PAL already uses (has always used) 4:2:0 color sampling. It actually works out to be more desirable than our 4:1:1 (I'll let you look up why, because I don't have time to type it all out - Adam W's 24p.com should spell it out). Granted 4:2:2 is much better, but I don't think it's the achilies heal of HDV (or even a relative concern at that).
Tape dropouts are easily prevented with proper care in which type of lubricant you use in your deck/camera and or a dilligent cleaning schedule if you can't avoid it. Mp2 encoding is an acceptable distribution format, but like you eluded to, when you drop some of the data, you lose a big chunk of footage. DV is intraframe compression, so every frame has all the needed data to reproduce it. When you take a hit on a tape, you might lose a field (one pass of the head) or a whole frame (both heads - or a full rotation of the drum), but the next pass will be the next full field and the pass after that will be the remainder of the frame. This is not the case with HDV. It's interframe compressed and each group of pictures is a reference to the I frame and whatever deviations occurred since it was recorded, and whatever deviations happend to the next frame to the next and the next, etc. So If you lose one stripe, you essentially lose the rest of the GOP to the next I frame. This is perported to be 15 frames long - so now rather than a field or a frame (1/24th of a second or less, depending on shooting mode), you're now missing up to half a second!
Mpeg 1 audio compression is also not much of an issue. MP3 does a sufficient job for broadcast that it's certainly not a problem area to be recording in mp1. If anything, the uncompressed 48K we're used to is overkill.