- assessment factor
- breakout groups
- chemical classes
- climatic zone
- Environmental protection agency
- Hazard assessment
- marine ecotoxicity
- marine hazard assessment
- probabilistic approach
- probabilistic techniques
- species sensitivity distribution
- test species
- toxicity data
WR 15 : The Probabilistic Approaches for Marine Hazard Assessment | June 2009
In order to discuss the potential use of probabilistic techniques in marine environmental hazard assessment, a workshop was organised by ECETOC and the Environmental Agency of England and Wales. Nearly 40 scientific experts from industry, academia and governmental agencies participated in the meeting hosted by the Norwegian Pollution Control Authority and held in Oslo on 18th and 19th June 2008. Seven plenary sessions were followed by four syndicate sessions, each addressing specific issues.
The overall conclusion was that the assessment factor approach described in the EU Technical Guidance Document is suitable for screening purposes to derive the initial PNECmarine, while the probabilistic approach (using species sensitivity distribution [SSD]) provides a refined estimation of marine species sensitivity. Both of these approaches should be used in a tiered assessment manner. Both the assessment factor and SSD approaches should prioritise population relevant endpoints (e.g. survival, development, growth or reproduction) in order to directly relate to ecologically important effects. The workshop also recognised that the probabilistic approach is inherently complex, data intensive and requires advanced statistical expertise. With SSD approaches, variation in species sensitivities and some of the associated uncertainty can be taken into account. However, both the assessment factor and probabilistic hazard assessment approaches pay less attention to potential uncertainty due to selection of test species, analytical chemistry and variation between laboratories and tests. It was also highlighted that in some cases (e.g. when the identified species have protected status, specific economic value or are the key species for a specific habitat) it may be advisable to initiate a third tier assessment and perform a species-specific, biologically based assessment.
The participants agreed that there are several problems with combining species in SSDs, mainly when species are combined over large taxonomic distances. In addition, extrapolation from acute to chronic toxicity may be based on different endpoints in bioassays (e.g. LC50 for acute tests with marine organisms), while chronic tests are based on growth or reproduction (and expressed as NOEC [No Observed Effect Concentration]). Further research is needed on these issues.
Workshop participants also discussed whether it is scientifically justified to combine saltwater and freshwater effects data. The benefit of combining data is to render the probabilistic approach more robust (e.g. with a larger set of acute or chronic data for analysis). Nonetheless, due caution is required when choosing compounds that may interact with compounds present in seawater, (e.g. metals or organic chemicals with the potential to affect osmoregulation). For such compounds, the SSD approach should be used on marine specific data. Available data suggests that the toxicity data from organisms in one climatic zone may be extrapolated to another zone (e.g. use of data from temperate marine species in the absence of data for Arctic species). However, it is not clear whether the environmental fate and bioavailability of the substance is different in polar versus temperate regions. While there was not a consensus on this issue, several workshop participants did recommend that further research is necessary into this matter.
The probabilistic hazard assessment approach is beginning to be applied for freshwater, marine and terrestrial risk assessment scenarios worldwide. Some key questions still remain, however, related to the comparison of acute or chronic effects data between freshwater and marine groups of organisms (it was noted that the global marine environment contains 16 animal phyla not found in freshwater ecosystems). There is also a need to strengthen the evidence base for extrapolating marine ecotoxicity data from one climatic zone to another.
A programme of research was proposed that would help resolve a number of these issues and which would lead to improvements to the current approaches being used in marine hazard assessment.