors within the whole-plot factor (experimental sites). At each site, 36 aquariums (100 cm length × 65 cm width × 70 cm height) were used as the experimental containers.
To simulate the natural shoreline of a waterbody, washed sand was used as the sediment to form a slope; more details of the experimental setup are provided in Figure S1. The total nitrogen (TN): total phosphorus (TP) in the drinking water of Ezhou = 0.71:0.04 mg/L; the TN: TP in the water of Xinmi = 0.564:0.04 mg/L. We added 50 g slow-release Osmocote fertilizer (Osmocote Exact, SCOTTS, USA., containing 16 g TN, 9 g P2O5, 12 g K2O5, 2 g MgO, 0.02 g B, 0.05 g Cu, 0.031 g Fe, 0.02 Mo and 0.014 Zn per 100 g) to the sediment to ensure that sufficient nutrients were available to support plant growth and to eliminate nutrient limitations. The fertilizer was evenly spread on the surface of the sand in each aquarium. The containers were randomly assigned to six treatments: (1) At the end of the experiment, the larvae and adults of A. hygrophila and N. eichhorniae in each aquarium were counted. Then, the defoliation of A. philoxeroides and E. crassipes in herbivory treatments was measured by the method used in the field surveys. Last, the plants were harvested, and the dead ramets of each species in each aquarium were counted. All plants were then washed and dried at 70°C for more than 48 hr. Their biomass was then determined, and the relative growth rate (RGR) (g.g −1 .day −1 ) was calculated as

| Data analysis
For field surveys, we used generalized linear mixed models in the lme4 Package in R to analyse how the average air temperature influenced the biomass, coverage and defoliations of both invasive plants A. philoxeroides and E. crassipes and nested habitat type within quadrat as a random effect. Because the defoliation and coverage are the ratio data of the range from 0 to 1, and we chose the binomial distribution for the response variable of coverage and defoliation. The biomass value is positive real numbers; we chose Gamma distributions for the biomass of both invasive plants. Then, we applied smooth functions to demonstrate the trends of the biomass, coverages and defoliations of A. philoxeroide and E. crassipes to average air temperature in ggplot2 Package in R. Then, ANOVA was used to diagnose the differences in biomass between A. philoxeroides and E. crassipes growing together and in monoculture in each field quadrat, and Duncan tests were used to compare levels within factors for significance (p <.05).
The common garden experiments were performed in a split-plot design, with the experimental site (Xinmi, Ezhou and Ezhou greenhouse) as a whole-plot factor and the culture type (monoculture or mixed culture) and the herbivory treatment (no herbivory or herbivory by specialist enemies) as subplot factors within the whole-plot factor (the experimental site). At each experimental site, all aquariums were randomly assigned to each treatment combination. Then, mixed ANOVAs were applied to test the effects of the experimental site (whole-plot factor), culture type (subplot factor), herbivory treatment (subplot factor) and their interactions on the biomass and RGR of E. crassipes and A. philoxeroides. When significant interactive effects were detected, Duncan tests were used to compare levels within factors for significance (p < .05). Before data analysis, the biomass data were transformed using a Ln (x) function. As all leaves F I G U R E 1 Average daily temperature (a) and daily accumulated light intensity (c) at two common garden experiment sites and the temperature difference (b) and mean temperature difference at different times of day (d) inside and outside of the Ezhou greenhouse [Colour figure can be viewed at wileyonlinelibrary.com] were intact and no ramet death occurred in the control treatments (no herbivory), the defoliation and mortality of each plant species and the abundance of each herbivore were analysed only within the herbivory treatments. We used the Kruskal function within "agricolae" package conducting two-factor nonparametric test to determine the effects of the combination of the experimental site  In the field, with the increase in the average air temperature, the population biomass (R 2 = 0.228, p < .01, Figure 3a) and coverage (R 2 = 0.11, p < .01, Figure 3b) of A. philoxeroides showed downward trends overall and were low between 20°C and 25°C (north of 32°N). Both the population biomass (R 2 = 0.0121, p < .01, Figure 3a) and coverage (R 2 = 0.031, p < .01, Figure 3b) of E. crassipes showed significant increasing trends with the rising of average air temperature. Furthermore, the trends in the total biomass of A. philoxeroides and E. crassipes with average air temperature within their distribution ranges were not significant (R 2 = 0.003, p = .06, Figure S4). In the distribution area where the two plant species co-occur, the population coverage and biomass of E. crassipes were consistently higher than those of A. philoxeroides (Figure 3a,b and Figure 4). Moreover, we also found that the biomass in the plots with only E. crassipes was significantly higher than that in the plots with only A. philoxeroides ( Figure 4). In areas where E. crassipes was absent, the coverage and biomass of A. philoxeroides were higher than those in other regions (Figure 3a,b). Meanwhile, with the increase in the average air temperature, the defoliation of A. philoxeroides showed a gradually rising trend (R 2 = 0.085, p < .01, Figure 3c), while the defoliation of E. crassipes was consistently low and did not show an obvious change trend (R 2 = 0.011, p = .493; Figure 3c).

| Common garden experiment
The experimental site, culture type and herbivory had significant effects on the biomass and RGR of both plant species (Table 1).
The biomass and RGR of E. crassipes and A. philoxeroides were both higher in Ezhou than in Xinmi. However, extremely high temperatures (Ezhou greenhouse) reduced the biomass and RGR of E. crassipes, while, it increased the biomass and RGR of A. philoxeroides (Figure 5a,b). The biomass and RGR of A. philoxeroides were both lower in the mixed-culture and herbivory treatments than in the control treatments (Figure 5a,b). The significant interaction effects of culture type and experimental site and of herbivory and experimental site on the biomass and RGR of E. crassipes suggest that the effects of the culture type and herbivory on E. crassipes differed in the different scenarios of temperature rise (Table 1). Mixed culture and herbivory significantly reduced the biomass of E. crassipes in the extreme high-temperature environment (Ezhou greenhouse) (Figure 5a,b). The significant interaction effects of culture type and herbivory on the biomass and RGR of A. philoxeroides suggest that the effect of herbivory on A. philoxeroides differed between the monoculture and mixed-culture treatments (Table 1). In the monoculture treatments, herbivory reduced the biomass and RGR of A. philoxeroides more intensely than in mixed-culture treatments ( Figure 5a,b). Additionally, the biomass and RGR of E. crassipes were higher than those of A. philoxeroides in all treatments (Figure 5a,b), which indicates that E. crassipes has an advantage in competition with A. philoxeroides.
The abundances of N. eichhorniae and A. hygrophila differed at the different experimental sites ( Table 2). The abundances of N. eichhorniae and A. hygrophila increased with increasing temperature (daily average temperature in Xinmi <Ezhou < Ezhou greenhouse).
We found that the differences between the abundances of N. eichhorniae and A. hygrophila in Xinmi and Ezhou were not statistically significant (Figure 6a,b).
Moreover, there were no significant differences in the abundances of N. eichhorniae and A. hygrophila in the monoculture and mixed-culture treatments (Table 2, Figure 6a (Table 2). E. crassipes ramet mortality in the Ezhou greenhouse was significantly higher than that in Xinmi and Ezhou (Figure 7c). No E. crassipes ramets in the control treatments or A. philoxeroides ramets in any treatment died during the experiment.
The SEM indicated that in environments experiencing climate warming, the climate warming significantly improved the performance of E. crassipes and A. philoxeroides; the abundance of A. hygrophila, the combination of A. hygrophila herbivory and competition from E. crassipes negatively affected A. philoxeroides (Figure 8a). In the extreme high-temperature environment, the relationships were similar to those in the SEM for climate warming. Notably, with the increase in temperature, the effect of temperature on the performance of E. crassipes changed from positive to negative. However, under extremely high temperature, the effect of temperature on the performance of E. crassipes was non-significant ( Figure 8b). abundance of A. hygrophila is high at low latitudes (Lu et al., 2013). Therefore, the defoliation of A. philoxeroides is high, and its performance is limited. A. philoxeroides allocates more resources to growth or reproduction and fewer resources to defence due to its release from specialist enemies at high latitudes (Yang et al., 2021). In addition, our study showed that E. crassipes restricted the population expansion of A. philoxeroides in warmer ranges where both invasive plant species co-occur. A previous study also found E. crassipes can decrease the growth of A. philoxeroides and have a negative effect on its invasion success (Wundrow et al., 2012). Both now and in the future, most of the distribution areas of A. philoxeroides and E. crassipes in the world overlap (Hallstan, 2005), and they also frequently overlapped in the field in this study. Therefore, when studying and predicting the distribution and effects of A. philoxeroides, attention must be paid to the interaction between A. philoxeroides and E. crassipes and not only to climate, natural enemies, environmental conditions or resource availability.

| Effects of herbivory on the growth performances of the two plants and their interactions
Both in the fields where the specialist beetle A. hygrophila occurred and in the A. hygrophila herbivory treatment in the common garden experiments, the defoliation of A. philoxeroides was quite high. A. hygrophila affects the growth of A. philoxeroides not only by damaging its organs and tissues and reducing its photosynthetic area but also by decreasing the nitrogen concentrations and photosynthetic activity of the remaining intact tissue (Yu & Fan, 2018). Therefore,

F I G U R E 3 Population biomass (a), population coverage (b) and defoliation (c) of Alternanthera philoxeroides and Eichhornia crassipes in field surveys (each point represents the average for all quadrats in a location) and their trends with the increase in the average air temperature throughout their distribution in China (bands represent the 95% confidence interval) [Colour figure can be viewed at wileyonlinelibrary.com]
F I G U R E 4 Alternanthera philoxeroides and Eichhornia crassipes biomass (mean ± SE) in co-occurring and not co-occurring field quadrats. Not co-occurring means that the quadrats contained only A. philoxeroides or Eichhornia crassipes; co-occurring means that the quadrats contained both species. Means marked with the same letters were not different according to Duncan tests (p > .05) [Colour figure can be viewed at wileyonlinelibrary.com] A. hygrophila can effectively control the population of A. philoxeroides in China (Lu et al., 2013). In addition, many generalist herbivores in China feed on A. philoxeroides (Li et al., 1990(Li et al., , 2008. Although these generalists cause less morphological and physiological damage to A. philoxeroides than co-evolved specialists, they can also limit the population spread of A. philoxeroides and eliminate the competitive advantage of A. philoxeroides over native plants Fan et al., 2016;Yu & Fan, 2018).
The defoliation of E. crassipes was low in the field even in the areas where the specialist N. eichhorniae occurred. In the natural environment of the common garden experiment, the defoliation and mortality of E. crassipes ramet caused by N. eichhorniae were also relatively low. Notably, the distribution area of the specialist in China is much smaller than that of its host plant, E. crassipes. Although studies have reported that N. eichhorniae can control E. crassipes (Ding et al., 2001;Julien, 2001;Reddy et al., 2019), the control effects TA B L E 1 The effects of culture type, herbivory, experimental site and their interactions on the biomass and RGR of Eichhornia crassipes and Alternanthera philoxeroides

F I G U R E 5
Plant biomass (a) and relative growth rate (b) (mean ± SE) of Eichhornia crassipes and Alternanthera philoxeroides with and without biological control agents growing in monocultures or mixed cultures at different experimental sites. Means with the same letters were not different according to Duncan tests (p > .05). The biomass data were transformed using a ln(x) function [Colour figure can be viewed at wileyonlinelibrary.com] depend on the water nutrient quality, winter temperatures and interference from herbicide operations (Coetzee et al., 2011). Some generalist enemies in China can feed on E. crassipes, but the population of E. crassipes cannot be well controlled by these enemies . The higher leaf nitrogen content and thinner foliage of A. philoxeroides result in its palatability being higher than those of E. crassipes Yu & Fan, 2018), and generalist enemies will prefer A. philoxeroides over E. crassipes (Zhao et al., 2014). In addition, E. crassipes exhibits high tolerance to herbivory and can fully or even excessively compensate for light defoliation by improving its photosynthetic rates and changing its resource allocation strategy. Simulated herbivory experiments have confirmed that only heavy defoliation (80% clipping leaf) can reduce the growth performance of E. crassipes (Lyu et al., 2016;Soti & Volin, 2010). Based on our field observation, the natural enemies could not produce such the heavy defoliation, only caused approximately 2% defoliation. In the native range of E. crassipes, herbivory also has minimal effects on the coverage and biomass of E. crassipes and its ability to maintain growing populations (Adis & Junk, 2003;Franceschini et al., 2010;Gutiérrez et al., 2001). Our study speculate that herbivory facilitates the E. crassipes suppressed the population of A. philoxeroides, and the enemy release hypothesis cannot effectively explain the success of E. crassipes invasions, and our research also confirmed the speculations of other studies (Lolis et al., 2020).
In systems with multiple exotic species, it is impossible to intro- Myriophyllum aquaticum and Azolla filiculoides) declined after the introduction of these insects; however, the establishment and spread of submerged and emergent invasive plant species increased and still pose significant threats to aquatic ecosystems (Hill et al., 2020).
Furthermore, introduced biocontrol agents may also feed on native species (Lu et al., 2015).

| Effects of different scenarios of temperature rise on the growth performances of the two plants and their interactions
In the common garden experiment, the performances of both the plants and the populations of their enemies were better at the subtropical sites than at warm temperate sites, but the interactions among plants and between plants and their enemies were not altered. The overlapping range of the subtropical and warm temperate zones is the same as the northern margin of the distribution of E. crassipes in China, which cannot overwinter in environments with extreme cold temperatures or ice coverage (Madsen et al., 1993;Tyndall, 1982). Over the past decades, winters have become warmer in China, with both the maximum and minimum temperatures increasing (Zhou et al., 2004). In the future, climate warming will continue, especially in winter and at higher latitudes. An increase in temperature in winter can increase the survival rate of E. crassipes (You et al., 2013) and promote the expansion of the distribution of E. crassipes to higher latitudes (Hellmann et al., 2008;Rahel & Olden, 2008). Likewise, climate warming promotes the northward expansion of the distribution of A. philoxeroides (Lu et al., 2013).
Hence, we speculate that E. crassipes will replace A. philoxeroides as the dominant species in some communities in warm temperate zones. However, A. philoxeroides will also expand its range farther to the north in response to climate warming.
Extreme high-temperature events can affect competition among plants and trophic interactions, leading to complex responses at the community level (Sentis et al., 2013;White et al., 2001) and facilitating biological invasions (Diez et al., 2012).  (Julien, 2001). The increase in average air temperature gradually maintained growing N. eichhorniae population, especially extremely high-temperature environment. And, the large number of the specialists had negative effect on its host plant E. crassipes (Ding et al., 2001;Julien, 2001). Furthermore, the interaction of herbivory and culture style has significantly affected A. philoxeroides biomass and RGR. The reason is that herbivory from A. hygrophila have significantly negative on the biomass of A. philoxeroides, the result has been proved by other many studies (Lu et al., 2013(Lu et al., , 2016 Finally, in the common garden experiment, temperature rise is beneficial for the biomass of A. philoxeroides. But based on the result of field surveys, the biomass decreased with the increase of temperature rise. In other words, our common garden experiments and field surveys showed opposing trends. The reason is that the interaction of many environmental factors determines the distribution of A.philoxeroides, not just the temperature and interspecific competition from one species.

| Implications for the processes and management of biological invasion
The existence in interspecific competition between invasive plant species demonstrates that invasive plants can resist the invasion of other exotic plants just as native plants do. It also demonstrates that some invasive plants will be replaced by more powerful invasive plants as the dominant species in certain communities. Our study found that not every biological control agent can control its host effectively within the invasion range and that one biological control agent cannot control its host effectively in all areas of the host range.
These results indicate that there is a mismatch in the biological dis- Therefore, the introduction of enemies cannot completely eliminate plants invasions in freshwater ecosystems, and a holistic approach to controlling invasive plants will be required (Hill et al., 2020). In addition, our findings imply that predictions of the distribution, performance and effects of invasive plants in the future should be linked to their interactions with enemies and other invasive plants and to climate warming, especially extreme climate events. Fan for their valuable work in the field and the laboratory. We also thank our field guides Baobo Zhao and Mingcao Zhai.

CO N FLI C T O F I NTE R E S T
We have no conflict of interest to disclose.

PEER R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1111/ddi.13406.

DATA AVA I L A B I L I T Y S TAT E M E N T
The raw data required to reproduce these findings cannot be shared at this time as the data also form part of an ongoing study.

B I OS K E TCH
Chunhua Liu is a freshwater ecologist and a professor currently at the Wuhan University. She is mostly working on aquatic invasive species and freshwater ecosystems, and often use aquatic macrophytes and related organisms to test ecological concepts on biological invasions and biodiversity, conduct both a series of laboratory and field studies to evaluate the effect biological invasions on native freshwater ecosystem.

S U PP O RTI N G I N FO R M ATI O N
Additional supporting information may be found online in the Supporting Information section.