Why did the number of recovery ships decrease with time?

Question

This NASA page names the various ocean ships that have helped recover NASA spacecraft. There is a general trend that later missions used fewer ships. Read the link for details, but here are some highlights:

• Mercury-Atlas 6 (John Glenn, first orbital flight): 25 ships
• Mercury-Atlas 9 (Gordon Cooper, last Mercury flight): 24 ships
• Gemini 3 (first Gemini flight): 20 ships
• Gemini 12 (last Gemini flight): 11 ships
• Apollo 7 (first manned Apollo flight): 9 ships
• Apollo 11 (first lunar landing): 8 ships
• Apollo 13 (failed mission): 9 ships
• Apollo 17 (last lunar mission): 4 ships
• Skylab 2&3: 3 ships each
• Skylab 4: 2 ships
• Apollo-Soyuz: 1 ship!
• Shuttle: We don't need no stinkin' ships.

Exactly what allowed this reduction in the need for recovery ships? I am looking for a more substantial answer than "they got better at doing it" or "they were more confidant." Which improvements in technologies, procedures, or practices allowed this to happen? Sources, please.

---

Edit: These are recovery ships for landing, not those which may assist launch.

Answer 1

• The Mercury missions really did have the potential to land anywhere around the globe. There was some automation, including an automated retrofire system to initiate re-entry, but it was not yet tested. Glenn and Carpenter had problems with the automatic system and each switched to manual control. Schirra's re-entry was entirely automatic, but he was dismayed at the amount of fuel used. Cooper did his re-entry entirely manually.

• Recovery ships for Mercury were therefore spread out around the globe, in case of any contingency. They were spaced such that the closest ship would have a maximum recovery time of either 3 or 6 hours. Source: Postlaunch Memorandum Report for Mercury-Atlas No. 7

• Gemini capsules had a flight computer that (among other tasks) controlled the re-entry burn. This produced a more precise and controlled re-entry than manual control. Consequently, fewer recovery ships were needed.

The Apollo Program Summary Report specifically states:

The effectiveness of the overall recovery support was maintained even with a trend toward the use of fewer ships and aircraft as the program progressed. The force reductions were based on several factors: a continually increasing confidence in the reliability of the spacecraft and launch vehicles, the availability of a tracking ship (the U.S.N.S. Vanguard) that could serve as a recovery ship during the launch abort phase, the deletion of the requirement for quarantine, and the availability of long-range heavy-lift helicopters late in the program.

In addition to what is noted in the preceding quote,

• Apollo command modules had a flight computer that was derived from Gemini computer, and supported by computers at mission control. The re-entry automation was therefore largely already written, tested, and precise. Even the ill-fated Apollo 13 mission was controlled (a TEI burn and two course corrections) by calculations performed by mission control computers. The improved accuracy decreased the expected landing area.

• Better aircraft were available for later missions. HC-130 planes with a range of 5,178 miles were used for searching, refueling other aircraft mid-air, and the initial airdrop of frogmen and recovery equipment. Mid-air refueling extended the range of the HH-53C heavy-lift helicopters used to recover the crew and capsule to 1000 miles.

• By using aircraft to detect and locate the spacecraft -- rather than just the ship itself -- Apollo recovery ships had a greater search range than Mercury or Gemini ships.

  The AN/ARD-17 sets, developed especially for the Apollo program, were installed in HC-130 aircraft. Two aircraft were generally located approximately 200 miles up range and down range of the predicted landing point and offset from the command module ground track (fig. 7-6). The S-band tracking was used from the end of the communications blackout until approximately 1 minute before the predicted main parachute deployment. (The VHF recovery beacon was activated at that time.) The S-band tracking mode was used to help determine whether the landing would occur up range or down range of the particular aircraft. The set was then switched to the VHF mode to attempt recognition of the recovery beacon signal as soon as the beacon was turned on. Immediate recognition of the recovery beacon signal was desirable because the line-of-sight range was approximately 300 miles when the command module was at an altitude of 10 000 feet compared with a 195-mile range when the command module was on the water.

• As noted at the end of the quote above, Apollo recovery crews began searching for the spacecraft as it was descending, instead of of waiting until it was in the water. This increased the line-of-sight distance to the spacecraft.

• For Apollo, recovery times for each ship were increased to a range of 16 to 32 hours. As each ship covered more area, fewer were needed.

• The number of recovery zones were reduced in number and size, further reducing the number of ships needed. Apollo 7 to 9 had four recovery zones (two Atlantic and two Pacific). Apollo 10 to 17 had two recovery zones, one Atlantic and one Pacific. Skylab and ASTP had one zone each.

• Apollo recovery ships did more moving around than previous programs. After trans-lunar injection, they were re-deployed to the mid-Atlantic and mid-Pacific. Starting with Apollo 14, the post-TLI Atlantic landing zone was eliminated entirely, and "if an Atlantic Ocean landing had been necessary for these missions, recovery of the command module would have been effected by a ship of opportunity." Prior to splashdown, the recovery ships were moved again to the primary recovery zone.

• The Shuttle glided to a runway on land, and did not need recovery ships!