A horizon scan for temperate pastoral weed science–a New Zealand perspective

ABSTRACT Pastures represent about half of the global agricultural area and productivity losses from weeds are significant. The complex interactions between them and other pasture plants, livestock and the environment imply a need for innovative research that transforms pasture management. To this end, a horizon scan was conducted to identify relevant issues, questions, opportunities, and drivers. The drivers were ranked using three criteria: (1) is this a horizon (is the driver likely to become important in 10–20 years?); (2) will the research require stretchy science (is it currently not well addressed by the science community?); (3) is the research transformative (will successful scientific research in this area lead to significant changes to weed management in pastures?). We identified 11 major issues and 46 subordinate ones. The three highest ranked major issues were: (1) anticipated reductions in access to herbicides; (2) rethinking weed management under an ecosystem services paradigm; (3) responding to shifts in best practice and the regulations that are altering farm system planning to reduce farming’s environmental impacts. We conclude that fundamental interdisciplinary research is needed that addresses biosecurity and weed management issues, while reducing the environmental footprint of farming and maintaining productivity.


Introduction
Weeds have important negative impacts on global food security.Reviews of crop-protection for major staple crops shows their control can potentially mitigate crop losses by Pasture weeds are a problem due to their abundance, impact on pasture dry matter yields and animal production, and the difficulty of managing them with herbicides without damaging desirable pasture plant species.The number of weeds estimated to occur in pastures in New Zealand varies from 187 (Bourdôt et al. 2007;Saunders et al. 2017) to 245 species (Ghanizadeh and Harrington 2019).Some of these species may have unrecognised beneficial value as sward components and thus should not be considered weeds.For example, plantain (Plantago lanceolata L.), had been considered a weed of pastures, but after considerable plant breeding effort it is now sown in drystock pastures with white clover as a high-quality forage mix, and is increasingly sown in dairy pastures to mitigate nitrogen leaching (Cheng et al. 2018).
Weed impacts on farm productivity or costs have been measured in a variety of ways.For example, at densities of 10,000 plants per hectare of the thistle, Cirsium vulgare (Savi) Ten.,. can reduce sheep liveweight gain by 20% (Hartley 1983).Similarly, the thistle Carduus nutans, at 1000 plants per hectare, can reduce forage dry matter yield by 8% (Hartley 1983;Thompson et al. 1987).A recent farm systems analysis indicated that profitability of an East Coast hill country sheep and beef property was reduced by $87 NZD and $135 NZD per hectare by the presence of C. arvense and Silybum marianum (L.) Gaertn., respectively (Stevens et al. 2019).At the national scale, there is wide variation between species and their impacts, as illustrated by the losses attributed to 10 species for which data is available (in NZD million/year [2014]): C. arvense 702; Setaria pumila (Poir.)Roem.& Schult.258; R. acris 210; Ulex europaeus L. 72; Nassella trichotoma (Nees) Hack.ex Arechav.27; Rubus fruticosus L. 14; Cytisus scoparius (L.) Link 8, Rosa rubiginosa L. 7; Nassella neesiana (Trin.& Rupr.)Barkworth 0.013; and Hieracium spp.6 (Thompson et al. 1987;Bourdôt et al. 2007).In comparison rangeland weeds in the United States were estimated to cost ranchers USD $5 billion annually just for control (Pimentel et al. 2005), with the value of the productivity loss component likely to be several times higher.The estimated aggregate cost of New Zealand's pasture weeds based on the 10 species (and therefore a conservative estimate) is NZD 1.7 billion [2014] per year (Saunders et al. 2017).
New Zealand is known for relatively intense herbicide-based weed management in pastures.This has resulted in multiple cases of herbicide resistance in New Zealand (Carduus nutans L., C. pycnocephalus L., R. acris, N. neesiana), more species than in Canada, United States, Hungary, Czech Republic, and Switzerland which have cases from pastures, but not higher than Australia which has seven resistant species (Heap 2021).The increasing prevalence of herbicide resistance in other agricultural sectors in New Zealand is making weed management more difficult but the prevalence of resistant weeds in pastures has not been estimated (Buddenhagen et al. 2020(Buddenhagen et al. , 2021;;Ghanizadeh and Harrington 2021).
The AgResearch pastoral sector weeds research strategy briefly outlines key challenges for pastoral weed research in New Zealand and reviews the weeds research capacity in AgResearch Ltd., in New Zealand, and internationally (Bourdôt et al. 2018).The process of generating that strategy, although similar to a horizon scan, emphasised local stakeholder weed management priorities and did not specifically aim to identify transformational science needs.The development of this horizon scan involved eight of the participants that were involved in developing the research strategy.Some scientists and senior managers believe that weeds are a perennial problem and that the relevant stakeholders (farms and agricultural chemical companies) are best placed to drive the innovation and fund research (believing that public-good funding priorities should be elsewhere).Others go so far as to say that the technology-driven weed research is not science, and that the real science is based in ecological theory (Ward et al. 2014).The persistence of this belief is probably because weed science is largely an applied field with effective management being the goal.It is dominated by repetitive, broadly similar studies examining the effectiveness of different control technologies e.g.herbicide use, efficacy, and resistance (Ward et al. 2014).Such studies are mostly distinguished by species, crop and geography rather than by wide sweeping innovations (Fernandez-Quintanilla et al. 2008).A lot of weed science expertise is embedded in herbicide companies, and only partially visible to the public through product documentation or direct advice to farmers.Most weed science and extension efforts are driven by farmer concerns, address specific local issues, and information is often available in the difficult-to-access 'grey' literature.In the public sector, scientists often find themselves on a treadmill of seeking research funding driven by a 'publish or perish' paradigm, which may limit their ability to creatively develop theoretically driven transdisciplinary collaborations designed to solve the overarching problems (Fernandez-Quintanilla et al. 2008;Paasche and Österblom 2019).Despite the importance of weeds in food production, basic knowledge about weed biology and ecology (e.g.seed biology, population dynamics) is often lacking.In short, a focus on immediate farmer concerns may tip the balance away from the more transformative, transdisciplinary strategic ecological and genetic research (Fernandez-Quintanilla et al. 2008).In this context, strategic research to support weed management in pastoral systems is needed.The identification of transformative research that addresses emerging issues is a shared aspiration for weed scientists, policy makers, government agencies, industry groups, land managers, and funding agencies.We have adopted the horizon scan approach as a means to meet this aspiration (Sutherland and Woodroof 2009;Ricciardi et al. 2017;Neve et al. 2018;Sutherland et al. 2019).

Methods
Our horizon scan was designed to: . be medium to long term (10-20 years), .focus on emerging threats and opportunities, .raise awareness and provide momentum for innovation, .inform policy, .help set research agendas, and .stimulate actions that address plausible threats.
Horizon scans generally identify opportunities and threats that need to be addressed but may or may not involve a process for expert ranking and scoring of questions (Sutherland et al. 2011).In late 2019 two independently facilitated expert elicitation workshops were held, one at Hamilton (29/11/2019) and the other at Lincoln (12/12/ 2019) in New Zealand.Each workshop involved 9-13 scientists and breakout groups of 3-5 people.Group research interests covered social science, biosecurity, weed control, invertebrate pests, soils, agrichemical marketing, conservation biology, modelling, risk assessment, Māori agribusiness, population and evolutionary genetics, ecology, and molecular biology.Participants were introduced to the concept of horizon scans outlined in the bulleted list above.Each group was then allocated 20 min to write down their issues and questions under key themes (previously identified by workshop leaders) on Post-it Notes and poster paper, and then peer-reviewed the other groups' ideas.
The key themes used to prompt workshop participants were: (1) low-input future (fertilizer, herbicides, pesticides); (2) agriculture and global change (climate, water quality, soil, biosecurity); (3) fit for purpose plants and animals; (4) robotics/tech/big data, remote sensing; (5) omics (genomics, proteomics, metabolomics, microbiomics, gene manipulation and gene-silencing techniques); and (6) mātauranga Māori (traditional Māori knowledge, cultural practices, and principles).Ideas, issues and questions generated by the groups were documented before synthesising them into 11 major issues which were then grouped with 4-5 related subordinate issues and questions.The issues were later shared with three participants who checked that the synthesis reflected the intent of the workshop.
The 23 workshop participants were asked to independently rank five randomly assigned major issues or drivers (not all ten) and the indigenous (Māori) knowledge systems (Table 1).Then for each person's top three most highly ranked issues they were given a chance to rank the linked subordinate issues (Buddenhagen 2022).Every participant was given the chance to rank the mātauranga Māori theme as a driver of innovation because there are national and institutional initiatives to improve Māori participation and incorporate their world view into our science.The ranking process was carried out in October 2020.Randomly chosen issues or drivers and the mātauranga Māori were ranked on three criteria: (1) is this a horizon (is the driver, issue or question likely to become more important in 10-20 years)?; (2) will it require stretchy science (is the question or issue currently not well addressed by the science community)?; (3) is it transformative (will successful efforts in this area lead to significant changes to weed management in pastures)?If a respondent chose not to rank an issue, no score was assigned.Mean ranks for each issue and criteria were converted to scores by subtracting the rank from the number of items ranked plus one (in this way the lowest ranked item scored a 1 and the highest a 6), then by taking the square of the mean score for each of the three criteria.So, if the average 'horizon' rank is for a driver is 2.5, then the weight is (7 −2.5) 2 = 20.25.A similar score would be obtained for the stretch and management criteria all of which contributed to the overall score for each driver (Table 1, Figure 1).Finally, the importance of different issues was visualised using a Sankey plot from the networkD3 package (Allaire et al. 2017).

Results and discussion
Major issues selected and ranked A summary of the final ranks for the major high-priority issues is provided in Table 1 (data is available from: Buddenhagen 2022).Figure 1 shows the relative importance of the major issues (rank averages converted to scores or weights) described in terms of a short summary phrase that captures the issue concept while the full descriptions are provided in Table 1.From an original list of 11 items, the high priority issues by weighted scores were: 1st-reduced access to herbicides; 2nd-ecosystem services; 3rd-farm system changes; 4thlower chemical inputs; and 5th-internal and border biosecurity (Table 1; Figure 1).The lowest ranked issues were: 11th-Big data, automation, and technology; 10th-Indigenous knowledge/mātauranga Māori; 9th-water quality; 8th-fit for purpose plants and animals; 7th-climate change; and 6th-weed, animal and soil microbiomes.The top agricultural drivers of land use change in New Zealand in a recent survey were water quality, climate change, environmental concerns, agricultural policy and biosecurity (Driver et al. 2022).Meanwhile challenges (target research areas) identified by Bourdôt et al. (2018) that were similar included: increased suite of non-chemical control options (alternatives to herbicides); herbicide resistance evolution understood and controlled; best practice pastoral weed management adopted by farmers.However, they mentioned topics not ranked highly here Table 1.Mean ranks provided by the workshop participants for the 11 major issues raised in the horizon scan workshops ordered by mean rank.

Number of participants assigning a rank
(1) There is reduced availability of chemical weed control options due to environmental and consumer health concerns which lead to the removal of active ingredients from the market and a lack of new alternatives which are costly to develop 2.57 16 (2) A better understanding of ecosystem services/disservices could lead to changes in farmer perceptions and weed management.For example, perceptions of weediness have changed over time (e.g.chicory and plantain were once regarded as weeds but now planted in pasture) 2.93 10 (3) Farm system changes are increasingly required to encourage better management of multiple values, biosecurity risk, biodiversity, carbon, and nutrient run-off and production value.These plans result in spatial and temporal farm system changes that will alter weed ecology and weed management

8
(4) There is pressure to use lower chemical inputs (herbicides, pesticides and fertilizers) due to concerns about their perceived impacts to the environment and human health.This will alter pasture weed species composition, abundance and management 3.08 13 (5) The effectiveness of biosecurity at the border, and internally (between farms and regions) is increasingly important as economic/population growth, and international immigration (e. including improved internal biosecurity; weed population biology informs and disrupts weed management; weed control economics informs and disrupts weed management; and effective riparian weed management.
Major issues ranked under the horizon, stretch and management criteria Ranks differed depending on the criteria, and our weighting of the ranks emphasises the relative contribution under the three criteria (Figure 1).Under the 'horizon' criteria the anticipated shift toward lower chemical inputs was ranked highly (Buddenhagen 2022).This is related to policy and regulatory shifts that require farmers to address sustainability, and social-licence challenges (Rowarth et al. 2020;Clothier et al. 2021;Caradus et al. 2021).It is easy to see how this is related to the farm system changes, which is about changes to best practice and regulations and was scored highly.These changes are expected to drive far-reaching changes in farm and weed management.This is also related to the other highest ranked horizon issue, namely the recent reductions in the number of herbicide active ingredients that are available for farmers to use (Phillips 2020).Locally and internationally some regulators have removed herbicides from markets or restricted their use because of actual or perceived problems with their safety.Meanwhile, there has been a simultaneous slow-down in the development of new herbicides and modes-of-action because of the prohibitive costs of bringing new products to market (Phillips 2020).Effective, widely used, low-risk herbicides (e.g.glyphosate) are currently under review by regulators in New Zealand and elsewhere.Some say there is a growing regulatory emphasis on hazards (potential to harm at any use level) as opposed to risk, i.e. the potential to cause harm under recommended rates and use conditions (Duke 2012;Phillips 2020).
Looking at the ranks given for 'stretchy science' the ecosystem services concept, and microbiomes were ranked highly (Figure 1).Ecosystem services research provides rich opportunities for understanding the role (services and disservices) of weeds and/or their management in pasture systems (Harrington et al. 2006;Dale and Polasky 2007;Vilà et al. 2010;Blaix et al. 2018), potentially shifting farmer attitudes toward some weeds, as it has for chicory (Cichorium intybus L.) and plantain, which were once regarded as a weeds but are now planted as valued fodder (Dodd et al. 2017;Cheng et al. 2018).For microbiome research, its potential was framed in terms of understanding or altering allelopathic interactions, plant palatability, and plant ecology, which in turn could be applied to weed management (Masteling et al. 2019;Dahiya et al. 2020).Finally, participants believed that research addressing the reduced availability of herbicides and looking at ecosystem services could transform weed management and the concept of 'best practice'.
Other strategic reviews have attempted to identify emerging issues, innovations and opportunities in the weed and invasive species space (Fernandez-Quintanilla et al. 2008;Ward et al. 2014;Ricciardi et al. 2017;Westwood et al. 2018;Neve et al. 2018).Global food security, climate change, environmental degradation, escalating rates of plant invasion, higher global connectivity, naturalisation from introduced species pools (Hulme 2020; Brandt et al. 2021), evolution of resistance to herbicides and the systemic failure to adopt integrated weed management are converging problems requiring scientific input (Neve et al. 2018).Horizon scan participants in this study recognised the need to carry out transdisciplinary research, include social dimensions and link weed science to theory (e.g.ecological theories); themes that came up in other efforts (Ward et al. 2014;Neve et al. 2018).
Information technology, precision agriculture and robotics were identified as needing development in this study though the topic was ranked lowly compared to others (Westwood et al. 2018).For example, a lot of the technology solutions using image analysis or robotic weed control are better suited to simple monocultures rather than diverse pastures.Participants in our workshops debated the value of public investment in technology, and the extent to which public-good science funding in New Zealand should be directed at technology solutions given how small local markets are.New Zealand farmers are likely to adopt and adapt technologies developed for larger markets.Participants discussed how technology-based solutions like the use of robotic weeders, automatic weed detection, distributed applications, decision-making and information systems can create issues around cost, privacy, safety, and control (farmer-owned technology versus contracting specialists).Similarly classical biocontrol comes with widely understood and manageable risks (e.g.potential non-target impacts) and a suite of perceived risks and ethical considerations, with long lead-up times before agents can be released; this topic got surprisingly little attention in our analysis (Hayes et al. 2013).
Participants believed that the rapid development of genetic tools could create great opportunities to develop fit-for-purpose plants, through breeding or genetic modification.Variety development could improve crop/pasture competitive ability relative to weeds, improve food security, farm profitability, and potentially address important climate change issues.Some participants had expressed concerns that regulations against growing GMO plants in New Zealand could restrict opportunities in this area (Caradus 2022).Nevertheless, there has been a lot of research developing endophytes with agronomic benefits, studies that attempt to alter below ground root competitive ability of ryegrass (Yeates et al. 2019), as well as research into the development of ryegrass and clover varieties that reduce methane emissions from livestock and fix carbon (Grainger and Beauchemin 2011;Beechey-Gradwell et al. 2018;Cooney et al. 2021;Caradus et al. 2022).

Subordinate issues for the three highest ranked major issues and drivers
The top three major drivers across all criteria were: 1st-reduced access to herbicides; 2nd-ecosystem services; 3rd-farm system changes.Within these drivers, issues, and opportunities, several subordinate topics were grouped and are listed and described below.We also did this for indigenous knowledge and mātauranga Māori, which was ranked 10th, but this is included because the government and AgResearch Ltd. has prioritised Māori co-development and inclusion because their world-view, perspectives and collaborations add value to the land-focused research.
Reduced availability of herbicides-subordinate issues and drivers (in rank order): (1) We need to understand the viability of alternative weed control measures if important herbicides are banned.(2) Reduced availability of herbicides will result in more cases of herbicide resistance.
(3) New chemical herbicides can be developed with benign breakdown products.(4) Reduced availability of herbicides will impact our ability to respond to biosecurity incursions.
Weed control efforts in New Zealand often involve the use of herbicides and about NZD158M is spent annually on herbicides (Buddenhagen et al. 2020).Some commonly used herbicides are under review (e.g.glyphosate, 2,4-D), or their use has recently been strictly limited (e.g.paraquat; EPA NZ 2021).There is still a reliance on herbicides because of their effectiveness in improving productivity, and they have utility when responding to weed incursions.The potential removal of key herbicides from the market came up as the highest priority issue for pasture weed management.Workshop participants thought we should find viable alternatives as full or partial replacements to herbicides, and that more work is needed to allow us to continue to safely use existing herbicides.Some alternatives deserve to have continued scrutiny and testing, such as classical and inundative biocontrol (Bourdôt et al. 2007(Bourdôt et al. , 2018;;Bourdôt and Cripps 2018;Ghanizadeh and Harrington 2019), deferred grazing (Tozer et al. 2020(Tozer et al. , 2021)), or other regenerative agricultural practices.There was a hope that new herbicides with benign breakdown products could be developed in future, perhaps using innovative genetic approaches (Duke et al. 2019;Dayan 2019) or naturally occurring chemicals such as those involved in plant diseases, plant defence and allelopathy (Dayan et al. 2009).
(3) Weeds impact animal health/microbiomes.( 4) Society values grasslands with weeds, and wildflowers-can farmers receive benefits if they manage for these values?
As discussed briefly under issues ranked highly for the 'horizon' criteria, some plants considered to be weeds and hence a disservice to production, may also provide ecosystem service benefits (Vimercati et al. 2020).Research is needed that addresses the ecosystem services/disservices of plants considered to be weeds in pasture systems, since previous economic assessments focusing mostly on costs/disservices (Saunders et al. 2017).Undesirable plants (weeds) in ryegrass and clover pastures could become valued as late summer fodders or because they take up excreted nitrogen, fix carbon or reduce downstream nutrient leaching (Harrington et al. 2006;Dodd et al. 2017).Similarly, species promoted as robust pasture components capable of withstanding droughts like C 4 grasses (Crush and Rowarth 2007) often come to be considered as weeds (Lonsdale 1994;Driscoll et al. 2014) and it will be essential to manage these risks as part of sustainable agricultural practice.Many weedy plants can provide livestock nutrition (Harrington et al. 2006), but there is poor understanding of their practical value in the field as fodder, or as an enhancer of rumen or soil microbiomes.The benefits of weedy plants can vary, for example, they can serve as a food resource for pollinators of desirable species, improve soil biology, act as carbon sinks, or they may have value for aesthetic and cultural reasons e.g. as non-native wildflowers.Expertise in the areas of green economics, carbon budgeting, soil biology, social science, and ecosystem services fields can be quite specialised, and collaborations with weed and plant ecologists are needed.
Farm system changes-subordinate issues and drivers (in rank order): (1) We need to understand how to incentivize 'public good' farm management.
(2) The barriers to adopting best practice need to be understood.
(3) Certification schemes that emphasise good practices for weed management and biodiversity protection could create market incentives.(4) Grazing regimes (stocking rate, timing, and frequency of grazing) impact weed abundance and farm profits.(5) Emphasising pasture persistence over productivity alters farmer behaviour and weed impacts, and economic and environmental outcomes.
This farm systems issue captures shifts in best practice, agricultural policy and the regulatory environment as New Zealand responds to local and international drivers for land use change (Driver et al. 2022).For example, in New Zealand farm environment plans are increasingly required by national and local governments to promote good practice mainly to mitigate farm impacts on water use, contaminants, effluent, and water quality (Good Farming Practice Governance Group 2018).These plans or other regulatory processes may soon be extended to consider biodiversity, biosecurity, and greenhouse gas emissions.We need to understand what drives farmers to adopt best practices and innovate, including the interaction of the different actors, structures and functions that drive change in farmer behaviour (Espig et al. 2022).We need to collect evidence that the adoption of recommended practices do achieve the stated goals.Such work requires the interdisciplinary efforts of educators, farmers, marketers, and scientists with expertise in specialised chemical, biological, economic, and social fields.The hope is that such efforts can capture value chain opportunities beyond the farm gate (Saunders et al. 2011).A 'one-health' approach has been proposed as a unifying concept that aims to sustainably balance and optimise the health of people, animals, and ecosystems (OHHLEP et al. 2022).
Indigenous (Māori) knowledge systems-subordinate issues and drivers (in rank order): (1) Incorporating Māori world views could fundamentally change weed and farm management decisions.
(2) Best practices must be identified to ensure effective engagement between the science community and Māori.(3) We must understand which weed management issues matter more to Māori.(4) We must improve Māori science capacity and employment rates in our research institutions.
A Māori world view and improved engagement with Māori stakeholders and scientists could facilitate innovation in weed management.In this study experts in Māori agribusiness contributed through discussions during workshop planning, the workshops themselves and during write up, but they were not weed experts.We would have preferred a more far-ranging engagement with a broad range of Māori experts that would have enabled this analysis to represent a broad Māori view.A deeper look at this should be done but was not within the scope of this review.Limited participation of Māori in the workshop may have influenced the content generated and the low rank (10th) indicated here.In another study about the drivers of land use change Māori values were given high importance by 39% of respondents (Driver et al. 2022).
Surely, a Māori perspective and involvement are potentially transformative.As Harmsworth and Awatere (2013) explain, Māori view themselves as a part of ecosystems and recognise a reciprocal interdependent relationship between caring for the land (manaaki whenua) and caring for the people (manaaki tāngata).Māori owned enterprises have significant investments in the sheep, beef and dairy (Kingi 2013), and as a result have interests in weed control can be similar to others in the agricultural sector.Māori concern about weeds is inherently practical but cultural attitudes influence management decisions (McGowan 2021); a dialogue with Māori could lead to priority shifts, innovation of science or changes to management practices.However, long-term effective engagement between government funding bodies, scientists and Māori has been elusive.Hiring, mentoring, and sustaining Māori cultural experts, scientists, and technicians is one way to help long-term engagement (McAllister et al. 2019;McAllister et al. 2020b).In the context of weed research, best practices to ensure effective engagement with Māori in New Zealand should be followed.For example, scientists should codevelop projects with Māori, from project inception, through implementation, and on to the realisation of research outcomes.This engagement should incorporate mātauranga Māori, which is an indigenous knowledge system that integrates Māori philosophical thought, worldview, and practice,requires engagement with mana whenua (people with authority over the land) where any of the research is carried out (McAllister et al. 2020a).

Conclusions
For weed management, and farming in general, social licence is a cross-cutting issue for societal change, agribusiness generally and the pursuit of scientific and technical innovation in pastoral farming systems (Edwards and Trafford 2016;Camara et al. 2018;Norton et al. 2020;Driver et al. 2022).A key research area is to understand the barriers to the adoption of innovations and established best practice within a systems context (Espig et al. 2022).Importantly, social licence is required for the business-as-usual approach.Some of the key drivers identified in the horizon scan are a direct consequence of a perceived erosion of social licence because of the impacts of agriculture on natural resources.These impacts are often centred around chemical inputs (fertilizer and pesticides), irrigation, high stocking densities, animal welfare, effluent, contaminants, eutrophication, greenhouse gas emissions and biodiversity (Driver et al. 2022).Some concerns around contaminants may reflect our increasing ability to detect minute amounts in the environment (Tarazona et al. 2017).Many of these societal concerns are market driven (Lyons and Lawrence 2017), or come from government agencies through regulation and policy (Kaye-Blake et al. 2019), subsidies, or taxation (Barrett and Makale 2019).Thus, a mix of intrinsic and extrinsic motivations like these are expected to drive farmer efforts to reduce their environmental footprint.Addressing these concerns could impact weed management.
Scientists and regulatory officials need to effectively communicate while drawing on multiple lines of scientific evidence to counteract public distrust (Saba and Messina 2003).Responsible evidence-based decision making needs to involve transparent equitable processes, inclusive of the public, regulators, and scientists.The agricultural sector (and scientists working in the field) do appear currently to be working within an expanded ethical framework.We need to be guided by the goal of objectivity, even if it is hard to achieve, and be aware that our research occurs in a space between what is objective and subjective, value free and value laden, neutral and advocatory (Rykiel 2001).
This horizon scan endorses research into ecologically based weed management, based on solid autecology and investigating 'natural levers' such as biotic controls, which includes approaches such as imported and naturally occurring biological weed control agents, grazing management, and optimising interspecific competition in weed control.Regulatory levers will also play a role in shifting pastoral weed management and science should respond to that.Herbicides remain important tools in the control of weeds, but the authors anticipate that declines in their availability and acceptance will require us to innovate to manage weeds.Furthermore, there are potential avenues for shifting our view of weedy plants by understanding their utility as potentially beneficial components of pasture systems.The cost and benefits of internal and national border biosecurity risk mitigation measures need to be well understood so they can be pursued while they remain effective, but no longer than that.To enable effective participation and co-development of weed research with Māori, we need to forge a deep dialogue between the communities of practice and find common ground to pursue promising ideas.The path forward must involve local and international cross-disciplinary collaborations.

Figure 1 .
Figure 1.Across the three ranking criteria (LH boxes), the major issues (RH boxes) have their node boxes and link widths scaled by the sum of their weights = (mean(weight)) 2 under each ranking criterion.Randomly allocated drivers were ranked by 23 sector experts under each of the three ranking criteria (Horizon, Management, Stretch).