For that sort of job you need a SAR. (An image-forming radar, basically). There have been proposals to put a SAR in orbit around Mars but there are 2 challenges: -

1) High power consumption, requiring a large acreage of solar cells, or the dreaded demon of RTG.

2) Large data volumes, requiring improved Mars-to-Earth bandwidth (Or you do the MOC thing of putting a high-res intrument in ornit and only returning 1% of the possible data. If we had a good data pipeline to Mars, we'd have a MOC-resolution map of the whole surface by now!)

One good thing about a radar mission, though, is that you already have a powerful radio transmitter on board. With appropriate configuration, you can use the transmit radar as a signal to Earth.

I suspect that after MARSIS, designs for a dedicated surface-imaging and shallow-penetrating SAR will resurface, with flight about 2013 ("dedicated" means no other instruments on board to waste space and mass, and communication bandwidth, or very little, anyway). The reason for the delay is that so far we know very little about the radar properties of Mars' surface. After MARSIS (and to a lesser extent SHARAD, we will know a lot more and begin making sensible choices of frequency bands and predicting realistic penetration.

As you point out, the Saharan examples where the sand is ultra-dry are a good Mars analogue (ice doesn't count, it's transparent to radar). There's a lot to learn about Mars, and every new instrument/waveband teaches us something new, so long as we've got our eyes open and our brains engaged. I'm sure that when opportunity measures the local rocks and soil, we will all be surprised, me included!