Uncertainty assessment in coastal morphology prediction with a bayesian network

Research output: Chapter in Book/Conference proceedings/Edited volumeConference contributionScientific

75 Downloads (Pure)

Abstract

In the present time of sea-level rise and climate change a global shift has occurred toward sandy coastal protection measures and Building with Nature. These type of protection measures impose extra uncertainty on the instantaneous state of the coastal system over time for which present deterministic forecasting techniques are not capable of providing necessary information on uncertainties and hence could display a false sense of accuracy and skill. At present in long term morphological modeling a full systemic approach for uncertainty assessment has not yet been applied. This paper investigates the use of a Bayesian Network as a tool for uncertainty assessment in decadal scale morphological modeling for the evolution of a mega nourishment at the Dutch North-Holland coast, the Hondsbossche Dunes (HBD). The Bayesian Network is trained with an existing set of model data and field data of one year bed development. The Bayesian Network successfully transfers the bandwidth in input variables, model uncertainty and calibration uncertainty to an uncertainty bandwidth around the output parameter of choice.
Original languageEnglish
Title of host publicationProceedings of Coastal Dynamics 2017
Subtitle of host publicationHelsingør, Denmark
EditorsT. Aagaard, R. Deigaard, D. Fuhrman
Pages1909-1920
Publication statusPublished - 2017
EventCoastal Dynamics 2017 - Kulturværftet, Helsingor, Denmark
Duration: 12 Jun 201716 Jun 2017
http://coastaldynamics2017.dk

Conference

ConferenceCoastal Dynamics 2017
CountryDenmark
CityHelsingor
Period12/06/1716/06/17
Internet address

Keywords

  • Bayesian Network
  • uncertainty
  • morphodynamics
  • numerical modelling
  • Building with Nature

Fingerprint Dive into the research topics of 'Uncertainty assessment in coastal morphology prediction with a bayesian network'. Together they form a unique fingerprint.

Cite this