Partial answer: I can identify the manager responsible for the decision, and the date, but not the reason why.
LC-39 was the sole topic at a meeting of the Launch Operations Working Group on 18-19 July  that brought together 113 representatives from LOD, MSFC, and the launch vehicle contractors: Boeing, North American, Douglas, and General Electric. [...] Following Donald Buchanan's report on the crawler and launcher-umbilical tower, Chester Wasileski briefed the meeting on propellant systems. Although LC-39 would involve no new propellants, loading requirements would dwarf LC-34 operations. Each pad would need storage for approximately 3,407,000 liters of LOX, 946,000 liters of RP-1, 2,460,000 liters of LH2, and 946,000 liters of LN2.
Moonport, ch. 6
Chester T. Wasileski was the manager of the KSC facilities and systems management group. The above quote in turn cites
LOD, "Minutes of the Saturn C-5 Launch Operations Working Group Meeting, 18-19 July 1962," 8 Aug. 1962, pp. 1-5 and app. 9., which I have been unable to find.
The Kennedy Space Center Story describes the size of the tank (p. 39), but does not explain why.
The Apollo Spacecraft: A Chronology, Apollo Program Summary Report, and a search of NTRS yield no useful results.
I agree with Organic Marble that it is desirable to have enough propellant for more than one launch attempt. In addition, it is worth noting that the LOX was the first cryogen loaded into the Apollo-Saturn stack, and thus most susceptible to boiloff.