The Consortium on Risk in the Environment: Diagnostics, Integration, Benchmarking, Learning and Elicitation (CREDIBLE)

Lead Research Organisation: University of Bristol
Department Name: Civil Engineering

Abstract

Natural hazard events claim thousands of lives every year, and financial losses amount to billions of dollars. The risk of losing wealth through natural hazard events is now increasing at a rate that exceeds the rate of wealth creation. Therefore natural hazards risk managers have the potential, through well-informed actions, to significantly reduce social impacts and to conserve economic assets. By extension, environmental science, through informing the risk manager's actions, can leverage research investment in the low millions into recurring social and economic benefits measured in billions. However, to be truly effective in this role, environmental science must explicitly recognize the presence and implications of uncertainty in risk assessment.

Uncertainty is ubiquitous in natural hazards, arising both from the inherent unpredictability of the hazard events themselves, and from the complex way in which these events interact with their environment, and with people. It is also very complicated, with structure in space and time (e.g. the clustering of storms), measurements that are sparse especially for large-magnitude events, and losses that are typically highly non-linear functions of hazard magnitude. The tendency among natural hazard scientists and risk managers (eg actuaries in insurance companies) is to assess the 'simple' uncertainty explicitly, and assign the rest to a large margin for error.

The first objective of our project is to introduce statistical techniques that allow some of the uncertainty to be moved out of the margin for error and back into an explicit representation, which will substantially improve the transparency and defensibility of uncertainty and risk assessment. Obvious candidates for this are hazard models fitted on a catalogue of previous events (for which we can introduce uncertainty about model parameters, and about the model class), and limitations in the model of the 'footprint' of the hazard on the environment, and the losses that follow from a hazard event.

The second objective is to develop methods that allow us to assess less quantifiable aspects of uncertainty, such as probabilities attached to future scenarios (eg greenhouse gas emissions scenarios, or population growth projections). The third objective is to improve the visualisation and communication of uncertainty and risk, in order to promote a shared ownership of choices between actions, and close the gap between the intention to act (eg, to build a levee, or relocate a group of people living in a high-risk zone) and the completion of the act. In natural hazards this gap can be large, because the cost of the act is high, many people may be affected, and the act may take several years to complete.

Ultimately, everyone benefits from better risk management for natural hazards, although the nature of the benefits will depend on location. In the UK, for example, the primary hazard is flooding, and this is an area of particular uncertainty, as rainfall and coastal storm surges are likely to be affected by changes in the climate. A second hazard is drought, leading to heat stress and water shortages. Our project has explicit strands on inland flooding, wind-storms, and droughts. Other parts of the world are more affected by volcanoes or by earthquakes, and our project has strands on volcanic ash, debris flows as found in volcanic eruptions (ie lahars; avalanches are similar), and earthquakes. In the future, new hazards might emerge, such as the effect of space weather on communications. A key part of our project is to develop generic methods that work across hazards, both current and emerging.

Planned Impact

The uncertainty framework developed by the CREDIBLE project will bring new approaches into the field of natural hazards and adapt them to the specific needs of this field. It will create consistency and greater scientific rigour regarding the estimation of uncertainty in natural hazard risk assessment. Thus enhancing capacity, knowledge and skills of stakeholders from private and public sectors, and improve societal security through better and more consistently informed decision-making under uncertainty.

Beneficiaries of the proposed research include the whole range of sectors involved in risk assessment of natural hazards, which is reflected in our list of project partners. These included the insurance and finance sector (Willis Ltd., Lighthill Risk Network, Catlin Underwriting Agencies Ltd), consultants (HR Wallingford, JBA Ltd, RMS), and the UK government agencies (Environment Agency, Met Office, DEFRA).

The insurance industry will benefit from the more accurate pricing of contracts. As we explain elsewhere in this proposal, currently many uncertainties in natural hazards are acknowledged but not explicitly quantified, and enter into insurance premiums as a loading for risk, which is effectively a margin for error on top of the expected loss. Our intention in CREDIBLE is to move some of these uncertainties out of the margin for error and into explicit assessment, thus improving the pricing of risk premiums, and, ultimately, the economic performance of UK insurance companies connected to our project partners. We will also introduce standard statistical tools for improving efficiency in Monte Carlo simulations, and for quantifying variability in Monte Carlo estimates, which will be immediately taken up by CAT modelling companies (notably our project partner RMS), and feed through into regulation, for example through quantifying variability in the estimate of the 99.5th percentile, as required by the EU Solvency 2 directive.

Consultants and agencies will benefit from more powerful tools for assessing uncertainty, and for propagating it into the choice between actions. Our environmental consulting project partners (JBA Consulting and HR Wallingford) focus primarily on flooding, which is also a major concern for our agency project partners (the Environment Agency, the UK Met Office, and DEFRA). In flooding it is acknowledged that providing information about uncertainties is crucial, and this has been reflected in UK postcode-level flood maps produced by the EA (although technically these maps convey frequencies rather than probabilities). CREDIBLE's focus on (i) explicitly assessing more of the uncertainty, (ii) providing additional tools to represent less-quantifiable uncertainties, (iii) considering explicitly the link between potential actions, scenarios, uncertainties and risks, and (iv) visualisation and communication, will provide a more transparent and defensible assessment of different actions and consequences. This will promote a shared ownership of some of the very contentious issues that can arise in natural hazards, such as developments on flood-plains, or groups of people who choose to live in the high risk zones of active volcanoes.

This final aspect is an absolutely crucial part of CREDIBLE, and of our intention to have an enduring effect on natural hazards risk management, and the quality of life of people who are affected by natural hazards. Natural hazards interventions (eg building a levee or a barrage, changing building regulations, relocating a group of people) are almost always contentious, costly, and take several years. There is a large gap, therefore, between the intention to act, and the completion of the action. CREDIBLE can reduce the size of this gap by promoting a shared sense of ownership, of the science and of the decision.

Publications

10 25 50
publication icon
Almeida S (2015) Accounting for dependencies in regionalized signatures for predictions in ungauged catchments in Hydrology and Earth System Sciences Discussions

publication icon
Almeida S (2016) Dealing with deep uncertainties in landslide modelling for disaster risk reduction under climate change in Natural Hazards and Earth System Sciences Discussions

publication icon
Beven K (2015) Epistemic uncertainties and natural hazard risk assessment - Part 1: A review of the issues in Natural Hazards and Earth System Sciences Discussions

 
Description SAFE (Sensitivity Analysis for Everyone) Toolbox
The SAFE provides a set of functions to perform Global Sensitivity Analysis in Matlab (or equivalent in the free software Octave or in R).
SAFE currently has over 250 academic users (in a wide range of application areas). It is currently being trialled by industrial users, including catastrophe risk modelling company Risk Management Solutions (RMS), environmental consultants JBA, and Airbus.
"The SAFE toolbox, and the thinking behind it, are helping us to get more out of the models we use in assessing strategies for cost-effective river water quality monitoring, ultimately supporting decisions about catchment management to meet the requirements of the European Water Framework Directive" (JBA)

Example Case Studies

1. GSA to support calibration of large-scale simulation models: application to karst systems (Hartmann et al. 2015): Karst develops through the dissolution of carbonate rock and is a major source of groundwater contributing up to half of the total drinking water supply in some European countries. Previous approaches to model future water availability in Europe are either too-small scale or do not incorporate karst processes, i.e. preferential flow paths. This study presents the first simulations of groundwater recharge in all karst regions in Europe with a parsimonious karst hydrology model. The model is calibrated by means of a novel parameter confinement strategy, based on Regional Sensitivity Analysis, which combines a priori information with recharge-related observations (actual evapotranspiration and soil moisture) at locations across Europe while explicitly identifying uncertainty in the model parameters. Simulation results are consistent with independent observations of mean annual recharge and significantly better than other global hydrology models that do not consider karst processes (PCR-GLOBWB, WaterGAP) and systematically over-estimate actual evapotranspiration and surface runoff.

2. GSA to support model calibration and evaluation using different data products: application to a land surface model (Pianosi et al. In prep): In this work, we use Variance-Based Sensitivity Analysis and Density-Based Sensitivity Analysis (PAWN) to support the calibration and evaluation of the Joint UK Land Environment Simulator (JULES), a global land surface model developed and currently employed by the UK Met Office. Sensitivity estimates are used to narrow down uncertainty in parameter ranges; to identify a set of parameterisations that would enhance the model accuracy with respect to default parameterisation, and that might be used as starting point for a finer model calibration; and to assess the information content in soil moisture data from different sensor types, including the novel cosmic-ray neutron sensor technology.
Exploitation Route The SAFE toobox can be used by any industry using numerical models. We are currently in dicsussion with several potential industry users.
Sectors Aerospace, Defence and Marine,Environment,Financial Services, and Management Consultancy,Transport

URL http://www.credible.bris.ac.uk
 
Description The SAFE provides a set of functions to perform Global Sensitivity Analysis in Matlab (or equivalent in the free software Octave or in R). SAFE currently has over 250 academic users (in a wide range of application areas). It is currently being trialled by industrial users, including catastrophe risk modelling company Risk Management Solutions (RMS), environmental consultants JBA, and Airbus. "The SAFE toolbox, and the thinking behind it, are helping us to get more out of the models we use in assessing strategies for cost-effective river water quality monitoring, ultimately supporting decisions about catchment management to meet the requirements of the European Water Framework Directive" (JBA)
First Year Of Impact 2015
Sector Aerospace, Defence and Marine,Environment,Financial Services, and Management Consultancy
Impact Types Economic

 
Description CREDIBLE Summer School
Geographic Reach National 
Policy Influence Type Influenced training of practitioners or researchers
Impact Improved statistical skills in insurance and other sectors.
 
Description CREDIBLE Summer School
Geographic Reach Asia 
Policy Influence Type Influenced training of practitioners or researchers
 
Description Influenced training of practitioners or researchers - CREDIBLE Summer School
Geographic Reach National 
Policy Influence Type Influenced training of practitioners or researchers
Impact Improved statistical skills in insurance and other sectors.
 
Description Influenced training of practitioners or researchers - CREDIBLE Summer School
Geographic Reach National 
Policy Influence Type Influenced training of practitioners or researchers
Impact Improved statistical skills in insurance and other sectors.
 
Description National Flood Resilience Science Advisory Group
Geographic Reach National 
Policy Influence Type Membership of a guideline committee
 
Description Volcanic ash cloud forecasting: combining satellite observations and dispersion modelling - Met Office Name User workshop, Met Office
Geographic Reach Local/Municipal/Regional 
Policy Influence Type Influenced training of practitioners or researchers
 
Description ENVIRONMENTAL RISKS TO INFRASTRUCTURE INNOVATION PROGRAMME
Amount £39,591 (GBP)
Funding ID NE/M008746/1 
Organisation Natural Environment Research Council 
Sector Public
Country United Kingdom
Start 10/2014 
End 03/2015
 
Description Environmental Change Challenge Fellowships
Amount £723,003 (GBP)
Funding ID EP/R007330/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Academic/University
Country United Kingdom
Start 07/2017 
End 06/2022
 
Description Innovation
Amount £10,747 (GBP)
Funding ID NE/N007700/1 
Organisation Natural Environment Research Council 
Sector Public
Country United Kingdom
Start 07/2015 
End 10/2015
 
Description Innovation
Amount £17,028 (GBP)
Funding ID NE/N004817/1 
Organisation Natural Environment Research Council 
Sector Public
Country United Kingdom
Start 01/2015 
End 06/2015
 
Description Knowledge Exchange (FEC)
Amount £146,270 (GBP)
Funding ID NE/M008878/1 
Organisation Natural Environment Research Council 
Sector Public
Country United Kingdom
Start 10/2014 
End 04/2016
 
Description Knowledge Exchange Fellowships
Amount £36,876 (GBP)
Funding ID NE/P013155/1 
Organisation Natural Environment Research Council 
Sector Public
Country United Kingdom
Start 10/2016 
End 03/2017
 
Title 'calibrater' 
Description A new R package freely available on-line 
Type Of Material Improvements to research infrastructure 
Year Produced 2015 
Provided To Others? Yes  
Impact TBC 
 
Title CURE 
Description CURE is a fully open source toolbox written in MATLABTM It employs a range of different Monte Carlo methods for forward UE (i.e. the forward propagation of uncertainties through a model: e.g. using prior estimates of input and parameter uncertainty) and conditioned UE (i.e. where UE results are conditioned on observations). Each approach and modelling application includes its own implicit and explicit modelling choices and assumptions recorded via the GUI. The GUI takes the form of a number of simple, sequential dialogue boxes where the user is asked to enter information, as text, in a structured way and which can be iteratively edited during any modifications to analyses. The toolbox structure is such that new methods can be easily added and it will be subject to ongoing development and augmentation with additional workflow examples. 
Type Of Material Improvements to research infrastructure 
Provided To Others? No  
Impact TBC 
 
Title SAFE 
Description The SAFE Toolbox provides a set of functions to perform Global Sensitivity Analysis in Matlab/Octave environment. It implements several methods, including the Elementary Effects Test, Regional Sensitivity Analysis, Variance-Based (Sobol') sensitivity analysis. Features: • Modular structure to facilitate interactions with other computing environments • Set of functions to assess the robustness and convergence of sensitivity indices • Several visualization tools to investigate and communicate GSA results • Lots of comments in the code and workflow examples to get started The SAFE Toolbox is freely available to academic and non-commercial users. Authors: F. Pianosi, F. Sarrazin, T. Wagener 
Type Of Material Improvements to research infrastructure 
Year Produced 2015 
Provided To Others? Yes  
Impact SAFE currently has over 300 academic users in a wide range of application areas. It is being trialled by industrial users, including catastrophe risk modelling companies, environmental consultants, technology and manufacturing companies. 
URL https://www.safetoolbox.info/
 
Title Calibrate R 
Description A package of useful functions for calibrating a computer simulator. 
Type Of Technology Software 
Year Produced 2014 
Open Source License? Yes  
Impact TBC 
URL http://www.maths.bris.ac.uk/~mazjcr/calibrate_0.4.tar.gz
 
Title SAFE - Sensitivity Analysis For Everybody 
Description A Matlab Package for sensitivity analysis of computer models. 
Type Of Technology Software 
Year Produced 2014 
Open Source License? Yes  
Impact TBC 
 
Title SAFER - Sensitivity Analysis For Everybody in R 
Description An R Package for sensitivity analysis of computer models. 
Type Of Technology Software 
Year Produced 2014 
Open Source License? Yes  
Impact TBC 
 
Description Member of the Cabinet Office Civil Contingencies Secretariat advisory group on effusive volcanic eruptions, also SAGE 
Form Of Engagement Activity A formal working group, expert panel or dialogue
Part Of Official Scheme? Yes
Geographic Reach National
Primary Audience Policymakers/politicians
Results and Impact TBC
Year(s) Of Engagement Activity 2013,2014
 
Description Panellist on Computer modelling: all about the image?, Sunday 20 October 2013, Battle of Ideas Festival 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Public/other audiences
Results and Impact Talk sparked questions and discussion afterwards.

TBC
Year(s) Of Engagement Activity 2013
URL http://www.battleofideas.org.uk/2013/session_detail/7900