For radiation sources it's sort of debatable too, because other non-DNA molecules also absorb UV and cosmic rays, so having a bigger DNA target wouldn't necessarily have a meaningful protective effect on other DNA. It's a bigger target now with the extra DNA. It's possible that it would protect against free radicals that actually get into the chromosome however.
That is indeed something that I intended to mention, if I didn't. Some repeating 'junk' base-pair sequences (e.g., for the sake of illustration, ATATATATATAT..., but it's probably not that, I'd have to go a-Googling to get the
exact details) act as 'lightning rods', actually being more attractive to radical ions and taking damage (with or without a handy molecular tool, nearby that spots damaged ATATATATATs and repairs/neutralises it).
Given the likelihood of ATATATATAT coding for anything
otherwise useful, and their tendency to be so 'attacked', that we have such strings in there (and repair functions to maintain/reinstate them) indicates that there is a 'use' which is beneficial to pass down (certainly, on balance, more useful than harmful), and current thinking (or at least as of a few years ago, when I heard about this) is that the lightning-rod function is at least part of it.
And yet, I think, the operational phrase there is "part of it". While multiple genetic expressions are likely used to maintain mission-critical functions, for redundancy's sake, if an organism can derive multiple useful functions from a single region (perhaps overlapping in function with differing alternate mechanisms, elsewhere) combines both resilience and the potential to develop new and useful mutations without losing old ones. Barring the "make use of errors" part, which is rarely good in computer data, consider it equivalent to a RAID0+1 striping-and-mirroring scheme for your server disc-set...
(And, I also forgot to say, DNA wraps up when not in use (unlike enzymes, which are proteins that wrap up in a mysterious way
to be in use), so 'internal' bits of DNA aren't inaccessible, particularly, at least not when it counts. But, if I understand the statement that led to this detail properly, someone is thinking that particular base-combos act as "pre-creased" bits of DNA that help direct the wrapping-up process... But I might be misunderstanding that part. It's new to me, and I might have the wrong end of the wrong stick.)