Early detection and eradication of invasive plants are more cost-effective than managing well-established invasive plant populations and their impacts. However, there is high uncertainty around which taxa are likely to become invasive in a given area. Horizon scanning that combines a data-driven approach with rapid risk assessment and consensus building among experts can help identify invasion threats. We performed a horizon scan of potential invasive plant threats to Florida, USA—a state with a high influx of introduced species, conditions that are generally favorable for plant establishment, and a history of negative impacts from invasive plants. We began with an initial list of 2128 non-native plant taxa that are known invaders or crop pests. We built on previous invasive species horizon scans by developing data-based criteria to prioritize 100 taxa for rapid risk assessment. The semi-automated prioritization process included selecting taxa “on the horizon” (i.e., not yet in the target location and not on a noxious weed list) with climate matching, naturalization history, “weediness” record, and global commonness. We derived overall invasion risk scores with rapid risk assessment by evaluating the likelihood of each of the taxa arriving, establishing, and having an impact in Florida. Then, following a consensus-building discussion, we identified six plant taxa as high risk, with overall risk scores ranging from 75 to 100 out of a possible 125. The six taxa are globally distributed, easily transported to new areas, found in regions with climates similar to Florida’s, and can impact native plant communities, human health, or agriculture. Finally, we evaluated our initial and final lists for potential biases. Assessors tended to assign higher risk scores to taxa that had more available information. In addition, we identified biases towards four plant families and certain geographical regions of origin. Our horizon scan approach identified taxa conforming to metrics of high invasion risk and used a methodology refined for plants that can be applied to other locations.

certainty, consensus building, Florida, horizon scan, invasion, prevention, rapid risk assessment

Invasive species can negatively impact ecosystems, economies, and human health (CBD 2009). Managing potential impacts of invasive species, and invasive plants in particular, is daunting given the many species introduced to novel areas each year, with rates predicted to increase in the future (Seebens et al. 2017). When governments and private landowners take action, they often manage invasive plants and mitigate negative impacts after establishment. However, preventing the introduction and initial spread of invasive plants is generally more effective and avoids potential ecological and economic losses (Keller et al. 2007; Sheley et al. 2015). Unfortunately, the benefits of prevention are difficult to quantify and involve high uncertainty, making post-invasion control the more common approach (Finnoff et al. 2007; Early et al. 2016). Thus, programs that help identify which non-native plant taxa have a high probability of becoming problematic invaders are essential for providing the first line of defense against plant invasions.

Horizon scanning is the systematic search to identify potential threats, emerging issues, and opportunities that can inform research and action (Sutherland and Woodroof 2009; Amanatidou et al. 2012). The goal of horizon scanning in conservation science is to preemptively identify threats so researchers can provide timely and informed input on policy and decision-making (Sutherland and Woodroof 2009). In Europe, horizon scanning of emerging invaders has involved acquiring lists of potentially invasive species for a specific region, assessing the likelihood of arrival, establishment, and impact for each species, and, in some cases, building consensus among experts around a list of species ranked by risk (Parrott et al. 2009; Matthews et al. 2014; Roy et al. 2014; Gallardo et al. 2016; Lucy et al. 2020). These horizon scans have informed policy and guided resource allocation towards research and prevention efforts.

Florida is one of the most important states for regulating invasive plants in the United States because nearly 85% of all non-native plants imported to the contiguous United States enter through one of Florida’s shipping ports or airports (Gordon and Thomas 1997). As international trade continues to grow, so too does the frequency of intentional and accidental introductions (Early et al. 2016). In addition to being an entry point for invasive species to the rest of the country, Florida is particularly vulnerable to the establishment of invasive plants due to its tropical/subtropical climate and diverse ecosystems (Simberloff 1997; Pyšek et al. 2017). Management of invasive plants in Florida’s conservation areas costs nearly $45 million annually (Hiatt et al. 2019) and invasive species (including plants, insects, and pathogens) cost Florida’s agriculture industry at least $179 million annually (Coffman et al. 2001). Identifying potential invaders before or soon after they enter Florida can reduce ecological and economic losses to the state as well as prevent the spread of invasive plants nationally.

Here, we developed a horizon scan approach to create a ranked list of non-native plants that are likely to arrive and establish in Florida and have impacts on native biodiversity, the economy, or human health in the near future. We started with a large initial list of plant taxa that were associated with invasion. We then developed criteria and used publicly available datasets to prioritize taxa for risk assessment. This step builds on previous horizon scans, which were able to assess all taxa on initial lists. We present a ranked list of potential invasive plant threats to Florida, which can be used to inform research, management, and policy aimed at reducing invasive plant impacts.

This horizon scan was part of the Florida Invasion Threats Horizon Scan, which assessed invasion threats of freshwater and terrestrial plants (reported here), marine taxa, freshwater invertebrates, terrestrial invertebrates, and non-marine vertebrates (Lieurance et al. in review). We adapted and revised the horizon scanning method outlined by Roy et al. (2014, 2015) to develop a ranked list of invasive plant threats and their potential pathways for arrival to the target location (Florida) in the near future (e.g., 5–15 years). We chose this time frame to prioritize upcoming threats, to establish a minimum frequency for updating the horizon scan with new information (once every 5–15 years), and to evaluate risk within current climate conditions (i.e., omitting future climate change scenarios). We kept this time frame in mind by considering current arrival pathways and environmental conditions in the target location.

Assessing and scoring the taxa

Nine assessors evaluated taxa using a rapid risk assessment tool modified from Roy et al. (2014). First, we used a species not included in the assessment list to evaluate the tool for clarity, timing, and assessment consistency. Then, we completed risk assessments with a standardized set of resources (Suppl. material 3). Because the risk assessments are designed to be completed rapidly, we aimed to spend less than two hours on each taxon.

We identified one or more potential pathways for taxa to arrive in Florida based on an established framework (Hulme et al. 2008; CBD 2014; Harrower et al. 2018). Briefly, the pathways included “release in nature” (intentional release, such as for erosion control), “escape from confinement” (intentional commodity that escapes, such as a horticultural taxon), “transport contaminant” (associated with the transport of a specific commodity, such as a seed contaminant), “transport stowaway” (other forms of unintentional transport, such as through soil on equipment), “corridor” (through human infrastructure linking previously unconnected areas, such as a waterway), and unaided (natural dispersal).

We scored the likelihoods of arrival, establishment, and negative impacts (environmental, socioeconomic, and human health) on a scale of 1 (very low) to 5 (very high; Fig. 1). To estimate the likelihood of arrival, we considered the current distribution of the taxon, the availability of the taxon for purchase, history of invasion by the taxon in other regions, and the presence of a plausible arrival pathway (Table 1).

Table 1.

Download as

CSV

XLSX

Rubrics for scoring likelihood of arrival, establishment, and impacts of potential invasive plants.

Category	Criteria	Score
Arrival†	Closest observation to target location‡ and closest online seller to target location are outside of region§.	1
	Closest observation to target location is within region, but not nearby§, and closest online seller to target location is outside of region.	2
	Closest observation to Florida and closest online seller to target location are within region, but not nearby or closest observation to target location is nearby, but not in target location, and closest online seller to target location is outside region.	3
	Closest observation to target location is nearby, but not in target location, and closest online seller is within region or nearby, but not in target location.	4
	The taxon has been observed or sold within target location.	5
Establishment†	No observations in areas with matching Köppen-Geiger (KG) zones to target location|.	1
	Few observations in one area with matching KG zones to target location.	2
	Many observations in one area or few observations in multiple areas with matching KG zones to target location.	3
	Many observations in multiple areas with matching KG zones to target location.	4
	Criteria for score 4 plus evidence of a biological strategy that aids establishment or evidence of establishment in target location.	5
Impact	Unlikely to cause negative impacts on the native biota or abiotic environment, human well‐being, or economic systems.	1
	Likely to cause (a) declines in the performance (e.g., biomass, body size) of native biota, but no decline in native population sizes or (b) income loss, minor health problems, higher effort or expense to participate in activities, increased difficulty in accessing goods, or minor disruption of social activities, but no significant impact on participation in normal activities.	2
	Likely to cause (a) declines in the population size(s) of native species, but no changes to the structure of communities or to the abiotic or biotic composition of ecosystems or (b) changes in the size of social activities, with fewer people participating, but the activity is still carried out. These changes to social activities could be linked to accessibility to the activity area or mild effects to human health (e.g., allergies).	3
	Likely to cause (a) the local or population extinction of at least one native species, leading to reversible changes in the structure of communities, the abiotic or biotic composition of ecosystems or (b) the local disappearance of a social or economic activity from all or part of the area invaded by the alien taxon, collapse of the specific activity, switch to other activities, abandonment of activity without replacement, emigration from region, or moderate effects to human health.	4
	Likely to cause (a) the replacement and local extinction of native species and will produce irreversible changes in the structure of communities and the abiotic or biotic composition of ecosystems or (b) local disappearance of a social or economic activity from all or part of the area invaded by the alien taxon or major effects to human health.	5

^† Arrival and Establishment rubrics were applied during the second review phase. Scores were adjusted by up to one point based on additional information in the assessments. ^‡target location = Florida, United States; observations based on GBIF.org (2020) or information provided by the assessor or reviewer. ^§For our purposes, “region” is contiguous United States and “nearby” are the states of Georgia, Alabama, South Carolina, North Carolina, Tennessee, and Mississippi. ^|Observations based on GBIF.org (2020) or information provided by the assessor or reviewer. Florida’s Köppen-Geiger zones include Af, Am, Aw, and Cfa (Kottek et al. 2006).

Figure 1.

Methods for selecting and evaluating taxa as invasive plant threats for a target location (Florida, United States). Data-based list processing led to the prioritization of 100 taxa for risk assessment. Rapid risk assessments performed by an expert panel included pathways for arrival and likelihood scores and certainty ratings for arrival, establishment, and impact. The three component likelihood scores were multiplied to get an overall score and certainty ratings were roughly averaged to get an overall certainty. Each risk assessment was evaluated with two rounds of review and a consensus-building discussion before the expert panel confirmed taxa rankings.

To estimate the likelihood of establishment (i.e., developing a self-sustaining population), we considered the distribution and number of records of the taxon within regions with Köppen-Geiger climate zones matching Florida (Table 1). This evaluation expands on the use of Köppen-Geiger climate zones to select taxa for our assessment list, in which records in only one matching location were needed to pass the criterion (Fig. 1). We also considered ecological properties of both the taxon and target location habitats, including time to reproductive maturity, reproduction rate, dispersal mechanism, propagule pressure, tolerance of a broad range of environmental conditions, resource availability, natural enemies, and amount of nurturing required (e.g., weeding, irrigation, fertilization, pest control; Petri et al. 2021). Geographic thresholds for arrival and establishment likelihood scores (Table 1) were chosen based on distance, ease of movement through ground transportation, and low barriers to introduction by travel or mail (USDA APHIS 2017a, b).

To estimate the likelihood of negative impacts, we used a scoring rubric modified from the Invasive Species Environmental Impact Assessment protocol (Branquart 2009), the Environmental Impact Classification of Alien Taxa (Blackburn et al. 2014; Hawkins et al. 2015), and the Socio-Economic Impact Classification of Alien Taxa (Bacher et al. 2018; Table 1). The overall risk score was the product of arrival, establishment, and impact likelihood scores (Fig. 1; Roy et al. 2015). We provided brief justifications for our scores and assigned certainty ratings that ranged from very low (i.e., all scores were equally likely) to high (i.e., could confidently eliminate all other scores). The overall certainty rating was the rating most consistent with three component certainty ratings (Suppl. material 3).

Data and code are available at https://doi.org/10.5281/zenodo.6211243.

Our horizon scan of invasive plant threats to Florida identified six taxa that have a high risk of becoming invasive in the state in the near future (5–15 years). The horizon scanning process helped us identify taxa that should undergo more thorough risk assessments and potentially receive policy restrictions or research priority. Our reliance on existing databases allowed us to quickly evaluate many taxa in a manner than can be applied to future horizon scans. Further, we used this case study to assess biases in the horizon scan process that should be taken into consideration in future horizon scans of invasive plants.

Although we used databases to reduce the number of taxa on our list, it was necessary to use expertise to perform rapid risk assessments, review, and consensus building. These expert-based processes are therefore not repeatable, but we aimed to increase transparency by providing the assessments and reviews (Suppl. material 3). The identity of the assessor significantly affected the overall risk scores. Two assessors who had 3–8 years of risk assessment experience scored taxa higher on average than two assessors who had less than one year of risk assessment experience. Because our sample size of assessors is small, we are unsure whether this outcome is coincidental (due to the taxa assessed by these individuals) or due to assessor experience. To address differences in experience, future horizon scans could calibrate scores among assessors with a set of test taxa, a more rigorous approach than our calibration with a single taxon, or derive composite scores from multiple assessors, for example through structured expert judgement (Wittmann et al. 2015). Discrepancies in experience highlight the importance of rubrics, peer review, and consensus building; although experience may have influenced assessors during the risk assessment phase, all assessors agreed on the final ranking of taxa.

Overall risk scores were positively related to overall certainty ratings. We hypothesize that this occurred because more available data can contribute to higher certainty and provide more evidence that a taxon may arrive, establish, or have impacts. Similarly, risk scores were negatively related to the year of the earliest U.S. record. We hypothesize that taxa with earlier and more records of occurrence in the United States are likely to be better represented in English-language texts than less common or more recently detected taxa, leading to more evidence for arrival, establishment, and impacts. Efforts to synthesize and standardize information about invasive species (Simpson et al. 2019; CABI 2021) could reduce these potential sources of bias. The relationships between risk scores and earliest record (negative) and number of records (positive) may also indicate that taxa with longer residence time and larger population sizes, respectively, have greater risk of arrival, establishment, and impact (Pyšek et al. 2009).

We evaluated taxonomic and geographic biases in the final horizon scan list and taxonomic biases in the initial CABI list. These biases may indicate shared characteristics of invasive plants or cultural biases in the CABI databases. While we cannot distinguish between these two causes, we look to previous studies for insights. The families Juncaceae (rushes), Poaceae (grasses), Polygonaceae (knotweeds), and Rosaceae (roses) were significantly overrepresented in both the final horizon scan list and the initial CABI list compared to the number of taxa in these families. These families are similarly overrepresented in global lists of naturalized plants (Daehler 1998; Pyšek et al. 2017). The overrepresented families may indicate shared characteristics of invasive plants. Taxa in these families are characterized by traits that can aid invasion, including high reproduction, broad environmental tolerance, and high human use frequency (Hummer and Janick 2009; Canavan et al. 2019; Ashby et al. 2020). In addition, mis-identified invasive rushes and grasses may go undiscovered for long periods, allowing them to establish self-sustaining populations before being controlled (Scott and Hallam 2003). Such general trends can help identify families on which to concentrate risk assessment resources.

Most of the taxa that made our final list were native to Europe, Asia, and North Africa. This result is likely a combination of shared characteristics of invasive plants and cultural biases in the initial CABI list. Europe is the native range for a disproportionately high number of naturalized plant species relative to the number of native plant species (van Kleunen et al. 2015), which may be influenced by plant adaptations to European pastoralism and cultivation—practices that have been widely adopted (MacDougall et al. 2018)—and historical exchange between Europe and other geographic regions (Pyšek et al. 2015). Temperate Asia is also a major source of global naturalized plant species (van Kleunen et al. 2015). Because Florida’s Köppen-Geiger climate zones most consistently overlap with Central and South America, central Africa, and southern and eastern Asia (Kottek et al. 2006), our final list likely omits key high risk taxa. Further, the scoring systems for arrival and establishment likelihoods may better estimate risk by including key locations outside of the contiguous U.S., such as Hawaii, Puerto Rico, and the U.S. Virgin Islands. Future horizon scans could focus more on taxa from geographic regions with a similar climate to the target location and strong trade and tourism ties. Although we did not evaluate the geographic ranges of taxa on the initial CABI list, our results from the final list indicate that this analysis could be an important initial step of the horizon scan process to identify whether invasive or naturalized species lists from underrepresented geographic regions need to be obtained.

Overall scores were calculated by multiplying likelihoods of arrival, establishment, and impact (Roy et al. 2015). By equally weighting these three processes, we assumed that each was crucial to a taxon becoming invasive (Blackburn et al. 2011, 2014; Bacher et al. 2018). Four taxa (Ligustrum vulgare, Cytisus scoparius, Phalaris arundinacea, and Avena fatua) had high overall risk scores with high certainty. Although we did not independently validate these results, staff at the University of Florida (including and trained by one of the authors) assessed these taxa with a more rigorous 49-question predictive tool and found them all to be high invasion risks (University of Florida, Institute of Food and Agricultural Sciences 2018). In our horizon scan, two taxa (Agrostis capillaris and Persicaria hydropiper) had high overall risk scores, but medium certainty. Because we were unsure how A. capillaris would fare in competition with native Florida grasses, competition studies could increase certainty. Similarly, agricultural impact studies of P. hydropiper, which interferes with crops and grazing in other regions, could increase certainty about the risk of this taxon. Taxa with high overall risk scores are included in noxious weed lists or laws for 1–14 states (EDDMapS 2021), raising the question of whether they have already arrived in Florida, but failed to establish. While we considered many environmental factors and plant traits in our assessment of establishment likelihood, establishment experiments may be valuable in informing how much resources should be allocated to preventing invasion of these taxa. On the other hand, the arrival and establishment of these taxa in Florida may be in a lag phase (Taylor and Hastings 2005; Aiko et al. 2010).

We identified “escape from confinement” as the most likely pathway for taxa on our final list to arrive in Florida’s natural areas, which is consistent with a global analysis of invasive plants (Hulme et al. 2008). This pathway includes escape from agriculture, botanical gardens, forestry, research facilities, horticulture, and ornamental purposes other than horticulture (CBD 2014). Domestication can select for traits that increase invasion risk, including fast growth rates, high fecundity, and the ability to hybridize (Petri et al. 2021). However, selection for traits that reduce invasion risk and do not interfere with the commercial purposes of plants could help prevent escape from confinement (Petri et al. 2021).

Taxa on our final list were also likely to arrive in Florida’s natural areas as transport contaminants or transport stowaways. Florida’s seaports are some of the most active in the country (U.S. Army Corps of Engineers 2018), hosting international and domestic trade, as well as millions of cruise passengers (Florida Department of Transportation 2017). Florida is also a top tourist destination, attracting well over 100 million visitors each year (VISIT FLORIDA 2020). These high movement rates provide ample opportunities for plant propagules to enter the state. The risk of introducing taxa through trade routes, however, can be mitigated by identifying steps in the process of importing, processing, and storing goods that can be modified to reduce plant survival (Hulme 2009).

This horizon scan of invasive plant threats to Florida provides a first step in reducing the impacts of invasive species on Florida’s natural systems. Like other horizon scans of potential invasive species, the generated list informs future research efforts and policy (e.g., Matthews et al. 2014; Roy et al. 2014; Gallardo et al. 2016; Lucy et al. 2020). Our horizon scan builds on previous invasive species horizon scans, however, in important ways. First, we began with a list of 2128 potential invasive taxa, which was too large a list to perform rapid risk assessments (approximately 2 hours each) in a reasonable timeline (approximately one year between initial workshop and all taxa consensus building; Lieurance et al. in review). We therefore developed data-based criteria to filter the list to 100 taxa. The databases and code we used are publicly available (Suppl. material 1, Kendig et al. 2022) and could be used for other horizon scans of potential invasive plants. Second, the rapid risk assessments and peer reviews led to enough consensus among experts that our final rankings relied entirely on scores from that process (e.g., in contrast to Roy et al. 2014; Lucy et al. 2020). Consensus building led to important methodological changes (i.e., removing a taxon with too much uncertainty, revisiting assessments with arrival and establishment rubrics), but did not directly alter the rankings. A major advantage of this approach is that the rapid risk assessment tool and rubric can increase transparency of the horizon scan process, especially as they become more refined with future horizon scans.

Here we presented a horizon scan of 2128 plant taxa, identifying six with a high invasion risk for Florida in the near future and 93 with medium or low invasion risk. The horizon scan process therefore can potentially reduce the number of taxa requiring thorough risk assessments by three orders of magnitude. The results provide researchers, regulators, and private and public land managers with a practicable list of high risk taxa to focus on. Given the substantial impacts and costs of invaders in Florida, the ability to differentiate and focus efforts on high probability threats is critical.

We thank Dale Laughinghouse and Seokmin Kim for help with processing the list, Julie Lockwood and Helen Roy for guidance on the horizon scan process, and all other participants of the Florida Invasion Threats Horizon Scan for constructive discussions. We thank Jane Molofsky, Gerry Moore, Curtis Daehler, and an anonymous reviewer for helpful feedback on the manuscript. Fig. 1 created with BioRender.com. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

We thank the Florida Fish and Wildlife Conservation Commission and UF/IFAS Dean for Research for funding this project.