Quite a lot of opinion built into OCIMF. Radars must be interswitched thereby ensuring interconnective route for total loss - also no requirement for SMS to mandate that the bridge prodecures accounted for any degradation in performance when and if operating interswitched (such as inappropriate PRF and length if master and slave on different ranges). Interswitching was developed for one purpose only - so the technician could quickly narrow down a fault nothing whatsoever to do with the operator and precious little to add redundancy.
Rain performance of a modern 3 cm radar is vastly better than of old, circular polarisation might even allow equality with 10 cm with superior definition at the same time. I had no qualms of retrofitting two X band radars when the regulations were seen on the horizon. SOLAS did NOT mandate S band for performance (although for rain/precipitation we are all aware that it is superior) but because the ITU were considering reallocation of the band to mobile telephones and shore microwave links Politics not physics. Almost universal but not optimum practice is also to run the S band for ARPA. In most conditions (and taking as read the EFFECTIVE length of the aerials) X band definition will allow the ARPA better accuracy.