I attended a very interesting conference in Houston in August 2001 where a large group of scientists got together to discuss "Geophysical Detection of Subsurface Water on Mars".

A team from ESA - specifically from the italian aerospace agency - was there to present their radar. They gave an optimistic talk about its capabilities, as one does, and then it was question time. A couple of very capable geophysicists asked them if they had factored-in the obvious observation that the surface dust of Mars and its rocks contain appreciable proportions of magnetic minerals, as seen for example in all the magnetic dust adherence experiments on the Mars landers. These minerals act to absob radio waves (including radar) and limit the penetration of the signal to a couple of km, rather than the 5km claimed by ESA. There were some red faces at this point, since no such calculation or test had been done. Subsequently some tests were run that established that the experts were pretty much right and that not enough signal power was available for deep probing. It was too late to redesign the whole probe, since higher power needs bigger solar panels, bigger batteries, and a better transmitter-receiver setup - all requiring huge mass margins.

The situation was potentially made worse by my own contribution, where I brought together various lines of evidence for temperature gradients on Mars. These include the strength of the crust in response to the load of Olympus Mons and other Tharsis strucures, the preservation of magnetic anomalies beneath 100 km diameter craters in the southern hemisphere, the radiogenic heat available in a planet with different composition than Earth (Mars' bulk density is 3.9, rather than 5.5, for Earth, so they clearly consist of a different mixture of elements and materials), and finally, the thermal conductivity of ice. Ice is an excellent conductor of heat, several times better than rock or water. Although only just emerging at that time, the Odyssey evidence for extensive shallow permafrost means that Mars has an thick conductive shell of icy rock just about everywhere. This ice will not reach melting point until 7km at the equator, and as much as 20 km at the poles, rather than the 1-2 km figure you see quoted by optimistic scientists.

The remaining hope is to find local hot spots where liquid water may be present at shallower depths than normal. However, the failure to find any active volcanoes or hot springs on Mars despite a 5-year search with two orbiting IR cameras suggests that these hot spots will be unlikely.

Thus, in my opinion, the MARSIS radar will never see the hoped-for signal of liquid water. I will be happy to be proved wrong, but as I interpret the evidence I see no easy way out of Mars being a deep-frozen ball of ice, with liquid water confined to great depth at the present day (and for most of geologic time, too).

This does not mean that the MARSIS radar will be useless. It will give us new maps of surface radar reflectivity, which will tell us interesting things about the geology of Mars. In the polar regions it WILL probe deeply, because clean ice is a good medium for radar transmission. We should see layering of the polecaps and perhaps get an estimate of their thickness, if we see a consistent basal reflection from rock-rich material at the base of the ice. There may also be some layering revealed in other parts of Mars, e.g. in crater fill and sedimented areas, or the layering of lavas in volcanic zones.

In short, there will be useful information, but the "Holy Grail" of water on Mars will remain elusive, if not wholly illusory.