iRIC Easy-performable

Long-wave Inundation
MOdel (ELIMO)∗
Users Guide
Ver. 2.0

July 14, 2015

∗
⃝2013
c Coastal and Ocean Engineering Laboratory, School of Engineering,
Hokkaido University
Introduction

The 2011 Tohoku earthquake tsunami brought great damages to human

lives, industries and economics in Japan, and provided an important lesson
for disaster reduction and evacuation. We have reconsidered the disaster
prevention policy and tried to renew multilateral regional disaster plans for
multiple scenarios of tsunami onsets. The estimation of regional tsunami
heights provide fundamental conditions to design the disaster plans; devel-
oping infrastructures, hazard maps and managing evacuation for various
scenarios. While reasonable tsunami computation to predict the tsunami
heights along coasts need to be performed, it requires technical knowledge
and experiences to operate computer simulators as well as expensive high-
performance computers. Whereas local government needs to make the local
disaster and evacuation plans to save the local resident against tsunamis, it
may be difficult for them to predict the heights by themselves because the
absence of human and computer resources, which may slow down procedures
of planning.
iRIC-ELIMO has been developed with a motto of ”simple and easy
tsunami computation” for people with minimum knowledge to operate Win-
dow softwares, which is capable of reliable estimations of tsunami generation,
propagation and evolution on coasts on easy-performable GUI environment
provided by iRIC. Users should carefully confirm the following specifications
of ELIMO and understand the model assumptions and limits of the appli-
cations. Any applications beyond the model conditions and computational
accuracies used in ELIMO are not supported. It should be noted that iRIC-
ELIMO cannot be held responsible for any disadvantages, loss, cost which
may be incurred by ELIMO. Confirm and agree Terms of Conditions before
using ELIMO.

ELIMO Development Team

Yasunori Watanabe, Hokkaido University
August, 2013

i
 ii

ELIMO 2.0 Specification (July 2015)

1. Governing equation; nonlinear long wave equation

2. Coordinate system; Spherical global coordinate

3. Computational method; Euler-Lagrangian finite difference method with

third order accuracy (CIP metod + Predictor-Corrector method)

4. Bathymetry; General Bathymetric Chart of the Oceans (GEBCO) al-

titude data

5. Shore boundary condition; Impermeable condition at shore (no in-

undation on land area) [Run-up computation may be available at
version 3.0]

6. Open boundary condition; Radiation condition + Sponge zone (addi-

tional energy dissipation)

7. Fault model; Elastic rectangular fault model (Okada 1985), Tsunami

generation model including horizontal fault displacements (Tanioka ad
Satake 1996).
Contents

1 What is ELIMO? 1
1.1 Governing equation . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Computational methods and boundary conditions . . . . . . . 4
1.2.1 Impermeable condition at a shore . . . . . . . . . . . . 4
1.2.2 Open boundary condition . . . . . . . . . . . . . . . . 5
1.2.3 Initial Conditions . . . . . . . . . . . . . . . . . . . . . 7

2 Tutorial 12
2.1 Acquisition of bathymetry data . . . . . . . . . . . . . . . . . 12
2.2 Format conversion of the bathymetry data . . . . . . . . . . . 18
2.3 Operations on iRIC . . . . . . . . . . . . . . . . . . . . . . . . 20
2.3.1 Generation of computing grids . . . . . . . . . . . . . 20
2.3.2 Setting computational conditions and Executing ELIMO 22
2.3.3 Visualizing the results . . . . . . . . . . . . . . . . . . 24

3 The 2011 Tohoku earthquake tsunami 30

3.1 Overview of tsunami . . . . . . . . . . . . . . . . . . . . . . . 30
3.2 Comparisons with the observed surface elevation . . . . . . . 31

4 Scenario tsunami 34

iii
CONTENTS iv

Update Records

1. July 2013 Release of ELIMO 1.0

2. November 2013 ELIMO 1.1; Release a format conversion software

”gebco conv” for GEBCO 1 minuet altitude grid data that is down-
loadable from British Oceanographic Data Center. The user’s guide
updated.

3. June 2015 ELIMO 2.0; ”gebco2s2014” for 2014 format of GEBCO

data has been provided. The elastic rectangular fault model by Okada
(1985) and the tsunami generation model by Tanioka and Satake (1996)
have been introduced. The arbitrary fault models with asperities can
be used for the ELIMO computation by using a file input function for
the fault parameters.
Chapter 1

What is ELIMO?

A tsunami computation is a fundamental way to provide tsunami heights and

arrival times along coasts for past and future earthquakes, which has been
used for designing disaster prevention plans. The computations have been
generally performed at high performance computers owned by Universities
and research institutes, and commonly used by researchers and engineers
who have scientific knowledge and experience of the numerical computation.
There may be no computational tool that is accessible to the public. While
the disaster prevention and evacuation plans for tsunami inundation are
required to develop by local governments, they have no way to estimate the
tsunami properties by their own human and computational resources. Under
these background, we have tried to develop technologies for producing a
compact and reliable software to compute tsunami for arbitrary earthquake,
which is executable on home PC. We hope this tsunami software named
ELIMO will be utilized for estimating the tsunami onsets in order to achieve
disaster reduction and efficient evacuation.
iRIC is a computational platform to support computations of river flows
and floods with the pre-processing functions of computational griding and
the post-processing of visualizing and plotting the results on GUI environ-
ment. The tsunami solver ELIMO (Easy-performable Long-wave Inundation
MOdel) is designed to perform fundamental tsunami computations on the
iRIC environment.
This users guide for ELIMO is composed of three chapters. In Chap-
ter 1, governing equations to be solved, boundary conditions, assumptions
and numerical methods used in ELIMO are explained. While researchers
and engineers who have scientific knowledges should confirm these technical
interpretations, general users may skip this chapter. In Chapter 2, the pro-

1
CHAPTER 1. WHAT IS ELIMO? 2

Figure 1.1: Spherical global coordinate system.

cedure to acquire bathymetry data, to compute tsunami and visualize the

results is explained. In Chapter 3, as an example of the tsunami computa-
tion, the 2011 Tohoku earthquake tsunami is simulated to compare with the
observed tsunami heights. The example computation for tsunami generated
by scenario fault model composed of multiple fault segments is introduced
in Chapter 4.
Scenario earthquake tsunami induced by the fault model composed of
multiple segments

1.1 Governing equation

The depth-averaged long wave equation is used as governing equation in
ELIMO (Fig. 1.1).

∂u u ∂u v ∂u 1 ∂pa g ∂ζ τ s − τλb
+ + =− − + λ
∂t R cos ϕ ∂λ R ∂ϕ ρR cos ϕ ∂λ R cos ϕ ∂λ ρ (h + ζ)
( 2
( ))
1 ∂ u 1 ∂ ∂u
+νh + cos ϕ
R2 cos2 ϕ ∂λ2 R2 cos ϕ ∂ϕ ∂ϕ
( )
u
+ 2Ω + v sin ϕ (1.1)
R cos ϕ
CHAPTER 1. WHAT IS ELIMO? 3

∂v u ∂v v ∂v 1 ∂pa g ∂ζ τϕs − τϕb

+ + =− − +
∂t R cos ϕ ∂λ R ∂ϕ ρR cos ϕ ∂ϕ R cos ϕ ∂ϕ ρ (h + ζ)
( 2
( ))
1 ∂ v 1 ∂ ∂v
+νh 2 2 2
+ 2 cos ϕ
R cos ϕ ∂λ R cos ϕ ∂ϕ ∂ϕ
( )
u
+ 2Ω + u sin ϕ (1.2)
R cos ϕ
1
∫ζ ′ 1
∫ζ ′
where the depth-averaged mean velocities u = h+ζ −h u dz, v = h+ζ −h v dz
(local velocities u′ , v ′ ), R is the distance from the center of the earth, Ω is
the earth’s rate of rotation，λ longitude, ϕ latitude, pa atmospheric pres-
sure, ζ surface elevation, νh coefficient of horizontal momentum exchange.
Mass conservation for the depth-averaged equation system is written by
∂ζ 1 ∂ 1 ∂
+ u (h + ζ) + v cos ϕ (h + ζ) = 0 (1.3)
∂t R cos ϕ ∂λ R cos ϕ ∂ϕ
Imposing zero gage pressure at the surface and ignoring surface shear, the
above equations are rewritten with an additional dissipation term Ad in a
sponge zone (explained in the following section) by
Du g ∂ζ τλb
=− − + Ad u
Dt R cos ϕ ∂λ ρ (h + ζ)
( ( ))
∗ 1 ∂2u 1 ∂ ∂u
+νh + cos ϕ
R2 cos2 ϕ ∂λ2 R2 cos ϕ ∂ϕ ∂ϕ
( )
u
+ 2Ω + v sin ϕ (1.4)
R cos ϕ

Dv g ∂ζ τϕb
=− − + Ad v
Dt R cos ϕ ∂ϕ ρ (h + ζ)
( ( ))
1 ∂2v 1 ∂ ∂v
+νh∗ + cos ϕ
R2 cos2 ϕ ∂λ2 R2 cos ϕ ∂ϕ ∂ϕ
( )
u
+ 2Ω + u sin ϕ (1.5)
R cos ϕ
( ) ( )
Dζ 1 ∂uh ∂u 1 ∂vh ∂v tan ϕ
=− +ζ − +ζ + v (h + ζ) (1.6)
Dt R cos ϕ ∂λ ∂λ R ∂ϕ ∂ϕ R
D ∂ u ∂ v ∂
where = + + , modified bottom shear τλb = Cd |u|u
Dt ∂t R cos ϕ ∂λ R ∂ϕ
and τϕb = Cd |v|v. While many values for the drag coefficient Cd have been
proposed for specific flows, Cd = 5 × 10−3 is used in ELIMO.
CHAPTER 1. WHAT IS ELIMO? 4

Figure 1.2: Impermeable condition at a shore

1.2 Computational methods and boundary condi-

tions
The third order accurate, finite difference computations are performed for
the nonlinear equation system Eq. (1.4)-Eq. (1.6) in the same manner as
Watanabe et al. (2012); Constrained Interpolation Profile (CIP) method for
the advection equation in a two-step fractional step framework, predictor-
corrector method for the Eulerian terms. We don’t explain the details of
these numerical methods in this users guide, which can be found in the
specialized literature (Goto, Okayasu and Watanabe, 2013).

1.2.1 Impermeable condition at a shore

The current version of ELIMO simply uses an impermeable condition the
shore; that is, virtual vertical walls are arranged along the shore (Fig. 1.2).
The impermeable condition at a shore is expressed by
∂η
=0 (1.7)
∂n
u·n=0 (1.8)

where n is the normal unit vector (see Fig. 1.2).

We are developing a runup model to simulate tsunami inundation, which
will be equipped to the next version of ELIMO.
CHAPTER 1. WHAT IS ELIMO? 5

Figure 1.3: Computational domain and region to place the sponge zone
(Tokachi area, Hokkaido, Japan).

1.2.2 Open boundary condition

When a finite computational domain is given in ocean area, an open bound-
ary condition imposed at the edges of the domain is important to trans-
mit incident waves across the boundary without any reflection. A flawed
condition may induce unrealistic reflected waves at the boundaries to form
multiple reflection system within the domain, which results in unacceptable
computations. While a very large domain has been commonly used in the
conventional tsunami computation to avoid the difficulties, e.g. a whole area
of Pacific Ocean, ELIMO employs a compact domain to reduce computa-
tional loads. Therefore, appropriate open boundaries need to be given to
achieve an available computation in ELIMO. While any boundary condition
to achieve perfect transmission of incident waves has yet been developed,
Sommerfeld’s radiation condition has been commonly used (Orlanski, 1976);

∂φ ∂φ
+c =0 (1.9)
∂t ∂n
CHAPTER 1. WHAT IS ELIMO? 6

Figure 1.4: Amplification rate for the dissipation term in the sponge zone.

where the arbitrary quantity φ, and the celerity c approximated by the finite
difference at the previous time step;
δφ δφ
c=− / (1.10)
δt δn
δ is the finite difference operator which is constrained by the conditions;

 0 (c < 0)
c= c (0 ≤ c ≤ ∆x/∆t) (1.11)

∆x/∆t (c > ∆x/∆t)
Significant reflection may appear under this radiation condition in the case
that the celerity and incident angle spatially change near the boundary. In
order to attenuate the incident waves near the boundary and to minimize
the wave reflection, a so-called sponge zone is arranged at an inner area
adjacent to the boundary (see Fig. 1.3)．
According to Cruz et al. (1993), the dissipation coefficient Ad is given
by
√ ( )N
Ad = Amax hg (N + 1) i−imax +dlayer
if i ≥ imax − dlayer
dlayer (1.12)
Ad = 1 else
where N is the oder of the distribution function. N = 1 is used in ELIMO.
As shown in Fig. 1.4, Ad linearly increased in the sponge zone enhances
CHAPTER 1. WHAT IS ELIMO? 7

Figure 1.5: Input window for computing conditions.

dissipation for Eq. (1.4) and Eq. (1.5). As additional reflected waves may
be produced within the sponge zone with too high gradients of Ad , suitable
(low) Ad should be provided in the computation. While a wider sponge
zone provides more dissipation, available computing area becomes smaller;
that is, there is tradeoff between computational cost and performance of
the boundary condition. User should carefully choose suitable number of
the computing cells comprising the sponge zone (dlayer) and the maximum
dissipation rate （Amax ） in the window for computational conditions (see
Fig. 1.5). It should be noted that tsunami is wrongly computed in coastal
area containing the sponge zone as the computation is available except the
sponge zone (see circles in Fig. 1.3). In this case, you should change the
locations of the boundaries and use larger domain.

1.2.3 Initial Conditions

Fig. 1.6 shows the distribution of the vertical displacement of sea floor due
to the 2011 Tohoku earthquake tsunami, which provided by Japan Mete-
orological Agency. The vertical displacement of faults directly contribute
to sea level rise while the horizontal displacement also results in sea level
variations over uneven sea floors.
In general, deformation of the elastic rectangular fault is determined by
the fault parameters; Rupture length (L), Rupture width (W ), Slip length
(U ), Rake angle (θ), Strike angle (ϕ), Dip angle (δ), Fault depth (d), Latitude
and Longitude of Fault Top（see Fig. 1.7 and Fig. 1.8）. ELIMO uses the
elastic fault model for the rectangular dip-slip deformation source, which
is derived by Okada (1985). This model provides the three-dimensional
CHAPTER 1. WHAT IS ELIMO? 8

Figure 1.6: Displacement of sea floor due to the 2011 Tohoku earthquake
tsunami; vertical displacement (left) and horizontal one (right) (web page
of Japan Meteorological Agency).

displacements at arbitrary coordinate (x, y).

U2 ( q )
ux = − − I3 sin δ cos δ (1.13)
2π (R )
U2 ỹq ξη
uy = − + cos δ arctan − I1 sin δ cos δ (1.14)
2π R (R + ξ) qR
( )
U2 ˜
dq ξη
uz = − + sin δ arctan − I5 sin δ cos δ (1.15)
2π R (R + ξ) qR
(1.16)
CHAPTER 1. WHAT IS ELIMO? 9

Figure 1.7: Fault parameters; slip length, slip angle, strike and dip angle.

where
µ ξ
I1 = − ( ) − I5 tan δ (1.17)
λ + µ R + d˜cos δ
µ
I2 = − log (R + η) − I3 (1.18)
λ+µ
 
µ  ỹ
I3 = ( ) − log (R + η) − I4 tan δ (1.19)
λ + µ cos δ R + d˜

µ 1 ( ( ) )
I4 = log R + d˜ − sin δ log (R + η) (1.20)
λ + µ cos δ
µ 2 η (X + q cos δ) + X (R + X) sin δ
I5 = arctan (1.21)
λ + µ cos δ ξ (R + X) cos δ
(1.22)

and p = y cos δ + d sin δ, q = y sin δ − d cos δ, ỹ = η cos δ + q sin δ, d˜ =

η sin δ − q cos δ, R2 = ξ 2 + η 2 + q 2 = ξ 2 + ỹ 2 + d˜2 , X 2 = ξ 2 + q 2 , U2 is the
dip-slip displacement source. These displacements are evaluated by using
the following relation;

f (ξ, η) = f (x, p) − f (x, p − W ) − f (x − L, p) + f (x − L, p − W ). (1.23)

Fig. 1.9 shows the vertical and horizontal displacements of the sea floor
for the fault parameters for the 2003 Tokachi off earthquake, estimated by
CHAPTER 1. WHAT IS ELIMO? 10

Figure 1.8: Fault parameters; longitude and latitude of the fault top, and
strike (in the case of the 2003 Tokachi-off earthquake)

the Okada’s model. Any sea floor displacements for arbitrary fault param-
eters can be determined by this model.
Tanioka and Satake (1995) proposed the tsunami generation model to
use as an initial sea level condition (ζ), based on bulk water displacement
due to horizontal sea floor displacement as well as direct water displacement
by vertical sea floor displacement;
∂h ∂h
ζ = uz + ux + uy (1.24)
∂x ∂y
ELIMO determines the initial sea level by the Tanioka-Satake model
using the Okada’s solutions for the input fault parameters. Users can choose
two options to input the parameters, ‘manual input’ and ‘file input’, on
ELIMO, and in particular, the latter option is useful for setting up spatially-
non-uniform displacements described by local deformation of multiple fault
segments, which will be explained in §2.2.2.
CHAPTER 1. WHAT IS ELIMO? 11

Figure 1.9: Vertical (left) and horizontal (right) displacements for the 2003
Tokachi off earthquake.
Chapter 2

Tutorial

This chapter describes a whole procedure to simulate and analyze tsunami;

acquisition of bathemetry data, format conversion of the bathemetry file,
generation of computing grids, setup of initial conditions (fault rapture pa-
rameters, computing conditions), execution of ELIMO, visualizing and an-
alyzing computed results.

2.1 Acquisition of bathymetry data

The bathymetry in the computing domain is determined by loading a text
file (tpo file) containing latitude, longitude and water depth (altitude) data.
While User can use any bathymetry dataset provided by any sources by
converting file format to the tpo file, ELIMO only supports use of General
Bathymetric Chart of the Oceans (GEBCO) that is downloadable from the
web page of British Oceanographic Data Center (BODC). This section ex-
plains how to get and convert the GEBCO bathymetry dataset. It should
noted that the file format of GEBCO data has been updated since 2014 and
thus the previous format convert software, gebco conv.exe, attached with
ELIMO ver. 1.X is not work for 2014 version of GEBCO.

1. Go to the web site of GEBCO on British Oceanographic Data Center

(BODC), https://www.bodc.ac.uk/data/online_delivery/gebco/.
Click ‘register’ (red circle 1 indicated in Fig. 2.1) to move to th page of
”User Registration”. If you have already completed the registration,
move to Procedure 4).

2. Fill in all user information on the User Registration (Fig. 2.2). Click
a button ”Next>” (red circle indicated in Fig. 2.2).

12
CHAPTER 2. TUTORIAL 13

Figure 2.1: Web site of GEBCO on British Oceanographic Data Center

3. You will received an email describing the confirmation code (green cir-
cle in Fig. 2.3) and URL to move next (red circle in Fig. 2.3). Fill the
confirmation code in the text box on the page ”Confirmation registra-
tion” (red circle in Fig. 2.4) and press ”Complete registration”(green
circle in Fig. 2.4).

4. Select a region to download on the world map at https://www.bodc.

ac.uk/data/online_delivery/gebco/(Fig. 2.1). Confirm selected
area (indicated by yellow) is appropriate as a computing domain (see
Fig. 2.5). As long time computation is required for a large domain,
we do not recommend to select very large area.

5. The data type is chosen at a box bellow the map (arrow in Fig. 2.5).
As the highest resolution data (30 second) has been confirmed
to contain errors, erroneous results may be computed when
the 30 second data is used without any correction. We rec-
ommend not to use the 30 second data without appropriate
corrections. ELIMO supports only one minuets data (arrow of Fig.
2.5).

6. Click the check box for ‘User defined area or global grids’ and ‘2D
netCDF’ (arrow 1 in Fig. 2.6). Click ‘Add data to basket’ to register
the selected data (arrow 2). Click ‘View basket’ (arrow 3) to confirm
CHAPTER 2. TUTORIAL 14

Figure 2.2: User registration

CHAPTER 2. TUTORIAL 15

Figure 2.3: Email automatically sent by British Oceanographic Data Center

Figure 2.4: Confirm registration

CHAPTER 2. TUTORIAL 16

Figure 2.5: Selection of the region to download

Figure 2.6: Registration of data

CHAPTER 2. TUTORIAL 17

Figure 2.7: Check out

Figure 2.8: Login

CHAPTER 2. TUTORIAL 18

Figure 2.9: Download of the data

the registered data (Fig. 2.7).

7. Click ‘Checkout your basket’ bellow Basket(Fig. 2.7).
8. Fill in your email address and password on the page of ”Log in” (Fig.
2.8). Press the bottom ”Login to proceed” bellow the text box.
9. Click the button ‘Download’ for the registered data on ‘Request status’
(Fig. 2.9) and save it.

2.2 Format conversion of the bathymetry data

The procedure to convert the downloaded GEBCO data file into iRIC com-
patible file format is explained in this section. ELIMO does not support the
previous version of GEBCO (provided earlier than 2014) .
1. The file convert software named ‘gevco2d2014.exe’ is generated at
the directory where iRIC installed; C:\ProgramFiles(x86)\iRIC2.
X(x86)\solvers\elimo_v2 (or C:\ProgramFiles\iRIC2.X\solvers\
elimo_v2. These paths may be changed depending on the versions of
iRIC.) Execute this software. Note that this software requires the dy-
namic link library ‘netcdf.dll’ to be in the same folder. Confirm the
presence of the file.
CHAPTER 2. TUTORIAL 19

Figure 2.10: Screenshot of gebco conv

2. Choose the downloaded file of GEBCO data in a file dialog (see Fig.
2.10), then the conversion will automatically start.

3. Confirm a new file of the same file name with the extension ‘.tpo’ is
generated. For instance, the original file whose name is ‘gebco 1min 138 35 146 4.nc’
is converted into the tpo file ‘gebco 1min 138 35 146 4.tpo’.
CHAPTER 2. TUTORIAL 20

Figure 2.11: Importing the tpo file (converted bathymetry data).

2.3 Operations on iRIC

The procedures to compute tsunami and visualize the results are explained
below.

2.3.1 Generation of computing grids

1. Run iRIC.
2. Choose ‘File’ → ‘New Project’. Choose ’ELIMO’ in the window ‘Select
Solver’.
3. Confirm ‘Pre-processing window’ is generated. Choose ‘Import’→‘Geographic
Data’→‘Elevation’.
4. Set a file filter located at the bottom right in the file dialog to be
‘Topography File (*.tpo *.anc)’ (see Fig. 2.11).
5. Choose the format converted tpo file, e.g. ‘ftp depth500 file 03 001.tpo’.
No need to filter the dataset(Click ‘OK’ button).
6. Confirm the imported bathymetry is identical with the area you want
to compute (Fig. 2.12). If you want to use multiple data areas for a
wide area computation, go back to 3 to import all bathymetry data.
7. ‘Grid’→‘Select Algorithm to Create Grid’. Choose ‘Create grid by di-
viding rectangular region (Longitude-Latitude)’ in the window ‘Select
Grid Creating Algorithm’.
CHAPTER 2. TUTORIAL 21

Figure 2.12: Example of imported bathymetry.

8. Define the computational domain by a mouse within the area of the

imported bathymetry (see Fig. 2.13). Note that the computational
domain has to be smaller than the imported bathymetry area, and
should be inside any bathymetry data boundaries.

9. Choose ‘Grid’→‘Create Grid’. Define a grid width ‘d’ to be the reso-

lution that you want to compute at (see Fig. 2.14). Since the global
coordinate system is used in ELIMO, x-axis represents longitude and
y-axis is latitude (i. e. d=0.1 degrees may indicate the grid width over
6 km). It should be noted that, even if you define very fine grid, a
substantial resolution for the computation may not be improved rather
than the original resolution of the bathymetry. As computational costs
increase with the resolution, User should choose the optimal grid width
in terms of resolution and cost. The grid width d to be in a range of
0.01 and 0.02 degrees (36 – 72 second) is recommended. Since the
GEBCO one minute data has the resolution of 1 minute = 1/60 de-
gree =0.0167 degree, if you want to compute at identical resolution to
the bathymetry data, the grid width should be less than 0.0167. We
have confirmed the grid system with d = 0.015 degree resolves rea-
sonably bathymetry and coastline (the grid system with d = 0.02 has
been also confirmed to provide unexpected strange coastline of Ojika
Cape, Japan as well as unreasonable very high tsunami height there).
Press ‘OK’.
CHAPTER 2. TUTORIAL 22

Figure 2.13: Generation of the computational domain.

2.3.2 Setting computational conditions and Executing ELIMO

1. ‘Calculation Condition’→‘Setting’.

2. There are two way to input the bathymetry data.

Manual input of the fault parameters

This method is useful to manually input the parameters for one or two
rectangular faults on the input window of ELIMO.
Confirm ‘Manual Input’ is selected for ‘Which manner to input fault
conditions?’ on the window of ‘File Input of Fault’ (Fig. 2.15). The
fault parameters should be provided on two input windows for comput-
ing conditions; First Fault Rupture Conditions shown in Fig. 2.16, and
Second Fault Rupture Conditions shown in Fig. 2.17 (see also §1.2.3).
Default parameters for the 2003 Tokachi-off earthquake tsunami have
been provided ( see also the parameters for the 2011 Tohoku earth-
quake tsunami presented in Chapter 3). For any fault model with a
single rapture, all parameters for ‘Second Fault Rapture Conditions’
should be given to be zero. It should be noted that ‘Fault depth’
is defined by depth of the fault bottom d in ELMO (see Fig. 1.7)
while depth at the fault top dtop may be used in some fault models for
scenario tsunami. d can be estimated with dtop by

d = dtop + W tan δ sin θ (2.1)

CHAPTER 2. TUTORIAL 23

Figure 2.14: Input a grid width ‘d’.

where δ is the dip angle，W is the rapture width and θ is the rake
angle.

File input of the fault parameters

Fault models for past and scenario tsunami, commonly provided, are
composed by many fault segments to describe asperity and local ori-
entation of the raptures (Fig. 2.18). In ELIMO, arbitrary fault pa-
rameters for multiple segments can be given by a batch file input. For
instance, if the parameters for 12 segments described in Fig. 2.18 are
provided by Table 2.1 (These parameters will be explained again in
§4), as shown in Fig. 2.19, the parameters for each segment are de-
scribed on each line in order of latitude, longitude, rapture length,
rapture width, slip-length, strike, dip, rake and fault depth with space
or comma separation. It should be noted that ‘Fault depth’ is defined
by depth of the fault bottom d in ELMO (see Fig. 1.7). If depth at
the fault top dtop is defined in past fault models, d estimated by Eq.
(2.1) needs to be provided in the file. Add lines for the parameters
of all segments; in the case of Fig. 2.18, every parameters of 12 lines
should be written and saved with arbitrary file name.
Choose ‘File Input’ for ‘Which manner to input fault conditions?’ on
the window of ‘File Input of Fault’ (Fig. 2.15). Click the button
adjacent to the file name box, and select the created text file on the
file dialog.
CHAPTER 2. TUTORIAL 24

Figure 2.15: File Input of Fault.

3. In ‘Computing Conditions’, input the maximum time step number,

time-step interval for saving output files (e.g. if 3 is given, the data
files are saved every 3 time steps) , number of grids contained in a
sponge zone (dlayer in §1.2.2), and the maximum dissipation rate in
a sponge zone (Amax in §1.2.2). dlayer and Amax may be adjusted if
unacceptable reflection occurs at the open boundaries (depending on
size and location of the computing domain).

2.3.3 Visualizing the results

How to visualize the results is explained in iRIC Users Manual; such as con-
tours, vector plots, animation of the results. In this chapter, only particular
operations required in ELIMO are explained.

1. In the Graphic User interface (GUI) environment of iRIC, the com-

puted results of ELIO are visualized on Post Processing Window in
‘Calculation Result’. The display contents ‘SElevation’ and ‘Velocity’
indicate the surface elevation of tsunami and depth-integrated mean
velocity, respectively （see Fig. 2.20 and Fig. 2.21）. In order to plot
the velocity vectors, ‘Draw’ → ‘Arrow’ to set up vector length and
sampling rate to plot.

2. Since Bird’s-Eye Post-Processing Window is not designated for the

CHAPTER 2. TUTORIAL 25

Figure 2.16: Input window of the fault parameters.

global spherical coordinate used in ELIMO, aspect ratio of the coor-

dinate needs to be adjusted; ‘Draw’ → ‘Z-direction Scale’ to change
scale to sufficiently small value less than 0.001 or magnify the image
(see Fig. 2.22).
CHAPTER 2. TUTORIAL 26

Figure 2.17: Input window of the fault parameters for the second fault.

Figure 2.18: Fault segments of the scenario tsunami of Hokkaido off earth-

quake.
CHAPTER 2. TUTORIAL 27

latitude longitude length width slip length strike dip rake depth (fault top)
(o ) (o ) (km) (km) (m) (o ) (o ) (o ) ( m)
Fault 1 147.5584 42.4135 70.0 70.0 35.0 235 10 90 5.0
Fault 2 146.8476 42.0703 70.0 70.0 35.0 235 10 90 5.0
Fault 3 146.1445 41.7226 70.0 70.0 35.0 235 10 90 5.0
Fault 4 145.4488 41.3704 70.0 70.0 35.0 235 10 90 5.0
Fault 5 144.7606 41.0140 70.0 70.0 35.0 225 10 90 5.0
Fault 6 144.1726 40.5693 70.0 70.0 35.0 225 10 90 5.0
Fault 7 147.0946 42.9421 70.0 70.0 30.0 235 20 90 17.0
Fault 8 146.3802 42.5962 70.0 70.0 30.0 235 20 90 17.0
Fault 9 145.6735 42.2457 70.0 70.0 30.0 235 20 90 17.0
Fault 10 144.9746 41.8907 70.0 70.0 30.0 235 20 90 17.0
Fault 11 144.1716 41.4608 70.0 70.0 30.0 225 20 90 17.0
Fault 12 143.5836 41.0132 70.0 70.0 30.0 225 20 90 17.0

Table 2.1: Fault parameters for the segments of the scenario tsunami of
Hokkaido off earthquake (corresponding to Fig. 2.18).

Figure 2.19: Parameter text file for the file input.

CHAPTER 2. TUTORIAL 28

Figure 2.20: Visualizing tsunami surface elevation ‘SElevation’

Figure 2.21: Visualizing depth-averaged velocity ‘Velocity’.

CHAPTER 2. TUTORIAL 29

Figure 2.22: Bird view of the tsunami surface elevation ‘SElevation’.

Chapter 3

The 2011 Tohoku earthquake

tsunami

The computed results for the 2011 Tohoku earthquake tsunami are presented
in this chapter.

3.1 Overview of tsunami

The Geospatial Information Authority of Japan provided the rectangle fault
model for the 2011 Tohoku earthquake with following parameters. In this
model, the depth of fault is defined at the top of fault. The fault depths,
defined at the bottom in ELIMO, are estimated by Eq. (2.1) to be 37.4
km and 31.8 km. Therefore, the estimated depths, 37.4 and 31.8, need to
be described on ‘First and Second Rapture Condistions’ instead of 10.0 and
10.1 in Table 3.1. As many fault models using the depth of fault top have
been proposed, Use should confirm the definition of the parameter used in
the model.
Fig. 3.2 shows the contours of the surface elevation of the tsunami until
one hour after the earthquake. It is seen that the tsunami radically spreads
at an early stage, arrives to Iwate prefecture at 10 – 15 minutes, Hokkaido
and Kita-Kanto areas after one hour. These computed arrival times of the
tsunami are consistent with the observed ones.

30
CHAPTER 3. THE 2011 TOHOKU EARTHQUAKE TSUNAMI 31

Figure 3.1: The rectangle fault model for the 2011 Tohoku earthquake (ini-
tial condition).

latitude longitude length width slip length strike dip rake Fault depth
(o ) (o ) (km) (km) (m) (o ) (o ) (o ) ( m)
Fault 1 39.00 143.49 199 85 27.7 202 18 97 10.0
Fault 2 37.21 142.51 176 82 5.9 201 15 81 10.1

3.2 Comparisons with the observed surface eleva-

tion
Kawai et al. (2013) reported the surface elevation of the 2011 Tohoku
tsunami measured by GPS buoys installed Tohoku off. Fig. 3.3 shows the
observed surface elevations presenting rapid evolution of the tsunami wave
shapes.
Fig. 3.4 shows the initial stage of the computed surface elevations at
the corresponding locations to the GPS stations. We find overall features of
the computed tsunami evolution are consistent with the observed ones de-
spite that the simplest rectangular fault model are used in the computation.
Improvement of accuracy is expected by introducing polished fault models
with asperity.
CHAPTER 3. THE 2011 TOHOKU EARTHQUAKE TSUNAMI 32

Figure 3.2: Propagation of the 2011 Tohoku earthquake tsunami (interval

of 20 minutes).
CHAPTER 3. THE 2011 TOHOKU EARTHQUAKE TSUNAMI 33

Figure 3.3: Observed surface elevations at the GPS tide stations (After
Kawai et al. 2013)．

Figure 3.4: Computed surface elevations at the GPS tide stations.

Chapter 4

Scenario tsunami

Figure 4.1: Fault segments of the scenario tsunami of Hokkaido off earth-

quake.

Many fault models for scenario tsunamis, expected to occur in future,

have been proposed by the governments and research institutes, which pro-
vides possible tsunami heights along coasts of Japan. Most models composed
of multiple fault segments can be used on ELIMO by the function of the
batch file input.
The government of Hokkaido, Japan, defined the scenario tsunami that

34
CHAPTER 4. SCENARIO TSUNAMI 35

latitude longitude length width slip length strike dip rake depth (fault top)
(o ) (o ) (km) (km) (m) (o ) (o ) (o ) ( m)
Fault 1 147.5584 42.4135 70.0 70.0 35.0 235 10 90 5.0
Fault 2 146.8476 42.0703 70.0 70.0 35.0 235 10 90 5.0
Fault 3 146.1445 41.7226 70.0 70.0 35.0 235 10 90 5.0
Fault 4 145.4488 41.3704 70.0 70.0 35.0 235 10 90 5.0
Fault 5 144.7606 41.0140 70.0 70.0 35.0 225 10 90 5.0
Fault 6 144.1726 40.5693 70.0 70.0 35.0 225 10 90 5.0
Fault 7 147.0946 42.9421 70.0 70.0 30.0 235 20 90 17.0
Fault 8 146.3802 42.5962 70.0 70.0 30.0 235 20 90 17.0
Fault 9 145.6735 42.2457 70.0 70.0 30.0 235 20 90 17.0
Fault 10 144.9746 41.8907 70.0 70.0 30.0 235 20 90 17.0
Fault 11 144.1716 41.4608 70.0 70.0 30.0 225 20 90 17.0
Fault 12 143.5836 41.0132 70.0 70.0 30.0 225 20 90 17.0

Table 4.1: Fault parameters for the segments of the scenario tsunami of
Hokkaido off earthquake (corresponding to Fig. 2.18).

may cause the maximum inundation on Pacific coasts of Hokkaido, and

provided the fault model. In this model for the extreme tsunami, 12 fault
segments with the parameters in Table 4.1 are assumed to adjacently locate
as shown in Fig. 4.1.
The text parameter file for Table 4.1 are shown in Fig. 4.2. The line
number corresponds to the fault number. The parameters for each seg-
ment,latitude, longitude, rapture length, rapture width, slip-length, strike,
dip, rake and fault depth with space or comma separation, are described
on each line. As already explained, ‘Fault depth’ is defined by depth of the
fault bottom d in ELMO (see Fig. 1.7). If depth at the fault top dtop is
defined in past fault models, d estimated by Eq. (4.1) needs to be provided
in the file.
d = dtop + W tan δ sin θ (4.1)
where δ is the dip angle，W is the rapture width and θ is the rake angle.
For example, for Fault 1 – 6, 5+70×tan(10×π/180)×sin(90×π/180) ≈
17.3, and for 7 -12, 17 + 70 × tan(20 × π/180) × sin(90 × π/180) ≈ 42.5
should be written in the file as ‘Fault depth’ instead of 5.0 and 17.0. Every
parameters of 12 lines are written and saved with arbitrary file name for the
batch input.
Fig. 4.3 shows the evolution of the Hokkaido scenario tsunami generated
by the model of Table 4.1 . The tsunami, generated by the fault displace-
ments over a wide area, radically propagates and reach to the Pacific coasts
after 10 – 20 minutes. The first wave of tsunami arrives at most coasts of
CHAPTER 4. SCENARIO TSUNAMI 36

Figure 4.2: Parameter text file for the file input.

Hokkaido and Tohoku after 30 minutes, and formations of edge waves are
confirmed after 50 minutes.
Fig. 4.4 presents the computed surface elevations (not by ELIMO) re-
ported by the government of Hokkaido. We find the analogous features of
tsunami elevations to Fig. 4.3.
I this way, ELIMO can simulate arbitrary tsunami generated by any
model without any limitation of number of segments.
CHAPTER 4. SCENARIO TSUNAMI 37

Figure 4.3: The surface elevations along Hokkaido coasts by Hokkaido sce-
nario tsunami, computed by ELIMO (10 minutes intervals).

Figure 4.4: The surface elevations along Hokkaido coasts by Hokkaido sce-
nario tsunami, reported by the local government of Hokkaido.
Conculding Remarks

Large earthquakes exceeding M8 are expected in near future, such as Tokai,

Nankai trough and Nemuro-off earthquakes. Multiple approaches for reduc-
ing tsunami disaster and minimizing damage expected for such earthquake
should be considered in a rush. We hope ELIMO that enables general users
to perform the tsunami computation supports developments of these ap-
proaches and improve education for disaster prevention.

August 8, 2013
ELIMO development team
Yasunori Watanabe (Hokkaido University)

Acknowledgment

Development of ELIMO is partially supported by research grant, River

Center of Hokkaido. We thank Prof. Yasuyuki Shimizu, Dr. Toshiki
Iwasaki, Dr. Tomoko Kyuka and Mr. Kazutake Asahi for technical sug-
gestions.

38
Bibliography

[1] Orlanski, I. (1976): A simple boundary condition for unbounded hy-

perbolic flows, J. Comp. Phys., Vol.21, pp.251-269.

[2] Goto H., Okayasu A. and Watanabe Y. (2013), Computational Wave

Dynamics, World Scientific, ISBN: 978-981-4449-70-0.

[3] Cruz E., Yokoki H., Isobe M. and Watanabe A. (1993)，Annual journal
of coastal engineering, JSCE, 40, pp. 46 –50.

[4] Watanabe Y., .Mitobe Y., Saruwatari A., Yamada T., Niida Y., (2012)
Evolution of the 2011 Tohoku earthquake tsunami on the Pacific coast
of Hokkaido, Coastal Engineering Journal, 54 (1), 1250002

[5] Okada Y. (1985): Surface deformation due to shear and tensile faults
in a half-space, Bull. Seism. Soc. Am., Vol. 75, pp.1135 - 1154.

[6] Tanioka Y. and Satake K. (1996): Tsunami generation by horizontal

displacement of ocean bottom, Geophysical Research Letters, Vol. 23,
No.8, pp.861-864.

[7] Hiroyasu Kawai, Makoto Stoh, Koji Kawakuchi and Katsumi Seki,
Characteristics of the 2011 Tohoku tsunami waveform aquired aroud
Japan by NOWPHASE equipments, Coastal Engineering Journal,
55(3), 2013.

39