Technical report 125

Discussion

In considering the EFSA framework developed for pesticides in a broader chemical context and in applying the framework to several case studies, the Task Force found the approach to be conceptually straightforward and logical. However, there were many points in the framework where additional information and more detailed guidance will be required for general applicability to all chemical sectors, including pesticides. A strong theme throughout the Task Force application of the framework was the importance of prioritising at each step in order to manage the time and effort required.  This discussion outlines the key development needs that the Task Force identified at each step.

Steps 1 and 2: Construct a habitat x ecosystem service matrix and assign importance rankings

The first two steps can be considered as i) the development of a reference table of habitats and ii) their importance for ecosystem service provision. This is essential information for the framework approach and although these two steps were considered in that order for the case studies, identification of which habitats would be expected to be exposed to specific chemicals would also be an initial step in applying the framework.

The habitat x ecosystem service matrix provides a flexible method for selecting relevant habitats and then drawing on expert assessments of the importance these habitats can be in providing ESs. The Task Force considered the EUNIS habitat typology a good, multi-level hierarchical classification. Although the case studies developed by the Task Force generally applied habitat classifications at a similar hierarchical level across all habitats, in principle, the hierarchy could effectively use different levels of resolution as required. It is clear that the matrix presented in Table 2.3 requires further work to extend the assessment to all combinations of habitats and ecosystem services. Levels of importance (+ to +++) were collated from key publications that compared multiple ecosystem services across multiple habitats.  Additional information was added by the Task Force where supporting knowledge was available to enable the case studies to better assess specific habitat importance. These were mostly the marine habitats (i.e marine inlets and transitional waters; coastal areas; shelf; open ocean). Sparsely vegetated land was not generally considered in the Task Force case studies, because of low exposure in most case studies and insufficient knowledge.

The matrix can be used with various levels of habitat resolution, for example, all fresh water habitats could be considered as one generic habitat or could be sub-divided into lotic and lentic habitats. Further differentiation of lotic or lentic habitats might also be appropriate for specific chemical emissions, although this level of information would require further development. The down the drain chemicals case study assessed lotic and lentic fresh water habitats separately since exposure of lotic systems was expected to be higher than lentic systems in most cases.

The use of all types of ecosystem services in the initial steps of the framework, as recommended by Maes et al (2014) and EFSA (2010), was considered important in identifying the key SPUs. The Task Force did not consider the completeness of the list but did not identify any gaps arising from the four case studies. Deviations from the EFSA approach included the combining of primary production with photosynthesis where the Task Force considered the SPUs to be essentially similar and the exclusion of abiotic ecosystem services such as oil (for fuel) and flowing water (for power generation), since these were not provided via biotic SPUs. Explicitly including SPUs that provide supporting and other intermediate services was considered a more explicit and informed approach to deriving key groups of SPUs and, therefore, in any subsequent identification of testing strategies for risk assessing the potential impacts on SPGs.

The treatment of biodiversity in the habitat x ecosystem service matrix was identified as a topic requiring further discussion. The Task Force recognised the importance of addressing biodiversity in relation to ecosystem services and adopted the position that biodiversity underpins the delivery of all ecosystem services that are dependent on biotic processes and specific components of biodiversity are explicitly addressed in many individual ecosystem services (e.g. genetic resources, ornamental resources, pollination, pest control, aesthetic value, etc).  Biodiversity, as defined by the Convention on Biological Diversity, was considered part of natural capital and not an ecosystem service per se as its inclusion as an ecosystem service would lead to the protection of ‘everything, everywhere’, which is too generic and vague to be useful for scientific risk assessment.  Therefore, the TF did not consider biodiversity as a discrete ecosystem service.  The Task Force identified potential confusion between genetic resources and biodiversity, i.e. both terms could be interpreted as meaning the same ecosystem service. These are defined as completely different ecosystem services and misinterpreting genetic diversity with biodiversity also adds to the issue outlined above. Familiarity with the definitions of ecosystem services and other terms is an important requirement if the EFSA framework is to be applied correctly and efficiently.

Step 3: Ranking potential impact for habitat x ecosystem service combinations using exposure and effects information

The Task Force found the preparation of schematic diagrams of potential routes of exposure helpful in assessing the relative level of exposure each of the habitats could experience from specific chemicals in the case studies. Inclusion of such schematic figures provides a simple and effective communication of exposure. The use of a three coloured traffic light approach proved adequate in ranking and differentiating levels of concern. However, the Task Force observed that different individuals scored (coloured) some cells differently in different case studies, i.e. there were differences in judgement of level of concern. Additional experience and guidance would help minimise such differences. In those case studies where chemical exposure or importance of the habitat for specific ecosystem service provision was negligible for a habitat or a specific ecosystem service, then that row or column in the matrix was dropped from further consideration, e.g. urban, woodland and forest, heathland and shrub, sparsely vegetated land, open ocean. Note that forest habitats could potentially be impacted by POPs, particularly in high altitude Alpine areas, in which atmospheric distillation takes place. Again, the use of a simple traffic light approach helps identify these cases.

The Task Force initially aimed to only use the relative level of exposure to rank the level of concern for each
habitat x ecosystem service combination. Although exposure was indeed acknowledged as the main driver along with importance of habitats for ecosystem service provision, additional chemical-related factors were also identified and applied. These refinements are described in Chapter 4 and include use of prior knowledge of chemical fate, behaviour and toxicity such as contamination of the (human) food chain, the range of potentially impacted species (more species exposed may lead to broader potential impacts, unless there is scope for compensation via functional redundancy) and the potential for both direct and indirect chemical impacts. There was also potential for additional factors to lead to reduced ranking, for example, due to the lower sensitivity of micro-organisms compared to higher organisms.

Assessing the level of potential impact due to chemical exposure was difficult for some ecosystem services. This was particularly pertinent for cultural services, for example education and inspiration could be considered likely to be always potentially impacted if the relevant habitats are exposed. Also, there can be differences in how different cultures perceive and value ecosystem services. Additional research to document and reference such differences would reduce uncertainties and inaccuracies in assessing levels of concern for impacts on cultural ecosystem services.

Although the identification of SPUs in this step is needed, the use of icons representative of the main taxonomic and functional groups was considered an appropriate level of resolution and a helpful summary at this step in the framework.

Step 4: Categorising the level of concern for exposed ecosystem services

In order to streamline the assessment of exposed habitat x ecosystem service combinations, the Task Force devised a prioritisation matrix (Chapter 2, Table 2.6). Only those combinations assessed as medium or high concern were investigated further in the case studies. This was simply to focus the Task Force resources on the combinations of higher concern, although including prioritising steps into the framework in general use is an important option to help align resources to the required level of assessment. Those combinations considered to be of low concern may still be important depending on the requirements of the specific protection goal description, i.e. how comprehensive it needs to be in informing subsequent risk assessment.

At this step, the Task Force ensured that potentially impacted service providing units in habitat and ecosystem service combinations identified as medium and high concern were identified at a suitable level of resolution for subsequent specific protection goal description. Access to reference tables of the key service providing units likely to occur in specific habitats helps complete this task and aids consistency. Since each habitat x ecosystem service cell requires relevant service providing units to be listed, the Task Force adopted a numerical superscript to simplify presentation of this information.

Step 5: Defining SPGs for ecosystem services of high and medium concern

Population and application of the tables reported in Chapter 6 was made with a high degree of uncertainty. This was because of the lack of detailed guidance and knowledge in deciding ecological entities, attributes and especially scale of potential impact. The Task Force considered that the six dimensions in EFSA’s guidance (ecological entity, attributes, magnitude of effect, temporal and spatial scale of effect and the degree of certainty required) provide a good basis for describing specific protection goals. However, more ecological knowledge is required in order for risk managers to define the acceptable magnitude and scale of impact based on underpinning science, together with other considerations. The Task Force did not assess the level of uncertainty required because of insufficient experience and guidance.

There is clearly a complex range of regulatory guidance to consider (see Chapter 3) but there remains a general lack of detail on specific protection goals in all but a few legal instruments. One notable exception is the derivation of ecological quality objectives under OSPAR, which clearly defines acceptable magnitudes, spatial and temporal scales of impact for key indicators or ecological populations in the North Sea (Table C1.3). For the most effective use of the ecosystem service approach, i.e. to utilise a habitat focus for setting specific protection goals, guidance on application of the various chemical sector specific regulations to land use scenarios is required. The Task Force considered that the use of legal requirement information should be made explicit, i.e. whether it is used to inform the specific protection goal or to use as additional information to ensure that a subsequent risk assessment is appropriately scoped.

The scope of the Task Force objectives effectively concluded with the derivation of specific protection goals for selected case studies. How these specific protection goals might be used in subsequent chemical risk assessment (prospective and retrospective) was not considered, but this is a key next step in practical application of the EFSA framework. In addition to the development of testing and modelling approaches needed to assess impacts on the service providing units that underpin specific protection goals, there is a need to define acceptable effects from unacceptable ‘adverse’ environmental effects, e.g. using retrospective or diagnostic methods.