64
3.4.1 Dealing with Uncertainty in Assessment–Management
Interactions
Estimates of fi sh stocks in the Baltic Sea, as well as elsewhere, have to deal with
complex cause–effect relationships involving many intervening variables. The risk
governance literature, which deals with complexities between cause and effect relations, calls for the best available involvement of experts in order to achieve prudent
assessment results (Renn 2008 ). With regard to EU fi sheries governance structures,
ICES takes up this task by devoting all its efforts to involve the best available (natural) scientifi c expertise in order to advice decision makers in a way that minimises
cognitive confl icts.
Uncertainty complicates complex governance arrangements of risk assessment–
risk management interactions in fi sheries. Uncertainty is a major issue in the science–policy interface of EU fi sheries management under CFP (cf. Dankel et al.
2012 ) and, as our results revealed, exists in several ways:
(a) Uncertainty in data gathering
(b) Uncertainty in data analysis
(c) Uncertainty impacts stemming from points ( a ) and ( b ) in framing, evaluation
and management
The fi shery sector is characterised by the so-called second-order uncertainty
(Renn 2008 ), i.e. a risk situation where circumstances might change in an unpredictable and unsystematic manner as in the case of fi sh stocks and the environmental
system of which they are part. Second-order uncertainty is diffi cult to communicate,
hence, our focus on it. In the case of biological assessments of fi sh stocks, it is helpful to make a distinction between aleatory and epistemic uncertainty (Renn 2008 :
71). Aleatory uncertainty characterises randomness in samples, which means that
only in the long run and with a large enough sample can the distribution of possible
values be identifi ed. Epistemic uncertainty on the other hand stems from a lack of
knowledge of dynamics or phenomena within the fi eld. Although extended data
gathering and research might decrease both aleatory and epistemic uncertainties,
with dynamic systems such as marine environments, uncertainty often prevails and
can even increase as a consequence of further research as it happened in 2014 with
the failure of ICES’ Eastern Baltic cod assessment (Eero et al. 2015 ).
Distinguishing between the two types of uncertainty and designing communication processes that make the distinction between the two obvious to the ICES
audience (the audience being broader than the clients alone) will increase transparency about different aspects of uncertainty and their role in assessment–management interactions. Such transparency about the type of uncertainty is especially
needed as uncertainty is always interpretable, i.e. a subject of ‘interpretative fl exibility’ (Meyer and Schulz-Schaeffer 2006 ), and therefore always a potential source
of confl ict in discussions amongst stakeholders in Baltic Sea fi sheries. This is most
notable in the Baltic RAC, where NGOs and fi sheries representatives read different
things from the scientifi c reports of ICES (cf. Linke et al. 2011 , 2014 ).
P. Sellke et al.
3.4.1 Dealing with Uncertainty in Assessment–Management
Interactions
Estimates of fi sh stocks in the Baltic Sea, as well as elsewhere, have to deal with
complex cause–effect relationships involving many intervening variables. The risk
governance literature, which deals with complexities between cause and effect relations, calls for the best available involvement of experts in order to achieve prudent
assessment results (Renn 2008 ). With regard to EU fi sheries governance structures,
ICES takes up this task by devoting all its efforts to involve the best available (natural) scientifi c expertise in order to advice decision makers in a way that minimises
cognitive confl icts.
Uncertainty complicates complex governance arrangements of risk assessment–
risk management interactions in fi sheries. Uncertainty is a major issue in the science–policy interface of EU fi sheries management under CFP (cf. Dankel et al.
2012 ) and, as our results revealed, exists in several ways:
(a) Uncertainty in data gathering
(b) Uncertainty in data analysis
(c) Uncertainty impacts stemming from points ( a ) and ( b ) in framing, evaluation
and management
The fi shery sector is characterised by the so-called second-order uncertainty
(Renn 2008 ), i.e. a risk situation where circumstances might change in an unpredictable and unsystematic manner as in the case of fi sh stocks and the environmental
system of which they are part. Second-order uncertainty is diffi cult to communicate,
hence, our focus on it. In the case of biological assessments of fi sh stocks, it is helpful to make a distinction between aleatory and epistemic uncertainty (Renn 2008 :
71). Aleatory uncertainty characterises randomness in samples, which means that
only in the long run and with a large enough sample can the distribution of possible
values be identifi ed. Epistemic uncertainty on the other hand stems from a lack of
knowledge of dynamics or phenomena within the fi eld. Although extended data
gathering and research might decrease both aleatory and epistemic uncertainties,
with dynamic systems such as marine environments, uncertainty often prevails and
can even increase as a consequence of further research as it happened in 2014 with
the failure of ICES’ Eastern Baltic cod assessment (Eero et al. 2015 ).
Distinguishing between the two types of uncertainty and designing communication processes that make the distinction between the two obvious to the ICES
audience (the audience being broader than the clients alone) will increase transparency about different aspects of uncertainty and their role in assessment–management interactions. Such transparency about the type of uncertainty is especially
needed as uncertainty is always interpretable, i.e. a subject of ‘interpretative fl exibility’ (Meyer and Schulz-Schaeffer 2006 ), and therefore always a potential source
of confl ict in discussions amongst stakeholders in Baltic Sea fi sheries. This is most
notable in the Baltic RAC, where NGOs and fi sheries representatives read different
things from the scientifi c reports of ICES (cf. Linke et al. 2011 , 2014 ).
P. Sellke et al.