King's College London

Research portal

Reducing Inconsistencies in Point Observations of Maximum Flood Inundation Level

Research output: Contribution to journalArticlepeer-review

Brandon L. Parkes, Hannah L. Cloke, Florian Pappenberger, Jeff Neal, David Demeritt

Original languageEnglish
Pages (from-to)1-27
Number of pages27
JournalEarth interactions
Issue number6
PublishedAug 2013

King's Authors


Flood simulation models and hazard maps are only as good as the underlying data against which they are calibrated and tested. However, extreme flood events are by definition rare, so the observational data of flood inundation extent are limited in both quality and quantity. The relative importance of these observational uncertainties has increased now that computing power and accurate lidar scans make it possible to run high-resolution 2D models to simulate floods in urban areas. However, the value of these simulations is limited by the uncertainty in the true extent of the flood. This paper addresses that challenge by analyzing a point dataset of maximum water extent from a flood event on the River Eden at Carlisle, United Kingdom, in January 2005. The observation dataset is based on a collection of wrack and water marks from two postevent surveys. A smoothing algorithm for identifying, quantifying, and reducing localized inconsistencies in the dataset is proposed and evaluated showing positive results. The proposed smoothing algorithm can be applied in order to improve flood inundation modeling assessment and the determination of risk zones on the floodplain.

View graph of relations

© 2020 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454