The Early Lunar Shelter: Stay Just a Little Bit Longer

Garrett AiResearch Lunar Shelter

The Garrett Early Lunar Shelter, showing its roots in the LM Truck and, in turn, the LM that actually landed on the Moon. The tanks draped around it are hydrogen and oxygen for the fuel cells, shelter pressurization, and recharging the astronauts’ suits after EVA. Public Domain image via NASA from Early Lunar Shelter Design and Comparison Study, Volume IV. Click for a larger view.

What it was: A two-astronaut shelter/living quarters for use with the Apollo program once it had progressed to needing 30-day stays on the surface, studied in 1966-67 by Garrett AiResearch at NASA’s request. Variants for three astronauts and for a mobile version that could be hitched to a lunar rover were also examined.

Details: Certain big names show up repeatedly in conjunction with the American space program: North American Rockwell, Grumman, Boeing, Lockheed, and so on. Around the fringes, though are less familiar names such as Bendix and TRW. Another one of the latter was Garrett AiResearch, a mid-sized aerospace pioneer best known (at least as far as the space program goes) for designing and building the atmosphere controls for Mercury, Gemini, and Apollo.

In 1966, NASA commissioned Garrett to move beyond what they’d done to that point, and work on a  full-fledged, if tiny, Moon base. Dubbed the “Early Lunar Shelter” (ELS), the intention was to build it following the J-Class missions—what turned out to be Apollo 15 through 17. Having progressed from short surface stays like Apollo 11 to longer ones that had a lunar rover to work with, like Apollo 17, the next step was to be month-long stays and that required more than a single LM.

From the beginning, the Moon landings had been quite restricted in mass, as much of an LM was taken up with astronauts, the consumables they needed, and the fuel and engines needed to get them back to the CSM for the flight home to Earth. If you could forgo all of that with an automated lander, you could haul a lot more equipment to the Moon—to wit, 4.67 tonnes of it.

The fruit of this thinking was the LM/T, or Lunar Module Truck, which was for all intents and purposes a rocket-powered mule that would head to the Moon some time before its associated astronauts would start their journey in their own LM (called the LM Taxi in this context). Landing close to the LM/T, the astronauts could walk over, unload everything, and enjoy a huge quantity of equipment as compared to Aldrin and Armstrong.

There were any number of configurations for the LM/T, constrained only by the volume an LM occupied on top of a Saturn V and the limit to the mass it could safely land, but the Early Lunar Shelter was the answer to one particular question: “Suppose we devote the Truck’s volume entirely to living quarters for two astronauts and the scientific equipment they’d use. What would that be like?”

What Garrett came up with was a stubby cylinder, 8.1 feet in diameter and 16 feet long (2.5 meters by 4.9 meters) which rested on its side above the LM/T’s descent stage, looking not unlike contemporary bathyscaphes. It would be launched atop a Saturn V along with a crewed CSM, which would dock to a hatch on its upper side, then ferry it to the Moon after the usual Saturn IV-B trans-lunar injection. After reaching their destination, the CSM would disengage and return its crew to Earth, while the ELS would land automatically.

The shelter could sit on the Moon for as much as six months before its astronaut-dwellers arrived (thanks to another Saturn V/CSM combination), with a minimum seven days prior to their launch for checkout of the shelter. A SNAP-27 radiothermal generator would power the ELS until activation. Once aboard, the minimum time the astronauts would use it was assumed to be 14 days, with 50 days being the upper end of possibility. The first day days of the mission would be devoted to the astronauts activating the shelter for their use, unloading it, switching the shelter to running off fuel cells (which would also supply water) and transferring the RTG to their LM Taxi so their ride home could be deactivated but kept “alive” until it was needed at the end of the mission.

Interior layout of the Early Lunar Shelter

The interior layout of the ELS, same source as previous. One presumes the outer hatch was closed when the toilet was in use. Click for a larger view.

The interior of the shelter was to be divided into two main areas. One was a lunar EVA airlock taking up one end, the CSM hatch on top being used solely for docking with a CSM. It would have been big enough for two astronauts at the same time as well as storage of two hard space suits. The bulk of the shelter was 628 cubic feet (17.8 m3) of living space. Though about half of this would be taken up with supplies, bunks, and spacesuit storage, its shirt-sleeve environment compared well with a regular LM’s 4.5 cubic meters of habitable volume. Alternatively, as the Moon does supply gravity, the ELS can be sized another way: it would have had 68 square feet of floor space (6.3 square meters).

The arrangement of bunks/radiation refuge quarters in the ELS.

The spartan arrangement of bunks/radiation refuge quarters in the ELS. No Apollo astronaut was taller than 71 inches. Same source as previous. Click for a larger view.

The shelter was double-walled aluminum and fiberglass (the latter in the inside), with 58 mils (0.058 inches, or 0.15 cm) between them for meteoroid protection—the usual tactic, as invented by Fred Whipple. The other major danger entertained was radiation, and the aluminum walls couldn’t be made thick enough to sustain 500 rads (a hypothetical solar flare) without weight close to a half ton more than was otherwise necessary. Accordingly the study suggested putting the necessarily numerous  PLSS recharging canisters (for the life-support backpack worn while on the surface) stored in water filled sleeves around the bunk area located at the opposite end from the airlock. Altogether, they, the walls, and the bunk material made an acceptable, if awfully cramped, radiation refuge for everyone on-board.

One final, intriguing safety touch was the dual-purpose boom attached near the airlock. While primarily intended for unloading instruments or a rover, it would also have been used to get an incapacitated astronaut up next to the entrance to the shelter.

Arranged around and behind the shelter were four tanks: one compressed gaseous oxygen, one liquid oxygen, and two liquid hydrogen. These weren’t intended for use with the Truck’s landing engine—it had its own tankage—but rather for use by the astronauts and the fuel cells (and so, accordingly, their water). Garrett pinpointed the storage of LOX and LH2 for up to six months before the astronauts arrived as the main technical challenge facing the ELS.

Another issue was what atmosphere they would breathe: pure oxygen at 5.0 psia, or nitrogen/oxygen mix comparable to Earth. The former was desirable for mass reasons, and to keep the ELS as close in technology to the rest of the Apollo program as possible, but Garrett were concerned that there were no medical studies of a pure oxygen atmosphere for a long period of time; the 30-day maximum they note was apparently just an educated guess. They ended up punting the problem down the road as essentially an issue of how much they could keep the ELS from leaking; if that could be minimized, the problem was moot. Safety concerns weren’t mentioned at all, and in fact the final filing of Garrett’s study was on February 8th, 1967, not even two weeks after the Apollo 1 fire. After that the CSM would switch to a oxy-nitrogen atmosphere for launch, though the LM would stay with the low-pressure pure oxygen.

Mobile ELS variant, hitched to a notional rover.

The Mobile ELS variant, hitched to a notional rover. Same source as previous. Click for a larger view.

As well as being a shelter, the ELS would have been a miniature scientific outpost. It would be equipped with a drill capable of getting 100 feet down into the Moon’s crust, carry explosive charges for seismic readings, and had three remote instrument stations that would be deployed far from the landing site thanks to the extended EVA capability the shelter would provide. All told, the shelter would come with 3470 pounds (1.57 tonnes) of science gear, while the shelter itself was a remarkably light 985 pounds (447 kg). Add in the expendables and altogether it could be successfully landed on the Moon by the LM/T with a mere pound and a half to spare. Let it not be said that they didn’t squeeze all the juice out of this one.

If the project had gone ahead, Garrett anticipated that the ELS would be operational in 1972. The study is silent on cost, apparently because the construction work was to be handed off to Grumman, and so it was their problem.

What happened to make it fail: It got caught up in the rapid ramping down of the Apollo program that started in 1968, not least the fact that Saturn V production was shut down and the rockets they had were all they were going to get.

By scrimping and saving (and cutting a couple of Moon landings) NASA managed to save Skylab, and eventually the detente-driven Apollo-Soyuz Test Project, but that was it. As any mission involving the Early Lunar Shelter was going to require two Saturn V launches it was an obvious target for a cut, taking up as it would two slots that could be used by two different, separate Moon missions. It was one of the first things to go, and did not make it out of 1968.

What was necessary for it to succeed: It’s interesting to compare the Early Lunar Shelter to the other Moon bases we’ve examined so far, Barmingrad and Project Horizon. Both were hugely ambitious and nowhere near happening in reality, while for this project the key word was early. A lot of people tend to conflate Moon bases with lunar colonies, or at least the next rank down of permanently inhabiting the Moon even if the personnel are swapped out periodically. What NASA put its finger on was that we’re not likely to make that big a leap all in one go. The first lunar bases are probably going to be temporary, just like the first space stations were before we worked our way up to Mir and the ISS.

On that basis it’s easy to get the ELS to fly, as it was a big part of the logical next step in lunar exploration (ignoring the elephant in the room that was automated exploration, mind you). With probably no more than some minor redesigning there could have been one on the Moon just a few years after when Garrett AiResearch pictured it: 1972.

As ugly as the post-1969 picture was for NASA’s funding, it’s not too much of stretch to see the three or four more necessary missions past Apollo 17 making it through the budget grinder and “Apollo ELS” flying sometime around late 1974 or early 1975. It’s a lot likelier than much of what NASA proposed post-Apollo 11, at least, if only because one mission like that would be as much or more of a punctuation mark at the end of the program as any other mission bar Apollo 11 itself.


Early Lunar Shelter Design and Comparison Study, Volume I and Volume IV, W.L. Burriss, N.E. Wood, and M.L. Hamilton. Garrett AiResearch. Los Angeles, California. 1967.

“Barmingrad”: The KBOM Lunar Base

KBOM Lunar Base, "Barmingrad"

A simple schematic of the KBOM lunar base, showing nine of the base module arranged in the proposed figure-8 pattern. Click for a larger view. Based on a blueprint diagram printed in Russia in Space.

What it was: An extensive late-60s/early 70s study of a Soviet lunar base to follow up on the N1-L3 lunar landing.

Details: American lunar base designs, and most Soviet/Russian ones, have generally been quite conservative. They usually consist of upgrades to lunar landers that allow astronauts to stay on the Moon for weeks or months, often with the aid of logistics landers that are more of the same. Detailed study of the construction of something more like a permanent settlement or an Antarctic base is actually quite rare. On the US side we have premature examples like Project Horizon, but in the USSR we had what was probably the most developed of the entire Space Race: Barmingrad.

OKB-1 was swamped with work by the mid-60s, a side effect of Sergei Korolev and Vasili Mishin’s instincts to hold on to as many crewed and automated programs in the aftermath of Vladimir Chelomei‘s grab for control in Khrushchev’s latter days. When in November 1967 the Soviet government launched the Galaktika program to study the exploration of the Moon, Mars, and Venus, they had already informally farmed off study of a lunar base to KBOM, headed by Vladimir Barmin. By March of 1968 this had crystallized into the Columb sub-study and KBOM really set to work developing what was informally dubbed “Barmingrad”.

The choice of KBOM was a bit surprising in that they were the bureau assigned to designing rocket launch facilities for the USSR—the moon base was their first non-terrestrial assignment. Even so, Barmin, his chief A. Chemodurov, and the people assigned to the work took the project with enthusiasm, probably extending far beyond what they were expected to design. Ultimately their work stopped only because the N1-based Moon program was cancelled in 1974.

What they came up with was an ambitious plan based around a multi-use module, which they studied in a variety of configurations before settling on one as the best. The module was 3.5 × 8.5 meters consisting of a rigid section and an expandable section. The expandable section would allow the module to be shipped as roughly a cube and then, once on the Moon, would double the module’s length. At each end as well as on one side of the rigid section was an adapter that  would join two modules together and serve as an access point between them, or allow the attachment of a specialized section, such as the airlock that was to serve as the base’s “front door” for EVA.

Nine of the basic modules would be shipped to the Moon and arranged as two rows of three, with the remaining three serving as “crossbeams”, altogether forming a figure eight. Excepting the aforementioned airlock, this section of the base would be surrounded by berms of regolith and covered with a layer of the same to a depth of 40 centimeters (16 inches), all in the name of radiation protection.

The base was to house 12 cosmonauts, with connections to X-ray and optical telescopes for scientific study, a power source (either a nuclear fission reactor or solar panels), three radiators to dump the base’s waste heat, a unit for cracking oxygen from lunar regolith, and a deep drilling rig. The cosmonauts could get around by walking or, if they needed construction equipment or wanted to travel longer distances, using one of several rovers based around a Lunokhod-like six-wheel chassis. The base would be resupplied by landing craft carrying a logistics module which could be docked to the base, unloaded, and then discarded. By 1974, the base module had reached the mockup stage and KBOM were exploring the ergonomics of their work.

That said, “Barmingrad” took on a life of its own, and KBOM carried on expanding their base design well in to the far future, ultimately using it as the core of a full-fledged Lunar colony with a population of 200, the radical increase of necessary living volume being accommodated by inflatable domes.

What happened to make it fail: When Mishin was replaced as head of TsKBEM (previously OKB-1) in May 1974, Valentin Glushko swept away all of the N1-L3 program in favor of his own ideas. This included a moonbase of his own, LEK, and so Barmingrad was cancelled as part of the coup.

What was necessary for it to succeed: Glushko in turn had his moonbase cancelled along with much of his proposed program about 18 months later, as the Soviet space effort pivoted towards Energia/Buran and space stations. If he’d not cleared the board when taking over from Mishin, that was an 18-month window in which to produce some success with the N1 that might have convinced the Soviet leadership to carry on—and there’s some reason to believe that the success would have come in that timeframe, even if a change in heart is more dubious. At the end of that line was the KBOM moonbase.


Zak, Anatoly. “Going to the Moon…to stay”, Russia in Space: The Past Explained, the Future Explored. Apogee Prime, 2013.

The Reusable Nuclear Shuttle: To the Moon, Again and Again (Integrated Program Plan, Part II)

Sample Nuclear Shuttle configurations

A 1971 slide prepared by Marshall Space Flight Center showing an unloaded Nuclear Shuttle (top) and two configurations with a various components docked to its forward end (middle and bottom). Public domain image by NASA via Wikimedia Commons. Click for a larger view.

What it was: The solution NASA envisioned to the difficulty of getting large payloads to anywhere much beyond Earth with mere chemical rockets. Something like a dozen of them would serve as the brute force “trucks” of the American space program beyond Low Earth Orbit.

Details: We’ve already discussed some aspects of the Integrated Program Plan, NASA’s ambitious 1969 proposal to follow up the Apollo Moon landings with a new goal and new technology. The new goal was a manned Mars Mission, but the new technology had two particular pieces that would do the grunt work of building a space station and a Moon base as intermediate steps to the red planet: a reusable orbiting space plane (not yet dubbed the “Space Shuttle”) and the Reusable Nuclear Shuttle (RNS), many of which would have been built. It would have been the space plane’s role to get astronauts and cargo into low Earth orbit, while the RNS would have been used for the “high frontier”, so to speak. If something was going to go higher a few hundred kilometers, it would be offloaded from the spaceplane to an RNS, and then sent on its way—potentially to the Moon, or even beyond.

The RNS was suited for this task and similarly restricted from landing on Earth for one reason: their engines were given oomph by a nuclear reactor, but approaching one too closely at the wrong angle would expose a person to a fatal dose of radiation.

Start with the Nuclear Shuttle’s advantages. A variety of factors affect the power of a rocket, but the dominant number is the specific impulse (ISP) of the propellants it uses (to be precise, it’s a proportional measure of how much propellant the rocket has to use to add or subtract a given amount of velocity, though confusingly its unit is the second). With variations due to several other factors, rocket engines that use UDMH and N2O4 produce a specific impulse in the neighbourhood of 280 seconds, while LOX/LH2 is much more efficient at around 450 seconds (the low density of liquid hydrogen hamstrings it, though, so it’s often only used in upper stages where the rocket is already well underway and moving fast).

Unfortunately, all chemical fuels with a better ISP than that are either fantastically explosive, corrosive, toxic, or some hellacious combination of all three of those characteristics. Even at that, the best known ISP ever obtained (with a tripropellant of lithium, hydrogen, and fluorine) is 542 seconds.

Ultimately this because chemical propellants depend on chemical bonds, and there’s only so much energy you can contain in those. Quite early on rocket engineers realized that a good way to higher ISP was to use a different source of energy. In the absence of real exotics like nuclear fusion and matter/antimatter reactions, nuclear fission was the way to go. Hydrogen heated by a nuclear reactor can have an arbitrarily high ISP; it’s just a matter of how much heat one can get away with before the physical components of the engine are melted away.

When John F. Kennedy made his famous 1961 speech that started the race to the Moon he made a largely-forgotten reference to the Rover nuclear rocket, a contemporary project that was working on a preliminary nuclear-fission powered rocket. This in turn led to successively more advanced nuclear engines with the colourful names KIWI, Phoebus, and Peewee-1. By the end of 1969, NASA had a design for a functional nuclear rocket engine, the NERVA-2.

NERVA-2 would have had a specific impulse of 825 seconds in vacuum, and be able to burn for 20 minutes and produce 399.5 kilonewtons of thrust. Compare this to the J-2, NASA’s comparable workhorse engine (used on the second stage of the Saturn V, among others): it produced 486.2 kN of thrust, but was far less efficient at just 421 seconds of ISP. Accordingly, even though the NERVA-2 was far larger and heavier than the J-2 (having an entire nuclear reactor on board does that), the savings on propellant mass and the mass of the tanks needed to store it would make any spacecraft using one smaller than the same spacecraft based around a J-2.

Getting to the Moon is considerably more difficult than getting to orbit—you need to add another 3 to 4 kilometers per second to your orbital speed—and so the radically reduced fuel consumption of a NERVA-2 engine was very useful. Enter the Reusable Nuclear Shuttle. This was a conceptually simple ship: a single large fuel tank containing LH2 would have a NERVA-2 attached to one end, while the other had a docking adapter that could connect up to a variety of payload containers. Attach your payload, light the engine, and the RNS would push the payload into high orbit, to the Moon, or even beyond. Ideally you’d also put it on a trajectory which would let it return to Earth orbit, as the NERVA-2 was designed for ten round trips before it would be unsafe to light up again.

The disadvantage of the RNS lay in the radiation environment it produced. The rocket’s exhaust was only marginally radioactive and so arguably acceptable to allow on a launchpad, but in the event of a containment breach on the ground or, worse, in the air the engine would have sprayed uranium all over the environment. Even in the heady days of the late 1960s this was considered too risky, so the plan was to launch an RNS on top of a Saturn rocket using conventional fuels—if the Saturn blew up, the reactors were sufficiently ruggedized that they could survive the accident intact and fall into the ocean safely (by 1960s standards anyway).

What was more problematic was the NERVA-2 in orbit. Once the reactor was up and running it needed a great deal of shielding to protect approaching astronauts. As shielding was heavy, the RNS wasn’t going to have much of it. Instead the approach chosen was the have a “shadow shield”, where the propellant tank and any propellant aboard would provide most of the shielding. This meant that humans getting close to an RNS had to approach it from the front at a fairly shallow angle, using the bulk of the RNS to cover them from the reactor. If they approached from the sides or, God forbid, the aft where the engine was located they were assured of radiation sickness or death. Even on top of the RNS, a crew member would get about the recommended annual maximum radiation dose each time the engine fired.

Nevertheless, the advantages of the RNS outweighed the disadvantages in NASA’s collective mind, and the Integrated Program Plan called for it to be the workhorse of the space program beyond Earth orbit. Each would be used up to ten times (with refueling gingerly taking place after each use), after which it would be discarded in a high orbit due to its extreme residual radioactivity. With it, crews and payloads could be sent to the Moon and returned, and ultimately the American manned Mars mission craft envisioned for the early eighties would be perched on top of three of them.

What happened to make it fail: As with much of the IPP, the nuclear shuttle never got built because of a combination of disinterest from the Nixon administration and the falling budgets that that caused. Of all its parts, only the re-usable Space Shuttle and its rocket stack made it off the ground.

The RNS has its own particular story embedded in this larger tale, though. For many years the nuclear rocket engine program had been championed by New Mexico Senator Clinton P. Anderson, as much of the work on NERVA had been done at Los Alamos. Just as NERVA-2 was ready to become operational he became seriously ill and unable to press his case as much as he had in the past. The White House convinced Congress to pull the plug on the nuclear rocket on the grounds that it would be the basis of a manned mission to Mars, a goal about which Congress was quite negative at the time. The plan was that the freed-up funds could be used for the more-practical Boeing 2707, a Mach 2.7 supersonic commercial passenger plane similar to the Concorde or the Soviet Tu-144. Ironically, Anderson had enough clout remaining in the Senate to apparently engineer a 51-46 vote against moving ahead with that project; the House of Representatives soon followed. While the exact maneuvering involved has never been documented, the vote was widely considered retaliation for the cancellation of NERVA.

Regardless, with its funding quickly dwindling despite Congressional efforts to keep it going, NERVA was cancelled on January 5, 1973, and the Reusable Nuclear Shuttle was dead.

What was necessary for it to succeed: Like much of the Integrated Program Plan, the RNS was doomed by the political currents in Washington, within NASA, and in the general public. When it came down to picking something to move forward on NASA picked the Space Shuttle and the hope that one day they would be able to move on to a space station from there. The RNS ranked third (with the Moon base and Mars landing fourth and fifth) on their priority list, and they even tried very hard to claim that without the Space Shuttle they would not be able to get any nuclear shuttles into space. This was not actually true as the initial plan to use NERVA involved an upgraded Saturn rocket, but it was a measure of NASA’s determination to do anything to get the Space Shuttle built.

Ultimately that’s the main route to getting the RNS into the sky. NASA engaged in a great deal of internal debate from 1968 to 1970 over whether to continue with ballistic capsules or move on to a reusable, winged orbiter. Related to this was the debate over whether or not to focus on Earth orbit as a testing ground or push hard into the rest of the solar system. If both debates had gone the other way, a nuclear engine would have been very attractive to planetary mission planners and the money would have been there to continue with NERVA and the RNS–despite Congress’ objections to Mars missions, the presidential Office of Management and Budget had considerable discretion to ignore how it was told to allocate the money it received until a post-Nixon backlash in 1975.

Instead the arguments settled around a winged orbiter and sticking close into Earth unless the mission was unmanned, and we got the space program that we did from 1975 to the first decade of the 21st century. Nuclear rockets were revived for a short while during the days of the Strategic Defense Initiative’s Project Timberwind, but again it never came to anything.

Even if the RNS got built, there’s the possibility that it would have been much restricted in use or even cancelled outright no matter what successes it scored. The Three Mile Island accident in 1979 soured the American public on nuclear power in general, and after the Challenger explosion in 1986 NASA became very leery about dangerous payloads–for example, deciding against the planned Centaur-B booster that was to be orbited aboard STS-61-G later in the same year for the purposes of getting the Galileo probe to Jupiter. While both were specific incidents, they were each the culmination of long-term cultural trends that likely would have choked off the use of the RNS no later than the mid-1980s, and possibly earlier if one of them was involved in an accident.

FLO: The First Lunar Outpost (Space Exploration Initiative, Part I)

FLO Base Lander

One of two landers for the First Lunar Outpost, this an unmanned one with an adapted Space Station module on top for use as a place to live during the mission. The astronauts would arrive at roughly the same time aboard a manned lander. This picture is somewhat incorrect in that the real lander would have had two large solar panels stretching to the right and left. Public domain image from NASA.

What is was: A 1992 benchmark mission for NASA to return to the Moon, using expendable launchers and a direct descent lander, and build a small periodically-inhabited lunar base on the Mare Smythii.

Details: Early during the presidency of George H. W. Bush, the White House directed NASA via Vice-President Dan Quayle to come up with a plan to go to Mars—a goal announced to the public as the Space Exploration Initiative (SEI). To say that NASA botched this opportunity is to put it mildly.

Their (admittedly non-mandatory) orders were that NASA should come up with a plan that could get to Mars relatively cheaply, probably using technology that had developed since the end of Apollo. Instead they put forward an obvious relative of von Braun’s 1969 Mars Expedition, ditching the nuclear rockets but otherwise following the path of building a big space station, a permanent Moon base, and then finally moving on to Mars. The total cost of the program was estimated to be about US$540 billion over about thirty years—or, to put it another way, a rough doubling of NASA’s annual budget through the next several presidential administrations. This was political suicide and the whole thing collapsed in acrimony almost the moment it was put forward. Richard Truly’s career as NASA administrator came to an end in large part because of the fiasco and he was replaced by Dan Goldin in 1992.

Goldin’s mantra for NASA was famously “faster, better, cheaper” and he arranged for another study that would attempt to recover the manned lunar exploration part of the SEI. It was explicitly to be based on new ideas from the so-called Stafford Report (properly known as America at the Threshold: Report of the Synthesis Group on America’s Space Exploration Initiative) from the previous year. Out of this study grew the First Lunar Outpost (FLO) proposal.

Comet rocket for FLO

An artist’s rendition of the proposed HLV for the mission, referred to informally as Comet. Public Domain image from NASA.

The first step to the outpost was literally getting there. The initial SEI plan had foundered in part because of its allegiance to the Space Shuttle which, as it could lift only 25 tonnes to LEO, meant that NASA needed to build their Moon craft in Earth orbit; that in turn required a space station. FLO was based on a return to an expendable launch vehicle, and it would have been a monster: the core of the launcher would either be an all-new rocket or a stretched version of the Saturn V (to the extent that that program could be revived 20 years after its end); either would have been flanked with two boosters. Its resulting payload to LEO was to have been in the vicinity of 200 tonnes. Contrast that with the Saturn V at 118 tonnes, or 88 tonnes for the Energia.

This massive increase in capability was to be used in two ways: the Moon craft would not have to reconfigure itself in Earth orbit like the Apollo arrangement did, and it would land directly at its destination on the Moon rather than sending down a landing craft followed by a lunar orbit rendezvous before returning to Earth. As well as making the missions safer by allowing more ways to abort and opening up more of the lunar surface for exploration, this simplicity was believed to be the route to a cheaper mission despite the upfront cost of the rocket that launched it.


The FLO spacecraft on top of its TLI stage in Earth orbit. Public domain image from NASA.

Nuclear thermal engines were studied for the trans-lunar injection stage of the FLO spaceship, but it was assumed that it would probably use a J-2S LOX/LH2 engine—essentially the same as was used by the Apollo S-IVB injection stage, though slightly upgraded to use a de Laval nozzle. The lander itself would have used four RL-10s, repurposed from the tried-and-true Centaur.  Again, these choices were made with an eye to saving money by using what the American aerospace industry already had to offer.

The direct-descent/direct-return profile of the actual landing forced the lander to be quite different, though. Admittedly a scaled-up version of the Apollo CM was perched on top of it, where it would carry four astronauts in the relative comfort of 11.3 cubic meters—somewhat larger than the old CM and LM taken together. Below that, though, was a much bigger spacecraft.

It would have packed no less than ten propellant tanks, four smaller ones in an upper tier for the ascent stage, and six larger ones underneath for the lander itself. Sitting on a relatively robust landing truss and four very long legs the whole arrangement would have been 56.7 tonnes with propellant, which is more than four times the mass of the Apollo LM. It would have towered 14.1 meters above the lunar surface, and been 18.8 meters from landing-leg foot to landing-leg foot.

Another big change from the Apollo program was actually a return to what had been planned for the original Moon landings post-Apollo 20. A second, unmanned lander would have been sent prior to the manned one and landed within an easy Moon rover drive, no more than two kilometers. Its entire ascent stage would be swapped out and replaced with a 35-tonne habitation module made in the manner of a Space Station Freedom module with as few changes as possible—again as a nod towards cost.


A sketch of the interior of the FLO habitation module on top of the unmanned lander Note the solar panels. Public domain image from NASA.

This module would have been the actual base. The crew of the manned mission launched in tandem with it would live there for 45 days, exploring the region within 10 kilometers using the aforementioned rover driven by astronauts, and up to 100 kilometers driving it by remote control from the habitat. The explorers would then return home to Earth but the base would not be closed up permanently. Powered by two solar arrays that brought the width of the base craft to just over 41 meters, the intention was that further groups of astronauts could be landed nearby as often as every six months and would find themselves with usable living quarters right away.


Leaving the Moon in the Ascent/TEI stage, leaving behind the landing stage. Public domain image from NASA.

Once the lunar surface mission was over, the astronauts would return to their original landing craft. Its central stack would ignite a hypergolic N202/MMH engine (hydrogen being too tricky to hold on to for 45 days on the lunar surface) and head directly for home. The final twist on the Apollo mission design would have seen the FLO capsule land on dry land, rather than splash down into the ocean.

By sticking as much as possible to technology they already had, or at the very least were already developing, the cost of the project to the end of the first landing mission was estimated at US$25 billion, with the unmanned base touching down around 2000 and the manned follow-up soon after. Just over half of this money would be for the development of the launcher and building three rockets. Even making allowances for the inevitable cost and schedule overruns, it was a remarkably different result from the original SEI.

What happened to make it fail: George H. W. Bush lost the 1992 presidential election and the Clinton administration was noticeably less interested in manned space exploration for its own sake. NASA reoriented itself toward keeping people in LEO, primarily building what had now become the International Space Station, and unmanned space probes beyond Earth orbit.

Extended manned lunar missions did creep back onto the agenda over the next few years, particularly as part of George W. Bush’s “Vision of Space Exploration” which pictured them as a test-bed for an ultimate Mars mission. But the discovery of water ice at the Moon’s south pole by the Clementine satellite in 1994 changed the nature of all future Moon base proposals by slewing them heavily towards using that water. Despite its generally innovative approach to a lunar landing, the First Lunar Outpost turned out to be the last gasp of an older paradigm for exploring the Moon.

What was necessary for it to succeed: This one is more speculative than most, but it’s interesting to consider the First Lunar Outpost in terms of what happened to Space Station Alpha in the same time period. The station came perilously close to cancellation and was only saved by a foreign policy decision: to turn it into the International Space Station, specifically in partnership with Russia in an attempt to absorb the time and skills of the Russian space engineers freed up by the collapse of the Soviet Union.

If you were looking to start an American make-work project in 1993 that capitalized on Russian expertise, a space station made the most sense. After all, Mir was beyond anything the United States had ever accomplished. But it’s not too hard to picture the busy-work being fulfilled by a different major space program. Since a manned Mars mission was out of the question due to expense, the relatively cheap First Lunar Outpost might have been the choice if the Clinton White House had been more interested in the inspirational side of space exploration than its nuts and bolts. They wouldn’t have been the first administration to feel that way.

Project Horizon (Part III): Landing Soldiers on the Moon and Keeping Them There


A depiction of the construction of the second, larger part of the 12-man Moon base proposed as the end goal for Project Horizon. The initial part of the base, shown covered in lunar material at left, was to be built within two months of the first manned landing. The proposed Lunar Landing Craft can be seen soft-landing in the background. Public domain image from Project Horizon: Volume I. Click for a larger view.

What it was: The culmination of the US Army’s Project Horizon proposal of 1959: sending a direct descent/ascent spaceship to the Moon, then building and populating a twelve-man Moon base shortly thereafter.

Details: Having taken off from Christmas Island Launch Facility in the Pacific aboard a Saturn I to the Minimum Orbital Station (MOS), two Army astronauts would receive further fuel launches and then finally an unloaded Moon craft perched on top of a Saturn with a specialized third stage. The third stage has already burned through its fuel to get the heavy direct descent ship into orbit, so after matching orbits with the MOS the Moon crew and the other men living longer-term on the station refuel it. Then the two men bound for points further afield climb aboard and use the stage to burn for their trans-lunar trip.

As well as the TLI stage, the proposed Horizon lunar craft consisted of two more stages. One soft-landed the spaceship on the Moon, and the other would detach from that one (leaving it behind) and return its astronauts to Earth directly. It in turn would separate from a crew return capsule used for re-entry into the atmosphere and splashdown into the ocean.  Altogether this two-stage vehicle would have been some 16 meters long and weighed 64 tonnes. This is huge: the Apollo CSM/LM combination was 45 tonnes, and even at that carried three men instead of two. Even a Saturn V (which was still in its early development during the times of Project Horizon, and is only roughly spec’ed out as a “Saturn II” here) wouldn’t be able to lift that off of Earth, and so the need for refueling in orbit.

To make up for this, there were actually two different types of landers suggested, one of which could be launched directly from Christmas Island on a Saturn I. To meet that requirement, this second type would have been relatively small: 12 tonnes with a payload to the Moon of 2.5 tonnes, a figure made possible only by the fact that they didn’t have to return to the Earth. One would be sent before the first two astronauts started on their journey to the Moon, carrying construction materials for the base. By the end of 1966, there would be four in all sent on their way.

The first manned lunar landing, of two men, would be in April 1965, guided into the site where the base is to be built. In the 1959 report, the Army even had three possible sites picked out: “the northern part of Sinus Aestuum, near Eratosthenes, in the southern part of Sinus Aestuum near the Sinus Medii, and on the southwest coast of the Mare Imbrium, just north of the Appennies”; the last of these is actually not far from where Apollo 15 landed. The Army astronauts’ job would be to explore the immediate area and make sure that the site was acceptable for building a base. They would live in their landing craft until the construction crew arrived in July 1965 (ninety days or so, as compared to Apollo 11’s 21 hours and 34 minutes) at which point they would head home.


A cross-sectional view of the initial two compartment quarters that would house nine to twelve men while they built the larger remainder of the base. Public domain image from Project Horizon, Volume II. Click for a larger view.

This construction crew would consist of nine men, and they would get to work using explosives and tools to dig a deep trench in which they would build their own quarters within fifteen days (or at least no more than thirty) and then cover them with lunar material for protection; a ramp at either end would allow entry and exit from their quarters’ airlock. These accomodations would necessarily be Spartan, but when done the crew would have a small underground base with a cylindrical cross-section (double-walled with vacuum between for insulation), while leftover cargo pods and the like would be buried nearby to hold LOX/nitrogen tanks and waste. The crew would also set up two nuclear reactors to power the base and erect communications equipment so they could stay in permanent contact with Earth.

Now enhanced to twelve men by another landing, the construction crew would get down to building a second, larger cylindrical section at a right angle to the first. When completed the living quarters would be moved here, and it would also contain an office and a sickbay. The original cylinder would be fitted out as two labs, one for biological studies (the proposal charmingly suggests it could be used to check for life on the Moon) and one for physical experiments. The sections of the base wouldn’t link up, but the ramps on one end of each would touch for relatively easy access between them, or as easy as having to put a spacesuit on to walk a few feet can be. A diagram of a remarkably odd-looking spacesuit is included for reference in the report; it has mechanical hands (the astronauts’ real hands were to be cocooned inside the sleeves), and large plates attached to his feet to support him if the lunar dust turned out to be thick.


A diagram of the Horizon spacesuit. Note the odd mechanical hands and “lunar dust walking” plates. Public domain image from Project Horizon: Volume I. Click for a larger view.

By the end of 1968 there would have been ten manned missions in all to the Moon, and eight returns, meaning that the Horizon base would still be inhabited into the indefinite future after that.

What happened to make it fail: The first part of Horizon that we discussed, the launch facility, was perfectly well laid out and not overly different from what ended up being built at Cape Canaveral (though it was bigger and had more pads). The second part is more speculative and is fairly different from what actually happened, but this is mostly because of the change to a Lunar Orbit Rendezvous mission by NASA a couple of years after Horizon was proposed. Still, in retrospect it looks as if they didn’t quite think their station through.

When we get to this third part, though, the speculative nature of what the Army wants to do is front and center. The ways they propose to build and maintain a Moon base are bizarre to modern eyes, mostly because they literally didn’t know the extent of the problems they would face. A close read of the relevant documents reveals a large number of weasel words embedded in every attempt to describe the way things would be done on the Moon if Horizon got the go-ahead. Even Horizon’s summary report admits that the Army wanted another eight months and US$5.4 million dollars just to nail things down before moving on to starting the hardware development for the program.

Having a Moon base was a possible ultimate goal for an American space program, but planning one down to the point of having dates and proposed sites was very premature. Ultimately Project Horizon didn’t fool enough people into thinking that the Army knew what they were doing. Even looking past the previously-discussed antipathy that President Eisenhower had for the military in space, he was known to have used the words “Buck Rogers” more than once to describe the nebulous plans he got from the Army and others, and he was justified in saying so.

What was necessary for it to succeed: The Project Horizon proposal wasn’t actually about how to get to the Moon. It was an attempt by the US Army to establish precedence over the other armed services and, later, the upstart NASA. With Horizon filed away in various Washington bureaucracies, they could point at their long-standing work on manned space travel and plausibly say “Why give money to these newcomers? They’d be starting from scratch and you’d have to pay for that! Give it to us; it’s the wise course to take”.

Then, if anybody bought what they were selling, the way they actually went about it would conform to Horizon only incidentally as they got around to determining how to build this part of their empire. They could always go back and get more money and more time if they needed it, once the US committed to doing it through them.

So Horizon Base was never going to get built. It’s not an appropriate way for housing 12 men on the Moon because when it was designed the proper ways to do it were literally unknown, and would remain unknown for some time. But all it needed to succeed at its actual goal was to fend off Eisenhower long enough for someone in Congress to step up for them and ram through a bill giving the Army control of manned space exploration. It was a decent bet, just one that didn’t pay off.