New pasture plants intensify invasive species riskDon A. Driscolla,b,1, Jane A. Catforda,b,c,d, Jacob N. Barneye, Philip E. Hulmef, Inderjitg, Tara G. Martina,h, Aníbal Pauchardi,j,Petr Pyšekk,l, David M. Richardsonm, Sophie Rileyn, and Vernon Visserm

aNational Environmental Research Program Environmental Decisions Group and ARC Centre of Excellence for Environmental Decisions, Australia; bFennerSchool of Environment and Society, Australian National University, Canberra, ACT 2601, Australia; cSchool of Botany, The University of Melbourne,Melbourne, VIC 3010, Australia; dDepartment of Ecology, Evolution and Behavior, University of Minnesota, Saint Paul, MN 55108; eDepartment of PlantPathology, Physiology, and Weed Science, Virginia Tech, Blacksburg, VA 24061; fBio-Protection Research Centre, Lincoln University, Lincoln 7647, NewZealand; gDepartment of Environmental Studies and Centre for Environmental Management of Degraded Ecosystems, University of Delhi, Delhi 110007,India; hCommonwealth Scientific and Industrial Research Organisation Land and Water, Brisbane, QLD 4001, Australia; iFacultad de Ciencias Forestales,Universidad de Concepción, Casilla 160-C, Concepción, Chile; jInstitute of Ecology and Biodiversity, Chile; kInstitute of Botany, Department of InvasionEcology, Academy of Sciences of the Czech Republic, CZ-25243 Pr�uhonice, Czech Republic; lDepartment of Ecology, Charles University in Prague, CZ-12844Prague, Czech Republic; mCentre for Invasion Biology, Department of Botany and Zoology, Stellenbosch University, Matieland 7602, South Africa;and nFaculty of Law, University of Technology, Sydney, NSW 2007, Australia

Edited by David B. Wake, University of California, Berkeley, CA, and approved October 3, 2014 (received for review May 19, 2014)

Agricultural intensification is critical to meet global food demand,but intensification threatens native species and degrades ecosys-tems. Sustainable intensification (SI) is heralded as a new ap-proach for enabling growth in agriculture while minimizingenvironmental impacts. However, the SI literature has overlookeda major environmental risk. Using data from eight countries on sixcontinents, we show that few governments regulate convention-ally bred pasture taxa to limit threats to natural areas, eventhough most agribusinesses promote taxa with substantial weedrisk. New pasture taxa (including species, subspecies, varieties,cultivars, and plant-endophyte combinations) are bred withcharacteristics typical of invasive species and environmentalweeds. By introducing novel genetic and endophyte variation,pasture taxa are imbued with additional capacity for invasion andenvironmental impact. New strategies to prevent future problemsare urgently needed. We highlight opportunities for researchers,agribusiness, and consumers to reduce environmental risks asso-ciated with new pasture taxa. We also emphasize four mainapproaches that governments could consider as they build newpolicies to limit weed risks, including (i) national lists of taxa thatare prohibited based on environmental risk; (ii) a weed risk assess-ment for all new taxa; (iii) a program to rapidly detect and controlnew taxa that invade natural areas; and (iv) the polluter-paysprinciple, so that if a taxon becomes an environmental weed, in-dustry pays for its management. There is mounting pressure toincrease livestock production. With foresight and planning,growth in agriculture can be achieved sustainably provided thatthe scope of SI expands to encompass environmental weed risks.

agriculture policy | biological invasions | environmental weed |invasive species | sustainable intensification

Livestock production is already the largest land use on earth,accounting for 30% of global land area (1). Nevertheless,

growing demand means that production must rise more than50% by 2050 (2) as global population size and per capita con-sumption increase (2–5). Responding to this demand, agribusi-ness* is developing and marketing new taxa† of forage plantsdesigned to increase pasture productivity. Through artificial se-lection and hybridization, public and private organizations aredeveloping plant taxa that are more productive and more tol-erant of disease and environmental extremes. At the same time,there is a strong campaign for sustainable intensification (SI) ofagriculture. One approach to SI is to increase production onsome lands while sparing others for conservation (5, 6). Agri-cultural intensification using new pasture taxa may thus be anefficient way to help meet rising demand and reduce some of thesocial and environmental costs of traditional agriculture (5).However, perversely, it may drive another environmental prob-lem because pasture plants can invade the native ecosystems that“land sparing” is designed to protect (7).

Environmental weeds are invasive alien plants that establish innatural areas (e.g., remnant native vegetation and conservationreserves), usually to the detriment of native species (8). Envi-ronmental weeds threaten biodiversity, compromise ecosystemfunction, and cost billions of dollars to manage each year (9–15).Many have been introduced as pasture forages (7). For example,in Australia, the introduced pasture species Andropogon gayanus(gamba grass) increases wildfire intensity fivefold, reducing car-bon stores and transforming species-rich native savannah to ex-otic-dominated grassland. Predicted to invade up to 380,000 km2

of northern Australia (16), gamba grass invasion has increasedthe cost of fire management by an order of magnitude, from lessthan AUD$2000 for each fire to as much as AUD$43,000 perfire (16). The possibility that SI may worsen problems like thesewarrants serious consideration, yet the topic remains contro-versial and the risks are not fully acknowledged (Fig. 1) (5, 17).

Significance

Governments spend billions of dollars each year managing in-vasive plant species. Many invasive plants have escaped frompastures and now degrade natural areas and transform eco-systems. New pasture taxa are promoted to help achieve sus-tainable intensification of agriculture by increasing productionwithout using more land. However, plant characteristics thatincrease production also increase invasion risk. Combined withinadequate regulation and management to establish largefeed-plant populations, new taxa will likely exacerbate prob-lems with invasive species. Livestock production accounts for30% of the world’s land area. Risks associated with invasivefeed-plants have been largely overlooked, even by studiesexplicitly critiquing the environmental risks of sustainable in-tensification. We suggest a suite of protocols to reduce theserisks in sustainable intensification of agriculture.

Author contributions: D.A.D., J.A.C., J.N.B., P.E.H., Inderjit, T.G.M., A.P., P.P., and D.M.R.designed research; D.A.D., J.A.C., J.N.B., P.E.H., Inderjit, T.G.M., A.P., P.P., and V.V. per-formed research; D.A.D. analyzed data; D.A.D., J.A.C., J.N.B., P.E.H., Inderjit, T.G.M., A.P.,P.P., D.M.R., S.R., and V.V. wrote the paper; D.A.D. led the design; J.A.C. provided majorcontributions at the design stage; D.A.D., J.A.C., J.N.B., P.E.H., Inderjit, T.G.M., A.P., P.P.,and V.V. gathered and contributed data; and S.R. contributed information about biose-curity laws in Australia.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1To whom correspondence should be addressed. Email: don.driscoll@anu.edu.au.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1409347111/-/DCSupplemental.

*Any private or government organization that develops pasture taxa for eventualcommercial deployment or that sells seeds of pasture taxa.

†Any taxon developed or marketed for pasture, including species, subspecies, varieties,cultivars, and plant–endophyte associations.

16622–16627 | PNAS | November 18, 2014 | vol. 111 | no. 46 www.pnas.org/cgi/doi/10.1073/pnas.1409347111

Here, we take a global perspective to consider whether newpasture taxa are likely to become environmental weeds (here-after “environmental weed risk”) and whether there are mech-anisms in place to limit potential risks. Although we focusspecifically on the risk of new pasture taxa becoming environ-mental weeds, we acknowledge that very similar risks, and likelysolutions, apply to other systems of production including bio-energy (18, 19), carbon sequestration (20), forestry (21), andhorticulture (21, 22). We find that increased environmental weedrisk from new pasture taxa presents a major challenge to in-creasing livestock production in a way that is consistent with SI(5). Nevertheless, there are practical solutions to reduce theserisks that can be informed by new research and extend fromgovernment regulation to responsible product development andconsumer choice (Fig. 2).

What Are the Risks?A previous track record as an environmental weed is often usedas a key indicator that an introduced taxon could become

a problem in a new area (23). Many pasture taxa currently on themarket have a track record as environmental weeds. From oursurvey of plants developed or promoted by 17 organizations ineight countries on six continents (Tables S1–S3), the majority oftaxa assessed were known to be environmental weeds somewherein the world. On average, 91% (SD, 10%) of taxa developed byagribusinesses were listed as weeds, with 141 weed taxa of 178taxa in total (excluding repeats; Tables S1 and S3). Further,many taxa were recognized as environmental weeds in thecountry where they were being actively developed and marketed,including in Australia, Canada, Chile, India, New Zealand, andthe United States (13/35 taxa; Table S2).The risk that pasture plants will invade natural areas can be

further elevated through pasture management. Taxa that area good match to local environments are deliberately selected,and pastures are managed to ensure large populations becomeestablished. Large, vigorous populations facilitate invasion be-cause stochastic extinctions are avoided (12) and masses of seedscan flow into surrounding environments, increasing propagulepressure (24–26). Although pasture management sometimesaims to minimize seed production, this aim may not always beachieved. Further, self-seeding may be a key to longer-termpasture persistence, particularly for annual species (27). Pasturesare often managed to achieve high densities of seeding plantsover large land areas, which means that high propagule pressureis likely to increase the risk of pasture taxa becoming environ-mental weeds (28).Although predicting which plants may be invasive and become

environmental weeds is difficult (12, 23), it is well establishedthat particular plant characteristics are associated with invasionand environmental impacts (29–32). Pasture breeding organ-izations actively select for characteristics that might inadvertentlylead to environmental impacts outside of pastures, includinghigher growth rates and tolerance to environmental stress (TableS2) (29). Using polyploidy or endophytes (symbiotic fungi andbacteria; Tables S1 and S2), plant breeders alter features such asgrowth, reproduction, disease resistance, and risk of seed pre-dation (33–35). Altering these characteristics has the potential tocreate forage plants that are more environmentally damaging(35, 36).Widespread establishment of enhanced pasture taxa is likely to

exacerbate the current environmental weed problem (17, 34, 35,37). New taxa may interbreed with existing weed populations,with potential to worsen environmental impacts (37–39). In-creased genetic diversity in pasture plants and other weedsenhances their capacity to invade natural areas across a broadrange of conditions (40–42). A diverse genetic base also facili-tates adaptation and subsequent invasion. For example, in a CO2enrichment experiment, Bromus madritensis (introduced forpasture in parts of the United States) (43) rapidly adapted todrier conditions, increasing its potential to spread in arid eco-systems as CO2 increases (44). Introducing new taxa of existingenvironmental weeds is therefore likely to increase impacts (35,37) and facilitate spread into areas previously unsuitable due toenvironmental limitations such as those found in mountains (45)or regions with high salinity (46, 47), soil deficiencies (42), lowrainfall (44), or low temperature (48).We acknowledge that risks will vary among regions. For ex-

ample, the two Czech companies that were sampled promoted 14native and 7 alien taxa (Table S1), none of which are regarded asinvasive in the Czech Republic. For some European countries,plant development for livestock industries may not pose a majorenvironmental weed threat due to a reliance on native pasturespecies. Nevertheless, European countries should not be com-placent. Increasing aridity in Europe motivates introduction ofC4 grasses in pastures, and several C4 species from the NewWorld (e.g., knotgrass Paspalum paspaloides) have become in-vasive in parts of the continent (49).

Fig. 1. The Undoolya Wattle Acacia undoolyana, nationally listed as vulnera-ble, standing dead in a sea of invasive pasture grasses (largely buffelgrassPennisetum ciliare). This species occurs in several small populations in a 165-km2

area of the East McDonnell Ranges in central Australia. Fires in 2013 killedmany of the trees in the N’Dahla Gorge population (pictured). Although thewattle is threatened by hot fires, the dominant role that buffelgrass playsin altering the fire regime is not mentioned on the information board. In-troduced pasture grasses are contentious in the Australian rangelands becausethey are used by the cattle industry, but also are highly invasive, fueling intensefires that kill woody plants and transform ecosystems. Photograph by D.A.D.

Driscoll et al. PNAS | November 18, 2014 | vol. 111 | no. 46 | 16623

SUST

AINABILITY

SCIENCE

Does Agribusiness Guard Against Environmental Weed Risk?Despite the risks of new pasture taxa becoming environmentalweeds, only 1 of the 17 organizations we reviewed undertookformal weed risk assessments of the taxa that it promoted [theAustralian Future Farm Industries Cooperative Research Centre(FFICRC); Table S1]. Under FFICRC policy, if the weed riskwas rated very high, promotion ceased. This policy was a world-leading advance, illustrating how agribusiness could take re-sponsibility for the products that it develops, using policy thatlinks environmental risk to development and managementchoices. Although this system may not always have preventedweeds from escaping into natural areas (SI Text), it demonstratedan important step forward and laudable industry precedence.The only other evidence that agribusiness considers the envi-ronmental weed risk of their products are weed risk assessmentsof Kochia spp. in the United States (Table S1) and weed riskresearch in an Australian organization that is conducted sepa-rately from pasture development research (Table S1). Un-fortunately, the innovative approach pioneered by the FFICRChas not received continued funding, and thus its long-term legacyin screening potential weed risks is unknown.The vast majority of agribusinesses, including government

agencies and private companies, do not manage the environ-mental weed risk of taxa they promote (Table S1). Why is thisthe case? The answer probably lies in the way costs are allocatedand assessed. Agribusiness is not accountable for environmentalcosts: it is not financially liable for environmental impacts orcontrol of pasture taxa that invade natural areas (50). Instead,the public pays to manage environmental weeds that have es-caped from pastures (7, 51). The risks for agribusiness areminimal, and therefore there is little incentive to address po-tential environmental impacts of new pasture taxa.

Although the public purse bears the cost of weed manage-ment, government-based or funded agribusinesses rarely con-sider the environmental weed risk of their products. The capacityfor cost-effective decision-making at a whole-of-governmentlevel is undermined by limited environmental policy integration(52). There is often inadequate communication among differentparts of government, and different sections have vastly differentcultures, funding sources, and motivations (53, 54). The agri-business section of government is often entirely separate fromthe section addressing environmental weed risk. Governmentagribusiness is protected from any environmental costs andtherefore has the same lack of incentive to consider environ-mental weed risk as private agribusiness. The lack of motivationfor agribusiness to consider environmental weed risk stems fromthe regulatory frameworks set by governments.

Does Government Guard Against Environmental Weed Risk?All of the countries we examined regulate entry of at least someplant species (SI Text). Weed risk assessment is a cost-effectivebiosecurity measure (55, 56) and is a component of plant reg-ulation in Australia, Canada, New Zealand, South Africa, andthe United States. These countries also maintain lists of pro-hibited species that cannot be imported. A key feature of theselists is that environmental issues are considered when evaluatingprohibited species. Chile also has a risk assessment and a list ofprohibited species, although the assessments are biased towardmaintaining agricultural productivity rather than addressingenvironmental concerns. Although a large number of pasturespecies could become environmental weeds (Table S1), thenumber prohibited from countries we surveyed ranged from0 (Czech Republic) to 22 (South Africa) (SI Text). Consequently,many environmental weeds are not prohibited from most

Fig. 2. Pathways influencing the risk that pasture taxa will invade natural areas and become environmental weeds. Currently, (A) economic models in-adequately accommodate long-term social and environmental costs. Governments impose few or no regulations on new pasture taxa despite having toprovide public funds to manage environmental weeds that were initially introduced as pasture. Most research into new taxa does not consider environmentalweed risk. With little self-regulation, agribusiness may therefore inadvertently increase the environmental weed risk. Solutions to these problems (B) includecloser interaction and feedback among researchers, government, and industry, government initiatives to promote low-risk pasture development, andindustry-led certification enabling consumers to reward environmentally responsible pasture development. *See Fig. 3 regarding protocols for weed riskassessment.

16624 | www.pnas.org/cgi/doi/10.1073/pnas.1409347111 Driscoll et al.

countries that we surveyed, although most countries assess theweed risk of new species proposed for import. At a subspecieslevel, there are few barriers to importing new taxa of permittedpasture species (SI Text and Table S2).With the exception of a voluntary system in Canada, new taxa

resulting from the development of species already present ina country are not subject to risk assessment (SI Text). Plantdevelopers in Canada may seek risk assessments for subspecifictaxa with novel traits, including traits arising through conven-tional breeding. However, relying on proponents to self-nomi-nate likely reduces the effectiveness of this legislation. To date,no conventionally bred taxa have been nominated. Notwith-standing Canada’s progressive approach, the paucity of regula-tion surrounding new plant taxa poses serious biosecurity risks(17, 34, 35, 40).Governments might be justified in not regulating most pasture

taxa if the benefits substantially outweigh the costs (57), butcurrent methods of cost–benefit analysis face major challenges.Economic assessments have had limited value for developingnational strategies for invasive species because the assessmentstend to be poorly implemented, focus on postinvasion effects,and do not address prevention or uncertainty (58). Economicassessments are also challenged because nonmarket goods likehabitat loss and ecosystem transformation can be hard to valuein monetary terms (59). In addition, it is unclear what discountrate (used to transform future costs or benefits into presentvalue) should be applied to environmental values (55, 59).However, combined with the lag time in environmental impacts(60), the values set for the discount rate can determine theoutcome of economic assessments (51). Resolving these uncer-tainties is critical before cost–benefit approaches can effectivelysupport decisions about planned introductions. New approachesto decision-making that can equitably accommodate environ-mental, social, and economic values are needed. In this respect,multicriteria decision analyses have the potential to be importanttools in the future (61–63).

What Needs to Be Done?Opportunities for Researchers. We identify three research needsthat are a priority for helping to prevent future weed problems.First, further development of methods for biosecurity risk anal-ysis is a priority because these can help guide better policydecisions immediately. Approaches such as multicriteria meth-ods can accommodate long-term, indirect, and off-site environ-mental and social costs, enabling new plant taxa to be fairly andaccurately assessed (Fig. 2) (51, 64). Accurate assessment alsodepends on accurate information. Therefore, a second priority isto develop and routinely apply improved methods for assessingenvironmental weed risks, such as tiered approaches that usefield and glasshouse trials, species distribution modeling, andweed risk assessment questionnaires (37, 65).A third research priority is to identify plant characteristics that

distinguish between the naturalization and spread impact stagesof invasion (23, 36). This distinction is crucial because pastor-alists desire feed-plants that form self-sustaining populations(i.e., naturalization), whereas limiting the spread and impact oftaxa will reduce environmental costs. Improved understanding istherefore needed of how traits associated with increased survival,growth and dispersal influence naturalization, spread, and im-pact (23, 36). Learning how those traits interact with ecosystemcharacteristics such as soil nutrients, disturbance, or herbivory isalso important (66–68). These areas of research will have themost impact if undertaken in conjunction with pasture devel-opers (Fig. 2).

Opportunities for Governments. Besides addressing the funda-mental drivers of the problem (Fig. 2), governments could buildon examples from around the world to regulate and discourage

environmental weed risks. There are four components to acomprehensive regulatory framework that might be considered(Fig. 3) (18, 19): (i) a prohibited list of taxa that considersimpacts on the environment and society; (ii) weed risk assess-ment for new taxa, considering evidence of past impacts in

Fig. 3. Proposed regulatory framework to reduce the risk that pasture taxawill invade natural areas (including new species, subspecies, varieties, culti-vars, and plant–endophyte combinations). The framework includes four keycomponents (shaded boxes): (i) a list of prohibited taxa; (ii) a weed risk as-sessment; (iii) postrelease early detection monitoring with the capacity forrapid control if the taxon becomes a weed; and (iv) a polluter-pays system topay for control or eradication of taxa that become environmental weeds.The protocol illustrates how weed history and characteristics related to weedrisk can be used to inform decisions to exclude taxa. There are well-estab-lished and tested criteria for undertaking weed risk assessments (70, 74, 78).Weed risk assessment provides a basis for determining which taxa are ac-cepted for field trials. In our simplified example of the weed risk assessmentstage, taxa that have no record of invading natural areas, have no newcharacteristics that are associated with environmental weeds, or have char-acteristics that would limit the risk of becoming an environmental weed canproceed to field trials. If field trials reveal no evidence that natural areasadjacent to experimental fields are invaded, taxa proceed to the releasestage. However, continued monitoring is essential because field trials do notalways identify the invasive capacity of a taxon (23). If economic assessmentsfully account for long-term environmental, social, and other costs, agri-business has the option to pay those costs so that they can continue usingcommercially valuable but environmentally damaging pasture plants (curveddashed line). Assessments to exclude a taxon may lead to revision of the listof prohibited taxa (straight dashed line). New research continually informsthe weed risk assessment, including improved assessments of the charac-teristics linked to weed risk. Where benefits are smaller than the cost of fieldtrials or the cost of an early detection and control program, development ofthat taxon would cease.

Driscoll et al. PNAS | November 18, 2014 | vol. 111 | no. 46 | 16625

SUST

AINABILITY

SCIENCE

natural areas, risks associated with particular plant character-istics (updated regularly as new research is completed), and ev-idence from experimental trials (37, 69, 70); (iii) a strongpostrelease monitoring program for early detection and rapidcontrol (necessary because it remains difficult to accuratelypredict which species will become environmental weeds despiteresearch gains in this area; 23, 71, 72); and (iv) if a taxonbecomes an environmental weed, industry pays for its manage-ment, with costs allocated to industry organizations, governmentand private plant breeders, seed companies, and farmers relativeto their culpability and ability to pay (Fig. 3). Insurance or en-vironmental bonds may be practical mechanisms for linking therisk of environmental impacts with commercial responsibility(50, 51), following the well-established polluter-pays principle(64, 73). This fourth component will likely motivate agribusinessto pursue strategies that reduce environmental risk.Almost all government regulation currently occurs at the spe-

cies level and is applied at national borders (SI Text). A keyregulatory challenge is to develop policy mechanisms to alsoregulate taxa below the species level (subspecies, varieties, culti-vars, plant–endophyte combinations) and to develop postborderregulation. The Canadian approach to regulating novel taxa servesas a useful starting point to build effective pre- and postborderregulation of pasture taxa below the species level of classification.Another role for government is to determine the spatial scale

over which weed risk assessments should be made and to whichthe outcomes apply. Methods for weed risk assessment can beadjusted to suit local flora and climates (74). Such adjustmentsprovide the scope for new taxa to be permitted in regions with lowenvironmental weed risk but excluded from other regions withinthe same country where the risk is high. However, local release ofplants with the potential to become weeds in other places facesmajor challenges (SI Text). First, changing climates may alter thegeographic location of high-risk areas (75). Second, increases inextreme weather (76), trade, and human movement (77) meanthat the chances of transportation outside of the target productionregion is likely to grow. It will be very difficult to contain taxawithin specific regions within a country or continent in the absenceof major biogeographic barriers to prevent natural spread andstrict quarantine enforcement to prevent human-assisted spread.One implication is that all governments, at all levels, will need tocooperate to ensure that taxa permitted in one jurisdiction do notspread to other jurisdictions where they are prohibited.

Opportunities for Agribusiness. Agribusiness could make theirproducts safer by integrating weed risk assessment with de-velopment of new taxa and by only releasing taxa where the riskof invading natural areas is low. Low invasion risk could beachieved by developing species with a track record of naturali-zation but not spread or impact, by using taxa native to the area,and by breeding for characteristics that will limit environmentalweed risk (18). Identifying characteristics that limit weed riskrequires research investment. Industry organizations (e.g., In-ternational Seed Federation) could develop a certificationscheme for plant taxa with low environmental weed risk. Com-panies that achieve industry certification could be rewarded witha market advantage via product eco-labeling.

Opportunities for Farmers and Consumers. Farmers could championplants that are not environmental weeds by raising awarenessand through their purchasing choices, such as buying seeds oftaxa with low risk of becoming environmental weeds. Farmersand other land managers may contribute to early detection andrapid response programs through industry-funded and govern-ment-regulated land management agreements. Consumers couldalso contribute to supporting sustainable intensification of agri-culture if eco-labeling extended to animal products from farmsthat use pastures with low environmental weed risk.Effective communication among these stakeholders will be

essential for addressing the feed-or-weed challenge; there areopportunities to build on existing protocols for achieving this(17). Reducing the risk of further invasions of natural areas bypasture taxa is important to avoid escalating costs of weed con-trol and minimize future environmental impacts. To claim sus-tainability, the scope of sustainable intensification must expandto include potential environmental weed risk.

MethodsIn eight countries located across six continents (Australia, Canada, Chile,Czech Republic, India, New Zealand, South Africa, and the United States), weassessed biosecurity measures and pasture plant development and sales.Countries were selected to represent each continent and where (i) thepasture industry was well organized; (ii) biosecurity was a serious consid-eration of government; and (iii) the environmental weed flora was wellassessed (SI Text). Our selected countries include temperate and tropicalclimates. Compared with nonsurveyed countries, our sample spans a widerange of meat production rates and proportion of land area under grazing(Fig. S1). In each country, two of the largest private or public organizationsinvolved in pasture development or sales were identified. Three companieswere used in Chile because one company specializes in a single species. Morethan two thirds of these organizations have international sales and pur-chasing links, emphasizing that our results have implications beyond theeight countries that we surveyed (Table S1). Of the taxa developed or soldfor pasture by each organization, three were randomly selected usinga random number generator. If available, information about the charac-teristics of each taxon and biosecurity measures was gathered from theorganization’s website (accessed November 2013), the most recent annualreport, publications, and discussion with senior members of each organiza-tion. Publications were searched for using Google Scholar with the searchterms: organization name (“invasive” OR “biosecurity” OR “weed”) and theplant name. We gathered details about government biosecurity measures,with a focus on the extent to which new taxa are regulated, from govern-ment websites and by discussion with staff in agencies that manage nationalbiosecurity in each country.

ACKNOWLEDGMENTS. We thank Yvonne Buckley, Mark Burgman, TonyGrice, Laura Meyerson, Hugh Possingham, John Scott, Andy Sheppard, JohnR. Wilson, and two anonymous reviewers for comments on earlier drafts. Wethank, anonymously, numerous people who provided helpful discussionsand feedback on sections of the manuscript from an agribusiness andquarantine perspective. P.P. was supported by Project RVO 67985939 and aPraemium Academiae award from the Academy of Sciences of the CzechRepublic and institutional resources of the Ministry of Education, Youthand Sports of the Czech Republic. J.A.C. was supported by the AustralianResearch Council (Grant DE120102221). A.P. was funded by IniciativaCientífica Milenio P05-002 and Comisión Nacional de Investigación Científicay Tecnológica Grant PFB-23. D.M.R. acknowledges funding from the NationalResearch Foundation (Grant 85417).

1. Steinfeld H, et al. (2006) Livestock’s Long Shadow. Environmental Issues and Options

(Food and Agriculture Organisation of the United Nations, Rome).2. Alexandratos N, Bruinsma J (2012) World Agriculture Towards 2030/2050: The 2012

Revision (Food and Agriculture Organization of the United Nations, Rome).3. Godfray HCJ, et al. (2010) Food security: The challenge of feeding 9 billion people.

Science 327(5967):812–818.4. Foley JA, et al. (2011) Solutions for a cultivated planet. Nature 478(7369):337–342.5. Herrero M, Thornton PK (2013) Livestock and global change: Emerging issues for

sustainable food systems. Proc Natl Acad Sci USA 110(52):20878–20881.6. Garnett T, et al. (2013) Agriculture. Sustainable intensification in agriculture: Premises

and policies. Science 341(6141):33–34.

7. Lonsdale WM (1994) Inviting trouble: Introduced pasture species in northern Australia.

Aust J Ecol 19(3):345–354.8. Richardson DM, et al. (2000) Naturalization and invasion of alien plants: Concepts and

definitions. Divers Distrib 6(2):93–107.9. Silvério DV, et al. (2013) Testing the Amazon savannization hypothesis: Fire effects on

invasion of a neotropical forest by native cerrado and exotic pasture grasses. Philos

Trans R Soc Lond B Biol Sci 368(1619):20120427.10. Simberloff D, et al. (2013) Impacts of biological invasions: What’s what and the way

forward. Trends Ecol Evol 28(1):58–66.11. Gilbert B, Levine JM (2013) Plant invasions and extinction debts. Proc Natl Acad Sci

USA 110(5):1744–1749.

16626 | www.pnas.org/cgi/doi/10.1073/pnas.1409347111 Driscoll et al.

12. Mack RN, et al. (2000) Biotic invasions: Causes, epidemiology, global consequences,and control. Ecol Appl 10(3):689–710.

13. DiTomaso JM (2000) Invasive weeds in rangelands: Species, impacts, and manage-ment. Weed Sci 48(2):255–265.

14. Pimentel D, Zuniga R, Morrison D (2005) Update on the environmental and economiccosts associated with alien-invasive species in the United States. Ecol Econ 52(3):273–288.

15. Sinden J, et al. (2005) The economic impact of weeds in Australia. Plant Prot Q 20(1):25–32.

16. Setterfield SA, et al. (2013) Adding fuel to the fire: The impacts of non-native grassinvasion on fire management at a regional scale. PLoS ONE 8(5):e59144.

17. Grice AC, Friedel MH, Marshall NA, Van Klinken RD (2012) Tackling contentious in-vasive plant species: a case study of buffel grass in Australia. Environ Manage 49(2):285–294.

18. Quinn LD, Barney JN, McCubbins JSN, Endres AB (2013) Navigating the “noxious” and“invasive” regulatory landscape: Suggestions for improved regulation. Bioscience63(2):124–131.

19. Davis AS, et al. (2010) Screening bioenergy feedstock crops to mitigate invasion risk.Front Ecol Environ 8(10):533–539.

20. Lindenmayer DB, et al. (2012) Avoiding bio-perversity from carbon sequestrationsolutions. Conservation Letters 5(1):28–36.

21. Richardson DM, Rejmanek M (2011) Trees and shrubs as invasive alien species: Aglobal review. Divers Distrib 17(5):788–809.

22. Bradley BA, et al. (2012) Global change, global trade, and the next wave of plantinvasions. Front Ecol Environ 10(1):20–28.

23. Hulme PE (2012) Weed risk assessment: A way forward or a waste of time? J Appl Ecol49(1):10–19.

24. Warren RJ, Ursell T, Keiser AD, Bradford MA (2013) Habitat, dispersal and propagulepressure control exotic plant infilling within an invaded range. Ecosphere 4(2):art26.

25. Kempel A, Chrobock T, Fischer M, Rohr RP, van Kleunen M (2013) Determinants ofplant establishment success in a multispecies introduction experiment with native andalien species. Proc Natl Acad Sci USA 110(31):12727–12732.

26. Nunez MA, Moretti A, Simberloff D (2011) Propagule pressure hypothesis not sup-ported by an 80-year experiment on woody species invasion. Oikos 120(9):1311–1316.

27. Neto AB, et al. (2014) Italian ryegrass establishment by self-seeding in integrated croplivestock systems: Effects of grazing management and crop rotation strategies. Eur JAgron 57:77–83.

28. Fensham RJ, Donald S, Dwyer JM (2013) Propagule pressure, not fire or cattle grazing,promotes invasion of buffel grass Cenchrus ciliaris. J Appl Ecol 50(1):138–146.

29. van Kleunen M, Weber E, Fischer M (2010) A meta-analysis of trait differences be-tween invasive and non-invasive plant species. Ecol Lett 13(2):235–245.

30. Keller RP, Kocev D, Dzeroski S (2011) Trait-based risk assessment for invasive species:High performance across diverse taxonomic groups, geographic ranges and machinelearning/statistical tools. Divers Distrib 17(3):451–461.

31. Marco A, Lavergne S, Dutoit T, Bertaudiere-Montes V (2010) From the backyard to thebackcountry: How ecological and biological traits explain the escape of garden plantsinto Mediterranean old fields. Biol Invasions 12(4):761–779.

32. Martin LJ, Murray BR (2011) A predictive framework and review of the ecologicalimpacts of exotic plant invasions on reptiles and amphibians. Biol Rev Camb Philos Soc86(2):407–419.

33. te Beest M, et al. (2012) The more the better? The role of polyploidy in facilitatingplant invasions. Ann Bot (Lond) 109(1):19–45.

34. Uchitel A, Omacini M, Chaneton EJ (2011) Inherited fungal symbionts enhance es-tablishment of an invasive annual grass across successional habitats.Oecologia 165(2):465–475.

35. Rout ME, et al. (2013) Bacterial endophytes enhance competition by invasive plants.Am J Bot 100(9):1726–1737.

36. Catford JA, Jansson R, Nilsson C (2009) Reducing redundancy in invasion ecology byintegrating hypotheses into a single theoretical framework. Divers Distrib 15(1):22–40.

37. Godfree RC, Woods MJ, Young AG (2009) Do virus-resistant plants pose a threat tonon-target ecosystems? II. Risk assessment of an Australian pathosystem using multi-scale field experiments. Austral Ecol 34(5):525–544.

38. Ellstrand NC, Prentice HC, Hancock JF (1999) Gene flow and introgression from do-mesticated plants into their wild relatives. Annu Rev Ecol Syst 30:539–563.

39. Lu BR, Snow AA (2005) Gene flow from genetically modified rice and its environ-mental consequences. Bioscience 55(8):669–678.

40. Lavergne S, Molofsky J (2007) Increased genetic variation and evolutionary potentialdrive the success of an invasive grass. Proc Natl Acad Sci USA 104(10):3883–3888.

41. Matesanz S, Horgan-Kobelski T, Sultan S (2014) Contrasting levels of evolutionarypotential in populations of the invasive plant Polygonum cespitosum. Biol Invasions16(2):455–468.

42. Rice K, Gerlach J, Jr, Dyer A, McKay J (2013) Evolutionary ecology along invasionfronts of the annual grass Aegilops triuncialis. Biol Invasions 15(11):2531–2545.

43. Reid CR, Goodrich S, Bowns JE (2008) Cheatgrass and red brome: History and biologyof two invaders. Proceedings–Shrublands Under Fire: Disturbance and Recovery ina Changing World, eds Kitchen SG, Pendleton RL, Monaco TA, Vernon J (US De-partment of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Col-lins, CO), pp 27–32.

44. Grossman JD, Rice KJ (2014) Contemporary evolution of an invasive grass in responseto elevated atmospheric CO(2) at a Mojave Desert FACE site. Ecol Lett 17(6):710–716.

45. Pauchard A, et al. (2009) Ain’t no mountain high enough: Plant invasions reachingnew elevations. Front Ecol Environ 7(9):479–486.

46. Grattan SR, et al. (2004) Evaluation of salt-tolerant forages for sequential water reusesystems: I. Biomass production. Agric Water Manage 70(2):109–120.

47. Mesleard F, Ham LT, Boy V, Vanwijck C, Grillas P (1993) Competition between anintroduced and an indigenous species - the case of Paspalum paspalodes (Michx)Schribner and Aeluropus littoralis (Gouan) in the Camargue (southern France).Oecologia 94(2):204–209.

48. Li H, Qiang S, Qian YL (2008) Physiological response of different croftonweed (Eu-patorium adenophorum) populations to low temperature. Weed Sci 56(2):196–202.

49. DAISIE (2009) A Handbook of Alien Species in Europe (Springer, Berlin).50. Perrings C, Dehnen-Schmutz K, Touza J, Williamson M (2005) How to manage

biological invasions under globalization. Trends Ecol Evol 20(5):212–215.51. Yokomizo H, Possingham HP, Hulme PE, Grice AC, Buckley YM (2012) Cost-benefit

analysis for intentional plant introductions under uncertainty. Biol Invasions 14(4):839–849.

52. Ross A, Dovers S (2008) Making the harder yards: Environmental policy integration inAustralia. Aust. J. Pub. Admin. 67(3):245–260.

53. Bundred S (2006) Solutions to silos: Joining up knowledge. Public Money Manage26(2):125–130.

54. Howes M, et al. (2014) Towards networked governance: Improving interagencycommunication and collaboration for disaster risk management and climate changeadaptation in Australia. J Environ Plann Manage, 10.1080/09640568.2014.891974.

55. Keller RP, Lodge DM, Finnoff DC (2007) Risk assessment for invasive species producesnet bioeconomic benefits. Proc Natl Acad Sci USA 104(1):203–207.

56. Springborn M, Romagosa CM, Keller RP (2011) The value of nonindigenous speciesrisk assessment in international trade. Ecol Econ 70(11):2145–2153.

57. Witt ABR (2010) Biofuels and invasive species from an African perspective: A review.GCB Bioenergy 2(6):321–329.

58. Born W, Rauschmayer F, Brauer I (2005) Economic evaluation of biological invasions -a survey. Ecol Econ 55(3):321–336.

59. Hanley N, Spash CL (1993) Cost-Benefit Analysis and the Environment (Edward ElgarPublishing Ltd., Cheltenham, UK).

60. Essl F, et al. (2011) Socioeconomic legacy yields an invasion debt. Proc Natl Acad SciUSA 108(1):203–207.

61. Cook DC, Fraser RW, Waage JK, Thomas MB (2011) Prioritising biosecurity investmentbetween agricultural and environmental systems. J Consum Prot Food Saf 6:3–13.

62. Davies AL, Bryce R, Redpath SM (2013) Use of multicriteria decision analysis to addressconservation conflicts. Conserv Biol 27(5):936–944.

63. Grechi I, et al. (2014) A decision framework for management of conflicting pro-duction and biodiversity goals for a commercially valuable invasive species. Agric Syst125(0):1–11.

64. Meyerson LA, Mooney HA (2007) Invasive alien species in an era of globalization.Front Ecol Environ 5(4):199–208.

65. Barney JN (2014) Bioenergy and invasive plants: Quantifying and mitigating futurerisks. Invasive Plant Science and Management 7(2):199–209.

66. Maron JL, Vila M (2001) When do herbivores affect plant invasion? Evidence for thenatural enemies and biotic resistance hypotheses. Oikos 95(3):361–373.

67. Houseman GR, Foster BL, Brassil CE (2014) Propagule pressure-invasibility relation-ships: Testing the influence of soil fertility and disturbance with Lespedeza cuneata.Oecologia 174(2):511–520.

68. Eskelinen A, Harrison S (2014) Exotic plant invasions under enhanced rainfall areconstrained by soil nutrients and competition. Ecology 95(3):682–692.

69. Leung B, et al. (2012) TEASIng apart alien species risk assessments: A framework forbest practices. Ecol Lett 15(12):1475–1493.

70. Gordon DR, Onderdonk DA, Fox AM, Stocker RK (2008) Consistent accuracy of theAustralian weed risk assessment system across varied geographies. Divers Distrib14(2):234–242.

71. Strayer DL (2012) Eight questions about invasions and ecosystem functioning. EcolLett 15(10):1199–1210.

72. Cousens R (2008) Risk assessment of potential biofuel species: An application for trait-based models for predicting weediness? Weed Sci 56(6):873–882.

73. Tobey JA, Smets H (1996) The Polluter-Pays Principle in the context of agriculture andthe environment. World Econ 19(1):63–87.

74. Dawson W, Burslem DFRP, Hulme PE (2009) The suitability of weed risk assessment asa conservation tool to identify invasive plant threats in East African rainforests. BiolConserv 142(5):1018–1024.

75. Crossman ND, Bryan BA, Cooke DA (2011) An invasive plant and climate changethreat index for weed risk management: Integrating habitat distribution pattern anddispersal process. Ecol Indic 11(1):183–198.

76. Bhattarai GP, Cronin JT (2014) Hurricane activity and the large-scale pattern of spreadof an invasive plant species. PLoS ONE 9(5):e98478–e98478.

77. Hulme PE, et al. (2008) Grasping at the routes of biological invasions: A framework forintegrating pathways into policy. J Appl Ecol 45(2):403–414.

78. Stone LM, Byrne M, Virtue JG (2008) An environmental weed risk assessment modelfor Australian forage improvement programs. Aust J Exp Agric 48(4):568–574.

Driscoll et al. PNAS | November 18, 2014 | vol. 111 | no. 46 | 16627

SUST

AINABILITY

SCIENCE

Supporting InformationDriscoll et al. 10.1073/pnas.1409347111SI Text

FFICRCThe FFICRC pioneered integration of environmental weed riskassessment with agricultural plant development and promotion(1). In one case when the system was applied, perennial veldtgrass (Ehrharta calycina) was rated as a very high risk of be-coming an environmental weed. The rating discouraged furtherpromotion of this species within the organization. In a secondcase, tall wheatgrass (Thinopyrum ponticum) was rated as a veryhigh risk in one Australian state (Victoria), but as high risk inthree other southern states. If the benefits are considered tooutweigh the risks, high-risk species may continue to be de-veloped and promoted by the FFICRC. Management guide-lines to minimize risk of escape were provided for this species.However, the weed risk assessment differentiates among statesonly by the extent of the area of native vegetation that thetaxon could invade. The risk of environmental impacts insuitable environments is the same. Thus, release of tallwheatgrass in the high-risk states of New South Wales andSouth Australia could facilitate invasion of the adjacent state ofVictoria where the weed risk is very high. Although integratingweed risk assessment with pasture development is an importantadvance, the detail of how weed risk is assessed (2), the scale towhich it is applied, and the industry and government* responseto those assessments, strongly influences whether environ-mental impacts are likely to be prevented. The FFICRC ceasedoperations in June 2014.

Does Government Guard Against Environmental Weed Risk?All countries surveyed have lists of excluded species and in mostcases include a small number of pasture species (South Africa, 22;Australia, 19; New Zealand, 13; United States, 12; India, 6; Chile,3; Canada 1; Czech Republic, 0). Below we provide details of thelegal frameworks that contribute to regulating pasture plants ineach country.

Australia.Law/regulation. Quarantine Act, 1908 (Cth)Environment Protection and Biodiversity Conservation (EPBC)

Act 1999 (Cth)Australian Weeds StrategyAustralia’s Biodiversity Conservation Strategy 2010–2030Weeds of National SignificanceStandards Australia HB 294:2006 National Post-Border Weed

Risk Management Protocol, CRC Australian Weed Management,Adelaide, and Standards Australia International Ltd., Sydney(a nonbinding guideline)Department responsible. Department of Agriculture, Forestry andFisheriesDepartment of the EnvironmentIndividual State Agencies

Extent to which new taxa are regulated. Quarantine Act 1908 (Cth),Proclamation 1998 (as amended), Schedule 5 lists permittedspecies (which includes all subspecific taxa of each species). TheQuarantine Act 1908 (Cth), Proclamation 1998 (as amended),Schedule 6 lists species that cannot be imported, including two

pasture species and six pasture genera.† Species within three ofthese genera are also listed on Schedule 5, limiting exclusions tosubsets of the genera. Considering the range of potential pasturespecies,† the number of excluded pasture species is 19. Speciesthat are not on the permitted or excluded list are subject toa weed risk assessment.Federal, state, and territory policies are reactive and not

preventative; they can require eradication once a species is listedas a threatening process, under the EPBC Act for example; orimpose obligations on land-mangers to eradicate and controlweeds, as occurs under weed legislation such as the NoxiousWeeds Act 1993 (NSW). Notably, once the impacts of a speciesare listed as a threatening process, no further action is takenunless the relevant minister thinks that adopting a threatabatement plan is a feasible and cost-effective means of abatingthe threat; see for example, sections 74 and 83 of the ThreatenedSpecies Conservation Act 1995, (NSW) and section 270A of theEPBC Act 1999 (Cth). The absence of remedial action despitelisting is exemplified by the recent listing in Victoria of invasion bytall wheatgrass (Thinopyrum ponticum) as a potentially threat-ening process. This species continues to be promoted by theVictorian government for use in agriculture.

Canada.Law/regulation. International Plant Protection ConventionOffice International des EpizootiesInternational Maritime OrganizationConvention on Biological DiversityWorld Trade Organization Agreement on Sanitary and Phy-

tosanitary MeasuresHealth of Animals ActCanadian Environmental Protection ActEnvironmental Assessment ActSeeds ActPest Control Products ActForestry ActNatural Resources ActTransportation of Dangerous Goods ActOceans ActFisheries ActCanada Wildlife ActWild Animal and Plant ProtectionRegulations of International and Interprovincial Trade ActCanada National Parks ActPlant Protection Act

Department responsible. Ministry of Agriculture and Agri-FoodMinistry of Fisheries and OceansMinistry of Natural ResourcesMinistry of the Environment

Extent to which new taxa are regulated. A permit from the CanadianFood Inspection Agency is required to import plant material intoCanada, and this includes a weed risk assessment that considersweed risk alongside agricultural pest risks.New taxa, imported or developed within Canada, must be

referred to the Safety Assessment Process if they meet the def-inition of plants with novel traits (these must be declared by theimporter/developer). This assessment includes consideration of

*Unrelated to the FFICRC, the Victorian Government continues to promote the use of tallwheatgrass, despite the species being listed as a potentially threatening process in thatstate. Although the weed risk is legally recognized, the feed paradigm dominatesregulation.

†To ensure a consistent method for identifying pasture species on lists of prohibitedspecies, we used the www.feedipedia.org database, accessed 22 January 2014. At thattime, 599 feed species were listed. We included all species that are used as fodder forlivestock, including fodder trees and fodder crops in that database.

Driscoll et al. www.pnas.org/cgi/content/short/1409347111 1 of 11

invasion risk. To date, approved plants with novel traits weregenetically modified and no conventionally bred pasture taxa havebeen nominated.Canada maintains a list of species that are not permitted

(pests), which includes one pasture species.† There is also a list ofProhibited Noxious Weed Seeds in Canada under the WeedSeeds Order (2005), with no pasture species listed. Plants may beimported from the United States without a permit, provided theyare not on the excluded lists and are not plants with novel traits.

Chile.Law/regulation. Application of Sanitary and Phytosanitary (ASPS)regulations, as dictated by the World Trade Organization (WTO)The ASPS refers to the International Plant Protection Con-

vention (IPPC)Department responsible. Service of Agriculture and Livestock(SAG), Ministry of AgricultureExtent to which new taxa are regulated. The main aim of IPPC is toprotect wild and cultivated plants by preventing the introductionand spread of pests and to minimize interference with the in-ternational movement of goods and people. Chile follows IPPCInternational Standards for Phytosanitary Measures (ISPM),among which the ISPM No. 11 stands out because it provides for“Pest Risk Analysis (PRA) for quarantine pests, environmentalrisks, and living modified organisms” and ISPM No. 5, whichincludes a glossary of phytosanitary terms and terminology of theConvention on Biological Diversity (CBD).New species introduced to Chile need to undergo a modified

version of the Australian weed risk assessment. However, mosteffort is directed to detect agricultural weeds. Environmentalconcerns are just recently being considered. New taxa need to bedeclared and registered, but no risk assessment is required.Chile has a list of quarantine pests, which includes 3 pasture

species† of 22 excluded plant species. New regulations are beingdiscussed for environmental weeds under the new Ministry of theEnvironment.

Czech Republic.Law/regulation. EU: CBD Article 8h, Bern ConventionBird Directive 79/409/EHSHabitat Directive Sm�ernice 92/43/EECAct no. 114/1992 Sb. On protection of nature and landscapeAct no. 326/2004 Sb. on plant protectionAct no. 289/1995 Sb. § 32 odst. 1 on forest protectionAct no. 334/1992 Sb. on protection of agricultural landAct no. 99/2004 Sb. on fisheries

Department responsible. Ministry of EnvironmentMinistry of AgricultureState Phytosanitary Administration Agency

Extent to which new taxa are regulated.The Czech Republic has a list ofenvironmental weeds that are classified based on scientific criteria.A “black list” for state authorities has been prepared and may beadopted by the government soon. The black list does not includeany pasture species, whereas a gray list (permitted but warrantsclose monitoring) includes three pasture species.† Species that arenot on this list will not be subject to biosecurity consideration.

India.Law/regulation. Destructive Insects and Pest Act 1914 andamendmentsThe Plant Quarantine (Regulation of Import into India) Order

2013International Plant Protection ConventionWTO-SPS AgreementIndian Forest Act 1927Biological Diversity Act 2002

Department responsible. National Biodiversity Authority, Chennai,India

Plant Quarantine Organization of IndiaState Biodiversity BoardBiodiversity Management Committees

Extent to which new taxa are regulated.The Indian institute must signa material transfer agreement with the institution from whereseeds or plants were sourced, and then an import permit isobtained from the National Bureau of Plant Genetic Resources(NBPGR). Imported/introduced plant material passes througha quarantine process.Lists of excluded and permitted species are maintained under

the Plant Quarantine Order 2013; these focus on excluding dis-eases and weeds of crops. Schedule IV lists 13 excluded species,none of which are typical pasture plants (grasses or legumes),although 6 are sometimes regarded as forage species† (banana,cassava, taro, oak, sugarcane, and the tuberous liana Dioscoria).Although new subspecific taxa of pasture plants pass through aquarantine process, regulation does not consider environmentalrisks of the new taxa.

New Zealand.Law/regulation. The Biosecurity Act 1993Hazardous Substances and New Organisms Act 1996

Department responsible. Ministry for Primary Industries (MPI)Extent to which new taxa are regulated. The Biosecurity Act 1993reformed the laws relating to pests and other unwanted organ-isms. It was a world first. In the act, an unwanted organism isdefined as one that “is capable or potentially capable of causingunwanted harm to any natural and physical resources or humanhealth” and a restricted organism means “any organism forwhich a containment approval has been granted in accordancewith the Hazardous Substances and New Organisms Act 1996.”Part 5 of the act provides for a National Pest ManagementStrategy and Regional Pest Management Strategy.MPI’s Plants Biosecurity Index identifies approved species for

importation by their scientific name. Only species listed in thePlants Biosecurity Index with a valid seed for sowing import spec-ification can be imported. The Plants Biosecurity Index also lists 414species that are prohibited from entry into New Zealand including13 pasture species.† Importation of species not on the Plant Bio-security Index requires full risk assessment for potential weed risk.

South Africa.Law/regulation. National Environmental Management: BiodiversityAct, 2004 (Act No. 10 of 2004) (NEMBA)Conservation of Agricultural Resources Act, 1983 (Act No. 43

of 1983) (CARA)Genetically Modified Organisms Act, 1997 (Act No. 15 of

1997) (GMOA)Plant Breeders’ Rights Act, 1976 (Act No. 15 of 1976) (PBRA)Plant Improvement Act (Act No. 53 of 1976) (PIA)

Department responsible. Department of Environmental AffairsDepartment of Agriculture, Forestry, and Fisheries

Extent to which new taxa are regulated. NEMBA provides for theplacement of species under two categories: 1a species must becontrolled/removed, and 1b species require a management pro-gram to be drawn up. NEMBA provides for the placement ofspecies under four categories, all of which may not be importedinto the country, sold, spread, grown or released in the country(unless with a permit). Category 1a species must be controlled/eradicated and control must be maintained, and 1b species mustbe contained. Management programs must be developed forthese species, and landowners must implement the plan. Category2 species require a permit to be on a landowner’s property andmust be contained within the boundaries of the property. Cate-gory 3 species may not be imported into the country.CARA species are also categorized (1–3) and may only occur

within biological control reserves and/or demarcated areas.

Driscoll et al. www.pnas.org/cgi/content/short/1409347111 2 of 11

NEMBA and CARA both prohibit the import of listed species,and NEMBA requires a risk assessment for the import of anyalien species. In terms of both of these laws, it is possible that ifa pasture species was to become problematic, it could be moti-vated to be added to the NEMBA list. However, there is noprovision in these laws for new cultivars of species already in thecountry to have to undergo any sort of risk assessment.The three other acts deal with the importation of new species or

taxa into the country: The GMOA requires permits for geneticallymodified species and risk assessments for these species. Taxa withPlant Breeders’ Rights may not be imported into the countrywithout permission from the PBR owner (PBRA), but there is noneed for risk assessment. The PIA exempts a number of agri-cultural species from the need for import permits. All otherspecies require permits in terms of this law, but these merelyrequire e.g., minimum germination standards.One pasture species is excluded by CARA, and 21 pasture

species† are excluded by NEMBA.

United States.Law/regulation. Plant Quarantine Act of 1912

Federal Plant Pest Act of 1957Federal Noxious Weed Act of 1974Plant Protection Act of 2000

Department responsible. US Department of Agriculture, Animaland Plant Health Inspection Service (USDA APHIS)Extent to which new taxa are regulated. Only species on state orfederal noxious weed lists are regulated in the United States.USDA APHIS, and more specifically the Plant Protection andQuarantine program, is responsible for determining which speciesare added to the federal list, which follows a weed risk assessment.At the federal level, species for planting can be added to the NotAuthorized Pending Pest Risk Analysis list, followed by a weedrisk assessment and final judgment for listing by APHIS officials.Individual states vary greatly in their process: some have formalrisk assessment systems, but most do not. Regulatory authorityat the state level is not consistent among the 50 states (fora detailed description of noxious weed lists, see ref. 3). Thefederal noxious weeds list includes 12 pasture species,† whichcannot be imported.

1. Stone LM, Byrne M, Virtue JG (2008) An environmental weed risk assessment modelfor Australian forage improvement programs. Aust J Exp Agric 48(4):568–574.

2. Hulme PE (2012) Weed risk assessment: A way forward or a waste of time? J Appl Ecol49(1):10–19.

3. Quinn LD, Barney JN, McCubbins JSN, Endres AB (2013) Navigating the “noxious”and “invasive” regulatory landscape: Suggestions for improved regulation. Bioscience63(2):124–131.

Fig. S1. Meat production (A) and area subject to livestock grazing (B) for target countries compared with the rest of the world. To describe how livestockproduction in our surveyed countries compares with global production, we accessed the Food and Agriculture Organization of the United Nations tool,FAOSTAT (faostat3.fao.org/faostat-gateway/go/to/home/E) in July 2014. We summed production of beef, buffalo, sheep, and goats as an estimate of meatproduction. To estimate the percentage of land area subject to grazing by livestock, we subtracted the percentage of land permanently under crops from thetotal percentage under agriculture. The remaining agricultural land is therefore used for grazing or both grazing and cropping. Box plots show the median,interquartile range, and maximum and minimum values. Country names are positioned by their y axis value and offset horizontally to avoid text overlap.

Driscoll et al. www.pnas.org/cgi/content/short/1409347111 3 of 11

Table

S1.

Descriptionoftaxa

dev

eloped

andbiosecu

rity

mea

suresundertake

nbyag

ribusinessesthat

dev

elopan

d/orsellpasture

taxa

Country

Organ

ization

Descriptionoftaxa

beingdev

eloped

Descriptionofbiosecu

rity

assessmen

ts

Australia

Agribusiness1.

Future

Farm

Industries

Cooperative

Resea

rchCen

tre(FFICRC),Gove

rnmen

tan

dindustry

funded

research

andproduct

delivery

colla

boration,closedin

June20

14.Itsproducts

weremarke

tedthroughaco

mmercial

agribusinesswhichex

portsto

ove

r50

countries

andisitselfasubsidiary

ofoneoftheworld’s

lead

inggrass

seed

organ

izationswhichmarke

tsinto

almost

100co

untriesan

dhas

subsidiaries

inTh

eNetherlands,Fran

ce,United

Kingdom,

Belgium,Lu

xemburg,Po

land,Den

mark,

Italy,

Russia,USA

,Argen

tina,

Brazil,Chile

,New

Zealan

d,Australia

,Chinaan

dSo

uth

Africa.

TheFF

ICRCdev

eloped

peren

nialshrubsto

improve

anim

alproductivityin

low

andhighrainfallarea

s,also

dev

elopingtropical

andsubtropical

forages.Ta

xadev

eloped

included

cocksfoot(D

actylis

glomerata*

),tall

fescue(Sch

edonorusarundinaceu

s*),phalaris(Phalarisaq

uatica*

),pan

icgrass

(guinea

grass)(U

roch

loamax

ima*

),asalt-tolerantgrass

(Distich

lisspicata*

)an

dthelegumes

Lotusco

rniculatus*,tedera(Bituminaria

bituminosa*),messina(M

elilo

tussicu

lus*,includinganew

salt-tolerant

Rhizobium

strain),Culle

nau

stralasicu

m†an

dCulle

npallid

um

†.Th

eyalso

promoteduse

ofthesalt-toleranttaxa

old

man

saltbush

(Atriplex

nummularia*

,†)an

dtallwhea

tgrass

(Thinopyrum

ponticu

m*).

Protoco

lswereestablished

torankwee

drisk.Pa

mphlets

wereproduced

describingwee

drisk

forsomespecies.

Forve

ryhighrisk

species,FFICRCpolicy

was

toceasepromotion.Fo

rhighrisk

species,man

agem

entinform

ationwas

provided

tohelpland-holdersan

dothersto

assess

andman

agetherisk.

Australia

Agribusiness2.

Gove

rnmen

tan

dindustry

funded

research

organ

ization.Pa

sture

taxa

promotedin

thePa

cificincluding:New

Zealan

dan

dFiji;

SEAsiaincluding:Ph

ilippines,Malay

sia,

Thailand,

Indonesia;Africaincluding:So

uth

Africa,

Ken

ya,

Tanza

nia,Zimbab

we,

Zambia,M

alaw

i,Botswan

a,Moza

mbique.

Dev

eloped

nea

rlyallAustralia

nPh

alaris*va

rieties.Ph

alarisaq

uatica*

,Ph

alarisarundinacea

*,Trifoliu

msubterran

eum*

Programsthat

tackle

inva

sive

species

issues

arenotlin

kedto

pasture

dev

elopmen

twithin

theorgan

ization.

Can

ada

Agribusiness3.

Federal

gove

rnmen

torgan

ization.

New

cultivarsofalfalfa(M

edicag

osativa

*),cicermilk

vetch(A

stragalus

cicer*),orchardgrass

(Dactylis

glomerata*

),sainfoin

(Onobrych

isviciifolia

*)Noev

iden

cefound

Can

ada

Agribusiness4.

Associationfunded

byindustry.

Surinam

grass

(Uroch

loa(form

erly

Brach

iaria)

decumben

s*)introducedfrom

Africa;

new

commercial

varietiesofalfalfa(M

edicag

osativa

*)Noev

iden

cefound

Chile

Agribusiness5.

Gove

rnmen

tfunded

agen

cy.

Pasture

grasses

andlegumes.O

nenativespecies(Bromusva

ldivianus*

,†),one

exoticspecies(Trifoliu

mpratense*)

hav

ealread

ybee

npaten

tedas

Chile

anva

rieties.Also,recentwork

onTrifoliu

mrepen

s*an

dLo

lium

peren

ne*

.Prev

iouslyworked

onLo

tustenuis*an

dLo

tusulig

inosus*.

Noev

iden

cefound

Chile

Agribusiness6.

Privateco

mpan

y.Se

llsto

Argen

tina,

Boliv

ia,Brazil,Ecuad

or,Pe

ru.

Specialisingin

Lucern

(Med

icag

osativa

*).A

rangeofva

rietiesav

ailable,

imported

largelyfrom

theUSA

.New

varietiesareproducedfrom

hyb

rids

ofim

ported

taxa

with“local”

taxa

.

Noev

iden

cefound

Chile

Agribusiness7.

Privateco

mpan

y.Se

llsgrasses

andgrass

hyb

rids(Loliu

mperen

ne*

,Sched

onorus

arundinaceu

s*,Bromusva

ldivianus*,Lo

lium

multiflorum*,

Festuca

pratensis*

XLo

lium

multiflorum*,

Festuca

arundinacea

*XLo

lium

multiflorum*,

Festuca

peren

nis),Le

gumes

(Pisum

sativu

m,Vicia

atropurpurea,

Trifoliu

mpratense*,

Trifoliu

mrepen

s*,Trifoliu

msubterran

eum*,

Trifoliu

mincarnatum*),an

dBrassica(B.oleracea*

,B.nap

us*

XB.oleracea*

).Allseed

sarebroughtdirectlyfrom

Den

mark

andNew

Zealan

d.S

eedmixes

aremad

ein

Chile

.Noseed

multiplicationor

dev

elopmen

tin

Chile

.

Noev

iden

cefound

Driscoll et al. www.pnas.org/cgi/content/short/1409347111 4 of 11

Table

S1.

Cont.

Country

Organ

ization

Descriptionoftaxa

beingdev

eloped

Descriptionofbiosecu

rity

assessmen

ts

Cze

chRep

ublic

Agribusiness8.

Privateco

mpan

y.Dev

elopscu

ltivars

forgrass

andfodder

mixtures.Ex

portsto

German

y.

They

offer

15grass

andfodder

species(ofwhich12

arecu

ltivarsofnative

species).Ex

otics

includeclove

r(Trifoliu

mhyb

ridum*),Italianryeg

rass

(Loliu

mmultiflorum*),an

dbromegrasses

(Bromussitchen

sis*,

B.m

arginatus).T

rifoliu

man

dLo

lium

originatefrom

Europe.

Nativespecies

includethegrasses

Loliu

mperen

ne*

,†,Dactylis

glomerata*

,†,Fe

stuca

rubra*,

†,Cyn

osuruscristatus*

,†,Agrostisstolonifera†,Po

anem

oralis

†,

Alopecuruspratensis*

,†,Bromusinermis*,

†,an

dthelegumes

Trifoliu

mpratense*,

†,Trifoliu

mrepen

s*,†,Lo

tusco

rniculatus*

,†.

Noev

iden

cefound

Cze

chRep

ublic

Agribusiness9.

Privateco

mpan

y.Ex

portsto

German

y,Sw

itze

rlan

d,Austriaan

dSlova

kia,

Ukrainean

dFran

ce.

Theco

mpan

ypromotesfourex

oticspeciesan

dsixcu

ltivarsofnativegrasses

andlegumes.Ex

otics

areredtopben

tgrass

(Agrostisgigan

tea*

),Italian

ryeg

rass

(Loliu

mmultiflorum*),sainfoin

(Onobrych

isviciifolia

*)an

dalfalfa

(Med

icag

osativa

*).Allex

otics

exceptalfalfaoriginatefrom

Europe.

Nativespeciesincludethegrasses

Anthoxa

nthum

odoratum*,

†,Lo

lium

peren

ne*

,†,an

dthelegumes

Anthyllis

vulneraria

†,Melilo

tusalba*

,†,

Trifoliu

mrepen

s*,†an

dTrifoliu

mpratense*,

†.

Noev

iden

cefound

India

Agribusiness10

.Gove

rnmen

torgan

ization.

Resea

rchfocu

sedonim

provingseed

yieldin

arangeoflegumes

andgrasses.

Grasses

(Axo

nopusco

mpressus*,Bothrioch

loainterm

edia

†,Bothrioch

loa

pertusa

†,Brach

iariabriza

ntha*

,Brach

iariamutica*,

Cen

chruscilia

ris*,

Cen

chrussetigerus*

,†,C

hlorisgay

ana*

,Chrysopogonfulvus†,C

oix

lacrym

a*,†,

Cyn

odondactylon*,

†,Dichan

thum

annulatum*,

Digitaria

decumben

s*,†,

Diplach

nefusca*

,†,E

ragrostiscu

rvula*,

†,E

ulalio

psisbinata†,H

eteropogon

contortus,Iseilemalaxu

m†,La

siurussindicus†,Melinisminutiflora*,

Panicum

antidotale*,

Panicum

max

imum*,

Panicum

turgidum,Pa

spalum

dila

tatum*,

Paspalum

notatum*,

Pennisetum

clan

den

stinum*,

Pennisetum

ped

icellatum*,

Sehim

anerosum,Se

tariasphacelata*

,Tripsacu

mdactyloides,Uroch

loamosambicen

sis*,Vetiveria

ziza

noides*.

Legumes

(Atylosiascarab

aeoides*,

†,Calopogonium

mucu

noides*,

Cen

trosema

pubescens*,Clitoriaternatea

*,Desman

thusvirgatus*,Desmodium

intortum*,

Indigofera

hirsusta,

Lablabpurpureus*,Lo

tononisbainesii†,

Macroptiliu

matropurpureum*,

†,Macroptiliu

mlathyo

ides*,

Macrotyloma

axillares*,

Neo

ntonia

wightii*,Pu

eraria

phaseo

loides*,

Stizolobium

dee

ringianum*(M

ucu

naprurien

s),Stylosanthes

guianen

sis*,S

tylosanthes

ham

ata*

,Stylosanthes

humilis*,Stylosanthes

scab

ra*,

Vignaluteola).In

addition,20

4accessionsof10

speciesofthelegumeLe

ucaen

a(136

accessionsbelonged

toL.

leuco

cephala*

).

Noev

iden

cefound

India

Agribusiness11

.Privateco

mpan

y.A

subsidiary

of

oneoftheworld’slead

inggrass

seed

organ

izationsthat

marke

tsinto

almost

100

countries.

Foragespecies:Lo

lium

multiflorum*,

Sched

onorusarundinaceu

s*,Pa

nicum

max

imum*,

seve

ralSo

rghum*hyb

rids,Trifoliu

malex

andrinum*,

mak

han

grass

(Brach

iariaramosa*),Ta

llfescue(Festuca

arundinacea

*),Annual

ryeg

rass

(Loliu

mmultiflorum*),Pe

rennialryeg

rass

(Loliu

mperen

ne*

),Orchardgrass

(Dactylis

glomerata*

),Mea

dow

grass

(Festuca

pratensis*),

Timothy(Phleum

pratense*),Annual

brome(Bromusjaponicas*),Sm

ooth

brome(Bromusinermis*),Ken

tuckyblueg

rass

(Poapratensis*),W

hea

tgrass

(Agropyron*sp),Red

clove

r(Trifoliu

mpratense*),Whiteclove

r(Trifoliu

mrepen

s*),Chicory

(Cichorium

intybus*),So

rghum*spp,M

ulticut

bajra

(Pen

nisetum

sp.),Oats(A

venasativa

*),Lu

cerne(M

edicag

osativa

*).

Noev

iden

cefound

Driscoll et al. www.pnas.org/cgi/content/short/1409347111 5 of 11

Table

S1.

Cont.

Country

Organ

ization

Descriptionoftaxa

beingdev

eloped

Descriptionofbiosecu

rity

assessmen

ts

New

Zealan

dAgribusiness12

.Privateco

mpan

y.Resea

rches,

dev

elopsan

dmarke

tspasture

taxa

inNZan

dAustralia

.

Dev

elopsmultiple

varietiesofdiploid

andtetrap

loid

peren

nialRye

grasses

(Loliu

mperen

ne*

),an

dhyb

ridswithan

nual

ryeg

rasses

(L.multiflorum*),

threead

ditional

grasses

(cocksfoot(D

actylis

glomerata*

),Bromus

catharticu

s*,F

estuca

arundinacea

*,ribwort

plantain

Plan

tagolanceolata*,

Chicory

(Cichorium

intybus*),five

speciesofclove

r(Trifoliu

mrepen

s*,

T.pratense*,

T.frag

iferum*,

T.michelianum*,

T.resupinatum*),an

dthe

legumeMed

icag

osativa

*.Varieties

inco

mmonwithAgribusiness13

.

Noev

iden

cefound

New

Zealan

dAgribusiness13

.Privateco

mpan

y.Plan

tbreed

ing

andresearch

linke

dto

oneoftheworld’slead

ing

grass

seed

organ

izationsthat

marke

tsinto

almost

100co

untries.

Pasture

plants

includediploid,tetrap

loid

andhyb

ridLo

lium

xbouch

eanum

Kunth

(Loliu

mmultiflorum*xL.

peren

ne*

)peren

nialryeg

rass

(Loliu

mperen

ne*

),somewithen

dophytes,Italianan

dan

nual

ryeg

rasses

(Loliu

mmultiflorum*),allwithmultiple

cultivarsav

ailable.Sixother

grass

species

(Bromusva

ldivianus,Bromuscatharticu

s*,cocksfoot(D

actylis

glomerata*

),tallfescue(Sch

edonorusarundinaceu

s*),wild

oats(A

venasativa

*),

timothy(Phleum

pratense*).Nongrasses

arech

icory

(Cichorium

intybus*),

plantain

(Plantagolanceolata*),whiteclove

r(Trifoliu

mrepen

s*)an

dred

clove

r(T.pratense*).Varieties

inco

mmonwithAgribusiness12

.

Noev

iden

cefound

South

Africa

Agribusiness14

.Privateco

mpan

y,sells

seed

sdev

eloped

byaprecedingorgan

ization.B

uys

seed

from

Africa,

theAmericas,Australia

andNew

Zealan

d.Ex

portsto

>30

countries.

Preced

ingorgan

izationdev

eloped

Erag

rostiscu

rvula*,

†,Erag

rostistef*,

Sched

onorusarundinaceu

s*,Lo

lium

xbouch

eanum

(Loliu

mmultiflorum*

xL.

peren

ne*

),Lo

lium

multiflorum*,

Loliu

mperen

ne*

,Rap

han

ussativu

s*va

r.oleiform

isan

dTrifoliu

mrepen

s*.Alsosells

Ave

nasativa

*,Chloris

gay

ana*

,†,Dactylis

glomerata*

,Digitaria

eriantha*

,†,Hordeu

mvu

lgare*

,Med

icag

osativa

*,Pa

nicum

max

imum*,

†,P

aspalum

notatum*,

Pennisetum

clan

destinum*,

Pennisetum

glaucu

m*,

Sorghum*hyb

rids,Triticum

xSe

cale

hyb

rids,Trifoliu

malex

andrinum*,

Trifoliu

mpratense*.

Noev

iden

cefound

South

Africa

Agribusiness15

.Privateco

mpan

y.New

cultivarsare

dev

eloped

onaco

ntractbasiswithuniversities

or

research

centers,orbytheco

mpan

yitself.Se

llsinto

man

yAfrican

countries,withbranch

esestablishingin

Zambia,Rwan

da,

Zimbab

we,

Moza

mbique,

Angola,Ken

yaan

dTa

nza

nia.

Hav

ehad

new

cultivarsdev

eloped

forErag

rostiscu

rvula*,

†,Ornithopus

sativu

s*an

dSo

rghum

bicolor*

,†,althoughthey

sellabroad

errangeof

taxa

,includinggrasses

(Anthep

hora

pubescens†,Bromuscatharticu

s*,

Cen

chruscilia

ris*

,†,Chlorisgay

ana*

,†,Dactylis

glomerata*

,Digitaria

eriantha*

,†,Erag

rostistef*,Echinoch

loacrus-galli*

,Lo

lium

multiflorum*,

Loliu

mperen

ne*

,Pe

nnisetum

glaucu

m*,

Sorghum*hyb

rids,Pa

nicum

max

imum*,

†,Ph

alarisaq

uatica*

,Sched

onorusarundinaceu

s*)legumes

(Astragaluspelecinus,Crotalariabreviden

s,Desmodium

intortum*,

Lablab

purpureus*,Le

sped

ezajuncea,

Lotusco

rniculatus*,Lo

tusped

uncu

latus*,

Lupinusalbus*,Lu

pinusan

gustifoliu

s*,Lu

pinusluteus*,Macroptiliu

matropurpureum*,

Med

icag

olittoralis*,

Med

icag

opolymorpha*

,Med

icag

osativa

*,Med

icag

otruncatula*,

Neo

notonia

wightii*,Ornithopus

compressus*,Pisum

sativu

m,Se

curigerava

ria*

,Stylosanthes

guianen

sis*,

Trifoliu

malex

andrinum*,

Trifoliu

mfrag

iferum*,

Trifoliu

mhirtum*,

Trifoliu

mincarnatum*,

Trifoliu

mmichelianum*,

Trifoliu

mpratense*,

Trifoliu

mrepen

s*,Trifoliu

mresupinatum*,

Trifoliu

msubterran

eum*,

Trifoliu

mve

sicu

losum*,

Vignaunguiculata*,

†,Vicia

ben

ghalen

sis*,Vicia

faba*

,Vicia

sativa

*,Vicia

villo

sa*)

andother

species(A

triplexnummularia*

,Atriplexhalim

us,Atriplexsemibaccata*,

Betavu

lgaris*,

Brassicarapa*

,Brassicanap

us*,Rap

han

usraphan

istrum*,

Sanguisorbaminor*).

Noev

iden

cefound

Driscoll et al. www.pnas.org/cgi/content/short/1409347111 6 of 11

Table

S1.

Cont.

Country

Organ

ization

Descriptionoftaxa

beingdev

eloped

Descriptionofbiosecu

rity

assessmen

ts

USA

Agribusiness16

.Privatelyfunded

organ

ization.

Conductspasture

dev

elopmen

tresearch

.Has

staff

in25

countries,an

dco

llaborateswithtw

ointernational

consortia

that

sellinto

USA

,Can

ada,

Mex

ico,A

rgen

tina,

Australia

,Brazil,Finland,U

.K.,

andUruguay

.

Intheirplantbreed

ingresearch

program

they

aredev

elopingcu

ltivarsof

fodder

cropsan

dgrasses

(oat

(Ave

nasativa

*),rye(Loliu

mperen

ne*

),tall

fescue(Sch

edonorusarundinaceu

s*),orchardgrass

(Dactylis

glomerata*

),sw

itch

grass

(Pan

icum

virgatum*,

†),Cerea

lrye(Secalecereal*),Triticale

spp.an

dDigitaria*spp.),an

dthreeforagelegumes

(Trifoliu

mpratense*,

T.repen

s*,Med

icag

osativa

*).

Noev

iden

cefound

USA

Agribusiness17

.Gove

rnmen

torgan

ization.Does

notselldirectlyto

other

countries,but

colla

borateswithag

ribusinessin

China,

and

productsaresold

throughseed

consortia.

Abroad

rangeofstate-based

programs,includingwork

on:Agropyron

frag

ile*,

Dactylis

golm

erata*

,Koch

iaprostrata*,

Panicum

virgatum*,

†,

Agropyroncristatum*,

Pseu

doroeg

neria

spicata†,Pa

scopyrum

smithii*

,†,

Astragalusfilip

es†,Dalea

purpurea†,Ach

illea

millefoliu

m*,

†,Le

ymus

cinereu

s†,Nassella

viridula*,

†,Ach

natherum

hym

enoides

†,Po

asecu

nda†,

Elym

ustrachycau

lus*

,†,Elym

uswaw

awaien

sis†,Elym

uselym

oides*,

†,

Bromusbiebersteinii*

,Psathyrostachys

juncea*

.Colla

borative

research

includes

trialsan

ddev

elopmen

tof58

1speciesin

Mongolia

,with12

3seed

colle

ctionsreturned

totheUSA

foruse

inbreed

ingprograms.Work

also

includes

endophytes

intallfescue(Sch

edonorusarundinaceu

s*)an

dFe

stuca*interspecific

hyb

rids.

ThisResea

rchLa

bhas

published

two

pap

ersoninva

sionrisk

ofKoch

iaprostrata.T

hey

concludelim

ited

spread

potential

withsomeex

ceptions.

Dataarebased

onourreview

oftheweb

sites,themost

recentan

nual

report,p

ublications,an

ddiscu

ssionswithseniorstaffofea

chag

ribusiness.Agribusinesseswithinternational

salesan

dpurchasinglin

ksare

indicated

.Allag

ribusinessesaresubject

tothegen

eral

biosecu

rity

legislationap

propriateto

each

country(SITe

xt).Agribusinessesarelistedan

onym

ouslybecau

seourintentionis

notto

criticizeindividual

compan

iesbutto

highlig

httheglobal

eviden

cethat

environmen

talwee

dsarewidelydev

eloped

andpromotedforuse

inpastures,often

across

multiple

countries.

*Environmen

talwee

dslistedin

Ran

dall’s

Global

Compen

dium

ofW

eeds(w

ww.hea

r.org/gcw

/).

†Nativein

atleastpartofthisco

untry.

Driscoll et al. www.pnas.org/cgi/content/short/1409347111 7 of 11

Table

S2.

Sample

oftaxa

dev

eloped

orso

ldbyag

ribusinessesfrom

eightco

untries,highlig

htingtheplantch

aracteristicscu

rren

tlypossessedby,

orbeingdev

eloped

in,p

asture

taxa

Species*

Commonnam

eCountry

Organ

ization

(tax

onnam

e)†

Dev

eloped

characteristicsassociated

withwee

drisk

‡

Other

species

characteristics

Recognized

aswee

din

country(reg

ulatory

response)§

Global

wee

dstatus{

Agropyroncristatum

(L.)Gae

rtn.

Crested

whea

tgrass

USA

Agribusiness17

1:highnutrientco

ntent;6:

high

establishmen

t,highsurvival

No

CE

Agropyronfrag

ile(Roth)P.

Can

dargy

Siberianwhea

tgrass

USA

Agribusiness17

1:droughttolerance;6:

high

establishmen

t,highsurvival

No

Agrostisgigan

tea

Roth

Red

top

Cze

chRep

ublic

Agribusiness9

4:highgrowth

rate;5:

highyield

Palatable

No

ACEN

jj

AgrostisstoloniferaL.

Creep

ingben

tgrass

Cze

chRep

ublic

Agribusiness8

Ornam

ental

No

ACE

AlopecuruspratensisL.

Mea

dow

foxtail

Cze

chRep

ublic

Agribusiness8

noinform

ationprovided

Easier

sowing

No

ACE

AstragaluscicerL.

Cicer

Milk

vetch

Can

ada

Agribusiness3

1:co

ldtolerance;5:

highyield;6:

high

survival

Competitive

Yes

(nothing)

AC

Bromusstam

ineu

sDesv.

Road

sidebrome,

Grazing

brome

Chile

Agribusiness5(Bromus

valdivianus)

4:highgrowth

rate;5:

larger,highyield;

6:highsurvival,fecu

nd

No

E

Dactylis

glomerataL.

Orchardgrass,Ascherson’s

orchardgrass;co

cksfoot

Australia

Agribusiness1.

Future

Farm

Industries

CooperativeResea

rch

Cen

tre

1:droughttolerance,broad

edap

hic

tolerance

Yes

(man

agem

ent

inform

ationprovided

byFF

ICRC)

ACEN

Dactylis

glomerataL.

Orchardgrass,co

cksfoot

Can

ada

Agribusiness3

1:highnutrientco

ntent,co

ldtolerance;

6:fecu

nd,disea

seresistan

cePa

latable

Yes

(nothing)

ACEN

Dactylis

glomerataL.

Orchardgrass,Ascherson’s

orchardgrass;co

cksfoot

New Ze

alan

dAgribusiness12

1:droughttolerance;4:

highgrowth

rate

Palatable

Yes

(nothing)

ACEN

Dactylis

glomerataL.

Orchardgrass,Ascherson’s

orchardgrass;co

cksfoot

New Ze

alan

dAgribusiness13

5:larger,highyield

No

ACEN

Dactylis

glomerataL.

Orchardgrass,Ascherson’s

orchardgrass;Cocksfoot

USA

Agribusiness17

1:co

ldtolerance,highelev

ation

tolerance,broad

edap

hic

tolerance;5:

larger;6:

highsurvival

Yes

(nothing)

ACEN

Erag

rostiscu

rvula

(Sch

rad.)Nee

sW

eepinglove

grass

South

Africa

Agribusiness15

1:broad

edap

hic

tolerance

No

ACEN

Erag

rostistef

(Zuccag

ni)Trotter

Teff

South

Africa

Agribusiness14

1:droughttolerance,broad

edap

hic

tolerance;2:

larger

leav

es;3:

high

above

-groundbiomass;4:

highgrowth

rate;5:

larger,highyield

No

A

Leucaen

aleuco

cephala

(Lam

.)deW

it

Leucaen

a;W

hitelead

tree

India

Agribusiness10

4:highgrowth

rate;6:

highsurvival

Yes

(nothing)

ACEN

jj

Loliu

mperen

neL.

Perennialryeg

rass

New Ze

alan

dAgribusiness12

1:en

dophyteman

ipulation

No

ACE

Loliu

mperen

neL.

ssp.

multiflorum

(Lam

.)Husnot

Italianryeg

rass;Annual

ryeg

rass

Chile

Agribusiness7(Loliu

mmultiflorum)

1:en

dophyteman

ipulation,co

ldtolerance;3:

highab

ove

-ground

biomass;4:

highgrowth

rate;5:

larger;

6:highsurvival

Palatable

Yes

(nothing)

ACE

Driscoll et al. www.pnas.org/cgi/content/short/1409347111 8 of 11

Table

S2.

Cont.

Species*

Commonnam

eCountry

Organ

ization

(tax

onnam

e)†

Dev

eloped

characteristicsassociated

withwee

drisk

‡

Other

species

characteristics

Recognized

aswee

din

country(reg

ulatory

response)§

Global

wee

dstatus{

Loliu

mperen

neL.

ssp.multiflorum

(Lam

.)Husnot

Italianryeg

rass;Annual

ryeg

rass

Cze

chRep

ublic

Agribusiness9

(L.multiflorum)

4:highgrowth

rate

No

ACE

Loliu

mperen

neL.

ssp.multiflorum

(Lam

.)Husnot

Italianryeg

rass;Annual

ryeg

rass

India

Agribusiness11

(L.multiflorum)

1:highnutrientco

ntent,co

ldtolerance;

3:highab

ove

-groundbiomass;4:

high

growth

rate;5:

larger,highyield;6:

fecu

nd

No

ACE

Loliu

mperen

neL.

ssp.multiflorum

(Lam

.)Husnot

Italianryeg

rass;Annual

ryeg

rass

New Ze

alan

dAgribusiness13

(L.multiflorum)

5:larger;6:

highestablishmen

tTe

trap

loid;

palatab

leNo

ACE

Loliu

mx

bouch

eanum

Kunth

Hyb

ridrye

South

Africa

Agribusiness14

(L.multiflorum

xL.

peren

ne)

1:co

ldtolerance,en

han

ced

photosynthesis;3:

highab

ove

-ground

biomass;4:

highgrowth

rate;5:

larger,

highyield;6:

disea

seresistan

ce

No

—

Med

icag

osativa

L.Alfalfa

Can

ada

Agribusiness4

1:broad

edap

hic

tolerance;6:

high

survival;disea

seresistan

ceNo

ACE

Med

icag

osativa

L.Alfalfa

Chile

Agribusiness6

4:highgrowth

rate;5:

larger,highyield;

6:highsurvival

No

ACE

Onobrych

isviciifolia

Scop.

Sainfoin

Can

ada

Agribusiness3

1:droughttolerance,co

ldtolerance;5:

highgrowth

rate;6:

highsurvival

No

AC

Ornithopussativu

sBrot.

Commonbird’s-foot;Pink

serrad

ella

South

Africa

Agribusiness15

1:broad

edap

hic

tolerance,grazing

tolerance

No

ACE

Pennisetum

cilia

re(L.)Link

Buffelgrass

India

Agribusiness10

(Cen

chruscilia

ris)

1:droughttolerance;4

:highgrowth

rate;

6:highsurvival

No

ACEj

j

Phalarisaq

uaticaL.

Bulbouscanaryg

rass,

Hardinggrass

Australia

Agribusiness2

1:droughttolerance,co

ldtolerance;6:

highsurvival

Yes

(nothing)

ACEN

Phalarisarundinacea

L.Ree

dCan

aryg

rass

Australia

Agribusiness2

1:droughttolerance,co

ldtolerance;6:

highsurvival

Yes

(nothing)

ACEN

jj

Sched

onorus

arundinaceu

s(Sch

reb.)Dumort.,

nom.co

ns.

Tallfescue

Australia

Agribusiness1.

Future

Farm

Industries

CooperativeResea

rch

Cen

tre(Loliu

marundinaceu

m)

1:droughttolerance,broad

edap

hic

tolerance,highnutrientco

ntent;4:

highgrowth

rate;6:

highsurvival

Yes

(nothing)

ACEN

jj

Sched

onorus

arundinaceu

s(Sch

reb.)Dumort.,

nom.co

ns.

Tallfescue

Chile

Agribusiness7(Festuca

arundinacea

)1:

droughttolerance,en

dophyte

man

ipulation,co

ldtolerance,flood

tolerance;3:

highab

ove

-ground

biomass;4:

highgrowth

rate;5:

larger;

6:highsurvival

Yes

(nothing)

ACEN

Sched

onorus

arundinaceu

s(Sch

reb.)Dumort.,

nom.co

ns.

Tallfescue

South

Africa

Agribusiness14

(Festuca

arundinacea

)1:

droughttolerance;6:

highsurvival,

disea

seresistan

ceNo

ACEN

Driscoll et al. www.pnas.org/cgi/content/short/1409347111 9 of 11

Table

S2.

Cont.

Species*

Commonnam

eCountry

Organ

ization

(tax

onnam

e)†

Dev

eloped

characteristicsassociated

withwee

drisk

‡

Other

species

characteristics

Recognized

aswee

din

country(reg

ulatory

response)§

Global

wee

dstatus{

Sched

onorus

arundinaceu

s(Sch

reb.)Dumort.,

nom.co

ns.

Tallfescue

USA

Agribusiness16

1:en

dophyteman

ipulation

Yes

(nothing)

A

Secale

cerealeL.

Cerea

lrye;

Rye

USA

Agribusiness16

1:grazingtolerance;5

:highyield;6

:high

survival

Palatable

Yes

(warning

inform

ationprovided

)ACEN

Stylosanthes

ham

ata

(L.)Ta

ubert

Chee

sytoes

India

Agribusiness10

1:grazingtolerance;4:

highgrowth

rate

Tetrap

loid;

highgrowth

under

elev

ated

carbon

dioxide

No

AE

Trifoliu

mpratense

L.Red

clove

rChile

Agribusiness5

4:highgrowth

rate;5:

larger;6:

high

survival

No

ACE

Trifoliu

mrepen

sL.

Whiteclove

r;La

dinoclove

r;Dutchclove

rCze

chRep

ublic

Agribusiness8

1:nutrientuse

efficien

cyNo

ACEj

j

Trifoliu

mrepen

sL.

Whiteclove

r;La

dinoclove

r;Dutchclove

rCze

chRep

ublic

Agribusiness9

1:droughttolerance,nutrientuse

efficien

cy;4:

highgrowth

rate;5:

larger,highyield;6:

highsurvival,

disea

seresistan

ce

Competitive

No

ACEj

j

Trifoliu

mrepen

sL.

Whiteclove

r;La

dinoclove

r;Dutchclove

rNew Ze

alan

dAgribusiness13

1:droughttolerance;2:

larger

leav

es;5:

larger,highyield;6:

disea

seresistan

ceNo

ACEj

j

Trifoliu

mrepen

sL.

Whiteclove

r;La

dinoclove

r;Dutchclove

rSo

uth

Africa

Agribusiness15

1:grazingtolerance

No

ACEj

j

Trifoliu

mrepen

sL.

Whiteclove

r;La

dinoclove

r;Dutchclove

rUSA

Agribusiness16

5:larger;6:

highsurvival

Competitive

Yes

(nothing)

ACEj

j

Trifoliu

msubterran

eum

L.Su

bterran

eanclove

rAustralia

Agribusiness2

1:nutrientuse

efficien

cy;3:

highab

ove

-groundbiomass;4:

highgrowth

rate

Yes

(nothing)

AE

Thewee

dstatusofea

chtaxo

nglobally

andwithin

theco

untryitisdev

eloped

orsold

inisalso

indicated

.—,nospecies-specific

inform

ation.

*Authority:USD

APlan

tsDatab

ase(plants.usda.gov/java

/,accessed

1Decem

ber

2013

).†Ta

xonnam

eusedbyorgan

izationprovided

inbracketsifitdiffers

from

thespeciesnam

e,althoughsomeco

mmercial

nam

esex

cluded

.‡Weclassified

characteristicsthat

areassociated

withwee

drisk

based

onthemeta-an

alysisbyva

nKleunen

etal.(9):1,

physiology;

2,leaf-areaallocation;3,

shootallocation;4,

growth

rate;5,

size

;6,

fitness.

§Th

ecu

rren

tstatusofspeciesin

each

countrywas

determined

acco

rdingto

thefollo

wingcriteria:Y

,leg

ally

reco

gnized

asawee

drisk

(appea

rsonawee

dlistthat

carrieslegislative

forcefortheco

untryorat

least

onejurisdictionwithin

that

country);y

,recognized

asan

environmen

talw

eedrisk

inpublished

material(scientificliterature,floras,wee

dlists

curatedbyplantsocieties,en

vironmen

talm

anag

emen

tag

encies

and

experts)in

theco

untryoraregionwithin

that

country,

butnotlegally

reco

gnized

;n,noev

iden

cethat

itisan

environmen

talwee

drisk

(thetaxo

nisab

sentfrom

thereference

materialab

ove

orspecifically

reco

rded

asnotinva

dingnativeecosystem

sin

themateriallistedab

ove

).Th

eregulatory

response

(bygove

rnmen

tunless

otherwiseindicated

)associated

withspecieslistedorkn

ownto

been

vironmen

talw

eed

risksisin

brackets.

{ Thestatusofea

chspeciesglobally

was

based

onRan

dall’s

Global

Compen

dium

ofW

eeds(w

ww.hea

r.org/gcw

/).A,ag

ricu

lturalwee

d;C,cu

ltivationescape;

E,en

vironmen

talwee

d;N,noxiouswee

d.

jj Specieslistedin

theGlobal

Inva

sive

SpeciesDatab

ase(w

ww.issg.org/datab

ase/welco

me/,accessed

1Decem

ber

2013

).

Driscoll et al. www.pnas.org/cgi/content/short/1409347111 10 of 11

Table S3. Enumeration of taxa developed or sold for eachagribusiness and whether those taxa are recognized as weeds,based on data in Table S1

Agribusiness no.Number of taxadeveloped or sold

Number listedas weeds* Weeds/total

1 12 10 0.832 3 3 1.003 4 4 1.004 2 2 1.005 6 6 1.006 1 1 1.007 15 12 0.808 15 12 0.809 10 9 0.9010 52 39 0.7511 22 21 0.9512 13 12 0.9213 12 12 1.0014 22 21 0.9515 61 55 0.9016 11 10 0.9117 21 14 0.67Mean (SD) 0.91 (0.10)No. taxa 178 141 0.79

No. taxa indicates the total number of unique taxa in the data set; taxaused by more than one agribusiness were only counted once.*Environmental weeds listed in Randall’s Global Compendium of Weeds(www.hear.org/gcw/).

Driscoll et al. www.pnas.org/cgi/content/short/1409347111 11 of 11