Open AccessProbabilistic flood hazard maps for J akarta derived from a stochastic rain‐storm generator
Author(s) -
Nuswantoro R.,
Diermanse F.,
Molkenthin F.
Publication year - 2016
Publication title -
journal of flood risk management
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.049
H-Index - 36
ISSN - 1753-318X
DOI - 10.1111/jfr3.12114
Subject(s) - flood myth , storm , quantile , flooding (psychology) , environmental science , probabilistic logic , event (particle physics) , 100 year flood , meteorology , computer science , hydrology (agriculture) , statistics , mathematics , geography , geology , psychology , physics , geotechnical engineering , archaeology , quantum mechanics , psychotherapist
Generally, the methods to derive design events in a flood‐modelling framework do not take into account the full range of extreme storm events and therefore do not take into account all aleatory uncertainties originating from rainfall intensity and spatial variability. The design event method uses a single simulation in order to represent an extreme event. The study presents a probabilistic method to derive flood inundation maps in an area where rainfall is the predominant cause of flooding. The case study area is the J akarta B asin, I ndonesia. It typically experiences high‐intensity and short‐duration storms with high spatial variability. The flood hazard estimation framework is a combination of a M onte C arlo ( MC )‐based simulation and a simplified stochastic storm generator. Several thousands of generated extreme events are run in the S obek rainfall–runoff and 1 D ‐2 D model. A frequency analysis is then conducted at each location in the flood plain in order to derive flood maps. The result shows that in general, design events overestimate the flood maps in comparison with the proposed MC approach. The MC approach takes into account spatial variability of the rainfall. However, this means that there is a need to have a high number of MC ‐generated events in order to better estimate the extreme quantiles. As a consequence, the MC approach needs much more computational resources and it is time‐consuming if a full hydrodynamic model is used. Hence, a simplified flood model may be required to reduce the simulation time.