Discovery - University of Dundee - Online Publications

Library & Learning Centre

Evaluating uncertain flood inundation predictions with uncertain remotely sensed water stages

Evaluating uncertain flood inundation predictions with uncertain remotely sensed water stages

Research output: Contribution to journalArticle

View graph of relations

Authors

Research units

Info

Original languageEnglish
Pages187-199
Number of pages13
JournalInternational Journal of River Basin Management
Journal publication date2008
Journal number3
Volume6
DOIs
StatePublished

Abstract

On January 2 2003 the Advanced Synthetic Aperture Radar (ASAR) instrument onboard ENVISAT captured a high magnitude flood event on a reach of the Alzette River (G.D. of Luxembourg) at the time of flood peak. This opportunity enables hydraulic analyses with spatially distributed information. This study investigates the utility of uncertain (i.e. non error-free) remotely sensed water stages to evaluate uncertain flood inundation predictions. A procedure to obtain distributed water stage data consists of an overlay operation of satellite radar-extracted flood boundaries with a LiDAR DEM followed by integration of flood detection uncertainties using minimum and maximum water stage values at each modelled river cross section. Applying the concept of the extended GLUE methodology, behavioural models are required to fall within the uncertainty range of remotely sensed water stages. It is shown that in order to constrain model parameter uncertainty and at the same time increase parameter identifiability as much as possible, models need to satisfy the behavioural criterion at all locations. However, a clear difference between the parameter identifiability and the final model uncertainty estimation exists due to ‘secondary’ effects such as channel conveyance. From this, it can be argued that it is necessary not only to evaluate models at a high number of locations using observational error ranges but also to examine where the model would require additional degrees of freedom to generate low model uncertainty at every location. Remote sensing offers this possibility, as it provides highly distributed evaluation data, which are however not error-free, and therefore an approach like the extended GLUE should be adopted in model evaluation.

Documents

Library & Learning Centre

Contact | Accessibility | Policy