:) The fossil record is available, and well studied (http://en.wikipedia.org/wiki/Black_Sea_deluge_hypothesis). The evidence suggests that the language actually arose as a result of trade among the inventors of agriculture along the once-freshwater Black Sea, which had been cut off from the ocean, and flooded when the rising Mediterranean overtopped the Bosporus strait. The deluge forced everyone to flee in different directions in a very short period of time, which might be why it's so hard to pinpoint the exact "birthplace" of the language.

In fact, the Turkey hypothesis for the language origin is not inconsistent with the Ukrainian one, if the two populations on either side of the sea were cut off from each other as a result of the flood. Thus, it may be that the real "birthplace" of the Indo-european languages is now underwater.

This theory is well supported by the geologic record, as detailed in "Noah's Flood" by William Ryan and Walter Pitman. Also here: http://en.wikipedia.org/wiki/Black_Sea_deluge_hypothesis

You can easily get around the diffraction limit by using several antennas in an array; this is already done routinely. As for the wavelength, it's probably X-band or K-band, so ~1-4cm. It does seem a little far-fetched to observe objects only 0.5mm across, but water is pretty reflective at those frequencies so it's not impossible -- even though the object size is smaller than the wavelength, it can still have a non-zero scattering cross-section. In this case, lacking further information I'm inclined to trust the NRL.

Yes, averaging and post-processing techniques using multi-frequency GPS receivers can easily achieve millimeter-level precision. For example, check out: http://pbo.unavco.org/instruments/gps