The global area planted with genetically engineered (GE) crops has consistently increased each year since GE crops were first commercially cultivated in 1996, reaching 90 million hectares in 2005.

Five countries (USA, Argentina, Brazil, Canada and China) are growing nearly 95% of the total area of these crops. In contrast the adoption of GE crops in Europe was much less intense. This situation is probably going to change, since the European Union (EU) entered the first GE maize varieties expressing insecticidal proteins from Bacillus thuringiensis (Bt) into the Common EU Catalogue of Varieties in September 2004. It is generally expected that Bt-maize will also be commercially grown in EU countries other than Spain, where commercial GE crop cultivation started in 1998. Several countries such as France, Germany, Portugal, and the Czech Republic started growing Bt-maize in 2005. Compared to Spain, where approximately 12% of the total maize area grown in 2004 (representing 58,000 ha) was planted with Bt-maize, the acreage in these countries is, however, very limited and accounts for less than 1,000 ha each.

GE crops, modern agricultural systems and the environment
Concerns have been raised that the commercial cultivation of GE crops could result in adverse effects on the environment.1 We have therefore reviewed the scientific knowledge on environmental impacts of GE crops deriving from ten years of worldwide experimental field research and commercial cultivation.2 Our study focused on the currently commercially available GE crops that could be relevant for agriculture in Western and Central Europe (i.e., maize, oilseed rape, and soybean), and on the two main GE traits that are currently commercialized, herbicide tolerance (HT) and insect resistance (IR). The sources of information included peer-reviewed scientific journals, scientific books, reports from countries with extensive GE crop cultivation, as well as reports from international organizations.

Potential impacts of GE crops should be put in relation to the environmental impacts of modern agricultural practices that took place during the last decades. Independent from the use of GE crops, modern agricultural systems have profound impacts on all environmental resources, including negative impacts on biodiversity. Several changes in the management of agricultural land over the last century have resulted in a decline in the biodiversity within agro-ecosystems.3,4

Effects of GE crops on non-target organisms
There are concerns that insect-resistant GE crops expressing Cry-proteins from Bacillus thuringiensis (Bt) could harm organisms other than the pest(s) targeted by the toxin. The published large-scale studies in Bt crops assessing possible non-target effects on arthropods have only revealed subtle shifts in the arthropod community, which can be explained by a lack of the target pest resulting from the effective control by the Bt crops.5 No adverse effects on non-target natural enemies resulting from direct toxicity of the expressed Bt toxins have so far been observed in laboratory studies and in the field. There is evidence that the Bt crops grown today are more target-specific and have fewer side effects on non-target organisms than most current insecticides used.

While the adoption of Bt maize has resulted in only modest reductions in insecticide applications due to the small area of conventional maize treated with insecticides against the European Corn Borer, the commercial cultivation of Bt cotton has resulted both in a substantial reduction in quantity and in number of insecticide applications.6 In addition to direct environmental benefits such as fewer non-target effects and reduced pesticide inputs in water, demonstrable health benefits have been documented for farm workers in developing countries due to less chemical insecticide spraying in Bt cotton.7

Impacts of GE crops on soil ecosystems
Similarly to non-target effects above ground, concerns were raised that Bt crops could have effects on soil organisms. Bt toxins enter the soil system primarily via root exudation and via plant remains after harvest. Both degradation and inactivation of the Bt toxin vary, depending on parameters such as temperature and soil type. The initial degradation of the toxin is rapid, while a low percentage (< 2%) may remain in the soil ecosystem following one growing season. Bt toxins have been shown to bind to clay and humic acid compounds; however, no accumulation of toxins has been observed after several years of cultivation of Bt crops.

Population sizes and community structure of soil organisms are subject to both natural seasonal variation and to variations caused by the agricultural system (soil type, plant age, crops, cultivars, and crop rotation). Neither laboratory nor field studies have shown lethal or sublethal effects of Bt toxins on nontarget soil organisms such as earthworms, collembola, mites, woodlice, or nematodes. Some differences in total numbers and community structure have been described for microorganisms. The ecological significance of the observed differences is not clear. Because most studies have not assessed the natural variation occurring in agricultural systems, it is generally difficult to establish whether the differences between Bt and non-Bt crops were exceeding this variation. The only study considering natural variation suggests that observed effects lie within this variation and that the differences between conventional cultivars outweigh the observed influences of Bt crops.8

Gene flow from GE crops to wild relatives
There is general scientific agreement that gene flow from GE crops to sexually compatible wild relatives can occur.9 Experimental studies have shown that GE crops are capable of spontaneously mating with wild relatives, however at rates in the order of what would be expected for non-transgenic crops. Few studies have shown that GE herbicide tolerant (GEHT) oilseed rape (Brassica napus) can form F1 hybrids with wild turnip (Brassica rapa) at low frequency under natural conditions. Questions remain whether these transgenes would cause ecologically relevant changes in recipient plant populations. Although there is a low probability that increased weediness due to gene flow could occur, it is unlikely that GEHT weeds would create greater agricultural problems than conventional weeds. Farmers can generally choose among several herbicides for the cultivation of a given crop and they have further a set of options within a crop rotation to control or manage weeds.

In natural habitats, no long-term introgression of transgenes into wild plant populations leading to the extinction of any wild plant taxa has been observed to date. Transgenes conferring herbicide tolerance are unlikely to confer a benefit in natural habitats because these genes are selectively neutral in natural environments, whereas insect resistance genes could increase fitness if pests contribute to the control of natural plant populations.

Invasiveness of GE crops into natural habitats
Despite the concern that GE crops could invade natural habitats, brought up early in the discussion on potential environmental risk related to the release of GE crops, it seems that modern crop varieties generally stay domesticated. There is no evidence at present that the extensive cultivation of GEHT oilseed rape over several years in Western Canada has resulted in a widespread dispersal of volunteer oilseed rape carrying herbicide-tolerance traits. Although one study found triple-herbicide resistant, and another study reported double-herbicide resistant, oilseed rape volunteers in Western Canada, the general lack of reported multiple-resistant volunteers suggests that these volunteers are being controlled by chemical and non-chemical management strategies, and are therefore not an agronomic concern to most farmers. Furthermore, there is currently no evidence that GEHT oilseed rape has become feral and invaded natural habitats.

Impacts of GE crops on pest and weed management
Impacts of GE crops on pest and weed management practices and their potential ecological consequences are usually difficult to assess, because they are generally influenced by many interacting factors and often only show up after an extended period of time. Numerous weed species evolved resistance to a number of herbicides long before the introduction of GEHT crops.10 The experiences available from regions growing GEHT crops on a large scale confirm that the development of herbicide resistances in weeds is not primarily a question of genetic modification, but of the crop- and herbicide management applied by farmers. Despite extensive cultivation of GEHT oilseed rape in Canada, no weed species has so far been observed being tolerant to the herbicides glyphosate and glufosinate. In continuously cultivated GEHT soybeans in the United States, in contrast, many fields have been treated only with glyphosate, which increases the pressure for the selection of resistant biotypes. As a consequence, three years after the introduction of GEHT soybean varieties, glyphosate-resistant horseweed (Conyza canadensis) has been detected. Although farmers have to add another herbicide to glyphosate to control the resistant weed species, there are alternatives to glyphosate for most weed species that are highly effective and provide good flexibility in application timing. There is, however, no question that glyphosate-resistant weeds will increase the costs of weed management for farmers.

The adoption of GEHT crops has allowed the use of a single broad spectrum herbicide that may reduce the need for costly herbicide combinations. Glyphosate and glufosinate are generally considered toxicologically more benign, being in particular less toxic to humans and the environment than many of the herbicides they replace. In addition, the adoption of GEHT crops has often facilitated the change to conservation tillage agriculture. Growers using conservation tillage have reduced their tillage operations, thus preventing soil erosion and soil degradation.

The results of the UK Farm Scale Evaluations (FSE) showed that weed biomass and numbers of some invertebrate groups were reduced under GEHT management in sugar beet and oilseed rape and increased in maize compared with conventional treatments.11 These differences were related to the weed management of both conventional and GEHT systems. Highly effective weed control practices, such as those chosen for the GEHT crops in the FSE, lead to low numbers of weed seeds and insects. Fewer insects and decreased weed seed might reduce the numbers of birds that depend on these insects and seeds as a food source. The FSE assumed no other changes in field management than GEHT crops replacing non-GE varieties. However, other cropping systems such as conservation tillage are possible, resulting in a greater availability of crop residues and weed seeds and, in consequence, improving food supplies for insects, birds, and small mammals.12

Conclusions
The risks GE crops pose for the environment, and especially for biodiversity, have been extensively assessed worldwide during the past ten years of commercial cultivation. Consequently, substantial scientific data on environmental effects of the currently commercialized GE crops is available today, and will further be obtained given that several research programs are underway in a number of countries. The data available so far provide no scientific evidence that the commercial cultivation of GE crops has caused environmental harm. Nevertheless, a number of issues related to the interpretation of scientific data on effects of GE crops on the environment are debated controversially. To a certain extent, this is due to the inherent fact that scientific data is always characterized by uncertainties, and that predictions on potential long-term or cumulative effects are difficult. Uncertainties can either be related to the circumstance that there is not yet a sufficient data basis provided for an assessment of consequences (the "unknown"), or to the fact that the questions posed are out of reach for scientific methods (the "unknowable"). Although some might argue that experience and solid scientific knowledge are still lacking, the debate is generally not purely due to a lack of scientific data, but more to an ambiguous interpretations of what is considered an ecologically relevant effect of GE crops. The interpretation of study results is thereby often challenged by the absence of a defined baseline to evaluate environmental effects of GE crops in the context of modern agricultural systems. There is thus a need to develop scientific criteria to assist regulatory authorities when deciding whether environmental effects of GE crops are considered relevant.

When discussing the risks of GE crops, one has to recognize that the real choice for farmers and consumers is not between a GE technology that may have risks and a completely safe alternative. The real choice is between GE crops and current conventional pest and weed management practices, all possibly having positive and negative outcomes. To ensure that a policy is truly precautionary, one should therefore compare the risk of adopting a technology against the risk of not adopting it.13 We thus believe that both benefits and risks of GE crop systems should be compared with those of current agricultural practices.

Acknowledgements
We would like to thank the Swiss Expert Committee for Biosafety for major funding of the study.

The study is publicly available on the internet via the following link: http://www.art.admin.ch/dms_files/03017_de.pdf

References
1.Conner AJ, Glare TR, and Nap J-P. (2003) The release of genetically modified crops into the environment. Part II. Overview of ecological risk assessment. Plant Journal 33, 19-46
2.Sanvido O, Stark M, Romeis J, and Bigler F. (2006) Ecological impacts of genetically modified crops: experiences from ten years of experimental field research and commercial cultivation, Agroscope Reckenholz Tänikon Research Station ART, Zurich, Switzerland
3.Robinson RA and Sutherland WJ. (2002) Post-war changes in arable farming and biodiversity in Great Britain. Journal of Applied Ecology 39. 157-176
4.Stoate C et al. (2001) Ecological impacts of arable intensification in Europe. Journal of Environmental Management 63. 337-365
5.Romeis J, Meissle M, and Bigler F. (2006) Transgenic crops expressing Bacillus thuringiensis toxins and biological control. Nature Biotechnology 24, 63-71
6.FAO (2004) The state of food and agriculture - agricultural biotechnology, meeting the needs of the poor? Food and Agriculture Organization of the United Nations, Rome, Italy
7.Huang JK et al. (2003) Biotechnology as an alternative to chemical pesticides: a case study of Bt cotton in China. Agricultural Economics 29. 55-67.
8.Griffiths BS et al. (2005) A comparison of soil microbial community structure, protozoa and nematodes in field plots of conventional and genetically modified maize expressing the Bacillus thuringiensis Cry1Ab toxin. Plant and Soil 275, 135-146
9.Stewart CN, Halfhill MD and Warwick SI. (2003) Transgene introgression from genetically modified crops to their wild relatives. Nature Reviews Genetics 4, 806-817
10.Heap I. (2006) The International Survey of Herbicide Resistant Weeds. http://www.weedscience.com
11.Squire GR et al. (2003) On the rationale and interpretation of the Farm Scale Evaluations of genetically modified herbicide-tolerant crops. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences 358, 1779-1799
12.Holland JM. (2004) The environmental consequences of adopting conservation tillage in Europe: reviewing the evidence. Agriculture Ecosystems & Environment 103. 1-25
13.Goklany IM. (2002) From precautionary principle to risk-risk analysis. Nature Biotechnology 20, 1075

--------------

Olivier Sanvido, Michèle Stark, Jörg Romeis and Franz Bigler
Agroscope Reckenholz-Tänikon Research Station ART,
Reckenholzstr. 191, CH-8046 Zurich, Switzerland
olivier.sanvido@art.admin.ch