A New Flood Inundation Model

1
A New Flood Inundation Model

• Yang Liu and Garry Pender
• School of Built Environment

Heriot Watt University
2
Contents

• Introduction of rapid flood spreading model
• A new conceptual model for maximum velocity
  prediction and application to an artificial
  digital elevation model.
• Current work

3
1.1 Methods
Speed up the time consuming 2D SWEMs (TUFLOW,
ISIS2D, MIKE21
Parallel Processing
Meta Model
Adaptive Grid
Rapid Flood Spreading Model
4
1.2 Objectives of developing RFSM

• Short time to run (Typically lt 5s)
• A good overall agreement of the final water depth
  and flood extent predictions between SWEM and
  RFSM.
• A good overall agreement of the maximum velocity
  prediction between SWEM and RFSM.
• Very useful to apply RFSM to probabilistic
  flood risk analysis (e.g. Bayesian Analysis) and
  real time forecasting.

5
1.3 Cellular Automata and RFSM
(1) Definition A cellular automata is a
collection of cells on a grid of specified shape
that evolves through a number of discrete time
steps according to a set of rules based on the
states of neighboring cells.
(2) Differences 2.1 Rapid Flood Spreading Model
uses a large irregular cell to save the
computational time. 2.2 Rapid Flood Spreading
Model uses one iteration and simple merging
process compared to CA iterations.
Neighbours rules
References (1) Wolfram, S. (1984) Cellular
automata as models of complexity, Nature. 311
419-24.
(2) Parson, J. Fonstad, M. (2007) A cellular
automata model of surface water flow,
Hydrological Processes, 21.
6
1.4 Basic RFSM algorithm

• Pre-calculation
• An array of flood storage cells is
  constructed from DEM
• Inundation
• A specified volume of flood water is
  distributed across the storage cells.

An example of pre-calculation process
Volume (cubm)
Water level (m)
An example of constant extra head (source Krupka
et al. 2007)
7
1.5 Existing RFSMs
RFSMs
Herriot Watt University Martin Krupka et al. and
ISIS Fast
HR Wallingford Julien Lhomme et al.
(1) Krupka M., Wallis S., Pender S., Neélz S.,
2007, Some practical aspects of flood inundation
modelling, Transport phenomena in hydraulics,
Publications of the Institute of Geophysics,
Polish Academy of Sciences, E-7 (401), 129-135.
(2) Lhomme J., Sayers P., Gouldby B., Samuels P.,
Wills M., Mulet-Marti J., 2008, Recent
development and application of a rapid flood
spreading model, River Flow 2008, September.
(3) Liu Y, Pender G (2010) A new rapid flood
inundation model, the first IAHR European
Congress, Edinburgh, UK.
8
1.6 Two different spreading algorithms
Next active grid Current active grid
(b)
(a)
One-directional RFSM
Multi-directional RFSM
9
1.7 Our improved RFSM

• Rules to provide accurate prediction
• Water will spread from high location to lower
  locations (one directional or multiple
  directional spilling algorithms) and has merging
  process.
• Dynamic Driving head based on inflow hydrograph
• Area with High Manning value n on the floodplain
  using a small driving head

Area 2
discharge
Area 1
t
10
1.8 Model parameters and evaluation functions
(1)
(2)
(3)
11
1.9 Application example
Inflow
Inflow hydrograph
3D plot
2D plot
17 flood cells
12
1.10 Compare RFSMs with ISIS2D
Flood extent using ISIS2D after 10 hours
Flood extent using MD-RFSM
Flood extent using OD-RFSM
13
1.11 ISIS2D simulation
14
1.12 One directional RFSM spilling process
15
1.12 Assumptions

• Time Series water depth can be predicted
  approximately accurate using RFSM
• Flow route needs to be predicted approximately
  accurate.

16
2.1 Maximum Velocity prediction using a new
conceptual model
17
2.2 Performance Comparison of the conceptual
model and ISIS2D
Maximum velocity using ISIS2D
Average Maximum velocity for 17 regions using
ISIS2D
Average Maximum velocity predictions for 17
regions using our proposed model
The conceptual model parameter C was calibrated
using one ISIS2D simulation when peak value
150cubm/s of sine inflow hydrograph.
18
2.3 Performance statistics
19
2.4 Current work about 2005 Carlisle flood event
20
Flood extent predictions Using ISIS2D and RFSM
Fig.2. Flood extent and water depth at 45.25
hours using RFSM. ( 5m grid resolution model will
take 2 seconds to run)
Fig1. Flood extent and water depth after 45.25
hours using ISIS2D. (15m grid resolution model
will take more than 1 hour to run)
21
Performance statistics
22
2.5 Current work

• The proposed method has been applied to
  Thamesmead, London.
• Test more locations.
• (3) Fast Rapid flood spreading Modelling using
  Cellular Automata.
• Targets
• (1) Time series of water depth and velocity
  prediction.
• (2) Run time lt 30 seconds using big irregular
  cells.

Thank you!