THE IMPORTANCE OF BEING UNCERTAIN

摘要

We are often faced with uncertainty when makingdecisions – from trivial decisions such as whether totake an umbrella, or major decisions such as whether tobuy that house. Appreciating the uncertainty in futureconditions (‘will it rain today?’; ‘will house prices continueto go up?’) is crucial to making good decisions. This is nodifferent for water resource managers, who need to makedecisions on flood prevention, climate adaptation or coalresource developments. As scientists, we strive to informdecision-makers about uncertainties in a comprehensive,unbiased and transparent manner.In this paper, we discuss some of the challenges andapproaches used to communicate uncertainty during ourcontributions to the Bioregional Assessments Program,a federally-funded research project to assess thepotential impacts of coal resource development on waterresources and water-dependent assets in central andeastern Australia.A first step in analysing potential impacts, is to identify thecausal pathways that detail how development activities canpossibly affect the groundwater and surface water systems,and how these changes might affect the economic, socialand ecological functioning of a region. This conceptualmodel provides the framework for detailed geological,hydrogeological, hydrological and ecological modelling.Predictions have traditionally been made using a singledeterministic model, a calibrated model that best fits theavailable observations. However, to assess the likelihoodof potential impacts, we used a stochastic approach tocreate an ensemble of possible predictions (hundredsand thousands of possible answers) that are all consistentwith the available observations. This results in a range or distribution of predictions. However, it is very challengingto communicate the range of model results, as well as allof the complexities and underlying assumptions, in a waythat is relevant and accessible to decision-makers..For bioregional assessments, we have worked withdecision-makers to improve communication ofuncertainty. This resulted in a consistent, calibratedlanguage, tables, plots of the range of predictionsand maps designed to convey probabilisticinformation in an intuitive manner. Further,model details and assumptions aredocumented in technical reports, andthe data, models and predictions aremade publically available to increasetransparency and reproducibility.The amount and technical detail ofthat information can be challengingfor decision-makers to identify whatis important and what is not. Tosupport decision-makers, we usea qualitative uncertainty analysisto summarise the rationale for, andeffect on, prediction of each majorassumption. This table, in combinationwith a plain English discussion, allowsreaders to rapidly appreciate thelimitations, as well as opportunities forfurther data collection or modelling.Bioregional assessments have highlightedthe importance of early consultation with targetaudiences, which has enabled us to tailor the uncertaintycommunication products for decision-makers, as well asto avoid the potential for biased interpretation of results,where decision-makers are drawn to the extremes.