general enquiries on this form should be made...

General enquiries on this form should be made to:Defra, Science Directorate, Management Support and Finance Team,Telephone No. 020 7238 1612E-mail: [email protected]

SID 5 Research Project Final Report

SID 5 (2/05) of 40

NoteIn line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The SID 5 (Research Project Final Report) is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra website. A SID 5 must be completed for all projects.

A SID 5A form must be completed where a project is paid on a monthly basis or against quarterly invoices. No SID 5A is required where payments are made at milestone points. When a SID 5A is required, no SID 5 form will be accepted without the accompanying SID 5A.

This form is in Word format and the boxes may be expanded or reduced, as appropriate.

ACCESS TO INFORMATIONThe information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

Project identification

1. Defra Project code AR0408

2. Project title

SUSTAINABLE WEED MANAGEMENT: DEVELOPMENT OF TECHNIQUES TO BALANCE BIODIVERSITY BENEFITS WITH RETENTION OF YIELDS

3. Contractororganisation(s)

Rothamsted ResearchHarpendenHerts AL5 2JQ

54. Total Defra project costs £ 443,644

5. Project: start date................ 01 October 2001

end date................. 30 September 2005

SID 5 (2/05) of 40

6. It is Defra’s intention to publish this form. Please confirm your agreement to do so...................................................................................YES NO (a) When preparing SID 5s contractors should bear in mind that Defra intends that they be made public. They

should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the SID 5 can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

(b) If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary7. The executive summary must not exceed 2 sides in total of A4 and should be understandable to the

intelligent non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.

Historically, research on weed biology and control has focussed on the aggressive species that threatened crop yields. However, it was becoming increasingly apparent, at the end of the last millennium that the impact of farming on the wider diversity of arable farmed areas had become unacceptably high and consequently a more ‘environmental’ approach to farming was needed. This change was formalised in the Government’s sustainable food and farming strategy (Working with the Grain of Nature, 2002). In terms of research into weed control, this new orientation meant a shift away from studying noxious species towards underpinning potential reductions in the intensity of weed control. A Defra desk study on the impact of herbicides on weed abundance and diversity (PN0940) had already identified that some arable plant species had greater biodiversity value than others. The suite of weed projects starting in 2001 focussed on 7 key species: fat hen, annual meadow-grass, knotgrass, groundsel, charlock, common chickweed, and scentless mayweed. The objective of the work of this commission was to ‘a) provide a better understanding of weed-crop competition and weed biology to support rotational approaches to weed management that are both economically sustainable and reduce environmental impact and b) to improve the practice of integrated management of arable weeds in support of better spatial/temporal targeting or reduced pesticide use. This project focussed particularly on part b) of this overall objective, aiming to investigate management options that could be used to increase the biodiversity value of arable crops, mainly winter wheat. The work included studying the tactical use of herbicides, both by the selection of appropriate products and doses, and the potential for spatially selective management. As a number of the environmentally beneficial species are primarily spring emerging (e.g. fat hen, knotgrass), additional work was done to establish the feasibility of using appropriate crop management (such as tine weeding in spring) to encourage these species within winter wheat. Finally, the project made limited progress with assessing the value of spring cropping to increase the presence of biodiversity valued species.

A) Autumn germinating species of biodiversity value Objectives 1 & 2. Tactical use of herbicide and mechanical weeding, and potential to exploit

lower doses to suppress weeds, rather than achieving complete control

1a) Tactical use of herbicides – product selection A series of seven field experiments explored the potential to select candidate herbicides that were not active on some of the seven non-target (beneficial) broad-leaved weed species, particularly, chickweed, mayweed and groundsel. For example, amidosulfuron did not control chickweed and pendimethalin would not control mayweed and groundsel. Partial control of mayweed and groundsel could be achieved with mecoprop and fluroxypyr. This

SID 5 (2/05) of 40

part of the project was quite successful. Conclusion – correct product choice can deliver appropriate selective control. These experiments identified several issues that need to be considered when implementing a low weed control approach: a) partial control of a desirable species was often associated with poor control of undesirable ones, b) commercially acceptability of such treatments was affected by the intensity of the weed infestation, c) the most popular herbicides currently used for the control of black-grass were very effective at controlling virtually all broad-leaved weed species. A further outcome was information on the impacts of surviving weeds on crop yields, which was used to estimate the economic consequences of not fully controlling all weed species. There was some evidence that crops would tolerate low levels of weeds with no associated yield loss, but once this ‘threshold’ was exceeded yield loss was linearly related to the increase in weed biomass in summer. Conclusion - crops can tolerate more weeds than they do in current crop production systems.

1b) Tactical use of herbicides - spatially selective weed control As weeds are frequently not uniformly distributed in fields a spatially located patch sprayer could be used to apply herbicides only to the denser weed patches, leaving other areas untreated to provide a resource for birds and invertebrates. Experiments in the project showed that it was possible to map weeds and use the map to apply spatially localised herbicide treatments. However, we were unable to demonstrate that the treatments resulted in an increase in desirable weed species in the unsprayed areas, although one would conclude that this was likely to be the outcome. This technique was not explored in detail, as the commercial development of patch spraying has been very slow. Spatially selective weed control has potential for environmentally beneficial weed management but other constraints on its acceptability to farmers have to be resolved first.

2) Potential to use reduced herbicide doses to manipulate the weed floraSix field and four pot experiments investigated the potential for using lower doses to reduce the degree of control of, primarily mayweed and chickweed (but also charlock and groundsel), by a range of commonly used winter cereal herbicides. Reduced doses lowered the degree of control but the dose response curves differed between experiments and it was hard to provide reliable guidance on the optimum dose. Climatic conditions and the relative vigour of crop and weed all influenced the final outcome. Conclusion - manipulation of herbicide dose is unlikely to be an effective method to enhance selectivities.

B) Spring-germinating species of biodiversity valueObjectives 3 and 4. Stimulation of spring emerging weeds in winter wheat and assessment

of the value of spring cropping for biodiversity.

3) Stimulation of spring weed emergence in winter wheat.Five experiments were done to assess the possibility of encouraging greater presence of fat hen and knotgrass in winter wheat. It was postulated that wide crop rows (24cm) and tine weeding in spring could increase emergence of these two species. The results did not support this hypothesis, as both techniques failed to stimulate emergence of these desired weeds. Conclusion - tine weeding in spring and wider rows are not effective at stimulating weed emergence in spring in winter wheat.

4) Benefits of spring croppingThe project has considered the role of spring cropping in changing the balance of weed species. Although weed infestations vary in spring crops depending on the competitive ability of the crop and weed management practices, there is some evidence that the increase in winter cropping has contributed to the decline in the arable weed flora. Thus, a return to more spring cropping will increase populations of a number of the more environmentally beneficial spring emerging species and additionally will provide over-winter stubbles, of benefit to birds. But, it must be remembered that gross margins of spring crops are lower than winter wheat, and so are less attractive to farmers. Conclusion - more desirable weed species were found in spring crops, even in winter crop dominated rotations.

SID 5 (2/05) of 40

Overall, the project showed that it was possible to use herbicides to retain some desirable weed species within winter wheat, but its implementation into practice will be quite difficult. Selective weed management will need to be based on choosing appropriate products, as manipulating low doses appeared unlikely to be successful. More development would be needed to introduce the more successful treatments into practice. Encouraging spring emerging weeds in winter wheat appears not be possible and so if these weeds are desired in cropping systems, then spring crops will have to be planted. A key element of the overall picture that seems to be missing is identification of the precise environmental benefits from increasing the presence of weeds in fields. How many, when and where should they be within the field? This project showed that leaving weeds in fields is likely to reduce wheat yields, although there may be a ‘threshold’ below which yields are not reduced. Additionally, associated increased weed seed production may impact on weed populations in subsequent years. Research in commission AR0407 and in the LINK project ‘Weed Manager (LK0916) has made progress with modelling this. However, as a consequence of these perceived negative impacts, farmers may be reluctant to adopt more environmentally beneficial weed management strategies without some form of financial support.

Project Report to Defra8. As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with

details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include: the scientific objectives as set out in the contract; the extent to which the objectives set out in the contract have been met; details of methods used and the results obtained, including statistical analysis (if appropriate); a discussion of the results and their reliability; the main implications of the findings; possible future work; and any action resulting from the research (e.g. IP, Knowledge Transfer).

1. Background

At the outset of this commission, research into weed management had been focussed strongly on improving understanding of the biology of the major aggressive weeds of arable cropping in the UK, in order to minimise their effects on crop production. However, it was becoming increasingly apparent, at the end of the last millennium, that the impact of farming on the wider diversity of arable farmed areas had become unacceptably high (Fuller et al, 1995; Preston et al., 2002; Siriwardena et al, 1998). As a consequence, Defra policy became oriented towards reductions of the impact of agriculture on arable ecosystems. This was formalised in the Government’s sustainable food and farming strategy (e.g. see ‘Working with the Grain of Nature 2002 (www.defra.gov.uk/wildlife-countryside/ewd/biostrat /index.htm). In terms of research into weed control, this new orientation meant a shift away from studying noxious species towards underpinning potential reductions in the intensity of weed control. A desk study on the impact of herbicides on weed abundance and diversity funded by Defra (PN0940) had already identified that some arable plant species had greater biodiversity value than others (Marshall et al 2001, 2003). This project provided information on plant species that should be retained in fields, so as to provide resources for invertebrates and birds (see Fig 1.1). Seven of these desirable weeds acted as the core of this and the other two commissions (AR0407, AR0409). These were Chenopodium album (fat hen), Poa annua (annual meadow-grass), Polygonum aviculare (knotgrass), Senecio vulgaris (groundsel), Sinapis arvensis (charlock), Stellaria media (common chickweed), and Tripleurospermum inodorum (scentless mayweed).

It is now clear that wild plants play a vital role in the arable ecosystem, providing the 'services' of food and shelter to higher trophic organisms (invertebrates, birds and small mammals) within the ecosystem food webs. There is debate about the respective contribution of plants in fields compared to those off field (field margins & hedgerows) but the reports of the Farm Scale Evaluation of GM crops

SID 5 (2/05) of 40

http://www.defra.gov.uk/wildlife-countryside/ewd/biostrat%20/index.htm

identify links between within field plant abundance and numbers of invertebrates, and hence their contribution to food webs of arable ecosystems (e.g. Heard et al., 2003, Hawes et al., 2003).

For arable farmers to retain plants of biodiversity value in fields, a balance has to be struck between this objective, and the prevention of harvesting problems and yield loss arising from competition between the weeds and the crop (i.e. a balance between environmental and economic goals). However, overuse of herbicides, endeavouring to eliminate all weeds within fields, at current input and commodity prices, will reduce profitability. Consequently, for both economic and environmental reasons there is a need to refocus weed management towards a more targeted approach. Damaging weeds need to be controlled but less damaging and environmentally beneficial weeds can be left in the field to provide resources for vertebrates and invertebrates that are part of the arable ecosystem. Such a targeted approach requires appropriate linkage of weed control practice with a good understanding of the competitivity, biology and population dynamics of the weeds. Users need information on the biology of weeds to have confidence that a reduced weed control strategy in one year will not cause unmanageable problems in future years. This programme integrates the data from AR0407 and AR0409, to explore alternative ways of managing weeds.

Figure 1.1 An example of information about plant:invertebrate associations that can be obtained from the Phytophagous Insect Data Base (Marshall et al, 2001). Species included in AR0408 are in darker shading.

Experiments in AR0408 investigated methods that could be used to manage fields or part fields to enhance plant biodiversity, in winter wheat, without jeopardising yields. The work has focussed on herbicide use, as such products are likely to be the mainstay of weed management for most farmers for the foreseeable future. However, other cultural practices, such as sowing date, cropping type, and mechanical weeding that can play a role in more targeted weed management, have also been included. The information from the project will feed into the SA LINK project (LK0916) Weed Management Support System (now Weed Manager), to enhance its value as a tool to encourage lower input weed management.

This project’s aim was to meet the Defra Objective in ROAME A to a) ‘provide a better understanding of weed-crop competition and population dynamics to support rotational approaches to weed management that are both economically sustainable and reduce environmental impact and b) to improve the practice of integrated management of arable weeds in support of better spatial/temporal targeting or reduced pesticide use’. This will contribute towards meeting the government goal of enhancing biodiversity and reducing the environmental impact of farming.

2. Scientific Aims and Objectives (from the project CSG7)

Overall objective: to test the potential for managing weeds that contribute to biodiversity in wheat without reducing economic sustainability (i.e. controlling damaging weeds and retaining those with biodiversity benefits). The programme will focus on the following species that have already been

SID 5 (2/05) of 40

identified as being ‘beneficial’ for biodiversity: Chenopodium album, Poa annua, Polygonum aviculare, Senecio vulgaris, Sinapis arvensis, Stellaria media, and Tripleurospermum inodorum.

A) Autumn germinating species of biodiversity value:1) Tactical use of herbicides and mechanical weed control:to establish the feasibility of managing weeds by selecting appropriate herbicides (or other weed control technique) to eliminate competitive weeds, whilst retaining those of biodiversity value. Work to include studies of the practicability of applying herbicides on a spatially selective basis, thus controlling high-density patches, whilst retaining plants on areas where densities are lower.

2) Manipulation of weed competitive effects:to investigate the possibility of using sub-lethal herbicide treatments and/or mechanical weeding to suppress the growth of target species, so that biodiversity value is retained but yield losses are reduced. Assess impact on viable seed production.

B) Spring germinating species of biodiversity value:3) Stimulation of spring weeds in winter wheat:to explore the possibility of using a tine weeder to stimulate the emergence of weeds that would be of environmental benefit. To assess success in terms of plant numbers and seed production. Check consequences for subsequent crops.

4) Benefits of spring cropping:to collate and summarise available information on the benefits of spring cropping, as it is already quite well established that spring cropping can have biodiversity benefits.

3. Potential to use reduced herbicide doses to manipulate weed flora (Objective 2) – Field experiments 2001-2003

Published information in the UK on the responses of weeds to low doses of cereal herbicides is rather limited. Although, independent research focussing on the commoner aggressive species such as Galium aparine was done in the 1980s and 1990s, much less work has been published on the environmentally valued species included in this project (e.g. Proven et al., 1993). Such work as has been done tends to focus on the performance of full rates and perhaps a half rate but does not explore the dose response fully (e.g. Wright et al., 1993; Cussans & Courtney, 1995). Research elsewhere in Europe, especially in Denmark, has explored dose responses in more detail (e.g. Christensen, 1994; Kudsk, 1999), but the focus of their work tends to be on spring crops, not winter wheat.

Thus, extensive data sets on the relative sensitivities of environmentally desirable broad-leaved weeds to a range of cereal herbicides are not available either from independent research groups or from the manufacturers’ labels. A lot is known about the performance of the recommended dose of products against common target weeds but much less is known about effects of lower doses, or on responses of non-target weeds. A major constraint on identifying selective treatments lies in the fact that many product labels only define responses of a limited number of species, but whether other species are resistant or there is ‘no information’ is commonly not distinguished (Table 3.1). Consequently, this part of the project was aimed at establishing the feasibility of increasing selectivities between species (to retain the desired species and minimise the aggressive ones) by reducing herbicide dose. It was not intended that a full list of potential selective products would be identified to achieve this objective, but rather that the principle of reducing doses to achieve environmental value would be established.

Six winter wheat field experiments in 2001/02 and 2002/03 explored the sensitivity of S. media and T. inodorum primarily to low doses of metsulfuron-methyl and fluroxypyr. Effects on crop and weed biomass and on crop yields were recorded. The aim of the work was to identify doses that would not control, or would only suppress, these environmentally desirable species. As impacts of any surviving weeds on crop yield were clearly of great relevance to the economic sustainability of the treatments, all plots were harvested at maturity.

This work was supported by more wide-ranging pot experiments at Rothamsted in 2001-04 to monitor the effects of the main spring-applied broad-leaved weed herbicides in winter wheat on the target weeds (see Section 4).

SID 5 (2/05) of 40

Table 3.1 Information on sensitivities of the six broad-leaved weeds to commonly used cereal herbicides, derived from product labels and the encyclopaedia of Weed Manager (NB S=90-100% control (susceptible): MS>76%-90% control; MR 51-75%: R <50% control (resistant))

Materials & Methods

Five of the six experiments were done in winter wheat with autumn-emerging weeds and included a range of doses of the selected herbicide and a series of densities of the target weed, creating a factorial design (Table 3.2). All were designed as randomised block experiments, with three replicates. Three focussed on S. media and two on T. inodorum. Weed seeds were broadcast onto the plots at appropriate densities, either just before (S. media) or just after (T. inodorum) drilling in the autumn. The sixth experiment at Rothamsted in 2001/02 used natural infestations of S. media (69 plants/m2) and G. aparine (18 plants/m2), which were treated with five doses of mecoprop-p (1.38, 1.03, 0.69, 0.34, 0.17 g ai/ha) or amidosulfuron (30, 22.5, 15, 7.5, 3.75 g ai/ha) (five replicate blocks).

Assessments were made of weed densities, of crop and weed biomass at the time of treatment. All experiments were sampled for crop and weed biomass once in May/June/July and yields were recorded, after harvesting with a small plot combine harvester. Extra assessments to cover specific aspects of the experiments, such as seed production, were also done, when appropriate.

Table 3.2 Details of experiments investigating effects of weed density and herbicide dose.

Experiment Weed species Weed densitiesplants/m2

Herbicide Date of appln

Herbicide doses(g ai/ha)

RRes 01/02 T. inodorum 0, 113, 255, 923

metsulfuron-methyl

25 March 02

0, 0.75, 1.5, 3.0, 6.0

ADAS 01/02 S. media 48, 107, 243, 794

metsulfuron- methyl

12 March 02

0, 0.75, 1.5, 3.0, 6.0

RRes 02/03 S. media 31, 90, 280, 1126

fluroxypyr 15 April 03

0, 25, 50, 100, 200

RRes 02/03 T. inodorum 1, 30, 27, 127*

metsulfuron- methyl

7 May 03 0, 0.38, 0.75, 1.5, 3.0

ADAS 02/03 S. media 16, 23, 38, 113

metsulfuron- methyl

6 May 03 0, 0.75, 1.5, 3.0, 6.0

* Spring counts on sub-sample of plots, as autumn counts greatly underestimated densities that were present at the end of the winter

SID 5 (2/05) of 40

Results

Crop and weed biomass in summerAll the S. media experiments were sampled at the end of May or early in June, prior to the onset of senescence of the untreated S. media. Data presented in Table 3.3a show the percentage reduction in plant dry wt (g/m2) achieved by the herbicide treatments in all experiments. All doses of amidosulfuron were ineffective on S. media, although the herbicide controlled G. aparine very effectively even at 25% of the full rate (Fig. 3.1a). In contrast, the mecoprop-p achieved over 80% control of both weeds (Table 3.3, Fig. 3.1b) at doses of 0.34 kg ai /ha or higher. It should be noted that neither herbicide was effective on V. persica. The other herbicide tested in this work, which is generally used to control G. aparine, was fluroxypyr. The test in 2002/03 showed that the full rate controlled S. media but lower rates were much less effective. Interestingly, 50% of the full 1l/ha rate of this product caused over 90% reduction in the dry weights of G. aparine, thus indicating that some selective management of the two weeds was possible with the 50% rate.

The two ADAS Boxworth experiments focussing on metsulfuron demonstrated somewhat different results as the 50% rate achieved over 90% reduction in biomass of S. media in one year (in 11 weeks) but only 30% in the second (in 4 weeks) (Table 3.3a). The differences between the two experiments suggest that reliable partial control would be difficult to achieve and consequently yield effects would be unreliable. These results contrast with the two Rothamsted experiments measuring the control of T. inodorum with the same herbicide. All treatments, even those as low as less than 10% of the full rate, caused more than 90% reduction in plant dry weight. Thus, this herbicide is not suitable for management aimed at the retention of low amounts of T. inodorum.

Figure 3.1 Dry weights of all weeds on the Rothamsted experiment 2001-02 on 23 May, following treatment with a) amidosulfuron or b) mecoprop-p (NB Vertical bars are SED’s of total weed weights)

Crop yieldsThe second key component of these experiments was the measurement of crop yield responses

to the different weed infestations and herbicide treatments. Clearly, if herbicides are to be used as management tools to reduce the intensity of weed control and leave a proportion of the weeds in the field, then it is essential to understand what the impacts of the remaining weeds will be on crop yields. Significant effects of herbicide dose on crop yields were recorded in all experiments except the two done at Rothamsted in 2002/03. There were a considerable number of ‘other weeds’ mainly annual grasses on these two experiments and certainly, these increased the yield losses, and diluted the effects of the fluroxypyr on the S. media experiment (Table 3.3b). Crop yields were also rather poor on these two experiments, probably due to poor crop vigour over the winter and drought in mid-summer. In the other experiments, other weeds were less abundant and significant yield losses were detected. A 5% yield loss threshold has been adopted in Table 3.4 (see shaded values), as this approximates to the average cost of herbicides used to control these weeds. Overall, the treatments resulting in the greatest yield losses in Table 3.3b were the mirror image of the ones causing more than 80% weed control in Table 3.3a. The untreated weeds in all comparisons caused more than 8% yield loss.

There seemed to be potential to retain S. media arising from treatment with the full rate of amidosulfuron, as yield losses were only marginally greater than the 5% threshold. Similarly, 25% and 12.5% rates of metsulfuron on the two ADAS experiments, left some S. media without causing large

SID 5 (2/05) of 40

yield losses, but it must be noted that the 2002/03 ADAS experiment yield loss from the weeds even on the untreated plots was relatively low (8.7%). This may have been due to drought limiting both wheat yields and S. media competition. Both the two experiments focussing on T. inodorum resulted in little yield loss on all the herbicide treated plots (Table 3.3), as metsulfuron, even at very low dose was highly effective on the weed.

All the above experiments with sown weeds had four comparative densities. Weed density had a significant effect on crop yields in three of them. This was either due to the increased yield loss in the untreated at the higher densities, or to the failure of the herbicide to control high densities of the weed effectively, thus leaving more individuals to compete with the crop. Any interactions between herbicide dose and weed density were either small or non-significant.

Table 3.3 Response of S. media and T. inodorum to a range of doses of several herbicides in seven experiments 2001-03: a) degree of control (% reduction in plant dry weight compared to untreated), b) percentage yield loss

ExperimentsSites RRes RRes ADAS ADAS RRes RRes RResYear 2001/02 2001/02 2001/02 2002/03 2002/03 2001/02 2002/03Weed Species S. media S. media S. media S. media S. media T.

inodorumT. inodorum

Herbicide mecoprop -p

amido-sulfuron

metsulfuron metsulfuron fluroxypyr metsulfuron metsulfuron

Full rate (kg ai/ha)

1.38 0.03 0.03 0.03 0.2 0.03 0.03

Product % ai or g/l

600 g/l 75% 20% 20% 200 g/l 20% 20%

a) % reduction in plant weightdate assessed 23 May 23 May 27 May 3 June 9 June 8 July 14 JulyProportion of rec. dose0.067 940.125 67 -26 57 18 -13 89 980.25 95 -2 71 16 -7 100 990.5 99 20 91 32 54 100 1000.75 100 461.0 100 57 98 80 86 100SED 8.8 15.3 5.9 19.7 12.2 na na

b) % Yield lossWeed free yields (t/ha)

9.09 9.09 8.76 8.34 6.73 9.59 6.74

0* 10.4 10.4 17.6 8.7 16.5 25.8 8.10.067 -6.10.125 5.1 9.1 6.3 3.1 18.9 3.1 -1.90.25 0.7 10.0 2.4 4.9 16.3 0.6 0.10.5 0.6 7.6 1.7 1.7 15.6 0.4 00.75 -0.9 6.21.0 0.2 5.2 0.1 3.2 8.1 -0.3SED 1.63 1.63 2.69 3.00 4.80 3.01 4.96

* 0 = untreated plotsShaded areas are treatments achieving more than 80% control and yield losses greater than 5%

There was clearly an inverse relationship between the % weed control (as reflected in weed biomass in summer) and % yield loss in the 2002 Rothamsted experiment (see Fig. 3.2). Similar responses were recorded in the other experiments on reduced doses, highlighting the problem of balancing retention of yield with retaining weeds to deliver ecosystem services. This and other experiments suggest that a low level of weeds can be tolerated without jeopardising yields. However, the variability in the data makes identifying such thresholds hard to prove statistically. This issue is discussed further in Section 9 (overview and conclusions).

SID 5 (2/05) of 40

Figure 3.2. Relationship between total weed biomass in May and crop yield loss, as influenced by five doses of mecoprop-p and amidosulfuron (Expt RRes 2002)

4. Potential to use reduced herbicide doses to manipulate weed flora (Objective 2) – Pot experiments 2001-2004

As it was not possible, within the resources available in this commission, to investigate in the field all the herbicides that could be used on the seven target weed species, a more extensive set of data on herbicide sensitivities was created from a series of pot experiments. These were aimed to update and widen the data set collected at the Weed Research Organisation and Long Ashton Research Station during the 1960s to 1990s (e.g. West 1988, 1994). These pot experiments provided data on the sensitivity of approximately 40 plant species to several doses of newly introduced products, but ceased in the 1990s and so new products have not been studied.

Materials & Methods

Pot experiments were done in the spring and summer in 2002-2004. Those in 2002 and 2003 studied the response of S. media to eight herbicides. A further experiment in 2003 measured the response of T. inodorum to five herbicides, and in 2004 four herbicides were applied to the previous two weeds and also to S. vulgaris and S. arvensis. Four plants/pot were grown in 12 cm pots in the glasshouse and these were transferred to an outdoor sand bed prior to treatment in 2002 and 2003. This was not done in 2004, as the pots were kept in the glasshouse throughout the experiment. In the first two years, there were four replicates and five herbicide doses, plus an untreated, whilst in the final year there were only 3 doses plus an untreated. All experiments were laid out in randomised blocks. Treatments commenced at the top dose (generally the field recommended rate) and were reduced by 50% at each subsequent dose, thus 100, 50, 25, 12.5% of full rate. In all experiments, treatments were applied using a glasshouse pot sprayer at a pressure of 2.45bar and volume rate of 254-284 l/ha. Plants were hand harvested 3-4 weeks after treatment and fresh weights/pot were recorded. Details of the weed species and herbicides used in the experiments are given in Table 4.1 and more details of doses and dates of treatment and harvest are in Appendix 4.1.

The data in 2002 and 2003 were analysed using the Maximum Likelihood programme (MLP), which calculated Log10ED50 values (dose required to reduce fresh weights by 50%) for each tested product. These were back-transformed to determine the relevant ED50 doses in terms of product/ha. In 2004 where there were only three doses and an untreated, ED50s were not calculated but analysis of variance was used to calculate the appropriate standard errors, and ED50s were approximated from the means.

SID 5 (2/05) of 40

Table 4.1 Details of the weed species and herbicides used in the four pot experiments

Year Weed species Herbicides tested2002 S. media amidosulfuron, bromoxynil+ ioxynil, cinidon-ethyl

fluroxypyr, mecoprop-p, metsulfuronthifensulfuron + metsulfuron

2003 S. media carfentrazone+ mecoprop-p, mecoprop-p, metsulfuron2003 T. inodorum bromoxynil+ ioxynil, florasulam, mecoprop-p, metsulfuron,

thifensulfuron + metsulfuron2004 S. media, S. vulgaris,

S. arvensis, T. inodorumamidosulfuron, fluroxypyr,mecoprop-p, metsulfuron

Results

Expt S. media 2002As can be seen in Fig 4.1, all doses of metsulfuron and metsulfuron+thifensulfuron achieved high levels of control of the target weed. As a consequence, it was not possible to calculate ED50s for these herbicides. In contrast, cinidon-ethyl failed to control this species at all doses (Fig. 4.1) and consequently the calculated ED50 was actually greater than the full dose (Table 4.2). Other products that are appreciably less effective on this weed at lower doses were amidosulfuron and fluroxypyr, which had ED50s at 15-20% of the full dose (Table 4.2).

Figure 4.1. Percentage reduction in S. media fresh weights (g/pot) caused by five doses of seven herbicides (pot experiment 2002)

Table 4.2 Calculated ED50s and detransformed values for the seven herbicides tested on S. media in 2002

Herbicide Product Log10 ED50* Detransformed ED50

% of full rec. rate

amidosulfuron Eagle (75%) 0.880 (0.074) 7.58 g/ha 19cinidon-ethyl Lotus (200g/l) -0.519 (0.149) 0.30 l/ha 100+fluroxypyr Starane 2 (200g/l) -0.849 (0.078) 0.14 l/ha 14ioxynil + bromoxynil

Deloxil (380 g/l) -0.801 (0.0865) 0.16 l/ha 7.9

mecoprop-p Duplosan (600 g/l) -1.065 (0.110) 0.086 l/ha 3.7metsulfuron Ally (20%) na <0.1thifensulfuron + metsulfuron

Harmony M (68+7%)

na <0.1

* Based on product dose, figures in brackets are estimated standard errors na = herbicide too effective at all tested doses

SID 5 (2/05) of 40

Expt S. media 2003One further herbicide was tested against S. media in 2003, and tests on two products from 2002 were repeated (Table 4.1). The starting dose of metsulfuron in 2003 was reduced to 25% of the recommended rate. Despite this, all doses of metsulfuron almost fully controlled the S. media again and it was not possible to calculate an ED50 for this product (Table 4.3). The ED50 for mecoprop-p in 2003 was similar to that calculated in 2002. Carfentrazone + mecoprop-p was less effective than mecoprop alone. As the mecoprop dose in this mixture was lower than that of mecoprop-p alone, the carfentrazone appeared to be providing little benefit, though the mixture may have potential to partially control S. media.

Expt T. inodorum 2003The experiment with T. inodorum demonstrated the high activity of the two sulfonyl urea herbicide products against this species, as well as against S. media. The ED50 values were less than 2% of the recommended dose (Table 4.4). The dose responses for these two herbicides and for mecoprop-p were very ‘flat’, as although low doses had considerable effects on the plants, the highest dose failed to control the plants completely. This is reflected in the high standard errors for the ED50s (Table 4.4). Florasulam, ioxynil+ bromoxynil and mecoprop-p had similar effects on T. inodorum, as the ED50 were all 5-6% of the recommended rate. Mecoprop-p was least effective at the full rate, suggesting some potential to retain this weed in fields treated with this herbicide.

Table 4.3 Calculated ED50s and detransformed values for the four herbicides tested on S. media in 2003


% of full recommended

ratecarfentrazone + mecoprop-p

Platform S (1.5+60%)

2.038 (0.091) 109 g/ha 10.9

mecoprop-p Duplosan (600 g/l) -1.186 (0.292) 0.065l/ha 2.8metsulfuron Ally (20%) na <0.1* Based on product dose, figures in brackets are estimated standard errors na = herbicide too effective at all tested doses

Table 4.4 Calculated ED50s and detransformed values for the six herbicides tested on T. inodorum in 2003


% of full recommended

rateflorasulam Boxer (50g/l) -2.287 (0.310) 0.005 l/ha 5.2ioxynil + bromoxynil

Deloxil (380 g/l) -0.946 (0.141) 0.113 l/ha 5.7

mecoprop-p Duplosan (600 g/l) 0.835 (1.292) 0.15 l/ha 6.4metsulfuron Ally (20%) -1.0667 (1.836) 0.086 l/ha 0.3thifensulfuron + metsulfuron

Harmony M (68+7%)

0.0043 (0.751) 1.01 g/ha 1.3

* Based on product dose, figures in brackets are estimated standard errors

Expts 2004 (T. inodorum, S. media, S. arvensis, S. vulgaris)The experiments in 2002 and 2003 focussed on a wide range of herbicides for single weed species. The 2004 experiment looked at the relative sensitivities of four of the target weeds to four herbicides (Table 4.1). Metsulfuron was highly active on S. media and T. inodorum (Fig. 4.2). S. vulgaris was least sensitive to this herbicide and to amidosulfuron, suggesting that this weed could be retained whilst the other three would be controlled. The normal field rate of fluroxypyr (200 g ai/ha) also failed to control S. vulgaris very effectively. This herbicide left some survivors at the 200 g/ha rate in all four species. Mecoprop-p, similarly failed to completely control S. media, S. vulgaris, and T. inodorum at its full rate, but the fourth, S. arvensis was seriously damaged.

SID 5 (2/05) of 40

metsulfuron

0102030405060708090

100

S. arvensis S. media S. vulgaris T. inodorum

plan

t fre

sh w

t (g/

pot)

untreated1.5 g/ha36

amidosulfuron

0102030405060708090

100


plan

t fre

sh w

t (g/

pot)

untreated7.5 g/ha1530

mecoprop

0102030405060708090

100


plan

t fre

sh w

t (g/

pot)

untreated0.35 kg/ha0.691.38

fluroxypyr

0102030405060708090

100


plan

t fre

sh w

t (g/

pot)

untreated100 g/ha200400

Figure 4.2. Response of T. inodorum, S. media, S. arvensis and S. vulgaris in the 2004 pot experiment (plant fresh weight/pot) to three doses of amidosulfuron, metsulfuron, mecoprop-p and fluroxypyr. (herbicide doses presented as g (or kg) ai/ha: vertical bars = 1 x SED))

Overview of field and pot experiments

S. media: Cinidon-ethyl failed to affect this weed, in the pot experiments, whilst metsulfuron and thifensulfuron + metsulfuron in field and pots were very active even at low doses (Table 4.5). The high responses to the sulfonyl-ureas had also been shown in several of the field experiments. Fluroxypyr and amidosulfuron may be selective at low doses (c. 25% of full rate). Mecoprop-p was rather variable giving high control at all doses in the field and in two pot experiments, but being somewhat less active in the third pot experiment (2004). Carfentrazone + mecoprop-p was not as active as mecoprop-p alone.

T. inodorum: Fluroxypyr had little effect on this weed in the 2004 pot experiment. Mecoprop-p also did not achieve full control at low doses, but its performance varied between experiments. Label recommendations indicate that this weed is moderately resistant to both herbicides. Our work confirms this but does show that if ‘weed suppression’ is the objective, then they may have some value. The field experiments showed high activity from metsulfuron on T. inodorum, similar to the conclusions of the two pot experiments. Florasulam and ioxynil + bromoxynil in the 2003 pot experiment, also showed high activity on this species. Amidosulfuron was also effective in the 2004 experiment.

S. vulgaris: The single pot experiment in 2004 with S. vulgaris indicated poor control from metsulfuron, but general field experience (Table 3.1) and other glasshouse studies suggest that it is sensitive at the recommended rate (Marshall et al 2001; Richardson et al 1984 ). Fluroxypyr (at the 200 g ai/ha dose) was not very active and amidosulfuron was also only partially effective. S. vulgaris was poorly controlled by the lowest dose of all the herbicides tested in the pot experiment (Fig. 4.2). Other experiments (not reported here) indicated that it could be controlled with florasulam and with high rates of ioxynil + bromoxynil.

S. arvensis: Of all the species tested in the 2004 pot experiment, S. arvensis seemed most sensitive to the products used. This species was not tested in the field although it was a major component of the indigenous flora on the 2005 ADAS experiment (see below).

SID 5 (2/05) of 40

Table 4.5 Overview of the performance of nine broad-leaved weed herbicide products tested in pot experiments, on four major non-target weeds in UK. Results presented are means based on actual ED50s from 2002 and 2003 and estimated ED50s from 2004. Values are % of recommended dose needed to achieve 50% control (reduction in fresh weight).

S. media T. inodorum S. vulgaris S. arvensisamidosulfuron 20 <10 20 20carfentrazone + mecoprop-p 22cinidon – ethyl 100florasulam 5fluroxypyr (1 l/ha dose) 10 100 60 20ioxynil + bromoxynil 8 6mecoprop-p 5* 20 25 20metsulfuron <5 15 85 25metsulfuron + thifensulfuron <5 <5

* this ignores poor control in 2004

The overall conclusion of the work in the field and in pots in 2002-2004 was that it was difficult to identify appropriate doses to selectively retain desired species in the field, except when the product was not fully effective even at the recommended rate. As has been shown in other herbicide studies, low doses are more vulnerable to variability in performance arising from such factors as weed size at application and weather conditions (e.g. Proven et al., 1993). It was concluded that the subsequent work in this part of the project (see Chapter 6) should focus on choosing appropriate herbicides that were intrinsically less active on the non-target weed species, rather than endeavouring to identify the suppressive potential of low doses of products that were highly active on these non-target weeds at their full rate.

5. Potential for mechanical weed control and other crop management practices to encourage the spring germination of broad-leaved weeds in winter wheat (Objective 3)

A number of the weed species thought to be of considerable value to invertebrates and birds (e.g. C. album, P. aviculare) emerge in the spring and so are much more frequent in spring planted crops. However, some do establish in autumn-sown crops but in such situations are much less abundant and the plants that do grow are much less vigorous. This lack of vigour is primarily related to the high level of competition exerted by the crop, which emerged in the previous autumn. In spring crops these species can be quite aggressive but in winter crops this is less of a problem. The project explored two approaches to encourage a greater presence of these spring emerging species in winter wheat: tine weeding and wide crop rows. The former moves the top layer of soil disturbing the roots of existing crop and weed plants. This disturbance could stimulate weed emergence, as cultivations and the associated exposure of ungerminated seeds to light and to greater variations in temperature are prime stimulants to germination. It was postulated that growing wheat on wide rows would increase the exposure of the seeds to light and would also provide more space for the newly germinated seedlings. Five experiments were done between 2001 and 2004 to establish the potential of these two techniques to increase the presence of spring emerging weeds in winter wheat and to monitor their effects on crop yields.

Materials & methods

Four field experiments were done at Rothamsted and ADAS Boxworth in the years 2001-02 and 2002-03 on sites known to have a seedbank of the desired weeds. All four had the same basic treatment structure: winter wheat crops either sown in 12 or 24cm rows, varying levels of autumn weed control and a comparison in spring of: a) no spring treatment, b) tine weeding in spring, and c) a selective spring herbicide. Further details of treatments are given in Table 5.1. The herbicide treatments varied between sites and years, depending on the background weed flora but all aimed not to affect the desired spring emerging species. Tine weeding on all experiments was done with a ‘finger weeder’ (e.g. Einbock). All experiments were designed either as randomised blocks or as split plots with four replications.

SID 5 (2/05) of 40

The final experiment at Rothamsted in 2003-04 simply compared winter wheat sown on 25cm rows, with that grown on 12cm rows and with spring wheat (on 12cm rows). Weeds were treated uniformly in the autumn (in the winter wheat) and then the plots were split into the three spring treatments used in the previous experiments (none, tine weeding, selective herbicide). Additionally, seeds of C. album and P. aviculare were broadcast onto all plots prior to sowing the winter wheat. The wheat was sown on 5 October and the spring wheat on 30 March. All the plots of winter wheat in this experiment received isoproturon and clopyralid in winter. The tine weeding and the spring treatment with florasulam were done in April. The spring wheat was tine weeded in early May and the herbicide treatment (metsulfuron) applied in early June.

The main assessments in all experiments were a count of weed densities (from a series of quadrats placed randomly on each plot) in late spring, a sample of crop and weed biomass in mid-summer (2 x 0.5m2 samples/plot) and crop yields from a small plot combine harvester or by hand harvesting.

Table 5.1 Response of weeds and wheat in two Rothamsted experiments to treatments comparing: 24cm vs 12cm rows; none, full and selective herbicide treatments in autumn; none, selective herbicide and tine weeding in spring. (‘all’ = all weeds present, ‘spring’ = spring germinating species only)

* This SED is based on the 12cm row, autumn full and selective treatments only

Treatments and experimental detailsSowing date 28 Sept 01 2 Oct 02Tine weeding 5 May 8 May Herbicides Autumn full pendimethalin pendimethalin

+ flupyrsulfuron (3 Nov) + flupyrsulfuron (26 Nov) Autumn selective IPU + CMPP (3 Nov) CTU + IPU (26 Nov) Spring selective bromoxynil + ioxynil (2 April) amidosulfuron (8 April)

IPU = isoproturon, CTU = chlorotoluron, CMPP = mecoprop-p,

Results

Expt 02 Rothamsted

In the absence of an autumn herbicide an appreciable infestation of annual grass weeds developed on the plots. These were not well-controlled in the spring and became very competitive, reducing crop yields (Table 5.1). Where selective and ‘full’ herbicide treatments had been applied in autumn, weed survival in spring was much lower. There was virtually no flush of spring emerging weeds on any treatment. Even the soil disturbance by the tine weeding treatment failed to stimulate

SID 5 (2/05) of 40

seedling emergence. The only treatment that resulted in the presence of some spring emerging species was the wide row treatment with selective weed control, where low numbers of P. aviculare and C. album were recorded (Table 5.1). This may have been due to the wider rows but could also be related to the lower crop density on this treatment, as in this experiment the wide rows were created by removing the wheat plants from alternate rows, after emergence. Thus, overall crop density was only 50% of that of the other treatments. As a consequence, crop vigour and yields were much lower than after the other treatments (this approach was not used in subsequent experiments).

Expt 02 ADASTotal weed levels were low on this experiment, except where no weed control was done in autumn where 33-44 g/m2 (dry weight) were recorded on 17 June after tine weeding or no spring treatment (Table 5.2). In the summer, fewer weeds survived on the plots treated with the selective herbicide (carfentrazone + mecoprop-p) but there was little reduction in weeds by the tine weeding. A very few spring emerging weeds (C. album, Atriplex patula, P. aviculare) were present in summer. There was a suggestion that more of these target species were present on the wide row treatments, especially in the absence of the spring herbicides. Tine weeding did not increase spring weed emergence in the dry conditions experienced in April 2002. Yields were not significantly different, but as yields were determined by hand harvesting, the standard error was high.

Table 5.2 Response of weeds and wheat in two ADAS experiments to treatments comparing: 24cm vs 12cm rows; none, full and selective herbicide treatments in autumn; none, selective herbicide and tine weeding in spring. (‘all’ = all weeds present, ‘spring’ = spring germinating species only)

X These SEDs refer to the Log10 transformed date (in parentheses)

Treatments and experimental detailsSowing date 27 Sept 01 8 Oct 02Tine weeding 24 April + 3 May 2 AprilHerbicides Autumn full triallate (27 Sept) pendimethalin +

(IPU + DFF) (26 Nov) flufenacet (31 Oct)DFF + IPU (7 Feb)

Autumn selective carf. + CMPP (20 Dec) carf.+ CMPP (7 Feb)Spring selective carf. + CMPP (19 Apr) carf.+ CMPP (22 Apr)

IPU = isoproturon, CMPP = mecoprop-p, DFF = diflufenican, carf = carfentrazone

Expt 03 Rothamsted In this experiment all plots received an autumn herbicide treatment. Neither the full nor the selective treatments were fully effective on the broad-leaved species present, although they did control the

SID 5 (2/05) of 40

annual grass weeds quite successfully. As a consequence, there was an appreciable infestation of weeds on all plots, especially those that received the selective autumn treatment. Neither of the spring treatments achieved high levels of control of all species present (Table 5.1). V. persica was particularly abundant in the summer on all treatments. Yields were low on all treatments, partly due to the degree of weed competition, and partly to the dry summer and the light soil on which the experiment was situated. There was a modest flush of spring emerging weeds, primarily C. album and Fallopia convolvulus (black-bindweed). Most were present on plots not treated in the spring and least where amidosulfuron had been applied in spring. Tine weeding did not stimulate extra weed emergence, nor did sowing the crop on wide rows.

Expt 03 ADASWeed levels were again low on this experiment on all treatments. Some spring emerging species were recorded (mainly F. convolvulus) but on 1 July the mean plant number was less than 2/m2 on most plots (Table 5.2). Weed numbers were little affected by treatments and yields were only lowered by the tine weeding treatment. Wide rows did not have this effect.

Expt 04 RothamstedThis experiment differed in design from the other four as it also included treatments to establish the role of spring cropping. Additionally, the natural seedbank was enhanced with supplementary C. album and P. aviculare seeds broadcast in the autumn.

Table 5.3 Response of weeds to spring and winter wheat (wide and narrow rows) and to spring weed control treatments – Rothamsted 04 (‘spring weeds’ = emerging in spring)

Treatments Density of spring weeds (plants/m2)

weed biomass (g/m2) July

Yields (t/ha)

Row width (cm) and crop

Spring weed control

P. aviculare

C. album

Spring weeds

Other weeds

12 (winter wheat)

None 14.0 0.5 2.9 54.2 7.09Selective 9.0 4.5 0.9 64.8 7.03Tine 9.0 1.0 2.5 35.5 7.52

24 (winter wheat)


12 (spring wheat)


SED 3.96 9.08 3.01 18.84 0.556

Populations of P. aviculare plants were established in the winter wheat but were less abundant in the spring wheat (Table 5.3). The low densities in the latter were, it is thought, linked to the destruction of germinating seedlings by the cultivations prior to the sowing of the spring wheat. The response of the C. album was the reverse, with more being present in the spring wheat. Overall biomass of these spring emerging weeds together did not differ between the treatments. The other weeds present on the plots, mainly S. media, P. annua and S. vulgaris, weighed appreciably more than the desired C. album and P. aviculare (Table 5.3), but again there were no statistically significant differences between treatments. The experiment did not detect benefits, as far as spring emerging weeds were concerned, from the wide rows or from the tine weeding. The spring wheat yielded less well than the winter wheat. The reasons for the contrasting behaviour of the two target weeds seems to be associated with their emergence patterns, as P. aviculare emerges earlier and has a shorter emergence window than C. album (Naylor, 2002). Consequently, the former could not respond to the favourable post-planting, spring, soil conditions whereas the C. album could.

Overview of experiments to stimulate greater emergence of spring germinating weeds in winter wheat

The five experiments in this part of the project demonstrated that it will be quite difficult to stimulate higher emergence of the desirable spring germinating weeds in winter wheat. There seems to be two basic reasons, firstly competition from the winter crop in early spring, for light and water, severely restricts resources for any seeds that do germinate at this time, so that they fail to establish, or remain

SID 5 (2/05) of 40

very small. Growing wheat on wider rows may help somewhat, as there were indications of such an effect in the two experiments in 2001/02. This aspect has also been explored in the SAFFIE project and that project also found little difference in weed cover between weeds in wheat sown on conventional and wide spaced rows (Jones & Boatman, 2004, Morris et al., 2004). Secondly, soil disturbance which is a fundamental stimulus for weed emergence was not very successful when simulated by tine harrowing/weeding in the spring. This may have been linked to our inability to disturb the soil adequately, because of the hardness of the soil surface in the dry springs that were experienced in 2002 and 2003. Another complication with this treatment was that the greater the vigorousness of the soil disturbance, the greater the crop damage and the greater the risk of yield loss. The final experiment demonstrated rather well that if C. album is to be present, the crop should be planted in spring. An experiment in Commission AR0409, designed to estimate seed production by P. aviculare in winter and spring wheat also clearly demonstrated the lack of vigour of this weed in winter wheat. P. aviculare plants in the spring wheat were 5-fold heavier (from biomass (g/m2) measured in July) than those in winter wheat, even though there were more plants in the winter wheat. One must conclude that if a flora based on spring-emerging species is required, then this has to be based on spring cropping and cannot be effectively established in autumn sown cereals.

6. Tactical use of herbicides to selectively remove undesired species and leave desirable species (Objective 1) – experiments 2003-05

The work in 2001-03 (chapters 3 & 4) had indicated that it was quite difficult to reduce the dose of a herbicide that was highly active on a particular weed species at its full rate, for it to be reliably less active on the same species. It seemed that it could well be more fruitful to explore the weed sensitivity spectra of commonly available herbicides to pick those products that were intrinsically less active on the species that were not to be controlled. Experiments reported in this section were carried out in the seasons 2003-2005. Additionally, the experiments provided a further opportunity to monitor the tolerance of winter wheat to competition from any surviving desired weed species and to continue to collect data on the seed production and seed viability of the partially controlled plants. As the main target for broad-leaved weed herbicides in winter wheat, in spring, is G. aparine, the field experiments have focussed on products used to control this species.

The 2004/05 research at ADAS explored a specific and highly relevant aspect of this approach: how to ensure the survival of desired weeds in a situation where the dominant weed was herbicide resistant A. myosuroides (black-grass). The work endeavoured to establish the potential to manage this pernicious grass weed and at the same time retain desirable broad-leaved species.

Materials and methods

Expts 2003/04 RothamstedTwo experiments (WW409, WW410) studied the management of T. inodorum in winter wheat. In both experiments a series of herbicides were applied in spring to manipulate the levels of this weed, and its survival and effects on crop yields were monitored. T. inodorum seeds were broadcast immediately after sowing the wheat on 29 September 2003. A ‘weed free’ treatment was created on both experiments through application of the recommended doses of a mixture of flupyrsulfuron and pendimethalin on 9 October. The other treatments all received an overall treatment of clodinafop-propargyl (plus adjuvant) on 8 December to control the grass weeds, mainly A. myosuroides. The following spring a range of herbicide treatments were applied to manage the target species and the other species present (Table 6.1). Owing to the presence of appreciable amounts of grass weeds in both experiments and Cirsium arvense on experiment WW409, intensive hand weeding was done in May.

Densities of weeds were assessed during the winter and a biomass sample was taken on a sample of plots immediately after the herbicides had been applied. A full sample to assess crop and weed biomass was taken on 14 June (Expt WW410) and 1 July (Expt WW409), based on 0.5m2

samples taken from each plot. The numbers of T. inodorum capitula were counted in ten 0.25m2

quadrats on 11 August, just prior to harvest. Final crop yields were determined from 2m2 areas hand harvested from each plot. The harvested material was processed through a static thrasher and grain yield determined.

SID 5 (2/05) of 40

Table 6.1 Details of herbicide treatments at Rothamsted in experiments 2003/04 (there were 4 replicates of each treatment laid out in randomised blocks)

Treatments Experiment WW409

Experiment WW410

Untreated Weed free (flupyrsulfuron+ pendimethalin) (Lexus + Stomp)

fluroxypyr (Starane) (400 g ai/ha) (100 g ai/ha mecoprop-p (Duplosan) (690 g ai/ha) (168 g ai/ha) carfentrazone + mecoprop-p (Platform S) (615 g ai/ha) (154 g ai/ha) metsulfuron (Ally) (0.4 g ai/ha) (0.1 g ai/ha)

Expts 2003/04 ADASTwo experiments were done by ADAS Boxworth in 2003-04, one to study the retention of S. vulgaris and the other T. inodorum. Winter wheat was sown on both experiments on 3 October, but because of dry soil conditions emergence was delayed. Seeds of the desired weeds were broadcast onto a proportion of plots when the crop was planted (Table 6.2) To provide a further test, oilseed rape seeds (Brassica napus) were also broadcast onto the plots to provide an infestation of an undesired weed that would need to be controlled.

Details of the treatments are given in Table 6.2. The pendimethalin was applied on 13 October and the mecoprop-p and amidosulfuron treatments in the S. vulgaris experiment on 23 January and 17 March, respectively. All other treatments, including the mechanical weeding (using a tine weeder) were applied in mid April. Additionally, all plots were treated with clodinafop-propargyl to control the A. myosuroides. However, some of this weed survived the clodinafop on the T. inodorum experiment and so required further treatment. Propoxycarbazone-sodium was applied to the experiment on 2 March, at 42 g ai/ha (60% of the full rate), to minimise the A. myosuroides.

The plots were monitored at intervals and a major sampling was done on each experiment in July (S. vulgaris 28 June – 2 July: T. inodorum 12-14 July). Plots were harvested with a small plot combine, and yields determined on 31 August-1 September.

Table 6.2 Details of weed control treatments at ADAS in experiments 2003/04 (there were 4 replicates of each treatment laid out in randomised blocks)

Herbicide treatments Experiment S. vulgaris

Experiment T. inodorum

Sown Unsown Sown UnsownUntreated

pendimethalin 1.32 kg ai/ha (Stomp)

carfentrazone + mecoprop-p 75 g ai/ha plus mecoprop-p 900 g ai/ha (Platform S +Duplosan)

amidosulfuron 15 g ai/ha (Eagle)

amidosulfuron 30 g ai/ha * *mecoprop-p 600 g ai/ha (Duplosan)

florasulam 7.5 g ai/ha (Boxer)

florasulam 3.75 g ai/ha

mechanical weeding (at crop GS30)

* amidosulfuron rate 22.5 g ai/ha

Expts 2004/05 Rothamsted

Expt WW515 The main experiment at Rothamsted in 2004/05 (WW515) was aimed at selecting post-emergence herbicides that could be used to retain more desirable weeds such as S. vulgaris and T. inodorum and at the same time eliminate competitive species like G. aparine. To achieve this, a

SID 5 (2/05) of 40

spectrum of weed seeds was broadcast onto the 3x9m plots at drilling (S. vulgaris, T. inodorum, G. aparine. S. media). One treatment was not seeded so it could act as a weed free control. There were four replicates. Winter wheat was sown on 29 September. Weed emergence was monitored and densities assessed once peak emergence had been reached. Species indigenous to the site were also recorded. A range of herbicide treatments was applied, primarily in the following spring, aimed at eliminating the G. aparine and retaining some of the desirable species (Table 6.3). The ‘weed free’ plots were also treated in autumn. Full details of the treatments are given in Appendix 6.1.

Main weed assessments were quadrat counts to determine plant densities for the main species in November/December, a biomass sample for S. vulgaris on 16 May and a full biomass sample on 28 June. In these samples 0.5m2 areas were harvested. T. inodorum capitula were counted in 16 x 0.1m2

quadrats/plot on 29 July. Wheat yields were recorded on 18 August by a small plot combine harvester.

WW516 This experiment aimed to manage a spectrum of weed species, such that aggressive members of the Cruciferae were minimised, along with G. aparine. Additionally, the aim was to retain some of the desired species, such as S. media and T. inodorum. Plots (8x12m) were sown with wheat on 22 Sept 04. The weeds studied were those present in the natural flora of the field. Three herbicide treatments were applied on 1 April (Table 6.4). Treatments were chosen for their activity on G. aparine and on members of the Cruciferae. There were four replicates.

Weed emergence was recorded on 8-9 December and plant biomass was calculated from 2 x 0.5m2 samples collected from each plot on 8 June. Wheat yields were recorded from a small plot combine harvester on 18 August. Expt 2004/05 ADASThere was one large experiment at ADAS Boxworth in 2004/05. This was established to study the potential for less intensive management of broad-leaved weeds in a field also infested with herbicide resistant A. myosuroides. This had been an issue in the T. inodorum experiment on the same farm in 2003/04 (see above). Desirable weeds S. vulgaris, T. inodorum and S. media were broadcast at drilling on 13 treatments. An undesirable species G. aparine was also broadcast onto these plots. Three further treatments intended to assess the responses of the indigenous flora received no extra seeds. The crop was drilled on 27 October. The herbicide treatments consisted of two categories, autumn treatments to control A. myosuroides and spring treatments to control surviving broad-leaved weeds. The latter was factorialised with the autumn treatments. Full details are given in Appendix 6.2 but information on the herbicides used is included in Table 6.5.

Weed infestation levels were counted on 19 January and a full sample of crop and weed biomass was taken on 20-23 June (2 x 0.25m2 samples/plot). Further assessments were done of S. media seed production (from sub samples of the material collected on 20-23 June) and for T. inodorum capitula number from samples collected on 3 August. Crop yields were determined on 16 August.

Results

Expts 2003/04 Rothamsted

Expt WW410 In this experiment the full weed control treatment successfully eliminated most weeds. The T. inodorum was greatly reduced by the highest dose of fluroxypyr but was only suppressed by the other treatments (Fig. 6.1). S. media was well controlled by most treatments apart from the lower rate of mecoprop-p alone and with carfentrazone. All treatments apart from the ‘weed free’ failed to suppress all the other weeds, P. annua being particularly abundant, but V. persica, A. myosuroides and Anisantha sterilis were also present.

Wheat yields reflected the degree of weed suppression achieved by the treatments. The weed free plots reached a yield of 8.6 t/ha (85% moisture) but as the total weed biomass recorded in June increased, yields decreased. In Figure 6.2, percentage yield losses are plotted against weed biomass recorded in June (this includes the other weeds, as well as the T. inodorum). A significant exponential response was identified for this experiment but there was a substantial amount of plot to plot variability and the % variance accounted for by the regression was only 34%. The presence of appreciable amounts of weeds other than T. inodorum has made the quantification of the impact of survivors of this species on crop yield hard to establish but as the quantities of other weeds were approximately equivalent across the treatments (Fig 6.1) one can surmise that the first approximately 5% of the yield loss was due to these weeds and not to the T. inodorum. Consequently, T. inodorum surviving the highest rate of fluroxypyr and mecoprop-p caused only a 2-4% yield loss. But the experiment does

SID 5 (2/05) of 40

show that when retention of desirable weeds is the target, undesired species must be managed effectively.

Figure 6.1 Biomass of weeds recorded on Experiment WW410 on 14 June, following application of the eight treatments (vertical bars = 1 x sed)

Figure 6.2 Relationship between total weed biomass in June and percentage yield loss for experiments WW409 and WW410 (response curve relates to WW410 only)

Figure 6.3 Relationship between T. inodorum biomass in June/July and number of capitula produced prior to harvest for Expts WW410 and WW409 (response curve is based on the raw data from both experiments)

SID 5 (2/05) of 40

The final assessment was the measure of seed production by the T. inodorum, based on the pre-harvest capitula counts. This showed a strong relationship with the biomass of the plants earlier in the summer (Fig 6.3). It seems that as the response is hyperbolic, that there was a high degree of intraspecific competition, which reduced flower production when infestations were high. As can be seen, the data from WW409 fit well with those from WW410. In WW410 the treatments that reduced yield the least (fluroxypyr and mecoprop-p at the higher rate) produced 108 and 256 capitula/m2 and with the estimated 230-300 seeds/capitulum, potential seed production would be 25,000-72,000 seeds/m2. This is a theoretical value as other work in 2005 concluded that only 14% of capitula had shed their seeds and a further 12% were nearly mature when the crop was harvested. Consequently, actual seed shed would only be approximately 4000-10000/m2, still an appreciable number of seeds. Both capitula and seeds provide resources for invertebrates (see report AR0409). Seed production/plant was calculated from the August sample of 15 single plants/plot. There was a strong linear relationship between plant weight and potential seed production, as had been shown in previous studies (Lutman 2002) (Fig 6.4). There was no evidence that herbicide treatment was affecting the biomass / seed production relationship, although plants on the untreated were bigger than those on the treated.

Figure 6.4 Relationship between plant weight and potential seed production by T. inodorum for three treatments of WW410. (Regression line is that generated from all 8 treatments)

Expt WW409 This experiment was in some respects less successful, as other weeds posed greater problems and the ‘weed free’ treatment had appreciable quantities of A. sterilis and A. myosuroides, despite the herbicide treatment and some hand weeding. Only the high dose of carfentrazone + mecoprop-p achieved some weed suppression (Fig 6.5). The other treatments were ineffective. As there was no true weed free plots, owing to the grass weeds on the intended weed free treatment, overall yields were lower than on WW410 and additionally, it was not possible to calculate true % yield losses from the experiment. Approximate values are presented in Fig 6.2, based on the use of the weed free yields from the adjacent WW410. The % yield losses from weed competition appeared a little lower than those of WW410 but followed a very similar trend.

Capitula and seed production by the T. inodorum plants matched that of WW410 very closely both in respect to the relationship between biomass and capitula number (Fig 6.3) and in relation to potential seed production. The formula for the log/log regression line of biomass and seed production was y = 0.68x + 3.20, which is very close to that of WW410 (Fig 6.4)

These two experiments demonstrated that management of the flora of winter wheat to retain T. inodorum was possible, but there were problems, particularly in relation to the management of grass weeds, as the graminicide used was not sufficient to manage all the grasses. Consequently, the value of the results from the range of broad-leaved weed control treatments was to some extent compromised. Certainly WW410 showed that higher doses of fluroxypyr and mecoprop-p could be used to minimise other broad-leaved species and retain the T. inodorum. In WW409 the doses of metsulfuron were too low. They had been selected from previous work on the basis that they would be partially effective on T. inodorum, but on the trial these treatments had little effect on any of the major weeds. This further confirms concerns about lower doses discussed in chapters 3 and 4.

SID 5 (2/05) of 40

Figure 6.5 Biomass of weeds recorded on Experiment WW409 on 1 July, following application of the five treatments (vertical bars = 1 x SED)

Seed production by the surviving T. inodorum plants was appreciable and was in line with previous data. However, the results predict potential seed production rather than actual production. Data from 2005 indicates that only 14% of capitula were viable at crop harvest. This information needs to be factored into further considerations of the impact of seed production from surviving weeds. There was no evidence that the herbicide treatments were markedly affecting the biomass/seed production relationships, though after some herbicides the treated plants were appreciably smaller. The presence of T. inodorum in wheat crops provides a resource for invertebrates but it can pose a threat to crop yield and may increase grain moisture and adversely affect harvesting operations. However, effects can be relatively low (yield losses 2-5%) provided other weeds are managed effectively.

Expts 2003/04 ADAS

Expt S. vulgaris Both oilseed rape and S. vulgaris established on the sown plots and the control of the grass weeds was adequate. Weed levels on the treated unsown plots were low. By the end of June treatment difference were clear, the untreated sown (and mechanically weeded) plots had appreciable quantities of S. vulgaris and oilseed rape (Fig. 6.6). Pendimethalin had had little effect on the S. vulgaris, whereas the other herbicide treatments (amidosulfuron and mecoprop-p +/- carfentrazone) had reduced weed levels substantially. All treatment had controlled the volunteer rape, except the mechanical weeding.

Figure 6.6 Biomass of weeds (Log10 transformed) recorded on the ADAS S. vulgaris experiment on 30 June, following application of the eight treatments (vertical bars = 1 x sed)

SID 5 (2/05) of 40

Figure 6.7 Relationship between total weed biomass in June/July and percentage yield loss (ADAS S. vulgaris expt 2003/04)

Yield losses tended to reflect the weed infestations recorded earlier, increasing as the weed infestation increased (Fig 6.7). Overall losses were relatively low and for the more effective treatments no yield loss could be detected. Yields of the weed free treatments were approximately 9.45 t/ha. The treatment leaving the most S. vulgaris, pendimethalin resulted in a 4.7% yield loss. This treatment, using the allometric relationship between plant weight and seed production produced in project AR0409, would have produced about 60,000 seeds/m2. As the mechanical weeding treatment was almost ineffective the yield loss after this treatment was similar to that from the untreated plots.

Expt T. inodorum As mentioned in the materials and methods, this experiment needed overspraying with propoxycarbazone in the early spring to eliminate surviving A. myosuroides plants. Most plants of this weed were killed but the treatment also eliminated all the volunteer rape and may have reduced the vigour and numbers of T. inodorum, as densities on the otherwise untreated plots declined from 156 plants/m2 in January to only 83/m2 in April. Later in the season, all treatments, apart from pendimethalin virtually eliminated the T. inodorum. Thus, all rates of florasulam, amidosulfuron and mecoprop-p (Table 6.2) controlled this weed. Weed biomass in July on the untreated plots, mostly T. inodorum, only reached 44g/m2, much lower than that recorded on the other 2004 ADAS experiment. Little weed was present on the other plots, apart from the ones treated with pendimethalin. Yields were all in excess of 10t/ha with evidence of a 4-6% yield increase from achieving full weed control. Pendimethalin did not control the T. inodorum, reflecting this product’s low activity on members of the Compositae seen in the previous experiment, in the tolerance of S. vulgaris to this herbicide.

Expts 2005 Rothamsted

Expt WW515 All the sown weeds, with the exception of S. vulgaris established well and these, together with a high indigenous population of some of the sown weeds, and Viola arvensis and Veronica persica, resulted in a very substantial infestation of broad-leaved weeds, approaching 900 plants/m2. Details of individual species are given in Table 6.3. This infestation posed a significant challenge for the herbicide treatments. The early samples taken of S. vulgaris showed that the sulfonyl urea-based products including florasulam (with the exception of amidosulfuron) had achieved full control, but some plants were still present on the other treatments (Table 6.5). At the later assessment this weed had senesced. Very large amounts of T. inodorum were present after the same treatments that were least effective on the S. vulgaris. Some reductions had been achieved by fluroxypyr, mecoprop-p and amidosulfuron, whilst the sulfonylurea mixtures and sequences, and florasulam had achieved virtually full control of this species. The aim of the trial had been to eliminate the G. aparine, and despite the very high infestation all except mecoprop-p and the metsulfuron/carfentrazone/mecoprop-p mixture had achieved good, if not total, control (Table 6.3). Amidosulfuron failed to control the S. media, whilst florasulam, fluroxypyr and the metsulfuron/ thifensulfuron/fluroxypyr mixture were less active on both V. arvensis and V. persica. The trial had focussed on broad-leaved weed control and although A. myosuroides was not a serious problem on most plots (one plot treated with fluroxypyr had a large amount) P. annua was very abundant across the experiment.

SID 5 (2/05) of 40

Untreated

Overall weed control was not adequate on most plots. Only the plots on the weed free treatment and those treated with metsulfuron/carfentrazone/mecoprop-p approached an acceptable degree of control. The substantial amounts of T. inodorum on many plots reduced yields appreciably and where this was well controlled, other species such as P. annua and V. persica compensated and so most treatments resulted in unacceptably high yield losses. From an environmental viewpoint many treatments resulted in the production of a lot of seeds of T. inodorum but few were recorded at the end of July on the plots treated with the metsulfuron-based mixtures, nor on those treated with florasulam (Table 6.3). Conversely, the cinidon/mecoprop-p and the untreated plots had over 1000 capitula/m2.

Table 6.3 Response of eight weed species and crop yields to ten treatments in Rothamsted experiment WW515. Biomass data collected on 28 June (apart from the S. vulgaris (see text)). Highlighted cells are those with appreciable survival of the studied weeds. Numbers immediately below the weed species are the densities (plants/m2) recorded in autumn 2004.

SENVU = S. vulgaris: TRIIN = T. inodorum: STEME = S. media: GALAP = G. aparine: VIOAR = Viola arvensis: VERPE = Veronica persica: POAAN = P. annua: ALOMY = A. myosuroides

This experiment exemplified the problems of endeavouring to achieve selective weed management. In this case, because weed pressure was so high, treatments aimed to be partially effective on the desired weeds gave inadequate control and so the crop yields were severely reduced. It also shows how control of one species provides space and resources for others. Thus, P. annua (another desirable weed that is generally not very competitive) was hardly present on the untreated plots but was most abundant on the plots aimed to be free of broad-leaved species. The treatments had successfully controlled the very high infestation of G. aparine (70 plants/m2) but had not managed to suppress the other weeds adequately, especially the T. inodorum. Where the sown weeds were well controlled, the indigenous V. persica and P. annua became more serious competitors. The experiment does clearly illuminate the strengths and weaknesses on the weed spectra of these products and mixtures, and thus their potential for more targeted weed management.

Expt WW516 A range of broad-leaved weed species were present on this experiment, along with some grass weeds (P. annua, A. myosuroides) (Table 6.4). The grasses were not well controlled by any of the three treatments, which were aimed at controlling broad-leaved weeds, especially G. aparine. Two of the three Crucifer weeds present (S. arvensis, C. bursa-pastoris) were well controlled

SID 5 (2/05) of 40

by all the treatments but amidosulfuron was less effective on Raphanus raphanistrum. All three treatments controlled T. inodorum, Geranium molle and S. media but amidosulfuron was weak on the latter and failed to control V. arvensis. Yields reflected the level of weed control achieved, with the untreated having a significantly lower yield and ioxynil, bromoxynil and mecoprop, having the highest. Thus, in the context of delivering biodiversity amidosulfuron has potential to retain some weed species, but there is a risk of lowering crop yields.

Table 6.4 Treatments and weed biomass responses (g/m2) plus wheat yields for experiment WW516 Rothamsted 2004/05. (there were 4 replicates of each treatment laid out in randomised blocks) Shaded areas indicate treatments that were ineffective on the weed species.

Expt 2005 ADASThis experiment set out to investigate the possibilities for retaining desired broad-leaved weeds when occurring with A. myosuroides. It was established in a field with a known background of this weed but although appreciable numbers of seedlings emerged prior to drilling few emerged in the crop. However, this does not detract from the main aim of the experiment. Good infestations of T. inodorum and some S. media were established but the S. vulgaris failed. Significant seedling populations of G. aparine and S. arvensis, especially the latter, were also present. In January the untreated sown plots contained means of 28 S. media, 58 T. inodorum and 4.6 S. vulgaris/m2. Occasional individuals of these species were recorded on the unsown plots. There was also an overall population of 97 S. arvensis and 13 G. aparine plants/m2.

By the following summer it was clear that the two autumn treatments including pendimethalin + flufenacet and the sulfonylureas (iodosulfuron + mesosulfuron, with or without a tank mix partner of pendimethalin or trifluralin) had virtually eliminated all the broad-leaved weeds (Table 6.5). The clodinafop, propoxycarbazone sequence was less effective on the broad-leaved weeds. The untreated weed flora was dominated by a substantial infestation of S. arvensis, but on the ‘sown’ plots T. inodorum was also present, along with G. aparine and S. media. A restricted analysis of variance involving only these plots and those treated with clodinafop and propoxycarbazone (Log10 transformed), showed that the S. arvensis had been eliminated by all the autumn treatments and that the spring treatments had reduced the other surviving broad-leaved weeds, somewhat. The Log10 analyses (Appendix 6.3) indicated that of the spring treatments, amidosulfuron was the most active on the main weed species present.

There was some seed production by both S. media and T. inodorum on the majority of plots treated with clodinafop + propoxycarbazone (Table 6.6), as well as on the untreated plots. Maximum biomass and associated seed production for S. media was on the plots treated in the autumn only. This was probably because competition from S. arvensis would have suppressed the S. media on the completely untreated plots. Maximum production in winter wheat was in the region of 30,000 seeds/m2

(Table 6.6). The relationship between Log10 flower number and Log10 plant dry weight was good

SID 5 (2/05) of 40

(flower number = 1.06 plant wt + 1.76: r2 0.84) and is close to the overall response measured in previous experiments (Lutman, 2002: y = 1.12x+1.82). The herbicide treatments reduced plant biomass and consequently reduced flower and seed production. Florasulam was less effective than the other two spring treatments. This may have been a seasonal effect as this herbicide was less effective than usual in other trials at ADAS Boxworth. Amidosulfuron was surprisingly effective on S. media. For T. inodorum, flower and seed production was based on flower numbers recorded in early August. The untreated plots were estimated also to have the potential to produce nearly 30,000 seeds/m2. The spring treatments reduced this potential production appreciably, especially amidosulfuron (Table 6.6). As with the S. media there was a strong relationship between plant biomass and flower/seed production (Log10 seed number = 0.67 Log10 plant wt + 3.37: r2 = 0.90). This response is very similar to Fig. 6.4 but is higher than that recorded in earlier work (y =1.05x +2.75). This may be because these data are based on potential seed production, whereas the former data were based on actual seed collected. It is known from other assessments that only a minority of flowers are actually fully ripe when wheat is harvested, thus potential production is never reached.

Table 6.5 Response of four weed species to sixteen treatments in ADAS experiment 2004/05. Weed biomass samples (g/m2) collected on 20 June and wheat yields recorded (t/ha) on 16 August.

*SED based only on the shaded treatments as other treatments were virtually zero but data still tended to be skewed (Log10 transformed data presented in Appendix 6.3)

Wheat yields were dramatically reduced on the untreated plots by the S. arvensis, augmented by the other sown weeds, where present (Table 6.5). Yields were apparently not statistically significantly lower on all the other treatments. However, a restricted analysis, omitting the untreated plots, demonstrated a small but significant yield reduction from the treatments based on clodinafop followed by propoxycarbazone, compared to the other two autumn treatments. There were no effects from the spring treatments.

Environmentally oriented management of weeds in wheat in the presence of A. myosuroides is clearly going to be more difficult, as the results of this experiment demonstrated that two of the standard ‘cocktails’ of products aimed at successfully controlling this weed, will result in the elimination of most broad-leaved weeds, irrespective of whether or not the autumn treatments were followed by a spring treatment. The only grass weed treatment that resulted in some broad-leaved weed survival was the treatment of clodinafop and propoxycarbazone. As A. myosuroides did not become a major

SID 5 (2/05) of 40

weed on the trial, it is not possible to assess how well this treatment would have worked on this species, but practical experience would suggest it was likely to be less effective than the other autumn/winter treatments. However, in this experiment it clearly offered a set of treatments that would permit the survival of some broad-leaved weeds, together with production of some seeds, without greatly affecting crop yields.

Table 6.6 Effect of herbicide treatments on flower and seed production by S. media and T. inodorum on the ADAS experiment 2004/05

Autumn treatment

Spring treatment

S. media T. inodorumBiomass

/m2Flowers /g DWt

Seeds /m2

Biomass/m2

Flowers /g DWt

Seeds /m2

Nil 26.7 76.7 19262 17.2 6.19 29653Nil (unsown) 13.8 66.3 8629 0.1 5.84 20479clodinafop + propoxycarb-azone

fluroxypyr 19.9 3.8 75 1.2 10.61 9591florasulam 38.3 53.3 19178 7.7 7.70 16217amidosulfuron 8.5 61.4 4895 0.2 11.21 3614None 44.3 70.7 29434 10.4 7.76 19645

7. Role of spatially selective weed control in the achievement of the retention of desired weeds within arable fields (Objective 1)

An alternative technique to achieve less intensive weed management would be to restrict herbicide application to those parts of fields where the most aggressive weeds were present. Most weeds are not uniformly distributed in fields. If treatments were only applied to weed patches, untreated areas, where weed infestations were less aggressive, could be left to contribute to the biodiversity of the field. Spatially selective weed control techniques have been in development over the last 10 years, associated with the increasing precision of global positioning satellite (GPS) systems (Miller, 2003). A LINK research project involving Rothamsted and Silsoe Research Institute developed a patch spraying system (Lutman et al., 2002) with the industrial sponsors providing the sprayers with the GPS and spatially selective application facility. These sprayers were available at the start of this project.

Two experiments were carried out to explore the potential to ‘patch spray’ target weeds. Both used the AGCO SpraCoupe self-propelled sprayer that had the potential to turn the whole boom on an off as directed by the weed map. The first investigated spatially selective control of patches of G. aparine, whilst the second studied the control of patches of A. sterilis.

G. aparine patch spraying experiment 2003. An area of a field at Rothamsted was mapped in February 2003 to assess the distribution of G. aparine, using GPS mapping system. This map was used to control the application of amidosulfuron on 1 April by the SpraCoupe sprayer. Four transects 110m long, were set out across the sprayed and unsprayed areas. Quadrats (1m2) were assessed every 15m. There were 11.5 G. aparine plants/m2 in the sprayed area but only 2.1/m2 in the unsprayed area, indicating that the herbicide had been applied to the weed patch area. Control of the weed was adequate, as G. aparine biomass at the end of May was only 4.7g/m2 in the sprayed area and 5.4g/m2 in the unsprayed, indicating that the higher, within patch, sprayed weed infestation had been well controlled. Very few other weeds were present, so it was not possible to assess the impact of treatment or no treatment on the ‘non target’ species.

A. sterilis patch spraying experiment 2004. A field at ADAS Boxworth was mapped in August 2003, just prior to wheat harvest, to locate the patches of A. sterilis. This map was used to determine the application of propoxycarbazone (70 g ai/ha) on 2 March 2004. At application four transects were set out c. 50m apart across the field, crossing sprayed and nonsprayed areas. A. sterilis and other weeds were counted in 1m2 quadrats placed every 8 m along the transects. The same places were counted again on 15 April. There were 1.8 A. sterilis plants/m2 pre-spraying on the sprayed area and this declined to 0.5/m2 after treatment. There were only 0.3 plants/m2 on the unsprayed area, indicating that the mapping had identified the core areas of the weed. Other weeds were present, especially Veronica hederifolia and V. persica. These too were recorded on the quadrats and declined from 34 plants/m2 to 5/m2 between the March and April assessments, with no differences in decline between the sprayed and non-sprayed areas.

SID 5 (2/05) of 40

Propoxycarbazone was not particularly active on broad-leaved species present and so the reason for their decline is not clear. The experiment showed that again the weeds were mapped effectively and the herbicide was applied to the correct areas but failed to demonstrate enhanced retention of broad-leaved species on the unsprayed areas.

Virtual patch spraying experiment 2004. Because of the difficulty experienced in the previous two experiments of establishing the value to biodiversity of using patch spraying to retain weeds in fields, when they were at densities that were not economically damaging, a specific ‘virtual’ experiment was established in 2004 at Rothamsted to compare the performance of florasulam, notionally targeted to control G. aparine, on other weed species present in sprayed and unsprayed areas. Eight plots were established and four were sprayed and four not. Weed levels were monitored on all plots pre- and post-spraying. The flora of the field was dominated by V. persica with very few other weeds present. Samples taken on 24 June showed a small, but significant, reduction in V. persica on the sprayed plots but few other weeds were present, apart from G. aparine (florasulam’s prime target species), which as expected, was less abundant on the sprayed plots. There were no differences in species number between the two treatments. It was disappointing that the site had such a poor weed flora and consequently the impact of the treatment on the wider weed flora could not be detected.

Overall, this series of experiments, with the collaboration of Silsoe Research Institute, did not fully deliver its anticipated outcomes. The work demonstrated that it was feasible to map weeds using GPS technologies and use the map to control the herbicide application from the sprayer, such that the required areas were treated. Thus, the technique was successful. However, in all three experiments we failed to demonstrate that spatially targeted treatments could deliver environmental value by retaining uncompetitive weeds in unsprayed areas. This was mainly due to the rather limited weed spectrum present on two of the experiments and unexplained decline in untreated weeds on the third. More resource was not put into this aspect of the programme because commercial uptake of the technology, anticipated in 2000 at the onset of the commission, has never materialised. The perceived need for automated detection of weeds, a technique not yet feasible for narrow row crops such as cereals, and the decline in farming profitability, have deterred commercial uptake of patch spraying technology. The basic philosophy is correct and, when the technology has wider commercial appeal, will be a tool that could deliver environmental gains with respect to the management of arable weeds.

8. Role of spring cropping in enhancing environmentally valued weed species (Objective 4)

The effect of cropping on the weed flora is well known, with spring crops having a flora that is different from autumn-sown crops. Some weed species can occur in both cropping types (e.g. S. media, P. annua), whilst others mainly emerge in autumn (e.g. G. aparine, Veronica hederifolia), and others mainly in spring (e.g. C. album, P. aviculare). This difference in emergence pattern is driven by the inherent dormancy characteristics of the species (Naylor, 2002). Weed infestations in spring crops depend on the competitive ability of the crop, the weed management practices and the nature of the weed flora. Maize and sugar-beet tend to contain few weeds, whilst spring rape and linseed tend to contain more. There is some evidence that the switch out of spring barley into winter wheat and winter barley that occurred in the UK in the 1970s and 1980s resulted in a decline in the arable weed flora. The winter crops are more competitive and there are more broader-spectrum herbicides available (Marshall et al, 2001). In the context of delivering environmental value, spring emerging species tend to be those that are associated with the greatest number of invertebrate species (Fig. 1.1).

A further benefit of spring cropping is that it provides an opportunity to leave an uncultivated stubble over the winter. Such stubbles are known to be a good source of food for overwintering birds, though their value depends very much on how intensive the weed control was in the previous crop: weedy crops provide better over-winter stubbles.

The limited amount of work in this project (and in the associated AR0407 and AR0409) confirms this perception. Section 5 in this report demonstrated that it is very hard to encourage the desired spring emerging weeds to emerge and grow satisfactorily in winter sown cereals. They much prefer to emerge in spring-sown crops. The experiments with P. aviculare seed production in winter and spring wheat (in Commission AR0409) showed that seed production by this weed was at least 8-fold greater in the spring crop. The modelling work in the Weed Management Support System has also shown that spring crops are needed if C. album and P. aviculare are to flourish.

SID 5 (2/05) of 40

More detailed modelling would be needed to predict the precise proportion of spring cropping required to deliver a more varied and environmentally valued weed flora, but work in Commission AR0407 has demonstrated quite clearly that different rotations involving a greater or lesser proportion of spring crops, can change the balance of weed species.

Table 8.1 Gross margins (2005) for a range of arable crops.

Crop Average Gross Margins £/ha (excluding single farm payment of approx. £254/ha)

Winter wheat 285Spring Barley/Wheat 215Spring Oats (Milling) 190Spring beans 220Peas 160Lupins 170Linseed 135From Nix, et al 2005

However, many spring crops are not as financially rewarding to the farmer as autumn sown ones (Table 8.1). The gross margins for spring combinable crops, tend to be lower than those of winter wheat. Consequently, there is a reluctance by growers to plant these crops. A further constraint is that on heavy land suited to winter wheat and oilseed rape, planting spring crops can be very risky, as rainfall at the ‘wrong time’ in the spring will prevent the creation of a good seedbed and consequently will jeopardise crop establishment. All these factors tend to discourage planting spring crops, especially under the new EU support regime, where it may now be financially more attractive to not plant a less profitable spring crop but leave the land unplanted.

9. Overview and conclusions of the research programme

This research project aimed to establish the feasibility of selectively managing environmentally desirable species of broad-leaved weeds in winter wheat. The prime species that were considered appropriate for such management, and hence retention within crops of winter wheat were S. media, T. inodorum, and S. vulgaris. The other two autumn germinating species P. annua and S. arvensis were not the prime focus for herbicide selection, though the former was present in several experiments and the latter occurred in some of the 2004/05 work. The remaining two species (C. album, P. aviculare) are primarily spring emerging species and were studied in specific relevant experiments. This species selection was based on the conclusions of the Defra project PN0940 (Fig. 1.1) (Marshall et al., 2001) and one of the issues that remains to be resolved is how useful these species actually are for invertebrates and birds. More work is needed to evaluate the spatial and temporal relationships between plants present in fields and the higher trophic groups, and how these are influenced by agronomic practices (time of sowing, cultivation etc). It is also unclear at this stage how large the weed infestation has to be to provide this ecosystem service. Such issues need to be resolved in order to understand fully the role that plants within fields play within the arable ecosystem.

A number of different techniques to enhance the broad-leaved weed flora within arable fields, were investigated in the project:

a) choosing appropriate herbicide products and dosesb) using mechanical weed controlc) manipulating production systems of winter wheat to increase spring emerging weedsd) exploiting spatial variation in weed distribution

Most emphasis has been put on selecting appropriate herbicide treatments, as this was considered to be the most likely method to be adopted by farmers endeavouring to retain some broad-leaved weeds in their fields for environmental purposes.

Herbicide choiceThe published information on the weed spectrum of some of the common herbicides used for broad-leaved weed control in cereals is limited (Table 3.1). This project could not possibly carry out enough field studies to establish conclusively the potential of the studied herbicides to manage the flora as required but it could indicate where there was potential to achieve this objective. The combination of

SID 5 (2/05) of 40

four pot experiments plus twelve field experiments has laid some sound groundwork and identified where the most fruitful avenues for development would be.

An overview of the conclusions of the field and pot experiments is presented in Table 9.1 and a fuller version is supplied in Appendix 9.1. Of the species included in Table 9.1, the herbicides failed most often on V. persica a species with little value for invertebrates and birds. Most products were developed to control G. aparine, which was generally well controlled. S. media was the most difficult of the ‘desired’ weed species to retain, with only amidosulfuron having no effect on this species. Control of the two members of the Compositae (S. vulgaris, T. inodorum) was variable, with mecoprop-p +/- carfentrazone and fluroxypyr, sometimes achieving reasonable suppression of these two species. Good control was achieved by florasulam and by amidosulfuron, whilst pendimethalin was virtually ineffective. Thus, amidosulfuron would be the preferred product if S. media was to be retained and pendimethalin for the two members of the Compositae. Mecoprop-p and fluroxypyr too would be appropriate for retention of the two members of the Compositae. The potential for achieving selectivity by reducing dose was explored in some of the early experiments with metsulfuron, in particular, but the results were not encouraging as the performance of the reduced doses seemed more variable than that of the full dose. Consequently, it was concluded that endeavouring to manipulate herbicide doses to achieve partial control, reliably, was unlikely to succeed and so the best approach was to select products with intrinsically low activity on the non-target weeds. Correct product choice can deliver appropriate selective control.

Table 9.1 Overview of the selectivities of the herbicides tested in this project in relation to the weed species to be retained or eliminated from winter wheat ( = well controlled, ? = variable control, x = no control)

Herbicide Desired species Unwanted species

Other species

S. media T. inodorum S. vulgaris G. aparine V. persicaamidosulfuron (Eagle) x ? xcarfentrazone + mecoprop-p (Platform S)

? ? x

florasulam (Boxer) xfluroxypyr (Starane) ? ? xmecoprop-p (Duplosan) ? ? xmetsulfuron (Ally) ? xpendimethalin (e.g. Stomp) x x

Two practical problems associated with this approach to broad-leaved weed management need to be considered. Firstly, if a product such as amidosulfuron is chosen as the main tool to manage broad-leaved weeds, it may ensure the survival of some of the desired species, but it may also permit the survival of undesired species too, such as V. persica. These survivors may reduce crop yields and may also compete strongly with the desired species, depending on their relative competitive abilities. Secondly, partial control may be successful when the infestation is moderate or low but can be inadequate if the infestation is high and result in unacceptably high yield reductions.

Several of the experiments clearly demonstrated links between weed survival and loss of crop yields (see Figs. 3.2, 6.2, 6.7). If all the data from the ten experiments that compared crop yields after application of a range of different herbicide treatments are combined, a significant relationship can be established: Y = 0.095 x – 5.52 (% variance accounted for = 64%): where Y = % yield loss and ‘x’ = weed wt (g/m2) in summer. Clearly such a generalised regression is vulnerable to different responses to different weed species, depending on their stature and phenology, and to variation in crop vigour, but it does give an indication of approximate yield responses to a range of weed infestations. It is not surprising that there is this inverse relationship, but such data do present quite clearly the key conflict between retaining weeds for their biodiversity value, and not jeopardising crop yields. There is some evidence from these data that the crop can tolerate modest levels of weeds, perhaps 50g/m2, without loss of yields. This conclusion may be an artefact of the analyses, as the errors in the yield estimates would mean that small yield losses were not detectable, but it could also be due to complementarity between the crop and weeds in the exploitation of resources. If the latter is correct, it provides a basis for demonstrating that weed suppression does not necessarily result in yield loss. Statistically significant differences in yield losses arising from the treatments were often in excess of 5%, because of the inevitable variability in the data from the experiments, which is often used as a ‘threshold’ of

SID 5 (2/05) of 40

tolerable reduction in yields from weeds, but with current crop values hardly exceeding £60/t, such a loss could well be the difference between the crop making a profit or a loss. So, careful risk assessment is needed whenever low herbicide doses are considered. Impacts of low weed infestations on harvesting efficiency, the need for grain drying and possible effects on following crops all also need to be taken into account. The latter is the emphasis both of commission AR0409 and of the rotational model of the weed decision support system (Weed Manager) (Collings et al, 2003). Crops can tolerate more weeds than they do in current crop production systems.

The final issue related to the use of herbicides to retain some broad-leaved weeds investigated in the project, was the impact of managing annual grass weeds such as A. myosuroides. On many farms the main driver of weed control decisions is the management of this pernicious annual grass weed. These weeds drive the initial selection of herbicides. In the 2005 experiment at ADAS Boxworth it was clear that an unintended consequence of the selection of many of the favoured products to control A. myosuroides, A. fatua and Lolium multiflorum, is the elimination of many broad-leaved species. The favoured products such as iodosulfuron, mesosulfuron and flufenacet + pendimethalin also controlled the non-target broad-leaved weed species include in the experiment. Only the products with a more restricted spectrum (clodinafop, propoxycarbazone) permitted the survival of some broad-leaved weeds. However, this more ‘benign’ sequence is known to be less effective on resistant types of A. myosuroides, making their use less attractive to many farmers. Thus, targeted management of broad-leaved weeds is going to be more difficult in fields where annual grass weeds are abundant.

There are two potential mechanisms that could be developed to resolve the dilemma of how to deliver biodiversity value without causing reductions in profitability. Firstly, the delivery of environmental value by retaining weeds in fields could be included within one of the Defra environmental schemes and as such would be appropriate for some form of compensatory payment. Secondly, once the spatial component of the delivery of within field biodiversity has been established then the proposed lower input broad-leaved weed management could be located in parts of fields or field headlands, instead of whole fields. Conservation headlands are already options in the current environmental schemes available to farmers and the work reported here could be used to fine tune the management of such areas.

Potential for mechanical weedingThe experiences of the researchers in this project with mechanical weeding were not very encouraging. Both the experiments to promote spring weeds (see below) and some of the experiments primarily involving herbicides, included tine weeding with an Einbock weeder. All the weeding treatments were done in late winter / early spring and the control of overwintered weeds was poor. It is likely that the treatments were applied too late, when the weeds were too well-developed to be vulnerable to uprooting by the weeder. Additionally, in the springs of 2002 and 2003, when much of the work was done, the soil was very dry and the tine weeder did not penetrate the dry and compacted soil adequately to remove the weeds. It should be pointed out that tine weeding should be done under dry conditions, otherwise uprooted weeds simply re-root. A final problem was that in our experience vigorous tine weeding in spring, reduced crop yields. From our experience tine weeding is clearly not an ideal tool to be used in spring to selectively manage broad-leaved weeds.

Encouragement of spring-emerging weeds in winter wheatThe project explored two techniques to achieve greater abundance of spring-emerging broad-leaved weeds in winter wheat; wide rows and tine weeding. As stated in the conclusions of chapter 5, neither technique was successful. Weed emergence in spring in winter wheat was always low and the plants that did emerge were of very low vigour. Consequently, if species such as C. album and P. aviculare are required to be present in crops, then it will be much more successful to encourage them within spring crops. Tine weeding in spring and wider rows are not effective at stimulating weed emergence in spring in winter wheat. Desirable spring-emerging weed species are more likely to be present in spring crops, even in winter crop dominated rotations.

Exploitation of spatial variability in weedsThere is much evidence that weeds are not uniformly distributed in fields and that they occur in patches. The project carried out several experiments to see if patch spraying, based on GPS maps, could be used to deliver environmental value. Although weed patches were mapped and sprayed with the support of Silsoe Research Institute, using the SpraCoupe patch sprayer, we failed to demonstrate differences in weed levels, post treatment, between sprayed and unsprayed areas. Thus although the

SID 5 (2/05) of 40

technique has, at least in theory, the potential to deliver environmental value, this was not demonstrated in the project. Emphasis on this aspect of the project was reduced as it became clearer that the potential for patch spraying, demonstrated at the end of the last millennium, was not going to be rapidly translated into purchases of commercial patch sprayers by farmers. Spatially selective weed control has potential but other constraints on the acceptability of the technique to farmers have to be resolved first.

Contribution of weeds to the arable ecosystemThe work in the project clearly demonstrated that environmentally beneficial weeds that survived lower-input weed management, were capable of producing high numbers of seeds. Thus, the surviving plants were providing leaf material and seeds for invertebrates and birds, within the crop and also in stubbles, post–harvest. Seed production from the herbicide treated plants seemed to follow the same allometric relationship between plant biomass and seed production, developed for non-herbicide treated plants, and data on seed viability (primarily presented in project AR0409) indicated no adverse effects on seeds, from earlier herbicide treatment. Modelling work in project AR0407 and included in the decision support system ‘Weed Manager’ have identified effects of increased seed rain on the population dynamics of the species concerned. Clearly, added seed rain can increase weed infestations in subsequent years, and this has to be considered when pursuing a lower input management approach, but sound rotational management can minimise any agronomic impact whilst retaining the environmental enhancement arising from the surviving plants and seeds.

The project has shown that management to achieve selective control of desired, biodiversity valued, weed species is possible but is not easy. Manipulating doses to improve selectivity seemed not to be a very fruitful technique but some herbicide products (e.g. amidosulfuron, fluroxypyr, pendimethalin) do seem to have a spectrum suitable for controlling unwanted species such as G. aparine, whilst having less effect on desirable ones. If spring emerging weed species are to be encouraged in arable areas, increases are not possible from changed management of autumn-sown crops (e.g. tine weeding in spring). Spring-sown crops are needed. Balancing retaining weeds for biodiversity value and not jeopardising the economic profitability of the crop is hard but not impossible to achieve. Overall analyses tentatively indicate that crops will tolerate low levels of weeds without an associated drop in yields. However, once this ‘threshold’ is exceeded yields will inevitably decline. A key unanswered question is how many weeds are needed within arable fields to provide the desired increase in the numbers of birds and invertebrates?

Future work

Key elements of future work emerging from this project are:a) There is a requirement to fine tune the herbicide regimes that could be used to selectively

retain desirable species and to confirm the potential selectivities emerging from the work in this commission. The issues of managing broad-leaved weeds for environmental value in fields where the flora is dominated by A. myosuroides needs more attention.

b) In order to establish the environmental benefits from retaining certain broad-leaved weeds in fields, there is a need for basic underpinning studies of the precise value of surviving weeds to higher trophic groups (e.g. invertebrates and birds), especially the spatial and temporal relationships. How many plants are needed to deliver environmental enhancement and where should they be located within a field (or indeed off the field)?

c) As retaining weeds as a resource for birds and invertebrates is likely to result in higher seed return, there is a need to further develop the population modelling techniques to establish rotational implications of alternative management practices. Such approaches already exist, and predictions are already being made (e.g. in Weed Manager) but the models need further validation and there are few parameters for some weed species, which need to be included.

d) It is clear from this project that there is little potential to encourage spring emerging weeds to emerge in winter wheat, and if these species are needed to meet environmental goals there is a requirement to establish the value of spring cropping within an arable rotation. Although some modelling work has been done, especially in relation to sugar beet, and parameter values are available for some species, more work is needed.

e) The delivery of environmental value needs to be put in a rotational and whole farm context. Winter crops and spring crops can each make a contribution towards enhancing diversity.

SID 5 (2/05) of 40

Similarly in–field and off-field management can also deliver components of the whole farm delivery of environmental value. The whole picture needs to be unified to establish an optimum environmental structure.

10. References

Christensen S (1994) Crop:weed competition and herbicide performance in cereal species and varieties. Weed Research, 34, 29-36.

Collings LV, Ginsburg D, Clarke JH, Milne AE, Parsons DJ, Wilkinson DJ, Benjamin LR, Mayes A, Lutman PJW & Davies DHK (2003). WMSS: Improving the precision and prediction of weed management strategies in winter dominant rotations. In: Proceedings of the BCPC International Congress – Crop Science & technology, Glasgow, UK, 1, 329-334.

Cussans GM & Courtney AD (1995) Cost-effective weed control in cereals. Home Grown Cereals Authority Project Report. 107, pp 105 , HGCA, London , UK.

Fuller RJ, Gregory RD, Gibbons DW, Marchant JH, Wilson JD., Baillie SR & Carter N. (1995). Population declines and range contractions among farmland birds in Britain. Conservation Biology, 9: 1425-1441.

Hawes C, Haughton AJ, Osborne JL, Roy DB, Clark SJ, Perry JN, et al. (2003) Responses of plants and invertebrate trophic groups to contrasting herbicide regimes in the Farm Scale Evaluations of genetically modified herbicide-tolerant crops. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences 358, 1899-1913.

Heard MS, Hawes C, Champion GT, Clark SJ, Firbank LG, Haughton AJ, et al. (2003) Weeds in fields with contrasting conventional and genetically modified herbicide-tolerant crops. I. Effects on abundance and diversity. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences 358, 1819-1832.

Jones N & Boatman N (2004) Effects of novel crop management techniques on weeds to improve biodiversity: preliminary results from the SAFFIE project Abstract Arable weeds and biodiversity AAB/BES Conference, 27-28 September 2004, York.

Kudsk P (1999) Optimising herbicide use – the driving force behind the development of the Danish decision support system. Proceedings The Brighton Conference, Weeds, 737–746.

Lutman PJW (2002) Estimation of seed production by Stellaria media, Sinapis arvensis and Tripleurospermum inodorum in arable crops. Weed Research 42, 359-369.

Lutman PJW, Perry NH, Hull RI, Miller PCH, Wheeler HC & Hale RO (2002) Developing a weed patch spraying system for use in arable crops. HGCA Project Report 291, 118pp , HGCA London, UK.

Marshall EJP, Brown V, Boatman N, Lutman PJW & Squire G (2001) The impact of herbicides on weed abundance and biodiversity. DEFRA Review PN0940, pp156.

Marshall EJP, Brown VK, Boatman ND, Lutman PJW, Squire GR & Ward LK (2003) The role of weeds in supporting biological diversity within crop fields. Weed Research 43, 77-89.

Miller PCH (2003) Patch spraying: future role of electronics in limiting pesticide use. Pest Management Science 59, 566-574

Morris AJ, Holland JM, Smith B & Jones NE (2004) Sustainable arable farming for an improved environment (SAFFIE): managing winter wheat sward structure for skylarks Alauda arvensis. Ibis 146, (Suppl. 2), 155-162)

Naylor REL (2002) Weed Management Handbook. Blackwell Publishing, Oxford, UK.Nix J, Hill P, & Edwards, A (2005) Farm Management Pocketbook, 36th Edition (2006), pub. Andersons, 263ppPreston CD, Pearman DA & Dines TD. (2002). New Atlas of the British and Irish Flora. Oxford University Press,

Oxford, UK. 910 pp.Siriwardena GM, Baillie SR, Buckland ST, Fewster RM, Marchant JH, Wilson JD (1998) Trends in the abundance

of farmland birds: a quantitative comparison of smoothed Common Birds Census indices Journal of Applied Ecology, 35, 24-43

Proven MJ, Davies DHK & Courtney AD (1993) An approach to developing appropriate herbicide options for winter wheat where cleavers are a problem. Proceedings Brighton Crop Protection Conference, Weeds, 1211-1216.

Richardson WG, West TM & White GP The activity and post-emergence selectivity of some recently developed herbicides: AC252214, DPX-T6376 and chlorazifop. Technical Report Weed Research Organisation, (74), pp 33

West TM (1994) The pre- and post-emergence activity and selectivity of the herbicide amidosulfuron (HOE 075032). Technical Report IACR – Long Ashton Research Station, (110), pp 25.

West, TM (1988) The activity and post-emergence selectivity of some recently developed herbicides: imazethapyr, BAS 51800H, DPX-L5300, triasulfuron, DPX-A7881. Technical Report IACR – Long Ashton Research Station, (104), pp 58.

Wright G McN, Lawson HM & Proven MJ (1993) Longer-term effects of reduced herbicide strategies on weed populations and crop yields in cereal rotations in England. Proceedings Brighton Crop Protection Conference, Weeds, 1229-1234.

SID 5 (2/05) of 40

Appendices

NB Numbers of Appendices refer to the relevant sections of the report

Appendix 4.1 Details of the weed species and herbicides used in the four pot experiments 2002-04

Year Weed species

Herbicide treatments

% ai or g/l

Maximumdosex

Dates of

g or l/ha sowing treatment harvest

2002 S. media amidosulfuron 75% 40 18 Dec 01

31 Jan 02 6 Mar 02bromoxynil+

ioxynil190+190 g/l

2

cinidon-ethyl 200 g/l 0.25fluroxypyr 200 g/l 2mecoprop-p 600 g/l 2.3metsulfuron 20% 30thifensulfuron + metsulfuron

68+7% 75

2003 S. media carfentrazone+ mecoprop-p

1.5+60% 1000 20 Jan 03

5 Mar 03 7 April 03mecoprop-p 600 g/l 1.5

metsulfuron 20% 7.5

2003 T. inodorum bromoxynil+ ioxynil

190+190 g/l

2 20 Jan 03

12 Mar 03 9 April

florasulam 50g/l 0.1 03mecoprop-p 600 g/l 3.0metsulfuron 20% 30thifensulfuron + metsulfuron

68+7% 75

2004 S. media amidosulfuron 75% 40 2-3 Feb 04*

1 Mar 04 17-19 Mar 04

S. vulgaris fluroxypyr 200 g/l 2S. arvensis mecoprop-p 600 g/l 2.3T. inodorum metsulfuron 20% 30

* Sinapis arvensis was resown on 12 Mar 04x maximum dose of product applied in the experiments

SID 5 (2/05) of 40

Appendix 6.1 Details of the treatments used in Rothamsted experiment 2004/05 WW515 and dates of applications

Treat-ments Products Active ingredients Dose

g/ha aiAppln Date

Weed free

Duplosan mecoprop-p (600 g/l) 1200 g /ha 9 NovQuantum DF tribenuron 75% (w/w) 15 g /ha 16 NovHarmony M + Starane 2

metsulfuron + thifensulfuron (7:68% w/w) and fluroxypyr (200 g/l)

45 g/ha + 100 g/ha 29 April

WB Boxer florasulam (50 g/l) 7.5 g ha 10 MarWE Eagle amidosulfuron (75% w/w) 30 g ha 10 Mar

WAD Ally Express + Duplosan

metsulfuron + carfentrazone (10:40% w/w) and mecoprop-p (600 g/l)

25 g /ha + 690 g /ha 10 Mar

WHS Harmony M + Starane 2

metsulfuron + thifensulfuron (7:68% w/w) and fluroxypyr (200 g/l)

22.5 g /ha + 150 g /ha 10 Mar

WD Duplosan mecoprop-p (600 g/l) 1380 g/ha 2 AprilWLD Lotus +

Duplosan cinidon-ethyl (200 g/l) and mecoprop-p (600 g/l)

50 g/ha + 690 g/ha 2 April

WP Platform S carfentrazone-ethyl + mecoprop-p (1.5:60% w/w) 615 g /ha 2 April

WS Starane 2 fluroxypyr (200 g/l) 200 g/ha 2 AprilW- untreated none

Appendix 6.3 Response of a range of weed species to several herbicide treatment ADAS Expt 2004/05. Weed biomass in summer (g/m2) - data transformed Log10 x +1 (see Table 6.5)

Autumn treatment Spring treatment

S. arvensis

T. inodorum

G. aparine

S. media Total weeds

Nil 2.78 1.05 1.18 1.35 2.87Nil (unsown) 2.55 0.06 0.18 0.95 2.60clodinafop + propoxycarbazone

fluroxypyr 0 0.30 0.58 1.00 1.10

florasulam 0.08 0.86 0.85 1.44 1.70amidosulfuron 0 0.09 0.02 0.79 0.81none 0.16 0.91 0.95 1.46 1.73

SED 0.156 0.199 0.270 0.320 0.264

SID 5 (2/05) of 40

Appendix 6.2 Treatment details for ADAS Experiment 2004/05

SID 5 (2/05) of 40

Appendix 9.1 Overview of the sensitivity of five main target and non-target weed species studied in the project ( = well controlled, ? = variable control, x = no control)

Herbicide ExperimentS.

mediaT.

inodorumS.

vulgarisG.

aparineV.

persica

amidosulfuron 02 RRes x x04 ADAS 04 ADAS 05 RRes x ? ? ?05 RRes ? Pot expts ? ?

carfentrazone + mecoprop 04 RRes ? x

04 RRes ?04 ADAS 05 RRes x x Pot expts ?

florasulam 04 ADAS 05 RRes xPot expts

fluroxypyr 03 RRes ? 04 RRes 05 RRes ? ? xPot expts x ?

mecoprop (2.3l/ha) 02 RRes x05 RRes ? ? ? ?Pot expts ?

mecoprop (1.15l/ha) 04 RRes x

04 ADAS ?

metsulfuron 02 & 03 ADAS 02 & 03 RRes 05 RRes ?Pot expts ?

pendimethalin 04 ADAS x x

SID 5 (2/05) of 40

References to published material9. This section should be used to record links (hypertext links where possible) or references to other

published material generated by, or relating to this project.

BOATMAN, N. & LUTMAN, P.J.W (2004) Assessing the risks of weed control on wider biodiversity. AAB/BES Conference, 27-28 September 2004, York.

LUTMAN P.J.W. 2001 Weeds and biodiversity Proceedings 38th Annual Review of Weed Control, BCPC, Farnham, 29-46.

LUTMAN, P.J.W., BOATMAN, N.D., BROWN, V.K. & MARSHALL, E.J.P. (2003) Weeds their impact and value in arable ecosystems. Proceedings BCPC International Congress - Crop Science and Technology, 219-226. Glasgow, BCPC.

SIMPSON, N., OSWIN, A., HULL, R. & LUTMAN, P.J.W. (2004) Targeted herbicide use in winter wheat for biodiversity and retention of yields. AAB/BES Conference, 27-28 September 2004, York.

SIMPSON, N.A, HULL, R.I., LUTMAN, P.J.W, & OSWIN A.L. (2005) The potential for managing weeds to minimise impacts on yields and enhance the ecological diversity of arable fields – a winter wheat case study. Abstracts EWRS Symposium 2005 (Bari).

Presentations

LUTMAN, P.J.W. Should we be leaving weeds in fields? (are there ‘good’ weeds?) Presentation at two HGCA Roadshows (winter 2003/04)

LUTMAN, P.J.W Weeds or Wild Plants: Positive and negative effects of weeds on production and the arable ecosystem. Presentation at two Abacus Organics meetings of Organic Farmers (April 04)

SID 5 (2/05) of 40

general enquiries on this form should be made...

Documents