1: \documentclass[12pt]{iopart}

2: %\usepackage{epsf}

3: \usepackage{graphics} 

4: \begin{document}

5: 

6: 

7: \def\om0{{\bf\Omega}}

8: \def\g0{\bi{g}}

9: \def\spd#1{\left<#1\right>_f}

10: 

11: \title[Geophysical studies...]{Geophysical studies with

12: laser-beam detectors of gravitational waves}

13: 

14: \author{ L.Grishchuk\dag\ddag\ , V.Kulagin\ddag\ ,

15: V.Rudenko\ddag\ , A.Serdobolski\ddag}

16: 

17: \address{\dag\ School of Physics and Astronomy, Cardiff University, 

18: Cardiff CF24 3YB, United~Kingdom}

19: 

20: \address{\ddag\ Sternberg Astronomical Institute, Moscow State University,

21: Moscow, 119899, Russia}

22: 

23: \ead{grishchuk@astro.cf.ac.uk}

24: 

25: \begin{abstract}

26: The existing high technology laser-beam detectors of gravitational waves 

27: may find very useful applications in an unexpected area - geophysics. 

28: To make possible the detection of weak gravitational waves in the region of 

29: high frequencies of astrophysical interest, $\sim 30 - 10^3~ Hz$, control 

30: systems of laser interferometers must permanently monitor, record and 

31: compensate much larger external interventions that take place in the region

32: of low frequencies of geophysical interest, $\sim 10^{-5} - 3 \times~10^{-3}~

33: Hz$. Such phenomena as tidal perturbations of land and gravity, normal mode 

34: oscillations of Earth, oscillations of the inner core of Earth, etc. will 

35: inevitably affect the performance of the interferometers and, therefore, the 

36: information about them will be stored in the data of control systems. We 

37: specifically identify the low-frequency information contained in distances 

38: between the interferometer mirrors (deformation of Earth) and angles between 

39: the mirrors' suspensions (deviations of local gravity vectors and plumb lines). 

40: We show that the access to the angular information may require some modest 

41: amendments to the optical scheme of the interferometers, and we suggest the 

42: ways of doing that. The detailed evaluation of environmental and instrumental 

43: noises indicates that they will not prevent, even if only marginally, the 

44: detection of interesting geophysical phenomena. Gravitational-wave instruments 

45: seem to be capable of reaching, as a by-product of their continuous

46: operation, very ambitious geophysical goals, such as observation of the 

47: Earth's inner core oscillations.

48: \end{abstract}

49: 

50: \pacs{0480N, 0710F, 0150P, 0340K}

51: 

52: \submitto{\CQG}

53:  

54: \section{Introduction}

55: 

56: Laser-beam detectors of gravitational waves are designed to explore 

57: the Universe in a new type of radiation and from a 

58: new perspective. With the already operating instruments, 

59: and coming soon online, we are expecting to witness 

60: the discovery of fascinating physics involved in powerful sources of 

61: cosmic gravitational radiation. The astrophysical aims of the 

62: gravitational wave (g.w.) science are well understood and 

63: comprehensively described in the literature (for recent reviews, 

64: see for example \cite{thorne, gr5, cthorne, gr}). Naturally,

65: g.w. community is focused on the ambitious target of `reaching for the 

66: black holes'. But what about a more modest goal of looking inside our 

67: own planet~? No doubt, black holes are fascinating and scientifically 

68: important objects. But it is also necessary to remember that, say, the 

69: still poorly understood and unpredictable earthquakes are claiming 

70: thousands of human lives per year. Is it possible that the 

71: cutting-edge technology of g.w. interferometers \cite{ligoweb, GEOweb,

72: virgoweb, TAMAweb} may help us with accurate geophysical studies, 

73: as a by-product of the continuous search for astrophysical 

74: gravitational waves ? 

75: 

76: This is the major question of the present work, and the 

77: answer is positive. We shall show in this paper 

78: that a laser-beam detector of gravitational waves is in fact 

79: automatically a valuable geophysical device. Without collecting, 

80: recording and processing environmental information of geophysical

81: origin, a g.w. laser interferometer would simply be incapable of 

82: working as a sensitive astrophysical instrument. Certainly, we have to 

83: make sure that the extraction of geophysical information 

84: does not compromise the astrophysical aims of the instrument.

85:  

86: A laser-beam detector of gravitational waves is a conceptually simple

87: installation. Each arm of the interferometer, typically of the

88: length $L= 3~km$, consists of two mirrors, and the distance variation 

89: between the mirrors is monitored by the laser beam. The mirrors

90: are hanging on wires in supporting towers. Each mirror is  

91: essentially a mass element of the pendulum placed in the local 

92: gravitational field of the Earth. The eigen-frequency of the pendulum is

93: normally in the range of $0.1 - 1~Hz$. The multi-stage pendulum system 

94: shields the mirrors from large uncontrollable displacements of the 

95: tops of the supporting towers. This makes the interferometer capable of 

96: measuring, in the region of relatively high frequencies of 

97: astrophysical interest $\sim 30 - 10^3~ Hz$, the incredibly small 

98: variations of distance between the mirrors, at the level of 

99: $10^{-16}~ cm$. The expected cause of these variations is the 

100: incoming astrophysical gravitational wave. 

101: 

102: The isolation from noises in the region of relatively high

103: frequencies is only a part of the story. To ensure successful 

104: performance of the interferometer as an astrophysical instrument, 

105: control systems of the interferometer should also register and 

106: compensate for large external interventions of geophysical origin that 

107: take place in the region of relatively low frequencies. For example, the 

108: tidal half-daily variations of distance beween towers separated by $3~ km$ 

109: are typically at the level of $10^{-2}~ cm$. This change of distance is 

110: 14 orders of magnitude larger than the anticipated astrophysical signal. 

111: If this sort of variations were allowed to affect distance between 

112: the mirrors, the interferometer would not be in the `locked' state,

113: and hence it would not be able to operate as astrophysical instrument. 

114: The `locking' of the interferometer requires that the distance between

115: the mirrors is maintained unchanged with accuracy of approximately one 

116: hundredth of the laser light wavelength, which amounts to 

117: $10^{-6}~ cm$ and less. This means that the low-frequency variations 

118: of distance between the mirrors should be monitored and largely 

119: removed by a control system called the adjustment system. 

120: A similar monitoring and compensation should be done with respect 

121: to low-frequency variations of the angle between the 

122: interferometer mirrors. This is being done by a control system 

123: called the alighnment system. Ideally, in order to reach the 

124: astrophysical goals, control systems should keep the interferometer 

125: in the working condition for the duration of time 

126: exceeding many months. 

127: 

128: Thus, the collected and recorded low-frequency information, 

129: which is vital for maintaining the operational state 

130: of the laser-beam detector of gravitational waves, 

131: inevitably makes the g.w. detector also a geophysical 

132: instrument. The time-scales of geophysical processes, some of 

133: which are believed to be crucial for global geodynamics, lie in the 

134: range from several minutes to several hours. In other words, we will be

135: interested in frequencies, which we call geophysical frequencies, 

136: somewhere in the interval $10^{-5} - 3 \times 10^{-3}~Hz$. 

137: 

138: In Sec.2 we consider a simple model of g.w. interferometer as

139: a geophysical instrument. It is assumed that the mirrors' suspension points 

140: can move and the plumb lines of hanging mirrors can vary. These changes

141: arise as a result of the Earth surface deformations and variations

142: of local gravitational field caused, for example, by the internal 

143: Earth dynamics. We derive general formulas for the distance between the

144: mirrors and the angles between the local plumb lines. Ideally, these are

145: two variables that are supposed to be monitored by, respectively, the 

146: adjustment and alignment systems of the interferometer.

147: 

148: In Sec.3 we consider a number of interesting geophysical phenomena which

149: inevitably affect the performance of a g.w. interferometer. The 

150: signatures of these phenomena are contained in the outcomes of the 

151: adjustment and alignment control systems. The geophysical effects to be 

152: studied include tidal perturbations, normal modes of Earth oscillations,

153: movements of the inner solid core of Earth, etc. We place the main

154: emphasis on the fascinating phenomenon of the inner core oscillations.

155: We estimate the useful geophysical signal accompanying this phenomenon, 

156: which will manifest itself in the variation of distance between the 

157: mirrors and in the variation of angle between the plumb lines of the

158: hanging mirrors. The guidance for the expected amplitude of the signal 

159: is provided by the reported in the literature indications that the 

160: inner core oscillations have been actually detected by other, traditional, 

161: methods. From the requirement that the signal to noise

162: ratio should be larger than 1, we define the level of tolerable 

163: noise in the proposed measurements. Specifically, the tolerable noise 

164: allows the detection of the useful signal, if the observation time 

165: exceeds 70, or so, inner core oscillation periods. 

166: 

167: A useful signal can be detected if the environmental and instrumental

168: noises are smaller than the calculated level of tolerable noise. In Sec.4 

169: we consider noises which we find most dangerous. We explicitely show 

170: that seismic, atmospheric and instrumental noises should not 

171: be capable of preventing the detection of inner core oscillations, 

172: even if only marginally. This refers both to distance and angle measurements.

173: There exists, however, a specific problem with the angle measurements,

174: related to the fact that the presently operating alignment systems 

175: are subject to a certain degeneracy. They cannot tell apart a tilt of the

176: mirror, which we are mostly interested in, and a latteral shift of the

177: mirror, which can be caused by a dull deformational noise. We analyze

178: this difficulty in great detail in Sec.5.

179: 

180: Since the angular measurements provide an important additional channel 

181: of geophysical information, we adress the problem of degeneracy in 

182: Sec.6 and suggest the ways of its circumvention. The desire to

183: keep the angular channel useful for geophysical applications may require 

184: some modest and harmless modifications of the optical scheme of g.w.

185: interferometers. We discuss at some length a few ideas with regard

186: to such modifications. It appears that, without interfering with

187: the astrophysical program of the instrument, certain geophysical

188: modifications are feasible.

189: 

190: In Sec.7 we emphasize some conclusions of the paper.

191: 

192: \section{Laser-beam detector of gravitational waves as a 

193: geophysical instrument}

194: 

195: To see better how a g.w. interferometer can work as a 

196: geophysical instrument, we will have to consider idealized, but 

197: representative, models. We start with a model of a pendulum, 

198: whose suspension point oscillates \cite{4a}, and which is placed in

199: a variable gravitational field. 

200: 

201: A pendulum of mass $m$ and length $l$ can oscillate in the

202: $x, y$ plane, see Fig.\ref{f0}. Its suspension point has 

203: time-dependent coordinates $x_0(t)$, $y_0(t)$, and the local 

204: gravity acceleration vector ${\g0}(t)$ is also a function of 

205: time: ${\g0}(t)=[g_x(t),\ g_0+g_y(t)]$. The local gravity 

206: field is spatially homogeneous on the scale of small oscillations 

207: of the mass, but it is not homogeneous on larger scales. Later, we 

208: will take into account its inhomogeneity (spherical symmetry) on the 

209: scale of two widely separated pendula and even on the scale of 

210: displacements of the suspension point of an individual pendulum.

211: 

212: The only dynamical variable in this problem is the angle $\alpha(t)$ 

213: between the suspension wire and the axis $y$. The coordinates of 

214: the mass are given by

215: 

216: \begin{equation}

217: \label{eq0}

218: x(t)=x_0(t)+l\sin\alpha(t),\ \ y(t)=y_0(t)+l\cos\alpha(t).

219: \end{equation}

220: 

221: \begin{figure}

222:   \begin{center}

223: \includegraphics{dyneq.eps}

224: %\epsfbox{dyneq.eps}

225: %\psfig{figure=dyneq.eps}

226:   \end{center}

227:   \caption{\label{f0} A pendulum with a moving suspension point

228:   placed in a variable gravitational field.}

229: \end{figure}

230: 

231: \noindent

232: The Lagrangian of the system can be written as

233: \begin{equation}

234: L=\frac{m}{2}(\dot x^2+\dot y^2)+mg_xx+m(g_0+g_y)y

235: \end{equation}

236: 

237: \noindent

238: Using expressions (\ref{eq0}) and ignoring non-dynamical

239: terms and total derivatives, one transforms the Lagrangian

240: to the final form: 

241: 

242: \begin{equation}

243: \label{finL} 

244: L=\frac{1}{2}ml^2\dot \alpha^2-ml(\ddot x_0\sin\alpha+\ddot

245: y_0\cos\alpha) +mg_x l\sin \alpha +m(g_0+g_y) l\cos \alpha.

246: \end{equation}

247: 

248: Now one can write down the equation of motion of the pendulum:  

249: \begin{equation}\label{dyneq0}

250: l\ddot\alpha=-\ddot x_0\cos\alpha+\ddot y_0\sin\alpha+g_x\cos\alpha

251: -(g_0+g_y)\sin\alpha.

252: \end{equation}

253: We assume that the time-dependent components of the gravity 

254: vector are very small, $g_x(t)\ll g_0$, $g_y(t)\ll g_0$, 

255: and the acceleration of the suspension point $\ddot y_0$ is much 

256: smaller than the unperturbed gravity acceleration $g_0$. 

257: We also assume that $\alpha\ll 1$, $\sin\alpha\simeq\alpha$,

258: $\cos\alpha\simeq 1$. Then, Eq. (\ref{dyneq0}) simplifies:

259: \begin{equation} \label{2}

260: \ddot\alpha+\omega_p^2\alpha=-\frac{\ddot x_0}l+\omega_p^2\alpha_0~,

261: \end{equation}

262: where $\omega^2_p=g_0/l$ is the square of oscillation frequency of 

263: the unperturbed pendulum. The angular variation of the 

264: local gravity vector $\g0$, that is, the angular deviation of the local

265: plumb line, is denoted $\alpha_0(t)=g_x(t)/g_0$. We see that, 

266: in the linear approximation in terms

267: of $\alpha$, the pendulum behaviour is affected only by changes in 

268: the gravity force direction, namely, by the component $g_x$ of the

269: gravity acceleration. The variations of the absolute 

270: value $|\g0|$ and the component $g_y$ of the acceleration 

271: produce smaller effects, which we neglect.

272: 

273: Equation (\ref{2}) is an inhomogeneous linear differential equation 

274: with constant coefficients. In the frequency domain, the relevant solution

275: can be written

276: \begin{equation}

277: \alpha_{\omega}= \frac{\omega^2\,x_{0\omega} + l\omega_p^2

278: \alpha_{0\omega}}{l\,(\omega_p^2-\omega^2)}, 

279: \label{3}

280: \end{equation}

281: where $\alpha_{\omega},\, x_{0\omega},\,\alpha_{0\omega}$ are 

282: complex Fourier amplitudes of the corresponding variables at 

283: frequency $\omega$. The divergent amplitude at the resonance

284: frequency $\omega = \omega_p$ will be tempered by friction, which

285: we have ignored. The regions of our interest are 

286: relatively high frequencies $\omega \gg\omega_p$ for

287: astrophysical applications and relatively low frequencies 

288: $\omega \ll \omega_p$ for geophysical applications. For orientation,

289: one can think of $\omega_p/2 \pi = f_p \approx (0.1 - 1)~ Hz$. This is 

290: true for the existing instruments.  

291: 

292: {\it a) High frequencies $\omega \gg

293: \omega_p$ -- astrophysical applications}. \\

294: The pendulum angular amplitude is given by 

295: \begin{equation} \label{4a}

296: \alpha_{\omega}\simeq \left[\left(\frac{\omega_p}{\omega}

297: \right)^2-1 \right]^{-1} \frac{x_{0\omega}}{l}- \left(\frac{\omega_p}

298: {\omega}\right)^2 \alpha_{0\omega} \simeq

299: - \frac{x_{0\omega}}{l}- \left(\frac{\omega_p}

300: {\omega}\right)^2 \alpha_{0\omega} ~.  

301: \end{equation}

302: The horizontal position of the pendulum mass is given by  

303: \begin{eqnarray} 

304: \label{5}

305: x_{\omega} & \simeq & x_{0\omega}+ l\alpha_{\omega} \simeq 

306: x_{0\omega} \left[1 + \frac{1}{(\omega_p/\omega)^2 - 1}\right] -

307: l \alpha_{0\omega}(\omega_p/\omega)^2 \nonumber \\ 

308: & \simeq & -\left(\frac{\omega_p}{\omega}\right)^2 

309: (x_{0\omega}+ l\alpha_{0\omega}) 

310:  =  -\left(\frac{\omega_p}{\omega}\right)^2 x_{0\omega}- 

311: \frac{g_{x\omega}}{\omega^2}. 

312: \end{eqnarray}

313: Eq.(\ref{5}) illustrates the well known method of vibrational 

314: isolation of interferometer's test masses-mirrors. The residual 

315: displacement $x_{\omega}$ of the pendulum mass is a factor 

316: $(\omega_{p}/\omega)^2$ smaller than the displacement 

317: $x_{0\omega}$ of the suspension point. The $N$ stages of isolation

318: reduce the external $x_{0\omega}$ by a factor $(\omega_{p}/\omega)^{2N}$.

319: This makes the mirror shielded from large deformational shifts 

320: of the suspension point, and therefore the interferometer

321: becomes capable of detecting increadibly small distance variations 

322: caused by astrophysical signals. 

323: However, the displacement caused by the local gravitational

324: environment (the last term in Eq.(\ref{5})), cannot be shielded.

325: If this gravitational noise is present, it creates certain 

326: problems for the astrophysical program, which we will return to later.

327: 

328: {\it b) Low frequencies $\omega \ll \omega_p$

329: -- geophysical applications }. \\

330: The pendulum angular amplitude is given by 

331: \begin{equation} 

332: \label{4}

333: \alpha_{\omega}\simeq \left(\frac{\omega}{\omega_p}\right)^2

334: \frac{x_{0\omega}}{l}+ \alpha_{0\omega}. 

335: \end{equation}

336: The contribution to $\alpha_{\omega}$ provided by movements 

337: of the suspension point (first term in Eq.(\ref{4})) is

338: suppressed by a factor $(\omega/\omega_p)^2$. At sufficiently

339: low frequencies, this term can become smaller than  

340: the second term in Eq.(\ref{4}). One can say that,

341: in this regime, the equilibrium position of the pendulum follows 

342: a new plumb line $\alpha_{0\omega}$ determined by the 

343: slow variation $g_x(t)$ of the local gravity force vector $\g0$.

344: 

345: The horizontal position of the pendulum mass, in the low-frequency

346: approximation, is given by 

347: \begin{eqnarray} \label{5new}

348: x_{\omega}\simeq x_{0\omega}+ l\alpha_{\omega} \simeq 

349: x_{0\omega} +\left(\frac{\omega}{\omega_p}\right)^2 x_{0\omega}+ 

350: l\alpha_{0\omega} \simeq x_{0\omega} + l\alpha_{0\omega}~.

351: \end{eqnarray}

352: It follows from Eq.(\ref{5new}) that the movements of the suspension

353: point transmit practically without change into the movements of the 

354: pendulum mass. In other words, the low-frequency deformations cannot

355: be shielded, so they should be monitored and compensated by the 

356: control systems.  

357: 

358: The interferometer arm includes two separated pendula, and 

359: it is the difference of displacements and difference of mirrors'

360: inclinations that are actually being monitored. We assume that the 

361: pendula are identical, but their environments are different,

362: see Fig.\ref{f1}. 

363: \begin{figure}

364:   \begin{center}

365: \includegraphics{fig1.eps}

366: %\epsfbox{fig1.eps}

367:   \end{center}

368:   \caption{\label{f1} Relative variations of plumb lines and mass`

369: positions in an interferometer.}

370: \end{figure}

371: We denote by symbol $\delta$ the variations at each site

372: and by symbol $\Delta$ the difference of variations at two sites.

373: It is important to know how different conditions are at two sites.

374: If the characteristic spatial scale $\lambda$ of perturbations is 

375: shorter than $L$, then the conditions at two sites are not correlated,

376: and $\Delta$-variations are typically of the same order of magnitude 

377: as the largest of individual $\delta$-variations. However, if both sites 

378: are covered by a long-wavelength perturbation $\lambda \gg L$, then

379: the $\Delta$-variations are much smaller than the $\delta$-variations.

380: Symbolically, we can write 

381: \begin{equation}

382: \Delta \simeq \frac{L}{\lambda} \delta,~~~~~~~~ {\rm if}

383: ~~~~~~~~ L \ll \lambda. 

384: \label{8} 

385: \end{equation}

386: 

387: Let the unperturbed coordinates of the suspension points be 

388: $x_{01},\,x_{02}$. Each 

389: point is subject to time-dependent deformational shifts 

390: $\delta{x_{01}},\,\delta{x_{02}}$. Of course, they are

391: much smaller than the unperturbed distance $L$, $L =x_{02}-x_{01}$.

392: We will also use the difference of the deformational shifts, 

393: $\Delta{x_0}(t)= \Delta L = \delta{x_{02}}(t)-\delta{x_{01}}(t)$.

394: The difference of angular variables of the two pendula is

395: $\Delta \alpha(t) = \alpha_2(t) -\alpha_1(t)$. The difference of local

396: values of $g_x$ and, hence, the angle between the local plumb lines 

397: at two sites is $\Delta \alpha_0(t)=\alpha_{02}(t) -\alpha_{01}(t)$. 

398: [Here we ignore the fact that the Earth's gravitational field is

399: centrally-symmetric, which leads to slightly different orientations

400: of the unperturbed gravity force vectors at two cites, but later 

401: we will take this fact into account.] Writing

402: equations (\ref{2}) at two sites, and taking their difference,

403: we arrive at the equation 

404: \begin{equation}

405: \ddot{\Delta \alpha}+\omega_p^{2} \Delta \alpha = 

406: -\frac{\ddot{\Delta x_0}}{l}+ \omega_{p}^{2}\Delta \alpha_0~.  

407: \label{6}

408: \end{equation}

409: Certainly, the relevant solution to this equation is given by

410: the formula similar to (\ref{3}), with obvious substitutions: 

411: \begin{equation}

412: \label{6a}

413: \Delta \alpha_{\omega}= \frac{\omega^2\,\Delta x_{0\omega} + l\omega_p^2

414: \Delta \alpha_{0\omega}}{l\,(\omega_p^2-\omega^2)}. 

415: \end{equation}

416: The variation of distance between the masses is 

417: \begin{equation}

418: \Delta d_{\omega} = \Delta{x_{0 \omega}} + l \Delta \alpha_{\omega}. 

419: \label{7}

420: \end{equation}

421: 

422: The high-frequency and low-frequency approximations to equations

423: (\ref{6a}), (\ref{7}) follow the lines of equations (\ref{4a}), 

424: (\ref{5}), (\ref{4}), (\ref{5new}). In particular, in the low-frequency

425: region, $\omega \ll \omega_p$, we have

426: \begin{equation} \label{9}

427: \Delta \alpha_{\omega}\simeq \Delta \alpha_{0\omega}+ 

428: \left(\frac{\omega}{\omega_p}\right)^2

429: \frac{\Delta x_{0\omega}}{l} 

430: \end{equation}

431: and

432: \begin{eqnarray} \label{10}

433: \Delta d_{\omega} \simeq \Delta x_{0\omega} + l \Delta \alpha_{0\omega}~.

434: \end{eqnarray}

435: It is seen from Eq.(\ref{10}) that the large deformational

436: contribution $\Delta x_{0\omega}$ leaks into the system. 

437: This is why the interferometer's adjustment system records and 

438: largely removes $\Delta x_{0\omega}$ from $\Delta d_{\omega}$ by 

439: applying force to the mirrors and mechanically adjusting their 

440: positions. (The adjustment system also tunes the frequency of the 

441: laser light). The circuits of adjustment system is one of the 

442: sources of geophysical information. 

443: 

444: The angle $\Delta \alpha_{\omega}$

445: between the suspension wires, Eq.(\ref{9}), is mostly the

446: reflection of the angle between local plumb lines, first term

447: in Eq.(\ref{9}). For sufficiently low frequencies, the second

448: term in Eq.(\ref{9}) is smaller than the first one. 

449: We assume that the angle between the wires is at the same time 

450: the angle between surfaces of the mirrors. Under the condition

451: that other contributions to Eq.(\ref{9}) can be 

452: identified and separated (one complication, related

453: to the centrally-symmetric character of the Earth's gravitational

454: field, is treated in detail in Sec.4c), the alighnment 

455: system, whose purpose is to monitor the angle between the mirrors, 

456: will be monitoring the angle $\Delta \alpha_{0\omega}$ between 

457: the plumb lines. This control system is another source of 

458: geophysical information. 

459: 

460: Variables of geophysical interest, $\Delta x_{0\omega}$ and

461: $\Delta \alpha_{0\omega}$, consist of two parts. 

462: One part is a useful geophysical signal, another 

463: part is environmental noise. In general, temporarily ignoring 

464: noises and other complications, the knowledge of the output 

465: signals $\Delta \alpha_{\omega}$ and $\Delta d_{\omega}$ allows 

466: one to solve equations (\ref{9}), (\ref{10}) with respect to 

467: $\Delta x_{0\omega}$ and $\Delta \alpha_{0\omega}$:

468: \begin{equation} \label{11}

469: \Delta x_{0\omega} \simeq \Delta d_{\omega}- l \Delta \alpha_{\omega}~,~~~~~~ 

470: \Delta \alpha_{0\omega} \simeq \Delta \alpha_{\omega}- 

471: \left(\frac{\omega}{\omega_p}\right)^2 \frac{\Delta d_{\omega}}{l}~.

472: \end{equation}

473: Ideally, this would be a reconstruction, as complete as possible,

474: of the acting geophysical perturbation. However, 

475: complications do exist, and we will perform more detailed 

476: analysis in the rest of the paper.   

477: 

478: In the end of this section we shall briefly discuss the

479: distance variations and plumb line deflections in terms of the 

480: Newtonian gravitational potential $U({\bf x},t)$ of the Earth.

481: The potential $U$ is the sum of the unperturbed potential

482: $U_0 = GM/r$ and perturbations caused by external and internal

483: gravitational fields. The external fields are mostly the 

484: tidal effects of the Moon, Sun and planets. The internal fields 

485: include such effects as normal oscillation modes of Earth and 

486: oscillations of the 

487: inner core of Earth. At the surface of an idealized rigid body, 

488: the local gravitational acceleration would be given by

489: ${\bf g} = \partial U/ \partial {\bf x}$. The true value of

490: ${\bf g}$ at the surface of real elastic Earth is somewhat 

491: different from this quantity. This happens because the 

492: gravitational field perturbations, originating from either external 

493: or internal sources, displace elements of the Earth's surface

494: in its own gravitational field and also give rise to a 

495: redistribution of the Earth's matter density.

496: In the point of observations, these changes create additional 

497: variations of the gravitational potential. In conventional geophysics 

498: these effects are taken into account by some numerical factors 

499: called the Love numbers \cite{c12}. For a perfectly rigid

500: body the Love numbers are zeros, while for a fluid body they are 

501: ones. On real deformable Earth, the Love numbers are roughly 

502: of the order of 1.  

503: 

504: The difference of variations of ${\bf g}$ at two sites separated

505: by $L$ is causing a change of distance between the sites, and

506: it also determines a relative deflection of local plumb lines. 

507: Let the common unperturbed ${\bf g}_0$ point out in the $y$ direction,

508: and the sites be located on the $x$ axis. Then, if $L$ is small

509: in comparison with the characteristic length $\lambda$ at which the

510: perturbation of the gravitational potential changes, one can derive: 

511: \begin{equation}

512: \Delta \alpha_0 = \frac{L}{g_0} 

513: \sqrt{\left( \frac{\partial g_x}{\partial x} \right)^2 +  

514: \left( \frac{\partial g_z}{\partial x} \right)^2 }=

515: \frac{L}{g_0}   

516: \sqrt{\left( \frac{\partial^2 U}{\partial x^2} \right)^2 +  

517: \left( \frac{\partial^2 U}{\partial z \partial x} \right)^2}.  

518: \label{pot}

519: \end{equation}

520: It is clear that we are dealing with gradients of gravity 

521: variations $\delta {\bf g}$. A projection of this formula on 

522: the $x,y$ plane returns us to 

523: the results discussed above (compare with Eq.(\ref{8})):

524: \begin{equation}

525: \Delta \alpha_0 = \frac{L}{g_0} \frac{\partial g_x}{\partial x}

526: \approx \frac{\delta g}{g} \frac{L}{\lambda} \approx \delta\alpha

527: \frac{L}{\lambda}. 

528: \label{8a}

529: \end{equation}

530: 

531: \section{Geophysical phenomena to be studied}

532: \label{s4}

533: 

534: A number of interesting geophysical phenomena are in fact 

535: geophysical signals that can be studied with the

536: help of laser-beam detectors of gravitational waves. The division

537: of geophysical phenomena into signals and noises is not 

538: clear-cut. For example, the low-frequency seismic perturbations 

539: can be assigned to either category. Likewise, periodic thermal 

540: deformations of land are probably the largest ones numerically, but 

541: they are not of gravitational origin, and we treat them as a 

542: predictable and removable hindrance rather than an interesting 

543: geophysical signal. In general, we shall regard signals some 

544: quasi-periodic processes with strong participation of gravity in them, 

545: whereas more random processes we regard noises, hampering detection of 

546: the signals. Noises are considered in the next section, while here we

547: focus on some quasi-periodic phenomena which are believed to be 

548: important for global geodynamics. They are accompanied 

549: by variations of local gravity vector and land deformations. Both 

550: of these perturbations influence the performance of laser 

551: interferometric detectors of gravitational waves, and should be 

552: monitored and removed by the control systems. 

553: 

554: 

555: \medskip

556: \noindent {\it a). The Earth tides.}

557: \medskip

558: 

559: Large perturbations of gravity ${\bf g}$ and deformations of land 

560: ${\bf u}$ are induced by the gravitational fields of the Moon and, 

561: to a smaller extent, Sun. The main tidal harmonics are well 

562: described and measured \cite{c12}. The values and directions of the 

563: perturbed quantities depend on the position of the observation point 

564: on the Earth surface, but we will be mostly interested in their typical 

565: (not exceptionally small and not exceptionally large) numerical  

566: amplitudes. The dominant contributor to the tidal effects is the 

567: lunar $M_2$ harmonic which has a period of $\tau\simeq 12~ hours$ 

568: and which is quadrupolar in terms of its angular dependence. 

569: The typical $M_2$ amplitude of the variations of the absolute

570: value of ${\bf g}$ is $\delta g\simeq 45~\mu Gal$ 

571: ($ Gal =  cm/sec^2,~\mu Gal = 10^{-6}~Gal,~nGal = 10^{-9}~Gal$). 

572: The unperturbed gravity field is

573: $g \simeq 980~ Gal$, so that for the $M_2$ harmonic we 

574: have $\delta g/g \simeq 5 \times 10^{-8}$. 

575: The main solar harmonic is $S_2$ with $\tau = 12~ hours$ and 

576: $\delta g \simeq 20~\mu Gal$. The higher frequency tidal 

577: harmonics are considerably weaker. For example:

578: $M_{3}$ with $\tau \simeq 8~ hours$, $\delta g \simeq 0.8 ~\mu Gal$ and 

579: $M_4$ with $\tau \simeq 6 ~hours$, $\delta g = 0.03 ~\mu Gal$. 

580: 

581: On the Earth surface, variations of gravity ${\bf g}$ and 

582: displacements ${\bf u}$ of land elements 

583: are linked by the Love numbers. If $W_2$ is 

584: the quadrupole component of the tidal potential, one can write

585: for the radial components $\delta g_r$ and $u_r$ \cite{c12}: 

586: \begin{equation}

587: \label{tides1}

588: \delta g_r = 2(1+h-\frac{3}{2} k) \frac{W_2}{R_e},

589: \end{equation} 

590: \begin{equation}

591: \label{tides2}

592: u_r =\frac{h}{g} W_2,

593: \end{equation}  

594: where $R_e$ is the Earth radius ($R_e \approx 6400~ km$) and $h, k$ are

595: the Love numbers ($h \approx 0.6,~ k \approx 0.3$). Since the radial and 

596: angular components of the displacement vector ${\bf u}$ are approximately

597: equal, we will be using for all of them the approximate relationship

598: \begin{equation}

599: \label{tides3}

600: u \approx 0.3~ \frac{\delta g}{g} R_e.

601: \end{equation}

602: This relationship is valid for any gravitational perturbation, not

603: necessarily a tidal one. Specifically for the $M_2$ tidal component,

604: a typical displacement amounts to $u \approx 10~ cm$. 

605: The observed maximal displacements are a factor $3 - 5$ larger than 

606: this evaluation. 

607: 

608: The characteristic spatial scale of the considered quadrupole 

609: perturbation is $\lambda \simeq R_e$. Therefore, assuming that 

610: the armlength of the interferometer is $L = 3~ km$, the 

611: $\Delta$-variation is smaller than the $\delta$-variation 

612: (see Eq.(\ref{8})) in proportion to $L/\lambda$, which reduces to 

613: \begin{equation}

614: \label{difffactor}

615: \frac{L}{R_e} \approx 5 \times 10^{-4}. 

616: \end{equation}

617: Combining the numbers, we arrive at  

618: \begin{eqnarray}

619: \label{tides4}

620: \Delta x_0 & = & u \frac{L}{R_e} =0.3~\frac{\delta g}{g} L 

621: \approx 5 \times 10^{-3}~ cm,~~~~~~{\rm and} \nonumber\\ 

622: \Delta \alpha_0 & = & \frac{\delta g}{g} \frac{L}{R_e} \approx 

623: 3 \times 10^{-11}~ {\rm rad}.

624: \end{eqnarray}

625: 

626: As was already mentioned, tidal perturbations of this level of

627: magnitude should be easily seen by the interferometer 

628: control systems. A successful performance of the 

629: interferometer requires a much better 

630: precision in monitoring the distance between mirrors, at the

631: level not less than $(10^{-6} - 10^{-7})~ cm$. The danger of 

632: tidal effects for the LIGO interferometers and the 

633: necessity of their removal have been long recognised \cite{raab}. 

634: We focus on the positive side of this necessary procedure: 

635: the existing network of interferometers, located

636: in different parts of the globe, will allow one to obtain useful 

637: information on amplitudes, spatial patterns and temporal phases of 

638: tidal perturbations. This collected and analyzed low-frequency 

639: information will help refine numerical values of 

640: Earth's parameters, such, for example, as its Love numbers. 

641:  

642: \medskip

643: \noindent {\it b). Free oscillations of Earth.}

644: \medskip

645: 

646: As every elastic body, our planet is capable of free oscillations

647: (see, for example, \cite{bullen}). In the

648: approximation of a non-rotating, uniform, spherically-symmetric Earth,

649: it is convenient to expand perturbations in terms of 

650: spherical harmonics $Y_{lm}(\theta, \phi)$, where $\phi$ is longitude

651: and $\theta$ is co-latitude:    

652: \begin{equation}

653: u(r,\theta,\phi) \sim~~  {_{n}R_{l} (kr)}\, Y_{lm}(\theta, \phi) \,

654: \exp\left( i \omega t\right). 

655: \end{equation}

656: The discrete values of the wave number $k$, $k_{ln} = 

657: \pi \alpha_{ln} / R_e$, are determined by the radial boundary 

658: conditions and the $n$-th root, $\alpha_{ln}$, of the 

659: corresponding boundary equation (see, for example, \cite{mf}). 

660: The eigen-frequencies $\omega$ do not depend on $m$ and 

661: are given by $\omega_{ln} = \pi \alpha_{ln} c_s /R_e$, where $c_s$ 

662: is the appropriate speed of sound. If the multipole order $l$ is much 

663: smaller than a (nonzero) $n$, the deformations have maximal amplitudes 

664: in the central region of the sphere. In the opposite case of $l$ being

665: much larger than $n$, the deformations are mostly concentrated near 

666: the surface of the sphere, and it also holds that 

667: $\pi \alpha _{ln} \approx l$. In the extreme regime $l \gg n$, 

668: the deformations look more like seismic perturbations rather than free

669: oscillations of Earth. The short-wavelength seismic perturbations

670: can propagate deep to the central regions of Earth, refract and 

671: reflect back. This is an important diagnostic of the central regions.     

672:   

673: There are two basic types of elastic deformations. One is accompanied 

674: by variations of volume, matter density and gravity ${\bf g}$,

675: and is called spheroidal modes $_{n}S_{lm}$. Another

676: is not accompanied by these variations; it represents purely twisting

677: motions, and is called toroidal (shear) modes $_{n}T_{lm}$.

678: Some of the shear perturbations cannot propagate in liquids, and their

679: observed attenuation in the central regions of Earth helped discover the

680: liquid (outer) core. Many hundreds of Earth's normal modes 

681: are now identified from the study of ground

682: motions and gravimeter records. The longest period normal mode is

683: $_0S_2$ with period $T=53.8~ min$. This period is of course 

684: consistent with our evaluations given above. Taking 

685: $c_s \approx 4~ km/sec$ and $\alpha_{20} \approx 1$, we arrive

686: at the right value for $T$. The frequency of the $_0S_2$ 

687: mode is $f \approx 3\times 10^{-4}~ Hz$, whereas the frequencies of 

688: higher order $S$ and $T$ modes extend to the $milli~Hz$ region. For 

689: example, the $_0S_{20} - {_0S_{40}}$ modes ($l = 20 - 40$) have equally 

690: spaced frequencies in the interval of approximately

691: $(1 - 5)~ mHz$, again in agreement with the evaluations given above.    

692: 

693: Free oscillations of Earth are best observed after large earthquakes.

694: In contrast to tidal harmonics, the amplitudes of free Earth

695: oscillations cannot be predicted in advance. However, it is

696: known that typical observed values of gravity variations in the 

697: fundamental $_0S_2$ mode are at the level of about 1 

698: $\mu Gal$ \cite{15a, 15b, 16}. The amplitudes of higher 

699: frequency registered modes are smaller, but they can reach 

700: $\delta g\sim 0.1$ $\mu Gal$. The amplitudes of modestly

701: excited modes between $_0S_{20}$ and $_0S_{40}$ were detected in the

702: long series of observations on seismically quiet days. The

703: acceleration amplitudes are at the level of $2~ nGal$, and 

704: higher \cite{tu}. There exists the permanently present noise 

705: in these modes with the amplitudes corresponding to $0.4~ nGal$. 

706: Interestingly, this background level of land deformations 

707: and gravity variations cannot be explained by cumulative effect 

708: of small earthquakes, despite the fact that there occur 

709: thousands of them per year. It is suggested \cite{tu} to seek 

710: the source of these continuous oscillations of Earth in the action 

711: of atmospheric pressure variations on the Earth surface. We will

712: return to this discussion in our analysis of atmospheric

713: noises in Sec.4b. 

714: 

715: The modestly excited higher frequency $S$ and $T$ modes seem to be 

716: within the reach of adjustment and alighnment systems of the

717: gravity-wave detectors. The $T$ modes do not affect the plumb 

718: line directions, but they are present in the distance variations.  

719: Let us take a representative normal mode with 

720: $f = 4 \times 10^{-3}~ Hz$ and the amplitude of gravity 

721: variations $\delta g = 2~ nGal$, i.e. $\delta g/g = 2 \times 10^{-12}$.

722: The wavelength $\lambda$ is $\lambda = c_s/ f \approx 10^{8}~ cm$,

723: that is, the spatial scale of perturbations $\lambda$ is smaller

724: than $R_e$, but still much longer than the armlength $L=3~ km$ of the 

725: interferometer. 

726: 

727: We shall first generalise Eqs. (\ref{tides3}), (\ref{tides4})

728: to the case, where the spatial scale of $\delta g$ is $\lambda$, not

729: $R_e$. Then, we will have to write 

730: \begin{equation} 

731: \label{gen}

732: \Delta x_0  \approx  0.3~\frac{\delta g}{g} R_e \frac{L}{\lambda}~~~~~~

733: {\rm and}~~~~~~ 

734: \Delta \alpha_0  \approx  \frac{\delta g}{g} \frac{L}{\lambda}. 

735: \end{equation}

736: Note that, on the surface of deformable Earth, there exists a universal 

737: relationship between the distance and angle variations, which directly

738: follows from Eq.(\ref{gen}):

739: \begin{equation}

740: \label{coupl}

741: \Delta \alpha_0 \approx \frac{\Delta x_0} {0.3~ R_e}. 

742: \end{equation} 

743: Substituting the numbers in Eq.(\ref{gen}), we obtain the

744: following evaluations for the considered normal mode of oscillations

745: at $f = 4 \times 10^{-3}~Hz$:

746: \begin{equation}

747: \label{oscil}

748: \Delta x_0 \approx 10^{-6}~ cm,~~~~~{\rm and}~~~~~~

749: \Delta \alpha_0 \approx 6 \times 10^{-15}~ rad.

750: \end{equation}

751: 

752: Obviously, if the amplitude of the excited mode is larger 

753: than $2~ nGal$, this will increase the expected 

754: amplitudes of $\Delta x_0$ and $\Delta \alpha_0$.

755: In addition, normal modes, as any other oscillations, are characterised

756: by certain quality factor $Q$. The mode can execute $Q$ cycles 

757: before its amplitude degrades by a factor of 3, or so. 

758: If the observation of the mode lasts for at least as long as 

759: its relaxation time $\tau$, $\tau = Q/ \omega$, the signal to noise 

760: ratio increases in proportion to $\sqrt{Q}$. This makes the 

761: prospects of detection of the Earth normal modes much better. 

762: The quality factor of individual modes is uncertain, but the 

763: literature often quotes the numbers 

764: well in excess of 100 (see, for example, \cite{c11}).     

765: 

766: \medskip

767: \noindent {\it c). The Earth inner core oscillations}

768: \medskip

769: 

770: The oscillations of the real Earth are complicated by Earth's aspherical

771: shape, its rotation, the presence of gravitational restoring forces

772: in addition to elastic forces, the stratification and variation 

773: of material properties with depth, etc. Very important modifications

774: are caused by the existence of the Earth core (see, for example,

775: \cite{c11}). It is believed that the inner core is solid and has

776: radius $r_s \approx 1220~ km$ and density $\rho_s \approx

777: 13~ g/cm^3$. The outer core is liquid and has radius 

778: $r_l \approx 3470~ km$ and density $\rho_l \approx 12.6~ g/cm^3$. 

779: [The current geophysical models claim a much higher 

780: precision of the Earth parameters than is actually needed 

781: for our evaluations.] 

782: 

783: The solid core is held in its equilibrium position within the

784: liquid core mainly by gravitational forces. It is reasonable

785: to expect that an earthquake, or some other cause, can 

786: occasionally excite oscillations of the inner core, predominantly 

787: along the Earth rotation axis (the polar mode) \cite{16, 17}. 

788: A displacement of a solid sphere within a liquid sphere is 

789: equivalent to a displacement of the effective mass 

790: $m = (4 \pi/3) (\rho_s -\rho_l) r_s^{~3}$. If the amplitude of 

791: the displacement is $A$, then the associated variation of 

792: gravity at the Earth surface is

793: \[

794: \delta g = \frac{2Gm}{R_e^3} A = \frac{8\pi}{3} G A (\rho_s -\rho_l)

795: \left (\frac{r_s}{R_e}\right)^3.

796: \]      

797: The evaluation of this expression using the parameters mentioned 

798: above gives

799: \[

800: \delta g \approx 2~ nGal~ \left(\frac{A}{cm}\right).

801: \]

802: 

803: Period $T_p$ of polar oscillations of the inner core mass 

804: $m_s$ is mainly determined by the gravitational restoring force:

805: \[

806: m_s {\ddot A} = - G \frac{m~\frac{4}{3} \pi \rho_l A^3}{A^2},

807: \]

808: which leads to $T_p^2 = 3 \pi \rho_s / G \rho_l (\rho_s - \rho_l)$.  

809: Corrections to this expression are taken into account by a

810: small numerical factor $\alpha_p$ \cite{17}, so that 

811: \begin{equation}

812: \label{ff1}

813: T_p \simeq \sqrt{ \frac{3\pi(\rho_s +\alpha_p \rho_l)} 

814: {G \rho_l(\rho_s -\rho_l)}}.   

815: \end{equation}

816: The evaluation of $T_p$ leads to a period of about 4 hours.

817: The Coriolis force of the rotating Earth splits this period

818: inito a triplet of periods, with the size of splitting 

819: determined by the Earth angular velocity.

820: 

821: The fascinating phenomenon of the inner core oscillations

822: will be in the center of our further discussion.

823: There exist indications that the inner core translational 

824: triplet has been actually detected \cite{18, 5aa}.

825: The analysis of long records from superconducting

826: gravimeters has shown three resonances with the central period

827: $T_C \approx 3.8~ hours$ and the side periods separated

828: from the central one by $ \sim \pm 10~ min$. The

829: amplitudes of each of the three oscillations are at the

830: level of $(5 - 6) ~nGal$. The quality factor $Q$ of each

831: resonance is somewhat higher than 100.    

832: 

833: Using evaluations similar to those that have been done in

834: previous subsections, we can estimate the effect of the

835: Earth inner core oscillations on the laser-beam detectors of

836: gravitational waves. We take the amplitude of gravity 

837: variations $\delta g/g = 6 \times 10^{-12}$,

838: the charecteristic spatial scale of gravity variations and land

839: deformations $\lambda_c \simeq 2~ R_e$, and the length of the 

840: interferometer arm $L = 3~ km$. Then, the distance and angular 

841: signal amplitudes amount to

842: \begin{equation}

843: \label{core}

844: \Delta x_0 \approx 3 \times 10^{-7}~ cm,~~~~~{\rm and}~~~~~~

845: \Delta \alpha_0 \approx 2 \times 10^{-15}~ rad.

846: \end{equation}

847: 

848: The detectability condition requires that the signal to noise

849: ratio $S/N$ should be better than 1. The resonance bandwidth of 

850: the core oscillations is $\Delta f = f_C /Q$, where the 

851: central resonance frequency is $f_C = 7.3 \times 10^{-5}~ Hz$.

852: We assume that the observation time of the signal 

853: is at least as long as $Q$ cycles of the signal. 

854: This means that the tolerable 

855: spectral noise ${\tilde N}$ per $\sqrt{Hz}$ is related to 

856: the signal amplitude $S$ according to

857: \begin{equation}

858: \label{sn}  

859: {\tilde N} ~ \sqrt{\Delta f} = N < S.

860: \end{equation}

861: Although the quality factor $Q$ for the Earth inner core 

862: oscilllations is somewhat larger than 100, we take it as $Q= 73$ 

863: in order to operate with the round number: $\sqrt{f_C/Q} = 10^{-3}$. 

864: 

865: Thus, we derive the allowed broadband 

866: noises in the region of resonances:

867: \begin{eqnarray}

868: \label{noises}  

869: {\Delta x}_{noise} \approx 3 \times 10^{-4}~ \frac{cm}{\sqrt{Hz}}, & & ~~~~~~ 

870: {\Delta \alpha}_{noise} \approx 2 \times 10^{-12}~ 

871: \frac{rad}{\sqrt{Hz}}, \nonumber\\  

872: & & {\delta g}_{noise} \approx 6 ~\frac{\mu Gal}{\sqrt{Hz}}. 

873: \end{eqnarray}

874: The relationship between the noise variables in Eq.(\ref{noises}) 

875: is of course the same as the relationship between the signal 

876: variables, Eq.(\ref{gen}) with $\lambda \simeq R_e$.  

877: 

878: Obviously, if for some reason the inner core gets excited to the

879: level much higher than $5 - 6~  nGal$, the requirements on the tolerable

880: noise could be significantly relaxed. Alternatively, if the noise 

881: remains at the level of Eq.(\ref{noises}), the $S/N$ gets much 

882: higher than 1.

883: For example, there exists evidence \cite{16} that the amplitude

884: of the inner core oscillations can be as large as $0.64~ \mu Gal$,

885: which would make the signal easily detectable. And, in any case, 

886: one can further improve $S/N$ by observing for longer time than 

887: the assumed 73 cycles.

888: 

889: 

890: \begin{figure}

891:   \begin{center}

892: \includegraphics{vectors2arch.eps}

893: % \epsfbox{vectors.eps}

894:   \end{center}

895:   \caption{\label{vectors} Gravity vectors describing tides and 

896: oscillations of the Earth inner core. Two distributions shown

897: on the panels are separated by a half of the tidal period.}

898: \end{figure}

899: 

900: Variations of gravity on the Earth surface within a patch of 

901: the size $L=3~km$ can be numerically simulated \cite{1a}. 

902: Fig.\ref{vectors} is the illustration of the gravity field

903: vectors disturbed by the inner core polar motion and 

904: by semidiurnal tides. To make the core contribution 

905: distinguishable on the graph, it was artificially 

906: enhanced by a factor 1000. 

907: 

908: 

909: \section{Environmental and instrumental noises}

910: 

911: The important geophysical phenomena listed in the previous 

912: section will certainly affect the laser-beam detectors of 

913: gravitational waves. However, they can only be measured if 

914: the signals are not swamped by various noises. In principle, 

915: any random fluctuations of the Earth surface and any random 

916: redistributions of the surrounding masses of soil, air and water

917: can alter the distances and plumb line directions, which we are 

918: interested in. In addition to these environmental noises, there 

919: always exist imperfections in the measuring devices themselves, 

920: that is, instrumental noises. Here, we will analyse the noises 

921: which we consider most dangerous for the proposed measurements.

922: 

923: \medskip

924: \noindent {\it a) Seismic perturbations}

925: \medskip

926: 

927: It has been long recognized that seismic waves propagating in the 

928: vicinity of a laser interferometer can limit its performance 

929: as a gravitational wave detector \cite{saulson, beccaria, 

930: hughesthorne}. The usual region of concern are frequencies

931: around $10~ Hz$, where the expected noise can be dangerous 

932: for the planned advanced g.w. interferometers. In contrast, we are 

933: concerned with perturbations at much lower frequencies, 

934: $\sim 10^{-4}~ Hz$, and with their effect on the control 

935: systems of the already operating initial interferometers. 

936: Some families 

937: of seismic perturbations are not accompanied by variations 

938: of matter density and, hence, they are not accompanied by 

939: variations of local gravity. This sort of perturbations cannot 

940: directly affect the plumb lines. However, all seismic 

941: perturbations can change the distance between the points on 

942: the Earth surface.  

943: 

944: We will extrapolate the often used spectra of the 

945: seismic (displacement) noise to the frequencies of 

946: geophysical interest. For a discussion of seismic noises, 

947: including those that were measured at the interferometer 

948: cites, see, for example, \cite{saulson, beccaria, hughesthorne,

949: AR, angew, 22b, daw}. The displacement noise is normally 

950: presented in terms of spectral

951: density $u_f$ written in units of $cm/\sqrt{Hz}$: 

952: $u_f = A f^{-\gamma}$. The mean-square value of the random

953: quantity $u$ within a bandwidth from $f_{min}$ to $f_{max}$ is

954: defined by

955: \begin{equation}

956: \label{meansq}

957: \langle u^2 \rangle = \int_{f_{min}}^{f_{max}} (u_f)^2 df.

958: \end{equation}    

959: The spectral index $\gamma$ vary from one spectral interval 

960: to another, but there exist indications that 

961: $\gamma \simeq 1$ at frequencies below $0.1~ Hz$, so we will adopt 

962: $\gamma = 1$. The coefficient $A$ is also uncertain. It 

963: varies from site to site, and there exist a broad range of

964: estimates. We will adopt some averaged value 

965: $A = 10^{-4}~cm /\sqrt{Hz}$ at $f < 0.1~Hz$. So, we shall be 

966: working with the spectrum   

967: \begin{equation} 

968: \label{seism}

969: u_f = 10^{-4} \left( \frac{10^{-1} Hz}{f}\right) ~\frac{cm}{\sqrt{Hz}}.   

970: \end{equation}

971: Surely, at very low frequencies, the concept 

972: of seismic perturbations overlaps with the concept of the Earth 

973: free oscillations, so the estimates should be consistent with each 

974: other and with the gravimeter records. 

975: 

976: To find the differential noise spectral density $\Delta x_{seism}$, 

977: we have to multiply Eq.(\ref{seism})

978: with $L/ \lambda = L f/ c_s$. Assuming that $c_s$ does not 

979: significantly depend on $f$, the frequency dependence cancels out

980: in the product, 

981: and the amount of seismic noise is expected to be approximately

982: equal at all geophysical frequencies. We want to use one simple formula

983: in the broad interval of frequencies, so we put $c_s = 1~ km/sec$

984: for this calculation. Then, using $L = 3~ km$, we arrive at 

985: \begin{equation}

986: \label{seism2}

987: \Delta x_{seism} \approx 3 \times 10^{-5}~ \frac{cm}{\sqrt{Hz}}.

988: \end{equation}

989: This estimate is a factor of 10 lower than the level of tolerable

990: noise in Eq.(\ref{noises}). It shows that this source of noise cannot

991: prevent the observation of the Earth inner core oscillations by the 

992: adjustment system of laser interferometers.

993: 

994: To check the consistency of our evaluations, it is instructive to 

995: estimate the upper bound of gravity variations caused by seismic 

996: perturbations. This evaluation of seismic $\delta g_f$ will also 

997: be helpful in the next subsection where we are considering atmospheric 

998: fluctuations. We will have to compare our estimations with

999: the actually measured (total) values 

1000: of the acceleration noise $\delta g_f$ at the core oscillation 

1001: frequencies around $f_C$ and normal mode $mHz$ 

1002: frequencies (compare, for example, with \cite{richetal, 5aa, 18, tu}):

1003: \begin{equation}

1004: \label{gravvar2}

1005: \delta g_f \approx 0.5~ \left(\frac{10^{-4}~ Hz}{f}\right)

1006: ~ \frac{\mu Gal}{\sqrt{Hz}}. 

1007: \end{equation}

1008: 

1009: A matter density wave, with the amplitude of density variations 

1010: $\delta \rho/ \rho$ and wavelength $\lambda$, produces a local

1011: variation of gravity with the amplitude 

1012: $\delta g/ g(\lambda) \sim \delta \rho/\rho \sim u/\lambda$, where

1013: $g(\lambda)/ g \approx \lambda/R_e$. We take into account the finite

1014: rigidity of Earth, expressed by the Love numbers, and write 

1015: $0.3~ \delta g/g(\lambda) \approx u/\lambda$, that is, 

1016: \begin{equation}

1017: \label{gravvar} 

1018: 0.3~ \frac{\delta g}{g} \approx \frac{u}{R_e}.

1019: \end{equation}  

1020: Surely, this formula is consistent with Eq.(\ref{tides3}). 

1021: Using Eq.(\ref{seism}) in this formula, we arrive at the estimate 

1022: for seismic $\delta g_f$, which turns out to be at the level of the 

1023: observed variations, Eq.(\ref{gravvar2}).

1024: 

1025: The estimate (\ref{gravvar2}) can now be translated into the 

1026: spectral noise contribution to $\Delta \alpha_0$ caused by seismic 

1027: perturbations. It is seen from Eq.(\ref{gen}) that

1028: the $f$-dependence cancels out, so we derive 

1029: \begin{equation}

1030: \label{angvar}

1031: \Delta \alpha_{seism} \approx 10^{-13}~ \frac{rad}{\sqrt{Hz}}. 

1032: \end{equation}

1033: This noise is lower that the tolerable level 

1034: of noise in Eq.(\ref{noises}). Therefore, the alighnment system 

1035: of laser interferometers can also be used, barring some complications

1036: discussed below, for observation of the inner core oscillations.

1037: 

1038: \medskip

1039: \noindent {\it b) Atmospheric noises}

1040: \medskip

1041: 

1042: It is somewhat surprising that the movement of light air 

1043: in the Earth atmosphere can be a more serious source of 

1044: noise for the proposed measurements than the 

1045: movement of heavy soil in seismic perturbations. 

1046: We will illustrate this statement with a simple theory of gravity 

1047: variations and land deformations caused by atmospheric 

1048: motions. 

1049: 

1050: Consider an air density inhomegeneity with the amplitude 

1051: $\delta \rho/\rho$ and a characteristic spatial scale $\lambda$. 

1052: Let this scale be shorter than the characteristic height of the 

1053: Earth atmosphere $H$. In a point of observation at the Earth surface, 

1054: the variation of $g$ is mostly determined by the variation of mass 

1055: $\delta m$ in a volume $\lambda^3$ with the center located at a 

1056: distance $\lambda$ from the observer. More distant volumes give 

1057: smaller contributions to $\delta g$ and can be ignored. 

1058: So, we could write:

1059: \[

1060: \delta g \approx \frac{G~ \delta m}{\lambda^2} \approx 

1061: G~ \delta \rho~ \lambda.

1062: \]

1063: 

1064: This expression is adequate for $\lambda < H$. However, we are mostly 

1065: interested in the opposite case $\lambda > H$. As will be clear from 

1066: discussion below, it is time-dependent atmospheric variations at these 

1067: longer scales that fall in the frequency range of geophysical interest. 

1068: Therefore, we have to consider variations of mass in a volume 

1069: $H \lambda^2$, rather than $\lambda^3$. The center of

1070: this flattened volume is located at a distance $\lambda$ 

1071: from the observer. The expression for $\delta g$, appropriate for

1072: scales and frequencies of our interest, should be written as

1073: \begin{equation}

1074: \label{vargatm}

1075: \delta g \approx \frac{G~ \delta m}{\lambda^2} \approx 

1076: G ~\delta \rho ~ H. 

1077: \end{equation}

1078: 

1079: Variations of air density $\delta \rho$ are more difficult to measure 

1080: than variations of pressure, so one usually operates with the 

1081: (spectral) pressure variations $\delta p_f = c_a^{~2}~ \delta \rho_f$, 

1082: where the square of the air sound speed is 

1083: $c_a^{~2} \simeq 10^{9}~ cm^2/ sec^2$.   

1084: Then, equation (\ref{vargatm}) takes the form  

1085: \begin{equation}

1086: \label{vargatm2}

1087: \delta g_f \approx G~ \frac{\delta p_f}{c_a^{~2}}~ H. 

1088: \end{equation}

1089: The question arises which $\delta p_f$ should be used 

1090: in this equation in order to estimate $\delta g_f$.

1091: 

1092: The observations indicate \cite{c10, c13, 22a, TA, GGM, tu} 

1093: that, during relatively quiet times, the

1094: spectral pressure variations behave, roughly, 

1095: as $\delta p_f \propto f^{-1}$. This frequency dependence is close 

1096: to the Kolmgorov-Obukhov law for atmospheric turbulence \cite{LL, GGM}

1097: $\delta p_f \propto f^{-7/6}$. In the frequency interval 

1098: $\sim (10^{-5} - 10^{-2})~Hz$, observational data

1099: can be approximately fitted by the formula 

1100: \begin{equation}

1101: \label{KO}

1102: \delta p_f = 200 \left(\frac{f}{10^{-4} Hz}\right)^{-1} 

1103: ~\frac{Pa}{\sqrt{Hz}}, 

1104: \end{equation}

1105: where the unit of pressure $Pa$ is $Pa = 10 ~g/cm~ sec^2$. The

1106: replacement of the spectral index $(-1)$ by the genuine 

1107: Kolmogorov-Obukhov index $(-7/6)$ leaves the fit equally good. 

1108: 

1109: A purely turbulent $\delta p_f$ raises certain concerns, as 

1110: it is not obvious that turbulence can directly contribute to 

1111: $\delta g$. Turbulence describes chaotic velocities in a medium and 

1112: variations of the medium's kinetic pressure, but these variations are 

1113: not necessarily accompanied by variations of the medium's mass 

1114: density $\delta \rho$ and gravity $\delta g$. [There exists,

1115: however, an indirect influence of air turbulence on $g$ 

1116: through the land deformations excited by the variable pressure.

1117: We will discuss this issue separately, at the end of this 

1118: subsection.] On the other hand, transportation and mixing of cold 

1119: and hot air, caused by the turbulence, does contribute 

1120: to $\delta \rho$ and, hence, directly contributes to $\delta g$. 

1121: Without analysing the true nature of the observed $\delta p_f$,

1122: we will be using the observed spectrum (\ref{KO}) in

1123: the expression (\ref{vargatm2}). 

1124: 

1125: Combining Eq.(\ref{KO}) and Eq.(\ref{vargatm2}), and 

1126: using numerical values of the participating quantities,

1127: including the effective height of the atmosphere $H \approx 10~km$, 

1128: we arrive at

1129: \begin{equation}

1130: \label{gravvara}

1131: \delta g_f \approx 0.2~ \left(\frac{10^{-4} Hz}{f}\right) 

1132: ~ \frac{\mu Gal}{\sqrt{Hz}}. 

1133: \end{equation}

1134: This evaluated $\delta g_f$ is smaller than, and is consistent with, 

1135: the observed (total) variations $\delta g_f$ approximated by 

1136: Eq.(\ref{gravvar2}).

1137: 

1138: To evaluate the contribution of the atmospheric $\delta g_f$ to

1139: the noises in distance and angular measurements 

1140: (see Eq.(\ref{gen})), we first need to find out the characteristic

1141: correlation length $\lambda$, as a function of frequency $f$.

1142: To do this, we have used experimental data which were kindly 

1143: provided to us by the Japanese Weather Association. These data

1144: were collected by the Meteorological National Geographical

1145: Institute at the set of meteo-stations near Tsukubo Scientific Center.

1146: The total area of $20 \times 20~km$ is covered by $10$ stations.

1147: The average distance between nearby stations is $10~km$. The pressure

1148: data were recorded at each station with a sample time $1~min$. The

1149: length of the analysed record was $1~month$.  

1150: 

1151: Using the experimental data, we have built temporal spectra of pressure

1152: $S_p(f)$ measured at each station, and also temporal spectra of difference 

1153: of pressure $S_{\Delta p} (f)$ at each pair of neighbouring stations. 

1154: Because of the spatial correlation of the data, the spectral components

1155: of the difference of pressure $S_{\Delta p} (f)$ prove to be smaller 

1156: than the approximately equal spectral components of pressure $S_p(f)$

1157: measured at each station. 

1158: For each pair of stations, the spatial correlation scale $\lambda(f)$ 

1159: was estimated from the relationship 

1160: \[

1161: \lambda (f) = r \frac{2 S_p(f)}{S_{\Delta p}(f)}.

1162: \]

1163: Then, the found quantities $\lambda (f)$ were avaraged over all pairs

1164: of stations. As a result, we have arrived at the mean value of 

1165: $\lambda (f)$:

1166: \begin{equation}

1167: \label{lamf}

1168: \lambda (f) = 500~ \left( \frac{10^{-4} Hz}{f} \right)~km.

1169: \end{equation}

1170: It also follows from the data that this number depends on overall 

1171: meteorological conditions and can change by a factor $2-3$. 

1172: 

1173: The evaluations (\ref{lamf}), (\ref{gravvara}) should now be used in 

1174: general formulas (\ref{gen}). We arrive at the expected atmospheric 

1175: noise contribution:

1176: \begin{equation}

1177: \label{atmn}  

1178: {\Delta x}_{atm} \approx 2 \times 10^{-4}~ \frac{cm}{\sqrt{Hz}}, ~~~~~

1179: {\rm and}~~~~ 

1180: {\Delta \alpha}_{atm} \approx 10^{-12}~ \frac{rad}{\sqrt{Hz}}. 

1181: \end{equation}

1182: These estimates are marginally consistent with the level of allowed 

1183: noise defined by Eq.(\ref{noises}).

1184: 

1185: It is satisfying that the differential displacement noise (\ref{atmn}) 

1186: is consistent with the direct estimate of land deformations, supposedly 

1187: excited by atmospheric turbulence \cite{tu}. The permanently present  

1188: amplitudes of acceleration $a_n$ at $mHz$ frequencies of the Earth's 

1189: eigen-modes are all approximately at the level of $0.4~ nGal$ \cite{tu}. 

1190: Taking $3~ mHz$ as the central frequency, the discrete set of

1191: these observed modal amplitudes $a_n$ can be translated into an

1192: interval of a continuous acceleration spectrum 

1193: \begin{equation}

1194: \label{acc}

1195: a_f \approx 0.4~ \frac{1}{\sqrt{3\times 10^{-3}}}~ \frac{nGal}{\sqrt{Hz}}

1196: \approx 7~ \frac{nGal}{\sqrt{Hz}}.

1197: \end{equation}

1198: The observed accelerations $a_f$ are associated with the 

1199: periodic displacements $u_f$ of the land elements: 

1200: \begin{equation}

1201: \label{displ}

1202: u_f \approx \frac{a_f}{(2 \pi)^2 f^2} \approx 10^{-9}~ \frac{1}{f^2}

1203: ~\frac{cm}{\sqrt{Hz}}.

1204: \end{equation}

1205: The corresponding spectral noise amplitude of the distance variation 

1206: between the interferometer's mirrors, separated by $L= 3~km$, will 

1207: amount to

1208: \begin{equation}

1209: \label{tun}

1210: \Delta x_{f} \approx u_f \frac{Lf}{c_s} \approx 3 \times 10^{-9} 

1211: ~\frac{1}{f}~ \frac{cm}{\sqrt{Hz}}.

1212: \end{equation}

1213: We extrapolate this formula to lower frequencies, including

1214: the interval around $f_C$. Then, at $f=f_C$, we obtain  

1215: \begin{equation}

1216: \label{tun2}

1217: \Delta x_{f} \approx 5 \times 10^{-5}~ \frac{cm}{\sqrt{Hz}}.

1218: \end{equation}

1219: This estimate is perfectly consistent with the independent evaluation 

1220: that we have derived in Eq.(\ref{atmn}). 

1221: 

1222: To conclude this subsection, we can say that the expected 

1223: atmospheric noise is marginally acceptable for the proposed 

1224: geophysical measurements with laser-beam detectors of gravitational 

1225: waves. It is likely that some rare events, such as passing  

1226: cyclones and atmospheric fronts, can produce larger disturbances than

1227: the estimates (\ref{atmn}), but these atmospheric events can be 

1228: anticipated and excluded from the data.    

1229: 

1230: \medskip

1231: \noindent {\it c) Coupling of noises in distance and angle measurements}

1232: \medskip

1233: 

1234: As was already discussed above, variations of the Earth's 

1235: gravitational field, caused either by external or internal

1236: sources, produce changes in, both, distance between the sites 

1237: and angle between the plumb line directions on the Earth surface. 

1238: It is only on the surface of an idealized perfectly 

1239: rigid body that the change of angle between the plumb lines is 

1240: not accompanied by distance variations. And it is only

1241: in an idealized homogeneous gravitational field, or during 

1242: spherically-symmetric oscillations of a homogeneous sphere, 

1243: that the change of distance between the sites is not accompanied by the 

1244: change of angle between the plumb lines. While describing geophysical 

1245: signals, we have already noticed this important link between 

1246: $\Delta x_0$ and $\Delta \alpha_0$, see Eq.(\ref{coupl}). 

1247: For example, as we know, the Earth inner core oscillations

1248: can lead to distance variations with the amplitude 

1249: $\Delta x_0 = 3 \times 10^{-7}~ cm$, and they will necessarily 

1250: be accompanied by the angle variations with the amplitude

1251: $\Delta \alpha_0 = 2 \times 10^{-15}~ rad.$ However,

1252: even in the absence of any geophysical signal, a similar link

1253: exists between the random (noisy) parts of $\Delta x_0$ and 

1254: $\Delta \alpha_0$.

1255: 

1256: Fig.\ref{f7} illustrates the appearance of noise in the

1257: channel of angle measurements, if there exists noise in 

1258: the channel of distance measurements. The fluctuating part

1259: $\Delta L$ of distance between the sites (independently of

1260: the reasons why this fluctuating part exists) translates into the 

1261: fluctuating part $\alpha_{gn}$ of angle between the plumb

1262: lines, $\alpha_{gn} \sim \Delta L/ R_e$. This coupling takes

1263: place simply because of the centrally-symmetric character 

1264: of the Earth's gravitational field. [We ignored this fact

1265: in derivation of Eq.(\ref{9}), but otherwise a term proportional 

1266: to $\Delta x_{0 \omega}/ R_e$ should have been added to the 

1267: right-hand-side of that equation.] Therefore, signals and 

1268: noises in the two channels are linked by essentially one 

1269: and the same relationship: 

1270: \begin{equation}

1271: \label{crel}

1272: \Delta \alpha \approx \frac{\Delta x}{R_e}. 

1273: \end{equation}

1274: This equation gives the minimal and unavoidable level of noise in 

1275: angles, which arises because of the noise in distances. Other reasons 

1276: (for example, related to local inhomogeneities on two sites) can 

1277: only make this relationship more complicated. However, even the simplified 

1278: link represented by Eq.(\ref{crel}) leads to an important conclusion: 

1279: there is no special advantage in measuring angles through the

1280: alignment control system, rather than measuring distances

1281: through the adjustment control system; the signal-to-noise ratios 

1282: are practically equal in these channels. On the other hand, the use of

1283: both channels, instead of only one of them, would certainly

1284: increase the reliability of geophysical detection. In particular,

1285: the combined information will help identify and remove the distance

1286: variations caused by

1287: quasi-periodic thermal deformations of the Earth surface. Moreover, 

1288: if both the distance and angles between mirrors are

1289: accurately measured (including their noisy contributions), 

1290: then the use of Eq.(\ref{crel}) for noises (or, preferably, the 

1291: use of a more precise version of this equation) would allow  

1292: one to subtract from the measured angle that part of the angle 

1293: which arises purely due to the coupling. This procedure would help

1294: identify that part of the measured angle which is caused specifically 

1295: by the variation of plumb lines.

1296: 

1297: It turns out, however, that the angle measurements by the 

1298: actually existing alignment systems have their own difficulties

1299: and ambiguities. We will fully address them in Sec.5. 

1300: 

1301: \begin{figure}

1302:   \begin{center}

1303: \includegraphics{fig7.eps}

1304: %\epsfbox{fig7.eps}

1305:   \end{center}

1306:  \caption{\label{f7} Coupling of signals and noises in distances and angles}

1307: \end{figure}

1308: 

1309: 

1310: \medskip

1311: \noindent {\it d) Instrumental noises}

1312: \medskip

1313: 

1314: It appears that instrumental noises will not be a major problem for

1315: the proposed geophysical studies. This is not surprising, as the 

1316: gravitational-wave detectors implement the best currently available

1317: technology aimed at measuring much weaker gravitational signals.

1318: Obviously, geophysical signals have different origin and 

1319: are concentrated in a substantially different frequency band, but

1320: in both cases we are dealing with manifestations of the same

1321: force - gravitation. 

1322: 

1323: The tolerable noise of Eq.(\ref{noises}) implies that the

1324: accuracy of the adjustment system is better

1325: than $3 \times 10^{-6}~ cm$ at frequencies around $10^{-4}~ Hz$.   

1326: Adjustment systems of the presently operating interferometers

1327: satisfy this requirement, but, in order to ensure a successful 

1328: performance of the interferometer as a g.w. detector, they may 

1329: not need the low-frequency precision higher than $10^{-7}~ cm$.  

1330: Certainly, the advanced interferometers will require an enhanced

1331: precision of the adjustment systems in the `control band' of 

1332: frequencies. This is necessary in order to 

1333: assure that only very small forces 

1334: are applied to the mirrors themselves \cite{sho}. In any case, 

1335: from the instrumental point of view, the accuracy of the 

1336: adjustment system can be made much better than presently achieved. 

1337: Even smaller variations, at the level of $10^{-13} cm$, can be 

1338: detected by the interferometer adjustment system, if the dominant

1339: noise is reduced to the photon shot noise \cite{MT1,MT2}. 

1340: The photon noise limited spectral sensitivity is basically determined 

1341: by the relationship

1342: \begin{equation}

1343: \label{Ed3}

1344: \Delta x_{instr} \simeq

1345: \frac{\lambda}{F} \sqrt{\frac{\hbar\omega}{\eta P}},

1346: \end{equation}

1347: where $\lambda$ is the wavelength of the laser light and

1348: $\omega= 2\pi c/\lambda$ is its frequency, $F$ is the 

1349: cavity finesse, $P$ is the optical power available at the photodector

1350: and $\eta$ is its efficiency. Taking reasonable parameters

1351: $\lambda = 10^{-4}~ cm$, $F = 100$, $P = 100~ mW$, $\eta =0.9$

1352: we arrive at  

1353: \begin{equation}

1354: \label{ed3a}

1355: \Delta x_{instr} \approx 1.5 \times 10^{-15} ~~\frac{cm}{\sqrt{Hz}}. 

1356: \end{equation}

1357: This number can be improved by the increase of $P$. Therefore, 

1358: we do not expect instrumental noises in

1359: the adjustment system to be a limiting factor for the measurement

1360: of geophysical signals described by Eqs. (\ref{oscil}), (\ref{core}).

1361: Obviously, we assume that the extra noises (above purely photon noise)

1362: are not excessively large.

1363: 

1364: The purpose of the alignment system is to monitor and remove

1365: possible misalignments between the laser beam and the 

1366: axis of optical resonator formed by the corner mirror and the end

1367: mirror. One reason for this misalignment to arise is precisely 

1368: the deviation of plumb lines caused by a geophysical signal and, 

1369: hence, the change of angle between mirrors. In principle, the existing

1370: alignment systems are very sensitive to a

1371: possible change of angle between mirrors. If the dominant noise

1372: is the photon shot noise, the angular spectral precision of the

1373: alignment system can reach \cite{8, 9, 1a, 9a}: 

1374: \begin{equation}

1375: \label{E3}

1376: \Delta \alpha_{instr} \simeq

1377: \frac{\lambda}{w_0} \sqrt{\frac{\hbar\omega}{\eta P}},

1378: \end{equation}

1379: where $w_0$ is the diameter of the laser beam `waist'.

1380: Taking the same parameters as before and $w_0= 0.1~ cm$, 

1381: we arrive at  

1382: \begin{equation}

1383: \label{e3a}

1384: \Delta\alpha_{instr} \approx 1.5 \times 10^{-12} ~~\frac{rad}{\sqrt{Hz}}. 

1385: \end{equation}

1386: This accuracy is at the level of the tolerable noise of Eq.(\ref{noises}),

1387: but it can be improved by the increase of $P$. We cautiously conclude that 

1388: the instrumental noise of alignment systems will not swamp the expected 

1389: geophysical signal. 

1390: 

1391: A more fundamental problem with the existing alignment systems 

1392: lies, however, in a different place. The error signal of the system 

1393: cannot tell us whether the misalignment arose because of the change of 

1394: angle between the mirrors or because of the change of lateral 

1395: position of the end spherical mirror. The first cause is

1396: what we are interested in, while the second cause could be simply 

1397: an unrelated noise. Without solving this problem, one would not be able

1398: to use angular measurements as an additional source of geophysical 

1399: information. This is why we address this problem in next sections.

1400: 

1401: \section{Measurement of mirrors' tilts}

1402: 

1403: Here, we will consider in more detail the response of a g.w.

1404: interferometer to the deformations of Earth's surface and 

1405: variations of local gravity. To be specific, we will do this

1406: on the example of the VIRGO observatory \cite{virgoweb}.  

1407: 

1408: Each arm of the VIRGO interferometer consists of an asymmetric

1409: Fabry-Perot optical resonator, which is formed by a flat front 

1410: mirror and a spherical end mirror. The length of each arm is 

1411: $L=3~km$. The laser beam enters the resonator through the partially 

1412: transparent front mirror. Each mirror is hanging on a multi-stage

1413: support, which behaves as a single wire pendulum at frequencies

1414: below $1~Hz$. The source of light itself is located very close 

1415: (at a distance less than $20~m$) to the front mirror, so one can

1416: neglect the gradients of local gravity and deformations of land at

1417: such short distances. Therefore, we shall accept, for simplicity, 

1418: that in the chosen coordinate system the front flat mirror does not 

1419: move, the laser beam is always orthogonal to the front mirror, and the

1420: beam enters the resonator from one and the same point on the flat mirror. 

1421: The problem is still sufficiently general, as the main trouble arises 

1422: when one considers the relative position and orientation of the remote 

1423: spherical mirror with respect to the front mirror and laser beam.     

1424: 

1425: The optical axis of the resonator is orthogonal to the flat mirror

1426: and goes through the center of curvature of the spherical mirror.

1427: The curvature radius $r$ of the spherical mirror is approximately

1428: equal to the distance between the mirrors, so we take $r= 3~km$.

1429: In the tuned working condition, the laser beam is aligned with 

1430: the optical axis. In this configuration, all light participates 

1431: in interference, and the sensitivity is maximal. A misalignment 

1432: between the beam and optical axis is sensed by the changing spatial 

1433: distribution of the electromagnetic field in the Fabry-Perot 

1434: cavity \cite{8, ward}. Since we have assumed that the

1435: beam is always orthogonal to the flat mirror, the misalignment

1436: can only arise if the spherical mirror inclines with respect to

1437: the beam, or changes its position in the plane orthogonal to 

1438: the beam. These possible 

1439: motions of the end mirror, in the $(x, y)$-plane containing the 

1440: beam, are shown in Fig.\ref{mirrors}.    

1441: 

1442: \begin{figure}

1443:   \begin{center}

1444: \includegraphics{mirrors.eps}

1445: %\epsfbox{mirrors.eps}

1446:   \end{center}

1447:  \caption{\label{mirrors} Misalignment of the optical axis in 

1448: the Fabry-Perot interferometer: a)~aligned configuration, 

1449: b) tilt of the spherical mirror, c) vertical shift of the 

1450: spherical mirror}

1451: \end{figure}

1452: 

1453: A technical advantage of this alignment control system (which turns 

1454: out to be a big disadvantage for geophysical applications) is in 

1455: that the system does not need to know whether the misalignment

1456: arose because of the tilt of the end mirror or because of its lateral

1457: shift. They both lead to similar displacements of the optical axis, 

1458: which will be corrected by the control system without ever

1459: distinguishing the true origin of the displacement. In other 

1460: words, there exists a degeneracy between tilts (configuration b) 

1461: in Fig.\ref{mirrors}) and shifts

1462: (configuration c) in Fig.\ref{mirrors}). They both can

1463: result in one and the same parallel displacement of the optical 

1464: axis (shown by a dashed line in Fig.\ref{mirrors}). The 

1465: error signal automatically arises when a misalignment develops, 

1466: but it will be the same signal for both configurations. In both cases, 

1467: the feedback mechanical system compensates the misalignment by 

1468: tilting the spherical mirror and returning the optical axis back 

1469: to its original position aligned with the laser beam.

1470:  

1471: Let the $x$-axis go along the beam and the $y$-axis be a vertical

1472: direction. Suppose that the plumb line tilts by an angle 

1473: $\theta_{xy}$  (see Fig.\ref{mirrors}). Then the curvature center

1474: of the hanging spherical mirror moves a distance 

1475: $r \theta_{xy}$. This means that the optical axis will take a new

1476: position displaced from the original one by $s_y = r \theta_{xy}$. 

1477: Now let the suspension point of the spherical mirror shift by 

1478: a distance $d_y$ (see Fig.\ref{mirrors}). Then the optical axis does 

1479: also shift by this distance, $s_y=d_y$. The parallel displacements 

1480: of the optical axis will be numerically the same, 

1481: if $d_y =r \theta_{xy}$.

1482: 

1483: The expected geophysical signal in angle variations is

1484: $\Delta \alpha = 2 \times 10^{-15}~rad$, see Eq.(\ref{core}). 

1485: The displacement of the optical axis caused by this

1486: useful signal will be $s_y = 6 \times 10^{-10}~ cm$. This is a

1487: measurable signal. However, such a displacement of the axis 

1488: can also be produced by a lateral shift of the suspension point 

1489: $d_y = 6 \times 10^{-10}~ cm$. These lateral shifts of the suspension

1490: point are likely to happen  - they are well below the 

1491: displacement noise of Eq.(\ref{noises}). All spatial components of

1492: the displacement noise have approximately equal amplitudes,

1493: so one can be sure that there is also a vertical component 

1494: of this level of magnitude. 

1495: In other words, a displacement noise, which is tolerable for 

1496: longitudinal distance measurements, turns out to be intolerable

1497: for angular measurements by the existing alignment systems.

1498: Although the control system will generate an error signal and

1499: will correct the misalignment, it is impossible to distinguish

1500: whether it was produced by a useful geophysical effect or by the 

1501: ever-present displacement noise. 

1502: 

1503: Since the channel of angular measurements is important

1504: for geophysical applications, we have to find a way of breaking the

1505: degeneracy, specific for spherical mirrors, between tilts and 

1506: shifts. This will require modest

1507: amendments to the existing optical configurations. Some possible

1508: ideas are described in the next section. It is also important to   

1509: note that the advanced interferometers may not need these 

1510: amendments at all. There exist strong arguments in favor of using

1511: non-spherical mirrors \cite{bonthorne} in which case the problem of 

1512: degeneracy will be automatically alleviated.

1513: 

1514: 

1515: \section{Modifications of the optical scheme}

1516: 

1517: It was shown above that the useful geophysical information on 

1518: $\Delta x_0$ is routinely stored in circuits of the adjustment 

1519: system. The extracton of this part of geophysical information does

1520: not require any hardware modifications, and it can be found at the 

1521: level of data analysis. In contrast, to receive appropriate information 

1522: on $\Delta \alpha_0$ one will first need to make certain modifications 

1523: of g.w. detectors with spherical mirrors, specially for the purpose of 

1524: geophysical applications. In principle, the necessary 

1525: angular information could be obtained by methods unrelated 

1526: to g.w. observations, but the modifications proposed here make

1527: use of the existing exceptional infrastructure of g.w. observatories, 

1528: such as long arms incorporating high-quality vacuum tubes, 

1529: some common for astrophysical and geophysical applications elements of 

1530: shielding and control, sensitive detection techniques, etc. Surely, 

1531: we keep in mind that any geophysical modifications should not jeopardize 

1532: the functioning of interferometers as astrophysical instruments.  

1533: 

1534: \subsection{Auxiliary interferometer for distance measurements}

1535: 

1536: One way of receiving information on $\Delta \alpha_0$ is based on

1537: Eq.(\ref{10}). The adjustment system responds to the low-frequency part of 

1538: the distance variation $\Delta d_{\omega}$ between mirrors. 

1539: As we know, the major contribution to $\Delta d$ is provided 

1540: by $\Delta x_0$, and this is why we intend 

1541: to extract information on $\Delta x_0$ from the data of the 

1542: adjustment system. However, the accuracy of this measurement can be so 

1543: high that it can reach the level of a much smaller contribution to 

1544: $\Delta d$: $l \Delta \alpha_0$, second term in Eq.(\ref{10}). 

1545: Taking $l=10^{2}~ cm$ for

1546: the suspension length and $\Delta \alpha_0 = 2 \times 10^{-15}~ rad$,  

1547: Eq.(\ref{core}), for the expected useful signal, we get the 

1548: estimate: $l \Delta \alpha_0 \approx 2 \times 10^{-13}~ cm$.  

1549: If one could measure the difference between $\Delta d$ and $\Delta x_0$ 

1550: with this sort of precision, and since the suspension length $l$

1551: is known, the result would have provided the necessary 

1552: information on $\Delta \alpha_0$. In principle, a direct measurement 

1553: of this difference is possible with the help of an additional 

1554: interferometer system, which we will now describe.

1555: This proposal is similar to the one first suggested by R. Drever

1556: in the context of an interferometer on magnetically suspended 

1557: mirrors \cite{DR, DA}.

1558: 

1559: The auxiliary interferometer is shown in Fig.\ref{f4}. One arm is

1560: formed by mirrors of the g.w. interferometer itself. Another arm

1561: is formed by mirrors attached to the suspension points, or, more

1562: practical, to the last stage of the anti-seismic filters. Such 

1563: configuration with two parallel arms is known as Mach-Zender 

1564: interferometer. Each arm of the Mach-Zender interferometer can 

1565: include a Fabry-Perot resonator.

1566:  \begin{figure}

1567:   \begin{center}

1568: \includegraphics{fig4.eps}

1569: %\epsfbox{fig4.eps}

1570:   \end{center}

1571:  \caption{\label{f4}  Auxiliary  interferometer. M$_1$ and M$_2$ are 

1572: mirrors of g.w. interferometer, M$_{1r}$ and M$_{2r}$ are additional 

1573: mirrors forming the extra resonator, BS and M$_0$ are beamsplitters 

1574: and mirrors changing the beam direction.}

1575: \end{figure}

1576: It is supposed that a small fraction of light is deviated from the 

1577: main beam to the additional interferometer. As the frequencies of 

1578: geophysical interest are relatively low, $\sim 10^{-4} Hz$, the 

1579: relaxation time of the Mach-Zender interferometer should be sufficiently 

1580: long. If the reflectivity of mirrors $M_{1r}$ and $M_{2r}$

1581: is sufficiently high, a small portion of light 

1582: can resonate in the cavity for long time, increasing the sensitivity of

1583: observation. The added interferometer is capable of measuring

1584: the difference of its arm-lengths with high precision.  

1585: It is interesting to note that this configuration has been actually 

1586: realised \cite{tsu}, even though the modification is motivated by the

1587: struggle with noises rather than by geophysical applications. 

1588: 

1589: The proposed scheme is differential, and therefore it has certain 

1590: advantages:

1591: 

1592: - seismic and other common for both arms noises cancel out due to 

1593:   differential character of measurements;

1594: 

1595: - sudden changes of laser beam position and direction (beam `walks' 

1596:   and `jitter') are not dangerous, as they are common

1597:   for both arms and cancel out; 

1598: 

1599: - the scheme is universal and can be used at any g.w. interferometer 

1600:   regardless of the actual construction of suspension.

1601: 

1602: Two other possible modifications, discussed next, are aimed at 

1603: direct breaking of degeneracy between shifts and tilts in 

1604: the alignment system. These modificatons do not imply a deviation of

1605: light from the main beam of interferometers and, therefore, 

1606: they are even less likely to compromise the g.w. performance of the 

1607: instruments.

1608: 

1609: \subsection{\label{sfl} Flat mirror behind spherical end mirror.}

1610: 

1611: The desired goal of telling the difference between shifts and tilts 

1612: of spherical end mirror can be achieved by installation of a 

1613: flat beamsplitter behind the end mirror. The 

1614: beamsplitter can be attached to the last stage of the anti-seismic

1615: filter. A scheme of this proposal is shown in Fig.\ref{f5}. It is 

1616: supposed that the beamsplitter (2) intercepts part of light traveling 

1617: toward the photodetector (3), normally used as an element of the

1618: alignment system. The intercepted light is directed through the focusing

1619: lens (5) toward the additional photodetector (7). This photodetector should

1620: be sensitive to the position of a spot of light focused on the photodetector.

1621: For example, the photodetector (7) could consist of two or four pieces,

1622: so that a slight movement of the focal spot would produce a differential

1623: signal. Since the spherical mirror (1) and the beamsplitter (2) are

1624: hanging very close to each other, they participate together in possible 

1625: shifts of the suspension point and in possible tilts caused by varying 

1626: local gravity.   

1627: 

1628: Imagine that the mirror (1) and beamsplitter (2) are shifted together in 

1629: vertical direction (configuration c) in Fig.\ref{mirrors}). This would

1630: lead to no change at the photodetector (7). In contrast, if the mirror (1) 

1631: and beamsplitter (2) are tilted together (configuration b) in 

1632: Fig.\ref{mirrors}), the focal spot will change its position on the     

1633: photodetector (7) leading to a measurable signal. In this way, the

1634: previously indistinguishable configurations b) and c) become 

1635: distinguishable. If the evaluation (\ref{e3a}) of the angular precision 

1636: remains valid for the proposed scheme, one will be able to extract 

1637: geophysical information on $\Delta \alpha_0$ from this measurement.

1638: \begin{figure}

1639:   \begin{center}

1640: \includegraphics{fig5.eps}

1641: %\epsfbox{fig5.eps}

1642:   \end{center}

1643:  \caption{\label{f5} A flat mirror (beamsplitter) (2) behind the spherical 

1644: end mirror (1). A photodetector of the main alignment system is 

1645: denoted (3), the focusing lens - (5), the last stage of anti-seismic 

1646: filter - (6), the additional photodetector - (7).}

1647: \end{figure}

1648: 

1649: \subsection{\label{sla} Additional laser behind the end mirror.}

1650: 

1651: The most natural solution to the formulated problem is, perhaps, the 

1652: installation of an additional source of light (for some extra details on this 

1653: proposal, see \cite{BBG}). Imagine that the additional laser is suspended 

1654: (possibly, from the same anti-seismic filter) behind the spherical end 

1655: mirror and shines inside the main cavity, see Fig.\ref{f6}. In order to 

1656: avoid any mixing of the new light with the already present light in the 

1657: cavity, the frequency of the auxiliary laser should differ from the frequency

1658: of the main laser. However, inside the cavity, the added light satisfies all 

1659: the usual requirements on mode matching, alignment of the light beam with the 

1660: optical axis, etc. A possible misalignment of the new beam with the optical 

1661: axis of the Fabry-Perot resonator will be sensed by the same alignment system 

1662: and in exactly the same manner as it takes place for the main beam. The only 

1663: difference is that the detected misalignment of a new beam will be recorded 

1664: for geophysical studies rather than used for correcting the inclination of 

1665: the end mirror. One will also need a new photodetector for the output light.  

1666: 

1667: The auxiliary laser and the end mirror are hanging very close to each other, so

1668: they react in the same way to shifts of the suspension point or tilts of the 

1669: plumb line. If the laser itself and the end mirror are both subject to a shift 

1670: of the suspension point (configuration c) in Fig.\ref{mirrors}) the alignment 

1671: of the additional beam with the optical axis of the Fabry-Perot resonator 

1672: remains intact. The system does not generate an error signal. In contrast, 

1673: a tilt of the plumb line (configuration b) in Fig.\ref{mirrors}) will 

1674: cause a tilt of direction 

1675: of the incoming beam of the auxiliary laser. This tilt will be detected by the 

1676: alignment system, thus distinguishing the configurations b) and c). The arising 

1677: error signal is what we are after. Assuming that the angular sensitivity is 

1678: still determined by Eq.(\ref{e3a}), one will get access to

1679: the information on $\Delta \alpha_0$. In fact, the installation of an extra 

1680: laser with properties similar to the ones described above was originally 

1681: discussed in a different context, with the aim of facilitating the initial 

1682: tuning of the main interferometer.

1683: \begin{figure}

1684:   \begin{center}

1685: \includegraphics{fig6.eps}

1686: %\epsfbox{fig6.eps}

1687:   \end{center}

1688:   \caption{\label{f6}  Additional laser behind the end mirror: (1) and

1689: (2) - the front and the end mirrors of the main interferometer; 

1690: (3) - beamsplitter; (4) - additional laser; (5) - additional photodetector.}

1691: \end{figure}

1692: 

1693: 

1694: \section{Conclusions}

1695: 

1696: The need for successful functioning of laser-beam detectors of 

1697: gravitational waves inevitably makes them valuable 

1698: geophysical instruments. The most immediate application of g.w. 

1699: detectors is accurate measurement of low-frequency distance

1700: variations between the central mirror and the end 

1701: mirrors. Essentially, the adjustment control system makes the 

1702: interferometer a low-frequency 3-point meter of deformations. The 

1703: principal advantage of the existing g.w. detectors in comparison 

1704: with traditional geophysical techniques is exceptionally

1705: long arms of interferometers (a few kilometers in comparison with 

1706: typical hundred-meter long geophysical interferometers) and high 

1707: sensitivity of measurements. This advantage leads to a much better

1708: strain sensitivity. The extraction of deformational part of the 

1709: geophysical signal does not imply any hardware modifications, and 

1710: it can be found in records of the adjustment system right at the 

1711: level of data analysis.

1712: 

1713: The alignment system of g.w. laser interferometers on suspended

1714: mirrors can potentially provide even more interesting geophysical 

1715: information - the relative variation of local gravity vectors 

1716: (plumb lines). This sort of information cannot be obtained from 

1717: geophysical interferometers which normally use mirrors attached 

1718: to the ground. It appears, however, that the extraction of this 

1719: information from the existing g.w. laser interferometers will require

1720: certain modest modifications of their optical scheme. Without 

1721: compromising the astrophysical program of the instruments, these 

1722: modifications seem to be possible, and we suggested several ways 

1723: of doing this. Their realization will allow one to use angular 

1724: measurements in addition to distance measurements. In future

1725: advanced g.w. interferometers based on non-spherical mirrors these

1726: modifications may be avoided.

1727: 

1728: The evaluation of interesting geophysical signals and 

1729: comparison with environmental and instrumental noises shows

1730: that the g.w. instruments can reach the most ambitious geophysical

1731: goals, such for example as detection of oscillations of the inner 

1732: core of Earth. 

1733: 

1734: It seems that further progress in this area can be achieved, 

1735: if a closer collaboration is established between geophysicists and 

1736: representatives of g.w. community. 

1737: 

1738: \section*{Acknowledgments} 

1739: 

1740: At various stages of this research we have benefited from discussions

1741: with P. Bender, R.~Drever, J. Hough, A. Kopaev, K. Strain, K. Tsubono, 

1742: H. Ward and, especially, A.~Giazotto. This work was partially 

1743: supported by a grant from the Royal Society (UK) for international 

1744: collaboration, RCPX331.  

1745: 

1746: \vspace{1cm}

1747: 

1748: {\it Note added:} In the very last days of preparation of this paper

1749: for publication we have tragically lost our young talanted colleague

1750: and friend - Andrej Serdobolski.

1751: 

1752: \section*{References}

1753: 

1754: \begin{thebibliography}{99}

1755: 

1756: \bibitem{thorne} Thorne K S 1987. In: {\it Three Hundred Years

1757: of Gravitation}, Eds. S. W. Hawking and W. Israel (Cambridge: 

1758: CUP) p~ 330

1759: 

1760: \bibitem{gr5} Grishchuk L P, Lipunov V M, Postnov K A, Prokhorov

1761: M V and Sathyaprakash B S 2001. Usp. Fiz. Nauk, {\bf 171} p~ 3

1762: [Physics-Uspekhi 2001 {\bf 44} p~ 1, {\it English transl.} ]  

1763: 

1764: \bibitem{cthorne} Cutler C and Thorne K S 2002. In: {\it General

1765: Relativity and Gravitation, GR16}, Eds. N. T. Bishop and S. D.

1766: Maharaj (World Scientific) p~ 72

1767: 

1768: \bibitem{gr} Grishchuk L P 2004. In: {\it Astrophysics Update},

1769: Ed. J. W. Mason (Springer-Praxis) p~ 281 (gr-qc/0305051)

1770: 

1771: \bibitem{ligoweb} LIGO website http://www.ligo.caltech.edu

1772: 

1773: \bibitem{GEOweb} GEO website http://www.geo600.uni-hannover.de

1774: 

1775: \bibitem{virgoweb} VIRGO website http://www.virgo.infn.it 

1776: 

1777: \bibitem{TAMAweb} TAMA website http://tamajo.mtk.nao.ac.jp

1778: 

1779: %\bibitem{1} VIRGO Final design report, (CNRS, INFN, issue 1, 1997)

1780: 

1781: %\bibitem{2} Abramovici A \etal 1992

1782: %Science, {\bf  256}, p~325

1783: 

1784: %\bibitem{3} Danzmann K \etal 1994

1785: %{\it Report 190,  Max-Planck-Institut  fur Quantenoptik}

1786: %(Berlin:Garching)

1787: 

1788: %\bibitem{4} Tsubono K \etal 1995

1789: %{\it Proc. of the  First E. Amaldi Conf.: Gravitational Wave Experiments}

1790: %(Singapore:W. Sci.)

1791: 

1792: \bibitem{4a} Landau L D and Lifshiz E M 1958

1793: {\it Mechanics} (Pergamon Press - London)  

1794: 

1795: \bibitem{c12} Melchior P 1983 

1796: {\it The Tides of the Planet Earth} (Oxford: Pergamon press)

1797: 

1798: \bibitem{raab} Raab F. and Fine M. {\it The effect of Earth tides

1799: on LIGO interferometers}, LIGO document T970059-01, 2/20/97.

1800: 

1801: \bibitem{bullen} Bullen K 1978

1802: {\it Earth Density} (Moscow: Mir, Russian Transl.)

1803: 

1804: \bibitem{mf} Morse P. M. and Feshbach H. {\it Methods of

1805: Theoretical Physics}, (McGraw-Hill, New York, 1953), Part II, Ch. 11. 

1806: 

1807: \bibitem{15a} Ness N. F., Harrison J. C. and Slichter L. B.

1808: 1961 J. Geophys. Research, {\bf 65}, p~ 2 

1809: 

1810: \bibitem{15b} Benioff H., Press F. and Smith S. 

1811: 1961 J. Geophys. Research, {\bf 65}, p~2 

1812: 

1813: \bibitem{16} Slichter L B 1961

1814: Proc. Nat. Acad. Sci. USA {\bf 47} p~186

1815: 

1816: \bibitem{tu} Tanimoto T. and Um J. 1999 J. Geophys. Research. {\bf 104},

1817: B12, p~723.28 

1818:  

1819: \bibitem{c11} Jacobs J A 1987 {\it The Earth's Core}

1820: (London: International Geophysics Series, Ed. W. L. Donn, Academic press)

1821: 

1822: \bibitem{17} Busse F H 1974

1823: J. Geophys. Research, {\bf 79} 5 p~753

1824: 

1825: \bibitem{18} Smylie D E 1992

1826: {\it Science} {\bf 255} p~1678

1827: 

1828: \bibitem{5aa} Courtier N \etal 2000

1829: Physics of the Earth and Planetary Interiors {\bf 117} p~3

1830: 

1831: \bibitem{1a} Kulagin V, Rudenko V., Pasynok S \etal 2000

1832: (Tokyo:Frontier Science Series ed. S.Kawamura, N.Mio,

1833: Univ. Acad. Press) {\bf 32} p~343

1834: 

1835: \bibitem{RP} Rudenko V and Pasynok S 1997

1836: Frontier Sci. Series (Tokyo: Univ. Acad. Press Inc.) {\bf 20} p~63

1837: 

1838: \bibitem{5} Rudenko V 1996

1839: Phys. Lett. A {\bf 223} p~421

1840: 

1841: %\bibitem{6} Kopaev A and Rudenko V 1994

1842: %{\it JETP Letters} {\bf 59} 9 p~662

1843: 

1844: \bibitem{saulson} Saulson P R 1984

1845: Phys. Rev. D {\bf 30} p~732

1846: 

1847: \bibitem{beccaria} Beccaria M \etal 1998

1848: Class. Quant. Grav. {\bf 15} p~3339

1849: 

1850: \bibitem{hughesthorne} Hughes S A and Thorne K S 1998

1851: Phys. Rev. B {\bf 58} p~1202

1852: 

1853: \bibitem{AR} Aki K and Richards P 1980

1854: Quantutaitive Seismology. Theory and Methods

1855: (San Francisco:ed. W.H.Freeman) {\bf 1}

1856: 

1857: \bibitem{angew} Angew D C 1986

1858: Review of Geophysics {\bf 3} p~579

1859: 

1860: \bibitem{22b} Rudenko V, Milyukov V, Nesterov V and Ivanov I 1994

1861: Astronomical and Astrophysical Transactions {\bf 5} p~93

1862: 

1863: \bibitem{daw} Daw E J {\it et al.} {\it Long term study of the

1864: seismic environment at LIGO}, gr-qc/0403046 

1865: 

1866: \bibitem{richetal} Richter B {\it et al.} 1995 Marees Terrestres, 

1867: Bulletin d'Informations, N~122, pp. 1963-1972  

1868: 

1869: \bibitem{c10} He-Ping Sun 1995

1870: {\it Static deformation and gravity changes at the Earth surface due to

1871: atmospheric pressure}

1872: (Bruxellis: Docteur en Sciences Thesis Observatuire Royal de Belgique)

1873: 

1874: \bibitem{c13} Hanada H 1996

1875: Publ. of the Nat. Astron. Observatory of Japan, Tokyo

1876: {\bf 4} 75--78

1877: 

1878: \bibitem{22a} Rudenko V, Serdobolski A and Tsubono K 2003

1879: Class. Quant. Grav. {\bf 20}, p~317

1880: 

1881: \bibitem{TA} Grachev A 1998

1882: Izv. Atmosph. Ocean. Phys. {\bf 34} p~570

1883: 

1884: \bibitem{GGM} Gurashvili V, Gusev G, Manukin A, Nikolaev A. 1999

1885: Fizika Zemli (Physics of Earth) {\bf 5} p~14

1886: 

1887: \bibitem{LL} Landau L D and Lifshitz E M 1959 {\it Fluid Mechanics}

1888: (Pergamon Press - London) 

1889: 

1890: \bibitem{sho} Schoemaker D. 2003 Class. Quant. Grav. {\bf 20}, 10, p~ S11

1891: 

1892: \bibitem{MT1} Drever R W P, Hall J L, Kowalski F V \etal 1983

1893: Appl. Phys. B {\bf 31} p~97

1894: 

1895: \bibitem{MT2} Drever R W P, Hall J L, Kowalski F V \etal 1983

1896: Appl. Phys. B {\bf 32} p~7

1897: 

1898: \bibitem{8} Anderson D 1984

1899: Applied Optics,  {\bf 23} p~2944

1900: 

1901: \bibitem{9} Sampas N and Anderson D 1990

1902: Applied Optics, {\bf 29} p~394

1903: 

1904: \bibitem{9a} Kulagin V, Pasynok S, Rudenko V and Serdobolsky A 2001

1905: {\it Large scale laser gravitational interferometer with suspended mirrors

1906: for fundamental geodynamics} {Moscow:Laser Optics'2000, Solid State

1907: Lasers, Proc. of SPIE} {\bf 4350} p~178

1908: 

1909: \bibitem{ward} Morrison E, Meers B J, Robertson D I, and Ward H 1994

1910: Applied Optics, {\bf 33} 50; {\it ibid} {\bf 33} 5037    

1911: 

1912: \bibitem{bonthorne} Bondarescu M and Thorne K S {\it A new family

1913: of light beams and mirror shapes for future LIGO interferometers},

1914: gr-qc/0409083 

1915: 

1916: \bibitem{DR} Drever R 1997

1917: Frontier Sci. Series (Tokyo: Univ. Acad. Press Inc.) {\bf 20} p~299

1918: 

1919: \bibitem{DA} Drever R and Augst S 2000

1920: Frontier Sci. Series (Tokyo: Univ. Acad. Press Inc.) {\bf 32} p~75

1921: 

1922: \bibitem{tsu} Tsubono K \etal 2004 Physics Lett. {\bf A 327}, 1

1923:  

1924: \bibitem{BBG} Bradaschia C, Braccini S, Giazotto A \etal 2001

1925: {\it Proc. MG-9, part 2} (World Scientific) p~345

1926: 

1927: %\bibitem{com} Gusev G, Manukin A 1985

1928: %Izvestia Rus. Ac. Sci., Physics of Earth, {\bf 9}, p~90

1929: 

1930: %\bibitem{22} Braccini S \etal 1998

1931: %Class. Quant. Grav. {\bf 15} p~3339

1932: 

1933: %\bibitem{tam} Kuroda K, Ohashi M, Fujimoto M K \etal 1999

1934: %Int. J. Mod. Phys. D {\bf 8} No.5 p~557-9

1935: 

1936: 

1937: \end{thebibliography}

1938: \end{document}     

1939: