6 Potential adverse agro-environmental effects across different spatial scale levels and implications for ERA and PMEM

As shown above, the combination of biological properties and the modified HR technology may entail various positive, neutral or adverse agro-environmental effects. These can be demonstrated to varying degrees by scientific observations and/or experiments, each of which were made on various scales of precision, space and time (Figure 3*). Compiling this information helps in the overall assessment of the evidence for potential agro-environmental effects (Table 1 and 2). This information is collected and investigated for the ERA of the HR sugar beet, and conclusions can be drawn for the PMEM. If information on specific aspects of the HR sugar beet is considered insufficient, further observations and experiments may be required.

An important aspect of the ERA procedure is to evaluate effects, i.e., are they adverse and do they constitute environmental damage. This is a function of hazard and likelihood of occurrence. The mere occurrence of a GM crop, for instance in arable non-GM fields or other biotopes, is not considered damage, but rather as an indicator for potential damage. According to Bartz et al. (2009*), an adverse effect can be defined as a reduction in valued attributes of one or more conservation resources. Moreover, environmental damage can be defined as a significant adverse effect on a biotic or abiotic conservation resource that has an impact a) on the environmental value of the conservation resource in whole or part, b) on the conservation resource as an ecosystem component, or c) on the sustainable use of the conservation resource or the ecosystem. This implies that components (composition, structure, functions) and scale levels (gene, species, ecosystem, landscape) of a conservation resource are considered in their entirety (Bartz et al., 2009, Figure 3*). Once an adverse effect is detected and sufficiently evidenced, its significance, which ultimately is an operational threshold value, must be assessed on a normative basis. It may then be considered as significant or not significant. If the significance is low or cannot yet be assessed, it can be allocated either to general surveillance (GS) or to a case-specific monitoring (CSM) (Figure 3*).

Figure 3: Pathways for ERA and PMEM and relevance of scales.

Table 3 presents the level of evidence of effects considered to be adverse, based on available findings listed in Tables 1 and 2. We provide indications of evidence as to whether an effect requires more research on the greenhouse- or field-scale level and/or whether it should be part of the PMEM carried out on the landscape scale. The indication of relevant scale levels depends on the type of effect or indicator group to be investigated (Graef et al., 2005b*; Hilbeck et al., 2008b). In general, the larger the scale of investigation, the more ecological relevance applies to an effect or indicator and the more challenging the risk management or control (Figure 1*). Some adverse effects such as invasiveness can be detected, if ever, only at larger scales. Furthermore, the larger the scale of investigation, the more challenging the experimental design and the less likely that standardised detection methods exist. Findings on the laboratory or greenhouse scale are not easily reproduced and/or confirmed on the field scale. Reasons include a) their poor relevance at larger scales due to higher variations and multiple influences of environmental factors, b) a lack of suitable field detection methods, or c) a poor statistical design. Based on our findings in Sections 3, 4, and 5, we outline research shortcomings at three different scale levels.

Laboratory or greenhouse scale: Lab or greenhouse experiments on the GM crop can involve survivability, genetic stability, as well as interactions with target and NTOs and the abiotic environment. Such experiments are a regular part of the GM crop development process. However, in general data on potential environmental effects of the GMHT sugar beet are few, inadequate or missing. Additionally, experimental results lack a sound statistical design and thus are not or hardly reproducible. For the acute toxicity tests in the risk assessment of NTOs we suggest to use the whole GM plant material instead of isolated microbially produced transgene products that are usually provided. We suggest prolonged ecotoxicity tests (Hilbeck et al., 2008a*), since in reality NTOs may be exposed over longer periods, for example one or several life stages or even over the whole life cycle. Moreover, they are exposed to the complete plant material of the GMP, not only to single substances. We also recommend incorporating realistic exposure pathways when testing NTOs (Römbke et al., 2010*). So far, the selection of test organisms is poorly founded. For instance, tests with insect pests of other crop species such as the European corn borer and the Colorado potato beetle are a questionable choice for investigating effects of a GMHR sugar beet on NTOs.

Field scale (limited in time and space): For the field scale, we suggest further experiments of adverse effects of a) the GMP and b) different glyphosate formulations on various trophic levels of field organisms, aquatic communities and soil microbial communities. The latter should focus on specific effects caused by the changed time frame of glyphosate spraying in GMHR sugar beet. To differentiate between herbicide effects and possible pleiotropic effects, part of the experiments should exclude pesticide application. Further research and more information for the ERA is also required on possible adverse effects for microbial and/or faunal activities and on soil degradation. Irrespective of the HR sugar beet, the relevance and potential adverse effects of horizontal gene transfer should be further clarified. Only little is known about possible combinatory effects after hybridisation and about pleiotropic and epigenetic genome effects of the HR sugar beet. More information is also required about potentially altered quality of plant parts after years of cultivation and about possible effects on various soil functions of the HR sugar beet.

Landscape scale: For the landscape scale, we suggest the CSM of possibly increased glyphosate formulation concentrations in soils and aquatic ecosystems as an indicator for adverse effects on field organisms and aquatic communities. Observation and eradication of sugar beet bolters and volunteers is usual agricultural practice. In some places and/or in some years, bolters or volunteers may be either overlooked or too frequent to overcome. This may lead to HR pollen flow, hybridisation with neighbouring beets and wild relatives, seed production and seed shed. We suggest the GS of possible persistence and invasiveness (triggered by a possible selective advantage and/or genetic drift) for HR cultivated, feral and weed beets, hybridised HR related species and/or the transgenes in the neighbouring environment. This applies especially to HR sugar beet seed production regions. A CSM is recommended for detecting the likely increase of HR in weeds and subsequently the likely increase and/or change in the herbicide application regime after several years of glyphosate application. A further recommendation is the CSM of the possible HR-related decrease in agrobiodiversity, including weed communities, herbivores, pollinators and beneficial species. GS should be done of the long-term sustainability of cropping systems and their rotations as well as of a potentially increased herbicide drift and pollution. Although current evidence is poor for potential adverse effects on migratory and mobile species, on soil functions and ecosystem food chain effects, and on larger scale biodiversity, literature findings do suggest GS of these issues.