The existing ERA is performed for one active substance or pesticide product that is applied once or a few times in a specific crop. However, the current cropping systems do not only receive one application of a pesticide. Their seeds might be already treated with a mixture of multiple systemic pesticides and several further products are applied on the growing plants or fruits during the season. In Germany in 2016 on average there were 6 pesticides applied (treatment index) in wheat, 7 in oilseed rape, 14 in potatoes, 22 in vine orchards, and 32 in apple production orchards (JKI, 2019). In the UK even more pesticides were used for the same crops: 11 pesticides for wheat, 13 for oilseed rape, and 21 for potatoes (FERA, 2017). Outside the EU maximum pesticide inputs as in banana in Costa Rica, where aerial applications are conducted in conventional plantations every 4 days, result in volumes of over 75 kg of active molecules/ha/year. It is obvious to every ecotoxicologist and ecologist that multiple, sequential field applications of biologically active chemicals are likely to cause more severe effects on a population of organisms than a single application event. However, the current risk assessment assumes that populations only face a single impact from a specific pesticide, with sufficient time following after application to allow the population to recover to former levels. In reality the same population is facing multiple pesticide impacts during the growing season. This is a worrying underestimation of the actual risk for biodiversity in the agricultural landscape resulting from pesticide use. Similar concerns of an underestimation of effects of contamination with multiple pesticides and other chemicals are also raised for human health (Leu and Shiva, 2014).

The current ERA scheme addresses the effects of a pesticide on each group of organisms separately. There are ERA sections on plants, on insects and spiders (arthropods) and birds. Field studies are sometimes performed for arthropods, where interactions between predatory insects and their prey is recorded. However, ERA does not include so-called indirect effects or interactions between trophic levels of different organism groups. An example can be seen in an herbicide that has no acute toxic effect on insects as well as birds and therefore passes the current risk assessment for both groups. However, the application of the herbicide leads to a reduction of “weeds” (as intended in the field) and of “non-target plants” (the same plant species growing outside the field), therefore reducing the amount of food for pollinators and herbivorous insects. This depletion can lead to further impacts on birds since herbivorous insect larvae, such as caterpillars, are smaller and less abundant after herbicide treatments (Hahn et al., 2014), reducing the insect biomass available to feed the birds offspring. Trophic interactions are fundamental features of ecosystems and therefore need to be considered in ERA.

The European ERA focusses on environmental effects that can occur in semi-natural structures outside the agricultural fields. Currently no ERA for the in-field risk is mandated. However, the scientific opinion for non-target arthropods, mentions “biodiversity has to be supported to a certain degree in the in-field areas (…) in order to provide important ecosystem services (EFSA, 2015).” However, the respective guideline is not addressing this issue and negative effects on biodiversity are therefore deliberately accepted in the cropping area where pesticides are directly applied at biologically effective rates. The agricultural cropping area that receives pesticide inputs in Europe represents 22% of the total land area, reaching more than 30% for example in Germany and France (for 2015, Eurostat, 2019). Therefore, in countries with a high proportion of cropped area almost a third of the terrestrial land surface is not evaluated regarding negative effects of pesticides on its biodiversity. To explain the observed decline in insect biomass in the agricultural landscape of Germany (Hallmann et al., 2017) the most parsimonious explanation (Occam's razor) is the annual application of insecticides on more than 30% of Germanys land area, the entire cropping area, since the 1970s. No other factor such as the suggested light pollution or soil sealing needs to be invoked to explain the observed reductions (BMU, 2018).

The ERA required for pesticide regulation is in most cases not addressing the impact of pesticide use in agricultural fields and does not include food-web related ecosystem effects. This fundamental misconception leads to an ERA scheme and a resulting pesticide regulation that is not protective for biodiversity. If we remain working with the ERA scheme in place, in our opinion we will continue to observe further declines of many groups of organisms such as farmland birds and insects in the agricultural landscape. Neglecting the three described factors can have far-reaching consequences at the ecosystem level that are likely larger than an underestimation of risk due to a lower uncertainty factor or a flaw in the experimental design of a field study. The misconception can also not be compensated by additional studies including new surrogate species or groups of organisms. The banning of certain insecticides or broad-band herbicides will also hardly improve the situation. We therefore urgently need to rethink our basis for the regulation of these biologically active substances and develop a holistic approach to include indirect effects caused by multiple pesticides applied in the agricultural productive land area of Europe. Since the current ERA for the regulation of pesticides is not addressing the real-world situation we ought to accept that the current practice of pesticide use in European agriculture is not sufficiently protective and therefore not safe for the terrestrial environment.

The development of a new systemic approach for ERA will take considerable time and require substantial resources. We therefore also need to discuss other options to at least halt the negative effects of pesticides on biodiversity of the agricultural landscape. Risk management to mitigate negative pesticide effects might be a helpful alternative until we are able to assess the true environmental risk of pesticide usage. Reducing pesticides in agricultural practice is an obvious option. It was estimated that total pesticide use could be reduced by more than 40% in almost 60% of 946 evaluated farms in a French network without any negative effects on both productivity and profitability (Lechenet et al., 2017). Integrated pest management should focus on using natural enemies of pests and crop rotations and agree on pesticides as a last option instead of current practices, where pesticides are prophylactically implemented in farming practices (e.g., seed-treatments of cereals). We additionally could extend the proportion of semi-natural habitats without pesticide inputs in the agricultural landscape, increase agri-environmental schemes and enlarge the area of organic farming. Many options are on the table and a strengthening of the greening of the common agricultural policy (CAP) is currently discussed for the coming period of European policy (Erjavec and Erjavec, 2015; Solazzo et al., 2016; Alons, 2017). Risk mitigation of pesticides needs to be implemented effectively and at a large scale to bend the curve of biodiversity decline in agricultural landscapes now. If we delay to change agricultural practice and its current pesticide use our efforts to stop the current biodiversity decline and restore it to former levels need to be much larger at a later stage.

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.