Let's assume the number of individuals that join mining pools grows by a lot. Let's say we have one mining pool with 50,000 members. The pool is successful and receives a block reward. The next step is to send to each pool member their share.
1. Is splitting the current 6.25 reward across 50,000 individuals very costly?
2. Is there a point when splitting the reward across all pool members becomes impractical?
3. Are there ways to improve the cost of distributing block rewards to the pool members?
I understand the lightning network is always helpful when discussing issues about increasing the number of transactions. But my focus with these questions is that it looks to me like pooled mining adds a large overhead of transactions that are not part of peoples' desire to trade or do the things they want with their bitcoin.
I don't see any problem here. I was hobby-mining altcoins with very small hash rate and I can tell that:
* a pool can very well use (internally) more than 8 digits after the decimal point for its own needs if he wants to, for more fair splitting of the reward or if the number of users becomes very big
* a pool doesn't have to send the reward to users after each block found, that would be a waste creating a huge number of inputs in user's wallet and the usual solution is to send only after a certain threshold is passed
All in all, imho you're worrying for not existing problems.
And LN may not be needed because:
* a mining pool can mine its own transaction, hence the tx fee is not a problem
* one tx will probably contain payment for many miners in the same time, it's more practical than a lot of small transactions
* because of that threshold I wrote about the sent amounts aren't that small