Transcript Slide 1

ENGAGING STAKEHOLDERS WITH
UNCERTAINTY
Hazel Faulkner &
Simon McCarthy
?
www.floodrisk.org.uk
Funders:
EPSRC Grant: EP/FP202511/1
FHRC research for FRMRC1/2
INTERVIEWS with
•Environment Agency professionals
•Insurers
•Floodplain planners
How can FRM optimise risk and uncertainty
communications at the professional interface?
Why the reluctance to use uncertainty tools
....barriers?
www.floodrisk.org.uk
Funders:
EPSRC Grant: EP/FP202511/1
BARRIERS?
MANY SCIENTIFIC ,PROFESSIONAL AND LAY STAKEHOLDERS
stakeholders at sources of information
‘science’
As webs of
influence vary,
agendas vary,
information needs
vary,
Stakeholders in information
pathways
professionals
floodplain
stakeholders
risk communication
strategies &
appropriate tools
will be totally different
stakeholders outside
floodplain
Information receptors, Category II
Information receptors, category I
www.floodrisk.org.uk
Funders:
EPSRC Grant: EP/FP202511/1
BARRIERS?
SCIENTIFIC AND DECISION MAKERS DIFFER IN THEIR
UNDERSTANDING OF THE SOURCES OF UNCERTAINTY
SCIENTIFIC UNCERTAINTY
Uncertainty in the science of flood forecasting and runoff
prediction models are largely associated with their assumptions,
structure, and boundary conditions, and confidence in validation
procedures given uncertainties about climatic and societal
futures
PROFESSIONALS experience BINARY DECISION
UNCERTAINTY (Decision rule uncertainty)
The implication of this is that scientific uncertainty is an
relatively unwelcome part of the risk message they are
charged with translating for the public
www.floodrisk.org.uk
Funders:
EPSRC Grant: EP/FP202511/1
BARRIERS?
THE LANGUAGE OF SCIENCE IS OPAQUE TO THE
NON-SCIENTIFIC PROFESSIONAL
If originally formulated in Bayesian terms, the language may be
too opaque for translation to be effective – do professionals
have sufficient statistical familiarity?
• Bayesian statistics ?
• Prior probability distributions?
• fuzzy set methods ?
• info-gap methods ?
• NUSAP?
match tool to communication
interface
The implication of this is that the language used to
communicate the uncertainty must match the needs
and agenda of the agencies involved in the
communication being undertaken
www.floodrisk.org.uk
Funders:
EPSRC Grant: EP/FP202511/1
THE WAY FORWARD: TRANSLATIONAL DISCOURSES?
LANGUAGE AND TOOLS TO DESCRIBE UNCERTAINTY VARY
Rt = H x V
stakeholders at sources of information
Phone warnings
Newspaper/TV
1;100/1;1000
fuzzy edged
‘science’
/twitter
Bayesian
uncertainty tools –
GLUE
Stakeholders in information
pathways
professionals
stakeholder consultation
’traffic lights’
webpages/leaflets
floodplain
stakeholders
stakeholders outside
floodplain
Information receptors, Category II
Information receptors, category I
www.floodrisk.org.uk
Funders:
EPSRC Grant: EP/FP202511/1
FINDINGS
FINDINGS
Our research (Faulkner et al 2007;McCarthy et al. 2009), both the
interviews undertaken with professionals and from the ‘experiment’
undertaken at the co-location workshop in Exeter, found that :
The power of VISUALISATIONS AND ANIMATIONS in realising the
uncertainty estimates was potentially great;
The professionals questioned initially struggled to comprehend
scientifically defined flood forecast uncertainties (probabilistic
and/or ensemble forecasts) without FURTHER TRANSLATION OF THE
SCIENCE. When this was available as perhaps a ‘translational
discourse’, the preparedness to embrace a more sophisticated
expression of the model’s uncertainties was welcomed.
Better DECISION-SUPPORT TOOLS are needed.
www.floodrisk.org.uk
Funders:
EPSRC Grant: EP/FP202511/1
The WIKI Decision-support tool
Decision Tree and Wiki
Pages at
http://www.floodrisknet.org.
uk/methods/
www.floodrisk.org.uk
Funders:
EPSRC Grant: EP/FP202511/1
A ‘TRANSLATIONAL DISCOURSE’
BRAINSTORMING
Brainstorming sessions involving professionals/scientists
and practitioners
A GUIDANCE MANUAL
how to involve all stakeholders, especially professionals at
local level is needed – professional input here important;
MORE DECISION SUPPORT TOOLS NEEDED......
www.floodrisk.org.uk
Funders:
EPSRC Grant: EP/FP202511/1
Acknowledgement
The research reported in this presentation was conducted as part
of the Flood Risk Management Research Consortium with support
from the:
– Engineering and Physical Sciences Research Council
– Department of Environment, Food and Rural
Affairs/Environment Agency Joint Research Programme
– United Kingdom Water Industry Research
– Office of Public Works Dublin
– Northern Ireland Rivers Agency
Data were provided by the EA and the Ordnance Survey.
www.floodrisk.org.uk
Funders:
EPSRC Grant: EP/FP202511/1