Nassella neesiana (Chilean needle grass) reduces farm profitability through its sharp seeds that damage wool, skins, hides and carcasses. It poses a biosecurity risk in New Zealand's Hawke's Bay region due to its seed dispersal via gravel and crushed aggregate from braided rivers. This study evaluated flotation and sieving to remove both dry and imbibed panicle seeds. We measured their length, width, mass, flotation, and passage through sieves with and without the awn (a hygroscopic, bristle-like structure). The dry awned seeds were, on average, 70.12 mm long and 1.13 mm wide, increasing to 74.07 mm long and 1.21 mm wide when imbibed. With imbibition, their mass with awn, increased from 9.41 to 14.06 mg, and from 3.96 to 5.04 mg with awn detached. Flotation was highest for the dry awned seeds (50%) and lowest for imbibed awned seeds (3%). The awn and imbibition restricted the seeds' passage through sieves while sand and a crushed aggregate (GAP 20 metal) facilitated their passage. The results indicate that flotation would be ineffective for removing N. neesiana seeds from river gravel but that an 8 mm aperture sieve could effectively remove them along with fine aggregate particles.

Variable efficacy of the acetohydroxyacid synthase (AHAS) inhibitor herbicide, flumetsulam, against dairy pasture populations of Ranunculus acris, has been attributed to resistance. We compared, over two years, the fitness of two resistant (R) clones and nine susceptible (S) clones treated or not with flumetsulam by measuring seed yields and survival of plants. Without the herbicide, all plants grew and survived. The seed yields of the R and S plants without the herbicide were similar (327 and 217 seeds/plant, respectively in Year 1, and 257 and 323 seeds/plant in Year 2) indicating no fitness penalty to the resistance. With the herbicide, growth and survival of the R plants was unaffected, whereas all S plants were defoliated before seedhead formation but 96% regrew from their rhizomes in Year 2. Seed yields of the treated R and S plants in Year 1 were 347 and 0 seeds/plant, respectively and 243 and 205 seeds/plant in Year 2. Survival of flumetsulam-susceptible R. acris plants following spraying would maintain susceptible AHAS alleles in the population, while without a fitness penalty, resistant plants would persist in non-spraying years facilitating further selection when the herbicide is used again.

The leaf-feeding beetle, Cassida rubiginosa, was introduced into New Zealand from Europe in 2007 as a biological control agent for Cirsium arvense (Californian thistle) and is now widely distributed. To determine the extent to which geographically variable climate might explain its observed sporadic impact, we developed a climate-niche model for each of the species using CLIMEX. The models reveal that the climate throughout most of New Zealand is variably suitable for both species, although everywhere relatively less suitable for the beetle. However, we found no evidence that this climate suitability bias favouring the thistle explains the sporadic impact of the beetle. First, 64 geographically separate beetle populations classified by field observers according to their impact on the thistle ('none', n = 29; 'low', n = 22; 'high', n = 13) showed no relationship with the modelled climate suitability (Ecoclimatic Index) of their locations for either the beetle or the thistle. Second, the two models reveal a similar climate suitability bias favouring the thistle throughout much of the species' 'eco-climatic suitability space' in their native range (Eurasia) where the beetle is a widespread natural enemy of the thistle. Hypothetical explanations for the sporadic impact of the beetle in New Zealand are explored, including lack of suitable overwintering habitat, predation, and incompatible thistle control operations. We conclude that in general, climate suitability, while a necessary condition, is alone insufficient for the success of a weed biological control program, and that non-climatic habitat factors may often explain spatially patchy impacts of natural enemies such as C. rubiginosa.

Step-point sampling, conceived by plant ecologist Dr Leonard Cockayne in ca. 1925 as a method for measuring the ground cover of species in New Zealand's modified tussock grasslands utilises a point marked on the observer's boot toecap. It is resource-efficient compared to other vegetation sampling methods and here we report on a unique evaluation of its accuracy and precision. The estimated cover of Taraxacum officinale in a grassy field (11%) was not significantly different from that obtained using line-point sampling (13%), but higher than that from line-intercept sampling (5.7%). It was unaffected across an order-of-magnitude range of sampling intensity (1600 to 160 observations/ha) and was acceptably precise with >= 400 observations/ha, although increased linearly with observation point size. These empirical results were supported by computer simulations enabling both accuracy and precision to be evaluated along with plant architecture. The simulations indicated that step-point sampling with ca. 400 observations per ha and an observation point diameter <= 1.1 cm (1.0 cm2), can provide ground cover estimates for a wide range of broadleaved species in pastures that are sufficiently accurate and precise for common pasture management applications. The method is deployed in the phone application, 'Grassland Cover Estimator' available at https://www.agresearch.co.nz/search?q = Grassland + cover + estimator.

Plant pathogens with a broad host range are commercially more attractive as microbial bioherbicides than strictly host-specific pathogens as a result of the wider market potential of a product capable of controlling multiple species. However, the perceived spatiotemporal disease risk to nontarget plants is a barrier to their adoption for weed control. We consider two approaches to managing this risk. First, we consider safety zones and withholding periods for bioherbicide treatment sites. These must ensure inoculum spreading from, or surviving at the site, exposes nontarget plants to no more inoculum than from natural sources. They can be determined using simple dispersal models. We show that a ratio of added:natural inoculum of 1.0 is biologically reasonable as an 'acceptable risk' and a sound basis for safety zones and withholding periods. These would be analogous to the 'conditions of use' for synthetic chemical herbicides aimed at minimizing collateral damage to susceptible plants from spray drift and persistent soil residues. Second, weed-specific isolates of broad host-range pathogens may avoid the need for safety zones and withholding periods. Such isolates have been found in many broad host-range pathogen species. Their utilization as bioherbicides may more easily meet the requirements of regulators. Mixtures of different weed-specific isolates of a pathogen could provide bioherbicides with commercially attractive spectrums of weed control activity. © 2023 The Authors. Pest Management Science published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.

Nassella neesiana (Chilean needle grass), an invasive 'sleeper weed' established in sheep and beef pastures in three of New Zealand's sixteen local government regions, has a potential geographic range amounting to 3.96 million hectares spanning all regions except the West Coast. It impacts the productivity, market value and welfare of livestock through its sharp penetrating that cause blindness and the downgrading of wool, hides, and carcasses. In this study we estimate the benefit of preventing its spread as the present value (PV) of local (regional) and national productivity losses that would accrue over 200 years under a 'do nothing' spread scenario. Using a 3% discount rate and two assumed spread rates, 201 and 100 years to 90% occupation of its potential range, we calculate national PV losses of NZ$ 192 million and NZ$ 1,160 million respectively. In a breakeven analysis, these losses, which equate to the national benefits of preventing the spread, justify annual expenditures of NZ$ 5.3 million and NZ$ 34 million respectively. Restricting the analyses to the regions with known infestations (Hawke's Bay, Marlborough, Canterbury) provided much lower estimates of the benefits (ranging from NZ$ 16.8 million to NZ$ 158 million) because spillover benefits from preventing spread to the other susceptible regions are not accounted for. These analyses support a nationally coordinated approach to managing N. neesiana in New Zealand involving surveillance and control measures respectively in the susceptible and infested regions.

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.

BACKGROUND:Weeds are among the most damaging pests of agriculture, causing ≈10% worldwide reduction in crop productivity each year. Over-reliance on synthetic chemical herbicides has caused weeds around the world to evolve resistance. Bioherbicides may be an alternative. However, among their many constraints including strict environmental requirements, complicated mass-production and high product costs, limited pathogenicity and a narrow spectrum of activity are frequently encountered and are major barriers to commercialization. RESULTS:We isolated a pathogenic fungus, HXDC-1-2, from diseased leaves of a gramineous weed, stiltgrass [Microstegium vimineum (Trin.) A. Camus], from the edge of farmland in Guizhou province, China. HXDC-1-2 was identified as the fungal species Bipolaris yamadae based on the morphological characteristics and ITS-GPDH-EF1α multiple primer analysis. Its potential as a bioherbicide was evaluated by determining its weed control efficacy and crop safety. The ED50 and ED90 values of HXDC-1-2 on Echinochloa crus-galli were 3.22 × 103 and 1.32 × 105 conidia mL-1 , respectively. Host range tests revealed that 20 gramineous weeds including Setaria viridis, Leptochloa chinensis, Eleusine indica, Pseudosorghum zollingeri, Leptochloa panicea, Bromus catharticus, E. crus-galli plants, were extremely susceptible whereas 77 crop species from 27 plant families including rice, wheat, barley, corn, soybean and cotton (excluding cowpea and sorghum) were unaffected. CONCLUSION:Bipolaris yamadae strain HXDC-1-2 has great potential to be developed as a commercial broad-spectrum bioherbicidal agent for controlling grass weeds in arable crops. © 2023 Society of Chemical Industry.

The phenoxy carboxylic acid, ALS-inhibitor and pyridine carboxylic acid herbicides vary in the magnitude and duration of Ranunculus acris control in dairy pastures and in collateral damage to clovers. For estimating the net economic benefit from a proposed herbicide treatment, we developed a model accounting for these sources of variation. Applied to a hypothetical dairy pasture with 12 tonnes dry matter/ha/year eaten and assuming present-day costs and prices (e.g. herbicides, nitrogen fertiliser, milksolids pay-out), the model illustrates the expected increase in net benefit with increasing pre-treatment R. acris cover. It also predicts lower breakeven covers (C-BE) for the phenoxys (MCPA C-BE = 3.72%; MCPB C-BE = -0.88%; MCPB + bentazone C-BE = 1.51%) and ALS-inhibitors (flumetsulam C-BE = 1.88%; thifensulfuron methyl C-BE = 1.50%) than for the pyridines (aminopyralid C-BE = 7.24%; aminopyralid + triclopyr C-BE = 5.72%), a result of their lower costs and lower and less-enduring clover damage compared to the pyridines. A greater uncertainty in the net benefit from the phenoxys and ALS-inhibitors results from a greater paddock-scale variation in their efficacy, a characteristic attributable to evolved resistance. The model is available as a weed control decision-support tool at https://giant-buttercup-ds-tool.azurewebsites.net/.

Weeds can be major environmental and economic burdens in New Zealand. Traditional methods of weed control including manual and chemical approaches can be time consuming and costly. Some chemical herbicides may have negative environmental and human health impacts. One of the proposed important steps for providing alternatives to these traditional approaches is the automated identification and mapping of weeds. We used hyperspectral imaging data and machine learning to explore the possibility of fast, accurate and automated discrimination of weeds in pastures where ryegrass and clovers are the sown species. Hyperspectral images from two grasses (Setaria pumila [yellow bristle grass] and Stipa arundinacea [wind grass]) and two broad leaf weed species (Ranunculus acris [giant buttercup] and Cirsium arvense [Californian thistle]) were acquired and pre-processed using the standard normal variate method. We trained three classification models, namely partial least squares-discriminant analysis, support vector machine, and Multilayer Perceptron (MLP) using whole plant averaged (Av) spectra and superpixels (Sp) averaged spectra from each weed sample. All three classification models showed repeatable identification of four weeds using both Av and Sp spectra with a range of overall accuracy of 70–100%. However, MLP based on the Sp method produced the most reliable and robust prediction result (89.1% accuracy). Four significant spectral regions were found as highly informative for characterizing the four weed species and could form the basis for a rapid and efficient methodology for identifying weeds in ryegrass/clover pastures.

A stage-structured matrix population model for the invasive perennial grass Nassella trichotoma (nassella tussock) was used to compare a wide range of hypothetical grubbing and biocontrol management options in a population of 19 visible plants ha(-1) on a typical sheep and beef cattle-grazed farm in North Canterbury, New Zealand. The model has two soil seed banks (ephemeral and persistent), five plant size classes, seed production and ground cover-dependent population growth reaching a carrying capacity of 35,000 plants ha(-1) (80% ground cover) in the absence of management. Elasticity analyses showed that population growth (lambda) was most sensitive to plant size class transitions and relatively insensitive to the soil seed bank transitions, seedling recruitment and seed production, implying management targeting the growth and/or survival of plants will be most effective. This conclusion was supported by population trajectory simulations which revealed that a biological control agent attacking seeds or seedlings would need to be 99.8% effective to achieve the population stability achieved under current management where 34.4% of the plants are grubbed annually. By contrast, a biological control agent attacking tillers and reducing plant transitions (growth) across all size classes by 30% annually gave a population trajectory equivalent to that achieved by annual grubbing. In the presence of annual grubbing, a combination of seed- and tiller-feeding agents giving 10% reductions in fecundity and plant growth, respectively, reduced the population density by 60%. Overall, the modelling shows that relaxing the current practice of annual grubbing (by reducing the rate or frequency of grubbing) whether in spring or autumn will result in the growth of N. trichotoma populations in North Canterbury. It also indicates that a biocontrol agent that impedes tiller production in N. trichotoma plants could potentially be an effective replacement for, or adjunct to, annual grubbing.

ABSTRACT Herbicide dose-responses of seedling progenies of Ranunculus acris populations varying in herbicide exposure history indicate that this weed has evolved multiple resistance to phenoxycarboxylic acid and acetolactate synthase (ALS)-inhibitor herbicides. To test the hypothesis that this ‘multiple-resistance’ can occur in the same plant, we conducted an experiment where 11 clones of R. acris, varying in parental population exposure to these two mode-of-action herbicides, were treated with full- and half-label doses of MCPA and flumetsulam. Aminopyralid and glyphosate (pyridine carboxylic acid and 5-enolpyruvylshikimate 3-phosphate (EPSP) synthase-inhibitor mode-of-actions respectively), both with label recommendations for R. acris, were included for comparison. Two clones, both from the same field population exposed historically to MCPA and to flumetsulam, and with anecdotal records of field control failures, were unaffected by flumetsulam and many plants survived treatment with MCPA, providing conclusive evidence for multiple herbicide resistance to these two herbicides in the same R. acris individuals. By contrast, there was no evidence among the 11 clones for resistance to either aminopyralid or glyphosate (86-100% mortality in all clones). The herbicide-resistant and -susceptible clones of R. acris provide a unique opportunity to investigate the biochemical mechanism(s) and genetics of phenoxycarboxylic acid and ALS-inhibitor resistance in this species.

The efficacy of control tactics for Ranunculus acris was quantified on poorly- and well-drained soils in a factorial experiment conducted over three years in 18 dairy pastures. Soil drainage, gibberellic acid and nitrogenous fertiliser (growth promoters), and a bioherbicide utilising Sclerotinia sclerotiorum had no effect on the cover of R. acris. The herbicides aminopyralid and aminopyralid+triclopyr, by contrast, gave long-lasting reductions in the cover of the weed and substantial temporary reductions in the clovers. Flumetsulam, thifensulfuron methyl, MCPA, MCPB and MCPB+bentazone were less effective overall. Pregraze mowing reduced R. acris as the frequency of mowing increased. For all herbicides, there was a 1:1 replacement of R. acris by grasses and clovers. The efficacy of the treatments varied greatly between pastures, possibly due to genetic differences between the R. acris populations and their historical exposure to the herbicides.

Abutilon theophrasti (velvetleaf, butterprint), a problematic weed of crops in some temperate countries, was introduced into New Zealand in 1948 as a potential fibre crop. It has naturalised in the North Island in the Waikato and Auckland regions but its status at the many sites where it was inadvertently sown in 2015 as a contaminant of Beta vulgaris seed lines remains unknown. To determine the potential distribution of the species in New Zealand we used an existing and a newly developed climate niche model, both constructed using the modelling software CLIMEX and a current global climate surface dataset. The Holt & Boose model, published in 2000, fitted only 54% of the known 3825 global occurrences of the species. It predicts that most of the North Island is climatically optimal and that most of the South Island is unsuitable. By contrast, the new model, parameterised using a geographically wider set of global occurrences, fitted 98% of the global occurrences. This more robust model predicts that almost all New Zealand, including all agricultural land, is currently climatically suitable for the species. We conclude that velvetleaf could become a weed throughout the country.

Herbicide resistance has repeatedly developed under intensive herbicidal weed management regimes globally with 255 species having resistant biotypes. In New Zealand, since 1979, resistance was found in 13 taxa, with >25 herbicides in 8 chemical classes showing reduced effectiveness (i.e. HRAC groups A, B, C, D, F, G, H, N and O). Cases included weeds in turf, pastures, orchards, vineyards, forage and arable crops. Surprisingly little is known about the spatial extent or frequency of this problem in New Zealand. We estimate that 14,000 farms have land-use histories like those favouring herbicide resistance historically. Sampling simulations of ≥10% farms provided good estimates of resistance prevalence for most regions and crop types. Acceptable sampling rates varied with target population size, actual resistance prevalence, and detection certainty. Our simulations provide a sobering caution regarding our ability to delimit the problem cheaply or accurately. Detection rates lower than 75% always give imprecise prevalence estimates. Sampling and screening involved 7.4 h labour and 27 km of travel costing $759 NZD per farm. Sampling 10% of farms would cost >$1 million NZD if lower risk farms were excluded, or >$3 million in exhaustive surveys. Regional farm and industry breakdowns could guide cost-sharing arrangements for surveys.

To quantify plant growth rates, responses to fertiliser and population responses to timing of grubbing (manual removal of plants) in a grassland population of Nassella trichotoma, three experiments were conducted. In Experiment 1, plant diameter growth, measured on 25 farms for up to 6 years, was unaffected by fertiliser and averaged 24 mm year(-1), enabling a rapid increase in seed output. In Experiment 2, autumn and spring grubbing, compared on six 5-ha plots on four farms over six years, gave reductions in population size of 71% and 83%, respectively. In Experiment 3, effects of time of grubbing and pasture disturbance in spring/summer on recruitment were measured and combined with published demographic process rates in a simple population model. The model predicts population stability when 33.75% of plants are removed (the current mean grubbing rate) before mid-November with 3% pasture disturbance, and also for later grubbing with higher grubbing rate or lower pasture disturbance.

Tradescantia fluminensis is an invasive plant species in New Zealand, Australia and parts of the USA. It reproduces vegetatively and can grow to form dense mats up to 60 cm deep. Growth is limited by available light, and shading is one of the few effective methods of control. In this paper, we develop a dynamic model of a vertical cross section of a T. fluminensis mat, capturing vertical variation in its biomass and internal light intensity. We measure both variables at different heights in experimental mats of the species and use these data to parameterize the model. The model produces realistic vertical biomass and light intensity profiles. We show that the mat grows to a steady-state biomass that depends only on: (i) the light absorption coefficient, which we estimate from experimental data and (ii) the ratio of photosynthesis to respiration rate. This steady state undergoes a transcritical bifurcation; when the ambient light intensity falls below a critical level, the biomass shrinks to zero and the mat cannot survive.

Intensification of pastoral farming in the temperate world has seen a dramatic shift from the botanically diverse native grasslands (e.g. tussock grasslands of New Zealand, prairie and Pacific grasslands of North America, pampas of South America and steppes of Europe) where many grasses, forbs and shrubs coexisted to botanically depauperate systems. In the extreme, such as in high-intensity flatland dairy farming, high farm production targets may be achieved with a pasture composed of as few as two species such as Lolium perenne and Trifolium repens. In less-intensive farm systems, such as sheep or beef cattle grazing on hill lands, a botanically more diverse pasture is common and often acceptable. Weed control is central to the establishment, maintenance and sustained productivity of these managed pastures. In particular, it is necessary to control non-palatable, injurious and poisonous species that would otherwise reduce livestock carrying capacity or impose other costs on the farm system such as those associated with reduced animal product quality, health and welfare (Fig. 1). These costs can be substantial (Jones et al., 2000; Sinden et al., 2004; Bourdôt et al., 2007a; Kaye-Blake et al., 2010).

In New Zealand, Regional Councils have the primary governmental responsibility for coordinating the management of weeds and other pests. Similarly, in Australia, it is the States that have this responsibility. In both countries, the national coordination and harmonisation of weed control efforts has been a perennial challenge, leading to initiatives such as the Weeds of National Significance in Australia and the National Policy Direction for Pest Management in New Zealand. Further, in both countries, the number of weeds that affect natural and productive ecosystems far exceeds the resources available to manage them such that prioritisation is necessary. In New Zealand, a favourable cost benefit analysis is required for a weed to be considered for inclusion in a Regional Pest Management Plan. The ability of New Zealand’s Regional Councils to undertake such analyses has been hampered by a lack of crucial biophysical data regarding the target weed, including its current distribution, potential future distribution and its rate of spread in the absence of control. To redress this, and to facilitate the required nationally consistent approach, scientists from AgResearch and CSIRO have collaborated to develop a platform of five tools: a national weed occurrence database (NWDD) that automatically harvests regional weed occurrence data; a database of CLIMEX models and their projections for New Zealand (CLIMENZ); a weed spread model (MDiG); a spatial weed risk assessment model (WRASP); and a cost benefit model (CBA for regional pest management). In this paper, we illustrate this decision-support system and comment on its adoption by Regional Councils in New Zealand and its contribution to targeting resources to the most deserving weed problems. We then consider how easily it might be extended for use in Australia.

Premise of The Study Simple sequence repeat (SSR) markers were developed for the study of genetic diversity of New Zealand Nassella trichotoma (Poaceae) and to support future studies in its native range. Methods and Results Genomic DNA was extracted from N. trichotoma leaf material and subjected to Roche 454 sequencing. From a total of 745 putative SSRs, 48 with di‐ to pentanucleotide repeats were screened, 32 primer pairs were designed, and 15 polymorphic markers were optimized for multiplex PCR on 105 N. trichotoma samples from four New Zealand regions. Each locus resulted in two to six alleles per locus, and four of the loci cross‐amplified in N. tenuissima. The mean observed and expected heterozygosity ranged from 0.00 to 0.90 and 0.00 to 0.50 per locus, respectively. Conclusions The novel SSR markers are valuable for the study of genetic diversity of N. trichotoma and might also be useful for closely related species.