'*'------NEWSANDVIEWS------.:N..::.Ac.:.TIJ;:_;RE=-..;V...:.O~L...;;,.31:..:.4...;..4.c..;AP:..;Rc:;.;ll:..:.1=98S the description of flow through porous Oceanography materials - 'the percolation problem'. The experiment of Nittmann et al. is also significant in that it gives insight into the Sonar classification of sea beds minimum requirements of a DLA-like structure. Although DLA seems quite from M.l. Somers simple, it has defied complete analysis. A major point of contention is whether noise AL THOUGH side-scan solar is used widely ween gravel and stone types. A similar due to the discrete nature of the particles in both scientific and engineering investi­ result could probably be obtained by ap­ making up the cluster dominates the gations of the sea bed, the task of extrac­ plying automatic gain control, removing dynamics. Dielectric breakdown structures ting a numerical classification of sea beds the mean and taking the auto-correlation have links of more-or-less fixed length and from the backscattered sound has proved function, though this would possibly be electrolytic deposits are polycrystalline. intractable. But an experienced human computationally less attractive. Experiments on these systems shed no light observer who has had access to the results It would be interesting to see the method on the question. But the fluid system of of comparisons between sonographs and applied to a wider selection of sea-bed Nittman et al. is continuous. Apparently, physical samples can rapidly make accurate echoes, using a number of frequencies in­ discreteness is not essential to make a DLA­ qualitative classifications, so there must be stead of the single 48-kHz frequency used like fractal. There has been speculation some quality of the backscattering that by the authors. An application which that this would indicate that DLA is an gives the necessary clues. On page 426 of comes to mind concerns the problem of example of chaos. In practical terms it may this issue, Z. Reut, N.G. Pace and M.J.P. mapping and assessing fields of manganese mean that solidification far from Heaton offer a solution to the problem and nodules in oceans that are several thousand equilibrium, which can also be mapped a method of classification into six sea-bed metres deep. This will require low­ onto DLA, could give rise to fractals - for types - mud, sand, clay, gravel, stones and frequency sound, such as used in the long­ example, if a snowflake were allowed to boulders. range side-scan sonar, GLORIA II . grow very large, it would probably be The backscattering of sound from the sea Manganese nodule fields occur very wide­ fractal, too. D bed is a complicated function of the graz­ ly in the Abyssal Plain regions of the deep I. Nittmann, J. , Daccord, G & Stanley, H. E. Nature 314, 141 ing angle of the sonar signal and the ocean and should be visible on a GLORIA (198S). roughness of the sea bed, with further com­ record. If so, it might at least be possible 2. Brady, R.M. & Ball, R.C. Nature 309,225 (1984) . 3. Matsushita. M. , Sano, M., Hayakawa, Y., Honjo, H. & plications thrown in by the acoustic im­ to delineate the boundaries of the fields, Sawada, Y. Phys. Rev. Lett. 53,286 (1984). pedance of the sea-bed material: if the ratio and perhaps even to give a measure of 4. Niemeyer, l., Pietronero, L. & Wiesmann, H. J. Phys. Rev. Lett. S2, 1033 (1984). of sea bed to water impedance is close to nodule abundance. 5. Mandelbrot, B. The Fractal Geometry of Nature (Freeman, unity (as in the case of mud), then there is Side-scan sonar requires oblique in­ San Francisco, 1982). appreciable penetration of sound into the cidence for its best visual effect, so that 6. Saffman, P.G. & Taylor, G.I. Proc. R. Soc. A24S, 312 (1958). sea bed and the possible production of a depth/range ratios need to be less than 7. Paterson, L. Phys. Rev. Lett. 52, 1621 (1984). boundary wave, which can interact with the about 0.2. But most systems working from 8. Witten, T.A. & Sander. L.M. Phys. Rev. Left. 47, 1499 (1981); Phys. Rev. 827, 5685 (1983). incident wave, profoundly modifying the the surface in deep water, or at high fre­ backscattering process (see Somers, M.L. quencies in shallow water, operate at near­ and Stubbs, A .R. IEE Proc. 131, Pt F; vertical incidence and the backscattering 1984). To circumvent these problems, Reut process is somewhat different. It will be in­ l.M. Sander is in the Physics Department, University of Michigan, Ann Arbor, Michigan et al. have concentrated on the rate of fluc­ teresting to see how the classification 48109-IJ20, USA. tuation of the backscattered sound over system of Reut et al. performs in such short periods of time. Their method is bas­ circumstances. ed on the observation that the fluctuations Turning from their concern with prob­ generally become more rapid as the ground lems of propagation and surface reverbera­ gets rougher. There are, of course, large tion, the underwater-acoustics communi­ differences in total energy, and the peak­ ty is devoting increasing attention to sonar to-valley ratios can actually be greater over scattering. Many theoretical papers and a soft or smooth sea bed than over a rough some reporting measurements or equip­ one, but the rate of fluctuation is still ment will be given at the International Con­ 100 Years Ago generally less than if the sea bed is rough. ference on Scattering Phenomena on 2-3 The method requires fairly careful nor­ April at the Admiralty Research Establish­ THE following account, we learn from Science, of unusual phenomena was received, March 10, at the malization, for which Reul et al. have ment, Portland, UK. Various applications Hydrographic Office, Washington, from the chosen to use cepstrum analysis, an outline of the analysis of sea-bed backscattering, branch office in San Francisco. The barque of which is given in their paper. By taking both civil and military, are being explored. lnnerwich, Capt. Waters, has just arrived at the mean of the power spectrum, all con­ Of the civil applications, many involve the Victoria from Yokohama. At midnight of February siderations of phase can be ignored, so en­ estimation of resources. Manganese 24, in latitude 37° north, longitude I 70° 15' east, suring that when the inverse transform is nodules have already been mentioned. the captain was aroused by the mate, and went on taken the result peaks at, or near to, zero Gravel provides a second example. Over ten deck to find the sky changing to a fiery red. All time delay. The effect of taking logs before per cent of the gravel used in the United appeared over the at once a large mass of fire Kingdom is dredged from the sea bed, and vessel, completely blinding the spectators; and, as doing the inverse transform is to de­ it fell into the sea some fifty yards to leeward, it emphasize any d.c. level in the signal, a rapid and reliable means of detecting caused a hissing sound, which was heard above the which shows up as a large peak at zero fre­ suitable gravel fields would be of con­ blast, and made the vessel quiver from stem to quency in the power spectrum. The two­ siderable value. The accuracy of acoustic stern. Hardly had this disappeared when a lowering power cepstrum integral parameters give a logs for ships that measure their speed by mass of white foam was seen rapidly approaching measure of how closely the cepstrum is con­ observing the Doppler shift of sound the vessel. The barque was struck flat aback; but, fined to low time-lag values. backscattered from the sea bed should also before there was time to touch a brace, the sails Justification of the method comes both be improved by application of the new had filled again, and the roaring white sea had from the foregoing qualitative arguments system of classification. D passed ahead. The master declares that the in its favour and from awfulness of the sight was beyond description and that can be mustered considers that the ship had a narrow escape from the fact that it seems to work - 120 sea­ M.L. Somers is at the Institute of Oceanographic destruction. bed areas are correctly classified into the Sciences, Wormley, Goda/ming, Surrey From Nature 31 514, 2 April 1885. six types with the only overlap being bet- GU8 5VB, UK. © 1985 Nature Publishing Group