1: %\documentstyle[epsfig,emulateapj]{article}

2: %\documentclass{aastex}

3: \documentclass[preprint]{aastex}

4: %\documentclass[preprint2]{aastex}

5: 

6: \usepackage{emulateapj5}

7: 

8: \newcommand{\gsim}{\mbox{\raisebox{-1.0ex}{$\stackrel{\textstyle >}

9: {\textstyle \sim}$ }}} 

10: \newcommand{\lsim}{\mbox{\raisebox{-1.0ex}{$\stackrel{\textstyle <}

11: {\textstyle \sim}$ }}}  

12: \newcommand{\gtsima}{$\; \buildrel > \over \sim \;$}

13: \newcommand{\ltsima}{$\; \buildrel < \over \sim \;$}

14: \newcommand{\simgt}{\lower.5ex\hbox{\gtsima}}

15: \newcommand{\simlt}{\lower.5ex\hbox{\ltsima}}

16: \newcommand{\himpc}{{\hbox {$h^{-1}$}{\rm Mpc}} }

17: \newcommand{\bfx}{{\mbox{\boldmath $x$}}}

18: \newcommand{\bfth}{{\mbox{\boldmath $\theta$}}}

19: \newcommand{\thf}{{\theta_{\rm f}}}

20: 

21: %\slugcomment{Submitted to Astrophys. J.}

22: 

23: %\shortauthors{Matsubara \& Jain} 

24: %\shorttitle{Genus Statistics from Gravitational Lensing}

25: 

26: \begin{document}

27: 

28: %\renewcommand{\thefootnote}{\fnsymbol{footnote}}

29: %\renewcommand{\theequation}{\mbox{\rm

30: %{\arabic{section}.\arabic{equation}}}} 

31: %\renewcommand{\theequation}{\mbox{\rm {\arabic{equation}}}} 

32: 

33: %MMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMM

34: 

35: \title{

36: The Topology of Weak Lensing Fields

37: }

38: 

39: \author{Takahiko Matsubara}

40: \affil{Department of Physics and Astrophysics, 

41: Nagoya University, Chikusa-ku, Nagoya 464-8602, Japan

42: }

43: \and

44: \author{Bhuvnesh Jain}

45: \affil{Department of Physics and Astronomy, University of Pennsylvania,

46: Philadelphia, PA 19104

47: }

48: 

49: \email{taka@a.phys.nagoya-u.ac.jp, bjain@hep.upenn.edu}

50: 

51: 

52: \begin{abstract}

53: The topology of weak lensing fields is studied using the 2-dimensional

54: genus statistic. Simulated fields of the weak lensing convergence

55: are used to focus on the effect of nonlinear gravitional evolution and

56: to model the statistical errors expected in observational surveys. For large

57: smoothing angles, the topology is in agreement with

58: the predictions from linear theory. On smoothing angles smaller than 

59: $10'$, the genus curve shows the non-Gaussian signatures of gravitational

60: clustering and differs for open and flat cold dark matter models. 

61: Forthcoming surveys with areas larger than 10 square degrees should

62: have adequate signal-to-noise to measure the 

63: non-Gaussian shape and the $\Omega$-dependence of the genus statistic. 

64: \end{abstract}

65: 

66: %MMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMM

67: 

68: \keywords{cosmology: theory --- gravitational lensing --- large-scale

69: structure of universe --- methods: numerical}

70: 

71: \setcounter{equation}{0}

72: \section{Introduction}

73: \label{sec1}

74: 

75: Mapping the mass density field with weak gravitational lensing is a

76: promising method to analyze large-scale structure in the universe. 

77: The statistical nature of the mass density field is

78: directly predicted by theories of the early universe. For instance,

79: most inflationary theories predict a random Gaussian initial mass

80: density field, while structure formation by cosmic defects predicts a

81: highly non-Gaussian field. As the density field evolves, gravitational 

82: clustering produces characteristic non-Gaussian features. 

83: Thus measuring the non-Gaussianity of the density field can be an 

84: important test of the origin and evolution of large-scale structure. 

85: 

86: The genus statistic is known to be a useful topological discriptor of

87: the the clustering pattern of the mass density field

88: \citep{got86,got89}. The ``genus curve'', which plots the measured

89: genus versus the threshold for iso-density contours, has a universal

90: shape for a random Gaussian field. Departures from this Gaussian shape

91: reveal the mechanism of the formation of nonlinear structures. So far,

92: the genus statistic has been applied to the distribution of galaxies,

93: and of clusters \citep{moo92,rho94,can98}. Since these tracers are

94: affected by biasing, it is desirable to directly use the mass density

95: field for the genus statistic. Fortunately, recent developments in

96: measuring weak lensing in wide area surveys suggest that this may be

97: feasible in the near future \citep{van00b,bac00,wit00,kai00}. As weak

98: lensing maps of the convergence field, $\kappa$, cover larger areas

99: with better signal-to-noise, statistical properties beyond just the

100: second moment can be measured. It is therefore important to obtain

101: theoretical predictions for the behavior of the genus statistic for

102: the weak lensing field, which we present in this paper.

103: 

104: The linear-scale fluctuations of the mass density field preserves the

105: Gaussianity of the initial state. However, measuring the linear-scale

106: fluctuations is difficult due to sample variance, unless the area

107: covered is extremely large. In forthcoming weak lensing surveys, the

108: convergence field is likely to contain non-linear effects of

109: gravitational clustering. The genus curve is distorted by the

110: non-Gaussianity induced by gravitational evolution, which

111: precludes a simple comparison of the observed genus with analytical

112: predictions.  Since this effect is dependent on the cosmological

113: model, it is of interest to establish whether forthcoming surveys will

114: allow for an adequate measurement of the non-Gaussian genus curve. 

115: 

116: We show in this paper that the effects of gravitational clustering on the

117: genus curve can be measured from high

118: quality data. For this purpose, we analyze the ray-tracing numerical

119: simulation of weak lensing developed by Jain, Seljak \& White (2000).

120: The noise due to the finite number of galaxies used to probe the weak

121: lensing field reduces the signal-to-noise on very small scales.

122: Therefore we have to balance sample variance on large scales with

123: discreteness noise on small scales to find the best length-scales for

124: measuring the genus. We study this issue by simulating the random

125: noise expected in observational surveys.

126: 

127: This paper is organized as follows. \S\ref{sec2} introduces the genus

128: statistic. In \S\ref{sec3}

129: the main results of our analysis from ray tracing simulations 

130: are presented. The genus

131: curves of simulated noise-free convergence fields are shown in \S\ref{sec3-1}

132: and those of noisy fields in \S\ref{sec3-2}. We conclude in \S\ref{sec4}. 

133: 

134: \section{Genus Statistics and Weak Lensing}

135: \label{sec2}

136: 

137: We consider the local convergence field $\kappa$ 

138: and calculate genus statistics of this field. In the weak lensing regime,

139: $\kappa \ll 1$, an area in the lens plane is magnified by a factor $1 +

140: 2\kappa$ compared to that in the source plane. The local convergence

141: field can be reconstructed from the shear field \citep{kai93}. For a 

142: distant observer, under the Born approximation, the convergence

143: along a line of sight 

144: can be expressed as a projection of the density perturbation $\delta$,

145: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

146: \begin{eqnarray}

147:    \kappa = \frac{3{H_0}^2 \Omega_0}{2}

148:    \int_0^\chi

149:    \frac{d_{\rm A}(\chi')d_{\rm A}(\chi-\chi')}{d_{\rm A}(\chi)} 

150:    \frac{\delta(\chi')}{a(\chi')} d\chi',

151: \label{eq2}

152: \end{eqnarray}

153: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

154: where $\Omega_0$ is the present value of the mass density parameter,

155: $H_0$ is the Hubble's constant, $a(\chi)$ is the scale factor at the

156: conformal lookback time $\chi$, $d_{\rm A}(\chi)$ is the comoving

157: angular diameter distance, $\chi$ the comoving radial coordinate. If

158: the radial distance is measured in units of ${H_0}^{-1}$, we can drop

159: the ${H_0}^2$ in the above equation. Integration is performed along

160: the unperturbed light path. In this case, the statistics of the

161: convergence field can be computed from the statistics of the density

162: field. 

163: 

164: To quantify the topology, we find contour lines of a threshold value

165: labelled by $\nu$. For a Gaussian random field, the threshold $\nu$ is

166: identified so that the field takes the value $\kappa = \nu\sigma$,

167: where $\sigma$ is the rms of the field, $\sigma^2 = \langle \kappa^2

168: \rangle$. The genus is defined as: (number of contours surrounding

169: regions higher than the threshold value) - (number of contours

170: surrounding regions lower than the threshold value)

171: \citep{adl81,col88,mel89,got90}. The genus per unit area $G_2(\nu)$,

172: as a function of the threshold is known as the genus curve.

173: 

174: The genus curve is analytically given by \citep{mel89}:

175: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

176: \begin{eqnarray}

177:    G_2 (\nu) = 

178:    \frac{1}{(2\pi)^{3/2}}

179:    \frac{\langle k^2 \rangle}{2} \, \nu \,

180:    e^{-\nu^2/2},

181: \label{eq1}

182: \end{eqnarray}

183: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

184: where $\langle k^2 \rangle$ is the square of the wave number $k$

185: averaged over the smoothed 2D power spectrum. Both sides of the

186: above equation have dimensions of $[{\rm steradian}]^{-2}$. 

187: If the 3D power spectrum is given by a power law with spectral 

188: index $n$, and a Gaussian window function $W(r) \propto

189: e^{-r^2/(2\thf^2)}$ is adopted, then $\langle k^2 \rangle =

190: (n+2)/(2\thf^2)$ for $n>-2$. 

191: 

192: The above analytic results are not expected to hold for 

193: ongoing weak lensing surveys because the survey areas are

194: not large enough to probe the linear scales. However, it is known that

195: the shape of the genus curve of mass density fields

196: is not strongly affected by nonlinear gravitational evolution

197: provided we use a rescaled labeling of $\nu$ \citep{got87,wei87,mel88}. 

198: For a non-Gaussian field, the threshold is defined

199: through the fraction of area $f$ on the high-density side of the contour

200: lines, which is related to the corresponding values of $\nu$ for a

201: Gaussian field as:

202: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

203: \begin{eqnarray}

204:    f =

205:    \frac{1}{\sqrt{2\pi}}

206:    \int_\nu^\infty e^{-t^2/2} dt.

207: \label{eq3}

208: \end{eqnarray}

209: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

210: For a Gaussian field, this definition gives the same result as

211: specifying the threshold directly in units of $\sigma$. For a

212: non-Gaussian field, it compensates for the horizontal shift of the genus curve

213: due to the shift of the probability distribution function. 

214: %It also makes the genus curve independent of the amplitude of the power

215: %spectrum, though it depends on the shape through the  $\langle k^2

216: %\rangle$ factor. 

217: We adopt this prescription for setting the threshold

218: values of $\kappa$ in the results shown below.

219: 

220: 

221: \section{Numerical Results}

222: \label{sec3}

223: 

224: \subsection{Genus without noise}

225: \label{sec3-1}

226: 

227: We use maps of convergence fields from the ray-tracing simulations of 

228: Jain, Seljak \& White (2000). The source galaxies are taken to be 

229: at $z=1$. We use 

230: %three cosmological

231: %models, an Einstein-de Sitter (EdS) model with $\Omega_0=1$, $\Gamma =

232: %0.5$, 

233: two cold dark matter models: a ``tau'' model with $\Omega_0=1$, and

234: shape parameter $\Gamma = 0.21$, and an open model with $\Omega_0 =

235: 0.3$, $\Gamma = 0.21$. The fields we use are $166.28$ arcminutes on a

236: side for the tau model, and $235.68$ arcminutes for the open model,

237: sampled with 2048x2048 grids \citep[see][for further details]{jai00}.

238: Figure \ref{fignoise1} shows a $\kappa$ map for the open model.

239: 

240: The genus curves are calculated using the CONTOUR2D code and related

241: routines, which were kindly provided by David Weinberg (see

242: \citet{got86} and \citet{mel90} for details of the methods). The key

243: relation used is a 2-dimesional analog of the Gauss-Bonnet theorem,

244: which expresses the genus in terms of the integral of the curvature

245: along a contour line, $G_2=\int C\, ds/2\pi$, where $C$ is given by

246: the inverse radius of curvature $r^{-1}$, and is negative or positive,

247: depending on whether a low or high density region is enclosed. For

248: each model, genus curves are obtained for 7-10 realizations. The error

249: bars on the genus curves shown in Figures \ref{figgenus1} and

250: \ref{fignoise2} give the $1\sigma$ deviations among those

251: realizations, and represent the $1\sigma$ error expected for surveys

252: of our sample area.

253: 

254: In the left panels of each plot in Figure~\ref{figgenus1}, the genus curves per

255: steradian of the simulated convergence fields are plotted for the open and

256: the tau models. Each plot has a different smoothing angle, $\thf = 1',

257: 2', 3', 10'$. 

258: %In Figure~\ref{fig2}, the standard model is also plotted

259: %for $\thf = 2'$. 

260: In the plots, thin curves are the Gaussian predictions with

261: normalizations fitted by minimizing chi-squares of the simulation

262: data. The smoothing angle of $10'$ corresponds to a scale on which

263: linear theory can be applied (e.g., see Fig. 3 of Jain \& Seljak

264: 1997). Within the sample variance error bars, the genus curves of the

265: simulation data are consistent with Gaussian predictions on this

266: linear scale.

267: 

268: On smaller, nonlinear scales of $\thf = 1', 2', 3'$, the genus curves

269: of the convergence field show a ``meatball shift'': a dominance of the

270: number of isolated clusters over isolated voids. This shift is

271: consistent with topological analyses of the galaxy distributions

272: \citep{got89,got92}, and is a common feature of cold dark matter

273: models \citep{spr98}. To quantify the degree of the meatball

274: shift, we define a meatball-shift statistic, $S=\int G_2(\nu)

275: d\nu/\int |G_2(\nu)| d\nu$, where the integration is limited from $\nu

276: = -2$ to $+2$. Likewise, the departure from Gaussianity can be 

277: quantified by a statistic like: $S_2=\int |G_2(\nu)-G_{\rm

278: gauss}(\nu)|/\int |G_{\rm gauss}|$. We discuss some results in the 

279: following sub-section.  

280: %Table~\ref{tab1} shows the numerical values for this

281: %statistic. The genus in the open model is shifted more than in the tau

282: %model becuase of the evolutionary effect.}\placetable{tab1}

283: 

284: 

285: The absolute values of the genus number of the high-convergence side

286: ($\nu > 0$) is larger for the open model, and smaller for the tau

287: model. Since our open and tau models have the same initial power

288: spectrum, the difference in the genus number is due to the difference

289: in the nonlinear gravitational evolution, and, more importantly, in

290: the space-time geometry. A volume factor $V(z)$ for a unit solid angle

291: is larger in the open model than in the tau model. Nonlinear structures

292: also form earlier in the open model. Both these effects cause the 

293: topology statistics for the open model to have different non-Gaussian

294: features from the tau model. 

295: 

296: \subsection{Genus with realistic noise}

297: \label{sec3-2}

298: 

299: The weak lensing fields in observational surveys contain

300: noise due to the finite number of source galaxies with randomly

301: oriented intrinsic ellipticities. 

302: We build simulated noisy maps of the convergence field by drawing each

303: component of the intrinsic ellipticity of a galaxy from 

304: a Gaussian with variance \citep[e.g.,][]{van00}

305: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

306: \begin{eqnarray}

307:    \sigma_{\rm noise}^{\,2} =

308:    \frac{\sigma_{\epsilon}^{\,2}}{2 N_g},

309: \label{eq4}   

310: \end{eqnarray}

311: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

312: where $\sigma_\epsilon$ is the rms amplitude of the intrinsic

313: ellipticity distribution and $N_g$ is the mean number density of

314: source galaxies in each cell. In the following, the number density of

315: source galaxies is $30$ per square arcminute, and their rms intrinsic

316: ellipticity is $\sigma_\epsilon = 0.2$.

317: 

318: Figure~\ref{fignoise2} shows the genus measured from a $\kappa$

319: field reconstructed from pure Gaussian noise which corresponds to a

320: map in which galaxies are randomly distributed with random

321: ellipticities without weak lensing signals. In fact, any white noise

322: has the same genus curve of equation (\ref{eq1}) regardless of its

323: variance. The agreement with the analytic curve for a random

324: Gaussian field (Eq.~\ref{eq1}) tests our reconstruction scheme and

325: genus measurement. In Figure~\ref{figgenus1} the right panels in each

326: plot show the genus curves of simulated signal+noise maps of the

327: convergence fields. 

328: %{\bf In Table~\ref{tab2}, the meatball-shift

329: %statistic for the same samples is shown.}\placetable{tab2}

330: 

331: For the large smoothing angle $\thf =10'$, the genus curves are again

332: consistent with the Gaussian predictions. For $\thf=3'$, the noise

333: effect is small, but the meatball shift is reduced from the noise-free

334: maps. For $\thf=2'$, the meatball shift is completely erased by noise

335: except in the vicinity of $\nu = 0$. For $\thf=1'$, there are

336: interesting features due to the noise. The relative effect of noise is

337: stronger for the open model because the lensing signal is small [see

338: Eq.~\ref{eq2}]. The high-convergence region of $\nu \sim 2$ is not

339: much affected by the noise. This is because most of the

340: high-convergence peaks come from strongly clustered regions. Around

341: $\nu \sim 1$, the noise contribution equals the signal. For the

342: low-convergence region of $-\nu \sim 1$--$2$ the genus is dominated by

343: noise because relatively large areas of low-convergence (which

344: corresponds to voids) have noise troughs superimposed on them. Such

345: complex situations make the genus curve have larger amplitude than the

346: Gaussian at negative $\nu$ for $\thf=1'$--the opposite of the case for

347: the noise-free maps. There is an interesting counterpart to this

348: effect in the statistics of peaks studied by \citet{jai00b}; they also

349: found an increase in the number of negative peaks (relative to the

350: Gaussian case) in the signal+noise maps.

351: 

352: These effects on the genus curve discussed above can be quantified

353: using the statistics giving the meatball shift and departure from

354: Gaussianity, $S$ and $S_2$. For the meatball shift statistic we find

355: that the open model shows a significant negative (sponge-like) shift

356: for $\theta=1'$: $S=-0.08\pm 0.01$. It shows a positive (meatball)

357: shift for large angles, e.g. for $\theta=3'$, $S=0.06\pm0.02$. The tau

358: model shows a similar trend but with lower amplitude. This is in

359: contrast to the noise-free field, for which there is a positive

360: meatball shift which decreases in amplitude with increasing angular

361: scale.

362: 

363: 

364: \section{Discussion}

365: \label{sec4}

366: 

367: We have analyzed the genus curve of weak lensing convergence fields. On

368: large angular scales, the genus curves are consistent with the random

369: Gaussian prediction. On scales of a few arcminutes, nonlinear

370: gravitational clustering produces a non-Gaussian ``meatball'' 

371: shift in the genus

372: curve. Its non-Gaussianity also differentiates the open and flat 

373: cosmological model for a range of smoothing angles and threshold values.  

374: The noise dilutes the meatball shift but it does not change

375: the genus curve to the random Gaussian one, 

376: even on the smallest scale $\thf=1'$. 

377: Over the range of scales tested, $\thf=1'$--$10'$, the open and flat

378: cold dark matter models can be distinguished even in the presence of noise. 

379: %The genus curve has a strange shape when both nonlinear and

380: %noise effects contribute to the sample as prominently shown in open

381: %model with $\thf = 1'$ in Figure~\ref{figgenus1}. However, when the noise

382: %effect is not large as in $\thf = 3'$, the shape of the genus curve is

383: %close to the random Gaussian one, even though the scales probed are

384: %well within a nonlinear regime. There are still slight meatball shift,

385: %consistent with the topology of the mass density fields.

386: 

387: We have analysed simulations of survey areas of about 10 square

388: degrees. When survey areas approach 100 square degrees, the error bars

389: shown in Fig.~\ref{figgenus1} should reduce by about a factor of

390: three. Thus, provided systematic errors can be kept to below the level

391: of the random noise modeled here, the genus statistic can provide a

392: valuable means of measuring the non-Gaussian features of the lensing

393: signal and distinguishing cosmological models. Further work to measure

394: other topological statistics and simulate a wider set of cosmological

395: models would be useful.

396: 

397: \acknowledgements

398: 

399: We are very grateful to David Weinberg for providing us useful

400: routines to calculate the 2D genus curve and for stimulating

401: discussions. We thank an anonymous referee for suggestions that 

402: helped improve the paper. T.\ M.\ acknowledges support

403: from JSPS Postdoctoral Fellowships for Research Abroad. B.\ J. \

404: acknowledges support from NASA through grants NAG5-9186 and NAG5-9220.

405: 

406: \begin{thebibliography}{}

407: 

408: \bibitem[Adler(1981)]{adl81} Adler, R. J. 1981, R. K. 1986, The

409: Geometry of Random Fields (Chichester; Wiley)

410: \bibitem[Bacon, Refregier \& Ellis 2000]{bac00} Bacon, D., Refregier,

411: A. \& Ellis, R. 2000, to appear in Proceedings of XXth Moriond

412: Astrophysics Meeting, "Cosmological Physics with Gravitational

413: Lensing", eds. J.-P. Kneib, Y. Mellier, M. Moniez and J. Tran Thanh

414: Van (astro-ph/0008248)

415: %\bibitem[Bartelmann \& Schneider 1999]{bar99} Bartelmann, M. \& Schneider,

416: %P. 1999, to be submitted to Physics Reports, astro-ph/9912508

417: %\bibitem[Bernardeau et al.(1997)]{ber97} Bernardeau, F., van

418: %Waerbeke, L. \& Mellier, Y. 1997, \aap, 322, 1

419: \bibitem[Canavezes et al.(1998)]{can98} Canavezes, A., Springel, V. et

420: al. 1998, \mnras, 297, 777

421: \bibitem[Coles(1988)]{col88} Coles, P. 1988, \mnras, 234, 509

422: \bibitem[Gott, Weinberg \& Melott(1987)]{got87} Gott, J. R., Weinberg,

423: D. H. \& Melott, A. L. 1987, \apj, 319, 1

424: \bibitem[Gott, Melott \& Dickinson(1986)]{got86} Gott, J. R., Melott,

425: A. L. \& Dickinson, M. 1986, \apj, 306, 341

426: \bibitem[Gott et al.(1989)]{got89} Gott, J. R. et al. 1989, \apj,

427: 340, 625

428: \bibitem[Gott et al.(1990)]{got90} Gott, J. R., Park, C.,

429: Juszkiewicz, R., Bies, W. E., Bennett, D. P., Bouchet, F. R. \&

430: Stebbins, A. 1990, \apj, 352, 1

431: \bibitem[Gott et al.(1992)]{got92} Gott, J. R., Mao, S.,

432: Park, C. \& Lahav, O. 1992, \apj, 385, 26 

433: \bibitem[Jain \& Seljak(1997)]{jai97} Jain, B. \& Seljak, U. 1997, 

434: \apj, 484, 560

435: \bibitem[Jain, Seljak, \& White(2000)]{jai00} Jain, B., Seljak, U. \&

436: White, S. 2000, \apj, 530, 547

437: \bibitem[Jain, van Waerbeke(2000)]{jai00b} Jain, B. \& van Waerbeke,

438: L. 2000, \apj, 530, L1

439: %\bibitem[Kaiser(1987)]{kai87} Kaiser, N. 1987, \mnras, 227, 1

440: %\bibitem[Kaiser(1992)]{kai92} Kaiser, N. 1992, \apj, 388, 272 

441: \bibitem[Kaiser \& Squires(1993)]{kai93} Kaiser, N. \& Squires, G.

442: 1993, \apj, 404, 441

443: %\bibitem[Kaiser(1998)]{kai98} Kaiser, N. 1998, \apj, 498, 26 

444: \bibitem[Kaiser, Wilson \& Luppino(2000)]{kai00} Kaiser, N., Wilson, G.

445: \& Luppino, G. A. 2000, submitted to ApJ Letters (astro-ph/0003338) 

446: %\bibitem[Mellier(1999)]{mel99} Mellier, Y. 1999, ARA\&A, 37, 127

447: \bibitem[Melott(1990)]{mel90} Melott, A. L. 1990, Phys. Rep., 193, 1

448: \bibitem[Melott, Weinberg \& Gott(1988)]{mel88} Melott, A. L.,

449: Weinberg, D. H. \& Gott, J. R. 1988, \apj, 328, 50

450: \bibitem[Melott et al.(1989)]{mel89} Melott, A. L., Cohen, A. P.,

451: Hamilton, A. J. S., Gott, J. R. \& Weinberg, D. H. 1989, \apj, 345,

452: 618

453: \bibitem[Moore et al.(1992)]{moo92} Moore, B. et al. 1992, \mnras,

454: 256, 477

455: \bibitem[Rhoads, Gott \& Postman(1994)]{rho94} Rhoads, J. E., Gott, J.

456: R. I. \& Postman, M. 1994, \apj, 421, 1 

457: \bibitem[Springel et al.(1998)]{spr98} Springel, V. et al. 1998,

458: \mnras, 298, 1169

459: %\bibitem[Squires \& Kaiser(1996)]{squ96} Squires, G. \& Kaiser, N.

460: %1996, \apj, 473, 65

461: %\bibitem[Schneider et al.(1992)]{sch92} Schneider, P., Ehlers, J.,

462: %Falco,. E. E. 1992, Gravitational Lenses, (Springer)

463: %\bibitem[Tyson, Wittman \& Angel 2000]{tys00} Tyson, J. A., Wittman,

464: %D. M. \& Angel, J. R. P. 2000, astro-ph/0005381

465: \bibitem[van Waerbeke(2000)]{van00} van Waerbeke, L. 2000, \mnras,

466: 313, 524

467: \bibitem[van Waerbeke et al.(2000)]{van00b} van Waerbeke, L. et al. 2000,

468: \aap, 358, 30

469: \bibitem[Weinberg, Gott \& Melott(1987)]{wei87} Weinberg, D.\ H.,

470: Gott, J.\ R. \& Melott, A.\ L.\ 1987, \apj, 321, 2

471: \bibitem[Wittman et al.(2000)]{wit00} Wittman, D.\ M., 

472: Tyson, J.\ A., Kirkman, D., Dell'Antonio, I. \& Bernstein, G.\ 2000, 

473: \nat, 405, 143 

474: 

475: \end{thebibliography}

476: 

477: \clearpage

478: 

479: %\centerline{\bf FIGURE CAPTIONS}

480: 

481: \begin{figure}

482: \epsscale{1.0} \plotone{fig1.ps} \figcaption[fig1.ps]{ Comparison of

483: the reconstructed and input maps of $\kappa$. The left panel shows the

484: input $\kappa$ maps for a field 3 degrees on a side, and the right

485: panel shows the same field reconstructed from simulated ellipticity

486: data that includes the noise due to the intrinsic ellipticities of

487: galaxies.

488: \label{fignoise1}

489: }

490: \end{figure}

491: 

492: \begin{figure}

493: \epsscale{.46} \plotone{fig2a.ps} \plotone{fig2b.ps} \epsscale{.46}

494: \plotone{fig2c.ps} \plotone{fig2d.ps} \figcaption[fig2a.ps]{Genus

495: curves per steradian for simulated convergence fields. The four plots

496: use smoothing angles $\thf = 1', 2', 3', 10'$. In each plot, fields

497: without noise (left panels) and with noise due to the intrinsic

498: ellipticities of galaxies (right panels) are shown. Symbols and thick

499: lines correspond to the genus measured from the simulation data. Thin

500: curves correspond to the prediction of random Gaussian fields with

501: normalization fitted to the simulation data. Two cosmological models

502: are shown: the open model (solid), and the tau model (dashed). 

503: \label{figgenus1}

504: }

505: \end{figure}

506: 

507: \begin{figure}

508: \epsscale{1.0} \plotone{fig3.ps} \figcaption[fig3.ps]{ The genus curve

509: of the pure Gaussian noise. This curve corresponds to a map in which

510: galaxies are randomly distributed with random ellipticities and there

511: is no weak lensing signals. The error bars are given by the $1\sigma$

512: deviations from 10 realizations. The area is the same as that

513: simulated for the open model and the smoothing angle is $\thf=1'$. The

514: solid curve shows the prediction for a random Gaussian field.

515: \label{fignoise2}

516: }

517: \end{figure}

518: 

519: %\begin{figure}

520: %\epsscale{1.0} \plotone{fig2.eps}

521: %\figcaption[fig2.eps]{caption 2 here.

522: %\label{fig2}}

523: %\end{figure}

524: 

525: % \clearpage

526: 

527: % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

528: % \begin{deluxetable}{ccccc}

529: % \tablecolumns{5}

530: % \tablewidth{0pc}

531: % \tablecaption{Meatball-shift statistic for pure lensing field.

532: % \label{tab1}} 

533: % \tablehead{

534: % \colhead{smoothing angle ...} & $1'$ & $2'$ & $3'$ & $10'$}

535: % \startdata

536: % OPEN ... & $0.26 \pm 0.03$ & $0.23 \pm 0.04$ & $0.18 \pm 0.05$ &

537: % $0.1 \pm 0.1$ \\

538: % TAU ...  & $0.20 \pm 0.03$ & $0.15 \pm 0.07$ & $0.15 \pm 0.07$ &

539: % $0.0 \pm 0.2$ \\

540: % \enddata

541: % \end{deluxetable}

542: % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

543: 

544: % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

545: % \begin{deluxetable}{ccccc}

546: % \tablecolumns{5}

547: % \tablewidth{0pc}

548: % \tablecaption{Meatball-shift statistic for noisy lensing field.

549: % \label{tab2}} 

550: % \tablehead{

551: % \colhead{smoothing angle ...} & $1'$ & $2'$ & $3'$ & $10'$}

552: % \startdata

553: % OPEN ... & $-0.08 \pm 0.01$ & $0.01 \pm 0.03$ & $0.06 \pm 0.02$ &

554: % $0.06 \pm 0.09$ \\

555: % TAU ...  & $-0.04 \pm 0.02$ & $0.05 \pm 0.06$ & $0.06 \pm 0.06$ &

556: % $0.0 \pm 0.1$ \\

557: % \enddata

558: % \end{deluxetable}

559: % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

560: 

561: \end{document}

562: