Why is that very likely? The majority of blocks which are currently being mined already take part in merged mining via the likes of RSK.
Work as in (work to solve the problem of mining incentives) i am aware of merged mining, got a dozen of those useless coins, all fees extracted by all the different merged coins throughout history are not close enough to being worth half the block reward, they are essentially useless to bitcoin, and I don't see how that would change.
But the example (RSK) is Bitcoin sidechain though, where you can "convert" between BTC and RBTC whenever you want.
Transactions last 24h
(Number of transactions in blockchain per day) 456,073
Reward (last 24h) 968.75+21.92 BTC ($28,921,512.6)
So 450k tx that need to cough up 28 million, that's 62$ per tx on average!
But what if the BTC price doubles every 4 years to compensate?
As I said above, you're just having the same number of security guards while doubling the area they need to patrol and twice the merchadise!
You read the data wrong though. Total BTC from TX fee is only 21.92
BTC (about $638.639 when i write this post) or $1.4 per TX on average.