Dear Abhinav,
It looks like either could be the second side-to-side tower-bending mode. Whether one is actually a torsion mode depends on the magnitude of the twist, which you haven’t plotted.
As I discussed in my Aug 29, 2014 post here, I recommend you set the distributed torsion stiffness high in BModes when creating modes for FAST anyway. In this way, I wouldn’t expect to see any low-frequency torsion modes.
You can use whichever modes you want in FAST, as long as they are representative of the response you are trying to simulate.
Best regards,