5 Direct and indirect effects on the agro-environment

The biological features, combined with the HR technology, as shown above, may entail direct, indirect, immediate, delayed and/or cumulative agro-environmental effects (Table 1 and 2).

Effects on the agro-environment may be induced by single or several different mechanisms; these may work singly or cumulatively (Graef, 2009*). The agro-environmental effects may be detected at a single scale or at multiple levels.

Section 6 discusses whether the effects are considered to be adverse, positive, not relevant or relevant for monitoring, whether they require further studies for the ERA or raise the concern of an environmental risk leading to deny approval, or whether they may constitute an environmental damage (Bartz et al., 2009*) that merits withdrawal of approval.

Effects directly or indirectly linked to sugar beet biology: Bolting, formation of groundkeepers, along with the high rate and distance of pollen spread and cross-pollination (Drießen et al., 2001; OECD, 2006) may trigger HR gene flow to neighbouring non-GM sugar beet, weed beets, wild beets and related wild species. The preconditions are spatial concurrence (Frese, 1998) and inefficient prevention of flowering by agricultural practice.

The potential increase of HR sugar beet in the seed banks, which may incorporate viable HR seeds over 10 years, may lead to persisting and invasive HR weed beets, wild beets and related wild species in fields and natural habitats. The preconditions: a selection advantage due to repeated glyphosate application, enhanced fitness parameters and genetic drift. Wilkinson et al. (2000) and Snow (2003) observed this for HR oilseed rape. Cultivated beet genes can persist in wild beet (Bartsch et al., 1999; Sukopp et al., 2005). If HR weed beets, wild beets and related wild species become invasive, this may variously impact habitats, food chains and biodiversity (Watkinson et al., 2000*; Züghart and Breckling, 2003*).

A horizontal gene transfer of HR from plant residues to soil microorganisms is rare but possible, irrespective of the HR trait (Heinemann and Traavik, 2004; Nielsen and Townsend, 2004), but its environmental implications are hard to determine. Another general ecological concern is the potential for a) adverse combinatory effects when GMHR sugar beet hybridises with weed beets, wild beets and related wild species that potentially affect the hybrid’s biology and/or herbivores and b) pleiotropic and epigenetic genome effects of the GM crop caused by the genetic modification procedure, as shown for instance with GM wheat (Zeller et al., 2010). These unintended and unanticipated effects, for instance on non-target organisms (NTOs), must be considered, especially when performing the ERA (GMO Panel, 2010*).

Effects directly or indirectly linked to the HR technology: More efficient weed suppression leads to less biomass, food and flowers for field organisms after spraying. This, in turn, entails lower abundances of various herbivores, pollinators and beneficial species (pest antagonists) (Bohan et al., 2005*; Heard et al., 2003b*), and may decrease agrobiodiversity (Watkinson et al., 2000*). A shift in weedy species (Heard et al., 2003b*,a*) and an increase of perennial weeds due to minimum-till practice is likely (Frick and Thomas, 1992*). Extending the HR technology may have various potential effects on field organisms and soil bio-geochemical cycles on the larger landscape scale (Benton et al., 2002*; Cerdeira and Duke, 2006*). Another ecological concern is possible adverse indirect effects on migratory and mobile species, leaf litter quality, crop competitiveness, and insect resistance (Squire et al., 2003*).

Applying glyphosate formulations compared to other herbicides has adverse effects on fields and neighbouring habitats. These include increased mortality of amphibians (Relyea, 2005a*,b*,c*) and mammals (Richard et al., 2005; Benachour and Séralini, 2009*) and, in combination with simultaneous exposure to parasites, reduced survival of freshwater fish (Kelly et al., 2010*). Adverse direct or indirect glyphosate effects were also reported for micronutrient uptake (Tesfamariam et al., 2009*), soil microflora and plant disease severity (Fernandez et al., 2009*; Heuer et al., 2002; Johal and Huber, 2009*; Kremer and Means, 2009*). This can hamper soil functions or bio-geochemical cycles (Züghart and Breckling, 2003*). Glyphosate formulations containing surfactants such as POEA (polyethoxylated tallow amine) are more toxic than the ai glyphosate alone (Benachour and Séralini, 2009*), in particular for aquatic organisms (Brausch and Smith, 2007). Some studies also indicate less herbicide toxicity and persistency than other herbicides (Agronomy Guide, 2010*; Cerdeira and Duke, 2006*; Squire et al., 2003*).

Until post-emergent spraying, more biomass is available for feeding organisms (Werner et al., 2000*; Strandberg et al., 2005*). After spraying, however, biomass drops compared to conventional spraying. Furthermore, spraying is usually done before weed seed development (Dewar et al., 2000), reducing the weed seed bank in the long term. Early band applications combined with late overall treatments may enhance biomass availability during the growing period (May et al., 2005). Nonetheless, seed development and abundance in arable flora are also reduced in the long run. Post-emergent spraying may increase herbicide drift into the agro-environment, for example due to increased spraying height (Johnson, 2001*). Post-emergent spraying also often entails a change in spray schedules of insecticides and fungicides, with potential implications for microbial and faunal activity (Champion et al., 2003*; Thorbek and Bilde, 2004*).

To control possible HR sugar beet volunteers in follow crops, modified crop rotations may be necessary. This may require farmers to change the tillage system (Schütte et al., 2004*), affecting field organisms and soil bio-geochemical cycles (McLaughlin and Mineau, 1995*; Orson, 2002). It may also require more ai, different types of herbicides or higher spraying frequency to control HR in weeds (Van Acker et al., 2003*). Again, this can impact agrobiodiversity. Coexistence measures to reduce gene flow may change agricultural practice and entail various environmentally relevant effects.