High risk of plant invasion in the understory of eucalypt plantations in South China

Eucalypt plantations expand rapidly out of their natural distribution zones, thus inducing a concern on their effects on biodiversity and ecosystem functions. We compare the understory plant diversity of 46 plots of eucalypt plantations, including early and later stages in rotation, with that of 21 plots of contrast vegetation, including other types of plantations and secondary shrub grassland, in Guangdong and Guangxi Provinces, South China. Although the overall plant diversity did not change significantly in eucalypt plantations relative to the contrast vegetation, the community structures changed dramatically. The Asteraceae family, which is the most important source of destructive invasive plant species in China, is ranked 3rd (7.42%) and 7th (3.14%) in species importance in the early and later stages in eucalypt plantations, respectively. Nevertheless, Asteraceae is ranked 15th (1.73%) in other types of plantations and 21st (0.94%) in secondary shrub grassland. Significant increases in the richness and frequency of invasive species were also observed in eucalypt plantations. Among the 20 invasive species recorded in the eucalypt plantations, 9 species were destructive invasive species and 7 of these species belonged to Asteraceae. This study highlights an enhanced plant invasion risk in eucalypt plantations in South China, particularly by Asteraceae.

Evidence illustrates that biodiversity contributes to ecosystem functioning 1,2 and that a loss of biodiversity alters ecosystem processes that are essential to the productivity and sustainability of ecosystems 3,4 . Natural forests, which provide habitat to the richest terrestrial biodiversity in the world, have been shrinking in size. This environmental change is largely attributed to anthropogenic causes, the most significant of which is deforestation to meet wood supply [5][6][7] . Plantation forests generally maintain lower biodiversity levels than natural forests [8][9][10] ; however, developing high-yield plantation forests in a sustainable manner may help preserve natural forests and biodiversity.
Eucalypts comprise a group of approximately 800 woody species 11 that belong to Eucalyptus, Corymbia or Angophora in Myrtaceae 12 , which are mostly native to Australia but not in China. Eucalypts are among the most fast-growing trees in the world; they adapt well in warm environments and provide materials for timber and pulp 13 . In South China, eucalypt plantations have expanded rapidly in the last decade, from 2.54 M ha in 2008, to 3.6 M ha in 2011 14 and to 4.40 M ha in 2013 15 . According to the Chinese State Forest Administration, eucalypt plantations in China constituted 22% of the world's eucalypt plantation area in 2013 and contributed 25% to Chinese wood production 15 . However, despite the remarkable growth of wood production in China, the self-sufficiency rate has declined in recent years (see Supplementary Fig. S1 online) mainly because of the rapid increase in wood consumption. The sustainable development of fast-growing eucalypt plantations is supported by the Chinese government, so as to enhance wood production and to preserve natural forests.
The expansion of eucalypt plantations out of their natural distribution zones has generated concerns about their effects on biodiversity and ecosystem functions. Previous studies have found that the understory plant diversity of eucalypt plantations are important for reducing nutrient loss and soil erosion 16 and maintaining soil microbial communities 17 ; such diversity may have positive effects on the productivity of eucalypts 18 . Eucalypt plantations in South China have been found to maintain a lower diversity of understory plant species than natural forests 9,19 . However, other types of plantations, such as Pinus massoniana and Dimocarpus longan plantations, and secondary shrub grasslands are more common than natural forests in South China, and theses may serve as potential field sources of eucalypt plantations. Whether eucalypt plantations maintain less understory plant diversity than other common types of vegetation remains unclear. Furthermore, previous studies usually focus on diversity indices. However, community structure, such as the relative importance of species or family, typically contains important messages and provides links with the ecological function of community. In another aspect, according to the theory of fluctuating resource availability proposed by Davis et al. 20 , the disturbance of understory communities during plantation, fertilization and short rotations of eucalypt plantations may provide opportunities for invasive species to capture light, water and nutrients. Nevertheless, questions like whether eucalypt plantations have more invasive species understory than other common types of vegetation and which environmental factors may affect the invasion risk have been rarely investigated.
We first compare the understory plant diversity, including diversity indices and community structure, in eucalypt plantations with those in other common types of vegetation in South China. Thereafter, we test if more invasive species exist in eucalypt plantations than in contrast vegetation. Finally, we detect the environmental factors that may affect plant invasion in both eucalypt plantations and contrast vegetation.

Results
Understory plant diversity of eucalypt plantations. We compared species richness, phylogenetic diversity, Shannon's index and Pielou's evenness index among four groups, namely, eucalypt plantations with growth year 1-4 (Euc14), eucalypt plantations with growth year 5-8 (Euc58), other types of plantations (CK_a) and secondary shrub grassland (CK_b) by using Tukey's multiple comparisons (Fig. 1). No significant differences in the aforementioned indices were found among the four groups except for the significant increase in Pielou's evenness index in the two groups of eucalypt plantations (i.e., Euc14 and Euc58) compared with that in secondary shrub grassland (both P < 0.05).
The 10 most important plant species were identified by averaging the importance values of each species across plots within each of the 4 groups ( Table 1). The rank of species importance demonstrated that Dicranopteris pedata of Gleicheniaceae was the most important species within each group. However, its importance was lower in eucalypt Shannon's index (c) and Pielou's evenness index (d) among two groups of eucalypt plantations (eucalypt plantations with growth years 1-4 (Euc14) and 5-8 (Euc58)) and two groups of contrast vegetation (common types of plantation (CK_a) and secondary shrub grassland (CK_b)). Groups sharing the same letter were insignificantly different at the 95% confidence level by Tukey's multiple comparisons. For each box-and-whisker plot, the box shows 25% median and 75% quantile of the given values. The whiskers extend to the most extreme data points that are not more than 1.5 times the interquartile range (length of the box) from the box. . Chromolaena odorata of Asteraceae, an invasive plant species, is ranked 9 th in species importance in Euc14. The rank of family importance revealed that the importance of Gleicheniaceae was lower in eucalypt plantations than in contrast vegetation (Table 2). Furthermore, Poaceae and not Gleicheniaceae was the most important family in Euc14. Moreover, the importance of Asteraceae was remarkably higher in eucalypt plantations than in contrast vegetation. Asteraceae ranked 3 rd (7.42%), 7 th (3.14%), 15 th (1.73%) and 21 st (0.94%) in Euc14, Euc58, CK_a and CK_b, respectively. Diversity of invasive species. Species richness and invasive species importance were compared between eucalypt plantations and contrast vegetation by using Tukey's multiple comparisons (Fig. 2). Although no significant differences among Euc14, Euc58, CK_a and CK_b were observed, significant increases in both invasive species richness (P < 0.05) and importance (P < 0.10) were noted in the combined eucalypt plantations (Euc) compared with those in the combined contrast vegetation (CK). Moreover, the frequency of invasive species that occurred in a plot was considerably higher in eucalypt plantations than in contrast vegetation: 79.31%, 47.06%, 15.38% and 25.00% in Euc14, Euc58, CK_a and CK_b, respectively.
A total of 20 invasive species were recorded in this study (Table 3), and all of them were noted in eucalypt plantations. C. odorata, Bidens pilosa and Praxelis clematidea were the 3 major invasive plants in eucalypt plantations; all of which belong to Asteraceae. Among the 9 destructive invasive species recorded, 7 belong to Asteraceae. Moreover, the importance of Asteraceae species among the total importance of invasive plants across plots was higher in eucalypt plantations than in contrast vegetation: 89.45%, 85.91%, 68.81% and 59.86% in Euc14, Euc58, CK_a and CK_b, respectively.

Environmental factors for plant invasion. Pearson's correlation between invasive species richness and
importance, and 6 environmental factors were performed across 46 plots of eucalypt plantations and 21 plots of contrast vegetation ( Table 4). The results indicated that invasive species richness in eucalypt plantations decreased with growth year after plantation (P < 0.05), canopy coverage (P < 0.05) and elevation (P < 0.10) but increased with annual total radiation (P < 0.05) and mean annual temperature (P < 0.1). Invasive species importance exhibited a significant positive correlation with total precipitation per year (P < 0.05). However, no significant correlations were found in contrast vegetation.

Discussion
Our results revealed that species richness or phylogenetic diversity does not decrease in eucalypt plantations compared with that in contrast vegetation in South China. The high Pielou's evenness index in eucalypt plantations suggests that the abundance of understory plants was more even among species in eucalypt plantations than in secondary shrub grassland. This finding concurs with the phenomenon that the importance of the dominant species (i.e. D. pedata) and Gleicheniaceae family decreases in eucalypt plantations. Considering the biodiversity indices, no reduction in understory diversity was found in eucalypt plantations compared with other common types of vegetation in South China. However, given the change in community structure and the enhanced importance of invasive species, the ecosystem function of eucalypt plantations may have been damaged. The analysis of community structure illustrated that the importance of the Asteraceae family increase remarkably in eucalypt plantations, particularly in Euc14, which is the early stage of rotation. Moreover, different from contrast vegetation, significant increases in invasive species richness, importance and frequency were observed in eucalypt plantations. Asteraceae comprises a major part in the total number of destructive invasive species and in the total importance of invasive plants in eucalypt plantations. According to a previous research, Asteraceae is the most important source family of invasive plants in China 21 , contributing 52.9% and 18.7% of destructive invasive species and total invasive species in China, respectively 22,23 . Hu et al. 9 indicated that P. clematidea (under the name Eupatorium catarium) and C. odorata were important understory species in eucalypt plantations in Hainan Province, China. Our results suggest an enhanced risk of plant invasion in eucalypt plantations in South China, specifically by Asteraceae species.
Furthermore, invasive species richness and importance in the understory are remarkable in Euc14 but tend to decrease in Euc58 as the growth year and canopy coverage increase. This finding concurs with the prediction using the theory of fluctuating resource availability 20 . The causes of the enhanced plant invasion risk in eucalypt Figure 2. Comparison of invasive species richness (a) and importance (b) within understory communities between eucalypt plantations and contrast vegetation. Groups sharing the same letter were insignificantly different at the 95% confidence level by Tukey's multiple comparisons. Euc stands for eucalypt plantations combined with Euc14 and Euc58, and CK stands for contrast vegetation combined with CK_a and CK_b. Significant difference between two groups at the 95% and 90% confidence level were marked with * and + , respectively. The abbreviations of the four groups and description for box-and-whisker plot are the same as in Fig. 1. plantations, particularly in the early stages by Asteraceae species, may lie in several aspects. First, anthropogenic disturbance during eucalypt plantation often remove small trees and shrubs and reduce the importance of the dominant species understory, thus providing vacant niches for the recruitment of herbaceous plants. Second, Asteraceae species often produce a large quantity of small seeds that are easily dispersed by wind or animals; hence, these species have a good chance to arrive at a vacant niche. Third, eucalypt trees often have narrow leaves hanging vertically in the canopy, thus causing a relatively high light and wind transmittance understory. The increased light availability and soil temperature but low soil water availability under eucalypt plantations may benefit the growth of drought tolerant pioneer plants 24 , such as C. odorata and P. clematidea. As the eucalypt plantation grows and canopy coverage increases, invasive species gradually lose their advantage against native species in resource competition, particularly with regard to light, and result in their exclusion from the understory community with age of plantation. Besides, our study reveals that eucalypt plantations tend to have an increased number of invasive species when they receive high radiation and precipitation and are located at low elevations; this finding is consistent with the pattern that regions with warm and moist climates tend to have a high number of invasive species in China 21 .
According to our results, to minimize plant invasion risk in eucalypt plantations, we should reduce anthropogenic disturbance of understory community, prolong the rotation length and employ natural succession, which have been reported to promote sustainable eucalypts production as well [16][17][18] Table 3. Invasive species recorded in eucalypt plantations and contrast vegetation, the importance values (Imp) averaged across plots within each group and invasive grade (G). Species belong to Asteraceae family were marked with ψ .     26 , 363 species from 229 genera in 85 families were recorded. Invasive plants were confirmed as species with an invasive grade of 1, 2, 3 and 4, which respectively denote destructive, serious, regional and general invasive in the "Checklist of the Chinese Invasive Plants" 23 . Statistical analysis. We applied four indices, namely, species richness, phylogenetic diversity 27 , Shannon's index (equation (1)) 28 and Pielou's evenness index 29 (equation (2)), to assess plant diversity in eucalypt plantations and contrast vegetation. To calculate phylogenetic diversity, we constructed a phylogeny tree with all of the recorded species by using the topology of "Phylomatic tree R20120829 for plants" (http://phylodiversity.net/ where C i is the coverage of the ith species, C is the total coverage of all species within a plot, D i is the number of the ith species, D is the total individual number of all species, F i is the frequency of the ith species in the subplots and F is the total frequency of all species within a plot. The importance value of each recorded species within each plot was calculated as the average of relative coverage, relative density and relative frequency in the community (equation (3)). The importance value of each family within each plot was calculated by using a similar method as species importance, with the coverage and number of each family calculated as the sum of species' values in which each family belongs to. The importance of invasive species within each plot was calculated as the sum of importance for all invasive species within a plot. The diversity indices, importance values and other statistics were calculated by using R version 3.1.2 33 with the package "vegan" (version 2.2-1) 34 .