Addressing uncertainties in the ERICA Integrated Approach
Introduction
As with any risk assessment, the evaluation of the impact of ionising radiation on non-human biota is complicated by a variety of sources of uncertainty. At all stages the assessments require the use of models, scenarios, assumptions and extrapolations. Knowledge bases are characterised by partly irreducible and often unquantifiable uncertainties on measurements and parameters (e.g. ecosystem variability and large data gaps on transfer coefficients for many radionuclides). Ecological risk assessments have a variety of mechanisms for addressing such uncertainties, including the application of safety factors, probabilistic risk analysis and the incorporation of conservatism within models, many of which have been applied within the ERICA Integrated Approach. ERICA has tried to deal openly with the various dimensions of uncertainty, recognising that uncertainty is intrinsic to complex systems and that not all uncertainties can be quantified numerically (van der Sluijs, 2002, van der Sluijs, 2007, Funtowicz and Ravetz, 1993).
During the development of the ERICA Integrated Approach, the importance of uncertainties has been raised a number of times, by both the consortium and the ERICA End Users Group (EUG). Recurring themes for the EUG were the significance of in-built conservatism in the model, the requirement for transparency in the choices and assumptions made in parameter selection, and ensuring that users neither trust the results too uncritically nor overcompensate for uncertainty (Oughton et al., 2004, Zinger, 2005, Oughton and Brevik, 2006, Forsberg and Oughton, 2006). For example, the “… need for transparency and traceability in the way the tool deals with uncertainty …” and awareness that “… ERICA has several types of intrinsic uncertainties and that some conservatism already is built-in to compensate for those …” (Oughton and Brevik, 2006). Other recommendations were that ERICA should address not only data issues (i.e. model parameters and input data) but also the uncertainties inherent in the ERICA tool (i.e. the model structure and scenario assumptions), and those associated with problem formulation and stakeholder involvement.
In order to make decisions about addressing uncertainties, it is important to understand the types of uncertainty, their significance under different situations and the options that exist for dealing with them – in general and within the specific context of the ERICA tool. These include development of approaches for mapping and prioritising the readily quantifiable as well as the broader aspects of uncertainty; and methods for addressing uncertainty within the various tiers of the assessment tool. Some of the methods are intrinsic to the tool and its use, while for others the tool simply provides guidance, recommending that the user considers the implications of uncertainties in the selection of parameters and interpretation of results. This paper will give an overview of these approaches. First, a summary of types and sources of uncertainty is presented, followed by an evaluation of different degrees of conservatism and approaches for addressing uncertainty in the three tiers. More detailed information on all the methodology presented can be found in the ERICA reports (Agüero et al., 2006, Oughton and Brevik, 2006, Beresford et al., 2007, Zinger et al., 2007, Brown et al., 2008).
Section snippets
Definitions and typology
Uncertainties can be categorised in various ways. One of the most conventional and widespread distinctions is that between uncertainty and variability (Warren-Hicks and Moore, 1998, Suter, 1993, Suter et al., 2000).
- •
Knowledge uncertainty (Type I uncertainty) is defined as a lack of scientific knowledge about specific factors, scenarios, parameters or models. It includes both quantitative (e.g. measurement errors) and qualitative sources (e.g. model misrepresentation) and can be often reduced
Uncertainties and assumptions in the ERICA tool
Within any assessment, the description of uncertainty can help to trace uncertainties, ease refinements of the analysis, and prioritise research needs. The key sources of variability and uncertainty in exposure and effect analyses need to be clearly identified and to be discussed to improve the transparency of any risk characterisation. Since the uncertainties and assumptions may be site and context dependent for any ecological risk assessment, the ERICA Integrated Approach has produced a
Conclusions
The ERICA Integrated Approach includes a number of methods for addressing various types of uncertainty during ecological risk assessment, and this paper has provided a brief summary of the approaches available. These range from simple characterisation of the different types and sources of uncertainty to probabilistic risk assessment and sensitivity analysis. Options are available for dealing with both the knowledge-based uncertainties (i.e. data gaps and model limitations) and variability-based
Acknowledgements
The authors would like to express their thanks to Carol Robinson, Enviros, for her contribution to work on uncertainty analysis within the ERICA Integrated Approach. The financial support of the European Commission through the 6th Framework ERICA project, Contract no. FI6R-CT-2004-508847, is gratefully acknowledged.
References (40)
- et al.
The ERICA Tool
J. Environ. Radioact.
(2008) Ecological risk assessment: risk of what? How do we decide?
Ecotox. Environ. Saf.
(1998)- et al.
Choosing an alpha radiation weighting factor for doses to non-human biota
J. Environ. Radioact
(2006) - et al.
Issues and practices in the use of effects data from FREDERICA in the ERICA Integrated Approach
J. Environ. Radioact.
(2008) Ethical issues in protection of the environment from ionising radiation
J. Environ. Radioact.
(2003)A way out of the credibility crisis of models used in integrated environmental assessment
Futures
(2002)Uncertainty and precaution in environmental management: insights from the UPEM conference
Environ. Modell. Softw
(2007)- et al.
Stakeholder interaction within the ERICA Integrated Approach
J. Environ. Radioact.
(2008) - et al.
- et al.
Study of the uncertainty in estimation of the exposure of non-human biota to ionising radiation
J. Radiol. Prot.
(2004)
Report of the Committee Examining Radiation Risks of Internal Emitters (CERRIE), London
A critical evaluation of safety (uncertainty factors) for ecological risk assessment
Environ. Toxicol. Chem.
EURACHEM/CITAC Guide: Quantifying Uncertainty in Analytical Measurement
Statistical Distributions
Species sensitivity distributions revisited: a critical appraisal
Hum. Ecol. Risk Assess.
Extrapolation in ecological risk assessment: balancing pragmatism and precaution in chemicals controls legislation
BioScience
A critique of the use of distribution based extrapolation models in ecotoxicology
Funct. Ecol.
Cited by (33)
Environmental radiological risk assessment of a coal ash and slag disposal site with the use of the ERICA Tool
2019, Journal of Environmental RadioactivityCitation Excerpt :It is linked to the FREDERICA Radiation Effects Database (Copplestone et al., 2008), which summarises the dose-effect relationships between radiation exposures and their effects on organisms. A detailed description of the Tool, its features and the related ERICA integrated approach can be found in other publications (Brown et al., 2016, 2008; Howard and Larsson, 2008; Oughton et al., 2008; Zinger et al., 2008). The Tool has already been used on different types and sources of real or potential environmental radiological concerns such as nuclear accidents (Fuma et al., 2015), nuclear power plant discharges (Vandenhove et al., 2013), uranium mining sites (Doering and Bollhöfer, 2016), coal-fired power plant (Čujić and Dragović, 2017), phosphate industry (Vandenhove et al., 2015), deep geological disposal facilities (Jaeschke et al., 2013) and releases from medical facilities (Veliscek Carolan et al., 2011).
Environmental impact and risk assessments and key factors contributing to the overall uncertainties
2016, Journal of Environmental RadioactivityCitation Excerpt :As demonstrated in mixed exposure experiments, low Se status can lead to reduced fertility in male mice (F0) and sterility for male offspring (F1), while the combination with gamma radiation aggravated these effects (Graupner et al., 2014). Risk quotients (i.e., PEC/PNEC ratios) for individual radionuclides are frequently used in environmental risk assessments such as the ERICA tool (Oughton et al., 2007). So far, attempts to assess the total risks from more than one stressor are made by summarizing risk quotients for mixed substances.
Evaluating variability and uncertainty in radiological impact assessment using SYMBIOSE
2015, Journal of Environmental RadioactivityApproaches to providing missing transfer parameter values in the ERICA Tool - How well do they work?
2013, Journal of Environmental RadioactivityCitation Excerpt :The probabilistic functionality of Tier 3 of the ERICA Tool was used to derive the 95th percentiles. The ERICA Tool default values which had been derived using extrapolation approaches were assumed to represent the arithmetic mean and the model run assuming that the underlying distribution was exponential; this is compatible with how these values were treated in the derivation of the ERICA Tool EMCL values (see Brown et al., 2008; Oughton et al., 2008). For the newly acquired empirical data, the arithmetic mean and standard deviation were entered and the underlying distribution was assumed to be log-normal, once more in line with the approach used by Brown et al. (2008) to derive EMCL values from empirical data.
Assessment of doses to game animals in finland
2013, Journal of Environmental RadioactivityMulti-criteria decision-making methods for the optimal location of construction and demolition waste (C&DW) recycling facilities
2013, Handbook of Recycled Concrete and Demolition Waste