Introduced non-native mangroves express better growth performance than co-occurring native mangroves

Mangroves are salt-tolerant woody species occurring in tropical/subtropical coastal habitats. Plantation of fast-growing non-native mangrove species has been used as a tool for mangrove restoration/reforestation in several countries. However, the fast-growth ability can make recently introduced species invasive as they can possibly replace co-occurring native mangroves through expressing higher growth performance and phenotypic plasticity. Therefore, quantifying growth differences between native versus non-native mangrove species is important for forest ecology and management. In this meta-analysis, we compared the growth performance of non-native and native mangrove species pairs by analysing all available results in the literature (33 studies). We found that non-native mangrove species performed better than co-occurring native mangrove species in their introduced regions (Log response ratio = 0.51 ± 0.05) and they also expressed higher trait plasticity. Therefore, these species can be potentially invasive owing to their greater competitive advantage. However, the growth difference was diminished at higher latitudes where native mangrove species seem to perform as well as non-native mangrove species do. This is the first meta-analysis on the growth response of mangroves and it has consequential management implications. We suggest that planting of non-native mangrove species should be avoided and their spread should be monitored.

. Detailed information on introduced woody species invading different mangrove ecosystems in the world. Note: References were listed correspondingly depending on the introduced region.
Non-native woody species can also be more competitive than native woody species by expressing greater phenotypic plasticity in biomass production which can be important in tree invasions [28][29][30] . Similarly, high phenotypic plasticity in biomass allocation in mangroves can be very beneficial for seedling establishment 31 . High plasticity in plant traits can allow mangroves to express higher tolerances to low temperature, aridity, salinity and nutrient scarcity 32 . Therefore, mangrove species with high phenotypic plasticity are more likely to adapt climate change by expressing optimal responses in changing environmental conditions. Functional traits are biochemical, morphological, physiological, phenological traits which can affect the performance and fitness of an individual 33,34 . Variation in a functional trait of an individual can determine the performance and fitness of the individual and its population which, in turn, can affect the community and ecosystem-level processes 35 . Therefore, scaling up from trait level responses to the ecosystem level is possible 36 . For example, specific leaf area (SLA) in woody species is positively correlated with the niche breadth of a species 37 .
Functional traits can also explain the coexistence of plants more than environmental conditions (e.g., rainfall) and evolutionary relationships (e.g., phylogenetic distance) 38 . Therefore, intraspecific trait plasticity can be a key factor to understand species fundamental niche and distribution 35 as well as the invasive potential of a species 39 . For example, several studies indicated that higher values in some traits can be used as indicators of invasive tree species ( Table 2). As a result, understanding variation in functional traits across non-native versus native mangrove species especially under changing environmental conditions caused by global climate change can be crucial for mangrove conservation.
Meta-analysis is one of the valuable methods to compare the growth performance of non-native and native mangroves 40 . Trait variation and plasticity comparisons between native versus non-native plants have been mostly studied in invasive herbaceous plants using meta-analyses 9,41-43 . However, woody invasive species were rarely examined (see Erskine-Ogden et al. 2016 for Mediterranean tree species) and, particularly no mangrove species were included in these meta-analyses. Although there are meta-analyses on mangrove productivity 44 , economic value 45 and carbon budget of mangroves 46 , there are no meta-analyses on the growth performance of mangrove species in the existing literature.
There are only a few number of experimental studies simultaneously compared the trait responses of both native and non-native mangrove species although biological invasion in mangroves had gained a lot of attention. Some studies focus on field surveys of non-native mangrove species, whereas others measured the response of either native or non-native mangrove species [47][48][49][50][51][52] . Quantifying growth differences between native and non-native species pairs under climate change is a knowledge gap in the literature. To address this knowledge gap, we compared the growth response of co-occurring native and non-native mangrove species across different regions by examining several key performance traits in a meta-analysis. We specifically asked: Do introduced non-native mangrove species express better growth performance than co-occurring native mangrove species in their introduced regions? Is there a latitudinal gradient in growth difference between non-native and native mangrove species? Moreover, we examined the current global status of introduced species in mangrove habitats and possible mangrove responses to climate change.

Methods
Literature search and study selection criteria. Quantitative and qualitative data were extracted from the literature using ISI Web of Science and China Knowledge Resource Integrated Database (CNKI) in 2019. Several keywords were used to find the most related papers for our meta-analysis including mangrove*, native*, naturalized, indig*, non-native*, nonnative*, alien*, nonindig*, non-indig*, exotic*, invasive* and combinations of these keywords. Following criteria were selected to be included in the analysis: (i) Comparative empirical Table 2. Detailed information on studies comparing invasive and native woody species in terms of growth performance. Note: Relative growth rate (RGR), specific leaf area (SLA), net assimilation rate (NAR), leaf area ratio (LAR), leaf mass ratio (LMR), leaf mass to area ratio (LMA), total leaf area (A L ), net CO 2 assimilation (A), leaf tissue construction cost (CC), water use efficiency (WUE), acclimation index (AI) and root:shoot ratio (RSR). studies must have at least two co-occurring mangrove species (e.g., a native and non-native species). (ii) Each study must measure at least one trait (either morphological or physiological trait). (iii) Each study must be at the species-level (community-level studies were excluded). "Native" or "non-native" groups were determined based on the species information presented in each publication. For example, Sonneratia caseolaris is native to Hainan, China but introduced to other regions in China 53 . Therefore, it was regarded as a native species in Hainan but non-native in Shenzhen, China.
In each study, a non-native mangrove species was paired with a co-occurring native mangrove species which was presented in the same study. In the case of several native and non-native species in a single study, several pairs were included in our analysis. These species pairs were regarded as independent data points 43 as each study has a unique response ratio including biological and environmental variation 54 . The use of co-occurring species pairs, instead of different species from separate studies, provides more realistic comparisons between non-native and native mangrove species because the data will be from the same experimental conditions (e.g., the same treatment and trait response across species).
Extracted data and effect size. Quantitative data belonging to several traits, representing the growth and photosynthetic performance of mangrove species, were extracted (e.g., relative growth rate-RGR, tree height, specific leaf area, net photosynthetic rate) from tables, figures and Supplementary Files/appendices (if there are any) as these traits can be predictors of invasions [55][56][57] (see also Table 2). The measuring tool in PDF-XChange Editor was used to detect values inside figures. Extracted data included species names, the origin of species (latitude), experiment type (field, survey, or greenhouse), treatment type (e.g., inundation, light, or salinity) and experimental conditions (e.g., habitat). To prevent pseudo-replication and include non-biased effect sizes, we extracted only one morphological and one physiological trait data from studies which met the criteria, and we preferably selected growth performance traits (e.g., RGR and net photosynthetic rate).
Log response ratio (LRR), that is widely used in meta-analysis studies, was preferred as an effect size to quantify the overall differences between growth differences between species pairs 58 : where X non is the mean value in a growth trait of a non-native mangrove species and X nat is the mean trait value in a growth trait of a native mangrove species. The use of LRR allows growth performance comparison between species pairs to assess which species (non-native or native) performs well in particular environmental conditions (e.g., under high salinity or low light). If the overall value of LRR is greater than zero (lower 95% confidence interval), growth performance in non-native woody species is greater than that in native species and vice versa. If LRR overlaps zero, there is no overall growth difference between non-native and native species. phenotypic plasticity calculation. In a subset of our dataset, we also used a plasticity index (PIv) 59 to quantify phenotypic plasticity values in growth performance traits (the same traits used for the LRR) of native and non-native mangrove species pairs in China: where X c is the trait mean under control treatment; X t is the trait mean under a treatment (e.g., salinity or light). PI V is a common index that allows plasticity comparisons across species by illustrating the overall absolute change in trait means 60,61 . PIv values are always between 0 and 1 (representing no plasticity and maximum plasticity, respectively).

Statistical analyses.
A two-way ANOVA was used to determine the effect of trait type, experiment type and their interaction on trait responses of mangrove species. In a Tukey's honestly significant difference (HSD) test, significant differences between trait, experiment type and their interaction were detected. In a paired sample t-test, non-native mangrove species were paired with a native species in each study and P-values were calculated at 0.05 level. Regression analyses were used to determine the relationship between the mean log response ratio (LRR) values and latitudes. All statistical analyses were performed in JMP software v. 13

Results
Dataset and trait responses in mangroves. We extracted data from four non-native mangrove species belonging to two families (33 studies) which met our selection criteria to compare the growth performance of co-occurring native and non-native mangrove species. In our dataset, there were 220 data points including nine traits, seven treatment types and 74 different experimental conditions (Appendixes A and B). In each study, non-native mangrove species were matched with a co-occurring native mangrove species. In total, there were 21 different combinations of non-native versus native mangrove species comparisons (Kandelia candel in China is recognized as Kandelia obovata 13,62 ) (Appendix C). We found a significant, positive LRR value (0.51 ± 0.05) across all studies ( Fig. 1 and Appendix A). That is, recently introduced non-native mangrove species expressed significantly higher values in growth-related morphological traits (e.g., RGR and leaf area) compared to co-occurring mangrove species. The growth performance difference between these species was much more pronounced in morphological traits than in physiological traits (Fig. 1). Similarly, in field and greenhouse experiments this difference was more apparent than in field survey studies. Moreover, in high latitudes (above 23°N), the effect size was much lower than in other latitude categories.

Scientific RepoRtS |
(2020) 10:3854 | https://doi.org/10.1038/s41598-020-60454-z www.nature.com/scientificreports www.nature.com/scientificreports/ Effect of trait and experiment types on mangrove responses. A two-way ANOVA test indicated significant differences between trait types and trait by experiment interaction (Appendix D). There were strong differences between studies comparing non-native and native mangrove species in terms of trait types and trait by experiment interaction (Appendix D). There were no greenhouse studies using mangrove species collected from low latitudes (less than 21°N) and therefore a three-way ANOVA test was not possible to run for further analyses. Different Tukey's HSD tests indicated that morphological traits are significantly greater compared to physiological traits (Appendix E). In survey studies, LRR values were lower than in field and greenhouse experiments but the difference was not statistically significant. Moreover, greenhouse experiments measuring morphological traits found much higher values than greenhouse experiments measuring physiological traits.
At the species level, the highest mean LRR values were in Lumnitzera racemosa which was introduced to Florida, USA (Table 3), followed by Laguncularia racemosa, Sonneratia apetala and S. caseolaris, respectively. In the dataset, most data points belonged to Sonneratia apetala which has been well-studied in South China. We also examined mean LRR values through regression analysis at the species level across latitudes in South China by taking averages of LRR values for each species in each location and experiment type (Fig. 2). As survey studies provide more realistic data from the field, experiment types were presented separately in the figure (survey versus field and greenhouse experiments). Interestingly, as the latitudinal gradient increased, the growth difference between non-native and native mangrove species tended to decrease in survey studies (Fig. 2). trait plasticity comparison between native and non-native mangroves. We compared phenotypic plasticity of recently introduced L. racemosa and S. apetala with that of native species using a subset of our dataset (Appendix F) which contains only empirical studies that conducted a control and treatment on co-occurring mangrove species in China. Plasticity index values (PIv) in native species were calculated using arithmetic means of different native species presented in each study (e.g., Laguncularia racemosa versus Rhizophora stylosa and Kandelia obovata). We found that, on average, recently introduced non-native mangrove species expressed higher trait plasticity than native mangrove species in China (Table 4); however, the difference was not statistically significant according to paired sample t-test results.

Discussion
performance of invasive and native mangrove species. High trait values in relative growth rate (RGR) and specific leaf area (SLA) are features of invasive tree species in the Mediterranean climate and these traits can be predictors of biological invasions 56 . In this study, we selected such key growth performance traits to assess the invasiveness potential of introduced mangrove species. Our results indicated that non-native mangrove species expressed higher values in growth performance traits (including RGR, height and photosynthetic traits) compared to co-occurring native mangrove species (LRR = 0.51 ± 0.05; Fig. 1). Therefore, these introduced   www.nature.com/scientificreports www.nature.com/scientificreports/ mangrove species can be potentially invasive as a higher competitive advantage can favor biological invasions through better ability to benefit from fluctuating resource availabilities triggered by climate change (e.g., invasive tree species and associated traits listed in Table 2).
Invasion of new habitats (e.g., mangroves) depends on how fast populations can adapt to changing environmental conditions 63 . The competitive ability of non-native species can determine their invasion success 39 . As taller trees can intercept more light, tree height affects the competitive ability. If non-native species grow faster and perform better than co-occurring native species do by expressing higher phenotypic plasticity 43 , they can outcompete and even replace existing native mangrove species as global environmental changes can favor the spread of invasive plants 9 . In the case of non-native S. apetala invasion, forest structure can be significantly affected after the invasion and mangrove ecosystem functioning (e.g., carbon accumulation) may be changed 64,65 .
In this meta-analysis, most of the data points were from two recently introduced mangrove species (Sonneratia apetala and Laguncularia racemosa) in China. L. racemosa was introduced from La Paz, Mexico where is in an arid climatic zone and freshwater availability can increase its growth and productivity 66 . In Gulf of Mexico, low rainfall and high salinity prevent the movement of mangroves to higher latitudes 67 . However, at its site of introduction (Hainan, China), the mean annual rainfall is almost 9 times higher and salinity is 3 times lower than in La Paz, Mexico [68][69][70] . L. racemosa can outcompete native mangrove species (Rhizophora apiculata) in Hainan and it is reported to be spreading outside afforestation sites 16 . Moreover, L. racemosa can express high trait plasticity under environmental stresses including salinity and shade 71 . Therefore, L. racemosa can shift its range towards higher latitudes in China where salinity is lower and become potentially invasive owing to fast-growing characteristics, high germination rate, buoyant propagules, cold resistance, the ability to form dense monospecific stands 70,72 . These characteristics make L. racemosa a model species for afforestation but invoke questions regarding its invasive potential as it has frequently been used in mangrove afforestation since 2002 in China.
Sonneratia apetala is also an introduced fast-growing pioneer species (from Bangladesh) that can replace native species through competitive exclusion as it can grow better than native species 73 . In recently reforested areas in China, introduced S. apetala occupies approximately 50% of the replanted mangrove areas 13 . Moreover, the efficiency of S. apetala in using photosynthetic energy is higher than that in native mangrove species which can make the species invasive in South China 74 . Two mangrove species from the Indo-Pacific region planted in the US in 1940 have been invading tropical Atlantic forests 12 and the same situation might be the case for L. racemosa in China whose origin is Atlantic East Pacific biogeographic zone.
When an invasive species is introduced to a new habitat, existing native plants may be more vulnerable to allelopathic chemicals originating from invasive species 75 because such biochemicals produced by a mangrove species can prevent the germination, growth and survival of other mangrove species 76 and therefore, determine mangrove forest succession (e.g., Kandelia species 77 ). For example, allelopathic properties of introduced mangrove species such as L. racemosa and S. apetala may reduce the germination rate of seeds belonging to native mangrove  Table 4. Paired sample t-test results of phenotypic plasticity index (PIv) for species pairs. Note: In each species pair group, a non-native species (e.g., L. racemosa or S. apetala) was matched with co-occurring native mangrove species. N is the sample size and SE stands for standard error. (2020) 10:3854 | https://doi.org/10.1038/s41598-020-60454-z www.nature.com/scientificreports www.nature.com/scientificreports/ species 78 and therefore, affect the mangrove ecosystem functioning. As our meta-analysis indicated introduced mangrove species can outperform native mangrove species in their introduced regions, the invasive potential of non-native mangroves should be considered during mangrove afforestation and conservation management.
Mangrove responses to climate change. We found that there might be a negative correlation between growth performance differences in mangroves and the latitudinal gradient in China (Fig. 2). In higher latitudes, the growth difference between non-native and native mangrove species becomes ambiguous as native mangrove species seem to perform as well as non-native species. One of the reasons behind this trend might be the adaptation to low temperatures in winter 79 . At high latitudes, naturally occurring native mangroves are expected to perform better than recently introduced species because biochemical adaptations to colder temperatures would be necessary 80 , and these adaptations can be costly for non-native species in terms of competitive ability 81 . However, as species at high latitudes experience a reduced number of freezing events as a result of climate change, cold adaptation in native species may be maladaptive.
Mangrove species expressing high phenotypic plasticity are expected to be favoured under global climate change because plasticity can allow cold tolerance and being widespread 82 . We found that introduced mangrove species in China (S. apetala and L. racemosa) express higher plasticity than co-occurring native species do (Table 4). Therefore, they might be more successful in adapting to changing environmental conditions. Therefore, findings of this meta-analysis can also predict the distribution of mangrove species and possible range shifts in mangrove species under climate change.
If certain native mangrove species cannot express appropriate performance under changing environmental conditions, they might become extinct or move to different habitats to alleviate the detrimental effects of global climate change 32,83 because mangroves can be highly sensitive to changes in climate 84 . Global climate change-induced cyclones, tsunamis, heat waves, and climatic extremes cause significant changes in mangrove mortality and recovery 85 . Such events can reshape species composition and species abundance within mangrove ecosystems, and understanding mangrove responses to these changes is essential for mangrove ecosystem services. potential implications for forest management. This study is the first meta-analysis in the literature on the growth responses of mangrove species and it has far-reaching implications for mangrove management (e.g., conservation and restoration). Interestingly, only a very limited number of studies carried out experiments using multispecies including non-native and native mangrove species pairs although several woody species were introduced to different mangrove ecosystems across the world (Table 1). Most studies on mangroves in the literature report the presence of invasive species (e.g., references in Table 1). Therefore, findings of this meta-analysis may be somewhat limited by the lack of available data and low sample size (e.g., low R 2 in Fig. 2). However, these findings can address the situation of mangroves in South China. Due to the use of non-native mangrove species (e.g., S. apetala) in afforestation in the past 30 years, the invasiveness of such species has been detected using either surveys or experiments 13,53 . The findings of our meta-analysis suggest the invasive potential of non-native mangrove species based on their superior growth performance compared to native species.
The relationship between growth performance at early life stages and invasiveness can be evaluated by the probability of dominance and dispersal which are two important aspects in predicting non-native species invasion in forests 86 . Most data in the present meta-analysis were from experimental comparisons on the growth and stress-tolerance of native and non-native species pairs, reflecting the early establishment of plants. According to our effect size (LRR), these non-native species performed better during the early establishment stages. Once they survive and dominate new habitats, they can be potentially invasive.
The early life stage is also an appropriate period to reduce the biological invasions by manipulating the degree of disturbance or productivity in forest management 87 . Shading, inundation or harvesting is the major measurements to control the invasiveness of a species 88,89 . Understanding how the introduced invasive and native species respond to changing environments will be a necessity for humankind to preserve robust ecosystem functioning. Therefore, in the future, more empirical studies are needed to better examine mangrove responses to biological invasions, particularly under global climate change.
The findings of this meta-analysis can be useful to predict the distribution of mangrove species and possible range shifts in mangrove species under climate change. On the global scale, climate change is a major challenge for woody plant invasion management which might accelerate the dispersal ability and growth performance of non-native species 90 . Our meta-analysis also examined stress tolerance to several environmental factors, such as inundation, salinity, light and temperature (Appendix A). As positive LRR values indicate, better performance of current non-native mangrove species can possibly lead to invasion. For instance, the poleward range limit of non-native S. apetala in China is around 28°N without reproduction, whereas it is 26.5°N with reproduction 91 . On the contrary, if certain native mangrove species cannot express better performance than non-native species (low LRR value) under changing environmental conditions, they might go extinct or move to different habitats to alleviate the detrimental effects of global climate change.
Our study also draws attention to the use of non-native mangrove species in afforestation practices and mangrove conservation management across the globe because introduced non-native species can outperform native mangrove species in their introduced ranges. We suggest that the invasion potential of non-native species should be considered and the use of introduced non-native mangrove species should be avoided in mangrove afforestation practices. Although ecological repercussions of non-native mangrove species invasion in mangrove forests are uncertain, such introduced non-native species should be monitored for their potential spread. Restoration and management of mangroves are crucial for socio-economic and scientific aspects which can be achieved through a better understanding of how mangrove species respond under global climate change and biological invasion. More empirical studies in the future can enhance our understanding of mechanisms behind invasions and possible methods to control the invasion in mangrove forests.