INTRODUCTION OF WALTER MUNK
Dr. L. Vere Shannon, President, IAPSO
The IAPSO President's Lecturer really needs no introduction, because I doubt whether there is a delegate here today who is not familiar with one or more aspects of Walter's fundamental contributions to physics. His more than 40 honours, prestigious awards and medals are ample testimony to his standing among his peers - as is the Science Citation Index.
Walter Munk can be considered as THE LEADER in the physics of the ocean. In 1982, on the occasion of the celebration of Walter's 65th birthday, Henry Stommel said, and I quote, "Walter has been number one for all these years". He was referring to Walter's pioneering work in so many fields - the measurement and spectral description of ocean waves - tsunamis - tidal phenomena - internal wave spectra - sound propogation in a real ocean with stratification, tides, internal waves and mesoscale eddies. (In respect to this he had explored the prospects for observing the ocean in the 1990s and had commenced his work on ocean acoustic tomography).
Fifteen years later where are we? Walter has extended the acoustic tomography into global scale underwater acoustics. He has linked this with satellite altimetry to provide the necessary observational underpinnings for global ocean models.
It therefore is a great honour and privilege for me to call on the man who is still "number one", to deliver a lecture entitled "Sampling the Oceans from Above and Beneath" - Professor Walter Munk.
SAMPLING THE OCEANS FROM ABOVE AND BENEATH
Walter Munk, Scripps Institution of Oceanography
June 1997, Melbourne
ABSTRACT. Probing the ocean from a few isolated research vessels has always been a marginal undertaking, and the first hundred years of oceanography could well be called "a century of undersampling". It came as a shock in the 1960's that the ocean, like the atmosphere, has an active weather at all depths with a scale of order 100 km (the mesoscale) which had slipped through the coarse grid of traditional sampling.
The most profound effect of satellite oceanography has not been the result the new sensor packages (and these have been remarkable), nor the global coverage, but rather that for the first time ocean processes were adequately sampled. TOPEX/POSEIDON altimetry is the outstanding success story. Its contribution towards understanding ocean processes goes well beyond anything that had been imagined. I will discuss the application of T/P altimetry to ocean tides. Although most of the effort here has been spent on removing tides so that other processes can be studied, there is new evidence that tides contribute significantly to the mixing of the abyssal oceans.
The deep ocean is opaque to electromagnetic radiation, and there are strong limits to the inferences that can be drawn with remote sensing from above. But the ocean is transparent to acoustic radiation, and low frequency signals can be monitored at global distances. I will discuss recent results of a study of ocean climate variability by a combined application of satellite altimetry and ocean acoustic tomography. The combined dataset is being used for testing climate models.
1. INTRODUCTION. Among scientists I have known, I have three heroes: Harald Sverdrup, Roger Revelle, and Sir Geoffrey Taylor. Harald was President of IAPSO 1946-48, and Roger 1963-67. In these older and better days, Association Presidents gave their own President's Lecture; it means a lot to me to be following in Harald's and Roger's footsteps.
2. A CENTURY OF UNDERSAMPLING. When I returned to the Scripps Institution at the end of W.W.II as a student of Harald Sverdrup, physical oceanography was still practiced in the tradition of the Challenger expedition, with a few isolated research vessels probing the world oceans. As a result, we were studying ocean climate rather than ocean weather.
Let me explain. Under climate I include the circulation in the subtropical and subpolar gyres, the western boundary currents such as the Gulf Stream, and the formation of deep and bottom water which refills the ocean basins every 3000 years. Representative climate scales are 5 to 10 megameters (1 Mm=1000 km), and 1 year or longer. Under weather I include mesoscale (or synoptic) eddies (the storms of the sea) and associated Rossby waves, with typical scales of 100 km and 100 days. At the time the mesoscale circulation had not yet been recognized even though it contains over 95% of the ocean's kinetic energy. A typical current record is better represented by 1 ± 5 cm/s than by 5 ± 1 cm/s ! The climatic mean is swamped by weather. But with the 5 ± 1 cm/s in mind, it was taken for granted if one just took enough Nansen casts, one would end up with a good representation of the general circulation.
So differences between stations at positions x1 and x2 taken at times t1 and t2 were ascribed to Dx. This means that the energetic ocean weather was aliased into an ocean general circulation of ever increasing complexity. The misinterpretation was aided by the golden rule of oceanography never to take the same station twice. (In the few cases of repeated stations, differences could always be ascribed to instrumental malfunctioning.) Still, it is almost unbelievable that the mesoscale circulation could have slipped for so long through the coarse grid of ocean sampling.
Sampling is a prime issue of this talk. By adequate sampling I understand strict obedience of the Nyquist limits: sampling at least twice in time for the shortest significant record period, and twice in space for the shortest significant length. From this point of view, the era starting with the Challenger expedition in 1870 can be viewed as a near-century of undersampling.
3. THE MESOSCALE REVOLUTION. The era came to an abrupt end with the discovery in the 50's of Gulf Stream meanders, and John Swallow's measurements with deep floats. I believe it was Fuglister who asked the incisive questions as to why the Gulf Stream always seems to be following oceanographic research vessels. Swallow acoustically tracked buoyant floats to test Stommel's theory of a southward-flowing counter current under the Gulf Stream. The theory was confirmed, but for any single drift the velocities greatly exceeded the mean flow, and the directions differed from one time to the next (having violated the golden rule). These early developments led directly to the MODE expeditions which defined the dimensions of the mesoscale circulation.
4. SATELLITES. The first ocean satellite, SEASAT, was launched in summer 1978. I believe that the principal contribution was not so much the development of new ocean-sensing devices (even allowing for their remarkable performance), nor the opportunity for a uniform global strategy, but the ability to sample the ocean properly. Times have never been the same.
I will confine my comments to the TOPEX/POSEIDON (T/P) altimetry measurements. This remarkable satellite has measured the topography of the sea surface to 3 cm accuracy at 7 km spacing for 3 years. Fig. 1 shows the average year-to-year change of sea level. The global mean by +2mm/year is largely the result of thermal expansion. But there are large departures, including some regions of sinking sea level. The appropriate scale of ocean climate is from 5 to 10 megameters; smaller scales are dominated by mesoscale noise, and larger scales smear climate variability. Global warming is not a good measure of climate change.
A "hot spot" centered on Hawaii is rising by 5 cm/year. This must not be interpreted as forerunner of greatly enhanced global warming. In figure 2 I have plotted Honolulu tide level for the 20th century. Superimposed on a mean rise by 2 mm/y are 10 cm decadal fluctuations, with T/P altimetry evidently showing the latest upward wiggle. A spatial (EOF) analysis by ( ) of global sea level shows two dominant spatial patterns associated with ENSO and "decadal" variability (figure 3 and figure 4), but it is not clear whether the Hawaiian hot spot is part of the wider decadal variation, or whether it is associated with the Hawaiian Island Ridge dynamics. This brings me to the next problem: tides.
5. TIDES. By and large the ocean community has regarded tides as a subject that went to bed with Victorian mathematicians. The advent of T/P altimetry did raise the issue of a more accurate knowledge of global tides, but largely in the context of their elimination so that other, more important, issues can be discussed. But not surprisingly, this new work has led to some new insight about the tides themselves.
After extracting the M2 tidal band from the 3-year T/P record (a rather heroic effort by Cartwright and others to avoid the aliases associated with a 10-day sampling of a 0.52 day oscillation), Ray and Mitchum plotted the filtered sea level along consecutive ascending T/P tracks (Fig. 5). The synthetic snapshot shows oscillations of order 5 cm amplitude with spatial scales of 100 km superimposed on the much longer surface tides. Subsequent spatial spectral analysis of the profiles here shown revealed prominent wave lengths of 150 and 85 km, and these were identified with the two gravest internal modes for M2 tides. The conclusion is that some of the surface tide energy is scattered into internal tides along the Hawaiian Island Ridge. The detection of internal tides from satellite altimetry came as a shock to the ocean community. Internal tides have small surface expressions, Dr/r times the internal amplitudes, and for most purposes one can get away with a "solid lid" surface boundary condition (hardly the way towards interpreting altimetry data). But taking internal amplitudes of order 50m and multiplying by Dr/r = 10-3 gives 5 cm, as measured.
Amplitudes diminish to both sides of the Hawaiian Ridge with a decay scale of 1 Mm, in agreement with previous internal tide detections 2 Mm to the northeast by ocean acoustic tomography (I will come to this in a moment). Presumably the energy contained in the internal tidal line spectrum is converted into the internal wave continuum. The measured internal wave intensity is known to be close to saturation, and presumably feeds into turbulence. The far-field internal tide radiation from the Hawaiian Ridge is roughly 15 gigawatts (1 GW=109W). The global total along all ocean ridges is estimated at fifteen Hawaiian ridges, or 200 GW.
For the last few decades oceanographers have estimated a diapycnal diffusivity kV = order 10-5 m2/s associated with pelagic (away from topography) turbulence. Independent estimates have come from measurements of microstructure by Cox and his collaborators, and from experiments by Ledwell and collaborators involving dye release. The global turbulent dissipation associated with kV = 10-5 m2/s comes again to 200 GW. One is tempted to conclude that pelagic turbulence is maintained by the dissipation of tides via the generation of internal waves. This radical suggestion is somewhat supported by the fact that both the diffusivity and the intensity of internal waves remain within a factor of two over a remarkably wide range of conditions, and one looks for a more reliable source than wind generation.
Figure 6 is an attempt to establish a budget of tidal dissipation. The total oceanic dissipation of 3.7 terawatts (3700 GW) has been determined by a number of independent measurements. The most accurate determination comes from the rate of 3.82 +- 0.07 cm/y at which the Moon moves away from the Earth by (measured by laser ranging on the retroreflector placed on the Moon in 1969 during the Apollo mission). There is no difficulty of accommodating 0.2 TW for the maintenance of pelagic turbulence. But about ten times this power is needed for downward diffusion (presumably along topographic features) to maintain the abyssal stratification against the formation of 25 Sverdrup's of deep and bottom water, for otherwise the ocean would fill up with cold water in 3000 years. The 3.7 TW is adequate to fill this need, but nearly all of it has been spoken for since 1919 when G.I.Taylor ( one my three heroes) estimated 41 GW of M2 dissipation in the bottom boundary layer of the Irish Sea, and Harold Jeffreys extended this to 2.2 TW (3 TW for Moon and Sun) of global dissipation.
All this is a very shaky zero-sum game. My conclusion is that oceanographers need to pay more attention to the Moon. Our Russian colleagues will be familiar with a fictitious character in their literature, Kozma Prutkov, a private in the Czar's army who is not very bright but usually gets it right. When asked which is more important, the Sun or the Moon, he replies: "the Moon, of course, because the Sun shines only in daytime when it is light anyhow".
6. OCEAN ACOUSTIC TOMOGRAPHY. The oceans are opaque to electromagnetic waves but transparent to sound. I have already mentioned the Swallow floats which were acoustically tracked. Following the mesoscale revolution with its greatly enhanced sampling requirements, Carl Wunsch and I were looking for a more explicit use of acoustics than for navigation and communication.
The speed of sound increases with temperature and with pressure. In temperate latitudes the result is a minimum in sound speed at roughly 1 km depth, the sound (SOFAR) channel; sound speed increases upwards because the temperature increases, and downwards because the pressure increases, thus forming a wave guide (figure 7). In ray language, steep rays spend least time in the low velocity channel and come in first, even though they have further to go. Axial rays come in last. In the example shown, early rays are delayed by a fraction of a second relative to a prediction based on the climatological mean profile; axial rays show no delays. The simplest interpretation is that the shallow ocean (above the sound channel) was cooler than the climatological average.
Inverse theory provides a systematic method for converting the measured delays t(t) to the soundspeed profile C(z), from which the temperature profile q(z) can be inferred. Perhaps more important, it provides error bars to all the estimates.
Figure 7 is associated with transmissions at 300 km range. The 1991 Heard Island experiment (figure 8) was conducted to explore whether acoustic tomography could be extended to the much larger ranges associated with climate variability. Research vessels from to 11 nations had volunteered to lower hydrophones into the sound channel to listen to powerful sources suspended near the uninhabited Australian island. (The site was chosen because it has great-circle access to all major ocean basins.) The experiment demonstrated the feasibility of global transmissions, but the quality of the receptions beyond 5 Mm appeared to be inadequate for the tomographic requirements. (The coded transmissions are far below background noise levels and their detection depends on phase-coherent signal processing.) Accordingly, plans were made for monitoring the north-east Pacific (fig. 8) at reduced ranges and reduced acoustic intensities. But the ATOC (for Acoustic Thermometry of Ocean Climate) coded transmissions were to be unexpectedly delayed for over three years.
The delay was due to concern about damage to marine life arising from our acoustic transmissions. Whales depend strongly on acoustic communications, and would suffer if these were interfered with. Accordingly, the Heard Island coded transmission were monitored by a team of biologists (including three Australians). At the initiative of some environmental groups, ATOC was placed under very severe restrictions that took three years to negotiate. (ATOC sources transmit 250 acoustic watts (about the level of a tanker), but are silent 98% of the time; intensities are reduced by a factor 106 between the source depth at the sound channel and the biologically important surface layers.)
In April 1994, Mikholevsky carried out a very successful 3.2 Mm transmission across the North Pole (Fig.9) which indicated that the .... water had warmed by 0.30C relative to some long-time climatological means. This was subsequently confirmed by direct measurements. And in 199?, Schott and Send's THETIS experiment in the Mediterranean for the first time combined satellite altimetry and acoustic tomography to test existing climate models.
Finally, in late 1995, the ATOC transmissions from a source on Pioneer Seamount off California to a series of horizontal arrays installed by the U.S.Navy as part of their submarine detection program (sketched in Fig. 8) got underway. Typical ranges are 5 megameters. The measurements show the seasonal variation of heat content, plotted in fig. 10 as mean temperature in the upper 1 km. A comparison with the variation along these same paths inferred (under certain assumptions) from T/P altimetry shows a pleasing agreement, but also important differences. For example, short-period barotropic oscillations are recorded by T/P but do not effect the acoustic travel times.
Evidently the simple principle underlying acoustic tomography, namely that sound travels faster in a warmer ocean, can be used to measure the mean temperature of the intervening waters at climate scales. The precision is high; changes by a few millidegrees are detectable. But the important consideration is that the spatial scale of the measurements is appropriate to the scale of climatic variability. With regard to the marine mammal problem, an independent team of biological observers have not detected any reactions to 15 months of ATOC coded transmissions.
7. THE PROBLEMS OF SEA-TRUTH. The IAPSO mission is "to promote the study of scientific problems relating to the Oceans". Previous Presidential Lecturers have taken this mission to heart, while I have had fun speculating about the results of experiments and expeditions.
We oceanographers have been slow to adapt to new methods. When Harald Sverdrup and I worked on wave predictions during W.W.II, we fitted waves to discrete sinusoids. Oceanographers were not aware of a more suitable representation by continuous spectra that had been common in optics and acoustics for a century.
In 1970 John Apel came to the Scripps Institution to look for advise and support in planning SEASAT. He got neither. When mentioning that satellite altimeters would measure dynamic heights, a well-known oceanographer replied: "if you gave it to me I wouldn't know what to do with it".
Closer to home, the physical oceanography community has been very reluctant to accept acoustic tomography as a tool. (It goes without saying that the environmental community has not shown any enthusiasm.)
One of the problems is that of "sea truth", the requirement to check out any new method against traditional methods. Consider for a moment that satellites had been here for a long time, but that Nansen bottles have just been invented. It would be difficult to justify dynamic heights against the standard of satellite altimetry. The more innovative the method, the more difficult it is to "certify" it by comparison with traditional methods.
8. CLOSING COMMENTS, MOSTLY FROM REVELLE. Roger Revelle gave the IAPSO President's Lecture in Berne in September 1967, ten years before SEASAT. His theme was that great periods of oceanography are determined by the development of new instrumental capacities. He mentioned:
That was thirty years ago. What would he say today? I have great difficulty coming up with an answer. One that comes to mind is that the Challenger moved at ten knots, and that modern research vessels still move at order ten knots. We can expect that ship-based sampling will be augmented by assigning some of the menial jobs to AUV's (Autonomous Underwater Vehicles). A few years later AUV's could go on autonomous missions. To continue with Roger's words:
"If my (Roger's) thesis is correct that advances in oceanography are dependent basically upon new kinds of instrumentation, the future of our science will be sterile unless we can continually find new things to observe and new ways to measure them.""What do these developments mean for oceanographers? It seems to me (Roger) we are somewhat in the position of the generals referred to by the great German military philosopher von Clausewitz when he said, "War is too important to be left to the generals". Our fellow men are now saying the ocean is too important to be left to the oceanographers. We must get over the notion that we somehow have a prior claim on the oceans, that only those who too hold a union card as certified oceanographers can work on the scientific problems of the sea."
These are Roger Revelle's words 30 years ago in his IAPSO President's Lecture. Thank you for inviting me to give this talk today.
Return to IAPSO Home page