Strikes me that I can add direct observation to help illustrate the challenge.

We scan a limited number of academic journals (~150) to compose the @SkepticalScience weekly #ClimateResearch review. It's a tiny sampling of journals but typical flow is about 1,500 articles/week.

Stipulating reasonably scrupulous review and a spectrum of paper content and editorial architecture, producing this output imposes a minimum requirement of something like 3,000 volunteer reviewers needing to commit perhaps 6,000 hours per week.

Overall, ~300,000 hours per year of diligent volunteeer effort to maintain this limited collection of journals.

[And a thing of beauty to behold.]