I guess I don't see the difference.
If logical corruption (rm -r * in the filepool or tsm screws up and deletes vols), the copy pool could be just as corrupt as the main pool. if the copy pool is in the same dd, then a dd failure would trash both anyway and you're in a DR situation. The snapshot is only over the local dd, so we cut snapshots on both dd's at the same time.
What I meant by logical corruption was more along the lines of maybe an aggregate bitmap somehow getting screwed up (which I've had) etc, and having to restore from a copy volume. A local copy should mitigate that, and you could still have the dd replication for a real DR. Low risk I know, and perhaps I'm being over cautious because I've had it happen to me. Steven