1: \documentclass[a4paper,12pt]{article}

2: %\documentclass{svmult}

3: 

4: \usepackage[english]{babel}

5: \usepackage[latin1]{inputenc}

6: 

7: \usepackage{indentfirst}

8: \usepackage[dvips,final]{graphicx}

9: \usepackage{amsfonts,amsmath,amssymb,amsthm}

10: 

11: \usepackage{color}

12: \definecolor{oneblue}{rgb}{0,0.0,0.75}

13: \usepackage[colorlinks,

14:             urlcolor=oneblue,

15:             linkcolor=oneblue,

16:             citecolor=oneblue,

17:             bookmarksopen=false,

18:             %pdfpagemode=FullScreen,

19:             pagebackref]{hyperref}

20: 

21: \hypersetup{

22:     pdftitle = Water waves generated by a moving bottom,

23:     pdfauthor = Denys Dutykh and Frédéric Dias,

24:     pdfsubject = Linear water waves

25:     pdfkeywords = {linear water waves, Cauchy-Poisson problem, tsunami, waves generation}

26: }

27: 

28: %\makeindex

29: 

30: \usepackage[dvips,final]{graphicx}

31: %\usepackage[sort,sectionbib]{natbib}

32: \usepackage{xspace} % To get the right spacings in front of : and so on

33: 

34: \usepackage{fancyhdr}

35: \pagestyle{fancyplain}

36: % section number and title

37: \setlength{\headheight}{28pt}

38: \renewcommand{\sectionmark}[1]{\markright{\thesection\ #1}}

39: \lhead[\fancyplain{}{\bfseries\thepage}] %

40:     {\fancyplain{}{\bfseries\rightmark}}

41: \rhead[\fancyplain{}{\scshape\leftmark}] %

42:     {\fancyplain{}{\bfseries\thepage}}

43: \cfoot{}

44: 

45: \vfuzz2pt % Don't report over-full v-boxes if over-edge is small

46: \hfuzz2pt % Don't report over-full h-boxes if over-edge is small

47: 

48: \def\Dv{\mathbf{D}}

49: \def\kv{\mathbf{k}}

50: \def\uv{\mathbf{u}}

51: \def\qv{\mathbf{q}}

52: \def\nv{\mathbf{n}}

53: \def\R{\mathbb{R}}

54: 

55: \newcommand{\pd}[2]{\frac{\partial#1}{\partial#2}}

56: \newcommand{\od}[2]{\frac{d#1}{d#2}}

57: \newcommand{\tens}[1]{\mathrm{#1}}

58: 

59: \newcommand{\abs}[1]{\left|#1\right|}

60: 

61: \title{Water waves generated by a moving bottom}

62: \author{\href{http://www.cmla.ens-cachan.fr/\~dutykh}{Denys Dutykh%

63: \footnote{Centre de Mathématiques et de Leurs Applications, 

64: \'{E}cole Normale Supérieure de Cachan,

65: 61, avenue du Président Wilson, 

66: 94235 Cachan cedex, France}} \and

67: \href{http://www.cmla.ens-cachan.fr/\~dias}{Frédéric Dias\footnotemark[1]}}

68: \date{}

69: 

70: \begin{document}

71: 

72: \maketitle

73: \begin{abstract}

74: Tsunamis are often generated by a moving sea bottom. This paper deals with the case 

75: where the tsunami source is an earthquake. The linearized water-wave equations are 

76: solved analytically for various sea bottom motions. Numerical results based on 

77: the analytical solutions are shown for the free-surface profiles, the horizontal 

78: and vertical velocities as well as the bottom pressure. 

79: \end{abstract}

80: \tableofcontents

81: 

82: 

83: \section{Introduction}

84: Waves at the surface of a liquid can be generated by various mechanisms: wind blowing on the free surface,

85: wavemaker, moving disturbance on the bottom or the surface, or even inside the liquid, fall of an object 

86: into the liquid, liquid inside a moving container, etc. In this paper, we concentrate on

87: the case where the waves are created by a given motion of the bottom. One example is the generation of

88: tsunamis by a sudden seafloor deformation.

89: 

90: There are different natural phenomena that can lead to a tsunami. For example, one can mention submarine 

91: slumps, slides, volcanic explosions, etc. In this article we use a submarine faulting generation

92: mechanism as tsunami source. The resulting waves have some well-known features. For example, characteristic wavelengths 

93: are large and wave amplitudes are small compared with water depth. 

94: 

95: Two factors are usually necessary for an accurate modelling of tsunamis: information on the magnitude and distribution 

96: of the displacements caused by the earthquake, and a model of surface gravity waves generation resulting 

97: from this motion of the seafloor. Most studies of tsunami generation assume that the initial free-surface deformation 

98: is equal to the vertical displacement of the ocean bottom. The details of wave motion

99: are neglected during the time that the source operates. While this is often justified because the earthquake rupture

100: occurs very rapidly, there are some specific cases where the time scale of the bottom deformation may become an 

101: important factor. This was emphasized for example

102: by Trifunac and Todorovska \cite{todo}, who considered the generation of tsunamis by a slowly spreading uplift

103: of the seafloor and were able to explain some observations. During the 26 December 2004 Sumatra-Andaman event,

104: there was in the northern extent of the source a relatively slow faulting motion that led to significant vertical

105: bottom motion but left little record in the seismic data. It is interesting to point out that it is the inversion of

106: tide-gauge data from Paradip, the northernmost of the Indian east-coast stations, that led Neetu et al. \cite{indiens2}

107: to conclude that the source length was greater by roughly $30\%$ than the initial estimate of Lay et al. \cite{Lay}. 

108: Incidentally, the generation time is also longer for landslide tsunamis.

109: 

110: Our study is restricted to the water region where the incompressible Euler equations for potential flow can be 

111: linearized. The wave propagation away from the source can be investigated by shallow water models which may or may not

112: take into account nonlinear effects and frequency dispersion. Such models include the

113: Korteweg-de Vries equation \cite{KdV} for unidirectional propagation, nonlinear shallow-water equations

114: and Boussinesq-type models \cite{bouss,peregr,bona}.

115: 

116: Several authors have modeled the incompressible fluid layer as a special case of an elastic medium 

117: \cite{podyapolsk1, kajiura, gusyakov, aleksgus, gusyakov3}. In our opinion it 

118: may be convenient to model the liquid by an elastic material from a mathematical point of view, but it is questionable 

119: from a physical point of view. The crust was modeled as an

120: elastic isotropic half-space. This assumption will also be adopted in the present study.

121: 

122: The problem of tsunami generation has been considered by a number of authors: see for example \cite{carrier, driess, bradd}.  

123: The models discussed in these papers lack flexibility in terms of modelling the source due to the earthquake. The present paper

124: provides some extensions. A good review on the subject is \cite{sabatier}.

125: 

126: Here we essentially follow the framework proposed by Hammack \cite{Hammack} and others. The tsunami generation 

127: problem is reduced to a Cauchy-Poisson boundary value problem in a region of

128: constant depth. The main extensions given in the present paper consist in three-dimensional modelling and more realistic source 

129: models. This approach was followed recently in \cite{todo, todo2}, where the mathematical model was the same as in

130: \cite{Hammack} but the source was different.

131: 

132: Most analytical studies of linearized wave motion use integral transform methods. The complexity 

133: of the integral solutions forced many authors \cite{kajiura, keller} to use

134: asymptotic methods such as the method of stationary phase to estimate the far-field behaviour of the solutions. In the present 

135: study we have also obtained asymptotic formulas for integral solutions. They are useful from a qualitative point of view, 

136: but in practice it is better to use numerical integration formulas \cite{filon} that take into account the oscillatory nature

137: of the integrals. All the numerical results presented in this paper were obtained in this manner.

138: 

139: One should use asymptotic solutions with caution since they approximate exact solutions of the linearized problem. The relative 

140: importance of linear and nonlinear effects can be measured by the Stokes (or Ursell) number \cite{ursell}:

141: $$

142:   U := \frac{a/h}{(kh)^2} = \frac{a}{k^2h^3},

143: $$

144: where $k$ is a wave number, $a$ a typical wave amplitude and $h$ the water depth. For $U \gg 1$, the nonlinear effects 

145: control wave propagation and only nonlinear models are applicable. Ursell \cite{ursell}

146: proved that near the wave front $U$ behaves like

147: $$

148:   U \sim t^{\frac13}.

149: $$

150: Hence, regardless of how small nonlinear effects are initially, they will become important. 

151: 

152: Section 2 provides a description of the tsunami source when the source is an earthquake. In Section 3, we review

153: the water-wave equations and 

154: provide the analytical solution to the linearized problem in the fluid domain. Section 4 is devoted to

155: numerical results based on the analytical solution. 

156: 

157: \section{Source model}

158: 

159: The inversion of seismic wave data allows the reconstruction of

160: permanent deformations of the sea bottom following earthquakes. In

161: spite of the complexity of the seismic source and of the internal

162: structure of the earth, scientists have been relatively successful

163: in using simple models for the source. One of these models

164: is Okada's model \cite{Okada85}. Its description follows.

165: 

166: The fracture zones, along which the foci of earthquakes are to be

167: found, have been described in various papers. For example, it has

168: been suggested that Volterra's theory of dislocations might be the

169: proper tool for a quantitative description of these fracture zones

170: \cite{stek2}. This suggestion was made for the following reason. If

171: the mechanism involved in earthquakes and the fracture zones is

172: indeed one of fracture, discontinuities in the displacement

173: components across the fractured surface will exist. As dislocation

174: theory may be described as that part of the theory of elasticity

175: dealing with surfaces across which the displacement field is

176: discontinuous, the suggestion makes sense.

177: 

178: As is often done in mathematical physics, it is necessary for

179: simplicity's sake to make some assumptions. Here we neglect the

180: curvature of the earth, its gravity, temperature, magnetism,

181: non-homogeneity, and consider a semi-infinite medium, which is

182: homogeneous and isotropic. We further assume that the laws of

183: classical linear elasticity theory hold.

184: 

185: Several studies showed that the effect of earth curvature is

186: negligible for shallow events at distances of less than $20^\circ$

187: \cite{ben1,ben2,smylie}. The sensitivity to earth topography,

188: homogeneity, isotropy and half-space assumptions was studied and

189: discussed recently \cite{tim}. A commercially

190: available code, ABACUS, which is based on a finite element model

191: (FEM), was used. Six FEMs were constructed to test the sensitivity of

192: deformation predictions to each assumption. The author came to the conclusion

193: that the vertical layering of lateral inhomogeneity can sometimes

194: cause considerable effects on the deformation fields.

195: 

196: The usual boundary conditions for dealing with earth problems

197: require that the surface of the elastic medium (the earth) shall

198: be free from forces. The resulting mixed boundary-value problem was

199: solved a century ago \cite{volt}. Later, Steketee proposed an

200: alternative method to solve this problem using Green's functions

201: \cite{stek2}.

202: 

203: \subsection{Volterra's theory of dislocations}

204: 

205: In order to introduce the concept of dislocation and for

206: simplicity's sake, this section is devoted to the case of an entire

207: elastic space, as was done in the original paper by

208: Volterra \cite{volt}.

209: 

210: Let $O$ be the origin of a Cartesian coordinate system in an

211: infinite elastic medium, $x_i$ the Cartesian coordinates

212: $(i=1,2,3)$, and $\mathbf{e}_i$ a unit vector in the positive

213: $x_i-$direction. A force $\mathbf{F}=F \mathbf{e}_k$ at $O$

214: generates a displacement field $u_i^k(P,O)$ at point $P$, which is

215: determined by the well-known Somigliana tensor

216: \begin{equation}\label{somigliana}

217:   u_i^k(P,O) = \frac{F}{8\pi\mu} (\delta_{ik} r_{,\: nn}

218:   - \alpha r_{,\:ik}), \quad \mbox{with} \;\; \alpha=\frac{\lambda+\mu}{\lambda+2\mu}.

219: \end{equation}

220: In this relation $\delta_{ik}$ is the Kronecker delta, $\lambda$ and

221: $\mu$ are Lam\'e's constants, and $r$ is the distance from $P$ to

222: $O$. The coefficient $\alpha$ can be rewritten as

223: $\alpha=1/2(1-\nu)$, where $\nu$ is Poisson's ratio. Later we will

224: also use Young's modulus $E$, which is defined as

225: \[ E = \frac{\mu\,(3\lambda+2\mu)}{\lambda+\mu}. \]

226: The notation $r_{,\:i}$ means $\partial r/\partial x_i$ and the

227: summation convention applies.

228: 

229: The stresses due to the displacement field (\ref{somigliana}) are

230: easily computed from Hooke's law:

231: 

232: \begin{equation}\label{hook}

233:   \sigma_{ij} = \lambda\delta_{ij} u_{k,k} + \mu

234:   (u_{i,j}+u_{j,i}).

235: \end{equation}

236: One finds

237: \[

238:   \sigma_{ij}^k (P,O) = -\frac{\alpha F}{4\pi} \left(

239:   \frac{3 x_i x_j x_k}{r^5} + \frac{\mu}{\lambda+\mu}

240:   \frac{\delta_{ki}x_j + \delta_{kj}x_i - \delta_{ij}x_k}{r^3}

241:   \right).

242: \]

243: The components of the force per unit area on a surface element are

244: denoted as follows:

245: \[

246:   T_i^{k} = \sigma_{ij}^k \nu_j,

247: \]

248: where the $\nu_j$'s are the components of the normal to the

249: surface element. A Volterra dislocation is defined as a

250: surface $\Sigma$ in the elastic medium across which there is a

251: discontinuity $\Delta u_i$ in the displacement fields of the type

252: \begin{eqnarray}\label{displdef}

253:   \Delta u_i & = & u_i^+ - u_i^- = U_i + \Omega_{ij}x_j, \\

254:   \Omega_{ij}& = & -\Omega_{ji}.

255: \end{eqnarray}

256: Equation (\ref{displdef}) in which $U_i$ and $\Omega_{ij}$ are

257: constants is the well-known Weingarten relation which states that

258: the discontinuity $\Delta u_i$ should be of the type of a rigid body

259: displacement, thereby maintaining continuity of the components of

260: stress and strain across $\Sigma$.

261: 

262: The displacement field in an infinite elastic medium due to the

263: dislocation is then determined by Volterra's formula \cite{volt}

264: \begin{equation}\label{volt2}

265:   u_k(Q) = \frac{1}{F}\int\!\!\!\!\int

266:   \limits_{\!\!\!\!\!\!\!\Sigma} \Delta u_i

267:   T_i^{k} \, dS.

268: \end{equation}

269: 

270: Once the surface $\Sigma$ is given, the dislocation is essentially

271: determined by the six constants $U_i$ and $\Omega_{ij}$. Therefore

272: we also write

273: \begin{equation}\label{eldisloc}

274:   u_k(Q) = \frac{U_i}{F}\int\!\!\!\!\int

275:   \limits_{\!\!\!\!\!\!\!\Sigma} \sigma_{ij}^k (P,Q) \nu_j dS +

276:   \frac{\Omega_{ij}}{F}\int\!\!\!\!\int

277:   \limits_{\!\!\!\!\!\!\!\Sigma} \{x_j\sigma_{il}^k(P,Q) - x_i

278:   \sigma_{jl}^k(P,Q)\}\nu_l dS,

279: \end{equation}

280: where $\Omega_{ij}$ takes only the values $\Omega_{12}$,

281: $\Omega_{23}$, $\Omega_{31}$. Following Volterra \cite{volt} and

282: Love \cite{love} we call each of the six integrals in

283: (\ref{eldisloc}) an elementary dislocation.

284: 

285: It is clear from (\ref{volt2}) and (\ref{eldisloc}) that the

286: computation of the displacement field $u_k(Q)$ is performed as

287: follows. A force $F\mathbf{e}_k$ is applied at $Q$, and the stresses

288: $\sigma_{ij}^k(P,Q)$ that this force generates are computed at the

289: points $P(x_i)$ on $\Sigma$. In particular the components of the

290: force on $\Sigma$ are computed. After multiplication with prescribed

291: weights of magnitude $\Delta u_i$ these forces are integrated over

292: $\Sigma$ to give the displacement component in $Q$ due to the

293: dislocation on $\Sigma$.

294: 

295: \subsection{Dislocations in elastic half-space}

296: 

297: When the case of an elastic half-space is considered, equation

298: (\ref{volt2}) remains valid, but we have to replace $\sigma_{ij}^k$ in $T_i^k$ by

299: another tensor $\omega_{ij}^k$. This can be explained by the fact

300: that the elementary solutions for a half-space are different from

301: Somigliana solution (\ref{somigliana}).

302: 

303: The $\omega_{ij}^k$ can be obtained from the displacements

304: corresponding to nuclei of strain in a half-space through relation

305: (\ref{hook}). Steketee showed a method of obtaining the six

306: $\omega_{ij}^k$ fields by using a Green's function and derived

307: $\omega_{12}^k$, which is relevant to a vertical strike-slip fault (see below).

308: Maruyama derived the remaining five functions \cite{maru}.

309: 

310: It is interesting to mention here that historically these solutions

311: were first derived in a straightforward manner by Mindlin

312: \cite{mindl1,mindl2}, who gave explicit expressions of the

313: displacement and stress fields for half-space nuclei of strain

314: consisting of single forces with and without moment. It is only

315: necessary to write the single force results since the other forms

316: can be obtained by taking appropriate derivatives. The method

317: consists in finding the displacement field in Westergaard's form of

318: the Galerkin vector \cite{wester}. This vector is then determined by

319: taking a linear combination of some biharmonic elementary solutions.

320: The coefficients are chosen to satisfy boundary and equilibrium

321: conditions. These solutions were also derived by Press in a slightly

322: different manner \cite{press}.

323: 

324: \begin{figure}[htbp]

325: \begin{center}

326: \input figok.tex

327: \end{center}

328:   \caption{Coordinate system adopted in this study and

329: geometry of the source model}\label{fig:okad}

330: \end{figure}

331: 

332: Here, we take the Cartesian coordinate system shown in

333: Figure~\ref{fig:okad}. The elastic medium occupies the region $x_3\leq

334: 0$ and the $x_1-$axis is taken to be parallel to the strike direction

335: of the fault. In this coordinate system,

336: $u_i^j(x_1,x_2,x_3;\xi_1,\xi_2,\xi_3)$ is the $i$th component of the

337: displacement at $(x_1,x_2,x_3)$ due to the $j$th direction point

338: force of magnitude $F$ at $(\xi_1,\xi_2,\xi_3)$. It can be expressed

339: as follows \cite{Okada85,mindl1,press,okada92}:

340: \begin{eqnarray}\label{okada1}

341:   u_i^j (x_1,x_2,x_3) & = & u_{iA}^j(x_1,x_2,-x_3) -

342:   u_{iA}^j(x_1,x_2,x_3) \\

343:   & & + u_{iB}^j(x_1,x_2,x_3) + x_3

344:   u_{iC}^j(x_1,x_2,x_3), \nonumber

345: \end{eqnarray}

346: where

347: \begin{eqnarray*}

348:   u_{iA}^j & = & \frac{F}{8\pi\mu}\left((2-\alpha)\frac{\delta_{ij}}{R}

349:   + \alpha \frac{R_iR_j}{R^3}\right), \\

350:   u_{iB}^j & = & \frac{F}{4\pi\mu}\Biggl(

351:   \frac{\delta_{ij}}{R} + \frac{R_iR_j}{R^3} +

352:   \frac{1-\alpha}{\alpha}\Bigl[

353:   \frac{\delta_{ij}}{R+R_3} +\\ & & + \frac{R_i\delta_{j3}-

354:   R_j\delta_{i3}(1-\delta_{j3})}{R(R+R_3)} -

355:   \frac{R_iR_j}{R(R+R_3)^2}(1-\delta_{i3})(1-\delta_{j3})

356:   \Bigr]\Biggr), \\

357:     u_{iC}^j & = & \frac{F}{4\pi\mu}(1-2\delta_{i3})

358:     \left(

359:     (2-\alpha)\frac{R_i\delta_{j3}-R_j\delta_{i3}}{R^3} +

360:     \alpha\xi_3\left[

361:     \frac{\delta_{ij}}{R^3} - 3\frac{R_iR_j}{R^5}

362:     \right]\right).

363: \end{eqnarray*}

364: In these expressions $R_1 = x_1-\xi_1$, $R_2=x_2-\xi_2$,

365: $R_3=-x_3-\xi_3$ and $R^2 = R_1^2 + R_2^2 + R_3^2$.

366: 

367: The first term in equation (\ref{okada1}), $u_{iA}^j(x_1,x_2,-x_3)$,

368: is the well-known Somigliana tensor, which represents the

369: displacement field due to a single force placed at

370: $(\xi_1,\xi_2,\xi_3)$ in an infinite medium \cite{love}. The second

371: term also looks like a Somigliana tensor. This term corresponds to a

372: contribution from an image source of the given point force placed at

373: $(\xi_1,\xi_2,-\xi_3)$ in the infinite medium. The third term,

374: $u_{iB}^j(x_1,x_2,x_3)$, and $u_{iC}^j(x_1,x_2,x_3)$ in the fourth

375: term are naturally depth dependent. When $x_3$ is set equal to zero

376: in equation (\ref{okada1}), the first and the second terms cancel

377: each other, and the fourth term vanishes. The remaining term,

378: $u_{iB}^j(x_1,x_2,0)$, reduces to the formula for the surface

379: displacement field due to a point force in a half-space

380: \cite{Okada85}:

381: \[

382: \left\{%

383: \begin{array}{ll}

384:     u_1^1 = \frac{F}{4\pi\mu}\left(

385:     \frac1R + \frac{(x_1-\xi_1)^2}{R^3} + \frac{\mu}{\lambda+\mu}

386:     \left[

387:     \frac{1}{R-\xi_3} - \frac{(x_1-\xi_1)^2}{R(R-\xi_3)^2}\right]\right), & \\

388:     u_2^1 = \frac{F}{4\pi\mu}(x_1-\xi_1)(x_2-\xi_2)\left(

389:     \frac1{R^3} - \frac{\mu}{\lambda+\mu}\frac1{R(R-\xi_3)^2}

390:     \right), &  \\

391:     u_3^1 = \frac{F}{4\pi\mu}(x_1-\xi_1)\left(

392:     -\frac{\xi_3}{R^3} - \frac{\mu}{\lambda+\mu} \frac1{R(R-\xi_3)}

393:     \right), &

394: \end{array}%

395: \right.

396: \]

397: 

398: \[

399: \left\{%

400: \begin{array}{ll}

401:     u_1^2 = \frac{F}{4\pi\mu}(x_1-\xi_1)(x_2-\xi_2)\left(

402:     \frac1{R^3} - \frac{\mu}{\lambda+\mu}\frac{1}{R(R-\xi_3)^2}\right), & \\

403:     u_2^2 = \frac{F}{4\pi\mu}\left(

404:     \frac1R + \frac{(x_2-\xi_2)^2}{R^3} + \frac{\mu}{\lambda+\mu}

405:     \left[

406:     \frac{1}{R-\xi_3} - \frac{(x_2-\xi_2)^2}{R(R-\xi_3)^2}\right]\right), &  \\

407:     u_3^2 = \frac{F}{4\pi\mu}(x_2-\xi_2)\left(

408:     -\frac{\xi_3}{R^3} - \frac{\mu}{\lambda+\mu} \frac1{R(R-\xi_3)}

409:     \right), &

410: \end{array}%

411: \right.

412: \]

413: 

414: \[

415: \left\{%

416: \begin{array}{ll}

417:     u_1^3 = \frac{F}{4\pi\mu}(x_1-\xi_1)\left(

418:     -\frac{\xi_3}{R^3} + \frac{\mu}{\lambda+\mu}\frac{1}{R(R-\xi_3)}\right), & \\

419:     u_2^3 = \frac{F}{4\pi\mu}(x_2-\xi_2)\left(

420:     -\frac{\xi_3}{R^3} + \frac{\mu}{\lambda+\mu}\frac{1}{R(R-\xi_3)}\right), &  \\

421:     u_3^3 = \frac{F}{4\pi\mu}\left(

422:     \frac1{R} + \frac{\xi_3^2}{R^3} + \frac{\mu}{\lambda+\mu}\frac1R

423:     \right). &

424: \end{array}%

425: \right.

426: \]

427: In these formulas $R^2 = (x_1-\xi_1)^2 + (x_2-\xi_2)^2 + \xi_3^2$.

428: 

429: In order to obtain the displacements due to the dislocation we need

430: to calculate the corresponding $\xi_k$-derivatives of the point

431: force solution (\ref{okada1}) and to insert them in Volterra's

432: formula (\ref{volt2})

433: \[

434:   u_i = \frac1F\int\!\!\!\!\int

435:   \limits_{\!\!\!\!\!\!\!\Sigma} \Delta u_j

436:   \left[

437:   \lambda\delta_{jk} \pd{u_i^n}{\xi_n} +

438:   \mu\left(\pd{u_i^j}{\xi_k} + \pd{u_i^k}{\xi_j}

439:   \right)\right]\nu_k \,dS.

440: \]

441: The $\xi_k$-derivatives are expressed as follows:

442: \begin{eqnarray*}

443:   \pd{u_i^j}{\xi_k} (x_1,x_2,x_3) & = & \pd{u_{iA}^j}{\xi_k}(x_1,x_2,-x_3) -

444:   \pd{u_{iA}^j}{\xi_k}(x_1,x_2,x_3) +\\ & & + \pd{u_{iB}^j}{\xi_k}(x_1,x_2,x_3) + x_3

445:   \pd{u_{iC}^j}{\xi_k}(x_1,x_2,x_3),

446: \end{eqnarray*}

447: with

448: \begin{eqnarray*}

449:   \pd{u_{iA}^j}{\xi_k} & = & \frac{F}{8\pi\mu}\left(

450:   (2-\alpha)\frac{R_k}{R^3}\delta_{ij} - \alpha

451:   \frac{R_i\delta_{jk}+R_j\delta_{ik}}{R^3} + 3\alpha

452:   \frac{R_i R_j R_k}{R^5}

453:   \right), \\

454:   \pd{u_{iB}^j}{\xi_k} & = & \frac{F}{4\pi\mu}\left(

455:   -\frac{R_i\delta_{jk} + R_j\delta_{ik} - R_k\delta_{ij}}{R^3} +

456:   3\frac{R_iR_jR_k}{R^5} \right. +\\ & & + \frac{1-\alpha}{\alpha}\Bigl[

457:   \frac{\delta_{3k}R + R_k}{R(R+R_3)^2}\delta_{ij} -

458:   \frac{\delta_{ik}\delta_{j3} -

459:   \delta_{jk}\delta_{i3}(1-\delta_{j3})}{R(R+R_3)} +\\ & & +

460:   \bigl(R_i\delta_{j3} - R_j\delta_{i3}(1-\delta_{j3})\bigr)

461:   \frac{\delta_{3k}R^2+R_k(2R+R_3)}{R^3(R+R_3)^2} +\\ & & \left. +

462:   (1-\delta_{i3})(1-\delta_{j3})\bigl(

463:   \frac{R_i\delta_{jk}+R_j\delta_{ik}}{R(R+R_3)^2} -

464:   R_iR_j\frac{2\delta_{3k}R^2 + R_k(3R+R_3)}{R^3(R+R_3)^3}

465:   \bigr)\Bigr]\right), \\

466:   \pd{u_{iC}^j}{\xi_k} & = & \frac{F}{4\pi\mu}(1-2\delta_{i3})\biggl(

467:   (2-\alpha)\Bigl[

468:   \frac{\delta_{jk}\delta_{i3}-\delta_{ik}\delta_{j3}}{R^3} +

469:   \frac{3R_k(R_i\delta_{j3}-R_j\delta_{i3})}{R^5}\Bigr] + \\& & +

470:   \alpha\delta_{3k}\Bigl[\frac{\delta_{ij}}{R^3} -

471:   \frac{3R_iR_j}{R^5}\Bigr] + 3\alpha\xi_3\Bigl[

472:   \frac{R_i\delta_{jk}+R_j\delta_{ik}+R_k\delta_{ij}}{R^5} -

473:   \frac{5R_iR_jR_k}{R^7}\Bigr]\biggr).

474: \end{eqnarray*}

475: 

476: \subsection{Finite rectangular source}

477: Let us now consider a more practical problem. We define the

478: elementary dislocations $U_1$, $U_2$ and $U_3$, corresponding to

479: the strike-slip, dip-slip and tensile components of an arbitrary

480: dislocation. In Figure \ref{fig:okad} each vector represents the

481: direction of the elementary faults. The vector $\Dv$ is the

482: so-called Burger's vector, which shows how both sides of the fault

483: are spread out: $\Dv = \uv^+ - \uv^-$.

484: 

485: A general dislocation can be determined by three angles: the dip

486: angle $\delta$ of the fault $(0\le\delta\le\pi)$, the slip or rake angle $\theta$ $(0\le\theta\le\pi)$, and the angle

487: $\phi$ between the fault plane and Burger's vector $\Dv$. When dealing with a

488: geophysical application, an additional angle, the azimuth or strike, is introduced in order to provide an orientation of the fault.

489: The general situation is schematically described in Figure \ref{fig:2}.

490: 

491: \begin{figure}[htbp]

492: \begin{center}

493: \input fig2mod.tex

494: \end{center}

495:   \caption{Geometry of the source model and orientation

496:   of Burger's vector $\Dv$}\label{fig:2}

497: \end{figure}

498: 

499: For a finite rectangular fault with length $L$ and width $W$

500: occurring at depth $d$ (Figure \ref{fig:2}), the deformation field

501: can be evaluated analytically by a change of variables and by

502: integrating over the rectangle. This was done by several

503: authors \cite{Okada85,okada92,chin,sat74,iwa79}. Here we give the

504: results of their computations. The final results are represented below in

505: compact form, using Chinnery's notation $\|$ to

506: represent the substitution

507: \[

508:   f(\xi,\eta)\| = f(x,p) - f(x,p-W) - f(x-L,p) + f(x-L,p-W), 

509: \]

510: where $p = y\cos\delta + d\sin\delta$. Next we introduce the notation

511: \[

512:   q = y\sin\delta - d\cos\delta, \quad \tilde{y} = \eta\cos\delta + q\sin\delta,

513:  \quad \tilde{d} = \eta\sin\delta - q\cos\delta

514: \]

515: and

516: \[

517:   R^2 = \xi^2 + \eta^2 + q^2 = \xi^2 + \tilde{y}^2 + \tilde{d}^2, \quad

518:   X^2 = \xi^2 + q^2.

519: \]

520: 

521: The quantities $U_1$, $U_2$ and $U_3$ are linked to Burger's vector

522: through the identities

523: \[

524:   U_1 = |\Dv| \cos\phi\cos\theta, \quad

525:   U_2 = |\Dv| \cos\phi\sin\theta, \quad

526:   U_3 = |\Dv| \sin\phi.

527: \]

528: For a strike-slip dislocation, one has

529: \begin{eqnarray*}

530:   u_1 & = & -\frac{U_1}{2\pi}\left.\left(

531:   \frac{\xi q}{R(R+\eta)} + \arctan\frac{\xi\eta}{qR} +

532:   I_1\sin\delta\right)\right\|, \\

533:   u_2 & = & -\frac{U_1}{2\pi}\left.\left(

534:   \frac{\tilde{y} q}{R(R+\eta)} + \frac{q\cos\delta}{R+\eta} +

535:   I_2\sin\delta\right)\right\|, \\

536:   u_3 & = & -\frac{U_1}{2\pi}\left.\left(

537:   \frac{\tilde{d} q}{R(R+\eta)} + \frac{q\sin\delta}{R+\eta} +

538:   I_4\sin\delta\right)\right\|.

539: \end{eqnarray*}

540: For a dip-slip dislocation, one has

541: \begin{eqnarray*}

542:   u_1 & = & -\frac{U_2}{2\pi}\left.\left(

543:   \frac{q}{R} - I_3\sin\delta\cos\delta

544:   \right)\right\|, \\

545:   u_2 & = & -\frac{U_2}{2\pi}\left.\left(

546:   \frac{\tilde{y} q}{R(R+\xi)} +

547:   \cos\delta\arctan\frac{\xi\eta}{qR} - I_1\sin\delta\cos\delta

548:   \right)\right\|, \\

549:   u_3 & = & -\frac{U_2}{2\pi}\left.\left(

550:   \frac{\tilde{d} q}{R(R+\xi)} +

551:   \sin\delta\arctan\frac{\xi\eta}{qR} - I_5\sin\delta\cos\delta

552:   \right)\right\|.

553: \end{eqnarray*}

554: For a tensile fault dislocation, one has

555: \begin{eqnarray*}

556:   u_1 & = & \frac{U_3}{2\pi}\left.\left(

557:   \frac{q^2}{R(R+\eta)} - I_3\sin^2\delta

558:   \right)\right\|, \\

559:   u_2 & = & \frac{U_3}{2\pi}\left.\left(

560:   \frac{-\tilde{d} q}{R(R+\xi)} - \sin\delta\left[

561:   \frac{\xi q}{R(R+\eta)} - \arctan\frac{\xi\eta}{qR}

562:   \right] - I_1\sin^2\delta

563:   \right)\right\|, \\

564:   u_3 & = & \frac{U_3}{2\pi}\left.\left(

565:   \frac{\tilde{y} q}{R(R+\xi)} + \cos\delta\left[

566:   \frac{\xi q}{R(R+\eta)} - \arctan\frac{\xi\eta}{qR}

567:   \right] - I_5\sin^2\delta

568:   \right)\right\|.

569: \end{eqnarray*}

570: The terms $I_1,\dots,I_5$ are given by

571: \begin{eqnarray*}

572:   I_1 & = &

573:   -\frac{\mu}{\lambda+\mu}\frac{\xi}{(R+\tilde{d})\cos\delta} -

574:   \tan\delta I_5, \\

575:   I_2 & = & -\frac{\mu}{\lambda+\mu}\log(R+\eta) - I_3, \\

576:   I_3 & = & \frac{\mu}{\lambda+\mu}\left[

577:   \frac1{\cos\delta}\frac{\tilde{y}}{R+\tilde{d}} - \log(R+\eta)

578:   \right] + \tan\delta I_4, \\

579:   I_4 & = & \frac{\mu}{\mu+\lambda}\frac1{\cos\delta}\left(

580:   \log(R+\tilde{d}) - \sin\delta \log(R+\eta)

581:   \right), \\

582:   I_5 & = & \frac{\mu}{\lambda+\mu}\frac2{\cos\delta}

583:   \arctan\frac{\eta(X+q\cos\delta)+X(R+X)\sin\delta}{\xi(R+X)\cos\delta},

584: \end{eqnarray*}

585: and if $\cos\delta=0$,

586: \begin{eqnarray*}

587:   I_1 & = & -\frac{\mu}{2(\lambda+\mu)}

588:   \frac{\xi q}{(R+\tilde{d})^2}, \\

589:   I_3 & = & \frac{\mu}{2(\lambda+\mu)} \left[

590:   \frac{\eta}{R+\tilde{d}} + \frac{\tilde{y} q}{(R+\tilde{d})^2} -

591:   \log(R+\eta)\right], \\

592:   I_4 & = & -\frac{\mu}{\lambda+\mu} \frac{q}{R+\tilde{d}}, \\

593:   I_5 & = & -\frac{\mu}{\lambda+\mu} \frac{\xi\sin\delta}{R+\tilde{d}}.

594: \end{eqnarray*}

595: 

596: Figures \ref{fig:dip}, \ref{fig:strike}, and \ref{fig:tensile} show

597: the free-surface deformation due to the three elementary dislocations.

598: The values of the parameters are given in Table \ref{parset}.

599: \begin{table}

600: \centering

601: \begin{tabular}{lc}

602:   \hline

603: {\it parameter} & {\it value} \\

604: \hline

605:   Dip angle $\delta$ & $13^\circ$ \\

606:   Fault depth $d$, km & 25 \\

607:   Fault length $L$, km & 220 \\

608:   Fault width $W$, km & 90 \\

609:   $U_i$, m & 15 \\

610:   Young modulus $E$, GPa & 9.5 \\

611:   Poisson's ratio $\nu$ & 0.23 \\

612: \hline

613: \end{tabular}

614: \caption[]{Parameter set used in Figures \ref{fig:dip},

615: \ref{fig:strike}, and \ref{fig:tensile}.} \label{parset}

616: \end{table}

617: 

618: \begin{figure}

619:   \includegraphics[width=0.9\linewidth]{dip.eps}\\

620:   \caption{Dimensionless free-surface deformation $z/a$ due to dip-slip faulting: $\phi=0$,

621: $\theta=\pi/2$, $\Dv=(0,U_2,0)$. Here $a$

622: is $|\Dv|$ (15 m in the present application). The horizontal distances $x$ and $y$ are expressed in kilometers.}

623:   \label{fig:dip}

624: \end{figure}

625: 

626: \begin{figure}

627:   \includegraphics[width=\linewidth]{strike.eps}\\

628:   \caption{Dimensionless free-surface deformation $z/a$ due to strike-slip faulting: $\phi=0$,

629: $\theta=0$, $\Dv=(U_1,0,0)$. Here $a$

630: is $|\Dv|$ (15 m in the present application). The horizontal distances $x$ and $y$ are expressed in kilometers.}

631:   \label{fig:strike}

632: \end{figure}

633: 

634: \begin{figure}

635:   \includegraphics[width=0.9\linewidth]{tensile.eps}\\

636:   \caption{Dimensionless free-surface deformation $z/a$ due to tensile faulting: $\phi=\pi/2$,

637: $\Dv=(0,0,U_3)$. Here $a$ is $|\Dv|$. The horizontal distances $x$ and $y$ are expressed in kilometers.}

638:   \label{fig:tensile}

639: \end{figure}

640: 

641: \subsection{Curvilinear fault}

642: 

643: In the previous subsection analytical formulas for the free-surface

644: deformation in the special case of a rectangular fault were given. In

645: fact, Volterra's formula (\ref{volt2}) allows to

646: evaluate the displacement field that accompanies fault events with much more

647: general geometry. The shape of the fault and Burger's vector

648: are suggested by seismologists and after numerical integration one

649: can obtain the deformation of the seafloor for more general types of events as well.

650: 

651: Here we will consider the case of a fault whose geometry is described by an elliptical

652: arc (see Figure \ref{fig:ellipse}).

653: \begin{figure}

654:   \includegraphics[width=0.8\linewidth]{ellipse.eps}\\

655:   \caption{Geometry of a fault with elliptical shape.}

656:   \label{fig:ellipse}

657: \end{figure}

658: The parametric equations of this surface are given by

659: $$

660:   x(\xi,\eta) = \xi, \quad 0\leq\xi\leq a, \qquad y(\xi,\eta) = \eta, \quad -\frac{c}2\leq\eta\leq\frac{c}2, 

661: $$

662: $$

663:   z(\xi,\eta) = -(b+d) + \frac{b}{a}\sqrt{a^2-\xi^2}.

664: $$

665: Then the unit normal to this surface can be easily calculated:

666: $$

667:     \nv = \left(

668:     \frac{b\xi}{\sqrt{a^4 + (b^2-a^2)\xi^2}}, 0, \frac{a\sqrt{a^2-\xi^2}}

669:     {\sqrt{a^4 + (b^2-a^2)\xi^2}}

670:     \right).

671: $$

672: We also need to compute the coefficients of the first fundamental

673: form in order to reduce the surface integral in (\ref{volt2}) to a

674: double Riemann integral. These coefficients are

675: $$ E = \frac{a^4 + \xi^2(b^2-a^2)}{a^2(a^2-\xi^2)}, \quad F = 0, \quad G = 1 $$

676: and the surface element $dS$ is 

677: $$ dS = \sqrt{EG-F^2} \,d\xi d\eta = \frac1a\frac{\sqrt{a^4+\xi^2(b^2-a^2)}}{\sqrt{a^2-\xi^2}}

678:   \, d\xi d\eta. $$

679: 

680: Since in the crust the hydrostatic pressure is very large, it is

681: natural to impose the condition that $\Dv \cdot \nv = 0.$

682: The physical meaning of this condition is that both sides of the

683: fault slide and do not detach. This condition is obviously satisfied

684: if we take Burger's vector as

685: $$

686:     \Dv = D\left(

687:     \frac{a\sqrt{a^2-\xi^2}}{\sqrt{a^4+\xi^2(b^2-a^2)}}, 0,

688:     -\frac{b\xi}{\sqrt{a^4+\xi^2(b^2-a^2)}}

689:     \right).

690: $$

691: It is evident that $D = |\Dv|$.

692: 

693: The numerical integration was performed using a $9$-point

694: two-dimensional Gauss-type integration formula.

695: %\begin{multline*}

696: %  \int\int\limits_{\!\!\!\!\!\!\!\!\!\!\!\![-1,1]^2} f(\xi,\eta)

697: %  d\xi d\eta = \frac1{81}\biggl(

698: %  64 f(0,0) + 40(f(0,\sqrt{\frac35}) + f(\sqrt{\frac35},0) +\\+

699: %  f(-\sqrt{\frac35},0) + f(0,-\sqrt{\frac35})) + 25(

700: %  f(\sqrt{\frac35},\sqrt{\frac35}) +

701: %  f(-\sqrt{\frac35},\sqrt{\frac35}) +\\+

702: %  f(\sqrt{\frac35},-\sqrt{\frac35}) +

703: %  f(-\sqrt{\frac35},-\sqrt{\frac35}))\biggr)

704: %\end{multline*}

705: The result is presented on Figure

706: \ref{fig:ellres}. The parameter values are given in Table \ref{parellipse}.

707: \begin{table}

708: \centering

709: \begin{tabular}{lc}

710:   \hline

711:   {\it parameter} & {\it value} \\

712:   \hline

713:   Depth event $d$, km & 20 \\

714:   Ellipse semiminor axis $a$, km & 17 \\

715:   Ellipse semimajor axis $b$, km & 6 \\

716:   Fault width $c$, km & 15 \\

717:   Young modulus $E$, GPa & 9.5 \\

718:   Poisson's ratio $\nu$ & 0.23 \\

719:   \hline

720: \end{tabular}

721: \caption[]{Parameter set used in Figure \ref{fig:ellres}.} \label{parellipse}

722: \end{table}

723: 

724: \begin{figure}

725:   \includegraphics[width=\linewidth]{ell.eps}\\

726:   \caption{Free-surface deformation due to curvilinear faulting. The horizontal distances $x$ and $y$ are expressed in kilometers.}

727:   \label{fig:ellres}

728: \end{figure}

729: 

730: The example considered in this subsection may not be

731: physically relevant. However it shows how Okada's solution can be extended. For a more precise modeling of

732: the faulting event we need to have more information about the earthquake source

733: and its related parameters. 

734: 

735: After having reviewed the description of the source, we now switch to the deformation of the ocean surface

736: following a submarine earthquake.

737: The traditional approach for hydrodynamic modelers is to use

738: elastic models similar to the model we just described with the seismic

739: parameters as input in order to evaluate the details of the seafloor deformation.

740: Then this deformation is translated to the free surface of the ocean and serves as initial condition of the

741: evolution problem described in the next section.

742: 

743: \section{Solution in fluid domain}

744: 

745: \begin{figure}

746: \centering

747:   \includegraphics[width=0.8\linewidth]{fluid.eps}\\

748:   \caption{Definition of the fluid domain and coordinate system}

749:   \label{fig:fluid}

750: \end{figure}

751: 

752: The fluid domain is supposed to represent the ocean above the fault area. 

753: Let us consider the fluid domain $\Omega$ shown in Figure

754: \ref{fig:fluid}. It is bounded above by the free surface of the ocean and below

755: by the rigid ocean floor. The domain $\Omega$ is unbounded in the horizontal

756: directions $x$ and $y$, and can be written as

757: $$

758:   \Omega = \mathbb{R}^2\times\left[-h+\zeta(x,y,t),\eta(x,y,t)\right].

759: $$

760: Initially the fluid is assumed to be at rest and the sea bottom to be 

761: horizontal. Thus, at time $t=0$, the

762: free surface and the sea bottom are defined by $z=0$ and $z=-h$,

763: respectively. For time $t>0$ the bottom boundary moves in a

764: prescribed manner which is given by

765: $$

766:     z = -h + \zeta(x,y,t).

767: $$

768: The displacement of the sea bottom is assumed to have all the properties

769: required to compute its Fourier transform in $x,y$ and its Laplace transform in $t$. 

770: The resulting deformation of the free surface

771: $z=\eta(x,y,t)$ must be found. It is also assumed that the

772: fluid is incompressible and the flow is irrotational. The latter

773: implies the existence of a velocity potential $\phi(x,y,z,t)$ which

774: completely describes this flow. By definition of $\phi$, the fluid

775: velocity vector can be expressed as $\qv = \nabla\phi$. Thus, the

776: continuity equation becomes

777: \begin{equation}

778:   \nabla\cdot\qv = \Delta\phi = 0, \quad (x,y,z) \in \Omega.

779: \label{laplacien}

780: \end{equation}

781: The potential $\phi(x,y,z,t)$ must also satisfy the following

782: kinematic boundary conditions on the free-surface and the solid

783: boundary, respectively:

784: \begin{eqnarray}

785:   \pd{\phi}{z} &=& \pd{\eta}{t} + \pd{\phi}{x}\pd{\eta}{x} +

786:   \pd{\phi}{y}\pd{\eta}{y}, \qquad z=\eta(x,y,t), \label{kscl} \\

787:   \pd{\phi}{z} &=& \pd{\zeta}{t} + \pd{\phi}{x}\pd{\zeta}{x} +

788:   \pd{\phi}{y}\pd{\zeta}{y}, \qquad z=-h + \zeta(x,y,t). \label{kbcl}

789: \end{eqnarray}

790: Assuming that viscous effects as well as capillary effects can be neglected, the dynamic condition to be 

791: satisfied on the free surface reads

792: \begin{equation}

793:   \pd{\phi}{t} + \frac12|\nabla\phi|^2 + g\eta = 0, \qquad

794:   z=\eta(x,y,t). \label{dcl}

795: \end{equation}

796: As described above, the initial conditions are given by

797: \begin{equation}\label{initialcond}

798:   \eta(x,y,0) = 0 \quad \mbox{and} \quad \zeta(x,y,0) = 0.

799: \end{equation}

800: 

801: The significance of the various terms in the equations is more transparent when the equations are written in dimensionless

802: variables. The new independent variables are

803: $$

804:     \widetilde{x} = \kappa x, \quad \widetilde{y} = \kappa y, \quad \widetilde{z} = \kappa z, \quad

805:     \widetilde{t} = \sigma t,

806: $$

807: where $\kappa$ is a wavenumber and $\sigma$ is a typical frequency. Note that here the same unit length is used in the

808: horizontal and vertical directions, as opposed to shallow-water theory.

809: 

810: The new dependent variables are

811: $$

812:   \widetilde{\eta} = \frac{\eta}a, \quad \widetilde{\zeta} = \frac{\zeta}{a}, \quad

813:   \widetilde{\phi} = \frac{\kappa}{a\sigma}\phi,

814: $$

815: where $a$ is a characteristic wave amplitude. A dimensionless water depth is also introduced:

816: $$ \widetilde{h} = \kappa h. $$

817: In dimensionless form, and after dropping the tildes, equations (\ref{laplacien}--\ref{dcl}) become

818: $$

819:   \Delta \phi = 0, \qquad (x,y,z) \in \Omega,

820: $$

821: \begin{eqnarray*}

822:   \pd{\phi}{z} &=& \pd{\eta}{t} + \kappa a \left(

823:   \pd{\phi}{x}\pd{\eta}{x} + \pd{\phi}{y}\pd{\eta}{y}

824:   \right), \qquad z = \kappa a \, \eta(x,y,t), \\

825:   \pd{\phi}{z} &=& \pd{\zeta}{t} + \kappa a \left(

826:   \pd{\phi}{x}\pd{\zeta}{x} + \pd{\phi}{y}\pd{\zeta}{y}

827:   \right), \qquad z = -h + \kappa a \, \zeta(x,y,t),

828: \end{eqnarray*}

829: $$

830:   \pd{\phi}{t} + \frac12 \kappa a |\nabla\phi|^2 + \frac{g\kappa}{\sigma^2}

831:   \eta = 0, \qquad z = \kappa a \, \eta(x,y,t).

832: $$

833: 

834: Finding the solution to this problem is quite a difficult task due to the

835: nonlinearities and the a priori unknown free surface. In this study

836: we linearize the equations and the boundary conditions by taking the limit as $\kappa a\to 0$. 

837: In fact, the linearized problem can be found

838: by expanding the unknown functions as power series of a small parameter

839: $\varepsilon:=\kappa a$. Collecting the lowest order terms in

840: $\varepsilon$ yields the linear approximation. For the sake of convenience,

841: we now switch back to the physical variables. The linearized

842: problem in dimensional variables reads

843: \begin{equation}\label{lapl}

844:   \Delta \phi = 0, \qquad (x,y,z) \in \mathbb{R}^2\times[-h, 0],

845: \end{equation}

846: \begin{equation}\label{kinfreesurf}

847:   \pd{\phi}{z} = \pd{\eta}{t}, \qquad z = 0,

848: \end{equation}

849: \begin{equation}\label{kinsolb}

850:   \pd{\phi}{z} = \pd{\zeta}{t}, \qquad z = -h,

851: \end{equation}

852: \begin{equation}\label{dynfreesurf}

853:   \pd{\phi}{t} + g\eta = 0, \qquad z = 0.

854: \end{equation}

855: 

856: Combining equations (\ref{kinfreesurf}) and (\ref{dynfreesurf})

857: yields the single free-surface condition

858: \begin{equation}\label{singlefreesurf}

859:   \pd{^2\phi}{t^2} + g\pd{\phi}{z} = 0,

860:   \qquad z = 0.

861: \end{equation}

862: 

863: This problem will be solved by using the method of integral transforms. We apply

864: the Fourier transform in $(x,y)$:

865: \begin{eqnarray*}

866:   \mathfrak{F}[f] & = & \widehat{f}(k,\ell) = \int\limits_{\mathbb{R}^2} f(x,y)

867:   e^{-i(kx+\ell y)}\,dx dy, \\

868:   \mathfrak{F}^{-1}[\widehat{f}] & = & f(x,y) = \frac1{(2\pi)^2}

869:   \int\limits_{\mathbb{R}^2} \widehat{f}(k,\ell)

870:   e^{i(kx+\ell y)}\,dk d\ell,

871: \end{eqnarray*}

872: and the Laplace transform in time $t$:

873: $$

874:   \mathfrak{L}[g] = \tens{g}(s) = \int\limits_0^{+\infty}

875:   g(t) e^{-st}\, dt.

876: $$ 

877: For the combined Fourier and Laplace transforms, the following notation is introduced:

878: $$ \mathfrak{F}\mathfrak{L}[F(x,y,t)] = \overline{F}(k,\ell,s) = \int\limits_{\mathbb{R}^2}

879: e^{-i(kx+\ell y)}\,dx dy \int\limits_0^{+\infty} F(x,y,t)

880:   e^{-st}\, dt. $$ 

881: After applying the transforms, equations (\ref{lapl}), (\ref{kinsolb}) and

882: (\ref{singlefreesurf}) become

883: \begin{equation}\label{lapltrans}

884:   \od{^2\overline{\phi}}{z^2} - (k^2+\ell^2)\overline{\phi} = 0,

885: \end{equation}

886: \begin{equation}\label{kinsolbtrans}

887:   \od{\overline{\phi}}{z}(k,\ell,-h,s) = s\overline{\zeta}(k,\ell,s),

888: \end{equation}

889: \begin{equation}\label{freesurftrans}

890:   s^2\overline{\phi}(k,\ell,0,s) + g\od{\overline{\phi}}{z} (k,\ell,0,s) = 0.

891: \end{equation}

892: The transformed free-surface elevation can be obtained from

893: (\ref{dynfreesurf}):

894: \begin{equation}\label{transfreesurf}

895:   \overline{\eta}(k,\ell,s) = -\frac{s}{g}\overline{\phi}(k,\ell,0,s).

896: \end{equation}

897: 

898: A general solution of equation (\ref{lapltrans}) is given by

899: \begin{equation}\label{genlapltrans}

900:   \overline{\phi}(k,\ell,z,s) = A(k,\ell,s)\cosh(mz) + B(k,\ell,s)\sinh(mz),

901: \end{equation}

902: where $m = \sqrt{k^2+\ell^2}$. The functions $A(k,\ell,s)$ and $B(k,\ell,s)$

903: can be easily found from the boundary conditions (\ref{kinsolbtrans}) and

904: (\ref{freesurftrans}):

905: \begin{eqnarray*}

906:   A(k,\ell,s) & = & -\frac{gs\overline{\zeta}(k,\ell,s)}{\cosh(mh)[s^2+gm\tanh(mh)]}, \\

907:   B(k,\ell,s) & = & \frac{s^3\overline{\zeta}(k,\ell,s)}{m\cosh(mh)[s^2+gm\tanh(mh)]}.

908: \end{eqnarray*}

909: From now on, the notation

910: \begin{equation}

911: \omega = \sqrt{gm\tanh(mh)}

912: \label{disprelation}

913: \end{equation}

914: will be used. The graphs of $\omega(m)$, $\omega'(m)$ and $\omega''(m)$ are shown in Figure \ref{fig:omega}.

915: \begin{figure}

916:   \includegraphics[width=0.9\linewidth]{omega.eps}\\

917:   \caption{Plot of the frequency $\omega(m)=\sqrt{gm\tanh(mh)}$ and its

918:   derivatives $d\omega/dm$, $d^2\omega/dm^2$. The acceleration due to gravity $g$ and the water depth $h$ have been

919: set equal to 1.}

920:   \label{fig:omega}

921: \end{figure}

922: 

923: Substituting the expressions for the functions $A$, $B$ in (\ref{genlapltrans}) yields

924: \begin{equation}

925:   \overline{\phi}(k,\ell,z,s) = -\frac{gs\overline{\zeta}(k,\ell,s)}

926:   {\cosh(mh)(s^2+\omega^2)}

927:   \left(\cosh(mz) - \frac{s^2}{gm}\sinh(mz)\right). \label{phihat}

928: \end{equation}

929: 

930: \subsection{Free-surface elevation}

931: 

932: From (\ref{transfreesurf}), the free-surface elevation becomes

933: $$

934:   \overline{\eta}(k,\ell,s) = \frac{s^2\overline{\zeta}(k,\ell,s)}

935:   {\cosh(mh)(s^2+\omega^2)}.

936: $$

937: 

938: Inverting the Laplace and Fourier transforms provides the general

939: integral solution

940: \begin{equation}\label{genintsol}

941:   \eta(x,y,t) = \frac{1}{(2\pi)^2}\int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

942:   \frac{e^{i(kx+\ell y)}}{\cosh(mh)}\frac1{2\pi i}

943:   \int\limits_{\mu-i\infty}^{\mu+i\infty}

944:   \frac{s^2\overline{\zeta}(k,\ell,s)}{s^2+\omega^2}e^{st}ds\; dk d\ell.

945: \end{equation}

946: One can evaluate the Laplace integral in (\ref{genintsol})

947: using the convolution theorem:

948: $$

949:   \mathfrak{L}[f_1(t)*f_2(t)] = \tens{f_1}(s) \tens{f_2}(s).

950: $$

951: It yields

952: $$

953:   \eta(x,y,t) = \frac1{(2\pi)^2}\int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

954:   \frac{e^{i(kx+\ell y)}}{\cosh(mh)}

955:   \int\limits_0^t (1-\omega\sin\omega\tau)\overline{\zeta}(k,\ell,t-\tau)d\tau \, dk d\ell.

956: $$

957: 

958: This general solution contains as a special case the

959: solution for an axisymmetric problem, which we now describe in detail. 

960: Assume that the initial solid boundary deformation is axisymmetric:

961: $$

962:     \zeta(x,y) = \zeta(r), \qquad r = \sqrt{x^2+y^2}.

963: $$

964: The Fourier transform

965: $\mathfrak{F}[\zeta(x,y)] = \widehat{\zeta}(k,\ell)$ of an axisymmetric

966: function is also axisymmetric with respect to transformation

967: parameters, i.e.

968: $$

969:     \widehat{\zeta}(k,\ell) = \widehat{\zeta}(m), \qquad m := \sqrt{k^2+\ell^2}.

970: $$

971: In the following calculation, we use the notation $\psi = \arctan(\ell/k)$. One has

972: \begin{multline*}

973:   \widehat{\zeta}(k,\ell) = \int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

974:   \zeta(r)e^{-i(kx+\ell y)}\;dx dy = \int\limits_0^{2\pi}d\phi

975:   \int\limits_0^{\infty}\zeta(r)e^{-ir(k\cos\phi+\ell\sin\phi)}rdr =

976:   \\=\int\limits_0^{2\pi}d\phi\int\limits_0^{\infty} r\zeta(r)

977:   e^{-irm\cos(\phi-\psi)}dr = \int\limits_0^{\infty} r\zeta(r)

978:   dr \int\limits_0^\pi (e^{-irm\cos\phi} +

979:   e^{irm\cos\phi})d\phi. 

980: \end{multline*}

981: Using an integral representation of Bessel functions \cite{gradshteyn} finally yields

982: $$ \widehat{\zeta}(k,\ell) = 2\pi\int\limits_0^{\infty}r\zeta(r)J_0(mr)dr \equiv

983:   \widehat{\zeta}(m). $$

984: It follows that

985: \begin{eqnarray*}

986:   \eta(r,t) & = & \frac1{(2\pi)^2} \int\limits_0^{2\pi}\;d\psi

987:   \int\limits_0^{+\infty}\frac{m e^{im r\cos(\phi-\psi)}}

988:   {\cosh(m h)}dm\int\limits_0^t(1-\omega\sin\omega\tau)

989:   \overline{\zeta}(m,t-\tau)\;d\tau \\

990:   & = & \frac1{2\pi}\int\limits_0^{+\infty}

991:   m\frac{J_0(m r)}{\cosh(m h)}dm

992:   \int\limits_0^t(1-\omega\sin\omega\tau)\overline{\zeta}(m,t-\tau)d\tau.

993: \end{eqnarray*}

994: The last equation gives the general integral solution of the problem

995: in the case of an axisymmetric seabed deformation. Below we no longer

996: make this assumption since Okada's solution does not have this

997: property.

998: 

999: In the present study we consider seabed deformations with

1000: the following structure:

1001: \begin{equation}\label{specrepr}

1002:   \zeta(x,y,t) := \zeta(x,y)T(t).

1003: \end{equation}

1004: Mathematically we separate the time dependence from the spatial coordinates.

1005: There are two main reasons for doing this. First of all we want to

1006: be able to invert analytically the Laplace transform. The

1007: second reason is more fundamental. In fact, dynamic

1008: source models are not easily available. Okada's solution, which was described in the previous section,

1009: provides the static sea-bed deformation $\zeta_0(x,y)$ and we will

1010: consider different time dependencies $T(t)$ to model the time evolution

1011: of the source. Four scenarios will be considered:

1012: \begin{enumerate}

1013:     \item {\bf Instantaneous}: $T_i(t) = H(t)$, where $H(t)$ denotes

1014:     the Heaviside step function,

1015:     \item {\bf Exponential}: $$

1016:     T_e(t) = \left\{%

1017:     \begin{array}{ll}

1018:         0, & t<0, \\

1019:         1 - e^{-\alpha t}, & t\geq 0, 

1020:     \end{array}%

1021:     \right. \quad \mbox{with} \;\; \alpha > 0,

1022:     $$

1023:     \item {\bf Trigonometric}: $T_c(t) = H(t-t_0) + \frac12[1-\cos(\pi t/t_0)]H(t_0-t),$

1024:     \item {\bf Linear}: $$

1025:       T_l(t) = \left\{%

1026: \begin{array}{ll}

1027:     0, & t<0, \\

1028:     t/t_0, & 0\leq t\leq t_0, \\

1029:     1, & t>t_0. \\

1030: \end{array}%

1031: \right.

1032:     $$

1033: \end{enumerate}

1034: \begin{figure}

1035: \centering

1036:   \includegraphics[width=0.8\linewidth]{time.eps}\\

1037:   \caption{Typical graphs of $T_e(t)$ and $T_c(t)$.

1038:   Here we have set $\alpha=6.2$, $t_0=0.7$.}

1039:   \label{fig:time}

1040: \end{figure}

1041: The typical graphs of $T_c(t)$ and $T_e(t)$ are shown in Figure \ref{fig:time}.

1042: Inserting (\ref{specrepr}) into (\ref{genintsol}) yields

1043: \begin{equation}\label{specintsol}

1044:   \eta(x,y,t) = \frac{1}{(2\pi)^2}\int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1045:   \frac{\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}{\cosh(m h)}\frac1{2\pi i}

1046:   \int\limits_{\mu-i\infty}^{\mu+i\infty}

1047:   \frac{s^2\tens{T}(s)}{s^2+\omega^2}e^{st}ds\; dk d\ell.

1048: \end{equation}

1049: 

1050: Clearly, $\eta(x,y,t)$ depends continuously on the source $\zeta(x,y)$. Physically it means that 

1051: small variations of $\zeta$ (in a reasonable space of functions such as $L^2$) yield small variations of $\eta$. 

1052: Mathematically this problem is said to be well-posed, and this property is essential for modelling the physical 

1053: processes, since it means that small modifications of the ground motion (for example, the error in measurements) 

1054: do not induce huge modifications of the wave patterns.

1055: 

1056: Using the special representation (\ref{specrepr}) of seabed

1057: deformation and prescribed time-dependencies, one can compute

1058: analytically the Laplace integral in (\ref{specintsol}). To

1059: perform this integration, we first have to compute the Laplace transform

1060: of $T_{i,e,c,l}(t)$. The results are

1061: $$

1062:   \mathfrak{L}[T_i] = \frac1s,\qquad \mathfrak{L}[T_e] =

1063:   \frac{\alpha}{s(\alpha+s)},

1064: $$

1065: $$

1066:   \mathfrak{L}[T_c] =

1067:   (1+e^{-st_0})\frac{\gamma^2}{2s(s^2+\gamma^2)} \;\; \mbox{with} \;\;

1068:   \gamma = \frac{\pi}{t_0}, \qquad

1069:   \mathfrak{L}[T_l] = \frac{1-e^{-st_0}}{t_0s^2}.

1070: $$

1071: Inserting these formulas into the inverse Laplace integral

1072: yields

1073: \begin{eqnarray*}

1074:   \frac1{2\pi i}

1075:   \int\limits_{\mu-i\infty}^{\mu+i\infty}

1076:   \frac{e^{st}s^2\tens{T_i}(s)}{s^2+\omega^2}ds & = &

1077:   \cos\omega t, \\

1078:   \frac1{2\pi i}

1079:   \int\limits_{\mu-i\infty}^{\mu+i\infty}

1080:   \frac{e^{st}s^2\tens{T_e}(s)}{s^2+\omega^2}ds & = &

1081:   -\frac{\alpha^2}{\alpha^2+\omega^2}\left(

1082:   e^{-\alpha t} - \cos\omega t - \frac{\omega}{\alpha}\sin\omega t

1083:   \right), \\

1084:   \frac1{2\pi i}

1085:   \int\limits_{\mu-i\infty}^{\mu+i\infty}

1086:   \frac{e^{st}s^2\tens{T_c}(s)}{s^2+\omega^2}ds & = & \frac{\gamma^2}{2(\gamma^2-\omega^2)} \\

1087:   & & \hspace{-0.5cm} \left(

1088:   \cos\omega t - \cos\gamma t + H(t-t_0)[

1089:   \cos\omega(t-t_0) + \cos\gamma t]\right), \\

1090:   \frac1{2\pi i}

1091:   \int\limits_{\mu-i\infty}^{\mu+i\infty}

1092:   \frac{e^{st}s^2\tens{T_l}(s)}{s^2+\omega^2}ds & = &

1093:   \frac{\sin\omega t - H(t-t_0)\sin\omega(t-t_0)}{\omega t_0}.

1094: \end{eqnarray*}

1095: 

1096: The final integral formulas for the free-surface elevations with

1097: different time dependencies are as follows:

1098: \begin{eqnarray*}

1099:   \eta_i(x,y,t) & = & \frac{1}{(2\pi)^2}\int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1100:   \frac{\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}{\cosh(m h)}\cos\omega t\; dk

1101:   d\ell, \\

1102:   \eta_e(x,y,t) & = & \frac{-\alpha^2}{(2\pi)^2}\int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1103:   \frac{\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}{\cosh(m h)}

1104:   \left(\frac{e^{-\alpha t} - \cos\omega t - \frac{\omega}{\alpha}\sin\omega t}

1105:   {\alpha^2+\omega^2}\right)\; dk d\ell,

1106: \end{eqnarray*}

1107: \begin{eqnarray*}

1108:   \eta_c(x,y,t) & = & \frac{\gamma^2}{(2\pi)^2}\int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1109:   \frac{\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}{2(\gamma^2-\omega^2)\cosh(m h)}

1110:   \\ & & \hspace{5mm} \left(\cos\omega t - \cos\gamma t+ H(t-t_0)[

1111:   \cos\omega(t-t_0) + \cos\gamma t]\right)\; dk d\ell, \\

1112: \eta_l(x,y,t) & = & \frac{1}{(2\pi)^2}\int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1113:   \frac{\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}{\cosh(m h)}

1114:   \left(\frac{\sin\omega t - H(t-t_0)\sin\omega(t-t_0)}{\omega t_0}\right)\; dk d\ell.

1115: \end{eqnarray*}

1116: 

1117: \subsection{Velocity field}

1118: In some applications it is important to know not only the

1119: free-surface elevation but also the velocity field in the fluid

1120: domain. One of the goals of this work is to provide an initial

1121: condition for tsunami propagation codes. For the time being,

1122: tsunami modelers take initial seabed deformations and translate them

1123: directly to the free surface in order to obtain the initial condition $\eta(x,y,0)$. Since a priori there is no

1124: information on the flow velocities, they take a zero velocity

1125: field as initial condition for the velocity: $\nabla\phi(x,y,z,0)=0$. The present computations show that it

1126: is indeed a very good approximation if the generation time is short.

1127: 

1128: In equation (\ref{phihat}), we obtained the Fourier transform of the velocity potential

1129: $\phi(x,y,z,t)$:

1130: \begin{equation}\label{potential}

1131:   \overline{\phi}(k,\ell,z,s) = -\frac{gs\widehat{\zeta}(k,\ell)\tens{T}(s)}

1132:   {\cosh(m h)(s^2+\omega^2)}

1133:   \left(\cosh(m z) - \frac{s^2}{gm}\sinh(m z)\right).

1134: \end{equation}

1135: 

1136: Let us evaluate the velocity field at an arbitrary level $z=\beta h$

1137: with $-1\leq\beta\leq 0$. In the linear approximation the

1138: value $\beta = 0$ corresponds to the free surface while $\beta=-1$ corresponds to the

1139: bottom. Next we introduce some notation. The horizontal

1140: velocities are denoted by $\uv$. The horizontal gradient $(\partial/\partial x,\partial/\partial y)$ is denoted by

1141: $\nabla_h$. The vertical velocity component is simply $w$. The Fourier transform parameters are denoted 

1142: $\kv = (k,\ell)$.

1143: 

1144: Taking the Fourier and Laplace transforms of 

1145: $$

1146:   \uv(x,y,t) = \left.\nabla_h \phi(x,y,z,t)\right|_{z=\beta h}

1147: $$

1148: yields

1149: \begin{eqnarray*}

1150: \overline{\uv}(k,\ell,s) & = & -i\overline{\phi}(k,\ell,\beta h, s)\kv \\

1151:  & = & i\frac{gs\widehat{\zeta}(k,\ell)\tens{T}(s)}

1152:   {\cosh(m h)(s^2+\omega^2)}\left(

1153:   \cosh(\beta mh) - \frac{s^2}{gm}\sinh(\beta mh)

1154:   \right)\kv.

1155: \end{eqnarray*}

1156: Inverting the Fourier and Laplace transforms gives the general formula for the

1157: horizontal velocities:

1158: \begin{eqnarray*}

1159:   \uv(x,y,t) & = &

1160:   \frac{ig}{4\pi^2}\int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1161:   \frac{\kv \widehat{\zeta}(k,\ell)\cosh(m\beta h) e^{i(kx+\ell y)}}{\cosh(m h)}

1162:   \frac{1}{2\pi i}\int\limits_{\mu-i\infty}^{\mu+i\infty}

1163:   \frac{s\tens{T}(s)e^{st}}{s^2+\omega^2}\;ds\; d\kv \\

1164:   & & -\frac{i}{4\pi^2}\int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1165:   \frac{\kv \widehat{\zeta}(k,\ell)\sinh(m\beta h) e^{i(kx+\ell y)}}{m\cosh(m h)}

1166:   \frac{1}{2\pi i}\int\limits_{\mu-i\infty}^{\mu+i\infty}

1167:   \frac{s^3\tens{T}(s)e^{st}}{s^2+\omega^2}\;ds\; d\kv.

1168: \end{eqnarray*}

1169: 

1170: After a few computations, one finds the formulas for the time dependencies $T_i$, $T_e$ and $T_l$. For

1171: simplicity we only give the velocities along the free surface ($\beta=0$):

1172: \begin{eqnarray*}

1173:   \uv_i(x,y,t) & = & \frac{ig}{4\pi^2}

1174:  \int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1175:   \frac{\kv\widehat{\zeta}(k,\ell) e^{i(kx+\ell y)}}{\cosh(m h)}

1176:   \frac{\sin\omega t}{\omega}\; d\kv, \\

1177: \uv_e(x,y,t) & = & \frac{ig\alpha}{4\pi^2}

1178:  \int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1179:   \frac{\kv\widehat{\zeta}(k,\ell) e^{i(kx+\ell y)}}

1180:   {(\alpha^2+\omega^2)\cosh(m h)}

1181: \left(e^{-\alpha t} - \cos\omega t + \frac{\alpha}{\omega}\sin\omega t \right) \; d\kv, \\

1182: %  \uv_c(x,y,t) & = & \frac{ig\gamma^2}{8\pi^2}

1183: %  \int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1184: %  \frac{\kv\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}

1185: %  {(\omega^2-\gamma^2)\cosh(m h)}

1186: %  \left(\frac{\sin\gamma t}{\gamma} -

1187: %  \frac{\sin\omega t}{\omega}\right)\; d\kv, \\

1188: \uv_l(x,y,t) & = & \frac{ig}{4t_0\pi^2}

1189:   \int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1190:   \frac{\kv\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}{\omega^2\cosh(m h)} \\

1191:   & & \hspace{1cm} \left(1-\cos\omega t - H(t-t_0)[1-\cos\omega(t-t_0)]\right)\;

1192:   d\kv.

1193: \end{eqnarray*}

1194: 

1195: Next we determine the vertical component of the velocity

1196: $w(x,y,z,t)$. It is easy to obtain the Fourier--Laplace transform

1197: $\overline{w}(k,\ell,z,s)$ by differentiating (\ref{potential}):

1198: $$

1199:   \overline{w}(k,\ell,z,s) = \pd{\overline{\phi}}{z} =

1200:   \frac{sg\widehat{\zeta}(k,\ell)\tens{T}(s)}

1201:   {\cosh(m h)(s^2+\omega^2)}\left(

1202:   \frac{s^2}{g}\cosh(m z) - m\sinh(m z)

1203:   \right).

1204: $$

1205: Inverting this transform yields

1206: \begin{eqnarray*}

1207:   w(x,y,z,t) & = & \frac1{4\pi^2}

1208:   \int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1209:   \frac{\cosh(m z)\widehat{\zeta}(k,\ell)}{\cosh(m h)}e^{i(kx+\ell y)}

1210:   \frac1{2\pi i}\int\limits_{\mu-i\infty}^{\mu+i\infty}

1211:   \frac{s^3 \tens{T}(s)e^{st}}{s^2+\omega^2}\;ds\;d\kv \\

1212:   & & -\frac{g}{4\pi^2}

1213:   \int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1214:   \frac{m\sinh(m z)\widehat{\zeta}(k,\ell)}{\cosh(m h)}e^{i(kx+\ell y)}

1215:   \frac1{2\pi i}\int\limits_{\mu-i\infty}^{\mu+i\infty}

1216:   \frac{s\tens{T}(s)e^{st}}{s^2+\omega^2}\;ds\;d\kv,

1217: \end{eqnarray*}

1218: for $-h <z\leq 0$. One can easily obtain the expression of the vertical velocity at a given vertical

1219: level by substituting $z=\beta h$ in the expression for $w$.

1220: 

1221: The easiest way to compute the vertical velocity $w$ along the free surface

1222: is to use the boundary condition (\ref{kinfreesurf}). Indeed,

1223: the expression for $w$ can be simply derived by differentiating the known formula for $\eta_{i,e,c,l}(x,y,t)$. 

1224: Note that formally the derivative gives the

1225: distributions $\delta(t)$ and $\delta(t-t_0)$ under the integral

1226: sign. It is a consequence of the idealized time

1227: behaviour (such as the instantaneous scenario) and 

1228: it is a disadvantage of the Laplace transform method. In order to avoid

1229: these distributions we can consider the solutions only

1230: for $t>0$ and $t\neq t_0$. From a practical point of view there is no

1231: restriction since for any $\varepsilon > 0$ we can set

1232: $t=\varepsilon$ or $t = t_0 + \varepsilon$. For small values of

1233: $\varepsilon$ this will give a very good approximation of the solution

1234: behaviour at these ``critical'' instants of time. Under this

1235: assumption we give the distribution-free expressions for the vertical velocity along

1236: the free surface:

1237: \begin{eqnarray*}

1238:   w_i(x,y,t) & = & -\frac{1}{4\pi^2}

1239:   \int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1240:   \frac{\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}

1241:   {\cosh(m h)}\omega\sin\omega t

1242:   \; d\kv, \\

1243:   w_e(x,y,t) & = & \frac{\alpha^3}{4\pi^2}

1244:   \int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1245:   \frac{\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}

1246:   {(\alpha^2+\omega^2)\cosh(m h)}

1247:   \left(

1248:   e^{-\alpha t} + \frac{\omega^2}{\alpha^2}\cos\omega t

1249:   - \frac{\omega}{\alpha}\sin\omega t\right)

1250:   \; d\kv, \\

1251:   w_c(x,y,t) & = & -\frac{\gamma^2}{4\pi^2}

1252:   \int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1253:   \frac{\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}

1254:   {2(\gamma^2-\omega^2)\cosh(m h)}

1255:   \bigl(

1256:   \omega\sin\omega t - \gamma\sin\gamma t \\

1257: & & \hspace{2cm} +H(t-t_0)[\omega\sin\omega(t-t_0) + \gamma\sin\gamma t]

1258:   \bigr)\; d\kv, \\

1259:   w_l(x,y,t) & = & \frac1{4t_0\pi^2}

1260:   \int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1261:   \frac{\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}{\cosh(m h)}

1262:   \left[\cos\omega t - H(t-t_0)\cos\omega(t-t_0)\right]\;d\kv.

1263: \end{eqnarray*}

1264: 

1265: \subsection{Pressure on the bottom}

1266: 

1267: Since tsunameters have one component that measures the pressure at the bottom (bottom pressure

1268: recorder or simply BPR \cite{Gonz}), it is interesting to 

1269: provide as well the expression $p_b(x,y,t)$ for the pressure at the bottom. The pressure $p(x,y,z,t)$ can be obtained

1270: from Bernoulli's equation, which was written explicitly for the free surface in 

1271: equation (\ref{dcl}), but is valid everywhere in the fluid:

1272: \begin{equation}

1273:   \pd{\phi}{t} + \frac12|\nabla\phi|^2 + gz + \frac{p}{\rho} = 0. 

1274: \label{bernou}

1275: \end{equation}

1276: After linearization, equation (\ref{bernou}) becomes

1277: \begin{equation}

1278:   \pd{\phi}{t} + gz + \frac{p}{\rho} = 0. 

1279: \label{bernou_l}

1280: \end{equation}

1281: Along the bottom, it reduces to

1282: \begin{equation}

1283:   \pd{\phi}{t} + g(-h+\zeta) + \frac{p_b}{\rho} = 0, \qquad z=-h. 

1284: \label{bernou_l_b}

1285: \end{equation}

1286: The time-derivative of the velocity potential is readily available in Fourier space. 

1287: Inverting the Fourier and Laplace transforms and evaluating the resulting expression 

1288: at $z=-h$ gives for the four time scenarios, respectively,

1289: \begin{eqnarray*}

1290:  \pd{\phi_i}{t} & = & -\frac{g}{(2\pi)^2}\int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1291:  \frac{\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}{\cosh^2(mh)}\cos \omega t \; d\kv, \\ 

1292:  \pd{\phi_e}{t} & = & \frac{g\alpha^2}{(2\pi)^2}\int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1293:  \frac{\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}{\alpha^2+\omega^2}

1294:  \left(e^{-\alpha t} - \cos\omega t - \frac{\omega}{\alpha}\sin\omega t\right) \; d\kv + \frac{\alpha^4}{(2\pi)^2} \\ 

1295:  & & \hspace{-0.5cm} \int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1296:  \frac{\widehat{\zeta}(k,\ell)\tanh(mh)e^{i(kx+\ell y)}}{m(\alpha^2+\omega^2)}

1297:  \left(e^{-\alpha t} + \Bigl(\frac{\omega}{\alpha}\Bigr)^2\cos\omega t + 

1298:  \Bigl(\frac{\omega}{\alpha}\Bigr)^3\sin\omega t\right)\; d\kv,

1299: \end{eqnarray*}

1300: %\pd{\phi_c}{t} & = & -\frac{g\gamma^2}{(2\pi)^2}\int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1301: % \frac{\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}{\gamma^2-\omega^2}

1302: % \\ 

1303: %& & \hspace{0.5cm} \left(\cos\omega t - \cos\gamma t + H(t-t_0)\left[\cos\omega(t-t_0)-\cos\gamma(t-t_0)\right]\right)\; d\kv \\ 

1304: % & & +\frac{\gamma^2}{(2\pi)^2}\int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1305: % \frac{\widehat{\zeta}(k,\ell)\tanh(mh)e^{i(kx+\ell y)}}{m(\gamma^2-\omega^2)}

1306: % (\omega^2\cos\omega t - \gamma^2\cos\gamma t \\ & & \hspace{1.5cm} + H(t-t_0)

1307: % \left[\omega^2\cos\omega(t-t_0)-\gamma^2\cos\gamma(t-t_0)\right])\; d\kv, \\

1308: \begin{eqnarray*}

1309: \pd{\phi_l}{t} & = & -\frac{g}{t_0(2\pi)^2}\int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1310:  \frac{\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}{\omega\cosh^2(mh)}

1311:  \left[\sin\omega t - H(t-t_0)\sin\omega(t-t_0)\right] \; d\kv.

1312: \end{eqnarray*}

1313: 

1314: The bottom pressure deviation from the hydrostatic pressure is then given by

1315: $$ p_b(x,y,t) = -\left.\rho\pd{\phi}{t}\right|_{z=-h} - \rho g\zeta. $$

1316: Plots of the bottom pressure will be given in Section 4.

1317:  

1318: \subsection{Asymptotic analysis of integral solutions}

1319: 

1320: In this subsection, we apply the method of stationary

1321: phase in order to estimate the far-field behaviour of the solutions. There is a lot of literature on 

1322: this topic (see for example \cite{erdelyi,murray,petr,BleisHandel,egorov}). This

1323: method is a classical method in asymptotic analysis. To

1324: our knowledge, the stationary phase method was first used by

1325: Kelvin \cite{kelvin} in the context of linear water-wave theory.

1326: 

1327: The motivation to obtain asymptotic formulas for integral

1328: solutions was mainly due to numerical difficulties to calculate the

1329: solutions for large values of $x$ and $y$. From equation (\ref{genintsol}), it is

1330: clear that the integrand is highly oscillatory. In order to be

1331: able to resolve these oscillations, several

1332: discretization points are needed per period. This becomes extremely expensive

1333: as $r=\sqrt{x^2+y^2}\to \infty$. The numerical method used in the

1334: present study is based on a Filon-type quadrature formula

1335: \cite{filon} and has been adapted to double integrals with $\exp[i(kx+\ell y)]$

1336: oscillations. The idea of this method consists in

1337: interpolating only the amplitude of the integrand at discretization points by

1338: some kind of polynomial or spline and then performing exact

1339: integration for the oscillating part of the integrand. This method seems

1340: to be quite efficient.

1341: 

1342: Let us first obtain an asymptotic representation for integral

1343: solutions of the general form

1344: \begin{equation}\label{trhomega}

1345:   \eta(x,y,t) = \frac1{4\pi^2}

1346:   \int\!\!\!\int\limits_{\!\!\!\!\!\R^2}

1347:   \frac{\widehat{\zeta}(k,\ell)e^{i(kx+\ell y)}}{\cosh(m h)} T(m,t) \;

1348:   dk d\ell, \;\ m = \sqrt{k^2+\ell^2}.

1349: \end{equation}

1350: Comparing with equation (\ref{specintsol}) shows that $T(m,t)$ is in fact

1351: $$

1352:   T(m,t) = \frac1{2\pi i}

1353:   \int\limits_{\mu-i\infty}^{\mu+i\infty}

1354:   \frac{s^2\tens{T}(s)}{s^2+\omega^2}e^{st}\;ds.

1355: $$

1356: For example, we showed above that for an instantaneous seabed

1357: deformation $T(m,t) = \cos\omega t$, where $\omega^2 =

1358: gm\tanh m h$. For the time being, we do not specify the

1359: time behaviour $\tens{T}(s)$.

1360: 

1361: In equation (\ref{trhomega}), we switch to polar coordinates $m$ and

1362: $\psi=\arctan(\ell/k)$:

1363: \begin{eqnarray*}

1364:   \eta(x,y,t) & = & \frac1{4\pi^2}

1365:   \int\limits_0^{\infty}

1366:   \int\limits_0^{2\pi}

1367:   \frac{\widehat{\zeta}(m,\psi)e^{im r\cos(\varphi-\psi)}}{\cosh(m h)} T(m,t)m \;

1368:   d\psi dm \\ & = &

1369:   \frac1{4\pi^2}

1370:   \int\limits_0^{\infty}

1371:   \frac{m T(m,t)}{\cosh(m h)}\;dm

1372:   \int\limits_0^{2\pi}

1373:   \widehat{\zeta}(m,\psi)e^{im r\cos(\varphi-\psi)}\;d\psi,

1374: \end{eqnarray*}

1375: where $(r,\varphi)$ are the polar coordinates of $(x,y)$. In the last

1376: expression, the phase function is $\Phi = m r\cos(\varphi-\psi)$.

1377: Stationary phase points satisfy the condition $\partial\Phi/\partial\psi = 0$,

1378: which yields two phases: $\psi_1 = \varphi$ and $\psi_2 = \varphi+\pi$.

1379: An approximation to equation (\ref{trhomega}) is then obtained by applying the

1380: method of stationary phase to the integral over $\psi$:

1381: $$

1382:   \eta(r,\phi,t) \simeq 

1383:   \frac1{\sqrt{8\pi^3r}}\int\limits_0^{\infty}

1384:   \frac{\sqrt{m}T(m,t)}{\cosh(m h)}\left(

1385:   \widehat{\zeta}(m,\varphi)e^{i(\frac{\pi}4-mr)} +

1386:   \widehat{\zeta}(m,\varphi+\pi)e^{i(m r - \frac{\pi}4)}\right)

1387:   dm.

1388: $$

1389: This expression cannot be simplified if we do not make any further hypotheses on the function $T(m,t)$.

1390: 

1391: Since we are looking for the far field solution

1392: behaviour, the details of wave formation are not important. Thus we

1393: will assume that the initial seabed deformation is instantaneous:

1394: $$

1395:   T(m,t) = \cos\omega t = \frac{e^{i\omega t} + e^{-i\omega t}}{2}.

1396: $$

1397: Inserting this particular function $T(m,t)$ in equation (\ref{trhomega}) yields

1398: $$

1399:   \eta(r,\varphi,t) = \frac{1}{8\pi^2}\bigl(I_1 + I_2\bigr),

1400: $$

1401: where

1402: \begin{eqnarray*}

1403:   I_1 & = & \int\limits_0^{\infty}

1404:   \frac{m\widehat{\zeta}(m,\psi)}{\cosh(m h)}

1405:   \int\limits_0^{2\pi}

1406:   e^{i(\omega t + m r\cos(\varphi - \psi))}\;d\psi dm, \\

1407:   I_2 & = & \int\limits_0^{\infty}

1408:   \frac{m\widehat{\zeta}(m,\psi)}{\cosh(m h)}

1409:   \int\limits_0^{2\pi}

1410:   e^{i(-\omega t + m r\cos(\varphi - \psi))}\;d\psi dm.

1411: \end{eqnarray*}

1412: The stationary phase function in these integrals is

1413: $$

1414:   \Phi (m,\psi) = m r\cos(\varphi - \psi) \pm \omega t,

1415:   \qquad \omega^2(m) = gm\tanh mh.

1416: $$

1417: The points of stationary phase are then obtained from the conditions

1418: $$

1419:   \pd{\Phi}{\psi} = 0, \quad \pd{\Phi}{m} = 0.

1420: $$

1421: The first equation gives two points, $\psi_1 = \varphi$ and $\psi_2 =

1422: \varphi + \pi$, as before. The second condition yields

1423: \begin{equation}\label{definerho}

1424:   \frac{r}{t}\cos(\varphi-\psi_{1,2}) = \mp\od{\omega}{m}.

1425: \end{equation}

1426: Since $d\omega/dm$ decreases from $\sqrt{gh}$ to 0 as

1427: $m$ goes from 0 to $\infty$ (see Figure \ref{fig:omega}), this equation has a unique solution for $m$

1428: if $\abs{r/t} \leq \sqrt{gh}$. This unique solution will be denoted by $m^*$. 

1429: 

1430: For $\abs{r}> t\sqrt{gh}$, there is no stationary phase. It means physically that the wave has not yet reached this region.

1431: So we can approximately set $I_1 \approx 0$ and $I_2 \approx 0$. From the positivity of the function $d\omega/dm$ one can deduce 

1432: that $\psi_1 = \varphi$ is a stationary phase point only for the integral $I_2$. Similarly, $\psi_2 = \varphi + \pi$ 

1433: is a stationary point only for the integral $I_1$.

1434: 

1435: Let us obtain an asymptotic formula for the first integral:

1436: \begin{eqnarray*}

1437:   I_1 & \approx & \int\limits_0^{\infty} \frac{m}{\cosh(m h)}

1438:   \left(

1439:   \sqrt{\frac{2\pi}{m r}} \widehat\zeta(m,\varphi+\pi)

1440:   e^{i(\omega t - m r)}e^{i\frac{\pi}4}

1441:   \right)\; dm \\

1442:   & = & \sqrt{\frac{2\pi}{r}}e^{i\frac{\pi}4}

1443:   \int\limits_0^{\infty}

1444:   \frac{\widehat\zeta(m,\varphi+\pi)}{\cosh(m h)}

1445:   \sqrt{m} e^{i(\omega t -m r)} \; dm \\

1446:   & \approx &

1447:   \sqrt{\frac{2\pi}{r}}e^{i\frac{\pi}4}\left(

1448:   \sqrt{\frac{2\pi m^*}{\abs{\omega''(m^*)}t}}

1449:   \frac{\widehat\zeta(m^*,\varphi+\pi)}{\cosh(m^*h)}

1450:   e^{i(\omega(m^*)t - m^* r)}e^{-i\frac{\pi}4}\right) \\

1451:   & = & \frac{2\pi}{t}

1452:   \sqrt{\frac{m^*}{-\omega'' \omega'}}\,

1453:   \frac{\widehat\zeta(m^*,\varphi+\pi)}{\cosh(m^*h)}

1454:   e^{i(\omega(m^*)t - m^* r)}.

1455: \end{eqnarray*}

1456: In this estimate we have used 

1457: equation (\ref{definerho}) evaluated at the stationary phase point $(m^*,\psi_2)$: 

1458: \begin{equation}\label{relrt}

1459:   r = t\left.\od{\omega}{m}\right|_{m=m^*}.

1460: \end{equation}

1461: 

1462: Similarly one can obtain an estimate for the integral $I_2$:

1463: $$

1464:   I_2 \approx \frac{2\pi}{t}

1465:   \sqrt{\frac{m^*}{-\omega'' \omega'}}\,

1466:   \frac{\widehat\zeta(m^*,\varphi)}{\cosh(m^*h)}

1467:   e^{-i(\omega(m^*)t - m^* r)}.

1468: $$

1469: Asymptotic values have been obtained for the integrals. As is easily observed from the expressions for $I_1$ and $I_2$,

1470: the wave train decays as $1/t$, or $1/r$, which is equivalent since $r$ and $t$ are connected by relation (\ref{relrt}).

1471: 

1472: \section{Numerical results}

1473: A lot of numerical computations based on the analytical formulas obtained in the previous sections have been performed. 

1474: Because of the lack of information about the real dynamical characteristics of tsunami sources, we cannot really conclude 

1475: which time dependence gives the best description of tsunami generation. At this stage it is still very difficult or even 

1476: impossible. 

1477: 

1478: Numerical experiments showed that the largest wave amplitudes with the time dependence $T_c(t)$ were obtained for

1479: relatively small values of the characteristic time $t_0$. The exponential dependence has shown higher amplitudes 

1480: for relatively longer characteristic times.

1481: The instantaneous scenario $T_i$ gives at the free surface the initial seabed deformation with a 

1482: slightly lower amplitude (the factor that we obtained was typically about $0.8\sim 0.94$). The water has a high-pass filter effect

1483: on the initial solid boundary deformation. The linear time dependence $T_l(t)$ showed a linear growth of wave amplitude from 

1484: 0 to also $\approx 0.9\zeta_0$, where $\zeta_0 = \max\limits_{(x,y)\in\R^2}\abs{ \zeta(x,y)}$. 

1485: 

1486: In this section we provide several plots (Figure \ref{fig:eta0}) of the free-surface deformation. 

1487: For illustration purposes, we have chosen the instantaneous seabed deformation since it is the most widely used. 

1488: The values of the parameters used in the computations are given in Table \ref{parameters}. We also give plots 

1489:  of the velocity components on the free surface a few seconds (physical) after the instantaneous deformation

1490: (Figure \ref{fig:uvw}). Finally, plots of the bottom dynamic pressure are given in Figure \ref{fig:press}.

1491: 

1492: \begin{table}

1493: \begin{center}

1494: \begin{tabular}{lc}

1495: \hline

1496: \bf{Parameter} & \bf{Value} \\

1497: \hline

1498: Young modulus, $E$, GPa & 9.5 \\

1499: Poisson ratio, $\nu$ & 0.27 \\

1500: Fault depth, $d$, km & 20 \\

1501: Dip angle, $\delta$, $^\circ$ & 13 \\

1502: Strike angle, $\theta$, $^\circ$ & 90 \\

1503: Normal angle, $\phi$, $^\circ$ & 0 \\

1504: Fault length, $L$, km & 60 \\

1505: Fault width, $W$, km & 40 \\

1506: Burger's vector length, $\abs{\Dv}$, m & 15 \\

1507: Water depth, $h$, km & 4 \\

1508: Acceleration due to gravity, $g$, $m/s^2$ & 9.8 \\

1509: Wave number, $k$, $1/m$ & $10^{-4}$ \\

1510: Angular frequency, $\omega$, Hz & $10^{-2}$ \\

1511: \hline

1512: \end{tabular}

1513: \caption{Physical parameters used in the numerical computations}

1514: \label{parameters}

1515: \end{center}

1516: \end{table}

1517: \begin{figure}

1518:   \includegraphics[width=1.05\linewidth]{eta4.eps}

1519:   \caption{Free-surface elevation at $t=0.01, 0.6, 3, 5$ in dimensionless time. In physical time it corresponds to one second,

1520: one minute, five minutes and eight minutes and a half after the initial seabed deformation.}

1521:   \label{fig:eta0}

1522: \end{figure}

1523: 

1524: \begin{figure}

1525:   \includegraphics[width=1.05\linewidth]{velocities.eps}

1526:   \caption{Components $u$, $v$ and $w$ of the velocity field computed along the free surface at $t=0.01$, that is one

1527: second after the initial seabed deformation.}

1528:   \label{fig:uvw}

1529: \end{figure}

1530: 

1531: From Figure \ref{fig:uvw} it is clear that the velocity field is really negligible 

1532: in the beginning of wave formation. Numerical computations showed that this situation does not change 

1533: if one takes other time-dependences.

1534: 

1535: \begin{figure}

1536:   \includegraphics[width=1.05\linewidth]{press4new.eps}

1537:   \caption{Bottom pressure at $t=0.01, 0.6, 3, 5$ in dimensionless time. In physical time it corresponds to one second,

1538: one minute, five minutes and eight minutes and a half after the initial seabed deformation.}

1539:   \label{fig:press}

1540: \end{figure}

1541: 

1542: The main focus of the present paper is the generation of waves by a moving bottom. 

1543: The asymptotic behaviour of various sets of initial data propagating in a fluid of uniform depth has been studied

1544: in detail by Hammack and Segur \cite{Segur2,Segur3}. In particular, they showed that the behaviours for an 

1545: initial elevation wave and for an initial depression wave are different.

1546: 

1547: %\bibliographystyle{unsrt}

1548: %\bibliography{tsun}

1549: %\input{references}

1550: \begin{thebibliography}{99.}

1551: 

1552: \bibitem{todo}

1553: {Todorovska MI, Trifunac MD} (2001)

1554: {Generation of tsunamis by a slowly spreading uplift of the sea-floor}.

1555: {Soil Dynamics and Earthquake Engineering}

1556: {21}:{151--167}

1557: 

1558: \bibitem{indiens2}

1559: {Neetu S, Suresh I, Shankar R, Shankar D, Shenoi SSC, Shetye SR, Sundar D, Nagarajan B} (2005)

1560: {Comment on ``The Great Sumatra-Andaman Earthquake of 26 December 2004''}.

1561: {Science}

1562: {310}:{1431a-1431b}

1563: 

1564: \bibitem{Lay}

1565: {Lay T, Kanamori H, Ammon CJ, Nettles M, Ward SN, Aster RC,

1566: Beck SL, Bilek SL, Brudzinski MR, Butler R, DeShon HR, Ekstrom G,

1567: Satake K, Sipkin S} (2005)

1568: {The great Sumatra-Andaman earthquake of 26 December 2004}.

1569: Science 308:1127--1133

1570: 

1571: \bibitem{KdV}

1572: {Korteweg DJ, de Vries G} (1895)

1573: {On the change of form of long waves advancing in a rectangular canal, and on a new type of

1574:   long stationary waves}.

1575: {Phil. Mag.} {39}:{422--443}

1576: 

1577: \bibitem{bouss}

1578: {Boussinesq MJ} (1871)

1579: {Th\'eorie de l'intumescence liquide appel\'ee

1580:   onde solitaire ou de translation se propageant dans un canal

1581:   rectangulaire}.

1582: {C.R. Acad. Sci. Paris} {72}:{755--759}

1583: 

1584: \bibitem{peregr}

1585: {Peregrine DH} (1966)

1586: {Calculations of the development of an undual bore}.

1587: {J Fluid Mech} {25}:{321--330}

1588: 

1589: \bibitem{bona}

1590: {Benjamin TB, Bona JL, Mahony JJ} (1972)

1591: {Model equations for long waves in nonlinear dispersive systems}.

1592: {Philos. Trans. Royal Soc. London Ser. A} {272}:{47--78}

1593: 

1594: \bibitem{podyapolsk1} 

1595: {Podyapolsky GS} (1968)

1596: {The generation of linear gravitational waves in the ocean by seismic sources in the crust}.

1597: {Izvestiya, Earth Physics, Akademia Nauk SSSR} {1}:{4--12},

1598: {in Russian}

1599: 

1600: \bibitem{kajiura}

1601: {Kajiura K} (1963)

1602: {The leading wave of tsunami}.

1603: {Bull. Earthquake Res. Inst., Tokyo Univ.} {41}:{535--571}

1604: 

1605: \bibitem{gusyakov}

1606: {Gusyakov VK} (1972)

1607: {Generation of tsunami waves and ocean Rayleigh waves by submarine earthquakes}.

1608: In: {Mathematical problems of geophysics}, vol {3}, pages {250--272},

1609: {Novosibirsk, VZ SO AN SSSR}, {in Russian}

1610: 

1611: \bibitem{aleksgus}

1612: {Alekseev AS, Gusyakov VK} (1973)

1613: {Numerical modelling of tsunami and seismo-acoustic

1614:   waves generation by submarine earthquakes}. In:

1615: {Theory of diffraction and wave propagation}, vol 2, pages {194--197}, {Moscow-Erevan}, {in Russian}

1616: 

1617: \bibitem{gusyakov3}

1618: {Gusyakov VK} (1976)

1619: {Estimation of tsunami energy}. In:

1620: {Ill-posed problems of mathematical physics and problems of interpretation of geophysical observations},

1621: pages {46--64}, {Novosibirsk, VZ SO AN SSSR}, {in Russian}

1622: 

1623: \bibitem{carrier}

1624: {Carrier GF} (1971)

1625: {The dynamics of tsunamis}. In:

1626: {Mathematical Problems in the Geophysical Sciences}, Lectures in Applied Mathematics, vol 13,

1627: pages {157--187}, {American Mathematical Society}, in Russian

1628: 

1629: \bibitem{driess}

1630: {van den Driessche P, Braddock RD} (1972)

1631: {On the elliptic generating region of a tsunami}.

1632: {J. Mar. Res.} {30}:{217--226}

1633: 

1634: \bibitem{bradd}

1635: {Braddock RD, van den Driessche P, Peady GW} (1973)

1636: {Tsunami generation}.

1637: {J Fluid Mech} {59}:{817--828}

1638: 

1639: \bibitem{sabatier}

1640: {Sabatier P} (1986)

1641: {Formation of waves by ground motion}. In: {Encyclopedia of Fluid Mechanics},

1642: pages {723--759}, {Gulf Publishing Company}

1643: 

1644: \bibitem{Hammack}

1645: Hammack JL (1973)

1646: A note on tsunamis: their generation and propagation in an ocean of uniform depth.

1647: J Fluid Mech 60:769--799

1648: 

1649: \bibitem{todo2}

1650: {Todorovska MI, Hayir A, Trifunac MD} (2002)

1651: {A note on tsunami amplitudes above submarine slides and slumps}.

1652: {Soil Dynamics and Earthquake Engineering} {22}:{129--141}

1653: 

1654: \bibitem{keller}

1655: {Keller JB} (1961)

1656: {Tsunamis: water waves produced by earthquakes}. In: {Proceedings of the Conference on Tsunami Hydrodynamics} 24,

1657: pages {154--166}, {Institute of Geophysics, University of Hawaii}

1658:   

1659: \bibitem{filon}

1660: {Filon LNG} (1928)

1661: {On a quadrature formula for trigonometric integrals}.

1662: {Proc. Royal Soc. Edinburgh} {49}:{38--47}

1663: 

1664: \bibitem{ursell}

1665: {Ursell F} (1953)

1666: {The long-wave paradox in the theory of gravity waves}.

1667: {Proc. Camb. Phil. Soc.} {49}:{685--694}

1668: 

1669: \bibitem{Okada85}

1670: Okada Y (1985)

1671: Surface deformation due to shear and tensile faults in a half-space.

1672: Bull. Seism. Soc. Am. 75:1135--1154

1673: 

1674: \bibitem{stek2}

1675: {Steketee JA} (1958)

1676: {On Volterra's dislocation in a semi-infinite elastic medium}.

1677: {Can. J. Phys.} {36}:192--205

1678: 

1679: \bibitem{ben1}

1680: {Ben-Menahem A, Singh SJ, Solomon F} (1969)

1681: {Static deformation of a spherical earth model by internal dislocations}.

1682: {Bull. Seism. Soc. Am.} {59}:{813--853}

1683: 

1684: \bibitem{ben2}

1685: {Ben-Mehanem A, Singh SJ, Solomon F} (1970)

1686: {Deformation of an homogeneous earth model finite by dislocations}.

1687: {Rev. Geophys. Space Phys.} {8}:{591--632}

1688: 

1689: \bibitem{smylie}

1690: {Smylie DE, Mansinha L} (1971)

1691: {The elasticity theory of dislocations in real earth models and changes

1692: in the rotation of the earth}.

1693: {Geophys. J. Royal Astr. Soc.} {23}:{329--354}

1694: 

1695: \bibitem{tim}

1696: {Masterlark, T} (2003)

1697: {Finite element model predictions of static

1698: deformation from dislocation sources in a subduction zone:

1699: Sensivities to homogeneous, isotropic, Poisson-solid, and half-space

1700: assumptions}.

1701: {J. Geophys. Res.} {108}(B11):2540

1702: 

1703: \bibitem{volt}

1704: {Volterra V} (1907)

1705: {Sur l'\'equilibre des corps \'elastiques multiplement connexes}.

1706: {Annales Scientifiques de l'Ecole Normale Sup\'erieure} {24}(3):{401--517}

1707: 

1708: \bibitem{love}

1709: {Love AEH} (1944)

1710: {A treatise on the mathematical theory of elasticity}.

1711: {Dover Publications, New York}

1712: 

1713: \bibitem{maru}

1714: {Maruyama T} (1964)

1715: {Static elastic dislocations in an infinite and semi-infinite medium}.

1716: {Bull. Earthquake Res. Inst., Tokyo Univ.} {42}:{289--368}

1717: 

1718: \bibitem{mindl1}

1719: {Mindlin, RD} (1936)

1720: {Force at a point in the interior of a semi-infinite medium}.

1721: {Physics} {7}:{195--202}

1722: 

1723: \bibitem{mindl2}

1724: {Mindlin RD, Cheng DH} (1950)

1725: {Nuclei of strain in the semi-infinite solid}.

1726: {J. Appl. Phys.} {21}:{926--930}

1727: 

1728: \bibitem{wester}

1729: {Westergaard HM} (1935)

1730: {Bull. Amer. Math. Soc.} {41}:{695}

1731: 

1732: \bibitem{press}

1733: {Press F} (1965)

1734: {Displacements, strains and tilts at tele-seismic distances}.

1735: {J. Geophys. Res.} {70}:{2395--2412}

1736: 

1737: \bibitem{okada92}

1738: {Okada Y} (1992)

1739: {Internal deformation due to shear and tensile faults in a half-space}.

1740: {Bull. Seism. Soc. Am.} {82}:{1018--1040}

1741: 

1742: \bibitem{chin}

1743: {Chinnery MA} (1963)

1744: {The stress changes that accompany strike-slip faulting}.

1745: {Bull. Seism. Soc. Am.} {53}:{921--932}

1746: 

1747: \bibitem{sat74}

1748: {Sato R, Matsu'ura M} (1974)

1749: {Strains and tilts on the surface of a semi-infinite medium}.

1750: {J. Phys. Earth} {22}:{213--221}

1751: 

1752: \bibitem{iwa79}

1753: {Iwasaki T, Sato R} (1979)

1754: {Strain field in a semi-infinite medium due to an inclined rectangular fault}.

1755: {J. Phys. Earth} {27}:{285--314}

1756: 

1757: \bibitem{gradshteyn}

1758: {Gradshteyn IS, Ryzhik M} (2000)

1759: {Tables of Integrals, Series, and Products}, 6th edition,

1760: Academic Press, Orlando, Florida

1761: 

1762: \bibitem{Gonz}

1763: {Gonz\'alez FI, Bernard EN, Meinig C, Eble MC, Mofjeld HO, Stalin S} (2005)

1764: {The NTHMP tsunameter network}. 

1765: {Natural Hazards} {35}:{25--39}

1766: 

1767: \bibitem{erdelyi}

1768: {Erd\'elyi A} (1956)

1769: {Asymptotic Expansions},

1770: {Dover Publications}

1771: 

1772: \bibitem{murray}

1773: {Murray JD} (1984)

1774: {Asymptotic Analysis}, {Springer}

1775: 

1776: \bibitem{petr}

1777: {Petrashen' GI, Latyshev KP} (1971)

1778: {Asymptotic Methods and Stochastic

1779:   Models in Problems of Wave Propagation},

1780:  {American Mathematical Society}

1781: 

1782: \bibitem{BleisHandel}

1783: {Bleistein N, Handelsman RA} (1986)

1784: {Asymptotic Expansions of Integrals},

1785: {Dover Publications}

1786: 

1787: \bibitem{egorov}

1788: {Egorov YuV, Shubin MA} (1994)

1789: {Elements of the Modern Theory. Equations with Constant Coefficients}. In:

1790: {Partial Differential Equations}, {Encyclopedia of Mathematical Sciences}, vol 2. {Springer}

1791: 

1792: \bibitem{kelvin}

1793: {Kelvin Lord (W. Thomson)} (1887)

1794: {On the waves produced by a single impulse in water of any depth, or in a dispersive medium}.

1795: {Phil. Mag.} {23}({5}):{252--255}

1796: 

1797: \bibitem{Segur2}

1798: Hammack JL, Segur H (1974)

1799: The Korteweg--de Vries equation and water waves. Part 2. Comparison with experiments.

1800: J Fluid Mech 65:289--314

1801: 

1802: \bibitem{Segur3}

1803: Hammack JL, Segur H (1978)

1804: The Korteweg--de Vries equation and water waves. Part 3. Oscillatory waves.

1805: J Fluid Mech 84:337--358

1806: 

1807: \end{thebibliography}

1808: 

1809: \end{document}