An Overview of Pre-Border Weed Risk Assessment and Post-Border Weed Risk Management Protocols

Printer friendly version


Bruce Auld

New South Wales Department of Primary Industries, Orange Agricultural Institute, Forest Road, Orange, NSW 2800, and
Charles Sturt University, Leeds Parade, Orange, NSW 2800, Australia.

Cite this article as:

Auld, B. (2012). An overview of pre-border weed risk assessment and post-border weed risk management protocols. Plant Protection Quarterly 27(3), 105-11.

General Introduction

The need for weed risk assessment has grown worldwide with increasing international trade and travel as well as new uses for plants such as biofuels. Australia has been at the forefront in developing a screening system for plants that has been adopted in other countries. However, refinements and improvements in the system have been suggested.

For species that have already invaded, a range of management options are available, depending on the risk posed and the feasibility of control. Jurisdictions with noxious plant laws inherently have some form of risk assessment, but this, and potential management strategies, are often not explicit nor well documented. In Australia, a National Post-Border Weed Risk Management Protocol (Standards Australia 2006), is currently being revised. It provides guidelines to prioritise species in which relative risk and feasibility of control are contrasted to arrive at suggested management actions. Future developments of the protocol are likely to take into account widespread weeds, contentious species and uncertainty in its various forms.

Pre-Border Weed Risk Assessment 


Although improved weed risk assessment protocols are in various stages of development in several countries, few are officially in operation or easily accessible. Many systems remain cumbersome.

In the United States of America (USA), Plant Protection and Quarantine (PPQ) uses a weed risk assessment procedure responding to questions in an open-ended narrative process that takes from two to eight weeks per species to complete (Koop et al. 2011). However, a recent United States Department of Agriculture (USDA) regulation (effective from June 2011) allows USDA’s Animal and Plant Health Inspection Service (APHIS) to prohibit imports of risky plants without needing to do an extensive weed risk assessment, placing them in a category referred to as ‘not authorised pending pest risk analysis’ (NAPPRA) (Jenkins 2011). APHIS plans to use the assessment system proposed by Parker et al. (2007), see USGPO (2011). This proposed ‘weed-ranking model’ for the USA is based on four main elements: invasive potential; geographic potential; damage potential; and entry potential. A ranking score is produced from the product of scores for each element (Parker et al. 2007). The inclusion of ‘entry potential’ in a quarantine system seems illogical, especially as a low or zero score would impart a low or zero score to the final ranking score. Presumably, a ranking score could be created from the other three elements alone.

In New Zealand, applications for plant imports come under the Hazardous Substances and New Organisms Act 1996. Applications made under this process are time consuming and expensive, and serve as a disincentive for would-be importers (Williams et al. 2010). In Canada, the Canadian Food Inspection Agency (CFIA) performs risk assessments under the International Plant Protection Convention (IPPC) rubric, (like many other countries) and is mostly concerned with the regulation of imports under international trade treaties (C. Wilson, personal communication). In South Africa, a Risk Assessment Framework was developed in 2004 to support the National Environmental Management: Biodiversity Act (NEMBA) but has never been promulgated; however, other risk assessments for plant imports are in place (L. Henderson, personal communication). Recently, in Great Britain, a Non-Native Species Secretariat (NNSS) has used a risk assessment template following the European and Mediterranean Plant Protection Organisation’s (EPPO) scheme for agricultural quarantine pests (Mumford et al. 2010). This protocol is discussed later in the paper.

Australia has played a leading role in the development of practical weed risk assessment and management protocols. The Australian Weed Risk Assessment system (AWRA) (Pheloung et al. 1999) has been widely evaluated and adapted for use in other countries and jurisdictions. In several retrospective analyses, the correct classification rate for weeds was generally better than for non-weeds, that is, the system commits more false positives than false negatives. In some cases, the correct classification of non-weeds has been improved by using a decision-tree model for species initially classified as requiring ‘further evaluation’ (Daehler et al. 2004).

Recent applications of the AWRA system

Chong et al. (2011) showed that scores achieved in the AWRA system at one locality could be used for other localities of similar climate when tested in four tropical and sub-tropical regions. They used the second screening suggested by Daehler et al. (2004) for some species requiring further evaluation. However, in Japan, Nishida et al. (2009) found no advantage in using a secondary test for species requiring ‘further evaluation’ when applying the AWRA system.

A modified version of AWRA has been used for east African rainforests, showing that the system was able to discriminate between invaders and non-invaders with accuracy comparable to similar assessments in sub-tropical and temperate regions (Dawson et al. 2009). Further modifications were suggested specifically for tropical forests.

McClay et al. (2010) evaluated the AWRA system for prediction of plant invasiveness in Canada, and found that the system correctly rejected all major and most minor weeds (86%), but also incorrectly rejected 44% of non-weedy species. Moreover, 23 of the 49 questions in the system were not significantly associated with local experts’ weediness ratings, suggesting that the system could be simplified. In Canada, expert ratings were strongly related to cold-hardiness, suggesting that performance of the system could be improved by making better use of climatic tolerances.

The AWRA system has also been tested in Spain on species already present (Gasso et al. 2010). The system rejected 94% of invasive species, but 50% of non-invasive (‘casual’) species were rejected and 29% required further evaluation. The low accuracy for non-invasive species was attributed to possible incorrect a priori expert classification, high weight of AWRA scores given to potential impacts and the short residence time for some potential invaders.

Another modified version of the AWRA system has been used in South America to evaluate mutual risk of plant invasions between Chile and Argentina (Fuentes et al. 2010), although no public risk assessments for screening new plant species exists for either country.

In Italy, the AWRA system has been tested on known invasive species and on prospective biofuel species (Crosti et al. 2010). The system correctly identified 93% of invasive species and 75% of non-invasive species. Further evaluation of 20% of the tested species was carried out using a secondary screening. The authors suggested that the AWRA system could be used as the basis for a Weed Risk Management Protocol for Italy.

Gordon and Gantz (2011) found that while the AWRA system correctly predicted all major invaders among 149 aquatic plant species in the USA, it identified only 1% of non-invaders. The system predicted that 83% of non-invaders would be invasive. A change in threshold (cut-off) values resulted in accurate identification of 89% of non-invaders but the correct primary identification of invaders was reduced to 76%. This risk-averse bias of the system may reflect the vulnerability of the vacant niches that open water surfaces represent. The estimated year of introduction to the USA did not have a significant influence on the accuracy for non-invaders but the authors concede that some of the apparently non-invasives “may have a high probability of becoming invasive over time”. Whether or not a species was invasive elsewhere was a more accurate predictor of invasiveness (87%) than was the complete AWRA system (34–58%). Gordon and Gantz (op. cit.) recommend either further testing to determine optimal threshold levels or a separate screening system for freshwater aquatic plants. In more recent paper, Gordon et al. (2012) have demonstrated that a modified version of the New Zealand aquatic plants weeds risk assessment system provides high accuracy in distinguishing non-invaders from harmful aquatic invaders in the USA.

Other approaches

Caley and Kuhnert (2006) explored classification and regression tree models as an alternative to the AWRA system. The optimal classification tree model included just four of the 44 attributes of the introduced plant species examined: intentional human dispersal; naturalisation beyond native range; weed elsewhere; and high level of domestication. They also point out the importance of propagule pressure in successful invasions. The performance of this system was slightly inferior to the current AWRA system, although this system required answers to only a few questions.

Koop et al. (2011) describe a new screening tool based on the AWRA system, using two elements of risk, establishment/spread potential and impact potential in a logistic regression model. It also includes a secondary screening tool to further assess those species categorised as ‘evaluate further’. Their model accurately identified 94% of major invaders and 97% of non-invaders. The analysis did not consider populations of minor invaders and excluded species with answers of ‘unknown’. Because of the correlation between establishment/spread and impact, the impact risk element was not a significant predictor of plant invasive status. However, it is difficult to understand why impact per se would predict invasive status. The latter could have been omitted from the analysis losing only 1% of the variability explained by the full model. Notwithstanding this, the authors chose to retain this element partly to create a ‘risk profile’ for each species rather than a single score. Once again, invasive status elsewhere had the greatest association between question response and a priori invasive status. The model excludes geographic and climatic suitability, although the authors state that this will be an important part of any weed risk management system developed for the USA.

A protocol to specifically assess weed risk to environmental assets in forage species has been proposed by Stone et al. (2008). This assessment model includes three modules: invasiveness, (negative) impact and potential distribution. Summed scores for each module (although potential distribution is a single score) are multiplied together, to produce a weed risk assessment score. While the protocol does not consider the potential benefits of species under examination, it could also be used to assess the risk of domesticated species in other land-uses such as biofuel production.

In the non-native species risk assessment scheme used in Great Britain (Mumford et al. 2010), risk assessments are commissioned from independent expert assessors and overseen by the Non-Native Species Risk Assessment Panel. Estimates of entry, establishment, spread and impact are made with measures of assessor confidence to give an overall semi-quantitative summary of risk. However, no attempt to sum the scores of the individual questions is made to give an overall score. This is partly because of the diverse nature of organisms embraced by the scheme. Instead, a risk profile is produced showing the cumulative probability of a magnitude of impact being reached five years into the future. This is generated by sampling component scores according to the confidence distribution for each component in a simulation of 5000 realisations of score values. Some estimate of confidence of scores based on uncertain information is desirable, perhaps without the need for this level of sophistication. The performance of this protocol is yet to be widely tested.

Hulme (2012) points out several shortcomings with current weed risk assessment systems, particularly the well recognised ‘low base rate’ problem. While acknowledging that this situation exists, like the ‘tens rule’ (Williamson 1993), it does not inform a decision made about an individual species. Hulmes’ claims of the high monetary costs of weed risk assessment are unsubstantiated and proposed alternatives such as ‘scenario planning’ are obscure and require further explanation. He suggests focussing investment towards early detection, mitigation and management as well as targeting sleeper weeds. These approaches ignore the pre-border barrier. Moreover, just which species should be targeted would surely require some form of weed risk assessment.

Recent evaluations of the AWRA system

A study of the AWRA system’s operation since 1998 showed that propagule pressure in the form of unintentional human dispersal is the strongest predictor of species rejection for importation (Weber et al. 2009). Only two non-human mediated variables were useful for prediction of the system’s outcome: “tolerates or benefits from disturbance” and vegetative reproduction. Reproduction in one year or less and broad climatic suitability were also important when human mediated variables were removed. However, the authors concluded that as many questions as possible should be answered in this or similar systems. A minimum number of positive responses are more important than responses to certain questions for increased confidence in decision making. Nevertheless specific questions may be especially relevant for particular species.

Onderdonk et al. (2010) considered evaluations of the AWRA system. They pointed out that methodological variation may influence results. Various tests of the system have used different a priori categories, there has been variability in the base rates of test species and the evidence required to answer questions in the system varies. Negative evidence is often not reported even when known, leading to more positive than negative answers. Typically, more information is available for invasive species (or species with a long history of cultivation) leading to higher scores for them. Considering this Onderdonk et al. concluded that the AWRA was a robust screening tool across geographies and should be used as uniformly and transparently as possible with methods and results reported clearly. In this spirit, Gordon et al. (2010) have published guidelines for addressing the 49 questions of the AWRA system together with lists of data sources and examples.

Future developments of the AWRA

The application of the AWRA system to a range of environments has proven to be relatively successful. While it is conservative and risk averse, its use may be preferable to allowing any invasive plants to be imported (Roberts et al. 2011). Many ‘non invasive’ species rejected by the system in other jurisdictions may not have had sufficient residence time to adequately reveal their invasiveness. Any of the various weed risk assessment systems are time specific (depending on current knowledge and land-use), unless the only criteria used for evaluation are inherent biological characters of a species. Moreover, while the system itself is somewhat subjective, retrospective analyses of the system are also subjective, employing expert opinion.

In the AWRA system, scores for answers to 49 questions are generally summed, although scores for some questions weight other scores within the system. Daehler and Virtue (2010) have suggested a modification in which the 49 questions are divided into two groups: likelihood (L) and consequences (C) in keeping with standard risk assessment procedure (Bowden et al. 2001). (This is similar to the division used by Koop et al. (2011) above.) In the system advocated by Daehler and Virtue (2010), a risk assessment score (R) is obtained by multiplying likelihood by consequence (L x C = R). Parker et al. (2007) also highlight the value of the multiplicative approach. Daehler and Virtue suggest that separation of L and C may give a clearer differentiation between major and minor weeds (‘risk profile’ (Koop et al. 2011)) and improve AWRA accuracy. The modified system was tested on previously assessed species from Hawaii. The R score gave a better prediction of weeds than the original AWRA system and identified non-weeds at the same rate as the original system. A problem with this approach, with the existing set of questions, is that there is some non-independence of L and C as noted above with Koop et al.’s (2011) system.

Although some questions in the AWRA have apparently more importance in allocating species to a category, this is partly due to the meta-analyses used; specific questions may be important for individual species. Weber et al.’s (2009) suggestion that as many questions as possible should be answered, is appropriate. If the modification of the system suggested by Daehler and Virtue (2010) is pursued, there may need to be some changes to the questions to increase independence between L and C. For instance, they use the ‘weed elsewhere?’ question to indicate consequences, whereas it could equally well indicate an ability to spread, that is, a likelihood.

The AWRA could also include some indication of potential variability around certain parameters as well as indications of uncertainty of knowledge regarding other parameters and responses weighted accordingly. For instance, the ‘weed elsewhere’ question is consistently the most accurate single predictor of invasiveness in a new environment. Apart from the obvious indication of invasiveness, its utility may be due, in part, to the fact that this piece of information, particularly if answered positively, is likely to be less variable and more certain than the answers to many other questions.

While refinement of the system is admirable, for a given species a simple yes/no/pending-further-evaluation decision regarding importation is ultimately made. The basis for this decision should be clear and transparent, and the current system largely achieves this.

Post-Border Weed Risk Management 


In contrast to pre-border weed risk assessment systems, there has been less development of new weed risk management (post-border) systems internationally. However, in Australia, a National Post-Border Weed Risk Management Protocol (Standards Australia 2006) has provided the basis for weed risk management systems in New South Wales, the Northern Territory and South Australia (see below). In addition, there are significant gaps on the translation of Weed Risk Management systems into policy and management responses. Notable exceptions include the application of prioritisation to protect biodiversity from weed species (Clarkson et al. 2010, Downey 2010, Downey et al. 2010c, NSW DPI and OEH 2011).

Risk assessment

Many post-border systems assess weed risks but do not consider management of these risks. For instance, Randall et al. (2008) developed an Invasive Species Assessment Protocol for non-native plants that negatively impact biodiversity in the USA. The system requires quite detailed information on ecological impacts, distribution, trends in distribution and management. The evaluation is limited to impacts on biodiversity and the protocol does not include climate matching as it is not intended for prediction.

An Alien Plants Ranking System has been developed for grassland and prairie parks in the central USA which relies on a set of 23 questions. The output is a graphical depiction of data as a dot on a two dimensional matrix comparing difficulty of control against impact, together with an indication of ‘pest score’ for a species, given as the size of the dot (USGS 2006).

In California, the Californian Invasive Plant Council has developed risk assessments for plants with predictions of potential spread of species including those present and potential threats (Cal-IPC 2012).

In Alaska, State and Federal agencies have developed a prioritisation scheme to evaluate the likelihood of a non-native species establishing and the consequences to the ecology and community in the three eco-graphic regions of Alaska. Following climate matching, invasiveness ranking is based on: ecological impacts; biological characteristics and dispersal ability; distribution; and feasibility of control (USDA FS 2012).

In Australia, risk assessment systems for some specific situations have recently been developed. These include evaluating the weed risk of plants in botanic gardens (Virtue et al. 2008), weed risk assessment for Australian nursery and garden industries (A. Kachenko personal communication) and the regional weed species-led prioritisations of weeds affecting environmental values in Australia in New South Wales (Downey et al. 2010b).

Management of weed risk

In the USA, a national early detection and rapid response system for invasive plants is being developed by the United States Geological Survey in cooperation with a number of interagency groups (USGS 2012).

Any jurisdiction that has noxious plant laws has, inherently, some form of risk assessment, and often, some form of risk management, although this may not be well documented. Individual states in the USA vary in their approach to weed risk management. In Washington for instance, each year, the State Noxious Weed Control Board adopts the State Noxious Weed List. This list determines which plants will be considered noxious weeds and where control will be required. The weeds are placed in four classes which indicate management required: eradication, containment, education of stakeholders or monitoring (WS NWCB 2010).

In Florida, the Institute of Food and Agricultural Sciences (IFAS), University of Florida, has developed an assessment tool for non-native plants in Florida’s natural areas. Information on each species is organised under four topics: ecological impacts; potential spread; management difficulty and economic value. Scores for these topics are used to place species in index categories from Very High to Low for each topic. The combinations of these categories, (e.g. High, Low, High, Low) are then used to indicate actions, within a ten year time horizon. These include: not considered a problem; caution-manage to prevent escape; and invasive and not recommended by IFAS faculty (Fox, A.M. et al. 2009). For plants not present in Florida, a modified version of the Australian Weed Risk Assessment system is used (Gordon et al. 2008).

In Canada, most provinces have weed control legislation including lists of regulated species that are subject to management (CFIA 2008). Although to date there have been no significant risk assessments at local level in the provinces, a number of pre- and post-border approaches are under development, for example in British Columbia and Alberta (C. Wilson, personal communication).

Although a system was proposed for the management of invasive alien plants in South Africa (Robertson et al. 2003), work is still in progress to develop an official system (M. Robertson, personal communication; J. Wilson, personal communication). The proposed system uses expert opinion in a Delphi method and takes account of uncertainty using a confidence score indicating the uncertainty and availability of data for each criterion.

In Japan, work is in progress to develop national post border weed management (T. Nishida, personal communication). Currently in Japan, Naturplant, an open forum established by Dr Hirohiko Morita, aims to maintain current awareness about naturalised plants in Japan (Naturplant 2012).

In the United Kingdom, a invasive non-native species risk management scheme includes a module: selection of risk management options for invasive non-native species, see DEFRA (2012). This module assumes that a risk assessment has already been undertaken, the species is invasive or potentially invasive and the risk is unacceptable. The module is then used to: determine the current invasion situation; assemble the key information that affects the choice of management options; and list management options comparing them for efficiency, cost, safety and acceptability. Management options are eradication, containment, suppression and restoration of invaded habitats. The management methods for plants are: physical, biological and chemical.

In Australia, the Victorian Pest Plant Prioritisation Process determines a plant pest score which is the sum of weighted scores for invasiveness plus present distribution/potential distribution plus impact (Weiss and McLaren 2002). Confidence ratings are assigned to information sources and weightings are applied to each specific question in the system. In a confusing complication, the ratio of present distribution to potential distribution is assigned a value between zero and one in an inverse relationship to the original ratio. These values are then assigned ratings from ‘very high’ to ‘extremely low’ (Weiss et al. 2004). Plans are in place to revise and clarify the system (J. Steel, personal communication). With this scheme and others that ultimately produce a single numerical score for a species, much of the information that would be useful in choosing management strategies and tactics is left behind.

In Queensland a “rapid evidence-based” system is used for assessing potentially invasive species that focuses on three key traits: history as a pest elsewhere; climatic suitability; and the extent of its native range. If a species has been tentatively assigned as ‘high risk’ a more detailed assessment is conducted. For species that are already widespread, a system based on the earlier unpublished Walton species assessment system is used (S. Csurhes personal communication 2011). It includes evaluation of economic, environmental and social benefits as well as negative impact assessment.

The Australian Post-Border Weed Risk Management Protocol

There have been significant new applications of the National Post-Border Weed Risk Management Protocol, in Australia by the Northern Territory (Setterfield et al. 2010; Rachor-Rossiter et al. 2012) and New South Wales (Johnson 2009a, b; Johnson and Charlton 2010), and internationally in a number of Latin and South American, South-East Asian and North African countries (FAO UN 2006).

The foundation for recent developments has been the system developed by Virtue for South Australia (e.g. Virtue 2008, 2010). In this system, comparative weed risk is the product of scores for three factors: invasiveness; impacts; and potential distribution. The magnitude of these factors is established from scores for answers to a series of questions. Similarly, feasibility of containment is assessed as the product of scores for control costs per unit area, current distribution and persistence. Total scores for weed risk and feasibility of containment are then used to place a species within a 5 x 5 weed management action matrix. This matrix has eight different management actions as well as an external ‘alert’ action for species not present that represent a significant threat.

In the system developed for the Northern Territory the 3 x 4 weed risk management matrix has 11 different management actions (Rachor-Rossiter 2012). In the system developed for New South Wales (Johnson 2009a, b) the 5 x 5 matrix follows the South Australian system with minor variations in management actions in some cells.

Thus, while these systems are often said to prioritise weeds for management, they actually categorise weeds into management action classes such as “eradicate”, “contain”, or “protect sites”. Within management action classes, species could be prioritised by score value, although practical factors such as access to sites may ultimately override minor score differences in deciding where to invest in weed management. With time and changes in circumstances or knowledge, species are reassessed and may be moved from one matrix cell (management action) to another.

A decision to attempt to eradicate weed species will ultimately depend on the amount of investment that can be made. Fox, J.C. et al. (2009) found that regardless of a carefully designed surveillance strategy, eradication of Nassella neesiana (Trin. & Rupr.) Barkworth in south eastern Queensland is implausible at current levels of surveillance and control and these efforts should be doubled to be successful. Panetta (2009) argues for a flexible approach to management allowing for a change of strategy when an initial plan, e.g. eradication, becomes untenable. Moreover, where there is a high level of uncertainty more than one management option may be indicated (see below). In considering the feasibility of containment strategies, Panetta and Cacho (2012) state that this should be viewed in terms of the effort required to reduce spread and the effectiveness of management actions. They conclude that management of dispersal pathways and timely detection of new infestation foci appear to be most critical. Thus scouting or surveillance must form part of a containment strategy. Surveillance may be ‘structured’ and/or ‘unstructured’, the latter including public awareness campaigns.

Januchowski-Hartley et al. (2011) have recently produced a spatially explicit decision method to identify management actions for invasive species aiming to minimise costs and the likelihood of reinvasions. At a regional level, Skurka Darin et al. (2011) have developed a tool for prioritising plant populations for eradication across a range of target weeds, although it would equally serve a suppression/containment program. These systems require detailed on-ground distribution data and control costs which are often not readily available for newly invading species.

Future developments

The Australian National Post-Border Weed Risk Management Protocol is currently under review (Auld et al. 2012). Areas being up-dated include those discussed below:

Widespread weeds and context   Williams et al. (2009) noted that widespread weeds impacting on biodiversity have often been overlooked in regional weed priority assessments. Just how widely the weed risk management net is cast will be important. Hence there is a need to clearly define the context of each evaluation. Downey et al. (2010a) suggest that the current Australian Weed Risk Management system is not aimed at managing widespread weed species. While certainly species-led rather than site-led, the current Weed Risk Management system can include widespread weeds as well as recent incursions and potential threats.

Given a group of weeds spread over a management matrix, prioritisations still need to be made, presumably on the basis of costs and benefits. So, while a newly introduced species may be a threat, return on investment may be greater from controlling a more widespread species. In practice, the situation becomes more complex when different actions would be undertaken by different institutions. Where a management action is site-based, protocols such as Asset-Protection Triage (Downey et al. 2010c) will be invaluable in prioritising actions.

Uncertainty   Downey et al. (2010c) have highlighted the difference between impact and threat (the likelihood of an impact occurring). Assessments surrounding potential future impacts of a newly introduced species are clouded by uncertainty compared with measureable impacts of existing and widespread species.

The New South Wales Weed Risk Management system introduces some accounting for uncertainty within the system. In this instance, uncertainty is assessed as the proportion of ‘do not know’ answers to the total number of answers (Johnson 2009a, b). The uncertainties surrounding ultimate distribution and impact of a species are of a different kind (model uncertainty) and potentially even more significant. Both kinds of uncertainty should be part of the architecture in future WRM developments (Benke et al. 2011). These uncertainties may be expressed as confidence intervals in some cases and combining them into an overall confidence range may mean that, for some species at least, more than one management option may be indicated.

Most assessment systems use expert opinion, either individuals or groups. Liu et al. (2011) have suggested a jury-based Deliberative Multi-Criteria Evaluation (DMCE) approach to incorporate uncertainty and to allow stakeholder participation and opportunities for diverse views to be expressed. While this is an admirable goal, funding constraints may often mean that assessments are made by a limited number of experts. Although experts may sometimes be overconfident in their judgements, there are procedures to reduce this effect (Speirs-Bridge et al. 2010).

Contentious species   These species may provide economic benefit in some situations and have negative impacts in other situations. A protocol to specifically assess the weed risk of forage species to environmental values has been proposed by Stone et al. (2008). It can be used for pre- and post-border assessments. This assessment model includes three modules: invasiveness, (negative) impact and potential distribution. Summed scores for each module (although potential distribution is a single score) are multiplied together, to produce a weed risk assessment score. While the protocol does not consider the potential benefits of species under examination, it could also be used to assess the risk of domesticated species in other land-uses such as biofuel production. The Australian Weed Risk Assessment System can also be used for this purpose as has been suggested in Italy (see above). The International Union for Conservation of Nature (IUCN) has recommended the AWRA system to screen plants targeted for biofuel crops (IUCN 2009).

There has been other significant new research into the management and risk assessment of contentious plants, for example, Johnson (2007); Grice et al. (2009); Clarkson et al. (2010); Ferdinands et al. (2010); Friedel et al. (2010); Grice et al. (2010); Johnson (2010) and Johnson (2012). This research extends into the area of potential biofuel source evaluation, for example Buddenhagen et al. (2009) and Ferdinands et al. (2011).

A dilemma in dealing with the contentious plants issue is just how far a risk management system should enter into any economic evaluation of potential new crops. The economic assessment of a putative new industry seems beyond the scope of a risk assessment system.

Links with other systems   The forthcoming Australian National Environmental Biosecurity Response Agreement (NEBRA) requires, among other things, prediction of the potential range of species using climate matching. This and other NEBRA requirements may need to become part of information collated in WRM systems for newly invading species.


I thank Stephen Johnson, Jackie Steele, Claire Wilson, Lesley Henderson, Stephen Darbyshire and the referees for critical comments on drafts of this paper. The work was part of a project updating the National Post-Border Weed Risk Management Protocol within the National Weeds and Productivity Research Program, funded by the Department of Agriculture, Fisheries and Forestry, managed by the Rural Industries Research and Development Corporation.


Auld, B., Johnson, S., Ainsworth, N. Brunel, S., Csurhes, S., Downey, P., Ferdinands, K., Hannan-Jones, M., Hennecke, B., Keese, P., Kriticos, D., Low, D., Munday, C., Panetta, F.D., Robold, A., Setterfield, S., Steele, J., Stewart, K. and Virtue, J. (2012). Further development of the National Weed Risk Management protocol. Proceedings of the 18th Australasian Weeds Conference, ed. V. Eldershaw, pp. 317-9. (Weed Society of Victoria, Melbourne).

Benke, K.K., Steele, J.L. and Weiss, J.E. (2011). Risk assessment models for invasive species: uncertainty in rankings from multi-criteria analysis. Biological Invasions 13, 239-53.

Bowden, A.R., Lane, M.R. and Martin, J.H. (2001). ‘Triple bottom line risk management: enhancing profit, environmental performance, and community benefits.’ (J. Wiley and Sons, New York).

Buddenhagen, C.E., Chimera, C. and Clifford, P. (2009). Assessing biofuel crop invasiveness: a case study. PLoS ONE 4(4), e5261. doi:10.1371/journal.pone.0005261

Caley, P. and Kuhnert, P.M. (2006). Application and evaluation of classification trees for screening unwanted plants. Austral Ecology 31, 647-55.

Cal-IPC, California Invasive Plant Council (2012). (accessed 26 September 2012).

CFIA, Canadian Food Inspection Agency (2008). Invasive Alien Plants in Canada-Technical Report. 71 pp.

Chong, K.Y., Corlett, R.T., Yeo, D.C. and Tan, H.T.W. (2011). Towards a global database of weed risk assessments: a test of transferability for the tropics. Biological Invasions 13, 1571-7.

Clarkson, J.R., Grice, A.C., Friedel, M.H., Setterfield, S.A. and Ferdinands, K.B. (2010). The role of legislation and policy in dealing with contentious plants. Proceedings of the 17th Australasian Weeds Conference, ed. S.M. Zydenbos, pp. 474-7. (New Zealand Plant Protection Society, Christchurch).

Crosti, R., Cascone, C. and Cipollaro, S. (2010). Use of weed risk assessment for the Mediterranean region of Central Italy to prevent loss of functionality and biodiversity in agro-ecosystems. Biological Invasions 12, 1607-16.

Daehler, C.C., Denslow, J.S., Ansari, S. and Kuo, H. (2004). A risk assessment system for screening out invasive pest plants from Hawaii and other Pacific islands. Conservation Biology 18, 360-8.

Daehler, C.C. and Virtue, J.G. (2010). Likelihood and consequences: reframing the Australian weed risk assessment to reflect a standard of risk. Plant Protection Quarterly 25, 52-5.

Dawson, W., Burslem, D.F.R.P. and Hulme, P.E. (2009). The suitability of weed risk assessment as a conservation tool to identify invasive plant threats in East African rainforests. Conservation Biology 142, 1018-24.

DEFRA (2012). Great Britain non-native species secretariat. (accessed 24 October 2012).

Downey, P.O. (2010). Managing widespread alien plant species to ensure biodiversity conservation: a case study using an 11-step planning process. Invasive Plant Science and Management 3, 451-61.

Downey, P.O., Johnson, S.B., Virtue, J.G. and Williams, P.A. (2010a). Assessing risk across the spectrum of weed management. CAB Reviews: Perspectives in Agriculture, Veterinary Science, Nutrition and Natural Resource 5, No. 38, 15 pp.

Downey, P.O., Scanlon, T.J. and Hosking, J.R. (2010b). Prioritising weed species based on their threat and ability to impact of biodiversity: a case study from New South Wales. Plant Protection Quarterly 25, 111-26.

Downey, P.O., Williams, M.C., Whiffen, L.K., Auld, B.A., Hamilton, M.A., Burley, A.L. and Turner, P.J. (2010c). Managing alien plants for biodiversity outcomes – the need for triage. Invasive Plant Science and Management 3, 1-11.

FAO UN, Food and Agriculture Organisation of the United Nations (2006). ‘Procedures for Post-Border Weed Risk Management’. (Food and Agriculture Organization of the United Nations, Rome). 21 pp.

Ferdinands, K.B., Setterfield, S.A., Clarkson, J.R., Grice, A.C. and Friedel, M.H. (2010). Embedding economics in weed risk management to assess contentious plants. Proceedings of the 17th Australasian Weeds Conference, ed. S.M. Zydenbos, pp. 482-5. (New Zealand Plant Protection Society, Christchurch).

Ferdinands, K., Virtue, J., Johnson, S.B. and Setterfield, S.A. (2011). ‘Bio-insecurities’: managing demand for potentially invasive plants in the bioeconomy. Current Opinion in Environmental Sustainability 3, 43-9.

Fox, A.M., Gordon, D.R., Dusky, J.A., Tyson, L., Stocker, R.K., Langeland, K.A. and Cooper, A.L. (2009). IFAS Assessment of Non-Native Plants in Florida’s Natural Areas: Status Assessment. (accessed 20 November 2011).

Fox, J.C., Buckley, Y.M., Panetta, F.D., Bourgoin, J. and Pullar, D. (2009). Surveillance protocols for management of invasive plants: modelling Chilean needle grass (Nassella neesiana) in Australia. Diversity and Distributions 15, 577-89.

Friedel, M.H., Grice, A.C., Clarkson, J.R., Ferdinands, K.B. and Setterfield, S.A. (2010). How well are we currently dealing with contentious plants? Proceedings of the 17th Australasian Weeds Conference, ed. S.M. Zydenbos, pp. 470-3. (New Zealand Plant Protection Society, Christchurch).

Fuentes, N., Ugarte, E., Kuhn, I. and Klotz, S. (2010). Alien plants in Southern America. A framework for evaluation and management of mutual risk of invasion between Chile and Argentina. Biological Invasions 12, 3227-36.

Gasso, N., Basnou, C. and Vila, M. (2010). Predicting plant invaders in the Mediterranean through a weed risk assessment system. Biological Invasions 12, 463-76.

Gordon, D.R. and Gantz, C.A. (2011). Risk assessment for invasiveness differs for aquatic and terrestrial plant species. Biological Invasions 13, 1829-42.

Gordon, D.R., Gantz, C.A., Jerde, C.L., Chadderton, W.L., Keller, R.P. and Champion, P.D. (2012). Weed risk assessment for aquatic plants: modification of a New Zealand system for the United States. PloS One 7(7): e40031 doi:10.1371/journal pone 0040031.

Gordon, D.R., Mitterdorfer, B., Pheloung, P.C., Ansar, S., Buddenhagen, C., Chimera, C., Daehler, C.C., Dawson, W., Denslow, J.S., LaRosa, A., Nishida, T., Onderdonk, D.A., Panetta, F.D., Pysek, P., Randall, R.P., Richardson, D.M., Tshidada, N.J., Virtue, J.G. and Williams, P.A. (2010). Guidance for addressing the Australian Weed Risk Assessment questions. Plant Protection Quarterly 25, 56-74.

Gordon, D.R., Onderdonk, D.A., Fox, A.M., Stocker, R.K. and Gantz, C. (2008). Predicting invasive plants in Florida using the Australian weed risk assessment. Invasive Plant Science and Management 1, 178-95.

Grice, A.C., Clarkson, J.R., Friedel, M.H., Ferdinands, K.B. and Setterfield, S.A. (2010). Containment as a strategy for tackling contentious plants. Proceedings of the 17th Australasian Weeds Conference, ed. S.M. Zydenbos, pp. 486-9. (New Zealand Plant Protection Society, Christchurch).

Grice, A., Friedel, M., Setterfield, S., Ferdinands, K., Clarkson, J., Rolfe, J. and MacLeod, N. (2009). Best practice for making strategic decisions about invasive plants of commercial value. Report No. 11/055. Rural Industries Research and Development Corporation, Canberra.

Hulme, P.E. (2012). Weed risk assessment: a way forward or a waste of time? Journal of Applied Ecology 49, 10-9.

IUCN, International Union for Conservation of Nature (2009). ‘Guidelines on Biofuels and Invasive Species’. (IUCN, Gland, Switzerland) 20 pp.

Januchowski-Hartley, S.R., Visconti, P. and Pressey, R.L. (2011). A systematic approach for prioritizing multiple management actions for invasive species. Biological Invasions 13, 1241-53.

Jenkins, P.T. (2011). Guide for the National Association of Exotic Pest Plant Councils. USDA’s New “Not Authorized Pending Pest Risk Analysis” Noxious Weed Regulation Program. (accessed 28 September 2012).

Johnson, S.B. (2007). When good plants turn bad! Resolving conflicts of interest between plant species that are both beneficial and weedy. Proceedings of the 14th Biennial NSW weeds conference. New South Wales Department of Primary Industries, Orange.

Johnson, S.B. (2009a). ‘New South Wales Weed Risk Management System. Instruction Book’. (Industry and Investment New South Wales, Orange).

Johnson, S.B. (2009b). ‘New South Wales Weed Risk Management System. Background Information’. (Industry and Investment New South Wales, Orange).

Johnson, S.B. (2010). A way to weigh dread weeds – a policy framework to estimate the costs and benefits of commercially valuable invasive species. Proceedings of the Global Biosecurity Conference 2010. pp. 122-3.

Johnson, S.B. (2012). Economic tools ≠ policy actions. Why benefit cost analysis are not a policy panacea for weedy but commercially valuable plant species. Proceedings of the 18th Australasian Weeds Conference, ed. V. Eldershaw, pp. 195-8. (Weed Society of Victoria, Melbourne).

Johnson, S.B. and Charlton, S.A. (2010). Evolution, not revolution. Getting stakeholders on board with the roll out of the National Post-Border Weed Risk Management protocols in New South Wales. Proceedings of the 17th Australasian Weeds Conference, ed. S.M. Zydenbos, pp. 493-6. (New Zealand Plant Protection Society, Christchurch).

Koop, A.L., Fowler, L., Newton, L.P. and Caton, B.P. (2011). Development and validation of a weed screening tool for the United States. Biological Invasions 13, 273-94.

Liu, S., Sheppard, A., Kriticos, D. and Cook, D. (2011). Incorporating uncertainty and social values in managing invasive alien species: a deliberative multi-criteria evaluation approach. Biological Invasions 13, 2323-37.

McClay, A., Sissons, A., Wilson, C. and Davis, S. (2010). Evaluation of the Australian weed risk assessment system for the prediction of plant invasiveness in Canada. Biological Invasions 12, 4085-98.

Mumford, J.D., Booy, O., Baker, R.H.A., Rees, M., Copp, G.H., Black, K., Holt, J., Leach, A.W. and Hartley, M. (2010). Invasive non-native species risk assessment in Great Britain. Aspects of Applied Biology 104, 49-54.

Naturplant  (2012).

Nishida, T., Yamashita, N., Asai, M., Kurokawa, S., Enomoto, T., Pheloung, P.C. and Groves, R.H. (2009). Developing a pre-entry weed risk assessment for use in Japan. Biological Invasions 11, 1319-33.

NSW DPI and OEH, New South Wales Department of Primary Industries and Office of Environment and Heritage (2011). ‘Biodiversity priorities for widespread weeds’. Report prepared for the 13 Catchment Management Authorities (CMAs) by NSW Department of Primary Industries and Office of Environment and Heritage, Orange. (Authors: Whiffen, L.K., Williams, M.C., Izquierdo, N., Downey, P.O., Turner, P.J., Auld, B.A. and Johnson S.B.). State framework and 13 parts.

Onderdonk, D.A., Gordon, D.R., Fox, A.M. and Stocker, R.K. (2010). Lessons learned from testing the Australian weed risk assessment system: the devil is in the details. Plant Protection Quarterly 25, 79-85.

Panetta, F.D. (2009). Weed eradication–an economic perspective. Invasive Plant Science and Management 2, 360-8.

Panetta, F.D. and Cacho, O.J. (2012). Beyond fecundity control: which weeds are most containable? Journal of Applied Ecology 49, 311-21. 

Parker, C., Caton, B.P. and Fowler, L. (2007). Ranking nonindigenous weed species by their potential to invade the United States. Weed Science 55, 386-97.

Pheloung, P.C., Williams, P.A. and Halloy, S.R. (1999). A weed risk assessment model for use as a biosecurity tool evaluating plant introductions. Journal of Environmental Management 57, 239-51.

Rachor-Rossiter, N., Setterfield, S., Ferdinands, K. and Elliot, L. (2012). ‘Northern Territory Weed Risk Management system. User guide’. (Northern Territory Government, Darwin).

Randall, J.M., Morse, L.E., Benton, N., Hiebert, R., Lu, S. and Killeffer, T. (2008). The invasive species assessment protocol: a tool for creating regional and national lists of invasive nonnative plants that negatively impact biodiversity. Invasive Plant Science and Management 1, 36-49.

Roberts, W., Harrod, O., Mitterdorfer, B. and Pheloung, P. (2011). Regulating invasive plants and use of weed risk assessments. Current Opinion in Environmental Sustainability 3, 60-5.

Robertson, M.P., Villet, M.H., Fairbanks, D.H.K., Henderson, L., Higgins, S.I., Hoffman, J.H., LeMaitre, D.C., Palmer, A.R., Riggs, I., Shackleton, C.M. and Zimmerman, H.G. (2003). A proposed prioritization system for the management of invasive alien plants in South Africa. South African Journal of Science 99, 37-43.

Setterfield, S.A., Ferdinands, K., Rossiter-Rachor, N. and Douglas, M.M. (2010). The Northern Territory’s Weed Risk Management System: an application of the National Post-Border WRM Protocol. Plant Protection Quarterly 25, 71-4.

Speirs-Bridge, A., Fidler, F., McBride, M., Flander, L., Cumming, G. and Burgman, M. (2010). Reducing overconfidence in the interval judgements of experts. Risk Analysis 30, 512-23.

Standards Australia (2006). HB 294:2006. National Post-Border Weed Risk Management Protocol. Standards Australia, Sydney and Standards New Zealand, Wellington.

Stone, L.M., Byrne, M. and Virtue, J.G. (2008). An environmental weed risk assessment model for Australian forage improvement programs. Australian Journal of Experimental Agriculture 48, 568-74.

Skurka Darin, G.M., Schoenig, S., Barney, J.N., Panetta, F.D. and DiTomaso, J.M. (2011). WHIPPET: a novel tool for prioritizing invasive plant populations for regional eradication. Journal of Environmental Management 92, 131-9.

USDA FS, United States Department of Agriculture Forest Service (2012). Alaska Region. (accessed 24 October 2012).

USGPO, United States Government Printing Office (2011). Federal Register Online, Volume 76, Number 103, pages 31171-210. (accessed 12 September 2012).

USGS, United States Geological Survey (2006). Alien plants ranking system. Version 5.1. (accessed 26 September 2012).

USGS, United States Geological Survey (2012). Early detection and rapid assessment of new invaders. (accessed 28 September 2012).

Virtue, J.G. (2008). ‘SA Weed Risk Management Guide: February 2008’. (Department of Water Land and Biodiversity Conservation, Adelaide, South Australia).

Virtue, J.G. (2010). South Australia’s Weed Risk Management System. Plant Protection Quarterly 25, 90-4.

Virtue, J.G., Spencer, R.D., Weiss, J.E. and Reichard, S.E. (2008). Australia’s botanic gardens weed risk assessment procedure. Plant Protection Quarterly 23, 166-78.

Weber, J., Panetta, F.D., Virtue, J. and Pheloung, P. (2009). An analysis of assessment outcomes from eight years’ operation of the Australian border weed risk assessment system. Journal of Environmental Management 90, 798-807.

Weiss, J., Edgar, R., Hunt, T. and Morfe, T. (2004). Victoria’s noxious weed review: roll out not fall out. Proceedings of the 14th Australian Weeds Conference, eds B.M. Sindel and S.B. Johnson, pp. 707-10. (Weed Society of New South Wales, Sydney).

Weiss, J. and McLaren, D. (2002). Victoria’s pest plant prioritisation process. Proceedings of the 13th Australian Weeds Conference, eds H. Spafford Jacob, J. Dodd and J.H. Moore, pp. 509-12 (Plant Protection Society of Western Australia, Perth).

Williams, M.C., Auld, B.A., Whiffen, L.K. and Downey, P.O. (2009). Elephants in the room: widespread weeds and biodiversity. Plant Protection Quarterly 24, 120-2.

Williams, P.A., Popay, I. and Gatehouse, H.A.W. (2010). New Zealand biosecurity legislation and the naturalization of exotic weeds. Plant Protection Quarterly 25, 95-8.

Williamson, M. (1993). Invaders, weeds and the risk from genetically modified organisms. Experientia 49, 219-24.

WS NWCB (2010). Washington State, Noxious Weed Control Board. (2010). (accessed 28 September 2012).


First published online: September 15, 2012