WP 1. To assemble the datasets required to construct PRAs valid for the whole of the EU
WP1 is primarily concerned with identifying, describing, assembling and enhancing the availability of the datasets required in the PRA process by creating a detailed inventory structured according to each stage of the PRA scheme and appropriate to the pests, pathways and receptors concerned. Sufficient metadata will be recorded for pest risk analysts to determine the relevance and suitability of each dataset. Access to the datasets will be determined and provided so that, where possible, direct links can be made through the computerised PRA scheme that therefore acts as a data portal. A high priority will be placed on obtaining the data from every EU member state required to produce PRAs that are representative for the whole of the EU. Trade with eastern Asia is rapidly increasing and many new pests are expected to invade the EU via this route. To create a new dataset of potentially damaging eastern Asian pests, WP1 will monitor European sentinel trees growing in Russian arboreta and planted in China with the help of Russian and Chinese subcontractors. WP1 will also assemble PRA schemes from different countries worldwide and obtain the data needed to test the developments to PRA techniques and draft examples of best practice. PRATIQUE will build on the DAISIE (Delivering Alien Invasive Species Inventories for Europe) database of non-native species established in Europe (DAISIE, 2009) and explore its capacity to support PRAs.
WP2. To enhance techniques for assessing economic, environmental and social impacts
The assessment of impacts is generally considered to be the most difficult section of any PRA. This arises because of the limited datasets available, the absence of clearly defined indicators and deficiencies in the existing tools. To enhance the techniques used in the EU, WP2 will first identify best practice by studying PRA schemes from other parts of the world and risk analyses from other sectors, primarily the environment, animal health, fisheries and food safety.
Much can be learned by analysing the characteristics of species that have already entered, established, spread and caused significant impacts in the EU. Through joint consortium membership, this project will exploit the analysis of species traits that lead to invasiveness undertaken within the ALARM EU project. PRATIQUE will extend the risk assessment tools developed and tested in ALARM to determine the extent to which trait analysis can provide prior warning of harmful pests not only in uncultivated habitats but also in cultivation.
Enhancements to the techniques for both qualitative and quantitative impact assessments are needed. Consistent qualitative assessments at five levels of scale are required for the PRA scheme and these will be prepared by providing indicators for minimal, minor, moderate, major and massive economic, environmental and social impacts. For quantitative assessments, computerised modules will be constructed to lead the assessor through the procedures required to predict pest spread and follow the partial budgeting and partial equilibrium approaches addressed in ISPM11. Emphasis will be given to two particularly challenging topics: the assessment of environmental and long-term impacts. Environmental impact methods will be enhanced not only by developing the economic methodology but also through multi-criteria analysis. The extent to which crops, cropping systems, uncultivated and amenity receptor environments can be assessed in terms of their vulnerability, with sufficient information for maps to be created in WP3, will also be determined. Predicting future impacts depends on techniques for scaling up impacts, e.g. from field to farm to industry, and a generic integrated model will be created that combines spread with the impacts. The methodology developed will be tested with a range of examples, representative of the diversity of PRA types.
WP3. To enhance techniques for standardising and summarising pest risk assessments
Four key risk assessment problems, (i) difficulties in standardising PRA production due to the lack of consistency in scoring responses to PRA questions, (ii) capturing and communicating uncertainty, (iii) mapping endangered areas and (iv) summarising risk, will be addressed by separate tasks in this work package. Each will begin with a review of best practice worldwide, taking particular care to explore the techniques used in other sectors and being developed by risk scientists, and finish with the development of protocols for integration into the PRA scheme.
Consistency will be enhanced by constructing a protocol with decision rules, illustrated by examples, for scoring each of the five levels of risk in every question of the EPPO PRA scheme. Within WP3, examples and/or values for each of the five levels of risk in the 17 questions of the establishment potential section of the EPPO PRA scheme will be developed. Scoring rules for impacts and pathway analysis will be generated by WP2 and WP4 respectively and coordinated by WP3.
Uncertainty arises in PRA due to missing, incomplete, inconsistent or conflicting data. Based on best practice worldwide, techniques will be developed and guidance given to assist risk analysts on the most appropriate steps to be taken so that assessments can be made even in situations of high uncertainty and the amount of uncertainty in PRAs can be captured, quantified and clearly communicated.
Currently, the mapping of endangered areas is based primarily on historical climate using the computer program CLIMEX. In this project we will compare CLIMEX with other techniques, such as Maxent and BIOCLIM, exploit new high resolution gridded climatologies representative of current and future climates and prepare guidance for the production of maps that integrate climate with the other biotic (e.g. hosts) and abiotic factors (e.g. soils) that are critical for pest establishment. Integration with land use and crop maps will greatly enhance not only the capacity to produce maps showing where pests can establish but also where economic impacts are likely to be most significant. Maps of endangered areas are an important method for summarising and communicating risk.
A variety of techniques for creating objective summaries of PRA have been proposed ranging from simple averages of scores to the application of conditional probabilities with or without Monte Carlo simulation. PRATIQUE will evaluate these different methods and determine which, alone or in combination, not only provide the most accurate and clear presentation of risk summaries but also communicate the summaries most effectively to decision makers.
WP4. To refine methods for pathway analysis and systems approaches
This work package will review best practice worldwide for the two components of PRA that are concerned with entry pathways: (i) assessing the potential for pests to move along pathways (pathway analysis) to enter the PRA area and (ii) devising a set of combined management options (systems approaches) that prevent or reduce the impact of such movement. Primarily because (i) occurs at the beginning of the risk assessment stage and (ii) is performed at the end of the analysis of risk management options there has been little attempt at linking the two procedures even though they are mutually dependent. Links between the two processes will be developed and provided to risk analysts in the computerised scheme. Guidance on the scoring of the five levels of risk in the entry potential section of the EPPO PRA scheme will be produced.
Pathway analyses often generate very long lists of organisms and these then need to be screened to identify those that may pose the highest risk and thus have the highest priority for PRA production. Pest risk analysts may also study one pathway and determine what species may be able to use this pathway to enter an area. The information from such pathway analyses is clearly highly relevant in determining the optimal selection of measures through systems approaches to prevent the entry of quarantine pests. The systems approach is a mitigation measure (in fact, the product of a series of measures acting in combination) to reduce the risks in a particular pathway. There is a limitation on the application of the systems approach in many instances because of the complexity of evaluating such a set of methods applied in varying combinations. By establishing improved computational tools (covering both qualitative and quantitative assessments of individual risks and mitigation components) it will be possible to include systems approaches more widely in pathway analyses.
This project will utilise existing global pest databases and compendia to generate species lists and, in a novel approach, determine whether neural networks combined with self-organising maps can be used to prioritise species for PRA. The concept involves the creation of a computerised linking system between key words or other items within distinct datasets or documents. So, for example, several sets of national PRA documentation could be searched by a neural network software engine to identify common elements (such as host species, climate references, geographical presence, pathway types, diction and control measures, etc) which form patterns of association that can then be used in a new risk analysis. Without a neural network to link otherwise unrelated documents and data, a new risk analysis would not benefit from a systematic analysis of other available information. The patterns of association may reveal risk factors, or mitigation opportunities, that would otherwise be missed. Imagine a matrix linking pest species to the countries and hosts where they are found. Each can be viewed as a series of nodes and links. Neural networks explore the underlying order in this matrix and make predictions as to which links are likely to exist but are as yet unrecorded. This may help the prediction of future pest risks.
Neural network techniques can be employed to provide links between information bits in documentation on a wide range of risks and mitigation measures. By creating explicit links between what is now scattered information (for example, on hosts, geographical range, trade volumes, mitigation performance, costs, etc) the process of pathway analysis will be made more effective and efficient. The neural network approach will benefit the pathway analysis approach both in conceptual terms, by demonstrating the ways in which information linkages should and could be considered, and by establishing techniques that can be directly applied to commonly held information in archives of PRAs and other databases. Because of the different formats of such information, a "soft" neural network link is preferred to a "hard" linkage that would require precise specification of the format and content of information bits in diverse national and international systems.
A framework for the generation of systems approaches within the EU PRA system will be developed with guidance on the identification and selection of efficient, feasible and reproducible measures that are operating together to reduce the risk of entry to an acceptable level. A significant issue to be addressed in the systems approach is how a combination of partially effective risk mitigation actions can, together, provide effective and efficient risk reduction. The wide range of measures that can be employed in a systems approach greatly increases the number of options that could be specified in risk management, and may include complex interactions and dependencies amongst the options. A modelling approach that can optimise, according to a range of policy objectives, will be used to develop tools that demonstrate the effectiveness of a systems approach to decision makers. Stochastic modelling with sensitivity analyses will be undertaken to explore the range of risk outcomes, to determine the degree of equivalence in management options and to provide appropriate weightings for system components.
WP5. Developing a decision support system for the eradication and containment of pest outbreaks
The PRA process is often considered to be complete once the pest risk has been assessed and, if the risk is unacceptable, reliable, cost-effective management options have been identified for preventing pest entry. However, internal measures, taken after the pest has entered and before the pest has established, may still be highly effective. In the EU, some EC directives describe measures to be taken following outbreaks of a few well known quarantine pests, e.g. Diabrotica virgifera virgifera, but there is no generic decision support scheme to help guide eradication or containment actions for all quarantine pests, whatever the pest, the habitat or the state of the outbreak when first discovered. To develop the first EU decision support scheme to guide actions at outbreaks, PRATIQUE will identify best practice by reviewing plant health contingency plans and conduct a meta-analysis of the successes and failures of eradication and containment campaigns worldwide. Measures that have been successful in combating pests that are particularly difficult to control due to problems of detection, pesticide resistance, high mobility and impacts caused at low population densities will be highlighted.
Particular emphasis will be given to the provision of guidance for assessing the costs and benefits of measures that can be taken to eradicate or contain pests at three key stages: when developing contingency plans, when a new outbreak has been discovered and when deciding whether to continue an official control campaign.
Pest surveying techniques play a key role in identifying new pest arrivals, first outbreaks and, once an outbreak has been detected, in defining its limits. This project will review pest surveying techniques worldwide and prepare a guidance document describing best practice for detecting new incursions and surveying pest outbreaks. New technologies with considerable potential in this field will be investigated. These include (i) multi-lure traps especially designed to allow simultaneous trapping of a large number of target species and (ii) 'smart-traps' that relay information automatically from pest surveillance systems, capture survey data remotely, include a spatial reference from global positioning systems and communicate between field and laboratory.
WP6. Ensuring that the PRA scheme is fit for purpose and user-friendly
The EPPO PRA scheme, which is used throughout Europe, directly follows ISPM11 and provides risk analysts with a comprehensive series of questions that explore all the factors that must be considered, will form the basis for PRATIQUE’s investigations and for the dissemination of its results. The principal outputs of PRATIQUE: datasets (WP1), guidance and enhancements to the PRA scheme (WPs 2-5) will be directly incorporated into the EPPO PRA scheme or provided as stand-alone modules and simple links. The EPPO PRA scheme, modules and links will all be available through a simple computerised interface.
Teams of independent experts from EPPO Panels, EPPO expert working groups and project partner institutes will test the revised PRA scheme including the draft protocols and guidance and their feedback will be used to improve the scheme and ensure it is appropriate in all circumstances irrespective of pest taxon, trophic level, habitat, whether the pest is intentionally or unintentionally introduced and whether a rapid or detailed analysis is required. Feedback on the PRA outputs will also be sought from those with experience in plant health regulation and stakeholders to ensure that the PRAs are easy to read and the key elements required for decision-making can readily be extracted. The scheme will be evaluated both by experts in PRA and those new to the subject to make sure it is useful at all levels of expertise. The testing phase will be completed six months before the end of the project to ensure that there is sufficient time for the key project deliverables to take the feedback into account.
The new computerised PRA scheme will provide risk analysts with:
- access to all technical project deliverables
- a validated, user-friendly computerised PRA scheme with detailed guidance, direct context-related links to relevant datasets, modules containing procedures to be adopted in difficult sections of the scheme, a manual and examples of best practice
Risk managers, phytosanitary regulators and policy makers will benefit because PRAs:
- will be much easier to read
- will more clearly highlight the key factors to take into account when developing phytosanitary measures, ensuring that the choices made are based on sound science, reflect uncertainties and represent the most cost-effective options while following IPPC principles of minimal impact, transparency and equivalence
- will also be much easier for stakeholders to read, so they can understand more readily the justification for the measures being proposed.
To improve take-up, further guidance will be provided with a written manual and examples of best practice. The examples of best practice will be chosen during PRATIQUE and will be selected from the PRAs used to test the methodology developed by WPs 2, 3, 4 & 5. In some cases, existing PRAs will be extended to illustrate the benefits of PRATIQUE outputs. Where appropriate, new PRAs will be constructed to ensure that the examples given are representative of the diversity of PRA types.