Hi Chandan,
Thanks for your question! I think you've pointed out the main issue. The difference in the scale of GRS and CRISM data products is quite significant. For that reason, you'd be comparing large regional areas (GRS) with many orders of magnitude smaller areas (CRISM). If the region was homogeneous, then no problem. However, heterogeneities exist on much smaller scales than GRS. A direct comparison with CRISM is not advisable for this exact reason.
-Scott