First of all, a hearty congratulations to NASA's Jet Propulsion Laboratory, for its success at landing a second rover, Opportunity, on the surface of Mars and sending back pictures.
As I wrote below when Spirit landed safely, putting a remote controlled craft on Mars is exceedingly difficult, and it's fantastic (and, no matter how excellent JPL's team, probably a little surprising) that both of them were successful landings. Kudos; and kudos again.

But, after an exciting initial splash and the return of some tantalizing data, the rover Spirit has encountered some problems. On Wednesday the communications signal was lost and all data transfer stopped. This naturally jeopardizes the rest of the mission and I'm sure JPL's engineers are working furiously to correct the problem and to do so as quickly as possible; the solar panels providing power to Spirit are only rated for a three month operating lifetime in the dusty, sandy Martian environment.

This kind of problem, while serious, is not uncommon. The entirely successful Galileo mission to Jupiter (and the Galileo Extended Mission to examine the four largest Jovian moons) was virtually dead on arrival at the giant planet when its high-gain antenna wouldn't unfurl. The initial plan had been to collect data just as fast as the spacecraft's sensors could, then use the high-gain antenna to transmit this data to earth in near real time (the data would pass through a buffer first, with the occasional buffer overrun captured on a backup tape recorder). Switching to the low-gain antenna would be like canceling your cable modem account and going back to a 2400-baud dialup modem, and I actually remember what using those was like.

The solution, which sounds simple but required extensive reprogramming and a significant change to the mission plan, was to capture all collected data directly to the tape recorder. This required the mission team to be much more selective about which data to collect, since the tape recorder's data capacity was not unlimited, and required huge blocks of time between close approaches for transmitting the contents of the tape recorder back to earth via the pokey low-gain antenna. It all worked brilliantly, and as the spacecraft lasted four times longer than anyone had expected, they eventually were able to collect all their data after all.

JPL announced yesterday that they had established some level of communication with Spirit, and had stopped the onboard computer from continuing to spontaneously reboot (probably calling Bill Gates at home on the weekend for tech support, or something similar). I figure since the spacecraft isn't totally non-responsive, it will be a matter of a few days until we have another successful mission-from-the-fire recovery story to talk about.

But all this brings us to President Bush's recently promulgated goal to send men to Mars, and the general skepticism which greeted the announcement. The main points were to (1) return to the moon; (2) establish a permanent presence there (the fabled moon base); (3) complete, then abandon in place, the International Space Station by 2008; and (4) use all the knowledge and technology developed in steps 1-3 to launch a manned expedition to Mars. Let's jump to the end and work our way backwards through these steps.

Why bother going to Mars at all? I wrote previously about a couple of the exciting prospects for life (or fossil remnants thereof) on the red planet. Let's stipulate that there are scientifically compelling reasons to explore the planet, and just consider whether it is better to do this with men or with remote control craft such as Spirit and Opportunity. The two main arguments I hear from the robots-only crowd are that sending human explorers is (1) too dangerous; and/or (2) that it adds complexity and cost to the mission which is not offset by a commensurate gain in expected return of scientific data.

The first of these, that it's too dangerous to risk human lives in this way, can be dispensed with swiftly. The tragedies of Apollo 1, Challenger, and Columbia notwithstanding, fear of loss is a poor reason to prevent pioneers from attempting to explore new terrain or develop new technologies. Lewis and Clark could have been killed by Indians or eaten by bears. The Wright brothers could have accidentally climbed to fifty feet and then flipped into a nosedive in the treacherous Kitty Hawk winds. But they, like modern astronauts, were volunteers for their particular missions, not ordered into harm's way against their will. The pioneer's personal safety is a poor excuse for society collectively to deny him permission to risk his own life in the advancement of human knowledge. The balance between personal risk and the mission's potential rewards is best evaluated by the pioneer himself, who will bear that personal risk. We would today scarcely credit Thomas Jefferson for forbidding the Lewis and Clark expedition, leaving the Pacific Northwest shrouded in continued mystery, for fear of the explorer’s safety.

The second is a bit tricky. There is no doubt that it adds cost and complexity to the mission, as food, water, and oxygen are all heavy and require much more fuel, and hence a much bigger and more expensive spacecraft, to carry them. And, while a robotic mission has only a few critical systems requiring redundant backups, a manned mission has many more critical systems which require triple-redundant backups. What, scientifically, do we gain from this?

A couple of things. Primarily, the mission gains massively in flexibility. A robotic mission has the ability to deeply explore one or two questions in a very narrow territorial area. But if one of those answers comes back with a negative answer, but a tantalizing possibility which begs of answering just a slightly different question, the robotic mission is over. NASA has to design a new spacecraft (2 years), build it (2 more years) and launch it (2 years flight time) before an answer can be had. And if that second answer demands a similar tangential inquiry, the six-year process begins anew.

A manned mission, with its larger, more complex (and, yes, expensive) facilities, could easily be equipped to handle a wide range of expected tangential inquiries by the crew. The results won’t necessarily be any better than would eventually be obtained robotically, but they will be available much faster. A human crew also has the ability to sample a much wider area, allowing qualified conclusions to be established much more decisively than after a more narrow sample (which would otherwise require a second follow up robotic mission six years hence). The extra cost involved greatly accelerates the whole learning process. We’re still, with the current rover missions, trying to answer the same fundamental questions as we were with the Viking missions in the 1970s—whether there is now, or once was, microbial life on Mars. Had the Viking mission been a well-equipped manned laboratory we would have had the answers to today’s questions thirty years ago. And if we land a manned lab on Mars today, we can immediately answer the questions we’d otherwise still be designing narrow-purpose spacecraft to answer in 2040.

The complexity of the spacecraft itself will also be somewhat offset by the improved ability of a human crew to land safely on the surface. The 20-minute lag of communications sent at light speed makes direct remote control of the landing impossible, so robotic craft have all their landing maneuvers preprogrammed. These programmed command sequences are not always successful, when presented with anything the least unexpected (ask the European Space Agency’s Beagle 2 team, or JPL’s Mars Polar Lander team, about this element).

Manned missions can’t replace robotic missions, and they don’t need to. The two are ideal complements to each other. There are places wholly impractical for manned missions (Europa, for example, with Jupiter's intense radiation so nearby). And for detailed global mapping, an unmanned orbiter can't be improved upon by sending humans along. And as a first inquiry, to see whether a planet is even interesting enough to study so intently, a cheap unmanned probe is superior. Both approaches to exploration have their merits and their place.

Now, moving backwards to the first three elements of George Bush’s Mars plan. It’s plain that a great big, Saturn V style of launch craft is probably not the best choice to go to Mars. We haven’t made anything nearly the size of the Saturn V (the Apollo launcher) in 30 years, and the Space Shuttle’s launcher is nowhere near as powerful. The Saturn V itself is not sufficiently powerful, nor does it have enough fuel capacity, to go directly to Mars anyway. And the type of craft, containing something of a laboratory, is far too heavy to launch with any rocket we have now or have ever built.

Payload weight presents a big problem in launches. For every pound of payload weight launched from earth, something like ¾ pound of fuel is required to elevate it to orbit (to say nothing of actually pushing it to Mars; this just escapes earth’s gravity). Rather perversely, this ¾ pound of fuel itself adds to the takeoff weight, requiring ¾ of its weight again in additional fuel (so ¾ of ¾ pound, or 0.5625 pound) just to lift the initial fuel amount, and so forth. While this calculation iterates infinitely (a third iteration requires ¾ of ¾ of ¾, or 0.421875 pound, etc), the resultant fuel increase at each iteration tapers off toward zero pretty quickly. To be all mathematically precise for a moment, this is known as a converging geometric series, so that as the limit of the number of iterations of this approaches infinity, the amount of additional fuel required for each iteration approaches the infinitely small (essentially zero). For a circumstance requiring ¾ pound of additional fuel per pound of payload, this ends up requiring approximately 1/(1-¾)=4 pounds of fuel per extra pound of payload all told. All this is approximate and just to illustrate the math concept involved; the actual numbers vary considerably depending on the precise type of fuel used, all of which have unique values for specific thrust and exhaust velocity and so forth. If you'd like to expand on the particulars for me, an email would be welcome.

So the idea of launching from a moon base makes some sense. The moon’s gravity is about 1/6 that of earth’s; if ¾ pound of fuel is required in the first iteration in the example above to lift one pound of payload to earth orbit, only 1/6 this amount (1/6 x ¾, or 1/8 pound) is required in the first iteration for a launch from the moon, so the total amount of fuel required per pound of payload to reach orbit from the moon is 1/(1-0.125)=1.14 pounds. Instead of requiring four pounds of fuel per pound of payload, only 1.14 pounds of fuel is required to launch a pound of payload from the moon. This eliminates the need for huge external fuel tanks like the shuttle uses, and enables a two-stage rocket (rather than a three-stage), further reducing the cost. And it may be more politically feasible to launch a fission-powered spacecraft (powering the trans-Mars trip, not the launch itself) from the moon than from earth, making the whole mission an even more efficient package.

The only problem with all this is the cost and expense involved in a moon base, and the lack of a really good reason to have one except for the Mars launch. It might be interesting to do it, just to practice our moon launch methodology with modern equipment, since the last time we were there we were still using vacuum tubes and slide rules; there are likely some scientific justifications for such a permanent presence there (if we can justify a permanent south pole base on the scientific merits, why not on the moon?). It just doesn’t seem all that compelling somehow. And part of the reason is the (rightly) much-maligned International Space Station.

I’ve kind of hated the ISS since its inception back in the 1990s, when funding was “found” for it in favor of funding the Superconducting Super Collider, which had the potential to advance fundamental physics knowledge in dramatic ways. The ISS is a money sinkhole, basically an artificially created purpose for NASA and its Russian partners. It’s an excuse to keep the aging shuttle fleet flying when they (let’s be frank) really aren’t doing much anymore apart from the occasional dramatic and valuable repair of a Hubble Telescope that couldn’t be done with cheap, disposable Delta rockets. It’s scientifically pointless. And yet…

The only reason to have it at all that makes any sense is as an assembly and docking point to launch a manned Mars mission. That’s been the one point that makes the whole project forgivable, in my opinion. There just has to be someone on that project team quietly scheming for when he can casually announce his station’s fitness for a jumping off point to Mars. From the ISS the orbit reaching fuel considerations above are manageable, because no matter how big the end product spacecraft is, it can be brought up to space in tiny parts, assembled at the ISS, and then launched on its way with whatever its final stage propulsion system would be, not a high-thrust low-efficiency takeoff stage. It seems a great and a cheap (compared at least to a moon base) solution, and frankly is so close to being ready (compared, again, to this proposed moon base) that I don’t see the point in going to the moon at all. Not, at least, if the main justification for it is to save launch weight for the eventual Mars mission.

I assume George Bush has people much smarter than I am advising him on this. I only hope they like the moon base because it is the technically best solution, not because of its ability to absorb all the money which can be thrown at it. I’m excited about the notion of a Mars mission, but I remain skeptical about this business of the moon base. We shall see.