Are Hong Kong and Taiwan stepping‐stones for invasive species to the mainland of China?

Abstract Understanding the origins and introduction pathways of invasive species is a fundamental issue for invasion biology, which is necessary for predicting and preventing future invasion. Once an invasive species is established in a new location, this location could serve as a stepping‐stone for further invasions. However, such “stepping‐stone” effect has not been widely investigated. Using the published literature and records, we compiled the first found locations of 127 top invasive species in China. Our study showed that the most common landing spots of these invasive species were Hong Kong (22 species) and Taiwan (20 species), which accounted for one‐third of the invasive species in China. Our analysis revealed that the invasive species in mainland China were more likely to transport from Hong Kong than Macau, a neighboring region with a similar area and colonial history. Similarly, more invasive species were also first landed on Taiwan than Hainan, a nearby island sharing similar climate conditions. Together, our findings indicate that Hong Kong and Taiwan are the most important stepping‐stones for invasive species to the mainland of China and suggesting that the increasing trade exchange of China's coastal ports constitutes a potential risk for the spread of more invasive species. We suppose that they would be the future stepping‐stones for invasive species to the mainland of China and these coastal ports regions where improved biosecurity is needed now.


It has been almost 60 years since Charles Elton published his classic book on biological invasions, The Ecology of Invasions by Animals and
Plants (Elton, 1958;Ricciardi & Maclsaac, 2008;Richardson, 2011;Richardson & Pysek, 2008). Ever since then biological invasions have become serious global environmental problems as a result of the globalization (Perrings, Dehnen-Schmutz, Touza, & Williamson, 2005). Nowadays, understanding the geographical distributions and spread pathways of invasive species has become a central goal of invasion biology and could benefit the prediction and prevention of future invasions (Cassey, Vall-Llosera Camps, Dyer, & Blackburn, 2015). For example, more than half of the naturalized plant species and invasive species originate from North and South America (Jiang et al., 2011;Xu et al., 2012). A comprehensive assessment of naturalized vascular plants worldwide has already been performed recently, which revealed that Northern Hemisphere is the major donors of alien species to other continents (van Kleunen et al., 2015).
It has been suggested that anthropogenic effects, especially global trades and transports, were major drivers of spread of invasive species worldwide. For example (Banks, Paini, Bayliss, & Hodda, 2015;Seebens et al., 2015), it was reported that the pet trade was a driver of introduction and establishment in alien birds in Taiwan (Su, Cassey, & Blackburn, 2016). It was also pointed that colonization pressure is key to understanding alien bird species richness at the global scale, alien bird species richness is currently highest at midlatitudes and is strongly determined by anthropogenic effects (Dyer et al., 2017). Global trade and transport network topology were used to study the geographical spread of invasive species and enable the development of more effective strategies to prevent invasions, utilizing this new insight and tools in a systemic approach can help decision-makers in managing threats to national and regional biosecurity and in safeguarding the world's natural and managed ecosystems (Banks et al., 2015). However, our knowledge of the spread pathways and landing spots of invasive species is still very limited, which may hinder our ability to make up effective policy and management strategies for invasion prevention (Dawson et al., 2017;Russell & Blackburn, 2017).
Once an invasive species successfully established in a new location via long-distance dispersal, such invaded location could serve as a stepping-stone for further invasions to the nearby locations.
Such "stepping-stone" effect has been observed for the spread and dispersal of invasive species in aquatic and terrestrial ecosystems (Buchan & Padilla, 1999;Suarez, Holway, & Case, 2001). With the growth in international trade and travel activities, some invasion hotspots would act as stepping-stones to determine the rate and spatial pattern of spread of invasive species (Floerl, Inglis, Dey, & Smith, 2009). However, unlike the extensive research on the geographical origin of invasive populations, few studies have focused on the first landing places of alien species in a new area and their introduction pathways, making it difficult to test the "stepping-stone" hypothesis.
As the third largest country and the biggest trading nation in the world, China faces grave challenges of biological invasion.
The economic losses caused by invasive alien species in China are estimated to be more than 14 billion USD per year (Xu et al., 2006). So far, a number of studies have investigated the origins, geographical distribution, invasion mechanism, ecological effects, and detrimental economic impacts of biological invasions in China (Liu, Liang, Liu, Wang, & Dong, 2005;Liu et al., 2006;Weber, Sun, & Li, 2009;Xu et al., 2006Xu et al., , 2012. Further, the province-based distribution of invasive plants in China has also been published (Feng & Zhu, 2010;Liu et al., 2005), making it possible to identify locations where invaders first establish their populations.
However, studies on the first landing places of invasive species in China and their spread pathway with a biogeographical approach are still lacking.
Using the published literature and records, we compiled the first landing places of 127 top invasive species in China, which provide an important dataset for revealing the potential stepping-stones for these invaders. We further compared the landing patterns of two trade hubs (Hong Kong and Taiwan) with their neighboring regions with similar area and climate conditions (Macau and Hainan) and hope that this effort will contribute toward offering insight into the prevention and control of biological invasions in China.

| Definition
Classifying invasion status of alien species is a controversial practice (e.g., Pyšek, Richardson, & Williamson, 2004). In previous literature, the concepts of alien, exotic, naturalized, and invasive species are often confounded. According to the definition of IUCN, we considered invasive species as those alien species that established in natural ecosystems that threaten native biological diversity, natural environment, economic, and human well-being (IUCN, 2000;McNeely, Mooney, Neville, Schei, & Waage, 2001;Shine, Williams, & Gundling, 2000). Therefore, only alien species with definitive evidence of negative environmental and social impacts are considered as invasive species here. In our analysis, we focused on the 127 top invasive species reported in Li and Xie (2002), as the invasion status, origins, and life forms of these invaders were well documented.  Tables S3 and S4). Using this approach, we acquired the data on the first detection locations for 108 of the 127 invasive species in China. The 19 species without known information on their first landing places were deleted from our analysis. Note that our analysis included invasive species that spread naturally as well as those assisted by human activities, such as trade exchanges. Further, we also acquired the trade volumes, population density, area, and GDP: Gross Domestic Product of the 34 provinces and regions of China, to assess the roles of trade and human factors in the distribution and spread of these invasive species.

| Data analysis
We first examined the spatial pattern of these top invasive species at the provincial level. For each province or region, we summarized the total number of invasive species and the number of invasive species that were first detected, to determine which province or region is the most common landing place for the invasive species. We further compared the numbers of invasive species that were first detected in Hong Kong to Macao, and Taiwan to Hainan. To see whether the invasive species in mainland China were more likely to come from Hong Kong than its neighboring region-Macao, we used a chi-square test to compare the observed number of first landed invasive species in Hong Kong and Macao to the null assumption that the number of first landed invasive species in these two regions is evenly distributed. We did a similar analysis for Taiwan and Hainan, to see whether these top invasive species were more likely to first land in Taiwan than Hainan. We used simple linear regression to model the number of top invasive species that first detected in Hong Kong in different provinces and regions, as a function of the geographical distance between these regions and Hong Kong, to assess whether the regions closer to Hong Kong contain more invasive species that first detected in Hong Kong. Further, we also used generalized linear mixed model (GLMM) with logit link and binomial error distribution to assess whether the top invasive species that first detected in Hong Kong have higher chance to present in the regions closer to Hong Kong. To do so, the presence/absence of invasive species first detected in Hong Kong in each region was considered as binary response variable, and the geographical distance of the region to Hong Kong was treated as fixed effects and invasive species identity as random effect. We perform the similar analysis for the invasive species first detected in Taiwan, to see whether these species more likely to present in the regions closer to Taiwan.
To confirm our findings, we also divided these top invasive species into different groups, according to their introduction pathway and taxonomic groups (see Table S1). We firstly divided top invasive plants and animals into intentionally and unintentionally introduced species, respectively. Then, we also divided animal invasive species into three taxonomic groups (insect, fish, and others), and plant invasive species into four taxonomic groups (vine, woody, dry herbs, and aquatic herbs). We perform the similar analysis for each group, to see whether the numbers of first detected invasive species of each group in Hong Kong and Taiwan were larger than Macao and Hainan, respectively (see Table S2). We also used simple linear regression to model the number of first detected invasive species that in different provinces and regions, as a function of the trade volumes, population density, area, and GDP of these regions, to see whether "stepping-stone" effect relates to the trade and other human factors.

| RESULTS
The analysis on the landing spots of the top invasive species in China (see Tables S3 and S4) showed that more invasive species were first discovered in Taiwan Table 2). We found similar pattern for the F I G U R E 2 The number of first detections (light red) and the total number (light blue) of invasive species in the 34 provinces and regions of China  Figure 3b).

The analysis on introduction ways of top invasive species in China
showed that about one-third top invasive species were intentional introduction. We found that the number of unintentionally introduced invasive species first found in Hong Kong and Taiwan was higher than Macao and Hainan, respectively (p < .10; Table 3). Similarly, the number of intentionally introduced invasive species first found in Taiwan was significantly higher than in Hainan (χ 2 = 11, p = .003; Table 3), but the number of intentionally introduced invasive species first found in Hong Kong and Macao did not show significant difference (χ 2 = 2, p = .480; Table 3). After we divided all top invasive species into different taxonomic groups, we found that for all the taxonomic groups first detected in the four regions we compared, Hong Kong and Taiwan always contained larger number of first detected top invasive species than Macao and Hainan, respectively (Table S2).
Our results showed that the number of first detections of invasive species significantly increased with the trade volumes, population density, and GDP of all the regions, whereas area has little effect on the number of first detections invasive species (see Figure 4). Further, we also showed that trade volume was the single best predictor (highest R 2 value) of the number of first detected invasive species, indicating that international tread may be the key factor in determining which locations become stepping-stones. To answer this question, we compared the number of first detected invasive species between Hong Kong and Macao, as well as Taiwan and Hainan (see Figure 1).

Macao is a neighboring region of Hong Kong with similar climate
conditions and similar colonial history ( Figure 2 , 2009b;Zhao, 2007), being closer to the mainland.
The areas of two islands are quite close, and they share similar climate conditions. Their locations are shown in Figure 2. Conceivably, they both could be the stepping-stone for biological invasions. However, our analysis shows the number of alien species first found in Taiwan is nearly ten times that in Hainan Island (see Table 1). Therefore, it can be concluded that the location, environmental, and climate factors alone could not make Hong Kong and Taiwan the stepping-stone of invasive species. There are some other factors.
Recent studies suggest that the colonial rule and the related historical events have accelerated biological invasions (Dyer et al., 2017;Wu, Hsieh, Chaw, & Rejmanek, 2004). In fact, Taiwan, as an island of beginning of the 20th century by Japanese during their presence on the island (1895-1945) (Li & Xie, 2002;Lu, Wu, Fu, & Zhu, 2007).
However, colonial history also could not fully explain the invasion pattern we found in China. Hong Kong and Macao were both colonized for a long time; Hong Kong was the colony of the Britain, and Macao was the colony of Portugal. It is surprising that the number of firstly detected invasive species in Hong Kong is much more than that in Macao (Table 1). We suggest that trade exchanges would be the driving force for the accumulation and spread of invasive species in Hong Kong. Hong Kong is a world hub for trade, logistics, and transportation, which have proven to be important driving forces for the introduction of alien species (Banks et al., 2015;Perrings et al., 2005;Seebens et al., 2015;Wilson, Dormontt, Prentis, Lowe, & Richardson, 2009;Zhao & Lu, 2007 (Seebens et al., 2015), Dyer et al. (2017) reported that recent introduction of alien bird species in whole world is a wider phenomenon, involving more species and countries, and driven in part by increasing economic activity, the number of introductions to a country in the fourth quartile is positively correlated with its per capita GDP.
Islands or peninsulas are more susceptible to species invasion compared with continents (Gimeno & Hulme, 2006 Kong) could rival Hong Kong as a point of entry for many introduced species that will become invasive in the future. We suppose that they would be the future stepping-stones for invasive species to the mainland of China and these cities where improved biosecurity is needed now.