It is becoming increasingly apparent that episodic magmatic intrusions at active spreading ridges on the ocean floor result in discharges of hot water, termed "megaplumes." Theoretically, such rapid expulsions of relatively buoyant fluid will perturb the height of the sea surface above the spreading ridge, enabling the use of satellite altimetry as tool for monitoring global submarine volcanism. The changes in steric height due to events observed in 1986 and 1993 are calculated, and TOPEX/POSEIDON altimeter data coincident with the 1993 events is examined in search of the predicted signal.
Because most of the EarthUs ocean ridges are submerged in thousands of meters of seawater, direct observation of individual eruptions at seafloor spreading centers is problematic. Although submarine eruptions are held to be the most common on the planet, no one has ever seen one in action anywhere along the Earth's 75,000 km of mid-ocean ridge (Reynolds, 1994). Nevertheless, recent research efforts and fortuitous encounters have come close. Remotely-operated vehicles, submersibles piloted by humans, and traditional oceanographic instruments have provided limited photographic and in situ observations of apparently recent --and sometimes violent-- volcanism at spreading ridges; sonar has generated bathymetric maps which reveal annual and decadal changes in geologic features on the seafloor; and data from the NavyUs acoustic submarine surveillance system have recently supplemented ship-board seismic detection of earthquake activity along some of the oceans' ridges.
However, none of these techniques promises to inform scientists of the nature of global submarine volcanism as easily as does a satellite-based technology. Using data from the TOPEX/POSEIDON satellite radar altimeters, it may be possible to observe submarine volcanic activity from space. Through a sudden reduction in the density of the surrounding seawater (due to an increase in temperature and/or a decrease in salinity) a submarine eruption may cause the dynamic height of the overlying sea surface to change enough to be detected by the highly accurate and precise satellite altimeters.
This working paper first outlines the theoretical relationship between changes in sea surface height and large bodies of anomalously warm water which were dubbed "megaplumes" when first observed in 1986 (Baker et al, 1987), and later associated strongly with submarine volcanic activity. After comparing the observational capabilities of the TOPEX/POSEIDON satellite radar altimeter with what is currently known about megaplumes, a preliminary examination of the altimetry data is made, searching for the signal expected from a series of 3 plumes observed in 1993.
Sea Surface Topography and Density Changes in the Ocean
Searching for a variation in sea surface height over a submarine eruption is complicated by the fact that sea level is not level at all! Indeed, as Koblinsky (1993) explains, the Earth's mean sea surface varies from a reference ellipsoid by ~100m though 99% of the variation is caused by the shape of the Earth's geoid, which is predominantly influenced by density inhomogeneities in the mantle. Consequently, in order to reveal the smaller (11 m) variations in sea surface topography related to ocean dynamics, satellite oceanographers examine the mean and time varying shape of the ocean surface relative to the geoid (Koblinsky, 1993).
Gill and Niiler (1973) explain that depression or uplifting of the sea surface can be specified by a dynamic height (eta) which has three components: variations due to atmospheric pressure, changes in the density of the water column, and alterations in bottom pressure. The variations caused by density changes, termed "steric effects," are primarily caused by shifts in the temperature and/or salinity of the seawater. Deviations in the steric height of the sea surface -- only one component of the overall dynamic height -- can have suprising magnitude relative to the scale of circulation features (11 m); in mid-latitude waters, local heating of the near-surface layers can cause vertical variations in the surface topography of 115 cm (Koblinsky, 1993).
Derivation of a relationship between temperature and steric height is not difficult. By combining the equations of Gill and Niiler (1973) with some elementary thermodynamic principles (as shown in Appendix A), the formula ( ) provides a rough estimate of how the overlying sea surface will respond to a thermally-induced density anomaly. However, before gathering measurements of temperature and salinity anomalies due to volcanic activity in the world's seas and calculating the associated changes in steric height, it is important to consider what sort of perturbation in the sea surface topography is detectable from space.
The TOPEX/POSEIDON Satellite Altimeters
The TOPEX/POSEIDON satellite altimeter mission was not designed to monitor the changes in steric height which may occur due to volcanism at the EarthUs active spreading centers. However, a comparison of the altimetric measurements and the nature of submarine eruptions may reveal that the data holds serendipitous utility in monitoring global submarine volcanism.
Foremost, the altimeters have extraordinary accuracy and precision. Through careful tracking of the satelliteUs position (using the Global Positioning System, DORIS beacons, and laser retroreflectors), and after considering atmospheric attenuation, the dynamic height of the sea surface can be obtained with a total error of 111.5 cm relative to the EarthUs geoid (see Fig. 1). Conveniently, the geoid error can be omitted for time dependent measurements --like the cyclic assessments of steric height changes over ocean ridges; in such cases, the total time dependent error is only 13.5 cm (Tsaoussi, 1994). This establishes the first constraint on the detectability of a sea surface deflection due to a submarine volcanic event: the steric height change between cycles must be greater than 3.5 cm.
Further constraints derive from BenadaUs (1993) description of the satelliteUs orbit. Carefully chosen to provide measurements which increase understanding of global ocean dynamics, the orbit provides near-global coverage (shown in Fig. 2) and has a periodicity of 9.9 days. Consequently, an observable volcanic event must generate a deflection of the sea surface which is located between 166.04! latitude and lasts on the order of weeks. An additional spatial constraint is the most unlikely to be met; because the distance along the equator between two adjacent pass intersections (called the Requatorial cross-track distanceS) is 315 km, and the altimeters' "footprint" is approximately 2 km in diameter, the surface effect of an observable eruption will have to be either hundreds of km in diameter, or located --at least in part-- directly under the path of the satellite.
Megaplumes: Source of a Signal?
In the absence of direct observations of submarine eruptions at spreading centers, an estimate of their size and duration --and therefore their observability-- can be gleaned by looking at similar phenomenon on land. In Iceland, a country straddling the mid- Atlantic ridge, a rifting event which began in 1975 demonstrates that "plates do not separate at a steady few cm per year but in intense surges of several meters over only a few years followed by perhaps a century of quiet" (McClelland et al, 1985). The rifting reportedly occurs "periodically in short active pulses at a few months intervals," accompanied by earthquake swarms, vertical ground movements of up to 2 m, and sometimes also volcanic eruptions and formation of new fumaroles (Bjornsson et al, 1979). While it may be reasonable to expect similar events to occur along the thousands of km of submerged spreading ridges, their spatial and temporal distribution has proven elusive.
However, since the discovery of hydrothermal fluids in the Red Sea in 1964, areas of hydrothermal activity have been found along numerous segments of the mid-ocean ridges (Rona, 1992). While it may be possible to detect the surface effect of relatively long term changes in thermal output over such areas, a more likely source of changes in steric height which would be noticible between 9.9-day cycles is a sporadic, more thermally-intense pulse of volcanic activity. As discussed in the previous section, the areal extent of the surface effect from such an event will have to be either large (10- 100 km in diameter) and/or contiguous with a satellite surface track; the duration must be on the order of 10 days; and a variation in steric height greater than 3.5 cm will be a prerequisite for detection.
In 1986, scientists from the National Oceanographic and Atmospheric AdministrationUs Pacific Marine Environmental Laboratory encountered an unexpected temperature anomaly above the Juan de Fuca Ridge (JDFR) (see Fig. 3). The "megaplume" was a roughly symmetrical, oblate spheroid of anomalous fluid, 20 km in diameter, 700 m thick, and centered 700 m above the seafloor (Baker et al, 1987). It had a mean temperature anomaly of 0.12 oC, representing a net heat exchange from seafloor to the benthic waters of almost 1017 J (Baker et al, 1989). Characterized as a brief, high-volume burst, the hydrothermal emission was associated with a sea-floor extension event (Embley et al, 1991). While the megaplume was formed in a few days, its duration is uncertain; it was not encountered at the same location 2 months after its formation (Baker et al, 1987).
Using the equations derived in Appendix A, the deviation in sea surface topography due to megaplume 1986 is expected to be about 5 cm (see Appendix B), just above the calculated limit of detectablity for time dependent altimeter observations. While the exact shape of the surface deformation above the plume has not been carefully modeled, if we expect the feature to mimic the extent of the plume, roughly 20 km in diameter, then at 45 oN latitude (where the cross-track separation is about 300 km), there is about a 1 in 15 chance of the altimeter recording the change in sea surface height. Assuming that the surface deflection due to the megaplume persists for a month or two, the probability of detection is slightly increased, for the plume may move horizontally, approaching or even crossing one of the adjacent satellite tracks. However, these estimations are moot because the megaplume occurred long before the satellite began recovering calibrated data in February, 1993 (Benada, 1993). A more recent event is necessary to enable a search for the signal in the sea.
In 1993, seismic activity was detected above the CoAxial segment of the JDFR using the U.S. Navy's acoustic submarine surveillance system. Persuantly, a series of in situ measurements was collected from oceanographic vessels, revealing 3 event plumes over the ridge. Each was cooler and smaller than the 1986 event, having maximum temperature anomalies no greater than 0.2!C, and diameters no greater than about 10 km (Baker et al, 1994). For the 1993 plumes, in which the heat flow was 1 to 2 orders of magnitude less than in the 1986 megaplume, the predicted change in steric height is only (XXX) (see Appendix B); additionally, it is likely to be evident over a significantly smaller area. However, because they are the only events known to have occurred when the TOPEX/POSEIDON altimeters were actively acquiring data, they provide a first opportunity to search for the theoretical signal.
Results from an Initial Search
Unfortunately, none of the 1993 event plumes formed at a latitude and longitude directly observed by the TOPEX/POSEIDON altimeters (see Fig. 3). Additionally, the theoretical change in steric height ( XXX) is below the time dependent error level. While it is therefore doubtful that the signal should be apparent in the altimetric data, the search is nevertheless motivated by a combination of pure curiosity and substantial uncertainty in the nature of the expected surface effect (its areal extent and propagation, and even its vertical variation).
Segments of the satellite tracks (#48 and #103) which passed closest to the site of seismic activity and megaplume formation (see Fig. 3) were extracted for a series of 6 cycles, providing observations of the same range of latitudes and longitudes at 9.9-day intervals (one cycle was unavailable and therefore skipped). The first cycles provide observations of the sea surface topography prior to both the initial seismic activity along the ridge and the formation of the event plumes (see Table 1). The subsequent series of cycles covers a period during which the plumes formed, rose above the seafloor, attained neutral buoyancy, and perhaps began to move horizontally.
After an initial examination of a surface plot of the regional topography, an attempt to remove some of the small scale noise was made. Errors up to 10 to 30 cm (Benada, 1993) can result from small cycle-to-cycle changes in the look angle of the altimeter antenna, especially in a region where the geoid is steep. Consequently, the along-track difference between the sea surface height (SSH) and a mean sea surface height (MSS) (derived from previous, less accurate altimeter missions) is examined; subtracting the MSS from the SSH removes the geoid, yielding the "residual surface" or "dynamic topography" (refer to Fig. 1) (Benada, 1993).
Figure 4 shows a series of residual heights along a segment of track #103. The sea surface topography is remarkably consistent between the range of cycles (covering a period of almost 50 days), displaying few and subtle cycle-to-cycle variations. One feature which resembles the expected volcanic signal is the "bulge" which appears in cycle 29, centered on sample #2381. It persists throughout the series of cycles and is revealed as a relative peak in the plots of "Delta Residual Height" shown in Fig. 7. Each plot of the change in residual height consists of about 40 measurements (samples) from each cycle differenced with spatially coincident residual heights observed in cycle 28 (taken as a reference period during which no seismic activity or megaplume formation took place).
An examination of cycle 27, however, places the relevance of the "bulge" in question, for Fig. 5 shows that cycle 28 itself is apparently more variant than most of the others; the bulge is clearly present in cycle 27. The relative dips and peaks in cycle 28 await explanation, but based on the temporal constraints provided by the PMEL observations, the variations cannot be correlated with the 1993 seismic and megaplume activity.
In this preliminary processing of the data, the hypothesis that the noise level is greater than signal expected from the 1993 plumes is confirmed. The 1993 events may have been too weak to form a detectable surface perturbation, or their effect may not have been observable because the satellite did not pass directly above the region of activity.
Conclusions and Further Study
One way in which to improve the feasibility of monitoring submarine volcanism from space is to learn more about spreading center eruptions, their correlation with megaplumes, and the characteristics of megaplumes themselves. Traditional oceanographic research may eventually reveal that the "typical" submarine eruptions and resultant megaplumes are much larger than the few events recorded to date in the scientific literature. After discussing the hypothetical heat sources which formed the 1986 megaplume, Baker et al (1987) concluded that "if a basaltic intrusion of only 0.01 km3 is sufficient to create a megaplume [the size of the 1986 event], hundreds to thousands of such plumes may be blossoming each year along the global ridge system."
Another method is to continue searching the TOPEX/POSEIDON database --and perhaps the archives of past satellite altimeter missions as well-- for the signal from a hitherto unobserved and especially large event. Based upon current oceanographic and geophysical understanding, the East Pacific Rise (Haymon et al, 1993, and MacDonald et al, 1989) and JDFR are the most promising regions in which to search, for they are established respectively as the fastest spreading and most thoroughly studied ocean ridges on the planet. Additionally, a search for longer term (month to year), large scale (hundreds of km) variations in steric height over volcanically active ridge segments may prove worthwhile, even within the geologically instantaneous lifetime of the TOPEX/POSEIDON satellite altimeter mission. The preliminary searches are easy to conduct and numerous, more complex statistical and computational strategies await exploration.
Acknowledgments
I would like to thank a number of individuals at the NASA Goddard Space Flight Center, the NOAA Pacific Marine Environmental Laboratory, and the Universities Space Research Association (USRA) for facilitating what was for me a uniquely formative summer. For the first time, I was allowed to maintain my interdisciplinary interests while creatively formulating and delving deeply into a specific scientific question. In my opinion, such a process is fundamental to the advancement of Earth System Science.
Chet Koblinsky welcomed me kindly into the Ocean and Ice Branch, and provided me with both the freedom and resources I needed to pursue research interests only partially integrated with his own. Liping Wang inspired me to leap fearlessly into the turbid sea of fluid dynamics and partial differential equations, where I will continue to swim for some time. Christine Gailliard-Boone saved me weeks of precious time which I would have otherwise spent running amok in the labyrinth of the Interactive Data Language, while Brian Beckley unlocked and held open the door to successful retrieval and comprehension of TOPEX/POSEIDON data. Lucia Tsaoussi graciously shared her knowledge of statistics, her geodesic expertise, and her computer. Karen Settle oriented me to both the virtual and real aspects of my "workstation." Thanks also to Christophe Menkes, Lina Lo, and Xia Chang for juggling entertainment, presentation support, and Chinese education, respectively.
Edward Baker and Bill Lavelle of PMEL provided me with not only cutting edge data from their observations of the 1993 event plumes, but also their professional opinions, suggestions, and calculations. I look forward to the possibility of learning more from them and their peers in the near future.
Final accolades go to Paula Webber and her cohorts at USRA, for simultaneously maintaining an atmosphere of logistical competence and personal devotion to the program.
References
Baker, E.T., Massoth, G.J., and Feely, R.A. Cataclysmic hydrothermal venting on the Juan de Fuca Ridge. Nature, Vol. 329, 149-151, 10 Sept. 1987.
Baker, E.T., Lavelle, J.W., Feely, R.A., Massoth, G.J., and Walker, S.L. Episodic Venting of Hydrothermal Fluids From the Juan de Fuca Ridge. Journal of Geophysical Research, Vol. 94, No. B7, 9237- 9250, July 10, 1989.
Baker, E.T., Massoth, G.J., Feeley, R.A., Embley, R.W., Thompson, R.E., and Burd, B.J. Hydrothermal Event Plumes from the CoAxial Seafloor Eruption Site, Juan de Fuca Ridge. June, 1994, for publication in Geophysical Research Letters.
Benada, R. Merged Geophysical Data Record (GDR) (TOPEX/POSEIDON) UserUs Handbook, Version 1.0. Physical Oceanography Distributed Active Archive Center (PO.DAAC). Sept. 13, 1993.
Bjornsson, A., Johnsen, G., Sigurdsson, S., and Thorbergsson, G. Rifting of the Plate Boundary in North Iceland 1975-1978. Journal of Geophysical Research, Vol. 84, No. B6, June 10, 1979.
Embley, R.W., Chadwick, B., Perfit, M., and Baker, E.T. Geology of the northern Cleft segment, Juan de Fuca Ridge: Recent lava flows, sea-floor spreading, and the formation of megaplumes. Geology, v. 19, 771-775, August, 1991.
Gill, A.E., and Niiler, P.P. The theory of the seasonal variability in the ocean. Deep-Sea Research, Vol. 20, 144-177, 1973.
Haymon, R.M., Fornari, D.J., Von Damm, K.L., Lilley, M.D., Perfit, M.R., Edmond, J.M., Shanks, W.C., III, Lutz, R.A., Grebmeier, J.M., Carbotte, S., Wright, D., McLaughlin, E., Smith, M., Beedle, N., and Olsen, E. Volcanic eruption of the mid-ocean ridge along the East Pacific Rise crest at 9o45-52'N: Direct submersible observations of seafloor phenomena associated with an eruption event in April, 1991. Earth and Planetary Science Letters 119, 85-101, 1993.
Koblinsky, C. "Ocean Surface Topography and Circulation" reprinted from Atlas of Satellite Observations Related to Global Change (Gurney, R.J, Foster, J.L., and Parkinson, C.L., eds.), Cambridge University Press, 1993.
MacDonald, K.C., Haymon, R., and Shor, A. A 220km2 recently erupted lava field on the East Pacific Rise near lat 8oS. Geology, v.17, 212-216, March 1989.
McClelland, L., Simkin, T., Summers, M., Nielson, E., and Stein, T.C., eds. Global Volcanism: 1975-1985. Smithsonian Institution, Scientific Event Alert Network (SEAN), p.8, 1985.
Reynolds, J. "Aftermath of a sea-floor eruption." Nature, Vol. 367, p.115, 13 January, 1994.
Rona, P.A. "The Dynamic Abyss" in Saving the Oceans (MacInnis, J., ed.), 1992.
Tsaoussi, L.S., and Koblinsky, C.J. An Error Covariance Model for Sea Surface Topography and Velocity Derived From TOPEX/POSEIDON Altimetry. Submitted to Journal of Geophysical Research, June 29, 1994.