Trends in the direction of global plant invasion biology research over the past two decades

Abstract Invasive plants are a growing ecological problem worldwide, but biases and patterns within invasive plant research may affect our understanding of invasive plant ecology. In this study, we analyzed 458 invasive plant papers sampled from the two journals dedicated entirely to the field of invasion biology, i.e., Biological Invasions and Neobiota. From these papers, we collected information on geographic coverage, climate, habitat, taxonomic coverage, plant functional type, and research topic to examine trends across a 21‐year time period from 1999 to 2020. Our analysis found that invasive plant research was consistently biased toward temperate grassland and forest ecosystems particularly within the Americas, Europe, and Australia, and toward smaller, herbaceous invasive plant species (i.e., forbs, grasses, and shrubs), with an increase in interest in invasive nitrogen‐fixing legumes over time. Our analysis also identified “hot” research topics in invasive plant research at specific time periods, such as a peak in the use of genetic analysis methods in 2014–2015 and a more recent focus on plant physiological and functional traits. While current models, concepts, and understanding of plant invasion ecology are still driven by such biases, this has been partially offset by recent increased research in understudied systems, as well as increasing awareness that plant invasion is heavily affected by their growth types, physiological traits, and soil interactions. As the field of invasion biology becomes ever increasingly important over time, focusing invasive plant research on understudied ecosystems and plant groups will allow us to develop a more holistic understanding of the ecology of invasive plants. In particular, given the outsized importance of the tropics to global biodiversity, the threats they face, and the dearth of studies, it is of critical importance that more invasive plant research is conducted within the tropics to develop a more globally representative understanding of invasive plant ecology.


| INTRODUC TI ON
Invasive species are an increasing ecological problem worldwide and can alter the diversity and composition of native ecosystems in a detrimental manner (Bellard et al., 2016;Murphy & Romanuk, 2014;Simberloff et al., 2013). Once established, invasive species often become difficult to remove, and their effects on native ecosystems are also often difficult to reverse (Kettenring & Adams, 2011;Mack & Lonsdale, 2002). Globalization has also increased opportunities for invasive species to be introduced to new ranges, and invasion is projected to be an increasing ecological problem worldwide (Seebens et al., 2015(Seebens et al., , 2021. Invasion biology (the study of invasive species), as an independent field of research in ecology, is a relatively young field (Simberloff et al., 2013). There has been a recent explosion of interest over the past 10-15 years in this topic (Chong et al., 2021;Figure 1), which now has two major journals dedicated exclusively to it-Biological Invasions, which started in 1999 (Carlton, 1999) and Neobiota, which started in 2011 (Kühn et al., 2011).
While invasive plant research is well represented in the general area of invasion biology (Pyšek et al., 2008), key areas within invasive plant research may be greatly underrepresented. Typical of many fields in ecology, the literature has historically been dominated by wealthier and developed countries (i.e., the Global North: America, Europe, Australia, and New Zealand), with a strong resultant bias in temperate ecosystems (Chong et al., 2021;Nuñez et al., 2022;Pyšek et al., 2008;Weidlich et al., 2020). As such, a recent analysis by Nuñez et al. (2022) found that a vast majority of all papers published within Biological Invasions were published by USA or Europeanbased scientists. Correspondingly, there is relatively widespread awareness of the harms that invasive plants can do in the temperate zone. "Classic" plant invasion examples that frequently attract attention and are frequently taught in university courses-for example, pines (Pinus spp.) in the temperate Southern Hemisphere, and kudzu (Pueraria montana) and the grass Bromus tectorum in North America-reflect such temperate biases in global recognition of invasive plant issues. Conversely, there have been few studies in tropical regions, except for particular island systems such as Hawaii (Chong et al., 2021;Weidlich et al., 2020), resulting in lower awareness of invasive plant issues in tropical areas. This is a significant problem, as tropical ecosystems contain much of the world's biodiversity (Barlow et al., 2018) and contain some of the world's most threatened ecosystems (Gardner et al., 2009;Laurance, 2013).
Furthermore, there may be other forms of biases and patterns within invasive plant research. As in all ecological research, fieldwork is important in invasive plant research. However, due to the inherent logistical difficulties of fieldwork, invasive plant research may be concentrated around easy-to-study and local plants and field sites.
For instance, there may be a bias toward easily accessed ecosystems like grasslands and some types of forest (Weidlich et al., 2020).
As an example, some of the more well-known research coming from the Americas (both North and South) in invasive plants look at grasslands-whether it be invasion of various types of plants in the South American Cerrado Savanna (Pivello et al., 1999), or invasion of the grass Bromus tectorum into North American grasslands (Vitousek et al., 1997). Furthermore, researchers may tend to prefer to study ecosystems which are more local or familiar to them, which may contribute to such biases. In particular, Nuñez et al. (2022) found that that most of the invasion biology manuscripts that were submitted, reviewed, or accepted for the journal "Biological Invasions" arose from wealthier temperate regions, such the North America, Europe, and Australasia. There may also be biases in the types of plants being studied (Weidlich et al., 2020). In particular, certain taxonomic and functional groups of plants may also receive more attention due to being seen as more likely to become invasive. For instance, nitrogen fixers (most notably legumes within the Fabaceae) are widely recognized as powerful invaders (Daehler, 1998) because of their mutualisms with nitrogen-fixing bacteria (Richardson et al., 2000), and research into invasive plants has the potential to pay disproportionate attention to this group. Finally, there may be biases in the topics and types of questions that are being studied within invasive plant research, and these could potentially change across time in a manner that follows the trends in ecological research as a whole. Two no-  (Chown et al., 2015) and an increasing recognition of the importance of various plant physiological and functional traits in invasive plant success (Drenovsky et al., 2012).
Literature biases are a problem that needs to be examined in detail in any field of scientific research-such biases can impact on the theories, the outcomes of meta-analyses and syntheses, and our fundamental views of any field. For example, substantive geographic bias of field sites has been shown to be a serious issue in F I G U R E 1 Number of invasive species papers published within the Web of Science (https://www.webof scien ce.com) over time. Search terms used (all within topic) were: "invasion biology" OR "invasive species" OR "biological invasion" OR invasive+ecology OR invasion+ecology.  1972-1981 1982-1991 1992-2001 2002-2011 2012-2021 Number of Papers Published Time Range (by Year) ecological research in the Arctic (Metcalfe et al., 2018), studies on how food webs respond to global change (Cameron et al., 2019), and work on the relationship between biodiversity and ecosystem functioning (Clarke et al., 2017). Such biases may also prevalent in the field of invasion biology (Chong et al., 2021;Nuñez et al., 2022;Pyšek et al., 2008), and as a relatively novel field of research, it is of interest as to whether these biases have changed over time in this rapidly expanding research area. In this paper, we look at the papers published within the two major journals devoted exclusively to pub- By addressing these questions, we aim to offer insights on how research foci and questions change over time in a rapidly developing and growing field of study. Some previous studies have considered the geographic biases of studies on invasion biology (Chong et al., 2021;Nuñez et al., 2022;Pyšek et al., 2008), but how these biases have changed over time, or other biases in invasion biology (e.g., taxonomic bias, or the types of questions that are being most commonly addressed) remain largely unstudied. The field of invasion biology will only become more important over time as the signature of human activity on the Earth's ecosystems becomes ever more apparent. Understanding the patterns of how invasive plant research has changed over the past two decades can help identify critical knowledge gaps and biases and inform and help predict future research directions within the field.

| ME THODS
All papers published between 1999 and 2020 in the journals Biological Invasions and Neobiota were sampled in 5-year steps, with each step spanning two full years (i.e., 1999-2000, 2004-2005, 2009-2010, 2014-2015, 2019-2020). Papers were sampled in 5-year steps because sampling every published paper across the two journals at the level of detail that we examined them would be intractable, and this provided enough separate steps (i.e., five) to meaningfully look at trends across this time period. From these papers, all papers involving collection of original data from field experiments, greenhouse (or similar) experiments, or observational studies of invasive plants, and excluding literature reviews, meta-analyses, theoretical modeling studies, or studies with a primarily sociological focus, were selected for analysis, with a total of 458 papers selected in total (See Table 1 for breakdown of number of papers by year). The papers were read to ensure they satisfy these criteria, and details about them were tabulated. Observed trends were analyzed by fitting the category of paper as a logistic regression against time, and the statistical significance of the slope was assessed with a chi-squared test. All analyses were conducted using R v3.6.3 (R Core Team, 2020). Data corresponding to each of the four questions were tabulated as follows:

| Geographical distribution
Coordinates and location/region details of the location of each published study were collected as follows. First, the field site(s)/ collection source(s) where the invasive plants were found or grown in the study were recorded. If exact co-ordinates of the locations described above were available within the paper or its supplementary documentation, these co-ordinates were converted to decimal degrees using the WGS84 datum (if necessary). If no coordinates were available within the paper, but a map of the sites were provided, then approximate co-ordinates were obtained by viewing the same area on Google Maps and recording the coordinates of the area indicated by the map. If no co-ordinates/ maps were available, or if there were too many (>50) co-ordinates indicated on a map, a single co-ordinate positioned in the center of the study area was used. A cluster analysis was performed with the Euclidean distances between all pairs of location co-ordinates, and the resulting dendrogram was cut to produce 45 unique clusters of locations. The latitude and longitude of the centroid of each cluster were then calculated from all the locations in that cluster and plotted onto a world map. Location data were also cat-  12  4  15  20  47  100  56  58  81  65 against time in 5 year steps. As studies may contain more than one location being studied, only unique paper location/co-ordinates combinations were analyzed/plotted.

| Climatic zone and ecosystem type
Climate and ecosystem type for each paper was either taken directly from the papers or inferred based on the location data.

| Plant taxonomic and functional identity
For this analysis, we looked only at papers studying specific invasive plant taxa, excluding observational community-level or survey studies. The taxonomic identity of the plant species that were studied was collected directly from each paper, and taxonomic synonyms were resolved by using the species name most commonly accepted within the scientific community for that species. Plant species in each paper were also assigned to broad functional types using the USDA plant database, with the following commonly accepted categories: Forb, Graminoid, Shrub, Tree, Vine, and (nitrogen-fixing) Legumes.
Nitrogen-fixing legumes were separated out as their own functional type because of the major role of this group as an ecosystem driver.
The number of studies in each taxonomic order and functional type were plotted against time. As studies may contain more than one species or functional type being studied, only unique paper-species/ functional type combinations were plotted. Unique paper-species combinations were used to account for different invasive plant species within the same order in the same paper.

| Types of topics being studied
The types of topics being studied in each paper were determined by reading the abstracts and contents of the papers and broadly classified into the following subcategories: Invader Establishment,

| Geographical distribution
The vast majority (

| Plant taxonomic and functional identity
The invasive plants that have been studied were generally diverse, with 34 total orders represented in the data set. Six orders (Asterales, Caryophyllales, Fabales, Lamiales, Poales, and Rosales) represented between 60% and 70% of all invasive plants studied  Table 2).

The Astarales/Asteraceae (Daisies and related plants) and
Poales/Poaceae (grasses, sedges, and reeds) are consistently among the most commonly studied invasive plants across time, and there is an increasing trend for scientists to study invasive legumes (i.e., Fabaceae) (χ 2 = 6.2, df = 1, p = .001) (Figure 7). These patterns are reflected in the data of invasive plant growth habits/functional types, which is consistently dominated by forbs, shrubs, and grasses (graminoids) and a more recent increase in the proportion of invasive legume studies (χ 2 = 19.4, df = 1, p < .001) (Figure 8). Of the other plant types, there was a distinct increase in number studies on invasive trees up until 2009-2010 data set, but with a decline thereafter.

| Types of topics being studied
There are clear trends in the type of topics being studied in invasive plant papers across time (Table 3). There is decreasing interest in establishment-type questions (i.e., "How did the invasive plant get/ survive here?") (χ 2 = 11.5, df = 1, p < .001), coupled with a nonstatistically significant increase in impact-type questions (i.e., "What is it doing to invaded ecosystems?") (χ 2 = 2.6, df = 1, p = .104). There as well as toward invasion biology (Liu et al., 2012;van Wilgen, 2020), and there could also be an increasing awareness of the problems invasive plants can cause in these regions (Fuentes et al., 2010;Speziale & Lambertucci, 2010;van Wilgen & Wilson, 2018). This is an important trend in this field, as invasive plants are projected to become a greater threat in developing countries due to globalization (Chong et al., 2021;Seebens et al., 2015Seebens et al., , 2021 and climate change (Dullinger et al., 2017), and a better understanding of the processes underlying plant invasions within these regions is crucial to understanding the ecology of plant invasions worldwide.

| Climatic zone and ecosystem type
The vast majority of invasive plant studies across the 22-year period of our analysis has been conducted in Mediterranean, subtropical, and temperate climatic zones. This is reflective of the geographical biases toward studies conducted in North American, European, and Australian regions, and these regions also account for most studies in arid climates (i.e., mostly in North American deserts) and cold climates (i.e., mostly alpine, subarctic, and boreal sites in North America and Europe). There is a conspicuous dearth of studies conducted in tropical climates, with only a recent uptick in 2019/2020.
While this may represent logistical issues in studying tropical field sites, it also demonstrates the relatively unknown extent of plant invasions and limited research performed within the tropics (Chong et al., 2021). While some available literature proposes that tropical ecosystems, being highly biodiverse habitats, may not be as invasible (i.e., easily invaded) (Lonsdale, 1999;Rejmánek, 1996;Teo et al., 2003; but see Chong et al., 2021;Fine, 2002), the shortage of studies conducted in tropical ecosystems make it difficult to determine the extent to which invasive plants are an issue in the tropics. This is further complicated by the fact that many existing tropical studies on invasive plants have been conducted on tropical islands, notably the Hawaiian Islands (Chong et al., 2021;Weidlich et al., 2020), which are unrepresentative of most tropical landmasses and may have very different ecological characteristics to continental tropical ecosystems (Essl et al., 2019). Tropical systems are also facing increased and substantial ongoing anthropogenic threats and disturbances (Barlow et al., 2018;Fine, 2002;Gardner et al., 2009;Laurance, 2013), and the resulting plant invasions may have a time lag of several decades before becoming apparent through an "invasion debt" (Chong et al., 2021;Seebens et al., 2015).
In terms of ecosystem types, most invasive plant research over the 22 years of our analysis has consistently been conducted within forests and grasslands. This may reflect ease of access and logistics, or that grasslands and forests receive more funding for invasion research due to public interest and/or economic reasons such as asset protection (Hiatt et al., 2019). While the proportion of studies being conducted across different ecosystems have remained fairly constant over time, many understudied systems (such as urban/anthropogenic and desert systems) have received increased research interest in recent years, reflecting awareness that few ecosystem types or habitats are immune to plant invasion.

| Plant taxonomic and functional identity
While there is a very large number of potentially invasive plant species, our analysis shows the vast majority of invasive plant research at least within the two journals devoted exclusively to invasion biology (Biological Invasions and Neobiota) has consistently focused on smaller plants such as forbs, shrubs, and grasses. Such plants often have traits that enable invasion, such as being easily dispersed and growing relatively quickly to maturity (Martín-Forés et al., 2018). They are also easily studied by researchers-the same traits that make them good invasive plants also make them easy to manipulate and grow in field and greenhouse experiments. In contrast, there has been a decline in invasive tree studies since 2009/2010-the longer-lived nature of trees may make them less amenable to short-term experimental studies within the duration of conventional funding cycles.
There has also been a steady increase in studies in nitrogenfixing plants (i.e., legumes within Fabaceae) over the 22 years covered by our analysis. Nitrogen fixation is well recognized as a major factor enabling invasion success, and nitrogen-fixing species feature disproportionately in many invasive floras (Daehler, 1998;Richardson et al., 2000). These species are also of interest to researchers because of their major role in driving plant-soil feedbacks (van der Putten et al., 2013;Wardle et al., 2004) and their transformative effect on soil nutrient cycles and ecosystem processes (Vitousek et al., 1987;Vitousek & Walker, 1989;Wardle & Peltzer, 2017). In particular, nitrogen fixers do especially well in N-limited and P-replete systems (Smith, 1992;Zheng et al., 2016), which are common in early successional and heavily disturbed systems (Craft, 1996;Smith et al., 2000), and invasive nitrogen fixers may therefore be well adapted to invade and dominate early in successional processes. topic in plant ecology research in recent years (Cornwell et al., 2008;de Bello et al., 2010;Diaz et al., 2004;Pantel et al., 2017), and this is accompanied by a continual increase in trait-focused studies in invasion biology over the past 20 years in our analysis. This literature is often connected to the idea that invasive plants have more acquisitive traits than do native species (Mathakutha et al., 2019;van Kleunen et al., 2010) that may impact on ecosystem processes including those in the soil (Vilà et al., 2011;Wardle & Peltzer, 2017), and that they possess phenological traits that enable faster, more widespread dispersal (Martín-Forés et al., 2018). The field of invasion biology will only become more important over time, and focusing invasive plant research on understudied and threatened ecosystems will allow us to develop a more holistic understanding of the ecological dynamics of invasive plants. In particular, given the outsize importance of the tropics to global biodiversity, the threats that their ecosystems face, and the dearth of studies in tropical regions, it is of critical importance that more invasive plant research is conducted within the tropics to better understand the threats they pose there. This study looked only at the existence of biases over time within invasive plant research, and this approach does not fully cover the many structural issues (such as logistics, lack of funding, and other reasons) that may cause such biases and hinder research in the tropics and other understudied areas (see Chong et al., 2021). It is therefore critical that the organizations and people (such as researchers, journals, and funding organizations) that enable scientific research understand these limitations, as well as work to reduce or eliminate the hurdles involved in carrying out such research in underrepresented regions. We conclude that a greater presence of underrepresented regions, ecosystems, and taxa in the literature is also essential in ensuring that the conclusions drawn from literature syntheses and meta-analysis and theory about invasion biology will not only be applicable or true for a subset of the world's ecosystems and invasive biota.

ACK N O WLE D G E M ENTS
We would like to thank two anonymous reviewers for their helpful comments and suggestions for our paper.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request (upon acceptance the raw data will be deposited in Dryad and the https://doi. org/10.5061/dryad.wdbrv 15ss. Data is currently being uploaded to the above DOI on Dryad).