PLANETARY MUSINGS

Entries for June 2005

A thin shell dynamo at Mercury?

Stanley, S, J Bloxham, WE Hutchison, and MT Zuber, Thin shell dynamo models consistent with Mercury's weak observed magnetic field, EPSL, 234, 27-38, (2005), doi:10.1016/j.epsl.2005.02.040.

Ever since Mercury's magnetic field was discovered by Mariner 10 back in 1974-1975 its origin has been somewhat of enigma. The basic observation was that the rate of increase in strength of the magnetic field as the spacecraft got closer to the planet was consistent with an internal field with a dipolar character. The strength of the field, however, is pretty weak compared to the Earth's. One possibility is that the magnetic field is generated by a dynamo, just one that doesn't behave exactly like the ours. Stanley et al [2005] focus on the idea that the character of the dynamo and magnetic field may be a function of the relative thickness of the liquid outer core. The Earth's inner core is only about 35% of the radius of the whole core, but that isn't necessarily true for other planets.

Stanley et al., argue that based on magnetostrophic balance assumtpions that the ratio of the strength of the dipolar component of Mercury's magnetic field to the inferred toroidal component is a factor of 10-1000 less than the Earth's if a dynamo is involved. They set out to assess whether a dynamo in a thin shell (inner core is 70% or more of radius of core) could explain this phenomena. The basic result was that yes it could if the modified Rayleigh number (a modified scaling of the ratio of buoyancy to Coriolis forces in this case) was modest.

One of the conclusions that they had that isn't clear to me is how "differential rotation" (apparently the mechanism for creating the toroidal field) can drive a dynamo. Where is the differential rotation, between the inner core and the mantle? They also suggested that it may be possible for the character of the magnetic field to vary with position inside and outside the tangent cylinder (imagine a cylinder placed around the inner core along the axis of rotation). If true that might be observable by the MESSENGER spacecraft when it gets to Mercury in 2011. Though I wonder what exactly we might look for in those data...

The shape of a plume

Farnetani, CG and H Samuel, Beyond the thermal plume paradigm, GRL, 32, L07311, doi:10.1029/2005GL022360, (2005).

The issue of mantle plumes (i.e., whether they exist) is a long-running problem that just keeps going back and forth on the teeter-totter. I've always thought that with as complex a planet as we have here that they yes or no approach was a bit limiting given that the fluid physics of thermal plumes is well-established... but I digress.

Farnetani and Samuel tackle the formation of plumes from a more general framework than is usually taken. The canonical view of plumes is that a mushroom-like head and thin conduit tail plume structure is created at a deep thermal boundary layer and traverses the mantle to the surface with modest entrainment of surrounding mantle. The head impacts the surface, creates a large igneous province and the conduit creates a following chain of volcanic centers as a plate moves over it. That's the broad framework, first-order and all. The Earth is more complex and this paper shows what happens if chemical buoyancy effects, mantle wind (caused by imposed plate motion at the surface), phase transitions, and heterogeneities are considered (if memory serves, many of these effects have been considered before, though maybe not all at once with the spatial resolution of this study). And in this more "relaxed" study of the relevant parameters, the authors have discovered that the plume head-tail structure is but one possible structure of a plume as it reaches the upper mantle. A plume head isn't even necessary, and a concentrically zoned plume tail isn't even necessary either. Basically, models of the Earth can be messy - consistent with the geochemical and seismological view.

Mantle plumes are just plain hard to avoid - there is heat coming out of the core, the boundary layer down at the core-mantle boundary is going to get unstable in some places every once in a while, and there are going to be plumes. The real questions are: what do they look like, how do they sample the mantle, and what happens when (if?) they reach the surface? These questions are still open. But this paper shows us that not every plume looks alike.

But... this view of plumes that a plume head is possible not necessary is interesting when extended to other planets. Now the "mantle wind" may not be as extensive on a planet without plate tectonics, but what is the driver behind the headless-plume? On Venus there are all the coronae and volcanic rises, many of which seem consistent with a thermal plume or diapir source (e.g., recent conceptual framework of Johnson and Richards, 2003). If even some of those, e.g. coroane, require head-like features, what does that mean for the mantle? Maybe it is just a question of scale... but the questions about mantles and their plumes seem to abound.

Hydrothermal activity on Europa

Lowell, RP and M DuBose, Hydrothermal systems on Europa, GRL, 32, L05202, doi:10.1029/2005GL022375, (2005), Article.

Europa is so fascinating, icy tectonics, subsurface oceans, tidal interaction, and who knows, maybe even some microbes... One of the open questions with regard to Europa is how to form "chaos" type regions and features that look like cryovolcanism. The two basic ideas out there are (1) warmer water from the ocean induces a melt-through event in the overlying (thin) ice, (2) tidal dissipation is focused in the (thick) ice shell which is convecting on its own, creating surface features. This paper focuses on the first of these two ideas and tries to understand the heat output from hydrothermal activity in the rocky layer beneath the ice and water ocean. They take estimate for Europa's total heat output from the literature on tidal dissipation and radiogenic heat output and calculate the portion of the heat flow carried by hydrothermal activity by analogy with Earth oceanic crust (seems reasonable to first-order). The punch line is that the total hydrothermal heat flux is similar to earth's, but the output of any given vent is about 10 times less than black-smokers on the bottom of the Earth's oceans. That leads them to conclude that there just isn't enough energy in the thermal anomalies created in the overlying ocean to rise up and melt through the overlying ice... in accord with other papers on the subject.

What strikes me though is that Europa's ocean must be salty, so it is conceivable that it is somewhat stratified like the Earth's oceans, and this may reduce even further the mass/buoyancy flux through the oceans. Also, the input parameters in this study are pretty liberal with the heat available, and yet their models don't predict plumes capable of melt-through, which seems like maybe another nail in that idea. Of course, if Europa's hydrothermal vent systems were more localized (along the trace of the sub-Jovian point or something like that) maybe a bit more energy would be available, but the salt-induced stratification may trump it anyway...

Lithospheric thermal structure

McKenzie, D, J Jackson, K Priestley, Thermal structure of oceanic and continental lithosphere, EPSL, 233, 337-349, (2005), http://dx.doi.org/10.1016/j.epsl.2005.02.005

McKenzie et al re-examine the problem of what controls the depth of earthquakes in the lithosphere (excluding subducted oceanic lithosphere) primarily by using a more sophisticated thermal model than has been typically used. Until Hofmeister's [1999] paper on the variation of thermal conductivity of mantle materials, no simple model for the variation of thermal conductivity with temperature and pressure really existed. However, with Hofmeister's work, follow-on papers by other workers, and pre-existing data used in the Hofmeister-framework, more realistic thermal models are possible, as was done in this paper. Though the results are certainly reasonable, I was a little surprised at the choice to drop the second term from equation (3) if its effect is about 5 degrees C given some of the other small observational variations that the paper discusses. However, I doubt any major conclusions would be changed.... The choice of a linear interpolation for inital crustal temperature was a surprise since it is pretty much emplaced at the melting temperature; though maybe the intial cooling is more rapid (and hence unresolved) by the time-steps used. Influence would only be near the ridge axis anyway. The inability to accurately reproduce heat flow through old oceanic plates is something that has apparently been known for some time, but is relatively new to me. I suppose there may be a change in the style of convection (small scale) beneath oceanic plates.

Temperature as the primary control on depth of earthquakes is not a surprising idea, but I wonder how much of it has to do with strain-rate?

Classic paper on gravity and lithospheric stress

Sleep, NH and RJ Phillips, Gravity and lithospheric stress on the terrestrial planets with reference to the Tharsis region of Mars, JGR, 90, 4469-4489, (1985), ADSABS Entry.

I recently re-read most of this classic paper. Trying to put together knowledge on gravity, topography, and tectonics in order to understand the state and evolution of planetary lithospheres is a long-standing problem. For small planets like Mars, membrane stresses (often known as fiber stresses to engineers thinking about pressure vessels) are important (not the case on Earth); these and a few other authors like Turcotte, Banerdt and their co-workers laid out these problems for Mars. Tharsis is most likely largely isostatically compensated, and this is an important first step to understanding circum-Tharsis tectonics. The treatment is relatively complete for its purpose, but we have a bit more data today and we worry about more than just isostatically compensated loading of the planet. However, the discussion of calculating gravitational potentials is lucid and useful and membrane stresses instructive.

Sumatra-Andaman Islands Earthquake

Bilham, R, A Flying Start, Then a Slow Slip, Science, 308, 1126-1127, (2005), http://dx.doi.org/10.1126/science.1113363

Lay, T, H Kanamori, CJ Ammon, M Nettles, SN Ward, RC Aster, SL Beck, SL Bilek, MR Brudzinski, R Butler, HR DeShon, G Ekstrom, K Satake, S Sipkin, The Great Sumatra-Andaman Earthquake of 26 December 2004, Science, 308, 1127-1134, (2005), http://dx.doi.org/10.1126/science.1112250

Science published a series of four papers on the Sumatra earthquake in their May 20, 2005 issue. We are reading the articles above in Journal Club this week.... The devastating Dec. 26, 2004 Sumatra earthquake has opened a new box of scientific questions about the earth because of how the earthquake proceeded. These articles indicate that there was no place on Earth that did not move at least a centimeter as a result of the earthquake. Other previous papers indicated that the Earth's rotation even changed, but imperceptively I think. Anyway, there are two main points in these articles that I think are worth noting. First, with the normal plate convergence in this region at ~14 mm/yr, the ~10 meters of slip that occured in seconds is phenomenal at a minimum, and these authors rightfully point out that would normally take ~700 years. So, how much of the total plate movement is due to large slips like this earthquake compared to the smaller, daily-to-yearly earthquakes? Following on from that is the second point: The variation in sliprate along the rupture from south to north. I would guess that variations in stress (and release of it) are the most important controls on this process, but are there others? For example, the total slip in the north according to Lay et al is of similar magnitude as the south (maybe a little smaller), but must have been slow (aseismic) because it was picked up by tiltmeters and GPS, not broadband seismometers. What controls this process? Might it be related to the 30 Myr variation in age of the crust along the rupture too?

Titan's interior and orbital evolution

Tobie, G, O Grasset, JI Lunine, A Mocquet, and C Sotin, Titan's internal structure inferred froma coupled thermal-orbital model, Icarus, 175, 496-502, (2005), http://dx.doi.org/10.1016/j.icarus.2004.12.007

A coupled thermal-orbital model of Titan's internal evolution to try to understand whether the satellite's current eccentricity (0.0292) can place some limits on the existence and composition of subsurface oceans. An interesting result is that they conclude that if Titan has a completely solid interior, it has always been that way, and the same of it has a liquid layer, there would always have been one since the separation of the silicate core. I was also intrigued by the notion that the later convection starts in the ice Ih layer, the smaller the initial eccentricity required to reach the present day eccentricity. I presume this is because the warmer temperatures (lower viscosities) consistent with convection also aid the tidal dissipation in the interior. It also appears that these models do not place much of a constraint on the initial eccentricity and may have some difficulty explaining the 4:3 resonance between Hyperion and Titan.

Martian magmas

Agee, CB and DS Draper, Experimental constraints on the origin of Martian meteorites and the composition of the Martian mantle, EPSL, 224, 415-429, (2004) http://dx.doi.org/10.1016/j.epsl.2004.05.022

Interesting experimental study of potential martian magmas with the aim to understand the super-chondritic CaO/Al203 ratios in shergottites. Basic, result is that apparently no one-stage melting process can create both the high CaO and Al203 concentrations as well as the CaO/Al203 ratios in the shergottites. Instead, the authors propose deep melting around 5 GPa (near their experimental conditions) that due to garnet produce appropriate CaO/Al203 ratios, followed by olivine crystallization at lower pressure to up the concentrations. They suggest that either a mantle-plume or magma ocean scenario might account for such a situation. (Note, none of these processes account for the FeO and Mg#'s using the proposed starting materials). They conclude with comments on studies that try to use the moment-of-inertia to constrain the composition of the planet.

Musings:
(1) I suspect that the pervasive upper mantle melting that I and others like Reese, Solomatov, Stevenson, etc have talked about, could be another mechanism for the two-stage differentiation. Normal upwelling (or plumes) generate deep melting, that proceeds to shallower levels where olivine crystallizes.
(2) Agee and Draper are probably right about the moment-of-inertia arguments. We don't know enough about the core at the moment, especially its composition. I would guess that once the high-pressure phase diagram of Fe-FeS is known better we might be able to place a minumum limit on the sulfur content of the core with thermal models, but it may be higher than current estimates because models (including mine) overestimate the melting pt depression of S for lack of data.