Models, decision-making and science

J. Doherty, R. Vogwill

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

2 Citations (Scopus)

Abstract

Management of groundwater systems relies heavily on groundwater models. They are often commissioned by one party and then used by another. Assurance of a model’s quality often rests on compliance with guidelines or standards. This chapter argues that integrity of model-based decision-support requires more than this. It argues that the decision-making process requires nothing less of modelling than that it implements the scientific method. This requires recognition of the stochastic nature of expert knowledge on the one hand, and the limitations of history matching in refining that knowledge on the other hand. Model predictions must therefore be viewed as probabilistic. As such they can form a basis for risk assessment, this being a vital component of any decision-making process. To fulfil this role models must be used in partnership with equally sophisticated stochastic and inversion software. However, regardless of the level of modelling sophistication, assessment of risk can only ever be subjective. In making the innumerable subjective decisions that the modelling process demands, a modeller’s reference point must necessarily be avoidance of failure of the modelling exercise. This happens if the risk of occurrence of an unwanted event is assessed to be lower than it actually is.
Original languageEnglish
Title of host publicationSolving the Groundwater Challenges of the 21st Century
EditorsRyan Vogwill
PublisherCRC Press/Balkema
Chapter7
Pages95-113
Number of pages19
ISBN (Electronic)9781315685182
ISBN (Print)9781138027473
Publication statusPublished - 2016

Keywords

  • Groundwater modeling
  • model-based decision-support
  • Model predictions

Fingerprint

Dive into the research topics of 'Models, decision-making and science'. Together they form a unique fingerprint.

Cite this