Introduction to selecting a reference surface
The surface of any celestial body can be anything but uniform. The oceans, where existing, can be treated as reasonably uniform, but the surface or topography of the land masses can exhibit large vertical variations between mountains and valleys. These variations make it impossible to approximate the shape of the celestial body with any reasonably simple mathematical model.
In case of the Earth, two main reference surfaces have been established to approximate the shape of it. One reference surface is called the geoid, the other reference surface is the ellipsoid. These are illustrated in the figure below:
The body's surface, and two reference surfaces used to approximate it: the geoid, and a reference ellipsoid.
Leaving further clarification to a few short excerpts from Wikipedia pages on geoid and reference ellipsoid, we get these descriptions:
- The geoid is the shape that the surface of the oceans would take under the influence of Earth's gravity and rotation alone, in the
absence of other influences such as winds and tides. All points on
that surface have the same scalar potential—there is no difference in
potential energy between any two.
- In geodesy, a reference ellipsoid is a mathematically-defined surface that approximates the geoid, the truer figure of the Earth, or
other planetary body. Because of their relative simplicity, reference
ellipsoids are used as a preferred surface on which geodetic network
computations are performed and point coordinates such as latitude,
longitude, and elevation are defined.
So we have defined two possible candidates for how to establish solid body reference surface, or as it's sometimes called, particularly when in reference to the mean sea levels, the vertical datum. For establishing the reference surface by using the ellipsoid method, this can be further differentiated to those measured over a local, or global area of the celestial body's surface:
The Geoid, a globally best fitting ellipsoid, and a regionally or locally best fitting ellipsoid, for a chosen region.
OK, so we have established possible methods, possible areas over which they apply, and we've also learned a bit on the terminology used. But how does this translate to actual uses on celestial bodies of our Solar system? This is what I was able to find out;
Zero elevation on Mars
Since Mars has no oceans and hence no 'sea level', it is convenient to
define an arbitrary zero-elevation level or "datum" for mapping the
surface. The datum for Mars is arbitrarily defined in terms of a
constant atmospheric pressure. During the Mariner 9 mission, this was
chosen as 610.5 Pa (6.105 mbar), on the basis that below this pressure
liquid water can never be stable (i.e., the triple point of water is
at this pressure). This value is only 0.6% of the pressure at sea
level on Earth, which forms the zero elevation datum for our planet.
Note that the choice of this value does not mean that liquid water
does exist below this elevation, just that it could were the
temperature to exceed 273.16 K.
Quoted excerpt source and Mars elevation map above: Wikipedia on Geography of Mars
Wikipedia is however a bit outdated and more recently, the data from the Mars Orbiting Laser Altimeter (MOLA) of the NASA's Mars Global Surveyor and a spherical-harmonic representation of an equipotential surface is recommended to be used as the reference for elevations by the Mars geodesy / cartography working group. MOLA transmits infrared laser pulses towards the Martian surface at a rate of 10 times per second, to produce a highly detailed map of the surface. This is an excerpt from Mars geodesy / cartography working group recommendations on Mars cartographic constants and coordinate systems (PDF):
MOLA topographic model will be used for accurate projection of images
and other remote-sensing data onto the planet, and a
spherical-harmonic representation of an equipotential surface will be
used as the reference for elevations.
Mars Orbiting Laser Altimeter (MOLA) Tracks
Zero elevation on the Moon
According to dr. Simon O'Toole from the Australian Astronomical Observatory, a similar instrument to the one employed to map elevation of Mars was used by NASA's Lunar Reconnaissance Orbiter spacecraft to provide a precise topographic model of the Moon.
Because it has no ocean or significant atmosphere, the zero elevation point of the Moon is the average diameter.
Colored global elevation map based on terrain data from the SELENE / Kaguya orbiter, Japan Aerospace Exploration Agency (Source: USGS)
According to The Unified Lunar Control Network 2005 article (PDF) published by USGS (U.S. Department of the Interior U.S. Geological Survey):
This report documents a new general unified lunar control network and
lunar topographic model based on a combination of Clementine images
and a previous network derived from Earth based and Apollo
photographs, and Mariner 10 and Galileo images. This photogrammetric
network solution is the largest planetary control network ever
completed. It includes the determination of the 3-D positions of
272,931 points on the lunar surface and the correction of the camera
angles for 43,866 Clementine images, using 546,126 tie point
Reference surface area of gas giants
With a fair bit of digging around, I have finally stumbled upon one explanation to where the reference surface of gas giants is measured from, according to which we can actually measure depth and heights of clouds and other weather phenomena forming on them, how deep our atmospheric probes have penetrated these planets before they finally gave in and imploded, and similar. In a IAU (International Astronomical Union) / IAG (International Association of Geodesy) working group report on cartographic coordinates and rotational elements from 2006 (PDF), this quote can be found:
The radii and axes of the large gaseous planets, Jupiter, Saturn,
Uranus, and Neptune in Table 4 refer to a one-bar-pressure surface.
The radii given in the tables are not necessarily the appropriate
values to be used in dynamical studies; the radius actually used to
derive a value of J2 (for example) should always be used in
conjunction with it.
The table it refers to doesn't really matter for us, but the definition of the reference surface area does. This means that, by international standards these two bodies (IAU/IAG) represent, the reference surface area is defined by the point at one-bar-pressure of their atmosphere. This same paper gives us another clue:
Reference surface area of minor planets and comets
For irregularly shaped bodies the ellipsoid is obviously useless,
except perhaps for dynamical studies. For very irregular bodies, the
concept of a reference ellipsoid ceases to be useful for most
purposes. For these bodies, topographic shapes are usually represented
by a grid of radii to the surface as a function of planetocentric
latitude and longitude (when possible, or also by a set of vertices
Another problem with small bodies is that two coordinates (i.e.
spherical angular measures) may not uniquely identify a point on the
surface of the body. In other words it is possible to have a line from
the center of the object intersect the surface more than once.
It goes on to describe examples and cartographers' ad hoc tricks for specific maps, and later gives a recommendation that longitudes on minor planets and comets should be measured positively from 0 to 360 degrees using a righthand system from a designated prime meridian, where the origin is the center of mass, to the extent known.
As for the international authoritative bodies defining standards of measure for the reference surface areas of celestial bodies, it appears IAU (International Astronomical Union) and IAG (International Association of Geodesy) are as good as it gets, and who's to argue that?