Andrey Zaytsev 1,2, Andrey Kurkin 1, Efim Pelinovsky1-4, and Ahmet C. Yalciner 5
1 Nizhny Novgorod State Technical University n.a. R.E. Alekseev, Nizhny Novgorod, RUSSIA 2 Special Research Bureau for Automation of Marine Researches, Yuzhno-Sakhalinsk, RUSSIA 3 Institute of Applied Physics, Nizhny Novgorod, RUSSIA 4 National Research University – Higher School of Economics, Moscow, RUSSIA 5 Middle East Technical University, Department of Civil Engineering, Ankara, TURKEY
BETA TESTING FOR INCREASED ACCURACY AND IMPROVED PERFORMANCE OF THE INDONESIAN TSUNAMI EARLY WARNING APPLICATION (Ina-TEWA) 169
Madlazim1,2, Tjipto Prastowo1,2
1Physics Department, The State University of Surabaya, Surabaya 60231, INDONESIA 2Center for Earth Science Studies, The State University of Surabaya, Surabaya 60231, INDONESIA Correspondence: [email protected]
EFFECTIVENESS OF GAME MODEL ON TSUNAMI DISASTER ANTICIPATION IN TWO PROVINCES OF INDONESIA, YEAR 2019 179
Widiastuti1,a, Siswo Poerwanto,2,b, Hernawan3,c , B.Firdiansyah4,d, Sugiharto5,e 1, 3, 4 Sports Education, Postgraduate Jakarta State University, INDONESIA
TSUNAMI SOCIETY INTERNATIONAL, 1741 Ala Moana Blvd. #70, Honolulu, HI 96815, USA. SCIENCE OF TSUNAMI HAZARDS is a CERTIFIED OPEN ACCESS Journal included in the prestigious international academic journal database DOAJ, maintained by the University of Lund in Sweden with the support of the European Union. SCIENCE OF TSUNAMI HAZARDS is also preserved, archived and disseminated by the National Library, The Hague, NETHERLANDS, the Library of Congress, Washington D.C., USA, the Electronic Library of Los Alamos, National Laboratory, New Mexico, USA, the EBSCO Publishing databases and ELSEVIER Publishing in Amsterdam. The vast dissemination gives the journal additional global exposure and readership in 90% of the academic institutions worldwide, including nation-wide access to databases in more than 70 countries.
OBJECTIVE: Tsunami Society International publishes this interdisciplinary journal to increase and disseminate knowledge about tsunamis and their hazards.
DISCLAIMER: Although the articles in SCIENCE OF TSUNAMI HAZARDS have been technically reviewed by peers, Tsunami Society International is not responsible for the veracity of any statement, opinion or consequences.
EDITORIAL STAFF Dr. George Pararas-Carayannis, Editor mailto:[email protected]
EDITORIAL BOARD Dr. Hermann FRITZ, Georgia Institute of Technology, USA Prof. George CURTIS, University of Hawaii -Hilo, USA Dr. Zygmunt KOWALIK, University of Alaska, USA Dr. Galen GISLER, NORWAY Prof. Kam Tim CHAU, Hong Kong Polytechnic University, HONG KONG Dr. Jochen BUNDSCHUH, (ICE) COSTA RICA, Royal Institute of Technology, SWEDEN Acad. Dr. Yurii SHOKIN, Novosibirsk, RUSSIAN FEDERATION Dr. Radianta Triatmadja - Tsunami Research Group, Universitas Gadjah Mada, Yogyakarta, INDONESIA
TSUNAMI SOCIETY INTERNATIONAL, OFFICERS Dr. George Pararas-Carayannis, President Dr. Carolyn Forbes, Secretary
Submit manuscripts of research papers, notes or letters to the Editor. If a research paper is accepted for publication the author(s) must submit a scan-ready manuscript, a Doc, TeX or a PDF file in the journal format. Issues of the journal are published electronically in PDF format. There is a minimal publication fee for authors who are members of Tsunami Society International for three years and slightly higher for non-members. Tsunami Society International members are notified by e-mail when a new issue is available. Permission to use figures, tables and brief excerpts from this journal in scientific and educational works is granted provided that the source is acknowledged.
Recent and all past journal issues are available at: http://www.TsunamiSociety.org CD-ROMs of past volumes may be purchased by contacting Tsunami Society International at [email protected] Issues of the journal from 1982 thru 2005 are also available in PDF format at the U.S. Los Alamos National Laboratory Library http://epubs.lanl.gov/tsunami/
NUMERICAL TSUNAMI MODEL NAMI-DANCE Andrey Zaytsev 1,2, Andrey Kurkin 1, Efim Pelinovsky1-4, and Ahmet C. Yalciner 5
1 Nizhny Novgorod State Technical University n.a. R.E. Alekseev, Nizhny Novgorod, Russia 2 Special Research Bureau for Automation of Marine Researches, Yuzhno-Sakhalinsk, Russia
3 Institute of Applied Physics, Nizhny Novgorod, Russia 4 National Research University – Higher School of Economics, Moscow, Russia
5 Middle East Technical University, Department of Civil Engineering, Ankara, Turkey
ABSTRACT
The numerical tsunami model NAMI-DANCE solving the equations of the nonlinear-dispersive theory of long waves is described. It is developed in cooperation of Turkish and Russian specialists and has been used to simulate tsunami characteristics since 2003. The basic model solver simulates the wave propagation with the well-known nonlinear shallow-water equations. The numerical scheme is based on the Leap-Frog method. The long wave dispersion related to the finiteness of water depth is modeled with the help of numerical dispersion by using specific conditions for spatial and temporal steps. The equations are solved in spherical (geographical) coordinates on rotated Earth taking into account nonlinear quadratic friction in the near-bottom layer. The NAMI-DANCE model is adapted to simulate the generation and propagation of tsunamis of various origin from: underwater earthquakes, submarine landslides, and meteo-tsunamis. For the seismic origin tsunami waves, the initial conditions for hydrodynamic equations are found from the Okada solution; the model allows simulating several sources from different fault segments in the earthquake zone. In the case of the meteo-tsunami, the spatial and temporal distribution of the atmospheric pressure is used as external force. The process of generating the landslide origin tsunami waves, is analyzed in the framework of the two-layer model with a lower viscous layer modeling the submarine landslide motion and an upper layer is the water body. The boundary conditions on the open boundaries correspond to the free-wave passage. The run-up process is also computed using the bathymetry and topography at the coastal boundaries. The NAMI-DANCE code has been verified with several benchmarks according to the NTHMP benchmarks. The difficulties of tsunami modelling connected with the lack of accuracy in the bottom bathymetry and the land topography are mentioned. The use of the developed code for the tsunami action analysis on the coasts and constructions is demonstrated.
Key words: tsunami, underwater earthquake, submarine landslides, meteo-tsunami, shallow-water equations, long wave theory, tsunami action on coasts and constructions
Vol. 38, No. 4, page 151 (2019)
1. INTRODUCTION
The historical information analysis on the world tsunami shows that, on average, a tsunami happens about once a month, and a catastrophic event approximately happens once a year. Only in 2018, two devastating tsunamis occurred in Indonesia with human casualties. The first of them happened on the Sulawesi Island after a strong earthquake on September 28, when the wave height reached 11 m and more than 2000 people died (Omira et al, 2019). The second was generated by the volcano Anak Krakatau eruption on 22 December 2018, which led to the wave up to 85 m high, which killed about a five hundred people; it was predicted in (Giachetti et al, 2012). Especially, we note the landslide origin tsunami that happened in Russia in December 2018 on the Bureya River with the wave up to 90 m high, fortunately it occurred in uninhabited places (Makhinov et al, 2019). Finally, the meteo-tsunami with the height of 1.5 m is worth mentioning, which occurred in Spain on July 16 (Science Global News, 2018). Even this simple event list for 2018 shows the whole variety of tsunami generation mechanisms.
There is a huge amount of literature describing tsunami modelling with the use of various numerical models. Here we describe our numerical model NAMI-DANCE that solves the equations of nonlinear-dispersive theory of long waves. We have been working with this model since 2003, and have applied it for modeling many historical and prognostic tsunami events.
2. NONLINEAR SHALLOW-WATER EQUATIONS SOLVED BY NAMI-DANCE
The catastrophic tsunami is often global in nature and, originating in one place, can remain very destructive at a distance of tens of thousands of kilometers from the source. Therefore, there is a need to solve the shallow water equations on the real Earth, taking into account its rotation, variable bottom bathymetry and the rugged coastline, as well as the friction influence in the turbulent bottom layer. Basic equations are those well-known nonlinear shallow water equations written on spherical coordinates with respect to flow discharges
! (1)
! (2)
! , (3)
where η is the water surface displacement, t is the time, M and N are the components of the water flow (discharge fluxes) along the longitude λ and the latitude θ on the rotating Earth, D = h(x, y) + η is the
Vol. 38, No. 4, page 152 (2019)
2
22 2
7/3
1 1 coscos cos
η ,
cos
M M MNt R D R D
gD gn M M N fNR D
⎛ ⎞∂ ∂ ∂ θ⎛ ⎞+ + +⎜ ⎟ ⎜ ⎟∂ θ ∂λ θ ∂θ ⎝ ⎠⎝ ⎠
∂+ + + =
θ ∂λ
2
22 2
7/3
1 1 coscos cos
η ,
N MN Nt R D R D
gD gn N M N fMR D
⎛ ⎞∂ ∂ ∂ θ⎛ ⎞+ + +⎜ ⎟⎜ ⎟∂ θ ∂λ θ ∂θ⎝ ⎠ ⎝ ⎠
∂+ + + = −
∂θ
( )1cos 0
cosM N
t R∂η ∂ ∂⎡ ⎤+ + θ =⎢ ⎥∂ θ ∂λ ∂θ⎣ ⎦
total basin depth, h(x, y) is the unperturbed water depth (defined by bathymetric charts), g is the gravity acceleration, f is the Coriolis parameter (f = 2Ω sinθ) and Ω is the Earth rotation frequency (the rotation period is 24 hours), R is the Earth radius. In spherical coordinates, the system of equations (1) – (3) is hyperbolic, so it is important to use the fast numerical algorithms.
In practice, it is necessary to take into account the energy dissipation in the bottom turbulent layer in the near-shore zone. The simplest model here is the parametrization of the bottom friction by a quadratic formula (the last terms in the left side of Eqs. (1) – (3)) with a bottom roughness coefficient – n (the so-called Manning formula). This parameterization is widely used in river hydraulics, where special laboratory experiments were carried out, and roughness coefficients were determined for various soil types. Tsunami waves are not stationary hydraulic flows, but their period is large enough for the quasi-static Manning formula to be used. In calculations, as a rule, the constant value n = 0.015 m-1s is used, which is the natural bottom characteristic (sand, small pebbles). It should be noted that in the problems of the wave run-up taking into account real buildings, a variable value of the roughness coefficient is now used (see, e.g. (Choi et al, 2012)), it is important for detailed zonation of the coastal flooding by tsunami waves.
The main advantage of the ‘spherical’ system (1) – (3) is that it is tied to geographical coordinates, which makes it easy to locate the computation results on the ground. In addition, the bathymetry is available in an electronic form and also in the marine charts in geographical coordinates, so that, in contrast, there is a problem of converting bathymetry to a “flat” Earth. The system of shallow water equations should be supplemented by boundary and initial conditions. At the sea boundaries of the computational regions (for example, in straits), the condition of free wave passage through open boundaries is set, which is exact in the framework of the linear theory of shallow water without taking into account the Earth rotation and the simplest geometry of the maritime border
! , (4)
where the derivative ∂η/∂n is calculated normally to the outer boundary of the computational domain. This condition allows us to carry out computations in a limited area, however, we should understand, that in reality any water basin is variable in depth, and the wave coming out from the computation region can return to it after the reflection in the outer basin. This circumstance imposes restrictions on the wave computation time in the basin (which, unfortunately, is not often paid due attention to), therefore, the maritime border should be moved as far as possible.
The shoreline is generally mobile and moves along with the wave climbing the beach. This point is also singular for shallow water equations (the hyperbolicity of the water wave equations is violated in it), moreover, the dissipative term coefficient becomes infinite. The natural boundary condition of the moving shoreline is
! , (5)
that allows studying the wave run-up on the beach and calculating the characteristics of the coastal flooding caused by dangerous sea waves. To calculate the tsunami wave run-up in real water areas, it is
Vol. 38, No. 4, page 153 (2019)
η η 0ght n∂ ∂
+ =∂ ∂
( , , ) ( , ) ( , , ) 0D t h tλ θ = λ θ +η λ θ =
necessary to have a very good seabed bathymetry and topography. The effect of the quality of the bathymetry data on the accuracy of tsunami computations is analyzed in (Kulikov et al, 2016). It is shown that modern digital bathymetry maps, for example, GEBCO, do not provide the adequate reproduction of the bottom topography in numerical models of the wave propagation, so it can lead to significant errors in estimating the maximum tsunami run-up on the coast. In the subsequent work (Ivanova et al, 2017), this effect was demonstrated by the simulation example of the 2006 and 2007 Simushir tsunamis in the central part of the Kuril Islands. It is shown that computations performed on grids with a resolution of 30 arc seconds provide only a qualitative estimate of the distribution of tsunami heights along the coast. At the same time, a quantitative agreement between the simulation results and the observational data can be obtained only for grids with a spatial resolution of at least 10 arc seconds. For instance, the resolution in 3 m is used for the created inundation map of Turkey (Dilmen et al, 2015, Tufekci et al., 2018).
If cliffs are located on the shore that extends almost vertically into the water, the natural boundary condition is the full reflection one
! or ! , (6)
here n is the normal to the coastline. The artificial vertical wall is often used at a depth of 5 - 20 m (at the last sea points of the computational grid), on which the boundary condition is (6). This makes it possible to exclude the run-up zone, which is rather laborious in numerical computations (a very fine grid). To determine the run-up characteristics, the computed displacement distribution along the artificial vertical wall is matched with the analytical 1D solution (Choi et al, 2011).
The system of shallow water equations described above is implemented in many numerical models. We chose the international code TUNAMI, described in (Goto et al, 1997) as the original numerical code. It is still widely used in different countries. The numerical scheme is based on finite differences of the “leap-frog” type. Due to the hyperbolicity of the shallow water equations, the time step is chosen from the Courant condition for the stability of the numerical scheme
! , (7)
where Δλ and Δθ are the steps of the computational grid along the longitude and the latitude; Δt is the time step; hmax is the maximum basin depth. The TUNAMI code was improved by our team, in particular, the program was rewritten in C++ (it is a more advanced language compared to Fortran, in which the original TUNAMI program was written) and both the MPI (Message Passing Interface) parallel programming methods and the Open Multi-Processing were used, which significantly reduced the counting time. A new program interface was developed, (Fig. 1) which allows us not only to compute tsunami waves, but also to process the received data. In particular, it is possible to plot spatial distributions of the water level displacement at different time moments, the wave field maxima and minima (which, for simplicity, we call positive and negative amplitudes) and the wave heights along the coast. We called this code NAMI-DANCE, its documentation can be found in (http://namidance.ce.metu.edu.tr/pdf/NAMIDANCE-version-5-9-manual.pdf; http://skbsami.ru/name-dance; http://lmnad.nntu.ru/ru/projects /namidance/). This numerical model is registered in Russia (the
certificate of the state registration of computer codes No. 2014611028 of January 22, 2014) and copyright held in Turkey.
!
Figure. 1. The interface of the tsunami numerical code NAMI-DANCE.
3. NONLINEAR DISPERSIVE MODEL OF LONG WAVES IMPLEMENTED IN NAMI-DANCE
It is known that for relatively weak earthquakes or landslide tsunamis, the characteristic tsunami wavelength is comparable with the depth of the basin. In this case, the shallow water approximation is insufficient, therefore, it is necessary to use the dispersive long wave model. The same thing may be important for the strongest tsunamis too, if the tsunami source is highly anisotropic. So, the source of the 2004 Indonesian tsunami extended for almost 1000 km along the meridian and about 50 km along the longitude. Consequently, the wavelength varies roughly from 50 km to 1000 km, and in its lower limit it reaches the ocean depth, especially at large distances. The discussion of the dispersion importance for tsunami waves can be found in (Mirchina and Pelinovsky, 1982; Pelinovsky, 2007; Glimsdal et al, 2013; Nosov, 2017).
Nonlinear dispersion long-wave models (the so-called Boussinesq systems) have been already widely used for a long time. Probably, the first of these is the Peregrine system, which is valid for weak nonlinearity and dispersion (Peregrine, 1967). Then a strongly nonlinear weakly dispersive model appears, and here we give some references (Madsen et al, 2003; Pelinovsky, 2007; Kirby et al, 2013; Fedotova et al, 2015; Shokin et al, 2015; Khakimzyanov et al, 2018). Some of them are implemented in the popular open-access codes FUNWAVE (Kirby et al, 1998) and COOLWAVE (Cheung et al, 2003), used to simulate tsunamis and storm surges. However, to implement numerical algorithms to solve the nonlinear dispersion theory equations we need smaller steps in space and time to calculate high-order derivatives. As a result, the computation time increases sharply and numerical errors are accumulated in the computation process over a long distance. It can be indicated, for example, that most of the
Vol. 38, No. 4, page 155 (2019)
numerical schemes of nonlinear dispersion Boussinesq type equations turn out to be unstable (Lovholt and Pedersen, 2009). Another difficulty is the lack of bathymetric maps with a detailed resolution, and their interpolation to small scales is not accurate.
Some other ways of dispersion generalization of shallow water equations that are not related to the introduction of explicitly high derivatives are suggested in (Yoon, 2002; Yoon et al, 2007). This approach is based on the use of numerical dispersion instead of the physical one. The main idea of the method is to change the numerical scheme of finite differences, discretizing the shallow water equations, for the value of the numerical error to lead to the water wave dispersion, as if the dispersion theory equations were strictly solved. It can be demonstrated on the example of one-dimensional Boussinesq equation, to which, in the simplest case, the dispersion long wave theory equations are reduced
! , (8)
where is the long wave velocity. If we now consider the wave equation (without the last term in (8)), solve it numerically by using the “leap-frog” method as it is done with the initial system of shallow water equations, we will get a different analogue of the wave equation. If we expand all the differences in a Taylor series and restrict them by the first few terms, the differential equation can be obtained again
, (9)
where the last term appears, cr = c0Δt/Δx is the known Courant number, Δt and Δx are the temporal and spatial steps of the numerical scheme. It is important to note that the “leap-frog” scheme makes it possible to exclude the appearance of the third derivative in (9), which would lead to either energy dissipation or instability of the numerical scheme (this is where the requirement to the smallness of the Courant number is followed). However, it is not possible to exclude the terms of the fourth derivative, describing the wave dispersion. Comparing equations (8) and (9) now, we obtain their identity if the step size is chosen equal to
or , (10)
which allows one to connect spatial and temporal steps between each other. This idea is also valid for a variable (but gradual variable) basin depth; in this case, it is necessary to use variable grids (Yoon, 2002). Technically, in (Yoon, 2002) it was proposed to use a constant time step value. As a result, there is no freedom in choosing the spatial grid step because of (10). At variable depths, the spatial grid can be very inhomogeneous, and, therefore, not always convenient for computations. We suggested changing the time step, making the spatial grid more homogeneous. This strategy was implemented in NAMI-DANCE, what allowed us to significantly increase the efficiency of using the numerical dispersion. The description of the scheme is given in (Kian et al, 2018).
Vol. 38, No. 4, page 156 (2019)
2 22 2 42 002 2 4 0
3c h
ct x x∂ η ∂ η ∂ η
− − =∂ ∂ ∂
0c gh=
( )2 2 2 4
2 2 20 02 2 4
( ) 1 012 rx
c c ct x x∂ η ∂ η ∂ η
− − − =∂ ∂ ∂
Δ
22 / 1 rx h c= −Δ 2 24 ( )x h gh t= +Δ Δ
We would like to mention that both versions of NAMI-DANCE (with/without dispersion) demonstrated good accuracy on various benchmarks (Velioglu et al, 2016; Lynett et al, 2017; Sogut and Yalciner, 2019).
4. TSUNAMI GENERATION BY EARTHQUAKES
In the case of the tsunami wave generation caused by underwater earthquakes, the seismic process in the source ends quickly enough, and the wave does not have time to leave the source. In this case, the seabed movements can be considered instantaneous, causing the same instantaneous water surface displacement. Then, the wave generation process is reduced to the setting of the initial conditions
! , ! . (11)
Some time ago, the information about the earthquake was very fragmentary (the coordinates of the earthquake and its magnitude were more or less given), so the initial conditions were determined through empirical formulas (see, for example, (Wells and Coppersmith, 1994; Pelinovsky, 1996)). Now the information is more complete (literally a few minutes after the earthquake) and includes the rupture plane position, the dislocation magnitude, etc. The characteristic parameters of the earthquake source, currently determined quite well, are shown in Fig. 2. They are the fault depth, the length (L) and the width (W) of the fault, the displacement (D), the angle between the meridian and the fault line (Strike angle, θ), the plate slope (Dip angle, δ), the shear angle of the plate (Slip angle, λ). Consequently, it is possible to perform a more complete calculation of residual displacements in the source, based on the theory of half-space elasticity; the corresponding formulas are given in a number of works. In most of our calculations, we use the formulas given in the paper (Okada, 1985) allowing us to determine the function η0(λ, θ) in (8) from the above earthquake characteristics (this approach is now generally accepted). Okada’s analytical formulas are very cumbersome (they occupy several pages), by now they have been computerized, therefore, the calculation of the initial water surface displacements can be fulfilled almost instantly.
!
Figure 2. The earthquake parameters
During the strongest earthquakes with magnitudes of the order of 9, the source rupture occurs not instantly, but within several minutes and even tens of minutes. These times are comparable to the characteristic period of tsunami waves, so it is necessary to solve the shallow water equations with the
right-hand side. In theoretical aspect, this problem is discussed, for example, in books (Pelinovsky, 1996; Levin and Nosov, 2009). In practice, however, the extended process of tsunami wave excitation is replaced by several ones, setting instantaneous movements in different time moments and in different points (obtained from the seismic data). The obtained solutions in the open part of the sea are added linearly, and then the general tsunami wave propagation process is analyzed. This procedure is implemented in the NAMI-DANCE model. There are other models used to describe the processes that occur under the influence of the earthquake, for example, a keyboard model developed in (Lobkovsky et al, 2006).
To conclude this section, we list several recently published articles where the NAMI-DANCE model has been used to compute real tsunamis and prognostic inundation maps (Dilmen et al, 2015; Aytore et al, 2016; Zaytsev et al, 2017; Kostenko et al, 2018; Tufekci et al, 2018).
5. TSUNAMI GENERATION BY UNDERWATER LANDSLIDES
According to the statistics, about 85% of tsunami waves are generated by underwater earthquakes, and only a small part is generated by landslides or rockfalls. Landslides often arise as a result of earthquakes, even weak ones, and they can already cause noticeable tsunamis. Such landslides were observed after the most famous catastrophic tsunamis of the seismic origin in recent years, including the last tsunami in Indonesia in 2018, which was mentioned in the Introduction (Sassa and Takagawa, 2019). The viewpoint that tsunamis are caused mainly by landslides and not directly by bottom movements, is rather popular (Yalciner et al, 2003).
In many cases, devastating landslides are produced by local processes in the absence of seismic events. Large amplitude waves associated with landslides were observed in Alaska, Japan and many of the fiords of Norway. Thus, the descent of the 300 million cubic meters avalanche to Lituya Bay (Alaska, July 10, 1958) led to the 60m high tsunami formation, the maximum run-up in the bay itself reached the record value of 524 m. It has already been said about the 40m high tsunami in the Bureya river valley in Russia in 2018. Even earlier, on September 1, 2017 in Baksan in the Caucasus (Russia), a landslide coming down into the river, had generated a wave of 1 m in height (one person died). A tsunami landslide often leads to catastrophic consequences, therefore, the problem of landslide generated waves is of great practical interest.
The numerical simulation of tsunami generation by underwater landslides is a difficult problem, and a number of models have been developed which depend on the rheology of the landslide material; they are discussed in (Imamura and Imteaz, 1995; Pelinovsky and Poplavsky, 1997; Kulikov et al, 1998; Garagash et al, 2003; Pudasaini and Hutter, 2006; Beisel et al, 2010). One of these “shallow-water” models called a two-layer model developed in (Imamura and Imteaz, 1995) is implemented in the NAMI-DANCE code. In the framework of this model, a landslide is assumed to be viscous-liquid and the problem geometry is presented in Fig. 3. The basic equations of the two-layer model have the following form (in Cartesian variables, excluding the Earth rotation)
Vol. 38, No. 4, page 158 (2019)
! , (12)
! , (13)
! , (14)
! , (15)
! (16)
! (17)
! Figure 3. The problem geometry: upper and lower layers correspond to water and landslide,
where η1 is the water surface displacement; η2 is the landslide height; M1 and N1 are the water flows along the x and y axes; D1 = h1 + η1 is the total water depth and h1 is the undisturbed water depth; M2 and N2 are the landslide mass flows along x and y; D2 = h2 + η2 is the total landslide thickness and h2 is the undisturbed landslide thickness; m1 and m2 are the roughness coefficients of the landslide surface (“liquid” bottom) and solid bottom, respectively, ρ1 is the water density and ρ2 is the landslide material density; β is the turbulent viscosity coefficient. It should be immediately noted that the landslide viscosity is taken into account phenomenologically only in one equation (15), providing the landslide spread and its surface smoothness. The vertical structure of the landslide flow is assumed to be uniform in the cross section as in the Savage – Hutter model (Pudasaini and Hutter, 2006). The applicability of such approximations is described in (Imamura and Imteaz, 1995).
The initial conditions determine the landslide location
, M2 = 0, N2 = 0, (18)
the water surface remains unperturbed (zero initial conditions for η1, M1 and N1). As a result, it is possible to compute the landslide motion and the waves caused by its movement.
Considering the fact that the wavelength of landslide tsunamis is usually much less than those caused by the earthquake, accounting for dispersion effects seems important. That is why we have improved the numerical scheme for the system of water layer equations (12) − (15), while the landslide is described by non-dispersive shallow water equations (16) − (17). This procedure is also implemented into the NAMI-DANCE software. The inclusion of dispersion in the water layer was carried out according to the same methodology as in Sec. 2 with the replacement of the physical dispersion by a numerical one.
This model was used to estimate the prognostic tsunami in the Nile Delta (Egypt) (Yalciner et al, 2014).
6. METEOTSUNAMI
Sometimes, large waves are generated by the sharp spatial and temporal changes in the atmospheric pressure. A low atmospheric pressure leads to an increase in the water level in the marine area part, and a high atmospheric pressure leads to a decrease in the water level of another zone (“the inverse barometer” law), so that the water level in the sea is distorted. The waves arising from it can be amplified due to the resonance properties of the coastal morphology (the Proudman and the Greenspan resonances). The observation review of the meteo-tsunami is given in (Pattiaratchi and Wijeratne, 2015). Meteorological long waves have the same wave period as conventional tsunamis (from a few minutes to 2-3 hours), they can cause human casualties and destructive effects in coastal zones, like ordinary tsunamis. For the meteorological tsunami, such generation mechanisms as spatial and temporal pressure distributions (the atmospheric fronts) and the atmospheric gravitational waves are important.
One of the widely known tsunami events of this kind was observed on June 23-27, 2014 in the Mediterranean and Black Seas from Spain to Ukraine (Sepic et al, 2015). In particular, on June 27, 2014, 1-2m high waves suddenly came to Odessa coast. Since significant earthquakes in the
Vol. 38, No. 4, page 160 (2019)
2 0η ( , , 0) η ( , )x y t x y= =
Mediterranean and Black Sea regions were not recorded during these events, the atmospheric disturbances are assumed to be a possible cause.
The equations of the long-wave model now include atmospheric effects; in particular, equations (1) − (2) are modified
! , (19)
! , (20)
and equation (3) remains unchanged. Here patm is the atmospheric pressure in Pascals, which is taken from meteorological maps, usually transmitted with an interval of one hour. Pressure distribution charts with an interval of 6 hours are available on the NASA website [https://gemini.gsfc.nasa.gov/aplo]. For computations, these maps are interpolated in time to obtain the pressure field at each time step.
Unfortunately, in this case, tsunami events are actually missing when a sharp change in the atmospheric pressure occurs in the given place within a short period of time, it just leads to the meteorological tsunami generation. Therefore, it is rather difficult to obtain the agreement with the observed data, especially if there are no more frequent observations in coastal locations. A strong amplification of meteorological tsunamis occurs at resonance, when the velocity of the atmospheric front is close to the long wave celery in the basin (the Proudman resonance) and such effects are noted in a number of regions of the World Ocean.
The NAMI-DANCE software modification with the atmospheric pressure inclusion in it is not difficult (the equation order does not change), with the exception of technical problems associated with the atmospheric maps described above. This modification has also passed verification on the well-known theoretical solutions of the linear theory of shallow water in basins with an idealized bathymetry (Metin et al, 2017).
7. TSUNAMI WAVE IMPACT ON COASTS AND STRUCTURES
The computations of wave characteristics (the water surface displacement and flow velocity) are now routine actions in tsunami modeling. The computation of tsunami force characteristics is only beginning. In the practice of wind waves loads, these characteristics are computed at the second stage, when the wave regime in the sea is known and it can be easily recounted to any point in the harbor. This way of work is not acceptable for a tsunami. Tsunami waves are long waves, the scale of which is comparable
to the size of coastal structures (harbors and ports), so it is impossible to recalculate the wave field from the open sea to the design area of construction using simple formulas in the framework of the ray tracing method.
Therefore, the design of large objects (port, dam, etc) in the coastal zone should always be accompanied by the wave regime modeling taking into account the planned structures, since the resonant properties of the water areas and the locations of the wave extrema change. Only for relatively simple small-scale structures, it is possible to carry out wave field computations independently of the planned structures, and then calculate the load on them separately. Such situation, in particular, is realized in the case of single vertical piles of a small diameter (compared to the tsunami wavelength). In some cases, the pile diameter is so small that it is “subgrid” in numerical meshes, and the perturbation inserted by the object practically does not affect the wave field. In such cases, the classical formulas for forces and moments can be used. In particular, the pressure and drag forces are
! , ! (21)
(ρw is the sea water density and S is the cross-sectional area of the structure against the direction of the wave propagation). It can be seen that they are proportional to the kinematic characteristics of the tsunami waves (the water displacement and the flow velocity), and all construction characteristics (the pile diameter, the wall width) determine only the numerical coefficients in (21). Therefore, the calculation of the tsunami impact on the small diameter objects can be performed only after the wave field analysis in the water area is fulfilled.
The calculation of changes in the seabed and coasts is more complicated, therefore, it requires using three-dimensional transport models. Meanwhile, the sediment transport can be roughly characterized by the Rouse number (Julien, 1995). It is a non-dimensional number in fluid dynamics used to define a concentration profile of suspended sediment and which also determines how the sediment will be transported in a flowing fluid. Mathematically, it is the ratio between the sediment fall velocity Ws and the upwards velocity on the grain as product of the von Karmán constant kand the shear velocity u*
! , (22)
where β is the ratio of vortex viscosity to vortex diffusion. The value of the particle fall rate Ws depends on their size and density, and is constant for each sediment type. The shear velocity can be roughly related to the water flow velocity u* ~ u, so that the Rouse number is inversely proportional to the water flow velocity. It is generally accepted that the sediment movement initiation occurs at Rouse numbers of about 7, and the bed-load movement occurs in the range 2.5 – 7.5. The movement of the suspended sediments corresponds to the Rouse numbers 0.8 – 1.2. At lower values of the Rouse number, a strong movement of the bottom sediments occurs, leading to significant bottom deformations. By calculating
the water flow velocity and determining the Rouse number, we can determine the character of the bottom processes. For areas of intensive changes in the bottom morphology, we can then perform a separate simulation of the quantitative characteristics of the sediment transport using various transport models.
We would like to note also the estimates of the possible house and various construction destruction on the shore produced by tsunami waves. For many areas, when the character of the buildings are taken into account, the so-called fragility curves are given that relate the percentage of damage or the relative number of deaths to the tsunami height on the coast (Nanayakkara and Dias, 2016). Such curves are determined for some areas of Japan, Indonesia, Egypt and the United States.
The calculation of all the parameters listed above is easily inserted into any computer model, and in particular, it has already been done in our NAMI-DANCE software (Ozer et al, 2015a,b; Kian et al, 2016). Let us note here that since 2018 in Russia the Code of Rules No. 292.1325800.2017 “Buildings and constructions in tsunami-hazardous areas. Design Rules”, approved on June 23, 2017 by the Ministry of Construction and Housing and Communal Services of the Russian Federation has been used.
8. CONCLUSION
This brief review provides the hydrodynamic models used in the tsunami problem which are solved with the help of the NAMI-DANCE software. They are based on the well-known two-dimensional equations of the shallow water theory, adapted to the rotating Earth, the available bathymetry database and the land topography. The numerical model takes into account the basic mechanisms of tsunami wave generation and dissipation. In some cases, it is necessary to take into account the water wave dispersion leading to the increase in the order of differential equations. In our computations, the physical dispersion is replaced by a numerical one, which allowed us not to increase the computation time significantly. The developed NAMI-DANCE software is actively used to simulate real events, prognostic tsunamis, inundation maps and tsunami force characteristics.
ACKNOWLEDGEMENTS
This study was initiated in the framework of the state task program in the sphere of scientific activity of the Ministry of Science and Higher Education of the Russian Federation (project No. 5.4568.2017/6.7 and 5.5176.2017/8.9) and financially supported by this program, the grant of the President of the Russian Federation for state support of leading scientific schools of the Russian Federation NSh-2685.2018.5 and grants RFBR (18-05-80019). EC project ASTARTE-Assessment, Strategy And Risk Reduction for Tsunamis in Europe—FP7-ENV2013 6.4-3, Grant 603839, UDAP-Ç-12-14 project granted by Disaster Emergency Management Presidency of Turkey (AFAD) and TUBITAK 113M556, 108Y227 and 213M534 Projects are also acknowledged.
Vol. 38, No. 4, page 163 (2019)
REFERENCES
1. Aytore, B., Yalciner, A.C., Zaytsev, A., Cankaya, Z.C., and Suzen, M.L. Assessment of tsunami resilience of Haydrapasa port in the Sea of Marmara by high-resolution numerical modeling. Earth, Planets and Space, 2016, vol. 68, 139-150.
2. Beisel, S.A., Khakimzyanov, G.S., and Chubarov, L.B. Simulation of surface waves generated by an underwater landslide moving along a spatially irregular slope. Computational Technologies, 2010, vol. 15, 39-51. 3. Cheung, K.F., Phadke, A.C., Wei, Y., Rojas, R., Douyere, Y.J.-M., Martino, C.D., Houston, S.H., Liu, P.L.-F., Lynett, P.J., Dodd, N., Liao, S., and Nakazaki, E. Modeling of storm-induced coastal flooding for emergency management. Ocean Engineering, 2003, vol. 30, 1353-1386.
4. Choi, B.H., Kaistrenko, V., Kim, K.O., Min, B.I., and Pelinovsky, E. Rapid forecasting of tsunami runup heights from 2D numerical simulation data. Natural Hazards and Earth System Sciences, 2011, vol. 11, No. 3, 707-714.
5. Choi, J.W., Kwon, K.K., and Yoon, S.B. Tsunami inundation simulation of a built-up area using equivalent resistance coefficient. Coastal Engineering Journal, 2012, vol. 54, 1250015-1-25.
6. Dilmen, D.I., Kemec, S., Yalciner, A.C., Düzgün, S., and Zaytsev, A. Development of a tsunami inundation map in detecting tsunami risk in Gulf of Fethiye, Turkey. Pure and Applied Geophysics, 2015, vol. 172, 921-929.
7. Fedotova, Z.I., Khakimzyanov, G.S., and Gusev, O.I. History of the development and analysis of numerical methods for solving nonlinear dispersive equations of hydrodynamics. I. One-dimensional models problems. Computational Technologies, 2015, vol. 20, 120-156.
8. Garagash, I.A., Lobkovsky, L.I., Kozyrev, O.R., and Mazova, R.Kh. Generation and runup of tsunami waves at an submarine landslide. Oceanology, 2003, vol. 43, 173-181.
9. Giachetti, T., Paris, R., Kelfoun, K., and Ontowirjo, B. Tsunami hazard related to a flank collapse of Anak Krakatau Volcano, Sunda Strait, Indonesia. Geological Society, London, Special Publications, 2012, vol. 361, 79-90.
10. Glimsdal, S., Pedersen, G.K., Harbitz, C.B., and Lovholt, F. Dispersion of tsunamis: does it really matter? Natural Hazards and Earth System Sciences, 2013, vol. 13, 1507-1526.
11. Goto, C., Ogawa, Y., Shuto, N., and Imamura, F. IUGG/IOC TIME Project: Numerical Method of Tsunami Simulation with the Leap-Frog Scheme. Manuals and Guides. Paris, Intergovernmental Oceanographic Commission of UNESCO, 1997. No.35. 126 p.
12. Ivanova, A.A., Kulikov, E.A., and Fain, I.V. On modelling 2006, 2007 Simushir tsunamis in the central Kuril region. Fundamentalnaya i Prikladnaya Gidrofizika, 2017, vol. 10, 56-64 [in Russian].
Vol. 38, No. 4, page 164 (2019)
13. Imamura, F., and Imteaz, M.A. Long waves in two layer: governing equations and numerical model. Science of Tsunami Hazards, 1995, vol. 13, No. 1, 3-24.
14. Julien. P.Y. Erosion and Sedimentation, Cambridge University Press, Cambridge, 1995.
15. Khakimzyanov, G., Dutykh, D., Fedotova, Z., and Mitsotakis, D. Dispersive shallow water wave modelling. Pt I: Model derivation on a globally flat space. Communications in Computational Physics, 2018, vol. 23, 1-29.
16. Kian, R., Horrillo, J., Zaytsev, A., and Yalciner, A.C. Capturing physical dispersion using a nonlinear shallow water model. Journal of Marine Science and Engineering, 2018, vol. 6, 84-1-11
17. Kian, R., Velioglu, D., Yalciner, A.C., and Zaytsev, A. Effects of harbor shape on the induced sedimentation; L-type basin. J Marine Science and Eng., 2016, No. 4, 55-65.
18. Kirby, J.T., Shi, F., Tehranirad, B., Harris, J.C., and Grilli, S.T. Dispersive tsunami waves in the ocean: Model equations and sensitivity to dispersion and Coriolis effects. Ocean Modelling, 2013, vol. 62, 39-55.
19. Kirby, J.T., Wei, G., Chen, Q., Kennedy, A.B., and Dalrymple, R.A. FUNWAVE 1.0. Fully Nonlinear Boussinesq Wave Model. Documentation and User's manual. Research Report No. CACR-98-06, Univ. of Delaware. 1998. 80 p.
20. Kostenko, I.S., Zaytsev, A.I., Minaev, D.D., Kurkin, A.A., Pelinovsky, E.N., and Oshmarina, O.E. The Moneron Tsunami of September 5, 1971, and its manifestation on the Sakhalin island coast: numerical simulation results. Izvestiya - Atmospheric and Ocean Physics, 2018, vol. 54, 1-9.
21. Kulikov, E.A., Gusiakov, V.K., Ivanova, A.A., and Baranov, B.V. Numerical tsunami modeling and the bottom relief. Moscow University Physics Bulletin, 2016, vol. 71, 527-536.
22. Kulikov, E.A., Rabinovich, A.B., Fain, I.V., Bornhold, B.D., and Thomson, R.E. Tsunami generation by landslides at the Pacific coast of North America and the role of tides. Oceanology, 1998, vol. 38, 323-328.
23. Levin, B., and Nosov, M. Physics of Tsunamis. Springer, 2009.
24. Lobkovsky, L.I., Mazova, R.Kh., Garagash, I.A., Kataeva, L.Yu., and Nardin, I. To analysis of source mechanism of the 26 December 2004 Indian Ocean tsunami. Russ. J. Earth Sci., 2006, vol. 8, ES5001.
25. Lovholt, F, and Pedersen, G. Instabilities of Boussinesq models in nonuniform depth. International Journal for Numerical Methods in Fluids, 2009, vol. 61, 606-637.
Vol. 38, No. 4, page 165 (2019)
26. Lynett, P.J., Montoya, L., Gately, K., Wilson, R., Arcas, D., Aytore, B., Dogan, G.G., Kian, R., Velioglu, D., Yalçıner, A.C., Bai, Y., Cheung, K.F., Heitmann, T.W., Bricker, J.D., Roeber, V., Castro, M.J., Escalante, C., Macías, J., Yamazaki, Y, David, C.G., Gonzalez-Vida, J.M., Grilli, S.T., Shelby, S., Horrillo, J., Pampell-Manis, A., Kanoglu, U., Sharghivand, N., Kirby, J.T., Shi, F., Tehranirad, B., Li, W., Thio, H.K., Nicolsky, D.J., Ortega, S., Park, Y.S., Tolkova, E., Zaytsev, A., and Zhang, Y.J. Inter-model analysis of tsunami-induced coastal currents. Ocean Modelling, 2017, vol. 114, 14-32.
27. Madsen, P., Bingham, H.B., and Schaffer, H.A. Boussinesq-type formulations for fully nonlinear and extremely dispersive water waves: derivation and analysis. Proc. R. Soc. Lond. A, 2003, vol. 459, 1075-1104.
28. Makhinov, A.N., Kim, V.I., Ostroukhov, A.V., and Matveenko, D.V. Large landslide in the valley of the Bureya River and tsunami in the reservoir of the Bureya hydroelectric power station. Vestnik of the Far East Branch of the Russian Academy of Sciences, 2019, no. 2, 35-44 [in Russian].
29. Metin, D., Pelinovsky, E., Yalciner, A., Zaytsev, A., Tarakcioglu, O.G., Yalciner, B., Kurkin, A. meteotsunami: generation, propagation and amplification. Proc. Thirteenth Int. Conference on the Mediterranean Coastal Environment, MEDCOAST 17, 2017, vol. 2, 1143-1154.
30. Mirchina, N.R., and Pelinovsky, E.N. Nonlinear and dispersive effects for tsunami waves in the open ocean. Int. J. Tsunami Soc., 1982, vol. 2, no. 4, D1 - D9.
31. Nanayakkara, K.I.U., and Dias, W.P.S. Fragility curves for structures under tsunami loading. Natural Hazards, 2016, vol. 80, 471-486.
32. Nosov, M.A. Applicability of long-wave approximation to the description of tsunami dynamics. Moscow University Physics Scientific Notes, 2017, no. 4, 1740503-1-7 [in Russian].
33. Okada, Y. Surface deformation due to shear and tensile faults in a half-space. Bulletin of the Seismological Society of America. 1985, vol. 75, 1135-1154.
34. Omira, R., Dogan, G.G., Hidayat, R., Husrin, S., Prasetya, G., Annunziato, A., Proietti, C., Probst, P., Paparo, M.A., Wronna, M., Zaytsev, A., Pronin, P., Giniyatullin, A., Putra, P.S., Hartanto, D., Ginanjar, G., Kongko, W., Pelinovsky, E., and Yalciner, A.C. The September 28th, 2018, tsunami in Palu-Sulawesi, Indonesia: A post-event field survey. Pure and Applied Geophysics, 2019, vol. 176, 1379-1395.
35. Ozer, S.C., Yalciner, А.С., and Zaytsev, А. Investigation of tsunami hydrodynamic parameters in inundation zones with different structural layouts. Pure and Applied Geophysics, 2015a, vol. 172, 931-952.
36. Ozer, S.C., Yalciner, A.C., Zaytsev, A., Suppasri, A., Imamura, F. Investigation of hydrodynamic parameters and the effects of breakwaters during the 2011 Great East Japan Tsunami in Kamaishi Bay. Pure and Applied Geophysics, 2015b, vol. 172, 3473-3491.
Vol. 38, No. 4, page 166 (2019)
37. Pattiaratchi, C.B., and Wijeratne, E.M.S. Are meteotsunamis an underrated hazard? Phil. Trans. R. Soc. A, 2015, vol. 373, 20140377-1-23.
39. Pelinovsky, E.N. Nonlinear – dispersive theory of tsunami waves: outlook after the hazardous Indian Ocean tsunami. Nonlinear waves’ 2006. Nizhny Novgorod, 2007, 393-407 [in Russian].
40. Pelinovsky, E., and Poplavsky, A. Simplified model of tsunami generation by submarine landslides. Phys. Chem. Earth, 1997, vol. 21, no. 1/2, 13-17.
41. Peregrine, D.H. Long waves on a beach. J. Fluid Mechanics, 1967, vol. 27, 815-882.
42. Pudasaini, S.P., and Hutter, K. Avalanche Dynamics: Dynamics of Rapid Flows of Dense Granular Avalanches. – Springer, 2006.
43. Sassa, Sh., and Takagawa, T. Liquefied gravity flow-induced tsunami: first evidence and comparison from the 2018 Indonesia Sulawesi earthquake and tsunami disasters. Landslides, 2019, vol. 16, 195-200.
44. Sepic, J., Vilibiс, I., Rabinovich, A.B., and Monserrat, S. Widespread tsunami-like waves of 23-27 June in the Mediterranean and Black Seas generated by high-altitude atmospheric forcing. Scientific reports, 2015, vol. 5, 11682-1-8.
45. Science Global News, 2018. http://scienceglobalnews.com/environment/meteotsunami-floods-islands-of-mallorca-and-menorca-spain (accessed 28 March 2019).
46. Shokin, Yu.I., Fedotova, Z.I., and Khakimzyanov, G.S. Hierarchy of nonlinear models of the hydrodynamics of long surface waves. Doklady Physics, 2015, vol. 60, No. 5, 224-228.
47. Sogut, D.V., and Yalciner, A.C. Performance Comparison of NAMI DANCE and FLOW-3D models in tsunami propagation, inundation and currents using NTHMP benchmark problems. Pure Appl. Geophys., 2019, vol. 176, 3115–3153
48. Tufekci, D., Suzen, M.L., Yalciner, A.C., and Zaytsev, A. Revised MeTHuVA method for assessment of tsunami human vulnerability of Bakirkoy district, Istanbul. Natural Hazards, 2018, vol. 90, 943-974.
49. Velioglu, D. Kian, R., Yalciner, A.C., and Zaytsev, A. Performance assessment of NAMI DANCE in tsunami evolution and currents using a benchmark problem. Journal of Marine Science and Engineering, 2016, vol. 4, 49-1-8.
50. Wells, D.L., and Coppersmith, K.J. New empirical relationships among magnitude, rupture length, rupture width, rupture area, and surface displacement, Bull. Seism. Soc. America, 1994, vol. 84, 974-1002.
Vol. 38, No. 4, page 167 (2019)
51. Yalciner, A.C., Pelinovsky, E.N., Okal, E, and Synolakis, C.E. (Eds) Submarine landslides and tsunamis. Kluwer, 2003.
52. Yalciner, A.C., Zaytsev, A., Aytore, B., Insel, I., Heidarzadeh, M., Kian, R., Imamura, F. A possible submarine landslide and associated tsunami at the northwest Nile Delta, Mediterranean Sea. Oceanography, 2014, vol. 27, 68-75.
54. Yoon, S.B., Lim, C.H., and Choi, J. Dispersion-correction finite difference model for simulation of transoceanic tsunamis. Terrestrial Atmospheric and Oceanic Sciences, 2007, vol. 18, 31-53.
55. Zaytsev, A., Kostenko, I., Kurkin, A., Pelinovsky, E., Pararas-Carayannis, G. Manifestation of the 1963 Urup tsunami on Sakhalin: observations and modeling. Science of Tsunami Hazards, 2017, vol. 36, 145-166.
Vol. 38, No. 4, page 168 (2019)
ISSN xxxx-xxxx
SCIENCE OF TSUNAMI HAZARDS
Journal of Tsunami Society International
Volume 38 Number 4 2019
BETA TESTING FOR INCREASED ACCURACY AND IMPROVED PERFORMANCE OF THE INDONESIAN TSUNAMI EARLY WARNING APPLICATION (Ina-TEWA)
Madlazim1,2, Tjipto Prastowo1,2
1Physics Department, The State University of Surabaya, Surabaya 60231, Indonesia 2Center for Earth Science Studies, The State University of Surabaya, Surabaya 60231, Indonesia
ABSTRACT Beta testing is a final stage for product examination following a series of product development.
Regarding this, beta testing performed for an increased accuracy of Indonesian tsunami early warning application (Ina-TEWA) was applied to real conditions of Indonesian earthquakes with magnitudes of 6.0 or greater that occurred during a time period from January 1, 2018 to July 1, 2019. This stage was conducted in real-time for use of tsunami early warnings officially managed by Centre for Research and Development, The Indonesian Agency for Geophysics, Climatology and Meteorology (BMKG). It is frequently preceded by a series of preliminary tests known as alpha testing usually designed only for laboratory-scaled examination that is limited and offline. While implemented in real-time tsunami warnings, beta testing is considered as examination of acceptance given by and hence direct feedback from users. The launch for this tsunami early warning application is of great importance in the sense that it measures directly tsunami parameters from the earthquake source. The main aim of this stage is to identify possible errors, if any, and to make them the errors minimum for an increased accuracy in tsunami assessment to a maximum value possible. Based on beta testing applied to 35 events during the time period examined here, it was merely 7 errors occurred and found for tsunami false warnings. It follows that the Ina-TEWA prediction for accurate tsunami warnings reaches to a value of 80% or false warnings by this application remains relatively high of 20%. Therefore, tsunami early warning application for future use requires a further increase in terms of accuracy.
Accurate and quick release of tsunami early warnings following major tsunamigenic earthquakes generated by tectonic movement may prevent local communities and surrounding infrastructures from severe catastrophes (Satriano et al., 2011; Suppasri et al., 2015). Soon after tsunami wave generation, tsunami passage in the open ocean and its continuing arrival times at coastal regions both near and far from the epicenter need to be predicted accurately (Prastowo et al., 2018) using rapid assessment of observed waveforms recorded by instrument, such as ocean bottom electro-magnetometer deployed at the ocean floor, Deep-ocean Assessment and Reporting of Tsunami (DART) surface buoys offshore, or tide gauges onshore. Within this context, tsunami waveform analysis requires a reliable software or application, which is technically developed for this matter.
Software development towards a reliable application or programming package includes a number of consecutive stages, that is, analysis of requirement, design of development, integration and testing, and deployment and maintenance. With respect to these, this study focuses on beta testing as part of sequential procedures for an increased accuracy and hence improved performance of the Indonesian tsunami early warning system (Ina-TEWS, a former system of tsunami warnings before Ina-TEWA), which has been operated since 2008 by the Indonesian Agency for Geophysics, Climatology and Meteorology (BMKG). However, as addressed by Madlazim and Prastowo (2016), the Ina-TEWS performance remains relatively low in terms of accuracy thereby requiring further improvement.
Kocbek and Heričko (2013) argued for the quality of a system that is driven by software-assisted performance. The performance is determined by a number of factors, such as accuracy, time-lapse, completeness, consistency, maintenance, security, safety, reliability, and usability. During the stages of system improvement, there are opportunities to find possible errors in various stages of product examination and development. Testing is thus generally applied for a specific purpose as a main tool to ensure a software of high quality, providing specific methods and corresponding techniques for error detection in such a system (Budnik, 2012; Kocbek and Heričko, 2013). However, software bugs are always present because the complexity of a developed software in general cannot be fully solved. It follows that design limitation can never be completely avoided, in particular when dealing with complex software-supported systems (Pan, 1999; Mohd and Shahbodin, 2015).
For improvement of performance, software testing in any stage of product development is carried out normally for the following main objectives: (1) quality assurance; (2) verification and validation; and (3) reliability estimation, where each of these refers to the suitability of specified product design requirement. When these are achieved, the high quality of product is attained in some circumstances. Debugging, a common way of software testing, is usually carried out to search for weaknesses of a program hence assuring the program being able to run well during performance. This process is of primary importance as it is impossible to produce an adequately complex program left uncorrected in
Vol. 38, No. 4, page 170 (2019)
its first-time use. According to Madlazim et al. (2015), another important purpose of software testing is verification and validation. In the process of fulfilling this purpose, software testing is a measure of its applicability hence reliability. Software reliability is therefore directly connected to the number of tests completed. Within the uncertainty during test measurements, software testing can therefore be used to effectively collect a set of data failures or unsuccessful performance for estimates of product reliability in future use (Pan, 1999; Madlazim et al., 2015).
As previously mentioned, software testing is complex, containing a series of tests. A preliminary test called alpha testing is performed by prospective users or independent programmers involved in the product development division. The primary aims of this test are to detect any error in early stages of development and to bring the corrected syntax embedded in a software to a further test for detailed examination, namely beta testing (Pradhan, 2012). In the present study, beta testing of the Ina-TEWA is reported. The primary purpose is to identify how many tests required for optimum performance to reveal a number of design defects in this tsunami early warning application and to reduce them all to a minimum since the application performance is measured in terms of its accuracy. We present here beta testing for 35 events during a time period from January 1, 2018 to July 1, 2019 in the country, recorded by a network of seismic stations managed by BMKG and then analysis all of the results. Discussions on these results are given before concluding remarks at the end of this work.
1. INDONESIAN TSUNAMI EARLY WARNING APPLICATION (Ina-TEWA)
Tsunami monitoring application in Indonesian territories, formally abbreviated as the Ina-TEWA, has currently used main earthquake-tsunami discriminants, such as rupture duration " , the duration exceed 65 seconds " and the dominant period " of the first arrivals of P-waves following a series of pioneering work by Lomax and Michelini (2009, 2011, 2012). The primary difference between the Ina-TEWA and these studies is the use of a network of local stations for the Ina-TEWA, rather than teleseismic observations, to record seismic activities generated by earthquakes of varying magnitudes. In the early stages before the Ina-TEWA was applied, the Ina-TEWS extracted seismic signals from a local network nearby an event and transformed it into a digital set of source parameters ready for analysis and assessment, including earthquake magnitude, origin time, epicenter and hypocenter. However, as claimed by Madlazim and Prastowo (2016), most Indonesian earthquakes with epicenters in the ocean with depths shallower than 70 km and magnitudes greater than 7.0 during 2007-2010, were falsely announced as tsunamigenic events. This requires evaluation of earthquake parameters used in the Ina-TEWS as a result of inaccuracy in tsunami assessment. This inaccuracy led to false warnings either tsunami alerts were issued as rapid responses to earthquakes of adequately large magnitudes or likewise occurrences with relatively small magnitudes were left to generate tsunami hence a series of destruction on its way to beaches and lands with no warnings. In regard to all of these false warnings, quick action after the main shocks is not the only important issue since accuracy in providing information is also significant to earthquake-tsunami assessment.
Vol. 38, No. 4, page 171 (2019)
TdurT65ex Td
To reduce the level of inaccuracy in false warnings formerly issued by BMKG for tsunami events in Indonesian regions, Madlazim et al. (2015) have then utilized numerical codes called Joko Tingkir for a better prediction of tsunami wave generation that may occur following ocean bottom ruptures by future events. These codes have incorporated the importance of geometry of a fault in parameterizing tsunami potential into play, which includes the length and width of a rupture. Since the length of a rupture is technically difficult to measure, " that is proportional to the length of a rupture is then used for a discriminant. Together with this parameter, " and " (determined from seismograms), are likely being used for a set of earthquake-tsunami discriminants for effective tsunami warnings. In addition to tsunami importance reported by Lomax and Michelini (2011), for events in Indonesian regions, Madlazim et al. (2015) argued that the discriminants are as follows, " ≥ 65 s, " ≥ 10 s,
≥ 1, × ≥ 650 s2 and × ≥ 10 s. If at least three of these inequalities are satisfied then a tsunami wave is possible and hence to issue tsunami alert.
2. METHOD FOR BETA TESTING
While being integrated into the existing monitoring system to support the Ina-TEWS performance with a direct supervision from BMKG, a programming package named Joko Tingkir application was validated using tsunami importance (Madlazim et al. 2015) and then tested using alpha testing during a period of 2014-2015 for early detection of possible errors. The results showed that the application has fulfilled requirement in an adequate level and confirmed that further improvement of performance for an increased accuracy hence perfectness is required. This is an essence of the current study, where beta testing of the Ina-TEWA is presented.
In the process of improvement in terms of accuracy, Madlazim et al. (2018) developed a method called Mfilter used to reduce seismic signals from unexpected sources of dominant noises. Although this method was not directly specified to increase the accuracy, it helped the Ina-TEWS performance better in that automatic computation of P-waves run faster, making the calculation process efficient (see for further use of automation in Kirubakaran and Karthikeyani, 2013). Therefore, beta testing was performed during the progress of this study whereas such a direct method of computation was also implemented in the automated system. For completeness, beta testing utilized a total of 35 earthquake events throughout Indonesian territories and its surrounding regions for the last two-years occurrences having magnitudes of 6.0 or greater and sources of 10 to 600 km deep beneath the surface, as depicted in Figure 1 below, where the majority of the events were located in the Indonesian eastern provinces and some of these occurred in Philippine and Papua New Guinea (PNG). Real-time measurements of the 5 earthquake-tsunami discriminants were carried out during the test using methodology detailed in previous work of Madlazim et al. (2015).
Vol. 38, No. 4, page 172 (2019)
TdurT65ex Td
Tdur TdT65ex Tdur Td Td T65ex
Before going further into the next section, here we provide the saline points of what is referred to as tsunami false warning. By definition, a false warning used in this study means that the Ina-TEWA suggests TSUNAMI POTENTIAL for an event of a particular magnitude, either hypocentered below the ground or the sea surface at a great depth, from which a tsunami alert may then be publicly issued by the relevant Indonesian authorities although field observations confirm no tsunami generated or the application says NO TSUNAMI POTENTIAL even though monitoring instruments offshore and onshore confirm tsunami generation. It is relatively easy to understand that the two possible ways of false warnings have to be avoided for putting people living and all properties nearby local beaches in danger. Even in some cases, places far away from the source remains vulnerable to tsunami hazard as the wave may travel a great distance across the ocean (Prastowo et al., 2018).
"
Figure 1. Distribution of epicenter of 35 earthquakes with magnitudes 6.0 or greater (marked orange circles) and varying depths from shallow to deep sources examined in this study during a time period
of January 1, 2018 till July 1, 2019.
3. RESULTS AND DISCUSSIONS
The Indonesian tsunami early warning application project has recently been extended into three versions that are managed by BMKG to date. The first version of the application was published in June 2013, followed by the second implementation in March 2016, and the third in November 2017. For each, a comprehensive computer testing was performed, as required. When the application is publicly used, there is no a turning point. Hence, this is why the process of internal-driven testing is of importance. For this reason, a group of people called an internal team responsible for beta testing (Buskey, 2005) has examined daily records on the use of the application, subsequently reviewing them all in detail. All errors, shortcomings, inconsistencies and feature requests were recorded by a Flyspray system, freely accessible at http://flyspray.org/ popularly known as a web-based bug tracking system, written in the form of a hypertext preprocessor for software development. The system allowed users to record few characteristics of the application, including a reported version, a
place of testing, a severity of test results, a priority of usability problems, and examiner names.
The group contains researchers utilizing this system for beta testing routinely once within a 5-days period of time. An external group named the Centre for Research and Development BMKG as the official authority responsible for beta testing obtains a weekly report. This group has examined all detected defects due to any imperfect design of the application. Before product release or launch for implementation in a real condition, it is required for the two groups to report all the problems found in a manner different from ways of guidelines and reporting engineering case studies recommended by Runeson and Höst (2008).
As mentioned, an internal team for beta testing plays a key role in the software development. Examiners within the team carry out a number of automatic computer-based tests as well as manuals on the devices and for earthquakes occurring throughout Indonesian territories and its surroundings. The primary purpose of this step is to identify design defects as many as possible to ensure tsunami early warning application working with high performance hence high accuracy, as expected. In order to make it real, a number of people within the division of human resources capable of doing the job on demand are needed. These people have examined the application in details within a framework of previous studies (Nielsen and Landauer, 1993; Nielsen, 2000; Jiang et al., 2017), who suggested that the best result of beta testing for optimum performance is obtained from only 5 internal users.
Adopting the formula for the number of usability problems found in a particular test (Nielsen and Landauer, 1993) and applying it in the context of tsunami assessment, the number of false warnings " given by the application under examination in beta testing with " users is calculated from
" (1)
where " is the total number of the problems detected in the software design and " is the proportion of design defects found when a single user is on work. A typical value of " is of 31% (Nielsen, 2000), on average, for a large number of projects under consideration.
Using the above value for " , the following results are then presented and subsequently discussed. The surprising result from Eq. (1) is that zero users with " give zero false warnings with " , as expected when there is no implementation. When the data were collected from a single test user with " , false warning rapidly increased to approximately one third of the total problems detected in the design. The large difference between a case of zero users and a few was deliberately shocking. When the second user was then in action (with " ), we found the same things being performed as the first single test user did, making an overlap in the learning process between the first two-users. However, there will always be something different from the two. For example, if the second user does a thing then the first user may not do it. Within this context, the second user provides new insight but
Vol. 38, No. 4, page 174 (2019)
Xn
X = N{1 − (1 − L)n}N L
L
Ln = 0 X = 0
n = 1
n = 2
it is not as much a small amount of new data, relatively compared with the previous two users. Keep it in mind, more and more users involved, then less and less lessons learned because the same things occur as the first user does. The third user (with " ) would then do many things compared with the first or the second user and even things observed twice, and again, would generate repeatedly. It is not necessary to keep the same things observed multiple times, and there exists enough reasons for being back to the drawing board and redesign the application to eliminate the usability problems. In short, by adding more users fewer false warnings are found because the same situation is repeated. After the fifth user, it is a time waste to observe the same problems or common findings repeatedly but not providing better insight much (Nielsen and Landauer 1993; Nielsen 2000). In other words, optimum performance of the test is achieved at a value of " (meaning that 5 users are involved) permitting us to set up the equal number of examiners for each group of internal and external origin during the test. Project on beta testing for the Ina-TEWA was carried out following such descriptions, where each group of examiners or testers was filled with 5 examiners or testers for the best results of optimum performance. Before testing was performed, the examiners or testers were knowledgeable by relevant instructions about which function of the application required to be tested. Each examiner or tester performed a number of tests, once every 5 days, where the Flyspray is the main logging tool. This device recorded a single design defect detected by an examiner or a tester during the test.
To facilitate data analysis, we consider the records collected in a week before the latest version (the third one) of the Ina-TEWA was launched. False warnings were classified into whether they were related to a level of accuracy in the application or a technical problem. All examiners or testers found 7 of 35 earthquakes examined in this study were falsely issued (5 problems of accuracy and 2 cases of inconsistencies). To validate this, we focus on the first 5 examiners or testers from the internal group and present the results in Table 1 below.
In consecutive, tester A found 7 errors in tsunami alert with no new false warnings reported. The same situation for tester B with zero new false reports issued, followed by tester D and E having the same experiences. Different from the others, tester C observed 6 tsunami alerts mistaken and a new error was recorded. We then move on the second 5 examiners or testers from the external group and present the results in a different panel below. A clearly similar performance is shown by the two groups for the same total number of events examined in this study, where the two have found 7 cases of falsely issued warning. It follows that 7 of 35 cases considered is categorized as false in tsunami alert reported. There remains challenging for future use of beta testing whether optimum performance using 5 testers for each examiner, as is the case in the present study, being efficient and consistent with more cases of earthquakes (Jiang et al. 2017).
Vol. 38, No. 4, page 175 (2019)
n = 3
n = 5
Table 1. False warnings reported by 5 testers from internal (FGHIJ) and external (ABCDE) groups of researchers responsible for beta testing using methodology of optimum performance.
The occurrence of 20% false warnings found using the 2017 Ina-TEWA version remains relatively high. This is due to only the dominant period " of the first P-wave arrivals as the main parameter for tsunami detection being filtered using techniques described in Madlazim et al. (2018). The other two crucial parameters, " and " , have not yet been directly calculated from filtered seismic signals, making them remains consisting of unexpected noises and thereby reducing the level of accuracy.
5. CONCLUSIONS
Beta testing for improved performance of the Ina-TEWA in terms of accuracy has been carried out using a total of 35 relatively large earthquakes of varying source depths in Indonesian and its nearby regions, particularly in the Indonesian eastern provinces where seismic activities have been found to
Vol. 38, No. 4, page 176 (2019)
Tester False WarningNew False Warning
A 7 0
B 7 0
C 6 1
D 7 0
E 7 0
Tester False Warning New False Warning
F 7 0
G 7 0
H 6 1
I 7 0
J 6 1
Td
Tdur T65ex
increase lately. These events are all considered tsunamigenic considering their scales in magnitude. After a series of stages during the test, we analyze all the results and conclude that this test is proved to increase the level of accuracy and hence to improve the Ina-TEWA performance by 80% since only 7 tsunami warnings are falsely issued. However, further refinement of the existing tsunami warning application remains challenging for future work. In the light of perfectness within the uncertainty or possible errors permitted in the process of computation, upcoming research on this subject may then also install M-filter used for the other two main tsunami parameters ( � and � ).
Acknowledgements
The authors would sincerely like to thank anonymous reviewer(s) for their best suggestions and invaluable comments upon this manuscript to be appropriate for publication in STH. Great thanks go to the Center for Research and Development BMKG, Indonesia, the authority with which all the data used are obtained and from which permission to run the application is given for this study.
References
Budnik, C. (2012), Software testing, software quality and trust in softwarebased systems, Proc. 36th Annual IEEE Computer Software and Applications Conference (COMPSAC), 253, Izmir, Turkey: IEEE, doi:10.1109/COMPSAC.2012.107.
Buskey, C. D. (2005), A software metrics based approach to enterprise software beta testing design, doctoral dissertation, 1-268, unpublished work, New York, US: Pace University Press, ISBN:0-542-34378-9.
Jiang, Z., K. P. Scheibe, S. Nilakanta and X. Qu (2017), The economics of public beta testing, Deci. Sci. doi:10.1111/deci.12221.
Kirubakaran, B. and V. Karthikeyani (2013), Mobile application testing: challenges and solution approach through automation, Proc. International Conference on Pattern Recognition, Informatics and Mobile Engineering (PRIME), 79-84, Periyar University, Tamilnadu, India: IEEE, doi:10.1109/ICPRIME.2013.6496451.
Kocbek, M. and M. Heričko (2013), Beta testing of a mobile application: a case study, the Second Workshop on Software Quality Analysis, Monitoring, Improvement, and Applications (SQAMIA), Novi Sad: Department of Mathematics and Informatics, University of Novi Sad, Novi Sad, Serbia.
Lomax, A. and A. Michelini (2009), Tsunami early warning using earthquake rupture duration, Geophys. Res. Lett., 36, L09306, 1-4, doi:10.1029/2009GL037223.
Lomax, A. and A. Michelini (2011), Tsunami early warning using earthquake rupture duration and P-wave dominant period: the importance of length and depth of faulting. Geophys. J. Int., 185(1), 283-291, doi:10.1111/j.1365-246X.2010.04916.x.
Lomax, A. and A. Michelini (2012), Tsunami early warning within five minutes. Pure Appl. Geophys., doi:10.1007/s00024-012-0512-6.
Madlazim, T. Prastowo and T. Hardy (2015), Validation of Joko Tingkir software using tsunami importance. Sci. Tsu. Haz., 34(3), 189-198.
Madlazim and T. Prastowo (2016), Evaluation of earthquake parameters used in the Indonesian Tsunami Early Warning System, Earthq. Sci., 29(1), 27-33, doi:10.1007/s11589-016-0143-6.
Madlazim, T. Prastowo, S. Rohadi and T. Hardy (2018), Filter-M application for automatic computation of P wave dominant periods for tsunami early warning. Sci. Tsu. Haz., 37(1), 26-32.
Mohd, CKNCK and F. Shahbodin (2015), Personalized learning environment: alpha testing, beta testing and user acceptance test, Proc.Soc. Behavioral Sci., 195, 837-843, doi:10.1016/j.sbspro.2015.06.319.
Nielsen, J. (2000), Why you only need to test with 5 users. retrieved on September 2, 2013 from http://www.nngroup.com/articles/whyyouonlyneedtotestwith5users/.
Nielsen, J. and T. K. Landauer (1993), A mathematical model of the finding of usability problems. Proc. ACM INTERCHI'93 Conference on Human Factors in Computing Systems, 206-213, Amsterdam, The Netherlands: CM DL, doi:10.1145/169059.169166.
Pan, J. (1999), Software testing, retrieved on September 2, 2013 from http://www.ece.cmu.edu/~koopman/des_s99/sw_testing/
Pradhan, T. (2012), All types of software testing, retrieved on September 2, 2013 from http://www.softwaretestingsoftware.com/all typesofsoftwaretesting/
Prastowo, T., L. Cholifah and Madlazim (2018), Analysis of travel time delay for large tsunamis across the Pacific and Indian Oceans, Sci. Tsu. Haz., 37(4), 195-212.
Runeson, P. and M. Höst (2008), Guidelines for conducting and reporting case study research in software engineering. Empir. Software Eng., 14(2), 131-164, doi:10.1007/s10664-008-9102-8.
Satriano, C., Y. M. Wu, A. Zollo and H. Kanamori (2011), Earthquake early warning: concepts, methods, and physical grounds, Soil Dyn. Earthq. Eng., 31(2), 106-118, doi:10.1016/j.soildyn.2010.07.007.
Suppasri, A., K. Goto, A. Muhari, P. Ranasinghe, M. Riyaz, M. Affan, E. Mas, M. Yasuda and F. Imamura (2015), A decade after the 2004 Indian Ocean tsunami: the progress in disaster preparedness and future challenges in Indonesia, Sri Lanka, Thailand, the Maldives, Pure Appl. Geophys., 172, 3313-3341, doi:10.1007/s00024-015-1134-6.
Indonesia is an archipelagic country prone to natural disasters, especially tsunami as the one that hit the Aceh Province in 2004 - which caused over 80,000 deaths and 125.000 injuries including elementary school children. A school preparedness survey in Indonesia indicated that among school-age children aged 7-12 years, their knowledge and preparedness in disaster emergencies was low. Consequently, dealing with disaster management, the government of Indonesia became responsible in determining the best strategies in building up preparedness. Given the existing problems, the present study aimed at developing and evaluating the effectiveness of the Game Model of Knowledge and Physical Fitness among Elementary School Children in Tsunami Disaster Anticipation for six tsunami-affected schools in Banten and W. Sumatra Provinces. The intervention study applied qualitative and quantitative approaches to develop and evaluate the effectiveness of the game model. Qualitative data was obtained from observations, interviews and secondary data documentation. The quantitative data obtained from the sample of 240 students was analyzed in accordance with the study objective. SPSS trial version 17 was used for the analysis. The results of paired and independent t-test on knowledge and fitness of the students showed that the model was effective. The effectiveness of the model in terms of improved scores was better among the intervention group than the control (p < 0.05), even though significant
improvement was also observed among the control. In conclusion, the policy implication suggested that the model may be applied and disseminated in other tsunami vulnerable areas in other provinces of Indonesia.
Keywords: Tsunami disaster management, effectiveness, game model, elementary school students.
1. INTRODUCTION
Indonesia is an archipelagic country prone to natural disaster, especially tsunamis generated by volcanic eruptions and earthquakes, resulting by collisions of three active tectonic plates which surround Indonesia. These are the Euro-Asian, the Indo-Australian, as well as the Philippine and the Pacific Ocean plates in the East, north, and south. Therefore Indonesia has high vulnerability to natural disasters such as volcanic eruptions, earthquakes and tsunamis (Anam et al. 2018). Data from the 2015-2019 National Disaster Management Agency showed that Indonesia was hit several times by tsunamis from 1818 to 2014 (Renas, 2014). In 2004, the Aceh province was hit by the biggest tsunami ever recorded, which resulted in over 80,000 deaths and 125.00 injuries, which including elementary school children. It was the fifth largest disaster in world history, after the 1964 Alaskan earthquake and tsunami (Fitriawan, 2017). In accordance with the Indonesian Law Number 24 of 2007 concerning Disaster Management, the government is responsible for preparing the best strategies related to disaster relief efforts in the country. In building preparedness, knowledge and understanding of disaster phenomena in general, are very important for the formal education of students. Consequently, after the 2004 earthquake and tsunami, in order to reduce the risk of tsunami disasters in Indonesia the government formed a Tsunami Warning System under the name of “Indonesian Tsunami Early Warning System” (Anam et al. 2018). It was obvious that tsunamis and natural disasters in general, have posed a social burden not only on the government but also on the communities.
The need to create a model that involves a learning process to enhance knowledge of tsunami disaster anticipation and of improving physical fitness of elementary school students has been supported by results of various studies. Research evidence from Malaysia showed that learning with new knowledge must include cultivating the value of a positive attitude, without leaving physical activity in order to stay fit, and able to continue the learning process (Horikawa and Shuto 2010). Results of research conducted by Stough and Sungyoon (Stough and Sungyoon 2018) stressed the importance of safety practices facing global disasters in schools for children by providing knowledge of emergency natural disaster procedures and not accompanied by actual action. A study by Hidayati (2016) proves that pusijump (Puzzle, Music and Magic) learning methods are effectively used in earthquake and tsunami disaster preparedness learning intended for mild mentally retarded students, While Burhany and Tendy (2016) found that discussing risk reduction in tsunami disasters is more suitable to be carried out by debriefing knowledge about anticipating the tsunami disaster. They stated that to reduce the number of fatalities due to natural disasters, one of the efforts is to provide knowledge in the form of information and of anticipated needed actions.
Vol. 38, No. 3, page 180 (2019)
Students or school-age children are the most vulnerable population, and yet play an important role in disaster preparedness, both at home and school (Anam et al. 2018). Children must have knowledge about human behavior and the risks caused by natural disasters, especially tsunamis (Suppasri et al. 2016), (Seneviratne, Amaratunga and Pathirage 2010). However, only a small number of students understand the preparedness that is needed for natural disasters, particularly of tsunamis (Kongko and Hidayat 2014). Survey data on school preparedness in facing disasters in Indonesia is lacking. Of 55 school children between 7 to 12 years old, 20% recognized the existence of the natural disaster early warning system, 13.3% of them had the ability to know emergency preparedness plans and 10.6% had skills in natural disaster emergency alert simulation (Syarif and Irawan 2015).
In addition, physical activity of children with appropriate training will benefit their cognitive abilities as well as their academic performance. Thus, participation of children in disaster simulation exercises not only involve cognitive intelligence, and are enhanced by their physical fitness. In brief, skills in tsunami alert emergency simulations involve cognitive knowledge, and physical fitness, both being prerequisites for success of the simulation game. The physical fitness component consists of various activities such as running, jumping, crawling, jogging, squatting, push ups, sit ups, and other motor movements aimed at increasing physical fitness.
The government provides students with the knowledge and skills through a formal educational syllabus, aimed at improving cognitive abilities as well as skills in simulating emergency anticipation of a tsunami disaster. The school curriculum for physical education aimed at achieving graduates with competence, i.e., the core competence (KIs) consisting of four competencies. KI-1 on faith and religion, KI-2 on character and personality, KI-3 on knowledge and KI-4 on skills. In accord with the 2013 Curriculum for Physical Education for Elementary School year 9, namely, KD 4.1 to 4.3, it was explained that students practiced a combination of locomotor and non-locomotor basic motion patterns based on the concept of motion in various simple forms of play (Kemendikbud 2013). Simple games in this case can be realized with a variety of fun activities without a coercive voluntary concept. The concept of the game is suitable for use in learning for elementary school age children (Singer 2015). All is required in the anticipated tsunami disaster activities.
Up to now there was no research in the development of tsunami disaster anticipation methods for primary school-aged children in tsunami-affected prone areas, as well as testing its effectiveness in improving cognitive abilities along with increasing physical fitness levels simultaneously. Therefore, the aim of the present study was to develop a tsunami disaster anticipation model for learning physical education and tested for its effectiveness in improving cognitive abilities and physical fitness among elementary school students in the tsunami prone areas. The outcome of this development study is a game model of tsunami disaster anticipation for elementary school children in the vulnerable areas of tsunami disaster.
Vol. 38, No. 3, page 181 (2019)
2. METHODS
The development study uses qualitative and quantitative approaches. The design for a model development adopted the ADDIE (Analysis, Design, Development, Implement and Evaluation) development step (Branch 2009) shown in Figure 1.
Figure 1. Steps to develop a model (Branch, 2009)
The study has been approved by the Ethical Committee of the Jakarta State University number 081/RE-PORS2/EC/III/2019 dated 6 March 2019. On behalf of the students, informed consent was given by their teacher after they understood the explanation given by the researchers. The simple random sampling was applied to obtain a sample from a population of elementary school children in 6 tsunami-affected schools in Banten and West Sumatra Provinces. There were 240 students randomly selected and consisted of 120 students assigned as the intervention group and 120 students was assigned as the comparison group. The intervention study design with two arms was implemented, in which the intervention group was compared with the control group. The intervention group received a developed model consisting of educational and learning module plus aerobic exercise, while the control group did not received development module but followed a conventional educational module given at the school. The aerobic exercise given was in aCcord with the 2013 Curriculum for physical education subjects in grade 3 elementary school, namely KD 4.1 to KD 4.3, whereas the knowledge-based educational and learning module was in addition to the existing standard module to achieve core competencies, namely KI-1 to KI-4. The KI-1 covers faith and religion, KI-2 covers character and personality, KI-3 covers knowledge and KI-4 covers skills.
Prior to the intervention, a qualitative approach was implemented and aimed to develop a model of game anticipation of a tsunami disaster. During the development phase, having done the initial needs analysis, the design of a model was developed and validated by a national natural disaster management expert, two regional disaster management experts, one sports sociologist and one physical education expert and sports game evaluation expert. In the developed model there are thirty items to be delivered to the intervention group, in which the items were based on the storyboard developed previously. The design consists of the items and objectives seen in Annex 1. As indicated elsewhere (Palar et al. 2015), a physical education learning process was an important element of the model and that the purpose of an exercise was to improve students' physical fitness by conducting one week of 2 to 3 meetings with an aerobic exercise duration of 60-90 minutes.
Vol. 38, No. 3, page 182 (2019)
Annex 1. Objectives and 30 modules’ item in the model of simulation game in anticipation of a tsunami disaster.
Objective 1: Understanding Preparedness and Anticipation Simulation
Objective 2: Understanding the simulation of tsunami early warning system activities
Objective 3: Understanding and translating BPBD information related to anticipating a tsunami disaster
1/. Preparedness simulation in anticipation of a tsunami disaster
2/. Simulation looking for and heading to a safe place.
3/. Simulation of tsunami disaster preparedness with compact and
4/. Eliminate feelings of panic, anxiety and worry
5/. Simulation actions improve life safety during
6/. Simulation of the tsunami early warning system according to tsunami disaster procedures
7/. Simulation of alertness and dexterity in the activities of the tsunami early warning system
8/. Learn and understand the BPBD information and
9/. Implement the instructions of the Regional Disaster Management
10/. Learn the BPBD information and instructions
11/. Understand the safe evacuation routes
12/. Finding someone else's safe evacuation
13/. Make a tsunami evacuation route
14/. Understanding of foreign words in anticipation of a
15/. Mention of foreign words in anticipation of the tsunami disaster
16/. Conduct activities from foreign words in anticipation of the
17/.Simple action simulation after the tsunami disaster
18/. Simulations avoid puddles, hazardous substances after the
Objective 7: The game that is applied contains the concept
Objective 8: Improving Physical Fitness component
The quantitative approach was implemented after the game model has been completed. The quantitative approach in terms of an intervention study was aimed at testing the model effectiveness through two outcome measures, namely, students’ knowledge and physical fitness before and after the intervention for about one month. The outcome was compared among the two groups at the end of the intervention (post and post test), as well as before and after (pre and post test) for each group. To test the effectiveness of the model a paired t-test for both group during pre- and post-intervention period amongst intervention group on both students’ knowledge and physical fitness, was used. In testing the physical fitness, the Indonesian Physical Fitness Test (TKJI) aged 10-12 years for Indonesian child standards was used, in which a composite standard score was used as a reference (see Table 1).
Table 1. Score of physical fitness in accord with TKJI (Indonesian Physical Fitness Test)
To test the knowledge score of the elementary school children, a questionnaire consisting of 18 questions in the model was evaluated to produce a composite score. The instrument was developed based on the indicators of tsunami early warning system and outdoor activity.The score was obtained from the test of knowledge on anticipation of tsunami disaster in the 18 questions with
multiple choice answers. The correct answer was scored 1, whereas wrong answer was scored 0. For each student the score range was from 0 to 18, and for each of either the intervention or control group, the score in percentage was obtained which was equal to [(correct answer/18) x 100].
3. RESULTS
As shown in Annex 1, out of 30 items that have been developed, as many as 28 items were implemented and another 2 were aborted due to limited infrastructure for implementation, and for inability of replacement by other games. The outcome of the model effectiveness was measured using both a composite knowledge score and a composite physical fitness score before and after intervention to intervention group only. To do the analysis, SPSS v.22 was used. Characteristics of the sample are shown in Table 2.
Table 2. Characteristics of the sample
Before a paired t-test was conducted, normality test of Kolmogorov-Smirnov was performed before and after the intervention for both data on knowledge and physical fitness. All data showed a normal distribution, which made feasible the paired t-test being being conducted. Results of descriptive statistics and paired t-test on knowledge and physical fitness test are shown in Table 2 and Table 3, respectively as follows.
Vol. 38, No. 3, page 185 (2019)
Characteristics
LocationTotal
Banten West Sumatra
n % n % n % row
Age 10 year 55 22,917 65 27,083 120 50
11 year 61 25,417 59 24,583 120 50
Gender Boy 49 20,417 71 29,583 120 50
Girl 62 25,833 58 24,167 120 50
K n o w l e d g e Score
High 29 12,083 30 12,500 59 24.6
Medium 47 19,583 67 27,917 114 47.5
Low 32 13,333 35 14,583 67 27.9
P h y s i c a l Fitness Score
Good 17 7,083 12 5,000 29 12.1
Medium 97 40,417 82 34,167 179 74.6
Poor 21 8,750 11 4,583 32 13.3
Group Treatment 55 45,833 65 54,167 120 50
Control 65 54,167 55 45,833 120 50
Table 3. Mean of knowledge and fitness score before (pre-) and after (post-) intervention among intervention and control students
KP : knowledge score among control group; IP : Knowledge score among trial group KF : fitness score among control group; IF : Fitness score among trial group
It can be seen in Table 2 that the mean results of the pre-test and post-test knowledge of the anticipation of the tsunami disaster were 38,658 and 78,188. Meanwhile, as seen from Table 2 the average physical fitness composite score during pre- and post-intervention among the intervention group showed scores of 8.86 and 17.18, respectively. Merely looking at the scores classification as shown previously in Table 1, the intervention model resulted in a significant improvement from ‘poor physical fitness’(range 5 – 9) to ‘moderate physical fitness’(range 14 – 17) .
The results of paired t-test of knowledge and fitness score before and after intervention among intervention and control groups are shown in Table 3. On the knowledge score, the mean difference between pre and post intervention was 39.53 (37.11, 41.94) with statistic t-count(df=119) = 32,368 and p-value = 0,00 (or <0.05), indicating that there was a 95% significance level difference in the knowledge score of anticipating the tsunami disaster before and after being given the treatment of the model game. It is therefore concluded that the game model is effective and can increase the knowledge of tsunami disaster anticipation among elementary school children.
Meanwhile, on the physical fitness score the mean difference between pre and post test was 8,317 (7,843-8.791) with a t-count(df=119) = 34,749 and p-value = 0.00 (or <0.05) with significancy level of 95%. Hence, it is concluded that the game model in the form of exercise is effective and can increase significantly the physical fitness of tsunami disaster anticipation among intervened elementary school children through a physical education learning module significantly.
The next study questions to be answered were : 1). What was the results among the control group?; and 2). Given the results of post-test on knowledge test and physical physical fitness test among the intervention and control group, was the effectiveness of module given to the intervention group better than the control?
Vol. 38, No. 3, page 186 (2019)
Group Module Pre & Post N Mean Std.
DeviationStd. Error
Mean
Intervention Pair 1 (Knowledge)
pre-IP 120 38,658 9,7079 ,8862
post-IP 120 78,188 9,8206 ,8965
Pair 2 (Fitness)
pre-IF 120 8,86 1,920 ,175
post-IF 120 17,18 1,986 ,181
Control Pair 3 (Knowledge)
pre-KP 120 40,878 9,5884 ,8753
post-KP 120 53,962 9,3413 ,8527
Pair 4 (Fitness)
pre-KF 120 8,95 1,454 ,133
post-KF 120 11,17 1,687 ,154
Looking back at the Table 2, it can be seen that among the control group the mean of the pretest and post-test knowledge of the anticipation of the tsunami disaster were 40,878 and 53,962 respectively. The average mean difference indicated that among the control students the knowledge on anticipation of tsunami disaster has increased significantly (p <0.05), as seen in Table 3.The mean difference between pre and post intervention showed that was 13,083 (11,57, 14.59) with statistic t-count(df=119) = 17,175 and p-value = 0,00 (or <0.05), indicating that there was a 95% significance level difference in the knowledge of anticipating the tsunami disaster before and after being given the treatment of the model game. It is therefore concluded that the game model is also effective and can increase the knowledge of tsunami disaster anticipation among control children. Compared with the intervention group, it was noted the improvement in the control group was worse than the intervention group, namely, 8,3 and 39,5 respectively.
Meanwhile, as seen from Table 2 the average physical fitness composite score during pre- and post-intervention among the intervention group showed scores of 8,95 and 11,17, respectively. Referring to the scores classification as shown previously in Table 1, the intervention model resulted in a slight improvement from ‘poor physical fitness’(range 5 – 9) to ‘deficient physical fitness’(range10 – 13) . The improvement was worse than that of the intervention group.
Overall, it is concluded that the applied exercise model can significantly improve the fitness level of primary school children through a physical education learning module within the game model. It also proves that the tsunami disaster anticipation among control students on physical education learning was also an effective effort to improve the physical fitness of control elementary school students who did not receive the module.
From Table 2 and Table 3, paired t-test indicated significant differences on both trial and control group on the knowledge and physical fitness score before and after intervention (p < 0.05). From these tables it was indicated that the increased score among trial group was higher than the control group.The knowledge of score difference after the trial period among intervention group was 50,56% as compared with 24.26% among the control group, indicating that the impact of the module given to the intervention group was greater than that of the control group. Similarly, the impact of the physical fitness module given to the intervention group was bigger than that of the control group, by as much as 49.63 points versus 19,87 point respectively. Given the above data, the hypothesis formulated that there was a significant larger increase on both knowledge and physical fitness score after the trial period among the intervention group than the that of the control group. To test the hypothesis the paired t-test was conducted. The results are presented in Table 4 and 5.
Vol. 38, No. 3, page 187 (2019)
Table 4. Significancy level of pre- and posttest on knowledge and fitness score before and after intervention among intervention and control students
KP : knowledge score among control group; IP : Knowledge score among trial group KF : fitness score among control group; IF : Fitness score among trial group.
Table 5. Summary of mean difference between trial and control student group on the increased score of knowledge and physical fitness
KP : knowledge score among control group; IP : Knowledge score among trial group KF : fitness score among control group; IF : Fitness score among trial group
Table 6. Results of paired t-test of Posttest of knowledge score and physical fitness score among Intervention group versus control group
KP : knowledge score among control group; IP : Knowledge score among trial group KF : fitness score among control group; IF : Fitness score among trial group
It is evident that the intervention group on anticipation of tsunami disaster compared well with the control who did not received the intervention module. The null hypothesis that there was no significant differences between intervention and control group at the end of the trial was rejected. Therefore, it was concluded that there was a significant bigger increase on both knowledge and physical fitness score after the trial period among the intervention group than the control group. It proves that the module of anticipation of tsunami disaster among school children was effective.
4. DISCUSSION
Based on the evidence obtained from this study, we were convinced that the developed game model was effective in significantly increases both the knowledge of tsunami disaster anticipation and physical fitness among elementary school children aged 10 to 12 years in Banten and West Sumatra Provinces. The magnitude of the improvement was significantly larger among the intervention group than the control. In other words, the game model developed has been proved effective in improving students' cognitive understanding of the anticipation of the tsunami disaster.
The results of this study was supported by other studies written by Bluma and Malgorzata (2018) which concluded that the impact of sports on pre-adolescents' cognitive functions was enhanced by sport activities. They also stated that sports conducted regularly on pre-adolescent children can have an impact on the development of cognitive functions of children. This can be explained by the fact that during playing, the children feel happy and enjoy beyond their conscious threshold. When the children are happy, the brain will issue reciprocity in the form of positive energy that is able to record information more quickly from outside to be recorded, processed and stored in the brain's memory in the long run (Al-Thaqib et al. 2018). Results of this study of model development infers that information obtained by individuals who carry out game-based activities
Vol. 38, No. 3, page 189 (2019)
Module ModulePost
& Post
Paired Differences
of Mean
Std. Deviation
Std. Error Mean
95% CI of the Difference t df
Sig. 2-tailed
Lower Upper
Knowledge
Pair 1 (Interventio
n vs Control)
post-IP VS
post-KP
24,22 12,47 1,138 21,97 26,48 21,27 119 ,000
Fitness
Pair 2 (Interventio
n vs Control)
post-IF VS
post-KF
6,00 2,68 ,245 5,52 6,49 24,52 119 ,000
will affect the long-term memory and other cognitive achievements much better than the individuals who obtained the information in a conventional manner.
Referring to an observed significant increase in physical fitness among elementary school students, the findings was also supported by another study conducted by Gabbett and Jenkis (2009) who stated that athletes who practice accurate game concepts can also improve physical fitness. The explanation was quoted as "collectively, the findings demonstrate the value of game-based training for sports team athletic skills and physical fitness"
Other similar research also revealed that there are several types of physical activity that can reduce adipose fat levels when the children do certain type of activity involving playing and game concept (Grace and David 2013), such as endurance and aerobic activities, sports-based games, sports training, and active play. Their findings explained that game activities in sports were able to improve skills, cognitive abilities, and the fitness of individuals who do it proportionally.
5. CONCLUSIONS
In summary, it is concluded: 1) that the game model is able to increase the knowledge of Tsunami disaster anticipation and improve the level of physical fitness of elementary school children aged 10 to 12 years old in a learning process; 2) that the impact of the game model was bigger among intervention group than the control in terms of knowledge score and physical fitness; 3) that the tsunami disaster anticipation game model developed in the form of 30 types of games in physical education learning in the form of books with an explanation of the tsunami disaster anticipation consisting goals, rules and how to play game systematically, can be applied and be effective; 4). The results suggest that the model can be applied and be disseminated in other tsunami vulnerable areas in other provinces along the three tectonic plates surrounding Indonesia.
ACKNOWLEDGEMENTS
We highly appreciate all members of the staff of the Faculty of Sports Science of the Jakarta State University who contributed to this study from the start to the end. We also appreciate the assistance of all the School Heads of the Six Elementary Schools in Banten and West Sumatra and their children for their willingness to participate in this study.
DISCLAIMER: All the views expressed in this article are our own and not an official position of the Jakarta State University.
Vol. 38, No. 3, page 190 (2019)
REFERENCES
Al-Thaqib, Abdulrahman., Fahad M., Abdullah J.K., Al-Zahrani, Shahid, Basid. (2018). Brain Training Games Enhance Cognitive Function in Healthy Subjects. Medical Science Monitor Basic Research. 3(11): 63-69. DOI: 10.12659/-MSMBR.909022.
Anam, Khoirul., Abdul., Aini., Andini., Febri., Befi. (2018). Kesiapan Institusi Lokal dalam Menghadapi Bencana Tsunami: Studi Kasus Kelurahan Air Manis dan Kelura-han Purus, Kota Padang. Jurnal Wilayah Dan Lingkungan, 6(1):15-29. http://dx.doi.org/10.14710-/jwl.6.1.15-29.
Bluma, Illona. B., Malgotzata L. 2018. Physical Activity and Cognitive Functioning of Children: A Syste-matic Review. International Journal of Environmental Research and Public Health, 15(4):33-41. DOI: 10.3390/ijerph15040800.
Branch, Robert Maribe. (2009). The ADDIE Approach : Instructional Design, New York-Springer, 2(-):1-199. DOI 10.1007/9780387-09506-6.
Burhany, Afif. and Tendy Ramadin. (2016). Antisipasi Bencana Alam dengan Memfasilitasi Sarana Pengetahuan di Indonesia. JG-Universitas Indonesia, 2(3):1-12.
Fitriawan, Rana Akbari. (2017). Jurnalisme Sains dan Sistem Peringatan Dini Bencana di Indonesia. Kajian Jurnalisme, 1(1):39-50. http://-jurnal.unpad.ac.id/kajian-jurnalisme
Gabbett, Tiem and David, Jenkis. (2009). Game-Based Training for Impro-ving Skill and Physical Fitness in Team Sport Athletes. International Journal of Sports Science & Coaching, 4(2):73-81. DOI: 10.-1260/174795409788549553.
Grace, O`Malley and David T. (2013). Physical Activity And Play In Children Who Are Obese. European Childhood Obesity Group : E-Book Eecg-obesity.eu.
Hidayati, Eri. (2016). Pendidikan Kesiapsiagaan Bencana Gempa Bumi Dan Tsunami Dengan Metode Play Therapymelalui Pusijump (Puzzle, Music Andmagic Jump) Untuk Siswa Tunagrahita. (Universitas Negeri Semarang). Retrieved from https://lib.unnes.ac-.id/27319/1/3201412048.pdf.
Horikawa, K.and Shuto, N. (2010). Tsunami Disaster and Protection Measures in Japan. Tsunamis-Their Science and Engineering, 3(5):9-22.
Kongko, W., and Hidayat, R. (2014). Earthquake - Tsunami in South Jogjakarta Indonesia : Potential , Simulation Models , and Related Mitigation Efforts, 2(3), 18–22.
Palar, Chrisly., Djon Wongkar, Shane H. R. (2015). Manfaat Olahraga Aerobik Terhadap Kebugaran Fisik Manu-sia. Jurnal E-Biomedik (eBM), 3(1):316-378.
Renas-PB. (2014). Rencana Nasional Penanggulangan Bencana 2015-2019. Badan Nasional Penang-gulangan Bencan.
Seneviratne, K., Amaratunga, D., Haigh, R., and Pathirage, C. (2010). Knowledge management for disaster resilience: identification of key success factors. International Journal of Disaster Risk Reduction, 8(2), 2-12.
Singer, Elly. (2015). Play and playfulness in early childhood education and care. Psychology in Russia: State of the Art, 8(2): 27-35. (doi: 10.11621-/pir.20-15.0203).
Stough, Laura and Sungyoon. (2018). Seven school-related disasters: Lessons for policymakers and school personnel. Education Policy Analysis Archives. 3(4):96-105. DOI:10.14507/epaa.26.3698.
Suppasri, A., Leelawat, N., Latcharote, P., Roeber, V., & Yamashita, K. (2016). International Journal of Disaster Risk Reduction The 2016 Fuku-shima earthquake and tsunami : Local tsunami behavior and recommendations for tsunami disaster risk reduction. Internati-onal Journal of Disaster Risk Reduction, 21(2017), 323–330. (https://doi.org/10.1016/j.ijdrr.2016.12.016).
Syarif, Andi and M.Irawan. (2015). Hubungan Self Efficacy Dengan Kesiapsiagaan Bencana Gempa Bu-mi Dan Tsunami Pada Siswa Sekolah Menengah Atas Negeri 2 Dan 6 Banda Aceh. Idea Nursing Journal, 6(2):53-62.
