Agronomic and Environmental Implications of the Establishment of GM Herbicide Tolerant

Final report:

Agronomic and environmental implications of the establishment of GM herbicide tolerant problem weeds

(project CPEC 45)

P.J.W. Lutman & K. Berry

(Rothamsted Research)

M.J. May & G.T. Champion

(Rothamsted Research – Broom’s Barn)

J.H. Clarke & S.K. Cook

(ADAS Boxworth)

BROOM’S BARN
Executive Summary

• This project has explored the biology of weed beet, volunteer rape and of several members of the Brassicaeae (e.g. Brassica rapa, Sinapis arvensis, Raphanus raphanistrum) and how their biology is affected by the presence of herbicide tolerance genes. Glyphosate and glufosinate tolerant sugar beet and oilseed rape, and imidazolinone tolerant (IMI) rape have all been considered. Models have been constructed of population dynamics of the three main species and predictions made of seed longevity. Impacts on adjacent crops and uncropped land, from gene flow from HT rape and beet have also been included.

• Hybridisation with wild relatives: Risks of hybridisation between oilseed rape and the Brassica weeds have been reviewed and the behaviour of hybrids assessed. It was concluded that B. rapa was the most likely species to hybridise with rape in the UK, and that the risks from the other species were much lower, such that changes in management to control them, were unlikely.

• Biology of weed beet, volunteer rape and rape/B. rapa hybrids: It was concluded from published data, primarily on oilseed rape, that HT weeds would be no more fit than non HT individuals. HT volunteer oilseed rape would be present in appreciable numbers for 6-10 years in fields that had been sown with HT rape crops. Hybrids between rape and B. rapa would be relatively uncommon and so it was concluded that management aimed at HT rape volunteers would also control any hybrids. There was some evidence that B. rapa seeds might be more persistent than volunteer rape, but the data were not really adequate to suggest the changed management might be needed for a longer period of time. Although, it was hoped that good management of bolting beet plants in HT beet crops would minimise the risk of the creation of HT weed beet, it was realised that eventually HT weed beet would occur, and management options were based on this premise. Weed beet seeds could survive longer than 10 years.

• Gene flow: Data on gene flow indicated that the risk of HT genes transferring to adjacent crops was low, especially if isolation distances proposed by Defra in their recent consultation document were implemented. Consequently, we did not think that management should be changed on adjacent fields, though some monitoring for unexpected effects would be appropriate. However, it was possible that very low levels of HT genes would be present, well below current EU thresholds (0.3-0.9%).

• Management options These have been developed for the control of the three HT weeds (weed beet, volunteer rape and rape/B. rapa hybrids). All major possible arable crops that would be grown in rotation with rape and sugar beet have been assessed, along with set-aside and between crop management options. The presence of glufosinate tolerant weeds was predicted to have virtually no impact on subsequent crop management. Management of glyphosate tolerant weeds would require some minor modifications to set-aside and between crop management for both rape and beet. Control of glyphosate tolerant weed beet in other crops would be unaffected. Some caution should be used when planting winter and spring field beans in rotations with herbicide tolerant rape volunteers, as control is not always adequate. Farmers should pay increased attention to their weed control in this crop. The presence of imidazolinone tolerant rape would preclude the planting of linseed and sugar beet because of reliance on related products in these crops for the control of this weed. Volunteer rape, weed beet or B. rapa individuals present on uncropped areas, such as field margins would need to be managed, both when the HT crop was being grown and thereafter. This could be by hand-cutting, pulling, or the use of selective herbicides.

• Environmental impacts: Two assessments have been done of the environmental implications of the changed management needs. Firstly, the environmental consequences of changing herbicide products have been assessed using the risk assessment tool, p-EMA, which calculates site specific values for each herbicide product in terms of eco-ratings and ecotox scores. These ratings combined all the toxicological data on the effects of the rates of products used on plants, invertebrates and vertebrates. Secondly, the consortium undertook an assessment of the environmental impacts of the management changes (cultural practices, rotations etc) anticipated for the management of the HT weeds, compared to the non HT ones. Potential impacts on the plants, invertebrates and birds were rated for the different management tools. The environmental impacts of all the herbicides that would be needed to control non HT and HT weeds were uniformly low (well below the p-EMA acceptable eco-rating of -40). So potential changes of intercrop management herbicides from glyphosate to glufosinate or paraquat/diquat (depending on the HT status of the weeds) had little effect. Similarly, where it would be necessary to change within crop herbicides for the management of IMI rape, the environmental impacts were low. Some of the management changes, encouraged for the management of HT weeds, such as preferring minimum tillage to ploughing, to promote more rapid decline of the initial seedbank of HT weeds, would have a positive environmental impact. The only management change that had a major negative environmental effect was the replacement of spring crops (where management of HT weeds was sometimes difficult) with winter wheat. This change resulted in the loss of environmentally beneficial over winter stubbles, with associated declines in plants, invertebrates and birds.

• Longevity of HT individuals: The only major issue that it was difficult to resolve related to how long farmers should avoid sowing a non HT crop after having planted an HT one. For example, as it was not possible to control HT rape volunteers in a non HT rape crop any volunteers would produce seeds that could cause the non HT crop to exceed the EU 0.9% admixture threshold. As stated above, this could mean not planting a non HT crop for 10 years. There was similar concern in relation to planting a different HT cultivar, as the presence of different HT volunteers would result in gene stacking. Problems were similar with sugar beet contaminated with HT weed beet, although, unlike HT rape, their presence would not affect the ‘purity’ of harvested non HT beet roots.

• Monitoring: The consortium concluded that there should be monitoring of fields that had been planted with HT crops and adjacent crops and uncropped field margins, following the planting of HT crops, to check for the presence of HT weeds. These could be detectable as a result of unexpected survival of individuals following the use of glyphosate (glyphosate tolerant) or sulfonylureas (IMI tolerant). Farmers should keep detailed records of crop management after the HT crop, as this would be helpful in explaining any subsequent issues.

• The overall conclusion of the project was that although HT volunteer rape and weed beet, would inevitably occur after the planting of HT rape or sugar beet, the subsequent management changes required to control them would be quite minor. The environmental impact of the changes both in herbicide choice and in cultural practices would also be low. The main issue would be when to replant oilseed rape and sugar beet on the same field and this would mean that it was vital to keep accurate records. Predictions of how long this should be are stronger for oilseed rape than they are for sugar beet. Management of the much less common rape/B.rapa hybrids would reflect the volunteer rape management options.

1) Introduction

The world-wide uptake of genetically modified crops continues to increase and in 2005 reached 90M ha (James, 2006). Even Europe (Spain, Germany, Portugal, France, Czech republic) is now growing small areas of GM maize and soya. There has been much discussion in Europe about the co-existence of GM and conventional/organic farming, as each country develops its own co-existence implementation strategy. Although much research has focussed on the environmental impact of the GM crops themselves, as exemplified by the comprehensive Farm Scale Evaluation (FSE) of GM crops, much less emphasis has been put on the environmental impact of any consequential changes in crop management in subsequent years and in adjacent fields. There has also been much publicity in the daily press about the possibility of gene escape from GM crops and the incorporation of the transgenes into wild relatives of the GM crop. Research such as that reported by Wilkinson (2003) has attempted to quantify this risk. The impact that any gene escape may have on farming and the environment will depend on the transgene present. The herbicide resistance trait, which still dominates GM crops worldwide, is particularly problematic as far as these issues are concerned. Volunteers from the GMHT crops and any hybrids created by cross fertilisation with wild relatives would exhibit the herbicide resistance trait and could therefore cause changes in farm management in subsequent years. These changes in crop and farm management could result in increases (or decreases) in the farm’s environmental profile.

Within the UK, the crops of primary concern are sugar beet and oilseed rape, as the third relevant GMHT crop, maize, has no related wild species in the UK and does not produce volunteers. It is already well established that bolted sugar beet can cross fertilise with sea beet (Beta maritima) to create annual beets (Bartsch et al., 1999) and similarly Wilkinson et al. (2003) in the UK and Joergensen et al. (1996) in Denmark have shown that oilseed rape can hybridise with Brassica rapa (turnip rape, wild turnip, bargeman’s cabbage). Hybrids with other members of the Brassicaceae may also occur. A further, and probably more critical issue with oilseed rape, arises from the persistence of seeds of the GMHT crop itself. The presence of these HT volunteers and hybrids may cause changes in crop management, which in turn may have environmental consequences. This project endeavours to quantify the risks of these problems arising and then to quantify the changes in crop and farm management that they may cause, and indicate the environmental impacts of these changes.

2) Objectives

1) To seek to predict the potential impact of HT problem weeds, focusing on weed beet in sugar beet and turnip rape in oilseed rape

2) To assess likely agronomic impacts of the presence of these HT weeds and identify any consequential environmental effects arising from changes in crop agronomy (herbicide use) caused by these HT weeds.

3) To consider additional effects arising from gene stacking

4) To recommend best practices to avoid the occurrence and spread of HT weeds

5) To propose monitoring measures

3) Reviews of current information on hybrids between GMHT crops (sugar beet, oilseed rape) and wild relatives, and on volunteer oilseed rape

3.1 Introduction

This project has focussed on:

a)hybrids between sugar beet and weed beet and sea beet

b)hybrids between oilseed rape and related members of the Brassicaceae, especially B. rapa

c)volunteer rape

The latter was not specified as a key part of the project at the outset but it became very clear that it would not be possible to consider changes in management associated with rape hybrids without also considering the management of volunteers. Although rape hybrids might not be common, GMHT volunteer rape will occur on every field that has grown GMHT rape. Consequently, it was highly likely that changes in management would be driven by the volunteer oilseed rape and that management of any rape hybrids would be associated with the volunteer management.

The reviews have collated published information on the biology of the wild species and, where available, of the biology of the hybrids, especially their fitness. They have identified how frequently hybrids will arise and how widespread the wild species are in the UK. In addition, they have considered issues relating to pollen flow and spatial gene flow, as this could impact on management in fields adjacent to GMHT crops. The following sections in Chapter 3 are a précis of the information produced in the project review (Berry et al., 2005).

3.1.1 Impact of herbicide tolerance on biology and population dynamics of beet and oilseed rape

One of the consequences of the introduction of ‘new’ genes into a crop is that it may change the ‘fitness’ of the plants, such that the crop may become better adapted to the natural environment, and hence become more invasive. This invasiveness could then be transferred to wild relatives. Evidence from work by Fredshavn & Poulsen (1996) and from Crawley et al., (1993, 2001) indicates that GM herbicide tolerant oilseed rape is not fitter than conventional varieties. Similarly, Crawley et al., (2001) could find no evidence of increased fitness with four crops that were transformed for either herbicide or insect resistance. Gene stacking of herbicide tolerant transgenes does not appear to affect the fitness of oilseed rape plants with double (glufosinate+imazethapyr, glufosinate+glyphosate) or triple tolerance (glyphosate+glufosinate+imazethapyr) (Simard et al., 2003, 2004). There seems to be no similar published data on sugar beet, though Bartsch (pers. comm. 2006) concludes that HT and conventional beet do not differ in their biology. Thus, it appears that HT volunteers will not be fitter that non HT ones, and increased fitness of hybrids between herbicide tolerant crops and wild relatives, as a result of the presence of the herbicide tolerance genes seems unlikely. Very recent information (Squire pers comm. 2006) has questioned this conclusion, as more GMHT volunteers have been found in fields 2-4 years after the HT crops were grown, compared to conventional non HT crops. The causes are unclear.

3.1.2 Impact of herbicide tolerance in crops on the weed flora

The main reason for the introduction of HT crops, at least for farmers, is the simplification of weed management, so that one or two treatments of a broad-spectrum post-emergence herbicide (primarily glyphosate, or glufosinate) can achieve very high levels of weed control. These treatments often replace complex series of pre- and post-emergence herbicides. One of the consequences to emerge from the widespread adoption of herbicide tolerant crops in N. America has been the appearance of weed species that are hard to control with the broad-spectrum herbicides, particularly glyphosate, associated with the HT crops. The appearance of these weed species can arise from two different mechanisms, weed shifts and the development of true resistance. Weed shifts occur due to selection pressure on the weed community from cultivation and herbicide regimes (Owen & Zelaya, 2005) and can occur in two ways. Firstly, they can arise through ecological adaptation to the cultivation techniques used in association with the HT crop. For example, different tillage systems can cause specific population shifts, with no tillage having the most dramatic effects on the weed community (Tuesca & Papa, 2001). American researchers differ in their opinion of the likelihood of a population shift due to selection pressure based on the continuous use of glyphosate, but it is clearly a possibility (Owen & Zelaya, 2005). Secondly, since the introduction of herbicide tolerant crops it has become apparent that there are a number of weed species which have inherent tolerance to the herbicides used in the HT crops (mainly glyphosate), despite this herbicide’s overall broad-spectrum activity. Owen and Zelaya (2005) have identified a number of species which may become, or are becoming, a serious problem in the USA for this reason.

The introduction of glyphosate tolerant crops may have increased the possibility of the evolution of glyphosate resistance, although there is much discussion on this issue within the scientific community. At present there are only eight species recorded as resistant to glyphosate (Heap, 2006; Owen & Zelaya, 2005). Conyza canadensis has developed resistance to glyphosate in 14 US states since 2000 and within 3 years has become a major problem in no tillage cotton production (Owen & Zelaya, 2005). This problem may be attributable to the way the glyphosate has been used within the GMHT cotton. Other weeds of cotton have also been reported to have become resistant (e.g. Amaranthus palmeri, Ambrosia artemisiifolia) (Heap, 2006).

In the context of the environmental impact of GMHT crops, there is a need to consider the consequences of weed shifts and the development of resistance for weed management. Clearly, when these effects arise, any environmental benefits arising from the GMHT crop technology may be diminished, as cultural practices may have to be changed and additional herbicides applied. Indeed, these management changes could be similar to those adopted to manage herbicide resistant volunteers or HT crop/weed hybrids. Thus, if there was a similar appearance of resistant weeds in the UK, following the adoption of herbicide tolerant crops, the solutions proposed in this report for volunteers and HT hybrids would contribute towards the resolution of the resistance problems.