The amplification of plant disease risk through ecological restoration

Ecological restoration, which encompasses habitat restoration and creation is essential to “bend the curve of biodiversity loss.” However, well‐intentioned actions could alter plant pest/pathogen distributions resulting in unanticipated detrimental impacts on both host plants and associated biodiversity. Awareness among practitioners of this risk is unknown but likely to be limited. Using an online survey, with responses from 224 participants involved in ecological restoration in the United Kingdom, we assessed: (1) their awareness of plant health biosecurity risks; (2) whether they had risk assessments and biosecurity guidance/best practices for plant pests/pathogens; and (3) what, if any, new guidance was required. Participants perceived that their neighbors' activities, rather than their own, were the most likely source of pests/pathogens establishing. Their perceptions of the activities most likely to be a source of pests/pathogens did not match the literature, with the introduction of both mature plants and seeds viewed as the same level of risk. Over half the participants did not have a risk assessment for plant pests/pathogens. More of the participants (70%) had biosecurity guidance. However, of those participants that had biosecurity guidance 22% failed to check whether this guidance was followed, while 40% did not have anyone in their organization responsible for biosecurity. The most frequent requests from participants for guidance were: (1) what biosecurity to put in place for staff/contractors?; (2) the current legislation on plant pests/pathogens and movement of plants/soil/equipment. We conclude that risk assessments for plant pests/pathogens must become standard practice in ecological restoration.


Introduction
Habitat creation and restoration, termed ecological restoration throughout, is essential to alleviate the biodiversity crisis and "bend the curve" of biodiversity losses (Mace et al. 2018;Leclere et al. 2020), with global and national goals for restoration of habitats (Convention on Biological Diversity 2010). However, well-intentioned actions could alter plant pest/ pathogen distributions resulting in unanticipated detrimental impacts (Simler et al. 2019). Plant diseases can cause considerable ecological, economic, and social impacts (Freer-Smith & Webber 2017;Hill et al. 2019;Mitchell et al. 2022). The awareness of their potential impact is increasing in the horticulture and forestry sectors ) but in habitats, outside of woodlands, plant diseases are rarely considered (Mitchell 2023). Plant diseases are defined here as the plant health consequences of biotic agents, that is, pests and pathogens, referred to as pests throughout.
Novel plant pests may be introduced to new areas during ecological restoration via dirty equipment and translocation of plants. Movement of equipment ranging from boots and hand tools to large machinery is a common occurrence during ecological restoration and, unless cleaned before movement, risks the movement of pest-infested soils. Plant translocation is also often a feature of ecological restoration, and it is difficult to ensure that only the intended species is translocated. In reality, a "biological package" (Davidson & Nettles 1992) is moved that contains the plant but also microorganisms or other organisms either on or in the plant, including species regarded as pests. If the soil in which the plant is growing is also translocated, then an even wider range of organisms may be translocated, and it is difficult to know if this may include pests. The introduction of novel plant pests may impact plant species but also have cascading effects on the biodiversity associated with the infected plants (Freer-Smith & Webber 2017;Hultberg et al. 2020;Mitchell et al. 2022), and drive changes in the ecosystem services delivered if the abundance of the host plant declines substantially (Boyd et al. 2013;Mitchell et al. 2016). The introduction of non-native plant pests has been described as an insidious, mostly overlooked threat to biodiversity (Jonsson & Thor 2012).
Mature plants for ecological restoration are often sourced from plant nurseries which can be a major source of plant pests/pathogens (Osterbauer et al. 2013). An example is the detection of Phytophthora tentaculata for the first time in the United States in native plant nurseries in four California counties, and in restoration sites on orange sticky monkey flower Diplacus aurantiacus, coffeeberry Frangula californica, and sage Salvia spp. (Rooney-Latham et al. 2015). Following this discovery, a wider survey found that Phytophthora species were common on nursery stock grown for restoration and revegetation purposes in California (Rooney-Latham et al. 2019) and that 25 new Phytophthora species, including P. tentaculata, P. cactorum, and other new or new hybrid Phytophthora species, had been unintentionally but extensively introduced into restoration areas in the greater San Francisco Bay Area (Eshleman et al. 1998;Garbelotto et al. 2018). This led to an extensive response to coordinate efforts to reduce their spread . In Britain, Riddell et al. (2020) suggested that Phytophthora austrocedri infections of native Juniper Juniper communis arose due to the introduction of infected nursery stock into populations. A further example is that of young sessile oak Quercus petraea seedlings grown by commercial nurseries; seeds collected from an ecologically sensitive site in Scotland and destined for translocation back to that site, once grown, were found to harbor the root pathogen Phytophthora quercina . Gyeltshen et al. (2021) suggests that in Australia nurseries supplying seedlings for restoration sites should screen their plants for the presence of Phytophthora cinnamomi which has driven extensive declines of native vegetation in forest ecosystems and is limiting the species that can be established during ecological restoration (Sena et al. 2018a(Sena et al. , 2018b. Despite these examples of pest spread during ecological restoration, there are large inconsistencies in recognizing plant disease risk and biosecurity in ecological restoration guidance. In Australia, in particular, there are examples where plant disease risks are recognized, for example in rare plant translocation guidance (Commander et al. 2018). However, plant disease risk is rarely mentioned in many other guides. For example, The Society for Ecological Restoration's International principles and standards for the practice of ecological restoration (Gann et al. 2019) briefly mentions that restoration sites should be protected from the introduction of undesirable pathogens but provides no details of how to do this and there is no explicit mention of the risks posed by plant material and soil in transporting plant diseases. Similarly, the Primer on Ecological Restoration, fails to mention plant health (SER 2004). There are few international guidance documents for ensuring that native seeds used in ecological restoration have the same standards of quality assurance that are a regular practice in the crop and horticultural industries but see Pedrini and Dixon (2020).
The most effective way to limit the risk of new pests impacting restored or newly created habitats is to stop their establishment. It is, therefore, key that those involved in ecological restoration are aware of the potential risks and have protocols in place to limit any accidental introductions. However, our knowledge of the awareness of practitioners about the risks plant health poses to natural habitats is limited. We used a survey of stakeholders involved in ecological restoration in the United Kingdom to better understand: (1) their awareness of plant health biosecurity risks; (2) what risk assessments and biosecurity guidance/best practices they currently use; (3) what, if any, new guidance is required.

Survey design
A questionnaire was designed with 19 questions (Supplement S1) split across four sections. The first section aimed to find out about the participant's role in ecological restoration, termed habitat restoration and creation in the questionnaire, and their knowledge of plant health risks. The second section asked questions about what the participants perceived as the greatest risks in relation to plant pests. The third section asked for information about what risk assessments and biosecurity guidelines/best practices the participants used in relation to plant health. The final section asked about knowledge gaps in relation to plant health risk assessments and biosecurity guidance/best practices. The questionnaire was approved by the James Hutton Institute's ethics committee before circulation. Informed consent was given by all participants. The questionnaire started by providing guidance on some terminology for the participants (Table 1). The questions were of four types: (1) participants were asked to score different habitats, drivers of biodiversity loss, or sources of pests from, for example, 1-very low to 5-very high; (2) participants were asked to select No, Do not Know, Yes; (3) participants able to select as many options as relevant to them from a list, for example, habitats in which they worked; (4) participants asked to assess if the risks of plant pests were greater, about the same or lower than a range of options. For all questions there was an option to select "Do not know" and participants did not have to answer all questions. For some questions, there was an additional free text box to allow participants to provide further information if they wished. The questionnaire was conducted via the software Qualtrics (Qualtrics 2005).

Survey participants
A web link to the questionnaire, together with an explanatory email was sent to 245 individuals and 88 organizations working within the United Kingdom. The individuals were known to be involved in habitat creation or restoration. For the organizations, we requested that they circulated the questionnaire to Restoration Ecology July 2023 appropriate staff members: those involved in habitat creation and restoration. The organizations included relevant professional bodies, environmental non-governmental organizations (NGO) (charities working in the environment), and government organizations with responsibility for the environment. The organizations operated at a range of scales: county level, devolved nation (Scotland, England, Wales, Northern Ireland), and national (UK-wide). In addition, the questionnaire was promoted via social media and websites. The survey ran from 29 September to 30 November 2021.
There were 224 responses to the questionnaire, although not all participants answered all questions.
Most of the participants carried out habitat restoration and creation work in England (125 participants) and Scotland (87 participants), with 14 participants working in Northern Ireland and 24 in Wales. Participants could select more than one country in which they worked, hence more responses than participants.
Most participants had moderate knowledge of biosecurity and habitat restoration. However, the majority said they only had basic knowledge about risk assessments and either basic or moderate knowledge about plant pests (Fig. S1). Most participants were involved in ecological restoration by providing advice, doing practical work on the ground, and/or planning and designing the ecological restoration work. Some participants were also involved through policy, as the landowner or agent, or in other aspects (Fig. S2a). They largely worked for environmental NGO or government agencies but some were self-employed, landowners, contractors, or working for consultancies (Fig. S2b). Most participants worked on ecological restoration in woodlands, hedges, and trees outside woodlands, grasslands, and freshwater habitats (Fig. S2c).

Data analysis
All data analyses were carried out in R software version 3.6.2 (R Core Team 2018). Ordinal data, that is, data that is ordered categorical data such as when participants were asked to score something from 1-very low to 5-very high, were analyzed by the ordinal package in R using a cumulative link model (clmm) (Christensen 2019). Ordinal data came from scoring different habitats or sources of pests. The participant was included as a random effect. For each analysis a cumulative link model including the habitat, or source was tested, using a chi-squared test, against a null model without this factor. When analyzing how participants scored habitats, whether the participant worked on that habitat and the interaction term between whether the participant worked on that habitat and the habitat were included as fixed effects. This model was tested against a model without the interaction term to determine if the participant working on a habitat influenced how the participant ranked the habitat. Tukey's pair-wise comparisons were used to determine differences between pairs of options within any one question (such as differences between habitats) using the package emmeans and p values were adjusted using the Tukey correction method for multiple tests (Lenth 2019). The emmean values were used to rank the options, for example, which habitats were most at risk from pests or which sources of pests were the highest risk. The scores from participants were used to calculate the median likelihood value and hence the median score given by the participants. Spearman rank correlations were used to test the correlation between two groups of categorical variables.
Binomial data, questions with a yes/no response, were analyzed using generalized linear models with a binomial distribution using the lme4 package (Bates et al. 2015). When Table 1. Terminology used in the survey about plant health risks and ecological restoration.

Terminology
Definition used in survey Biosecurity guidance/ best practice Actions or procedures on the ground (e.g. at restoration sites or in propagation facilities for growth of plant material used in restoration projects) that should limit the risks of plant pests and pathogens spreading (e.g. cleaning of footwear and equipment). These may or may not be compulsory to follow.

Ecological restoration
The process of assisting the recovery of an ecosystem that has been degraded, damaged, or destroyed by habitat creation or restoration. Plant health A decline in plant health is a decline in plant fitness due to pests and pathogens; it does not include a decline in fitness due to unfavorable growing conditions (such as weather, soil conditions, etc.), damage by vertebrate herbivores, or stress caused to the plant by unfavorable management practices. Pathogens A virus, bacteria, fungus, or fungus-like organism that causes disease. Pests Invertebrates that cause damage to a plant, either by direct action or by acting as a disease vector, for example, insects and nematodes. Plant pests and pathogens Include both native and non-native pests and pathogens, and include both those pests and pathogens already present within the United Kingdom and those that might establish. Habitat restoration Any management to restore an area of land to more favorable ecological status, which may include the translocation of plants.

Habitat creation
The creation of a new habitat which may include the translocation of plants.

Risk assessment
An assessment made before an activity is carried out to identify and assess the potential impacts and risks of that activity. In relation to plant health this would include assessments of where staff/equipment/plants to be translocated may come into contact with plant pests and pathogens and how the proposed activity may result in the plant pests and pathogens being transported/introduced to a new location. Translocation The movement of plants at any stage of their life cycle (seeds, cuttings, adult plants), including both vascular plants and bryophytes.
analyzing whether the participants had risk assessments, biosecurity guidance/best practices, or people responsible for biosecurity within their organization, who the participant was employed by (environmental NGO, government agency, selfemployed, private landowner, contractor, consultancy) was included in the analysis as a fixed effect to test if there were differences between employers. Once again differences between pairs were assessed using Tukey's pair-wise comparisons as described above.

Awareness of plant health risks and their impacts
Participants assessed the likelihood (from 1-highly unlikely to 5-highly likely) of pests establishing from 10 different sources (Fig. 1). Source was significant in explaining the scores given by participants (comparison of a model including source against a null model excluding source, χ 2 (7, N = 159) = 185, p < 0.001). Spread from neighboring sites was ranked highest risk whereas the introduction of seed was ranked lowest (Fig. 1). Calculating the median likelihood for each source showed that spread from neighboring sites, movement of machinery, walkers and recreation, and movement of soil were all ranked as high risk. Spread by deer and other animals, other land management operations, and the introduction of mature plants and seeds were all classed as medium risk.
Participants perceived the likelihood of pests establishing to significantly differ between habitats ( Fig. 2A). Habitat was significant in explaining the scores given by participants (comparison of a model including habitat against a null model excluding habitat χ 2 (9, N = 168) = 537, p < 0.001). Woodlands were ranked as being at significantly greater risk than all the other habitats except freshwater habitats. Alpine/montane habitats were ranked as being of lower risk than all other habitats. How the participant scored the likelihood of a pest establishing in a habitat was not influenced by whether the participant worked in that habitat ( p > 0.05 for the interaction term of whether the participant worked on that habitat and the habitat).
The potential impact of plant pests on biodiversity was perceived by participants to differ significantly between habitats (Fig. 2B). Habitat was significant in explaining the scores given by participants (comparison of a model including habitat against a null model excluding habitat χ 2 (9, N = 156) = 335, p < 0.001). The impact of plant pests on biodiversity in woodlands was ranked significantly higher than the impact on biodiversity in all other habitats and the impact on alpine/montane habitats was ranked significantly lower than all other habitats. How participants perceived the potential impact on biodiversity was not influenced by which habitats they worked on (p > 0.05 for the interaction term of whether the participant worked on that habitat and the habitat). There was a significant positive correlation (Spearman's rank correlation p < 0.001, rho = 0.655) between the participant's perceptions of the likelihood of a pest establishing in a habitat and the potential impact it might have on biodiversity.
Compared to other drivers of biodiversity loss, participants ranked the establishment of plant diseases as a lower threat to biodiversity (i.e., the potential to cause a decline in biodiversity) than agricultural practices, habitat loss, and land management. Plant pests were perceived to be about the same level of threat to biodiversity as climate change, grazing by wild herbivores, other invasive non-native species, nitrogen pollution, sulfur pollution, and urban development. Plant pests were not viewed as a greater threat to biodiversity than any of these other drivers of biodiversity loss.

Risk assessments, biosecurity guidance, and monitoring
Fifty-one percent of participants either did not know if they had or did not have a risk assessment for plant pests when carrying Restoration Ecology July 2023 out ecological restoration (Fig. 3A). More of the participants (70%) had biosecurity guidance/best practice for ecological restoration (Fig. 3B). However, of those participants that had biosecurity guidance/best practice 22% did not check if this guidance was followed (Fig. 3C). Forty percent of the participants either did not know or were not sure if there was anyone responsible for biosecurity within their organization (Fig. 3D). Nineteen percent of participants said that no one was responsible for biosecurity in their organization and 40% of participants knew that there was someone in their organization who had responsibility for biosecurity during ecological restoration. Who the participant was employed by had no significant impact ( p > 0.05) on any of these results.
Thirty-six percent of participants monitored the ecological restoration for plant pests following the completion of the work (Fig. 4A). Thirty-seven percent did no monitoring and the remaining 27% selected the "other" option (Fig. 4A). After all the questionnaires were completed the free text in the "other" box was used to categorize these responses (Fig. 4B). Of the 40 participants that selected the "other" box 42% of them said that while they did not do any specific monitoring for pests, they  = monitoring is not carried out; Yes = yes monitoring is carried out and the time frame indicating for how long after completion the monitoring is carried out. (B) uses the free text box of the "other" category in (A) to identify why the participants selected other. Yes = monitoring was carried out but they did not know for how long; general habitat monitoring = no specific monitoring for plant pests and pathogens occurring but the participants thought that any plant pests/pathogens would be picked up in general habitat monitoring; no responsibility = the participant stated that post-completion monitoring was not the responsibility of their organization; do not know = the participant do not know if post-completion monitoring for plant pests/pathogens occurred. thought that pests would be picked up in general habitat condition surveys. Other reasons for selecting the other box were that the participants said monitoring for pests did occur but they did not know over what time period (20%), it wasn't their organization's responsibility to do any post-construction monitoring (15%), or that they did not know if any monitoring was done (22%).

Sources of information and further guidance
Just over a quarter of participants (26%) felt that the guidance they had access to, about the plant health risks associated with ecological restoration, was sufficient. Thirty-seven percent thought it was insufficient and 36% did not know.
Participants got most of their information about plant health risks from colleagues, followed by government bodies, websites, and professional newsletters/magazines (Fig. 5). Few participants obtained such information via social media.
When asked what further guidance participants would like with respect to plant health, between 45 and 38% of participants wanted guidance on 9 of the 10 different options provided (Table 2). While biosecurity for staff and contractors was the top request (45% of participants) this was followed closely by more information on the current legislation on plant pests and movement of plants/soil/equipment (40% of participants) and biosecurity for the general public (39% of participants). Advice on propagating plants free of pests was the only type of guidance for which there was no high demand (Table 2).

Discussion
This study represents a first attempt to better understand the awareness of those involved in ecological restoration of the potential risks posed by plant pests. The study showed that: (1) participant's perceptions of risk do not match the literature with, for example, mature plants being ranked as a lower risk than suggested by the literature; (2) there is a lack of knowledge about the potential risks that plant pests pose to habitats outside of woodlands and to wider biodiversity; and (3) risk assessments identifying the risks to plant health, and biosecurity protocols to mitigate or reduce the identified risks, are not standard practice during ecological restoration and must become so.

Perceptions of risk
If neighbors are perceived as the most likely source of infection of pests rather than any activity carried out by the participants, this raises the question of how much participants are prepared to alter their activities. For example, they may feel it is not worth changing their activities to reduce the risk if the greatest risk is from their neighbors. This is similar to the attitudes of those involved in large scale-landscaping projects (Karlsd ottir et al. 2021). In that study participants often expected biosecurity expertise to come from elsewhere, mainly the landscape contractor or the supplier to source plants responsibly.
Contrary to the views of the participants in this survey, the literature suggests that the use of seed in ecological restoration is considered intrinsically lower risk than translocations involving living vegetative tissue, as many (but not all) plant pathogens are not transmitted by seed (Anderson et al. 2004). When mature plants are moved, a "biological package" (Davidson & Nettles 1992) is moved that contains the plant but also any organisms that are on that plant or in the soil surrounding the plant. This may provide a route for pests to establish during ecological restoration work (Migliorini et al. 2015). In addition, signs of pests may easily be mistaken for signs of stress due, for example, to over or under-watering, and may be ignored during the restoration process. Therefore, the use of mature plants in habitat restoration is a much higher risk than seed (Mitchell et al. 2023).
If risk assessments for plant pests are to become standard practice for ecological restoration, as we suggest it should be, it is important that the risks are correctly assessed. This survey suggests the perceptions of risk of those involved in ecological  Restoration Ecology July 2023 restoration may not be accurate, and further education/ knowledge exchange may be required to help ensure the risks are accurately assessed.

Lack of knowledge of potential risks and impacts
The participant's assessment of the habitats most at risk appeared to be related to those habitats about which there had been the greatest publicity. Woodlands were scored as having the greatest likelihood of pests establishing in them followed by freshwater habitats. These two habitats have had substantial publicity recently in the United Kingdom. In particular, there has been considerable publicity about the impact of tree pests (Fellenor et al. 2018;Hill et al. 2019). While this study cannot prove a direct link between publicity and the participant's assessments, it is suggested as likely. If there is such a link this is positive as it shows that the publicity was successful. However, it also highlights the lack of knowledge of most participants about the potential risks in other habitats. For example, Mitchell et al. (2023) found that 161 plant pests listed in the UK Plant Health Risk Register (Defra 2021) could be hosted by vascular plant genera that occur at more than 25% of cover on Scottish moorlands and heathlands. The lack of awareness of plant diseases in natural habitats compared to agriculture, horticulture, and forestry has been highlighted by Mitchell et al. (2023). The significant correlation between how participants assessed which habitat was most at risk from plant pests, and how participants assessed the potential impact of plant pests on biodiversity, suggests that they are unable to distinguish between the two. There could be a habitat that might be at low risk of pests establishing but, if a pest did establish, it would have huge consequences for biodiversity. For example, if a pest hosted by Calluna vulgaris established in moorlands or heathlands (Mitchell 2023).

Risk assessments and biosecurity procedures
The results from this study concur with Guberti et al. (2014), who states "when dealing with the conservation-disease interface, a standardized risk assessment procedure including risk mitigation has to become the rule." Our results highlight several potentially serious shortcomings with respect to risk assessments and biosecurity protocols. First, most participants either did not have or did not know if they had a risk assessment for plant pests. While over 70% of participants did have best practice guidance in place that might be expected to reduce the risks of pests establishing, if a risk assessment has not been carried out first then it is not clear that the biosecurity protocols in place were appropriate for the risks. Second, nearly a quarter of participants did not check if the biosecurity protocols were followed and thirdly most of them did not know or did not have someone responsible for biosecurity in their organization. Filling these three gaps would be a quick win in terms of improving biosecurity in ecological restoration.
The ranking of requests for further guidance suggests that practitioners do not understand the need for risk assessments and are more focused on the implementation of biosecurity. How to develop risk assessments was ranked 8 out of 10, while guidance for what biosecurity to put in place for staff/contractors and the general public was ranked first and third respectively. In animal health, there is formal guidance on how to conduct a disease risk analysis (Ewen et al. 2023;Sainsbury & Carraro 2023). Such guidance is lacking for plant health, although recently Mitchell et al. (2023) identified some key factors that should be considered in all plant conservation translocations with respect to assessing biosecurity risks, and these contain many of the elements of a disease risk analysis that would be applicable to ecological restoration. Mitchell et al. (2023) suggest factors to consider including (1) assign responsibility for plant health risk assessments and biosecurity; (2) assess whether there are known pests and pathogens of the species to be transplanted; (3) assess the risks of spreading pests and pathogens at the site the plant is taken from; (4) assess the risks of pest and pathogen infection during transportation and propagation phases; (5) assess whether the quarantine plan and pest and pathogen monitoring steps are adequate; (6) assess whether the translocated plants may spread disease at the destination site.
The lack of anyone being responsible for biosecurity means that biosecurity is often neglected with people thinking that it is someone else's responsibility. This result concurs with studies of biosecurity in other sectors such as horticulture (Marzano 2021) and large scale-landscaping projects (Karlsd ottir et al. 2021). If standardized risk assessments become the rule for ecological restoration, this would raise the profile of plant health within organizations and encourage organizations to establish who is responsible.

Next steps
There are multiple types of ecological restoration, and some will be riskier than others. There is thus a balance to be achieved between ensuring greater awareness of the plant health risks and not burdening low-risk ecological restoration projects with unnecessary biosecurity measures, which become a barrier to "bending the curve for biodiversity" and mitigating the biodiversity crisis.
As participants appear to gain most of their knowledge from colleagues, it is important that those carrying out ecological restoration are educated about the risks to plant health from pests. This will enable them to pass this knowledge on to their colleagues and make appropriate assessments about the risks to plant health ecological restoration. It will also ensure that wellintentioned actions do not alter pest distributions, resulting in unanticipated detrimental impacts and the amplification of plant disease risk through ecological restoration.

Acknowledgments
The work was funded through the Plant Health Centre, Scotland's Centre of Expertise Project PHC2020/03 and through the Scottish Government's Strategic Research programme 202-227 project JHI-D4-2. The author thanks Duncan Stone and Pete Hollingsworth for useful discussions which formulated many of the ideas in this manuscript and for their, and Mariella Marzano's comments on draft versions of the questionnaire used in this study. Altea Lorenzo-Arribas and Graciela Martínez S anchez from Biomathematics and Statistics Scotland provided helpful advice on the data analysis. Ian Toth provided helpful comments on an earlier version of the manuscript. The survey and the anonymous results from the survey on which this paper is based are available at: https://doi.org/10.5281/zenodo. 7669877.

Supporting Information
The following information may be found in the online version of this article: Supplement S1: The questionnaire. Figure S1: How participants ranked their knowledge about biosecurity, habitat restoration, plant health risk assessments and plant pests/pathogens. Figure S2: How the participants of the questionnaire were involved in habitat creation/ restoration a) The type of work they did; b) Their employer, c) The habitats they worked on.