At the European Food Safety Authority’s (EFSA’s) 10th anniversary conference (EFSA, 2012), it became apparent that EFSA’s environmental risk assessment (ERA) schemes have evolved independently in the different areas within its remit (see EFSA, 2011), and that further harmonisation is possible on specific topics. EFSA, therefore, mandated the Scientific Committee (under mandate M-2013-0098) to harmonise EFSA’s ERA schemes with regard to: (1) accounting for biodiversity and ecosystem services to define protection goals for ERA; (2) coverage of endangered species as nontarget organisms in single-stressor ERA; and (3) temporal and spatial recovery of non-target organisms for ERAs. The Scientific Committee prepared three separate scientific documents to address the above mentioned issues and this scientific opinion is specifically about (3) temporal and spatial recovery of non-target organisms for ERAs.
In ERA for potential stressors that fall under the remit of EFSA, two protection goal options can be distinguished, viz., the threshold option and the recovery option. The threshold option is the specific protection goal option accepting no (or negligible) population-level effects of exposure to a potential stressor. The recovery option is a specific protection goal option accepting some population-level effects of the potential stressor if ecological recovery takes place within an acceptable time-period.
The EFSA Scientific Committee collected and discussed relevant information from the diverse areas of ERA conducted by EFSA and from the scientific literature. From these discussions, a draft scientific opinion was proposed for public consultation and further adoption by the EFSA Scientific Committee (see Sections 1 and 2). For this assessment, the Scientific Committee proceeded in four steps.
First, the Scientific Committee provided clarification on terminology and concepts that are needed when addressing ecological recovery (see Section 3). In particular, definitions were provided for environmental stressors (i.e. physical, chemical and biological) including pulse and press disturbances (Section 3.1); direct and indirect effects (Section 3.2) and ecological recovery (Section 3.3), comprising actual and potential recovery (Section 3.3.1), recovery at the population level, including internal and external recovery (Section 3.3.2) and resilience at the ecosystem level (Section 3.3.3).
In the above clarifications, an analogy was made between the terms stress, disturbance and perturbation. Also, it was highlighted that multiple environmental stressors can act simultaneously or sequentially. It was concluded that, independently of the type(s) of stressor(s) and duration of stress, the normal operating range (NOR) of individuals, populations, communities and ecosystems becomes disrupted when the environmental stressors exceed a threshold of exposure. The actual recovery is related to the return to this NOR, whereas the potential recovery is defined as the disappearance of the stressor to a level at which it no longer has a direct (toxic) effect on the ecological entity (endpoint) and after which recovery theoretically can start. However, it is challenging to define and measure the NOR. Under field conditions, adverse effects of a stressor can remain unnoticed if the measurement endpoints show a relatively large variability due to effects of natural factors. For indirect effects, it was noted that they may persist longer than direct effects. Furthermore, clarifications were provided on the spatial dynamics operating when a stressor affects populations differentially in space and time, and on population stability, both of which are necessary to understand recovery in a landscape context. To predict recovery of populations of non-target organisms, it is necessary to understand actual population conditions during the period of stress.
Second, the Scientific Committee developed a conceptual framework for the assessment of ecological recovery (Section 4) and gathered knowledge on the key parameters that need to be considered when assessing ecological recovery, in particular (1) the properties of the types of potential stressors of concern that fall under the remit of EFSA (hereafter mentioned potential stressors), i.e. plant protection products (PPPs), genetically modified organisms (GMOs), feed additives and invasive alien species (IAS) that are harmful to plant health (see Section 5, and Appendices A and B); (2) the species and their traits, e.g. related to demography, dispersal and foraging behaviour as well as adaptation to potential stressors (see Section 6); and (3) the specific features of the landscape, i.e. variations in land use, and the types, spatial distribution and connectivity of habitats (see Section 7).
Regarding the properties of the PPPs, GMOs, feed additives and IAS that are harmful to plant health (described in Sections 5.1, 5.2, 5.3 and 5.4, respectively), the Scientific Committee summarised information on their patterns of use, or presence in the case of invasive alien species that are harmful to plant health, in space and time, and on how ecological recovery is tackled for each of these potential stressors in the European Union (EU) legislation. In addition, when available, studies providing data on ecological recovery from exposure to these stressors were described. Finally, impacts on food-web interactions and ecological recovery from these stressors were considered.
Regarding the species traits that may affect ecological recovery, demographic (life-history traits), recolonisation (dispersal traits) and other traits such as foraging behaviour are identified as being of utmost importance (Section 6.1). To illustrate this, some examples of specific traits for focal taxa are described (Section 6.2). The contribution of genetic diversity to recovery is discussed in the context of adaptation to stresses (i.e. in the sense of the selection and genetic inheritance of resistant genotypes) (see Section 6.3). According to the ecological insurance hypothesis, the more genetically diverse a population or community, the better they can withstand potential stressors and can continue providing ecosystem services. It is worth noting that tolerance acquisition resulting from adaptation to stress by different processes may or may be not associated with fitness costs.
Some specific features of agricultural landscapes that may affect ecological recovery (Section 7) are described for the terrestrial and aquatic (i.e. for surface waters that drain and/or irrigate agricultural landscapes) compartments. For the terrestrial compartment (Section 7.1), the spatial distribution and connectivity of treated fields in relation to non-treated areas and the variety of possible land uses in Europe are known to influence the likelihood of concurrent events (i.e. treatments in multiple fields) and therefore the level of exposure to potential stressors in the landscape. These features are all important to consider when selecting the spatial scale at which recovery needs to be assessed. It is also highlighted that these features are important for influencing recovery of organisms that move between in-field and off-field areas (due to the concept of ‘action at a distance’ – i.e. effects of potential stressors may occur outside of the spatial area occupied by these stressors). For the aquatic compartment (Section 7.2), the surface area drained by streams overall is considerably larger than that of ponds, whereas ditches have an intermediate position. In reverse, the retention time of water (i.e. the average length of the time that water spends in the system) increases when going from streams to ditches to ponds. In theory, both the potential of faster recovery following exposure to a potential stressor and the chance to suffer multiple potential stressors will be ranked in the order streams > ditches > ponds. Given the spatial and temporal variability of the European landscapes and also given the diversity of the data sets and classifications used to assess and record the landscape structure in Europe, it may be challenging to incorporate such variations when conducting an ERA and assessing ecological recovery (Section 7.3).
Third, taking into account the complexity of ecological systems comprising multiple variables (see Section 8), the Scientific Committee examined the pros and cons of experimental (Section 8.1) and modelling (Section 8.2) approaches to address ecological recovery of the appropriate focal species. Experimental model ecosystem studies (e.g. mesocosm studies) allow replication so that treatmentrelated effects on, and recovery of, populations, communities and functional endpoints can be evaluated statistically, but are limited in the ecological realism that can be investigated. The minimum detectable difference is suggested as an indicator of the statistical power of a semifield test. For modelling approaches, pros are mostly linked to the ability of models to simulate accurately complex ecological systems where potential stressors may cause multiple outcome changes due to feedback mechanisms. This requires a good understanding of the ecological processes influencing the responses of the assessed entity within its environmental context and a clear definition of the domain of the applicability of the model. Potential disadvantages are the high demand for data and expert skills for both the development and validation of models. However, in prospective risk assessment (e.g. in the case of invasive alien species that are harmful to plant health), neither experimental nor modelling approaches can provide complete information. In such cases, expert opinion elicitation is required. Finally, it is concluded that experimental and modelling approaches need to be linked to appropriately predict recovery processes at the appropriate spatial and temporal scales, whereas field monitoring is required as a reality check.
Fourth, from the information collected and described above, the conceptual framework as given in Section 4 is revisited and discussed, and an integrated approach for addressing ecological recovery for any potential stressor, and IAS that are harmful to plant health, is proposed (see Section 9). Initially, the factors affecting ecological recovery of vulnerable non-target organisms after exposure to different types of potential stressor (Section 9.1) and the relationships between recovery of structural and functional endpoints (Section 9.2) are clarified. Then, the integrated approach is described (Section 9.3) based on the conceptual framework described earlier and information is provided on how to select appropriate focal taxa and/or processes (Section 9.3.1) and appropriate spatial scales (Section 9.3.2) to address exposure, effects and ecological recovery. Finally, clarifications are provided to address specifically ecological resilience for systems impacted by IAS that are harmful to plant health (Section 9.3.3).
This scientific opinion proposes that a systems approach is required to appropriately address ecological recovery in ERA (Section 10). This systems approach allows the integration of the various species, environmental factors, scales and stressor-related responses necessary to address the context-dependency in ecological recovery. Although this may appear to generate an overly complex ERA, the systems approach allows the identification of realistic worst-case combinations of species and environmental scenarios that are necessary. To ensure confidence in this approach it is important that the tools (environmental scenarios and models) are developed as a common resource ensuring transparency and reliability. Thus, the complexity may be reduced to arrive at a manageable day-to-day approach for all parties in regulatory risk assessment. In this context it is important to note that the conservatism of the assessment depends upon the selection of appropriate scenarios and focal biological entities. To reject a systems approach on the basis of complexity would ignore the fact that current decisions based on general approaches may not provide adequate levels of protection (either over- or under-protective). To successfully implement a systems approach the following challenges should be addressed:
- Harmonisation of the procedure for selection of focal taxa and construction of environmental scenarios between different regions and different potential stressors. Common focal taxa need to be identified to reduce the number of models that need to be developed by using the same focal taxa in as many scenarios as possible.
- Make available resources to exploit and further improve databases to select focal taxa, to construct environmental scenarios, and to develop and validate related models representative of different regions in Europe (similar to the procedure adopted by Forum for the co-ordination of pesticide fate models and their Use (FOCUS) exposure scenarios).
- Case studies should be developed as proof of principle.
- Ensure an appropriate linkage to lower tier approaches, monitoring and other EU data collection initiatives already in place (e.g. the Sustainability Use Directive).
- Ensure the maintenance and updating of scenarios and models as new information becomes available and incorporate changes in agricultural systems over time. This needs to be coordinated by a version control group (possibly an EFSA activity).
This scientific opinion makes several conclusions and identifies key challenges for assessing ecological recovery of non-target organisms in ERA of potential stressors (see Section 10), followed by a series of recommendations (see Section 11). In conclusion, the following key challenges were identified:
- define the NOR of ecological entities (bearing in mind that this may vary in time and between different ecosystems);
- identify focal taxa, focal communities and/or focal landscapes;
- assess appropriately action at a distance in cases where the specific protection goal allows the recovery option for in-field habitats, but not for off-field habitats, particularly for mobile non-target organisms;
- predict the role of indirect effects on ecological recovery at the landscape level;
- select appropriate spatial and temporal scales, and key landscape properties for the assessment of impact and recovery of different organism groups and for determining the most optimal management and/or mitigation decisions;
- operationalise links between experimentation, modelling and monitoring, and between prospective and retrospective studies, to consolidate risk assessments;
- parameterise population and food-web models including uncertainty;
- establish predictive food-web and/or ecological interaction models that can be used in prospective ERA;
- develop good mechanistic effect models which are both manageable and realistic enough;
- integrate systems approaches and multiple (potential) stressors into ERA.