A peer-reviewed open-access journal @) NeoBiota Advancing research on alien species and biological invasions NeoBiota 99: 341-362 (2025) DOI: 10.3897/neobiota.99.153010 Research Article Invasive species risk assessment in practice: Insights from a survey of practitioners Susan Canavan'™®, Kim Canavan22®, Sabrina Kumschick**®, Doria R. Gordon®”®, John R. U. Wilson*=®, Deah Lieurance®® eaN Do FP WwW DY School of Natural Sciences, Ollscoil na Gaillimhe—University of Galway, Galway, Ireland Centre for Biological Control, Department of Zoology and Entomology, Rhodes University, Makhanda, South Africa Department of Plant Sciences and Afromontane Research Unit, University of the Free State, Phuthaditjhaba, South Africa Centre for Invasion Biology, Department of Botany and Zoology, Stellenbosch University, Stellenbosch, South Africa Kirstenbosch Research Centre, South African National Biodiversity Institute, Cape Town, South Africa Environmental Defence Fund, New York, NY, USA Department of Biology, University of Florida, Gainesville, FL, USA Department of Ecosystem Science and Management, The Pennsylvania State University, State College, USA Corresponding author: Susan Canavan (sucanavan@gmail.com) OPEN Qaceess Academic editor: Ingo Kowarik Received: 14 March 2025 Accepted: 5 June 2025 Published: 11 July 2025 Citation: Canavan S, Canavan K, Kumschick S, Gordon DR, Wilson JRU, Lieurance D (2025) Invasive species risk assessment in practice: Insights from a survey of practitioners. NeoBiota 99: 341-362. https://doi. org/10.3897/neobiota.99.153010 Copyright: This is an open access article distributed under the terms of the CCO Public Domain Dedication.. Abstract Risk assessment is an important tool in invasive species prevention and management, providing a structured approach to identify and evaluate the risks posed by non-native species. Despite the wide- spread development of risk assessment and risk analysis (collectively referred to as RA here) methods, there is a lack of information on which methods are used in practice. We conducted a global survey of RA practitioners from diverse regional, professional, and taxonomic contexts to identify the tools and databases used, the qualifications and experience of assessors, and the implementation and accessibility of RA results. 107 responses were received from practitioners focussed on all continents except Antarc- tica, with the most from the United States and the United Kingdom. Respondents reported using more than 46 different RA tools, with the Fish Invasiveness Screening Kit and the USDA Plant Protection and Quarantine Weed Risk Assessment the most commonly mentioned tools, based on the number of users. Plants were the most frequently assessed taxonomic group and terrestrial species, though repre- sentation from all taxonomic groups and systems exist. Respondents listed 107 generally open-access databases that they frequently use to conduct RAs, with the most commonly used sources being the CABI Invasive Species Compendium, occurrence records from the Global Biodiversity Information Facility (GBIF), and taxonomic information from the Integrated Taxonomic Information System (ITIS). Assessors typically had tertiary education, with the majority holding at least a master’s degree, though many did not believe that a post-graduate degree was necessary to be an effective assessor. After RAs were completed, assessments were predominantly reported to government agencies. Most finalized RAs included some measure of certainty and were usually publicly accessible, though few included a formal process for stakeholders to comment. Almost all respondents identified the importance of train- ing and certification programs to standardize qualifications for assessors. Based on the views expressed in the surveys we discuss the importance of: (1) training and capacity building, (2) open access data- bases and FAIR (Findable, Accessible, Interoperable and Reusable) data standards, (3) incorporating stakeholders in the process, and (4) standardisation of tools; if these measures are implemented, they may enhance the consistency, transparency, and effectiveness of RAs of non-native species. Key words: Alien species, biosecurity, GBIE impacts, Invasiveness Screening Kit, ITIS, non-native, weed risk assessment 341 Susan Canavan et al.: Risk assessment in practice Introduction Risk assessments for non-native species aim to predict the potential that species may establish and negatively impact natural and human systems. Risk assessments facilitate systematic evaluation or forecasting of the level of threat non-native spe- cies pose pre- or post-introduction. They have become a standard practice in many countries, with agencies and organizations using them to guide management de- cisions (e.g., EPPO 2011; USDA APHIS PPQ 2020), and consequently, assess- ments related to species introductions have been incorporated into international trade and environmental policy agendas, quarantine legislation, and procedures at national borders. For example, the International Standards for Phytosanitary Measures (ISPM) and the World Organisation for Animal Health (WOAH) have established standards and guidelines for risk assessments to restrict the movement of pests and protect human and animal health (FAO 2007a, 2007b). Risk assessment methods vary in design, scope, and intended use, being tailored to different taxa, regions, ecosystems, and pathways. Some focus on screening spe- cies for their potential invasiveness, while others provide detailed assessments of specific species, including full impact categorization, pathway analysis, and species distribution mapping. Examples include the Aquatic Species Invasiveness Screen- ing Kit (AS-ISK) for aquatic systems (Copp et al. 2016) and the Australian Weed Risk Assessment (AWRA) for plants (Pheloung et al. 1999). Region-specific assess- ments include the EPPO framework for Europe (EPPO 2011) and the Canadian Food Inspection Agency (CFIA) pest risk analysis (CFIA 2012). However, there is often a lack of consolidation and continuity among these approaches, with varying terminology (Roy et al. 2018). Terms such as “protocols”, “frameworks”, “kits”, “schemes”, and “systems” can incorporate elements of risk assessments to varying degrees, with some terms used interchangeably, though the term “risk analysis” typically encompasses risk assessment, risk management, and risk communication (e.g., FAO 2007a and b; Kumschick et al. 2020b). All of these incorporate a pre- dictive process intended for the management of biological invasions. For the sake of brevity, we will collectively refer to these tools and methods as “RAs”. The peer-reviewed literature includes numerous publications on RAs, offering various perspectives on the utility and function of the process (e.g., Leung et al. 2012; Kumschick and Richardson 2013; Roy et al. 2018; Lieurance et al. 2024). While the development of RA tools is well documented in academic literature, whether and how they are actually used in practice is less clear, as this often takes place in grey literature or internal agency processes and is not systematically tracked (Roy et al. 2018; Dean et al. 2024). As a result, it remains unclear which tools are used, by whom, and for what purposes. Understanding how RAs are applied is important, as previous comparative studies have shown that inconsistent tool use, limited stakeholder involvement, and uneven training can undermine transpar- ency, repeatability, and comparability in outcomes (McGeoch et al. 2016; Roy et al. 2018). Key questions include: Who develops these methods, and who actually implements them? What expertise do practitioners need, and what training should they have? Which taxa and ecosystems are being prioritized for assessment, and which data sources underpin these evaluations? Equally important, particularly for decision-makers and affected stakeholders, is understanding how clearly and effectively RA results are communicated. NeoBiota 99: 341-362 (2025), DOI: 10.3897/neobiota.99.153010 342 Susan Canavan et al.: Risk assessment in practice In this study, our goal was to gain a practical understanding of the real-world application of RAs as applied to non-native species. We conducted a global sur- vey of practitioners from diverse regional, professional, and taxonomic contexts to identify the tools and databases they use, the qualifications they hold, and how RAs are implemented and reported. Methods Survey development We developed a 29-question survey to gather information directly from risk prac- titioners, including a list of RA programs (including both risk assessment and risk analysis), databases, and other sources of information, as well as a practical descrip- tion of RA in practice. The survey was composed of various questions (Table 1; See Suppl. material 1: table S1 for full survey) Throughout the data collection process, standard procedures of survey design used in similar studies were followed (e.g., Gozlan et al. 2013). The survey was confidential, and respondents’ anonymity was maintained. The questions included a mixture of list-all-that-apply, check-all-that-apply, Likert-scaled, multiple choice, and open-ended questions (i.e., fill in the blank), with the survey taking an esti- mated 10-15 minutes to complete. The survey was approved by the University of Florida’s Institutional Review Board, ID #202001808. We used three approaches to solicit responses. 1) The survey was distributed on 17 November 2020, through two listserv groups: Ecolog-L, hosted by the Ecological Society of America, and Aliens-L, hosted by the Invasive Species Spe- cialist Group (ISSG) of the IUCN Species Survival Commission. At the time of distribution, Ecolog-L had approximately 27,000 subscribers, and Aliens-L had 1,470 subscribers. Both groups primarily consist of students, academics, and practitioners working in the fields of ecology (Ecolog-L) and invasive species (Aliens-L). 2) The survey was shared on Twitter on 17 November 2020 initially on two main accounts (@ifasassessment and @drdeahlieurance) and then shared from there by other users. The Twitter posts promoting the survey received a combined 16,278 impressions (i.e., times a user was served a tweet in their time- line or search results), excluding impressions from approximately 62 retweets. And 3) direct invitations were sent to corresponding authors of peer-reviewed papers on invasive species risk analysis or risk assessment. These authors were identified through a targeted search, with a focus on improving geographic rep- resentation. Efforts focused on reaching out to researchers from underrepresent- ed regions, particularly in Asia and South America, based on preliminary survey responses following the Twitter call, which indicated a deficit from these areas. The survey remained open for responses from November 2020 to March 2021. Data analysis All analyses and data visualization were performed using the programming language R (v4.0.4.). To extract the survey results, we used the package ‘qualtRics’ (Ginn et al. 2024). Responses that were at least 80% complete were included, as this threshold ensured that respondents had answered the majority of questions, while still allowing NeoBiota 99: 341-362 (2025), DOI: 10.3897/neobiota.99.153010 343 Susan Canavan et al.: Risk assessment in practice Table 1. Overview of survey topics and key findings. This table summarizes the structure and major results of the survey. For the complete list of survey questions and response formats, see Suppl. material 1: table S1. Theme Question topic Results summary General information Tools used 46 tools identified (Table 2; Fig. 2) Geographic scale Assessments done mostly at the country-level (40%) and regional scale (39%) Geographical coverage 89 countries assessed (Fig. 1) Taxonomic groups assessed Ecosystems assessed Experience with RAs Length of experience Number of RAs conducted Predominantly plants (Fig. 3a) Predominantly terrestrial ecosystems (Fig. 3b) Most respondents > 5 years’ experience (60%) (past 3 years) Most respondents conducted < 10 RAs (44%) Time required per assessment Typically weeks to months Organizations represented Qualifications & Opinions Education level of assessors Government agencies (48%), research institutions (29%) Majority hold master’s (24%) or doctoral (64%) degree Necessary education for effective assessment | Bachelor’s degree or lower (52%); Master’s degree or higher (39%) Availability of training programs Strong support for formal training and certification programs (73%) Data and Information Data sources used Data quality & difficulty Implementation of RAs Peer review included 107 databases listed, mostly open-access with CABI, GBIF, and ITIS being the most cited (Fig. 4; Suppl. material 1: table S3) Challenges include outdated or incomplete data 88% include peer review (internal/external) Uncertainty/confidence reported Included in 93% of assessments Public comment period Accessibility of RA results Results used in policy Rarely included (41% never, 20% always) Usually publicly accessible (40% always, 37.2% occasionally) Frequently inform policy recommendations for minor omissions that commonly occur in surveys. This yielded 107 usable surveys from the initial 262. We used the longitude and latitude coordinates of the IP addresses to estimate where survey respondents were based and plotted the data using the ‘maps’ package (Fig. 1). We then mapped the number of RAs conducted per country based on the results of the survey questions (Fig. 1), noting that some respondents may have used Virtual Private Networks (VPNs), reducing the accuracy of these location estimates. We used Likert scale analysis to summarize and visualize responses to specific survey questions regarding the taxonomic groups assessed, the factors that typically initiate a new RA, and the ecosystems evaluated. Additionally, we summarized the most fre- quently used databases (those mentioned by more than five respondents) that inform RAs for general, taxonomic, and occurrence information. To illustrate the frequency of database use, we created a heatmap, providing a clear visual representation of the data. We manually grouped RA tools using two objective criteria: AWRA affiliation and geographic region. First, each RA tool was categorized based on its relationship with the AWRA framework, one of the original and most widely adapted frameworks, into two groups: (1) AWRA or derivative tools, including those that directly imple- ment the original AWRA or are derived from it, either directly or indirectly. Indirect derivation includes tools like the NZAqWRA, and its direct derivatives like the US- AqWRA, a tool originally developed to address limitations of the AWRA for aquat- ic plants. (2) Non-AWRA tools, developed independently of the AWRA framework. NeoBiota 99: 341-362 (2025), DOI: 10.3897/neobiota.99.153010 344 Susan Canavan et al.: Risk assessment in practice 33 27 7g, =! 14 31313131313 1313 421212121212121212121212121212121212 | 10 | 7 ra | 6666 | @ ‘ = 9 Dt De De ya a 3333 CESSES TCOOSTGBCVBEMTDHOHFSTOCRVBDAZASCTCTCCOGD OTST H TOBE GOGO VTGIGTClCVWV#GBAlOewvseeweseo=y eco Tea rs) PRBS SCORSESE OSLCELTSSSGCCECtStv ec gQesreseacoeaeecpaecveaecsHeoveaaynetesoc BSOAGAOSENOVO CHS Seo ORS EC ESS SESE SS SHES OSS E SEER ESS SLE SS ESR SPERSFOESLSSSRSS EES Esse BE Be 2 2B SBOE eS EO BESS RBS EOS GP oS ON BSS Core POOF SS SSVI EDA Ga cor oa 5a om? A Pe @ aenn OZBIOA8 L5EOHo s2s Foot 73s se BE N BS S = os 5 x 2 WJ z