Invasion by alligator weed, Alternanthera philoxeroides, is associated with decreased species diversity across the latitudinal gradient in China

Abstract

Aims

Invasive species occurrence and their effects on biodiversity may vary along latitudes. We examined the occurrence (species cover) and relative dominance (importance value) of invasive alligator weed, Alternanthera philoxeroides, in its terrestrial habitat in China through a large-scale latitudinal field investigation.

Methods

We established 59 plots along the latitudinal transect from 21°N to 37°N. We recorded species name, abundance, height and individual species coverage of plants in every quadrat. We then measured α-species diversity variations associated with the A. philoxeroides community across the latitudinal range. We also analyzed the effect of latitude on plant species’ distributions in this community by using canonical correspondence analysis (CCA).

Important Findings

We found that species cover and importance value of A. philoxeroides increased in areas <35°N, but decreased at higher latitudes. Lower latitudes supported greater species diversity than higher latitudes. Small-scale invasion of A. philoxeroides was associated with higher species diversity, but community diversity was lower when A. philoxeroides species cover exceeded 36%. Community plant species changed from mesophyte to hygrophyte gradually from low to high latitude. Our research suggests that latitude had significant influences on community diversity which interacted with the biotic resistance of a community and impact of invasion. Consequently, A. philoxeroides may become more invasive and have greater negative impacts on community species diversity in higher latitudes as global climate changes.

INTRODUCTION

With the rapid development of economic globalization and global climate change, alien species invasions are increasingly critical threats to the environment, ecosystems and global biodiversity (Ding et al. 2008; Hulvey and Zavaleta 2012; Westphal et al. 2008). Biological invasions can lead to declines in native species diversity through resource competition, habitat change and changes in species interactions (D’Antonio et al. 2011; Farrer and Goldberg 2009; Scharfy et al. 2010; Wright et al. 2012). Global biodiversity naturally varies along latitudinal gradients, with a decline in species number and taxa from the tropics to the poles (latitudinal diversity gradient, LDG) (Fischer 1960; Jablonski et al. 2006). The spatial patterns of invasive species across latitudinal gradients have been examined in multiple regions (Liu et al. 2005; Martin et al. 2014), however, the effect of invasion on native species diversity along the latitude is less well known (Richardson et al. 2012, but see Martin and Wilsey 2015), especially within Asia (Pysek et al. 2008). Indeed, the degree of alien invasion and its interaction with latitudinal gradients of biodiversity on reductions in native species’ richness is an area that remains relatively unexplored, yet is of vital importance for managing future spread.

Latitudinal gradients generate variation in abiotic (e.g. solar radiation, precipitation, soil nutrition) and biotic factors (e.g. disturbance, human population density, species interactions), correspondingly, the occurrence of exotic species and their effects on biodiversity may vary in different latitudes (Fridley et al. 2007; Richardson et al. 2012; Stohlgren et al. 2005). For instance, Richardson et al. (2012) found that as abiotic stress (moisture and nutrient limitation) decreased at higher latitudes, an invasive grass inhibited the establishment of native plants, while at lower, less stressful latitudes, the invasive grass facilitated native plant establishment. In contrast, in a recent meta-analysis, Sorte et al. (2013) found that both native and invasive terrestrial species showed similar responses to environmental change across latitudes. However, for aquatic species, warming at higher latitude locations was more likely to negatively affect native versus nonnative species which were more likely to have positive responses to increased temperatures. Taken together, as invasive species richness can also vary along latitudinal clines with paradoxical associations with native species richness dependent on scale (Fridley et al. 2007), it is important to understand the relationships among the degree of invasion, latitude and native biodiversity.

Native to South America, alligator weed, Alternanthera philoxeroides, is a worldwide invasive weed whose distribution spans tropical, subtropical and warm temperate regions across a wide latitudinal gradient, approximately from 42°S to 37°N (including American, Australia, New Zealand, Southeast Asian, India and China) (Julien et al. 1995; Lu and Ding 2013; Telesnicki et al. 2011). It was initially introduced into China in the 1930s as a forage crop, and now occurs widely across the country (Lu and Ding 2010). A. philoxeroides reproduces clonally even as small fragments, spreads and establishes into a variety of environments (Geng et al. 2007), and it can occur in both aquatic and terrestrial habitats. A recent study found that A. philoxeroides has expanded its distribution region from 21.5°N to 36.8°N in mainland China (Lu et al. 2013). However, the effects of A. philoxeroides on native species diversity across latitudinal gradient are unknown.

In this study, we conducted a large scale field survey to examine variation in terrestrial A. philoxeroides invasion across a latitudinal gradient in China. We hypothesized that: (i) The negative impact of A. philoxeroides invasion on native species diversity would be aggravated by increasing latitude; (ii) There would be large differences in plant species’ distributions associated with A. philoxeroides across the latitudinal gradient. Specifically, we address the following questions: (i) Does the dominance of A. philoxeroides vary across a latitudinal gradient? (ii) Does species diversity within communities invaded by A. philoxeroides differ across latitudinal gradients? (iii) How does the degree of A. philoxeroides invasion affect species diversity in communities?

MATERIALS AND METHODS

Site selection

Our latitudinal gradient transected mainland China from N 21°31′ to N 36°45′. During August and September 2012, we chose locations of at least 100 m² invaded by A. philoxeroides to establish our sampling plots. In total, we established 59 plots along the latitudinal transect, most of which were in terrestrial habitats but a few plots incorporated aquatic habitats where A. philoxeroides spread from banks to shallow water. The 59 plots were selected along eight latitudinal clusters about 2° apart (see Fig. 1). In each latitude cluster, we selected 6–10 plots (10×10 m for each, at least 10 km apart) in two or three sampling cities. In each plot, we evenly set up two to four 10 m transects. For each transect, 5 quadrats (0.5×0.5m) were uniformly spaced 2 m apart. The whole study area covered 18 cities within 10 provinces in China and spanned tropical, subtropical and temperate regions.

Vegetation sampling and data collection

We recorded species name, abundance, height and individual species coverage of plants in every quadrat. For abundance, we counted asexual branches (clonal plants) or tiller numbers (graminaceous plants) or individual numbers (non-clonal herbaceous and a few woody plants) for each species. We randomly chose 10 individuals to measure plant height (the vertical distance from soil surface to the tip of the highest branch or longest leaf) for each species (if there were <10 individuals present, we measured all individuals) to calculate average height. To measure the percentage coverage of plants, we placed a 0.5×0.5 m metal frame with 100 cells (each 5×5cm) above the canopy in each quadrat, and summed all cells occupied by a particular plant species. Unknown plants were identified using ‘Flora of China’ (Wu and Chen 2004) and the website ‘Chinese Virtual Herbarium’ (http://www.cvh.org.cn/). In addition, we recorded the longitude, latitude and elevation of every plot using a handheld GPS receiver (Garmin eTrex 20, Garmin international incorporated company, Kansas, USA).

Data analysis

Importance value (IV) is a comprehensive quantitative index that measures the growth and relative dominance of species in a community. Relative IV was calculated using the following formulae, modified slightly from Wang et al. (2007) and Jing et al. (2014):

IV = (relative cover + relative height + relative abundance)/3

Where relative cover, relative height and relative abundance refer to the percentages of one species cover, mean height and abundances over the sum of all species cover, mean height and total abundances within a plot respectively.

Total IV was the sum of a plant species’ relative IV in the 59 plots.

To measure the patterns of species diversity, we employed four α species diversity indices as follow (Pruchniewicz and Zołnierz 2014; Zhang et al. 2015):

Patrick richness index: R = S

Simpson diversity index: λ = 1 − ∑P_i²

Shannon–Wiener diversity index: H = −∑P_i ln P_i

Pielou evenness index: E = (−∑P_i ln P_i)/ln S

Where S is the number of species within a plot and P_i is the relative abundance of species i.

To examine the effect of latitude on A. philoxeroides’ cover and relative IV, we performed a series of regression using SPSS16.0 software (SPSS Inc., Chicago, USA).We used the 11 common models (linear, logarithmic, inverse, quadratic, cubic, etc) provided by the ‘Curve Estimation’ procedure of SPSS16.0 to choose the best-fitting regression model which had the significant maximum of determination coefficient (R square).We also used this curve regression to examine the relationship between species cover of A. philoxeroides and four diversity indices. In addition, we used ‘One-way ANOVA’ and ‘Multiple Comparisons’ (Subset for α = 0.05) to examine differences in species diversity indices along the latitudinal gradient (SPSS 16.0), where our data met assumptions of normality and homogeneity of variance. Otherwise, we used the ‘Games-Howell’ non-parametric test (Yu et al. 2015).

We established a random species matrix containing 30 plants which had a total IV more than 0.200 (59×30; see online supplementary Table 1 for plant names) and a random environmental matrix containing three topographical factors (59×3; latitude, longitude and elevation) to explore species’ distributions within communities invaded by A. philoxeroides along a latitudinal gradient by using the software Canoco for Windows 4.5 (Microcomputer Power, Ithaca New York, USA). Detrended correspondence analysis (DCA) was applied to examine whether linear ordination or unimodal ordination would be appropriate to analyze the data. Results showed that the maximum gradients of the four axes was 3.448, higher than 3 standard deviations, therefore canonical correspondence analysis (CCA) was used (Ter Braak and Smilauer 2002).In addition, a Monte Carlo permutation test based on 499 random permutations was operated to test the significance of the eigenvalues of all CCA canonical axes.

RESULTS

Community species composition

We found a total of 170 plant species in 52 families and 134 genera in 59 plots with A. philoxeroides (see online supplementary Table 1). A. philoxeroides was the dominant community member and had the largest total IV (IV = 32.39) in all our communities (the selected plots were always dominated by A. philoxeroides), followed by Digitaria sanguinalis (IV = 4.35).

The Poaceae had the greatest richness, with 26 genera and 32 species represented in our plots, followed by the Asteraceae (15 genera 22 sp), Labiatae (10 genera 11sp), Cyperaceae (5 genera 9 sp) and Polygonaceae (2 genera 8 sp). These five plant families contained 43% of the total genera and 48% of the total species in our communities. After A. philoxeroides, Alternanthera sessilis appeared most frequently (16/59 plots), followed by Polygonum hydropiper (13/59), Digitaria sanguinalis (12/59), Monochoria vaginalis (11/59) and Roegneria kamoji (10/59).

Species diversity indices along latitudinal gradients

Species richness (Patrick index) varied across the latitudinal gradient (F_{7, 51} = 2.732, P = 0.0170), and was significantly higher in lower latitudes (21–23°N) than in higher latitudes (>25°N) (Fig. 2). Species diversity (Shannon and Simpson indices) differed across the latitudinal gradient (Shannon–Wiener index (F_{7, 51} = 3.660, P = 0.003) and Simpson index (F_{7, 51} = 3.581, P = 0.003)), and, in general, was greater in lower latitudes than in higher latitudes (Fig. 2). Species evenness (Pielou index) also varied across the latitudinal gradient (F_{7, 51} = 2.348, P = 0.037), and was highest in the lowest latitudes (21–25°N) compared to the higher latitudinal gradients (33–37°N).

Occurrence and dominance of A. philoxeroides along latitudinal gradients

Regression analysis (Fig. 3a and d) showed that the optimally fitting model between latitude and either A. philoxeroides’ average cover or relative IV was cubic (F_{2, 56} = 10.314, P < 0.001; F_{2, 56} = 8.329, P = 0.001). A. philoxeroides cover increased significantly with increasing latitude in the range of 21–35°N, peaked at 35°N, then dropped. Relative IV of A. philoxeroides increased between 21–34°N, but decreased at higher latitudes. A. philoxeroides height increased significantly with latitude from 21 to 32°N but decreased after 32°N (Fig. 3b, F_{2, 56} = 5.873, P = 0.005). Abundance of A. philoxeroides was positively associated with higher latitudes (Fig. 3c, F₁, ₅₇ = 4.757, P = 0.033).

A. philoxeroides species cover and species diversity

The optimal fitting relationships between A. philoxeroides’ species cover and Patrick index (F_{3, 55} = 6.118, P = 0.001), Shannon–Wiener index (F_{3, 55} = 33.766, P < 0.001), Pielou index (F_{3, 55} = 59.619, P < 0.001) and Simpson index (F_{3, 55} = 54.120, P < 0.001) were cubic (Fig. 4).However, peak thresholds for A. philoxeroides’ species cover were different for each diversity index (Patrick 36%, Shannon 15%, Pielou 7% and Simpson 16%)indicated by the black arrows in Fig. 4.

Ordination analysis of community plant species’ distributions

In our CCA ordination, the first three canonical axes explained 100% of the variance in the species–environment relation, and the cumulative percentage explanation of the first two axes was 84.5% (48.5% for axis 1 and 36.0% for axis 2, respectively). The Monte Carlo permutation test indicated that the eigenvalues were significant for the first two axes (the P values were all <0.01) but not significant for the third axis (P = 0.078), so we used the first two axes to draw two-dimensional CCA ordination diagram (Fig. 5). The dominant environmental variable correlated with the first axis is latitude (coefficient = −0.677, P < 0.01). Longitude had a strong correlation with the second axis (coefficient = −0.842, P < 0.01), while elevation had no significant correlation with the two axes.

Species associated with A. philoxeroides were mainly hygrophilous plants in higher latitudes (area I in Fig. 5), including Polygonum hydropiper (6), Scirpus planiculmis (24), Eleocharis dulcis (15), Echinochloa crusgalli (16), Marsilea quadrifolia (29) and Rumex patientia (11), which accounted for 67% of the total species in area I. There were fewer hygrophilous plants in area II compared to area I: species included Monochoria vaginalis (3), Cyperus microiria (10), Beckmannia syzigachne (21) and Leersia hexandra (14). Lower latitude areas (area III) supported the fewest hygrophilous plants, with only two species (Acorus gramineus (20) and Polygonum lapathifolium (25)). In general, at lower latitudes, plant types associated with communities invaded by A. philoxeroides community switched from hygrophyte to mesophyte.

DISCUSSION

Our results showed that the plant community varied across latitude, with the invader’s relative dominance (IV) generally increasing with latitude. We found complex relationships with A. philoxeroides invasions and community diversity, with small scale invasions associated with increased diversity but larger invasions associated with decreased species diversity (Fig. 4). We also found that, from low to high latitude, plant species in the community invaded by A. philoxeroides gradually changes from mesophyte to hygrophyte.

We found that both A. philoxeroides’ cover and relative IV increased with latitude from south to north and peaked at about 35°N. Compared to the tropical climate in its native range of South America, A. philoxeroides experienced a similar climate in low latitudes of our survey areas. In fact, A. philoxeroides growth was seriously suppressed by an insect natural enemy, Agasicles hygrophila (alligator weed flea beetle), in the lower latitudes (especially 20°–25°N) of mainland China (Zhao et al. 2015). Since A. hygrophila cannot overwinter and establish population above 32°N and has a lower tolerance to cold hardiness than A. philoxeroides (Lu et al. 2013; Zhao et al. 2015), the dominance of A. philoxeroides with increasing latitude up to a point could be mediated by this plant–insect–climate interaction. Furthermore, the solar radiation and precipitation in higher latitude are lower, while the ability of maintaining photosynthesis and capturing water of A. philoxeroides are greater than many natives (such as A. sessilis) due to relatively lower light compensation point and higher physiological plasticity (Chen et al. 2013; Geng et al. 2006; Pan et al. 2006). But in areas >35°N, low temperature may be less unsuitable for A. philoxeroides’ growth compared to the lower latitudes (as the relation of latitude-temperature showed in online supplementary Figure1). This phenomenon that biological invasion could be accelerated at higher latitude due to global warming has been suggested in other studies (Richardson et al. 2012; Sorte et al. 2013; Walther et al. 2002). With global warming, A. philoxeroides may continue to spread and invade at high latitudes (Barrett and Gray 2011; Lu et al. 2013; Wang et al. 2011).

The LDG of the communities associated with A. philoxeroides is likely formed by the comprehensive effects of insolation, thermal seasonality and other abiotic and biotic factors. Low latitudes in tropics usually have higher insolation (light and heat) and lower thermal seasonality which means higher environment stability, thus promoting species specializing and supporting greater species coexistence (Archibald et al. 2010; Clarke and Gaston 2006). In subtropical areas, as thermal seasonality increases, plants may have to exert energy to resist this climate perturbation. However, A. philoxeroides phenotypic plasticity and asexual reproduction ability may offset this energy loss, increasing their invasive spread in higher latitudes.

Invasive species usually threaten biodiversity, however, initial small invasion could improve species diversity (Lososová et al. 2012). This phenomenon had been found on some invasive plants such as Eucalyptus camaldulensis in South Africa and Sargassum muticum in Canada (Tererai et al. 2013; White et al. 2011). We also found that A. philoxeroides improved four diversity indices at the early invasion stage (Fig. 4). As a member of community, the invader itself can also increase community species richness (Lockwood et al. 2011). However, as its population increases, A. philoxeroides likely threatens biodiversity by competing with other plants for resources and space via excreting allele chemicals such as sterols, flavones and terpenes in areas where it outbreaks (Zhou et al. 2012; Zuo et al. 2012). As species cover of A. philoxeroides increased in our study, all four community indices showed a downward tendency towards lower biodiversity after an initial increase, suggesting that a small level of A. philoxeroides invasion could improve community species diversity, but large invasions are associated with negative diversity impacts (Fig. 4).

A. philoxeroides (1) locates centrally on the CCA ordination (Fig. 5), indicating that this invader has strong ecotypic adaptation and can grow and establish populations in a wider range of environmental conditions. The locations of A. philoxeroides (1) and Humulus scandens (12) in ordination diagram have some overlaps, suggesting that there may have strong interspecific competition between these two plants. Low latitudes have high solar radiation and intense moisture evaporation (Clarke and Gaston 2006), thus plant species may evolve homologous life forms in these areas. This may explain why there are more xerophyte plants in A. philoxeroides communities at lower latitudes than those in higher latitude. On the contrary, low moisture evaporation and temperature in higher latitude may provide more suitable environments for mesophyte and hygrophyte species. Species richness for Poaceae and Asteraceae were greater than for other families in this invaded community, it may because that they were the most species rich families in vascular plants (Hayasaka et al. 2012; Ramírez et al. 2007). In addition, A. sessilis, which is invasive in the USA and is in the same genera as A. philoxeroides, occurred more frequently with A. philoxeroides than other species, suggesting that these plants may have similar ecological niches.

Our findings have important implications for predicting A. philoxeroides invasion under global warming. As A. philoxeroides may continue to expand its distribution range to higher northern latitudes with increasing global temperatures, this invader’s effect on native species diversity may be accelerated at higher latitude. Maintaining higher habitat heterogeneity to improve community species evenness will be conducive for enhancing community resistance to A. philoxeroides invasion. Our results are also helpful for understanding species structure characteristics of invaded communities and for revealing the influence of environmental factors on biological invasion. In this study we only focused on the terrestrial form of A. philoxeroides, future studies are needed to examine the patterns of the aquatic form and the effects on native aquatic species along latitudes, as established literature had showed that the response of aquatic invasive species to global warming is more positive than their terrestrial counterparts (Store et al. 2013).

With global warming, especially with the high rate of temperature increase in higher latitudes, biological invasion may bring even greater threats to global biodiversity (Barrett and Gray 2011; Walther et al. 2002; Wasowicza et al. 2013). Global warming can also favor invasive species to expand their realized climatic niche along latitudinal gradients and increase their effects on native species (Lu et al. 2013; Webber et al. 2012). In this study we show that invasion by A. philoxeroides is associated with decreased species diversity across the latitudinal gradient in China. As many invasive plants such as Conyza sumatrensis, are also native to South America and now widely invaded into south and central China (Ren et al. 2010), studies on their occurrence and effects along latitudes could assist to predict their invasions and provide guidance for future management.

SUPPLEMENTARY MATERIAL

Supplementary material is available at Journal of Plant Ecology online.

FUNDING

Knowledge Innovation Program of Wuhan Botanical Garden (Y455437H05) and Juli Carrillo was supported by a National Science Foundation Postdoctoral Research Fellowship.

ACKNOWLEDGEMENTS

We acknowledge Xinmin Lu, Xu Shao, Jingzhong Lu for field assistance and the help in plant identification from Minghui Yan. We are grateful for comments by the two anonymous reviewers that improved the early version of this manuscript.

Conflict of interest statement. None declared.

REFERENCES