Geographical distribution of the dispersal ability of alien plant species in China and its socio-climatic control factors

Dispersal ability is important for the introduction, establishment, and spread of alien plant species. Therefore, determination of the geographical distribution of the dispersal ability of such species, and the relationship between dispersal ability and socio-climatic factors are essential to elucidate the invasion strategies of the alien plant species. Analytic hierarchy process and inventory, risk rank, and dispersal mode data available on Chinese alien plant species were used to determine their dispersal ability, the geographical distribution thereof, and the relationship between socio-climatic factors and dispersal ability. High-risk alien plant species had a higher natural dispersal ability (or several natural dispersal modes) but a lower anthropogenic dispersal ability (or few anthropogenic dispersal modes) than low-risk alien plant species. The geographical distribution of the dispersal ability of the alien plant species showed an inverse relationship with species density. Alien plant species with low dispersal ability (i.e., with fewer dispersal modes and distribution in the southeast) showed a tendency to adapt to environments with mild climates, while those with high dispersal ability (i.e., with more disposal nodes and distribution in the northwest) showed a tendency to adapt to harsh environments. It is essential for land managers and policy makers to understand the geographical distribution of the dispersal ability of alien plant species and their socio-climatic control factors to formulate strategies to control the natural and anthropogenic dispersal of such plants.

, particularly for those that have different dispersal modes, because each mode corresponds to a particular dispersal distance 15,20,22,23,27,28 . For example, a diaspore can use more than one dispersal mode based on its morphological characteristics. Species that are dispersed by wind (anemochory) can present with the following five types of dispersal modes based on their morphology: 1) chamaechory (an entire round plant that rolls on the ground); 2) cystometeorochory (balloon-like diaspores that glide or roll on the ground); 3) trichometeorochory (diaspores with pappi and that float in the air); 4) pterometeorochory (diaspores with wings and descend in a spiral); and 5) semachory (small diaspores with no special morphology and float in the air) 10,13,15,20,30 . Therefore, the dispersal mode-a combination of dispersal agents, plant traits, and environmental conditions-represents a considerable component of the ability of plants to extend their dispersal distance 15 . Several ranks of the risk from alien plant species were utilized to meet the requirements of policy makers who were responsible for developing strategies to manage invasive alien plants in China 3,18,[31][32][33][34][35] . Ma et al. 35 have reported that in China, alien plant species are classified into the following six risk ranks, from high-to low-risk, based on their biological traits, scale of possible invasion, and socioeconomic damages caused: I) malignant invaders, II) serious invaders, III) local invaders, IV) mild invaders, V) species requiring further observation, and VI) cultivated alien species. The ability of alien plant species in different risk ranks to extend their dispersal distance is important for invasion control 18 . Further studies should also be conducted to assess the dispersal ability of alien plant species that are in different risk ranks and those that use both natural and anthropogenic dispersal agents.
China spans across four precipitation regimes (humid, semi-humid, semi-arid, and arid) from east to west because of its remarkable geographical diversity and continental monsoon climate. From north to south, the country also covers five climatic zones (cold temperate, temperate, warm temperate, subtropical, and tropical) 36,37 . Additionally, regional population density and the gross domestic product (GDP) tend to decrease from southeast to northwest 38,39 . Therefore, certain socioeconomic factors associated with population may be closely related to the dispersal ability of alien plant species. Previous studies have shown that the density of alien plant species decreases from the southeast coast to the northwest inland region of China 2,3,11,40-42 . Factors influencing this regional distribution pattern of alien plant species density (or richness) can be natural (e.g., precipitation and temperature) and/or anthropogenic (e.g., density of human population, GDP, and GDP per capita) 2,11,42 . Therefore, both natural and anthropogenic factors should be considered as influencing factors that impact the regional distribution of the dispersal ability of alien plant species.
In this study, we used an inventory, risk ranks, and information of the dispersal modes of alien plant species to examine their dispersal ability, geographical distribution, and the relationship between socio-climatic factors and dispersal ability. Our objectives were to determine the following: 1) the characteristics of alien plant species' dispersal abilities, including those in different risk ranks, 2) the geographical distribution of the dispersal abilities across China, and 3) the relationships between socio-climatic factors and dispersal abilities. The results found based on these objectives might help land managers and policy makers to control alien plant species in China.

Methods
Data collection. Data on 562 species of alien plants and their distribution in China were collected based on The Checklist of the Chinese Invasive Plants 35 . The alien plant species were distributed in 34 provincial-level administrative regions-twenty-three provinces, five autonomous regions, four municipalities, and two special administrative regions 2,11,35,43,44 . We collected the data on morphological traits of diaspores from several sources, including the Flora of China (http:// www. iplant. cn/ foc/), the PLANTS database managed by the United States Department of Agriculture (https:// plants. usda. gov/ java/), and the Seed Information Database from the Royal Botanic Gardens, Kew (http:// data. kew. org/ sid/). We retrieved references from the China National Knowledge Infrastructure database (CNKI, https:// www. cnki. net/), Baidu (https:// www. baidu. com/), and Web of Science (http:// apps. webof nowl edge. com) using the Latin name of the alien plant species as our keywords, and reviewed them to obtain information related to dispersal modes.
Definition of risk ranks for alien plant species. The alien plant species in China were categorized into six ranks, from high-to low-risk, based on their biological and ecological characteristics, distribution, and influence on ecosystems and the national economy (Table 1) 35,41,45,46 . Selection of socio-climatic factors and collection of data. We selected the four climatic factorsmean annual humidity (MAH), mean annual precipitation (MAP), mean annual temperature (MAT), and mean frost days (MFD)-to examine the relationship between them and determined the dispersal ability of alien plant species 44 . The climatic data were obtained from a dataset containing information from 2,086 ground meteorological stations in China for the period 1981 to 2010 (http:// data. cma. cn/) 44 . The annual mean values for each climatic factor were obtained by collecting monthly data from these stations, which were distributed in the 34 provincial-level administrative regions.
Evaluation of diaspore dispersal ability. In this study, we defined the ability of alien plant species to extend their diaspore dispersal distance as dispersal ability without considering the roles of the number of diaspores dispersed due to lack of data on the latter. To overcome selective pressures arising from both biotic and abiotic factors, plant species rely on multiple dispersal modes to facilitate the spread of their diaspores 14 . Diaspore dispersal ability can be used to estimate the dispersal modes used 15 ; therefore, the dispersal ability of a species is a consequence of a combination of multiple dispersal modes. Using the technique of analytic hierarchy process (AHP), we first assigned a weight to each dispersal strategy and mode based on the dispersal distance they corresponded to by calculating the relative importance of two strategies and modes and by generating a pairwise comparison matrix of the dispersal strategies and modes in each of these (Supplementary Material 2). We then determined the dispersal ability of a species by adding the scores of the multiple dispersal modes used (Supplementary Material 2).

Data analysis.
We conducted regression analyses to determine the influence of socio-climatic factors on diaspore dispersal ability. Parameter fitting of the regressions and analysis of variance (ANOVA) and Duncan's tests for multiple comparisons of dispersal ability were conducted (with statistical significance as p < 0.05 for Duncan's test) using the software PASW Statistics 18.0 (IBM Inc., Armonk, New York, USA). Histograms of the changes in dispersal ability and the distribution of alien plant species at different risk ranks were illustrated using

Results
Dispersal ability of alien plant species in six risk ranks. Our results showed that there was a significant decreasing trend (p < 0.05) in the natural dispersal ability, but an increasing trend (p < 0.01) was observed in the anthropogenic dispersal ability for plants in ranks I to VI (Fig. 2, Table 2). Overall, there was an increase (p < 0.01) in both natural and anthropogenic dispersal abilities for plants in rank I to VI (Fig. 2, Table 2).
Geographical distribution of the dispersal ability of the alien plant species. The overall, natural, and anthropogenic dispersal ability of all alien plant species showed an increasing trend from southeast to www.nature.com/scientificreports/ northwest regions of China (Fig. 3 Ta, Tb, Tc). The variations in dispersal ability (i.e., the differences between the highest and lowest dispersal abilities) for all alien plant species within the 34 administrative regions studied were 3.0%, 1.5%, and 2.7%, respectively, overall and in natural and anthropogenic dispersal modes (Fig. 3

Relationships between the dispersal ability and socio-climatic factors. The overall, natural, and
anthropogenic dispersal ability of all alien plant species showed significantly negative relationships with GDPwith p < 0.01 and p < 0.05 for overall and natural dispersal ability, respectively-and MAH (Table 3). Natural dispersal ability was positively correlated with MAP (p < 0.01) for plants in rank II, MFD (p < 0.01) and MAT (p < 0.05) for those in rank III, and MFD (p < 0.01) for plants in rank V; however, it was negatively correlated with MAH (p < 0.01) for plants in rank IV. Anthropogenic dispersal ability was positively correlated with MAP (p < 0.01) for plants in rank I but negatively correlated with MAP (p < 0.01), MAH (p < 0.01), MFD (p < 0.05), and MAH (p < 0.01) for plants in ranks III, IV, V, and VI, respectively (Table 3). Overall dispersal ability was positively correlated with GDP (p < 0.01) and MFD (p < 0.01) for plants in ranks I and V, respectively, and negatively correlated with MAH for plants in rank III at p < 0.01; GDP for plants in ranks IV and V at p < 0.05 and p < 0.01, respectively; and MAH for plants in rank VI at p < 0.01 (Table 3).

Discussion
Changes in the dispersal ability of alien plant species in six risk ranks. Our results showed that the natural dispersal ability of alien plant species in ranks I to VI exhibited a decreasing trend; however, the total and anthropogenic dispersal abilities showed increasing trends (Fig. 2). These results implied that highrisk alien plant species (i.e., those that resulted in considerable economic and ecological loss) had a higher natural dispersal ability, and low-risk alien plant species (i.e., those that resulted in less or negligible economic and ecological loss) exhibited a lower natural dispersal ability. Possible reasons for this include the fact that high-risk alien plant species rely more on natural dispersal strategies, and rely less on anthropogenic strategies apart from accidental introduction (e.g., speirochory or introduced involuntarily via mixing with grain or fodder) 14,35,48,49 . For example, Ageratina adenophora (common name: crofton weed; rank I), which was accidentally introduced to China from Mexico, is a notorious weed that exerts considerable ecological impacts and causes substantial economic losses 50,51 . The species exhibits three natural dispersal modes-trichometeorochory, hydrochory, and zoochory-and two anthropogenic dispersal modes, agochory and speirochory ( Supplementary  Material 1 and 2) 14,35 . In contrast, cultivated alien species-which are widely introduced, domesticated, and cultivated by humans for applications in food, fodder, medicine, and building and gardening materials-rely more on anthropogenic strategies 11,37,52 . For example, Mangifera indica (common name: mango; rank VI) is a nutritious and popular fruit that originates mainly from the Indian subcontinent 53 . The species has two natural dispersal modes-barochory and frugivory-and three anthropogenic dispersal modes, namely agochory, ethelochory, and speirochory (Supplementary Material 1) 35,54 .
Geographical characteristics of the distribution of the dispersal ability of the alien plant species. The overall, natural, and anthropogenic dispersal ability of all alien plant species in the six risk ranks showed an increasing trend from the southeast to the northwest regions of China (Fig. 3). The geographical distribution of dispersal ability of the alien plant species was inversely related to the number and density thereof 2,3,11,[40][41][42] . The majority of alien plant species with high dispersal ability were distributed in inner areas of the northwest parts of the country; however, the majority of alien plant species with low dispersal ability were distributed in the southeast region. The dispersal ability of alien plant species is considerably dependent on the number and relative importance of each dispersal mode (Supplementary Material 2). Most alien plant species present with at least one anthropogenic dispersal mode 13,23,45,55,56 , although those with a high dispersal ability may possess more natural dispersal modes than those with a low dispersal ability 14 . Therefore, alien plant species distributed in the inner areas of northwest China often present with more dispersal modes than those in the southeast. Table 2. ANOVA results for the natural, anthropogenic, and overall natural and anthropogenic dispersal abilities of alien plant species in the six risk ranks.  www.nature.com/scientificreports/ Relationships between socio-climatic factors and dispersal ability. Dispersal ability was negatively correlated with socio-climatic factors (Table 3). Southeast regions of China demonstrate the presence of a mild climate, with high precipitation levels, temperatures, and humidity levels, and this supports the survival of a high density of alien plant species 3,41 . In contrast, northwest regions of China demonstrate the presence of a harsh environment-with low precipitation levels, temperatures, and humidity levels-and therefore supports the survival of a low density of alien plant species 40,41 . However, our findings are contrary to these relationships observed between alien plant species density and socio-climatic factors 2,3,11,[40][41][42] . Our results implied that alien plant species with a low dispersal ability (or fewer dispersal modes) could adapt to mild climatic environments as these might render easier invasion 11 . However, alien plant species with a high dispersal ability (or a greater number of dispersal modes) tend to adapt better to harsh environments and can thus invade these regions Table 3. ANOVA analysis and estimation of regression coefficients among the natural, anthropogenic, and overall dispersal abilities and socio-climatic factors for all alien plant species and each of the six risk ranks thereof.  www.nature.com/scientificreports/ successfully 3 . Therefore, studies should be conducted more on the investigation of the northwest regions of China compared to the southeast regions of China in terms of formulation of strategies to prevent the establishment of invasive species. Land managers and policy makers should develop more such measures to eliminate natural dispersal routes of alien plant species, based on their natural dispersal modes, and enforce stricter regulations to control anthropogenic dispersal routes. This study has an important limitation. Traditionally, researchers have assumed that the diaspore of a plant species uses one dispersal mode that is closely associated to its morphology 10 . This perspective ignores the probability that such diaspores can be dispersed via other dispersal modes. However, the data on the probabilities of this occurrence are unknown for alien plant species in China. Therefore, using the technique of AHP to determine the dispersal ability of the latter by weighting all dispersal modes can reflect a trend in their dispersal ability; however, application of such an approach cannot provide an understanding of the exact mechanism by which such alien plant species cause dispersal of their diaspores.

Summary
High-risk alien plant species had a higher natural dispersal ability (or more natural dispersal modes) but a lower anthropogenic dispersal ability (or fewer anthropogenic dispersal modes) than low-risk alien plant species. The dispersal ability of alien plant species increased from the southeast to the northwest regions across China, which was contrary to the geographical distribution of species density. Alien plant species with a high dispersal ability possess more dispersal modes than those with a low dispersal ability. Alien plant species with a low dispersal ability (or fewer dispersal modes) can adapt to mild environments; however, alien plant species with a high dispersal ability (or more dispersal modes) can adapt to harsh environments. Understanding the characteristics of the geographical distribution of the dispersal abilities of alien plant species and invasion strategies adopted to achieve dispersal is essential to enable land managers and policy makers to develop preventive measures for the elimination of natural dispersal routes and for enforcement of laws/policies to control anthropogenic dispersal routes. www.nature.com/scientificreports/ wrote the manuscript. All authors contributed substantially to the preparation of drafts and approved the final manuscript for submission.